ANALYSIS OF THE IMPACT OF VERTICAL PLUMES AND EXHAUST EFFLUENT ON AVIATION SAFETY

Transcription

1 ANALYSIS OF THE IMPACT OF VERTICAL PLUMES AND EXHAUST EFFLUENT ON AVIATION SAFETY Final Report for the Performed Scientific Analysis 1 October September 2010 Prepared for: The Federal Aviation Administration (FAA) The Airport Obstructions Standards Committee (AOSC) The Office of Airport Safety and Standards Airport Engineering Division (AAS-100) The Flight Procedures Standards Branch (AFS-420) Project Manager: Robert Bonanni, P.E. The AOSC and AAS-100 Phone: Robert.Bonanni@faa.gov Contractor: Science Applications International Corporation (SAIC) Prime Contract Number: DTFAWA-03-P Sub-task Order: 09D FAA AAS-101 Contract Item: Final report Sub-contractor: Scott Mitchell Scott Mitchell Consulting, LLC Phone: (703) Scott.Mitchell@braemarnet.com Principal Investigator: Alexander Praskovsky, PhD Anubis Enterprises, Inc. Phone: (954) apraskovsky@yahoo.com 30 September

2 Executive Summary The aviation community and small airport authorities expressed concerns that new industrial facilities could be built near airports and adversely affect safety of operations. The FAA decided to examine the plume-related issues and the Airport Obstruction Standards Committee (AOSC) initiated this research project to perform a thorough analysis of the impact of vertical plumes and exhaust effluent on aviation safety. The objective of the scientific analysis was to address the key question: Can the vertical plumes induce unacceptable risk level to flying through aircraft and aircrew? If the answer was found negative, no further actions would be required. The affirmative answer to the key question was obtained and the innovative methodology was developed for evaluating the minimum vertical and horizontal clearances which is the major outcome and deliverable of the performed analysis. The AOSC formulated and approved five Specific Project Tasks to be encompassed in the scientific analysis and those tasks were entirely completed. Task 1: Determine the impact to stabilized flight of plume phenomena to include but not be limited to such things as vertical velocity, shear, divergence and curl in different atmospheric conditions and winds All potentially hazardous plume phenomena were analyzed, the dominant impacts on aircraft and aircrew were identified, and preliminary quantitative limits were established for a severity of each identified impact Task 2: Research safety analyses of plume issues including any requirements from the Environmental Protection Agency (EPA) and/or the Occupational Health and Safety Administration (OSHA) No EPA or OSHA requirement was found for the exposure time of 20 sec or less which is the maximum expected duration of the aircraft travel through exhaust effluent Task 3: Examine the potential impact to both aircraft and aircrews to repeated exposure of flying through plume effluent (CO x,.no x, SO x, NH 3, and other potentially harmful gases, as well as particulate matter such as ash and soot) No adverse impacts of plume effluent neither on a health of aircrew or passengers nor on the aircraft performance were found to be expected Task 4: Examine the obscuration effects of plume induced condensation clouds. It was found that the plume induced condensation clouds do not affect aviation safety Task 5: Complete a detailed safety analysis related to the above plume issues which includes recommended requirements for minimum vertical and horizontal clearances to avoid negative effects of plume turbulence, condensation, and harmful effluents. The minimum vertical and horizontal clearances for avoiding detrimental plume effects were recommended to be equal to respectively the height and width of a plume-induced region with the unacceptable risk level for flying through aircraft An innovative methodology was developed for mapping the plume-induced risk and evaluating the height and width of the region with the unacceptable risk level 2

3 The methodology was used to evaluate the clearances for exhaust stacks of three power plants for Cessna 172 aircraft on final approach. Preliminary estimates for the minimum vertical and horizontal clearances are respectively 1,810 ft AGL and 330 ft for the operating Fort Martin Power Station near the Morgantown airport, WV; 1,240 ft AGL and 320 ft for the proposed Towantic Energy plant near Waterbury Oxford airport, CT; and 1,120 ft AGL and 230 ft for the proposed Mariposa Energy plant near the Byron airport, CA The following are the major conclusions of the performed study. Vertical plumes can generate an unacceptable risk level for an aircraft and aircrew over spatial regions with significant dimensions at rather typical conditions The minimum vertical and horizontal clearances for avoiding harmful plume effects are recommended to be equal to the height and width of the region with the unacceptable risk The developed innovative methodology for quantitative mapping of the plume-induced risk level provides physically substantiated, consistent and adequate guidance for evaluating the minimum vertical and horizontal clearances Dimensions of the plume-induced region with an unacceptable risk for Cessna 172 aircraft on the final approach and thus the minimum vertical and horizontal clearances can exceed respectively 1,300 ft AGL and 300 ft at typical conditions The plume-induced turbulence is the dominant cause of the detrimental plume impact on aircraft and aircrew 3

4 Table of Contents 1. Introduction Analysis Results Guidelines for Evaluating the Plume Impact on Aviation Safety Innovative Methodology for Evaluating the Minimum Clearances Identification of Potentially Hazardous Plume Impacts Quantitative Evaluation of the Impact Severity and Likelihood Plume Modeling Mapping the Plume-Induced Risk and Estimating the Clearances Conclusions 22 References 23 Appendix A: Glossary of Terms.. 27 Appendix B: Reports on the Plume-Related Incidents Appendix C: Existing Approaches to the Problem...32 Appendix D: Aerodynamic Criteria for Potentially Hazardous Air Disturbances.. 36 Appendix E: Limited Expert Analysis E-1: Frequency of stack overflight E-2: Definition of upset for light GA airplanes.. 40 E-3: Aircraft tolerance to vertical gusts.. 41 E-4: Impact of high temperature and oxygen-depleted exhaust on airplane.. 42 E-5: Airplane types, configurations, and flight conditions for modeling 42 E-6: Severity of selected physical impacts on airplane and aircrew 43 E-7: Quantitative characterization of other physical impacts.. 44 E-8: Pilot response time E-9: Obscuration effect of plume induced condensation clouds. 45 Appendix F: Safety Effects of Exhaust Effluent. 47 Appendix G: Modeling the Aircraft Dynamics Appendix H: Modeling the Plume-Induced Aerodynamic Field. 52 Appendix I: Evaluating the Plume-Induced Risk

5 List of Figures Figure 1: Overview of the performed study Figure 2: Overview of an innovative methodology for evaluating the minimum clearances. 12 Figure 3: Spatial distributions of the mean vertical velocity and temperature Figure 4: Schematic of vertical turbulent gusts and their impact on aircraft.. 17 Figure 5: Severity of the vertical acceleration for Cessna 172 at ambient wind 1 knot.. 19 Figure 6: Risk level for Cessna 172 at ambient wind 1 knot Figure 7: Map of a combined risk level with recommended clearances.. 20 Figure G1: Vertical acceleration of Cessna 172 along discrete vertical gusts. 51 Figure G2: The maximum vertical acceleration for Cessna Figure H1: The amplitude of the plume-induced turbulent vertical gusts Figure H2: Spatial distributions of the turbulent integral scale Figure H3: Flow characteristics at the plume centerline Figure H4: Spatial distributions of the external intermittency factor.. 57 Figure I1: Severity of the airframe-damaging vertical gusts at variable winds.. 59 Figure I2: Severity of the Cessna 172 vertical acceleration at variable winds 59 Figure I3: Hourly temperatures versus wind speeds 60 Figure I4: Likelihood of the plume-induced impacts at variable ambient wind speed 60 Figure I5: Risk level for the vertical acceleration of Cessna 172 at variable winds 61 Figure I6: Composite map of the plume-induced risk for Cessna List of Tables Table 1: Quantitative limits for severity of plume-induced physical impacts. 14 Table 2: Stack parameters and exhaust characteristics for considered plumes Table 3: The risk matrix from the FAA SMS Manual. 18 Table 4: Likelihood definitions reproduced from the FAA SMS Manual Table D1: The ICAO turbulence index 38 Table D2: Quantitative measures of turbulence intensity Table E1: Qualitative classification for the impact severity 43 Table E2: The expert-provided magnitude of potentially hazardous impacts. 44 Table E3: The expert-provided estimates for the pilot reaction time.. 45 Table E4: Limitations on flying under VFR at low visibility.. 46 Table F1: The NAAQS concentration limits for criteria pollutants

6 1. Introduction The impact of vertical plumes and exhaust effluent on aviation safety has been under attention of the Federal Aviation Administration (FAA) for a long time. In 2005 the Director of Flight Standards Service (AFS-1) tasked the Flight Procedures Standards Branch (AFS-420) of the Flight Technologies and Procedures Division (AFS-400) to perform a risk analysis of flights over vertical plumes. AFS-420 organized and led a safety risk analysis team consisting of the FAA experts and civilian contract personnel to examine the issue (see report AFS-420, 2006). Since the AFS-420 (2006) study, new data on potentially hazardous impact of exhaust plumes on aviation safety were brought to the FAA attention (e.g., CEC 2007, Pietrorazio 2010, and pilot and incident reports in Appendix B). The general aviation (GA) community and small airport authorities expressed concerns with new industrial facilities that could be built in the vicinity of airports and adversely affect safety of operations. Another issue brought to the FAA attention is the difference in defining an obstruction for protecting the navigable airspace in the US and abroad. The FAA Title 14 of the Code of Federal Regulations, Part 77 considers only physical dimensions of the obstruction 1 such as the height of a smoke stack. The Australian Civil Aviation Safety Regulations (CASR) Part adds the horizontal and vertical limits of gaseous efflux 2 to the height. Interpretation of the plume exhaust as a potential obstruction led to involving into the problem another FAA line of business, the Office of Airport Safety and Standards Airport Engineering Division (AAS-100). Figure 1: Overview of the performed study With the new data to consider, the FAA decided to re-examine the plume-related issues. The problem was delegated to the FAA Airport Obstruction Standards Committee (AOSC) that 1 The paragraph Standards for determining obstructions in the FAR Part 77 Objects Affecting Navigable Airspace, states: An existing object, including a mobile object, is, and a future object would be, an obstruction to air navigation if it is of greater height than any of the following heights or surfaces 2 The Australian Civil Aviation Safety Regulations (CASR) Part 139 Aerodrome Certification and Operation, Subpart 139.E Obstacles and Hazards, paragraph Monitoring of Airspace, defines an obstacle as any object, building, or structure; or any gaseous efflux having a velocity exceeding 4.3 meters per second. 6

7 works across all FAA lines of business. The AOSC initiated this research project to perform a thorough analysis of the impact of vertical plumes and exhaust effluent on aviation safety. The objective of the 12-months scientific analysis was to address the key question: Can vertical plumes create unacceptable risk levels to aircraft and aircrew? If the answer to the key question was found negative, no further actions would be required; Figure 1. The performed analysis provided the affirmative answer and the innovative methodology was developed for evaluating the minimum vertical and horizontal clearances to avoid negative plume effects. The preliminary estimates revealed that plumes can induce sufficiently large spatial regions with unacceptable risk requiring minimum vertical and horizontal clearances that can exceed 1,300 ft AGL and 300 ft at typical conditions. The AOSC formulated and approved five Specific Project Tasks to be encompassed during the scientific analysis and the tasks were entirely completed. Task 1: Determine the impact to stabilized flight of plume phenomena to include but not be limited to such things as vertical velocity, shear, divergence and curl in different atmospheric conditions and winds. Available incident and pilot reports were analyzed and it was found that the vertical plumes can generate spatial regions of unacceptable risk level to flying through aircraft and aircrew All potentially hazardous plume phenomena were analyzed and the dominant impacts on aircraft and aircrew were identified as the airframe-damaging vertical gusts, wing stall, abrupt changes in the vertical acceleration and aircraft pitch and bank angles, and abrupt ascend/ descend Preliminary quantitative limits were established for a severity of each identified impact Task 2: Research safety analyses of plume issues including any requirements from the Environmental Protection Agency (EPA) and/or the Occupational Health and Safety Administration (OSHA) No EPA or OSHA requirement was found for the exposure time of 20 sec or less which is the maximum expected duration of the aircraft travel through exhaust effluent Task 3: Examine the potential impact to both aircraft and aircrews to repeated exposure of flying through plume effluent (CO x,.no x, SO x, NH 3, and other potentially harmful gases, as well as particulate matter such as ash and soot). No adverse impacts of plume effluent neither on a health of aircrew or passengers nor on the aircraft performance were found to be expected due to a short exposure time and a fast dilution rate of the air contaminants Task 4: Examine the obscuration effects of plume induced condensation clouds. It was found that the plume induced condensation clouds do not affect aviation safety 7

8 Task 5: Complete a detailed safety analysis related to the above plume issues which includes recommended requirements for minimum vertical and horizontal clearances to avoid negative effects of plume turbulence, condensation, and harmful effluents (and identify current problem locations with suggested mitigations for each of those locations) 3. The minimum vertical and horizontal clearances for avoiding detrimental plume effects were recommended to be equal to respectively the height and width of a plume-induced region with the unacceptable risk level for flying through aircraft. A conceptually new methodology was developed for assessing quantitatively the plumeinduced risk and recommending the minimum vertical and horizontal clearances for avoiding harmful plume effects. The essence of the new concept is to evaluate spatial distribution (create a map) of the plume-induced risk level to an aircraft and aircrew and utilize the map for recommending the minimum clearances. Following the FAA Safety Risk Management (SRM) process, the risk was assessed as the composite of predicted severity and likelihood of the potentially hazardous plume impacts in the worst credible system state. The risk level and therefore recommended clearances were linked quantitatively to the stack location and dimensions, exhaust parameters, atmospheric conditions, and type and configuration of an aircraft. The developed innovative methodology is physically substantiated, consistent, adequate for the task, and includes: o An adequate, computationally efficient and numerically stable integral plume model that does not require highly skilled personnel for using it, o The newly developed comprehensive modeling of plume-induced turbulence, o Expert-identified potentially hazardous plume impacts, and o Expert-evaluated quantitative limits for the impact s severity The methodology was used to evaluate the clearances for exhaust stacks of three power plants for Cessna 172 aircraft on the final approach. Preliminary estimates for the minimum vertical and horizontal clearances are respectively 1,810 ft AGL and 330 ft for the operating Fort Martin Power Station near the Morgantown airport, WV; 1,240 ft AGL and 320 ft for the proposed Towantic Energy plant near Waterbury Oxford airport, CT; and 1,120 ft AGL and 230 ft for the proposed Mariposa Energy plant near the Byron airport, CA. The major result of the executed analysis is the affirmative answer to the key question: vertical plumes can generate an unacceptable risk level for aircraft and aircrew over regions with significant spatial dimensions. In accordance with the result, the requirements were recommended for the minimum clearances near exhaust stacks. It was presumed that the minimum vertical and horizontal clearances should ensure operational safety of flying public and were recommended to be equal respectively to the height and width of the region with the unacceptable risk level. The developed methodology uses the SRM process and utilizes the expert-identified hazardous impacts and quantitative limits for the impact severity for producing 3 The bracketed sub-task was eliminated by the AOSC from the Initial Analysis Plan on 17 December

9 the required height and width of such region. The methodology provides guidance for recommending the minimum clearances to avoid harmful plume impact and it is the major outcome and deliverable of the performed scientific analysis. It should be emphasized that some elements of the methodology utilize limited experimental and pragmatic data thus the methodology in its current state provides preliminary quantitative estimates for the minimum vertical and horizontal clearances. The major conclusions of the performed scientific analysis are the following. Vertical plumes can generate an unacceptable risk level for an aircraft and aircrew over spatial regions with significant dimensions at rather typical conditions The minimum vertical and horizontal clearances for avoiding harmful plume effects are recommended to be equal to the height and width of the region with the unacceptable risk The developed innovative methodology for quantitative mapping of the plume-induced risk level provides physically substantiated, consistent and adequate guidance for evaluating the minimum vertical and horizontal clearances Dimensions of the plume-induced region with an unacceptable risk for Cessna 172 aircraft on final approach and thus recommended minimum vertical and horizontal clearances can exceed respectively 1,300 ft AGL and 300 ft at typical conditions The plume-induced turbulence is the dominant cause of the detrimental plume impact on aircraft and aircrew 9

10 2. Analysis Results This section outlines major results of the performed scientific analysis on the impact of vertical plumes and exhaust effluent on aviation safety. The section is mainly focused on the key accomplishment of the analysis: an innovative methodology for quantitative evaluation of the plume-induced risk level to flying through aircraft (mapping the risk). The risk map provides guidance for recommending vertical and horizontal clearances to avoid harmful plume effects. The necessity of developing the new methodology is demonstrated and all steps in the methodology are outlined. Technical details are delegated to the proper appendices. 2.1 Guidelines for Evaluating the Plume Impact on Aviation Safety The major objective of the performed scientific analysis of the plume impact on aviation safety was to address the key question: Can vertical plumes create unacceptable risk levels to aircraft and aircrew? If the answer was found negative, no further actions would be required. The study began from analyzing available incident reports and the major findings of the analysis were as follows; Appendix B. The plume-related incidents are quite frequent in the airports that have exhaust stacks underneath of, or very close to typically used arrival/ departure flight routes. A majority of plume-related incidents remains unreported unless the incident creates quite a serious problem like severe turbulence or a pilot is asked to provide a report. Quite often pilots do not identify a plume as the cause for air turbulence because they do not see the stack or do not attribute turbulence to the exhaust plume. Invisible plumes are the most hazardous because the impact on aircraft is unexpected and takes the pilot-in-command (PIC) by surprise. The physical impact of exhaust plumes on aircraft is especially hazardous for inexperienced pilots. Calm winds and low air temperature intensify the impact of exhaust plumes on aircraft. The aircraft is exposed to the plume-induced high air temperature and/or oxygen-depleted air for too short time for affecting noticeably the engine performance at typical conditions. Impact of exhaust effluent on aircrew is negligibly small due to a short time of exposure and low concentration of hazardous gases away from the orifice at typical conditions. Although the incident reports indicated that a plume can be hazardous, they did not provide a spatial extent of the plume-induced regions with an unacceptable risk level for flying aircraft. To quantify the plume-induced risk level, an adequate approach to quantifying the impact of plumes on aviation safety was needed. The existing approaches to the problem such as statistical analysis of available databases in AFS-420 (2006) and the unique threshold for the vertical gust in the Australian Civil Aviation Safety Authority (CASA) Advisory Circular (AC) (0) 4, are considered in Appendix C. Conventional criteria for quantifying the impact of air disturbances on aircraft are discussed in Appendix D and include the National Aeronautics and 4 The CASA AC (0) Guidelines for Plume Rise Assessment 10

11 Space Administration (NASA) metric for windshear, the International Civil Aviation Organization (ICAO) turbulence metric, and often referred to measures of turbulence intensity from the Jeppesen Sanderson Training Products. It was found that intrinsic limitations of the approaches and criteria prevent their application to estimating the risk of plume overflight. The absence of an adequate approach and quantitative criteria to analyzing the impact of plumes on aviation safety forced the authors to develop an innovative methodology for addressing the problem. A conceptually new methodology was developed for assessing the plume-induced risk and recommending the minimum vertical and horizontal clearances to avoid harmful plume effects. The essence of the new concept is to evaluate quantitatively spatial distribution (create a map) of the plume-induced risk level to an aircraft and aircrew and utilize the map for recommending the minimum clearances. As noted in AFS-420 (2006), the FAA Safety Risk Management (SRM) procedural process contained in the FAA Air Traffic Organization Safety Management System (SMS) Manual provides generic and flexible guidance that suits for determining aviation risk for the problem in hand. The following steps of the process were applied for performing the present analysis: Identification of potential hazards, Risk analysis, Risk assessment, and Treatment (mitigation) of the risk, if required 2.2 Innovative Methodology for Evaluating the Minimum Clearances After the key question was answered affirmatively, the Project Task 5 has become the crucial element of this research: recommend requirements for the minimum vertical and horizontal clearances to avoid negative effects of plume turbulence, condensation, and harmful effluents. The only effective way of recommending rational requirements is to relate the clearances for a specific plume to spatial distribution of the plume-induced risk level for flying through aircraft and aircrew. Therefore getting the distribution (mapping the risk level) became the major goal of the analysis. An innovative methodology was developed for evaluating quantitatively the impact of plumes on aviation safety and thus recommending the minimum clearances. The methodology is the major deliverable of the performed study and it is illustrated in Figure 2. Following the FAA SRM process, the risk was assessed as the composite of predicted severity and likelihood of the potentially hazardous plume impacts in the worst credible system state; Section 2.6. Quantitative measures for severity and likelihood of the impacts are not specified in the government regulations thus they needed to be established. A conceptually new approach was developed for quantifying the impact severity and likelihood and utilized in the methodology; Section 2.4. Another unresolved issue of crucial importance was identification of the potentially hazardous plume impacts on aircraft and aircrew. Although the impacts to be examined were outlined in general terms in Specific Project Tasks 1 4, neither their relevance to nor relative severity for exhaust plumes are specified in the government regulation or scientific literature. All potentially hazardous plume phenomena were analyzed and the dominant impacts on aircraft and aircrew were identified as the airframe-damaging vertical gusts, wing stall, abrupt changes in the vertical 11

12 acceleration and aircraft pitch and bank angles, and abrupt ascend/ descend; Section 2.3. Magnitudes of these impacts were related to characteristics of the plume-induced aerodynamic field using modeling of aircraft dynamics; Section 2.4. Figure 2: Overview of an innovative methodology for evaluating the minimum clearances One can see in Figure 2 that the developed innovative methodology exploits exhaustively an extensive knowledge and experience of the subject matter experts for obtaining crucial data that are absent in the government regulations. The expert analysis was chosen as the most efficient, fast and inexpensive tool for filling out the gaps in government regulations and scientific literature. Ambiguous classifications, non-established quantitative limits, and concepts allowing different interpretations are typical issues that could be resolved only by utilizing empirical wisdom of the highly skilled specialists in the field. A limited expert analysis was executed during the performed study where seven highly experienced aviation safety experts, light aircraft engineers and general aviation pilots have been interviewed; Appendix E.. The major outcome of the interviews could be summarized as follows. The most hazardous physical impacts of exhaust plumes on aircraft were identified and their severity preliminary quantified, It was recommended to consider the impact likelihood to be equal to the probability of its occurrence, High temperature and oxygen-depleted air in exhaust plumes have been identified as nonhazardous to an aircraft and aircrew due to a short exposure time, and Dynamics of aircraft flying through turbulent gusts was recommended to be initially analyzed for Cessna 172 airplane on the final approach 12

13 The expert-identified hazards and aircraft modeling results were used as requirements for developing a model for the plume-induced aerodynamic field. An adequate model should provide characteristics of air disturbances that are needed for quantifying plume impacts on aircraft and aircrew. For the impacts identified in Section 2.3 and related to aerodynamic characteristics in Section 2.4, the minimum set of plume characteristics to be modeled includes the mean and turbulent components of the vertical velocity, the turbulent integral scale of the fluctuations, and the external intermittency factor. None of the existing models provides the necessary parameters of turbulence. A new model has been developed and it is described in Section 2.5. Figure 2 illustrates that the developed innovative methodology provides guidance for evaluating the minimum vertical and horizontal clearances that are linked quantitatively to the stack location and dimensions, parameters of exhaust, atmospheric conditions, and type and configuration of an aircraft. The methodology is physically substantiated, consistent, adequate for the task, and includes: An adequate, computationally efficient and numerically stable integral plume model that does not require highly skilled personnel for using it, The newly developed comprehensive modeling of plume-induced turbulence, Expert-identified potentially hazardous plume impacts, and Expert-evaluated quantitative limits for the impact s severity 2.3 Identification of Potentially Hazardous Plume Impacts Potentially hazardous plume impacts on flying through aircraft and aircrew were identified in quite generic terms in Specific Project Tasks 1 4. Tasks 2 and 3 require examining the health impact of exhaust effluent on aircrew and passengers of aircraft traveling through a plume. Requirements from the Environmental Protection Agency (EPA) and/or the Occupational Health and Safety Administration (OSHA) shall be considered in the process. The examination is described in Appendix F and the major findings are as follows. No EPA or OSHA requirement was found for the exposure time of 20 sec or less which is the maximum expected duration of aircraft travel through exhaust effluent The exposure time of aircraft, aircrew and passengers to exhaust effluent is too short for any noticeable effects on the human health or aircraft performance Concentration of the air contaminants decreases with distance from the orifice much faster than the amplitude of vertical gusts thus the expected contamination impacts are much weaker than the dynamic ones Effect of exhaust effluent from smoke plumes on a health of aircrew and passengers is expected much lower than that of a contaminated air in crowded airports The conclusion of the examination in Appendix F is that no adverse impacts of plume effluent neither on a health of aircrew or passengers nor on the aircraft performance are to be expected. Task 4 indicates another potential hazard: the obscuration effects of plume induced condensation clouds. The hazard was examined in Appendix E-9 where it was concluded that the plumeinduced clouds do not affect aviation safety. The reasoning underneath the conclusion was that reduced visibility does not affect airplane operations under the Instrument Flight Rules. Flying at 13

14 low visibility under the Visual Flight Rules (VFR) is strictly regulated by the FAA CFR, Title 14, Basic VFR weather minimums. It was presumed that a pilot obeys the regulation hence he/ she cannot be endangered by the plume condensation clouds although indirect negative effects may include restricted traffic patterns and cloud-avoidance maneuvers. Results of the AFS-420 (2006) study and incident reports in Appendix B led to a natural presumption that the most potentially hazardous plume impacts on aircraft and aircrew are those outlined in Task 1. All possible physical plume impacts were considered in the performed study including but not limited to upset, engine stall due to high temperature and oxygen-depleted air and the erroneous pilot actions that could result in incident/ accident. The identification was performed using the only effective and reliable approach: expert analysis. The most hazardous plume impacts on aircraft and aircrew were identified in Appendices E-2, E-3, E-4, E-9, E-6 and E-7 as the airframe-damaging vertical gusts, wing stall, abrupt changes in the vertical acceleration and aircraft pitch and bank angles, and abrupt ascend/ descend. Only those impacts are included in the developed methodology for mapping the risk of plume overflight; Figure Quantitative Evaluation of the Impact Severity and Likelihood The FAA SMS Manual provides only qualitative classification for the impact severity. To implement the identified in Section 2.3 potentially hazardous plume impacts on aircraft and aircrew into the methodology in Figure 2, quantitative limits for the impact severity had to be established. This problem was resolved with the expert analysis and the details are given in Appendices E-6 and E-7. The final outcome is summarized in Table 1. It is noteworthy that quantitative limits in Table 1 are not specified in any official document. Another noteworthy feature is that the limits depend on aircraft type and flight conditions. As indicated in Appendix E-7, this table is based on interviews of five experienced pilots and the magnitudes are just preliminary estimates. It means that the magnitudes for the minimum vertical and horizontal clearances obtained in the performed study are preliminary estimates as well. Severity Impact Airframedamaging vertical gusts Wing stall Plunge, feet Ascend, feet Change in the pitch angle, degrees Change in the bank angle, degrees Amplitude of abrupt changes in the vertical acceleration Minor g Major g Hazardous X X g Table 1: Quantitative limits for severity of plume-induced physical impacts for Cessna 172 aircraft on the final approach To utilize the limits in Table 1, one should relate the magnitude of each impact to parameters of the plume-induced aerodynamic field. This task was solved by modeling dynamics of aircraft flying through air disturbances with variable parameters. The modeling is described and illustrated for Cessna 172 aircraft on the final approach in Appendix G. It was concluded that the impact of discrete vertical gusts on aircraft depends on the gust s amplitude and width, and on 14

15 aircraft type and flight conditions. The conclusion indicates that the representative plume modeling must provide the amplitude and the width of vertical gusts. The experts suggested considering the impact likelihood as the total probability of its occurrence; Appendix E-1. To utilize dynamics of an aircraft flying through discrete vertical gust of specified amplitude and width, the total probability of occurrence of such gust should be evaluated and it includes four independent components. First, each run of plume modeling is executed at specific stack parameters, exhaust characteristics and ambient atmospheric conditions; see Figure 2 and Section 2.5. The stack parameters such as the height and orifice diameter are typically constant for a specific power plant. The exhaust characteristics such as the effluent temperature and velocity depend on the power plant operational regime and are typically taken at the full operational power (the worst case scenario). The ambient atmospheric conditions including temperature, wind speed and direction, and stratification are the major modeling variables. Probability of each combination of those variables is defined by the climatology at the stack location for the heights above the stack. The climatology-defined probability of weather conditions is constant for each modeling run. There are two other, presumption-defined probabilities that are the same for all locations in the plume. The first one depends on the adopted peak factor for the gust amplitude. The commonly used value of three standard deviations was chosen in the present study hence the probability of occurrence of the gusts with such or higher one-side amplitude is considering the turbulent vertical velocity as the random Gaussian process. The second presumption defines specified limits for the ratio of the gust width to the turbulent integral scale. Using experimental data for turbulent shear flows by Praskovsky (1982, 1983) and references therein, the probability of gusts with the amplitude above three standard deviations and the width between 0.6 and one integral scale was estimated as The last component of the total probability for the gust occurrence is the intermittency factor which varies with the location in the plume-induced field. The intermittency factor defines the probability of high-amplitude gusts with large scales far away from the plume centerline. It is of utmost importance for the plume analysis because such gusts are strongly pronounced in an area where the impact of mean plume parameters is already negligible although the plumeinduced turbulence could still significantly affect an aircraft; Section 2.5. In the developed methodology, the impact likelihood is defined as a product of the intermittency factor, the above described presumption-dependent probabilities and the climatology-defined probability of considered weather conditions. Thus the intermittency factor is one more plume parameter to be modeled for evaluating a local likelihood of the impacts. 2.5 Plume Modeling Utilized in the present study plume modeling is described in Appendix H. From a large variety of aerodynamic models of turbulent buoyant jets and plumes, the numerically stable integral model by Jirka (2004, 2006) was chosen for describing mean characteristics of the plume-induced aerodynamic field in accordance with the following criteria: Suitability: Provides the mean flow characteristics at special scales of the order of meters and temporal scales of the order of seconds Computational efficiency: Run time on a personal computer does not exceed few minutes Universality: Describes a single jet and merging jets issuing from multiple sources over realistic range of stack and exhaust parameters and atmospheric conditions 15

16 Flexibility: Allows easy modification and enhancement Reliability: Physically substantiated, tested under a wide range of modeling conditions, and verified with available high-quality experimental data Simplicity: Has relatively simple mathematical formulation The model was applied for predicting the mean plume velocity, temperature and concentration of pollutants as well as the plume width and trajectory from single or multiple exhaust stacks. Power plant Fort Martin Power Station, VW Mariposa Energy Project, CA Towantic Energy Project, CT Stack height 550 ft = 168 m 79.5 ft = 24.2 m 150 ft = 45.7 m Number of stacks Stack separation ft = 47 m 130 ft = 39.6 m Stack diameter 35 ft = 10.7 m 12 ft = 3.66 m 18.5 ft = 5.64 m Exhaust velocity 59.5 fps = 18.1 m/s 90.2 fps = 27.5 m/s 58.4 fps =17.8 m/s Exhaust temperature 131 F = 328 K 840 F = 722 K 201 F = 367 K Ambient temperature 22 F = 268 K 38 F = 276 K 16 F = 264 K Table 2: Stack parameters and exhaust characteristics for considered plumes The chosen model is capable to handle rather an arbitrary but steady vertical wind profile and covers a wide range of atmospheric conditions. As an illustration, aerodynamic fields for exhaust stacks of three power plants were modeled at neutral stratification which represents the worst case scenario. The modeling conditions are specified in Table 2. The reason for choosing the first plant located near the Morgantown Municipal airport, WV was a severe turbulence encounter by the United Express flight 6922 on 18 December 2008 described in Appendix B. The second plant is proposed to be built near the Byron airport, CA and it has already been a subject of few intensive studies; e.g., Mariposa (2010) and references therein. The third one is proposed to be built close to the Waterbury Oxford airport, CT and the concerns on its impact on aviation safety were raised in Pietrorazio (2010). The modeling parameters for the second and third plants were taken from Mariposa (2010) and AECOM (2009), respectively. An example of the mean velocity and temperature fields for the Fort Martin Power Station at calm winds is presented in Figure 3. The example illustrates that the temperature excess diffuses much faster than the vertical velocity. In particular, at 1,800 ft AGL the temperature excess drops about 22 times (from 60 K at the orifice to 2.7 K) while the velocity drops only about 5 times (from 18.1 m/s to 3.6 m/s). It is noteworthy that concentration of air contaminants dilutes at the same high rate as the temperature excess decreases; see also Mariposa (2010) and Senta (2010). As explained in Appendix G and illustrated in Figure G2, the impact of vertical gusts on aircraft is especially strong when the gust width is comparable to the size of aircraft. Shown in Figure 3 plume radius is defined in a conventional way as the location where the mean velocity decreases e times from its centerline value; e.g., Jirka (2004) and reference therein. Results in Figure 3 demonstrate that the plume radius grows rather fast and becomes much larger than a typical 16

17 size of light and medium aircraft at low heights above the stack; see also Katestone (2006) and (2010). Turbulence-generated gusts behave differently which is illustrated in Figure 4. Figure 3: Spatial distributions of the plume-induced mean vertical velocity in fps (left) and mean temperature in degrees Fahrenheit (right). Fort Martin Power Station, calm winds. Hereafter the solid lines show the conventional plume radius In this figure the mean velocity profile and plume diameter correspond are those for the Fort Martin Power Station at about 2,400 ft AGL at calm winds. An aircraft with a size of 60 ft like Saab-340B is shown for a comparison. The vertical turbulent velocity is simulated as a random process with the amplitude and scale corresponding to results of modeling. The schematic illustrates qualitatively the following statements. The amplitude of turbulent vertical gusts is at least twice larger than the mean velocity The width of strong turbulent gusts is about three times smaller than the plume radius Turbulent gusts have the largest velocity gradients and their impact on aircraft and aircrew is significantly more hazardous than that of mean velocities A plume area with strong turbulent gusts is more than twice wider than that with considerable mean velocities Figure 4: Schematic of plume-induced vertical turbulent gusts and their impact on aircraft The statements are supported quantitatively in Appendix H and allow concluding that the plumeinduced turbulence is the dominant cause of the harmful plume impact on aircraft. Therefore, an adequate turbulence modeling is necessary for evaluating potential plume impact on aviation safety. The term adequate means that modeling provides all turbulent characteristics that are 17

18 needed for quantifying plume impact on an aircraft. As shown in Section 2.4, the minimum set of such characteristics includes the amplitude of turbulent vertical gusts, characteristic scale of such gusts in the along-the-flight-track (horizontal) direction and the external intermittency factor. None of the existing plume models provides those characteristics hence new turbulence model was needed to be developed. The developed model needed to be consistent with the integral model by Jirka (2004, 2006) that is chosen for modeling the mean flow. The model has been developed that satisfies the above requirements; see Appendix H for details. It is based on the field model for buoyant turbulent flows presented in Hossain and Rodi (1982) which was chosen as the starting point for the following reasons. It employs the most reliable and well-tested two parametric k ε field model (k is the turbulent kinetic energy and ε is the rate of its dissipation) It was specifically modified for describing reliably a broad range of buoyant flows It was thoroughly tested by comparing with a diverse body of available experimental data 2.6 Mapping the Plume-Induced Risk and Estimating Clearances The essence of the new methodology is to assess quantitatively spatial distribution of the plumeinduced risk to an aircraft and aircrew, to create a map of the risk level. According to the SRM process, the risk level is assessed as low, medium or high in each location of the plume-induced aerodynamic field for each hazardous plume impact as the composite of predicted local values of the impact s severity and likelihood; see Table 3. The last column of the risk matrix in the SMS manual is not shown in Table 3: it describes catastrophic severity that is not present in the plume-induced field; see Appendices E-6 and E-7. Severity Likelihood Probable/ Probable (B) Remote (C) Extremely Remote (D) Extremely Improbable (E) Minimal 5 Minor 4 Major 3 Hazardous 2 Table 3: The risk matrix from the FAA SMS Manual: red high; yellow medium; green low risk As illustrated in Figure 2, the risk map is linked quantitatively to the stack dimensions, characteristics of exhaust, type and configuration of an aircraft, and climatology data in the stack location. The assessment begins from modeling relevant characteristics of the plume-induced aerodynamic field including the mean vertical velocity, the turbulence intensity, integral scale and the intermittency factor; Section 2.5 and Appendix H. The temperature field is also modeled and its effect on the mean vertical velocity and turbulence is comprehensively accounted for; an illustration is given in Figure 3. However, neither temperature nor air contaminants are considered as output parameters by themselves because they do not produce noticeable impact on aircraft and aircrew; Section 2.3. Quantitative evaluation of the impact severity is accomplished in two steps. First, the magnitude of each potentially hazardous plume impact defined in Section 2.3 is evaluated. Evaluation provides spatial distributions (maps) for the impact magnitude that would be experienced by aircraft flying through the plume at variable conditions. Two impacts were considered in the 18

19 performed study: the airframe-damaging vertical gusts and the aircraft vertical acceleration. It was found that spatial dimensions of the plume-induced region with hazardous airframedamaging gusts are much smaller than those of the region with the vertical acceleration of hazardous severity (Appendix I) and only the dominant impact, the aircraft vertical acceleration is considered below. The magnitude of the acceleration near exhaust stacks listed in Table 2 was mapped for Cessna 172 aircraft on the final approach. The magnitude of the acceleration is linked to the amplitude and width of the vertical gusts throughout the plume-induced aerodynamic field; Appendix G. In the second step, the magnitude of the impact (acceleration) was related to its severity using the expert-specified quantitative limits in Table 1. The process of mapping severity of the aircraft vertical acceleration is illustrated in Figure 5. Amplitude of vertical gusts, fps Vertical acceleration, g-units Gust width, ft Severity of the impact Figure 5: Severity of the vertical acceleration for Cessna 172 aircraft near the Fort Martin Power Station at an ambient wind speed of 1 knot. Color coding for severity: green minor, yellow major, and red hazardous The likelihood of the potentially hazardous plume impacts was estimated in accordance with the FAA SMS Manual. As recommended for the safety professional, the Flight Procedures and NAS Systems columns of the table were used; see Table 4. Likelihood Probable Remote Extremely Extremely Remote Improbable Probability p o p o > p o < p o < 10-7 p o < 10-9 Table 4: Likelihood definitions reproduced from the FAA SMS Manual The table provides quantitative limits for the impact likelihood that were applied for estimating the risk. 19

20 As explained in Section 2.4, the likelihood depends on the a priori estimated constants, climatology data for the stack location and a spatial distribution (map) of the external intermittency factor in the plume-induced turbulent field at variable conditions; Appendix I. Severity of the vertical acceleration impact The impact likelihood Green minor, Yellow major Red hazardous White extremely improbable Green extremely remote Yellow remote Risk level for the impact Green low, Yellow medium Red high (unacceptable) Figure 6: Risk level for the vertical acceleration of Cessna 172 aircraft near the Fort Martin Power Station at an ambient wind speed of 1 knot The maps for the impact severity and likelihood at variable conditions were used for mapping the risk level using the SRM rules in Table 3 and the process is illustrated in Figure 6. Figure 7: Map of a combined risk level for the vertical acceleration of Cessna 172 aircraft near the Fort Martin Power Station. Preliminary estimates for the minimum vertical and horizontal clearances of 1,810 ft AGL and 330 ft respectively are indicated by the horizontal and vertical dash lines Finally the risk maps for all impacts over the whole range of relevant conditions are combined into the risk map for analyzed stack and aircraft; Appendix I. The recommended vertical and horizontal clearances are chosen to be equal to the height and width of the region with the high 20

21 risk level which is classified in the SRM process as unacceptable. Combined risk map is shown in Figure I6 and the recommended clearances are illustrated in Figure 7. More details on evaluating the impact severity and likelihood and combining them for assessing the risk level for flying through plume aircraft and aircrew are given in Appendix I. Using the methodology, the minimum clearances were evaluated for exhaust stacks of power plants listed in Table 2 for Cessna 172 aircraft on the final approach; Appendix I. Preliminary estimates for the minimum vertical and horizontal clearances are respectively 1,810 ft AGL and 330 ft for the operating Fort Martin Power Station near the Morgantown airport, WV; 1,240 ft AGL and 320 ft for the proposed Towantic Energy plant near Waterbury Oxford airport, CT; and 1,120 ft AGL and 230 ft for the proposed Mariposa Energy plant near the Byron airport, CA. The estimates showed that even at rather typical conditions vertical plumes can generate spatial regions of an unacceptable risk level with significant dimensions. It seems therefore constructive to assess a potentially detrimental plume impact on aircraft and aircrew for industrial facilities with considerable efflux velocity and/or temperature. 21

22 3. Conclusions The performed thorough analysis of the impact of vertical plumes and exhaust effluent on aviation safety allowed drawing the following conclusions. Vertical plumes can generate an unacceptable risk level for an aircraft and aircrew over spatial regions with significant dimensions at rather typical conditions The minimum vertical and horizontal clearances for avoiding harmful plume effects are recommended to be equal to the height and width of the region with the unacceptable risk The height and width of such regions for Cessna 172 aircraft on the final approach and thus recommended minimum vertical and horizontal clearances can exceed 1,300 ft AGL and 300 ft respectively at typical conditions The developed innovative methodology provides physically substantiated, consistent and adequate guidance for evaluating the minimum vertical and horizontal clearances to avoid harmful plume effects on aircraft and aircrew The plume-related incidents are quite frequent in the airports that have exhaust stacks underneath of, or very close to typically used arrival/ departure flight routes A majority of plume-related incidents remains unreported The exposure time of aircraft, aircrew and passengers to exhaust effluent is too short for any noticeable effects on the human health or aircraft performance Concentration of the air contaminants drops down much faster than the amplitude of vertical gusts decreases thus the expected impacts of air contamination are much weaker than the dynamic ones High temperature and oxygen-depleted air in exhaust plumes cannot produce noticeable negative effect on an aircraft performance due to a short exposure time No adverse impacts of plume effluent neither on a health of aircrew or passengers nor on the aircraft performance are to be expected The plume induced condensation clouds do not affect aviation safety The plume-induced turbulence is the dominant cause of the detrimental plume impact on aircraft and aircrew An adequate turbulence modeling is the vital component of a representative study of the potential impact of vertical plumes on aviation safety 22

23 References AECOM, 2009: Towantic Energy Project - PM2.5 Modeling, Air Quality Study Report, AECOM Environment, Westford, MA. AFS-420, 2006: Safety risk analysis of aircraft overflight of industrial exhaust plumes, Safety Study Report, DOT-FAA-AFS Antoine, Y., F. Lemoine, and M. Lebouche, 2001: Turbulent transport of a passive scalar in a round jet discharging into a co-flowing stream, Eur. J. Mech. B Fluids, v. 20, Antonia, R.A. and Q. Zhao, 2001: Effect of initial conditions on a circular jet, Experiments in Fluids, v. 31, Argawal, A. and A.K. Prasad, 2003: Integral solution for the mean flow profiles of turbulent jets, plumes, and wakes, J. Fluids Eng., v. 125, Becker, H.A., H.C. Hottel, and G.C. Williams, 1967: The nozzle-fluid concentration field of the round, turbulent, free jet, J. Fluid Mech., v. 30, Best, P., L. Jackcon, C. Killip, M. Kanowski, and K. Spillane, 2003: Aviation safety and buoyant plumes, Proc. Of Clean Air Conference, Newcastle, New South Wales, Australia. Boeing, 1998: Aerodynamic Principles of Large-Airplane Upsets, AERO Magazine, vol. 03, The Boeing Co., Bowles, R.L., 1990b: Windshear detection and avoidance airborne systems survey, 29 th IEEE Conf. on Decision and Control, v. 2, Briggs, G.A., 1965: A plume rise model compared with observations, J. Air Pollution Control Assoc., v. 15, Briggs, G.A., 1975: Plume rise predictions, Chapter 3 in Lectures on Air Pollution and Environmental Impact Analysis, D. A. Haugen, editor, AMS, Burattini, P, R.A. Antonia, and L. Danalia, 2005: Similarity in the far field of a turbulent round jet, Phys. Fluids, v. 17, , 14 pp. CEC, 2007: Business Meeting Before the California Energy Resources Conservation and Development Commission, Transcript of 09/26/2007, California Energy Commission (CEC) Chen, C.J.,and W. Rodi, 1980: Vertical turbulent buoyant jets: a review of experimental data, HMT, Pergamon Press. Cornman, L.B., C.S. Morse, and G. Canning, 1995: Real-time estimation of atmospheric turbulence severity from in-situ aircraft measurements, J. of Aircraft, v. 32, no. 1, Dimotakis, P.E., R.C. Miake-Lye, and D.A. Papantoniou, 1983: Structure and dynamics of round turbulent jets, Phys Fluids, v.26, no. 11, El Hayek, M. D., 2004: Prediction of heated turbulent jet flow using advanced scalar Turbulence Method, Proc. of IMEC2004 International Mechanical Conference, December 5-8, Kuwait, EPA, 1989: User s Guide to the Complex Terrain Dispersion Model Plus Algorithms for Unstable Situations (CTDMPLUS): Vol. 1. Model Description and User Instructions, EPA/600/8-89/041, U.S. Envir. Prot. Agency, 23

24 EPA, 1995: User s guide for the industrial source complex (ISC3) dispersion models, v. 2 Description of model algorithms, EPA-454/B b, U.S. Environmental Protection Agency, EPA, 2004: AERMOD: Description of model formulation, EPA-454/R , U.S. Environ. Protection Agency, FAA, 2006: Turbulence reporting procedures, 24 th Meeting of IPACG/24, Honolulu, HI, USA. Fan, L.-N, and N.H. Brooks, 1969: Numerical Solutions of Turbulent Buoyant Jet Problems, Report No. KH-R-18, W. M. Keck Laboratory of Hydraulics and Water Resources Division of Engineering and Applied Science, Cal. Inst. Tech, Pasadena, California, Frost, W. and R.L. Bowles, 1984: Wind shear terms in the equation of aircraft motion, J. of Aircraft, v. 21, no. 11, Godden, D.; Lurmann, F., 1983: Development of the PLMSTAR Model and Its Application to Ozone Episode Conditions in the South Coast Air Basin, Southern California Edison Co.: Rosemead, CA. Heuwinkel, R., 2005: Automated air reporting, 8 th Meeting of METLINKSCG, Montreal, Canada. Hinze, J.O., 1975: Turbulence, 2 nd edition, McGraw Hill Co., 790 pp. Hossain, M.S., and Rodi,W., 1982: A turbulent model for buoyant flows and its application to vertical buoyant jets, in Turbulent Buoyant Jets and Plumes, v. 6, Ed. W. Rodi, Pergamon Press. Karachandani, P., L. Santos., C. Tonne, and C. Seigneur, 2000: Development and Evaluation of a State-of-the-Science Reactive Plume Model, Envir. Sci. Technol., v. 34, Katestone, 2007: Plume vertical velocity assessment of a proposed gas-fired power station at Russell City Energy Center, Report for job KE , Katestone Environmental PTY LTD, Toowong, Queensland, Australia. Katestone, 2010: Assessment vertical plume velocities for the Mariposa energy project, Report for job KE , Katestone Environmental PTY LTD, Toowong, Queensland, Australia. Kuznetsov, V.R., A.A. Praskovsky, and V.A. Sabelnikov, 1990: Intermittency and fine-scale turbulence structure in shear flows. Parts 1 and 2, Proc. of Inst. for Thermo-Physics, v. 223 and 224, Novosibirsk, 98pp. Kuznetsov, V.R., A.A. Praskovsky, and V.A. Sabelnikov, 1992: Fine-scale turbulence structure of intermittent shear flows. J. Fluid Mech., v. 243, Jirka, G.H., 2004: Integral Model for Turbulent Buoyant Jets in Unbounded Stratified Flows, Part 1: Single round jet, Environ. Fluid Mech., v. 4, Jirka, G.H., 2006: Integral Model for Turbulent Buoyant Jets in Unbounded Stratified Flows, Part 2: Plane jet dynamics resulting from multiport diffuser jets, Env. Fluid Mech., v. 6, Lester, P.F., 1993: Turbulence: A New Perspective for Pilots, Jeppesen Sanderson Inc. Lester, P.F., 1995: Aviation Weather, Jeppesen Sanderson Inc. Lewis, M.S., P.A. Robinson, D.A. Hinton, and R.L. Bowles, 1994: The relationship of an integral wind shear hazard to aircraft performance limitations, NASA TM

25 List E.J., 1982a: Turbulent Jets and Plumes, Ann. Rev. Fluid Mech., v.14, List, E.J., 1982b: Mechanics of turbulent buoyant jets and plumes, in Turbulent Buoyant Jets and Plumes, v. 6, Ed. W. Rodi, Pergamon Press. Mariposa, 2010: Mariposa Energy Project, Staff Queries, Set 1, Mariposa Energy, LLC, CA. Morton, B.R, G.I. Taylor, and J.S Turner, 1956: Turbulent gravitational convection from maintained and instantaneous sources, Proc. Roy. Soc. London A, v. 234, Muppidi, S., and K. Mahesh, 2007: Direct Numerical simulation of Round Jets in Crossflow. J. Fluid Mech., v. 574, NIWAR, 2004: Good Practice Guide for Atmospheric Dispersion Modeling, National Institute of Water and Atmospheric Research (NIWAR), Ministry for the Environment, New Zealand Pantokratoras, A., 1998: Horizontal penetration of inclined thermal buoyant water jets, Int. Comm. Heat Mass Transfer, v. 25, no. 4, Pietrorazio, R., 2010: Industrial Plant Plumes and Effects on Aviation, Presentation to FAA, FAA Headquarters, Washington D.C., 02/23/2010. Praskovsky, A.A., 1982: Statistics of turbulent bursts. Proc. of the USSR Inst. for Sci. Information, v. 4831, Moscow, Praskovsky, A.A., 1983: Statistical characteristics of turbulent bursts. Izv. Akad. Nauk SSSR, Prikl. Mech.i Tech. Phys., 3, Proctor, F.H., D.A. Hinton, and R.L. Bowles, 2000: A windshear hazard index, 9 th Conf. on Aviation, Range and Aerospace Meteorology, Orlando, Florida, AMS, Scire, J.S., D. Strimaitis, R.J. Yamartino, and E. M. Insley, 1997: User s Guide for the CALPUFF Dispersion Model (Version 5.0); EarthTech, Inc., Concord, MA. Schmidt, L.V., 1998: Introduction to Aircraft Flight Dynamics, AIAA Education Series, 397pp. Senta, 2010: Plume Effects on Local Aviation: Effects of Reduced Oxygen Levels, Report for the Mariposa Energy Project, Senta Engineering, LLC, Davis, CA. Spillane, K.T. and G.D. Hess, 1988: Fair weather convection and light aircraft, helicopter and glider accidents, J. of Aircraft, v. 20, no. 1, Stengel, R.F., 2004: Flight Dynamics, Princeton University Press, 752pp. Sumwalt, R.L., 2003: Airplane upset recovery training: a line pilot s perspective, Flight Safety Digest, vol. 22, no. 7-8, Sykes, R. I., R. S. Gabruk., 1997: A Second-Order Closure Model for the Effect of Averaging Time on Turbulent Plume Dispersion, J. Appl. Meteorol., v. 36, URS, 2006: Siting and feasibility for the proposed electrical generating plant at NCAS Miramar, Final report for contract N D-3032 X005, URS Corporation, San Diego, California. 25

26 Walton, W. D., K B. McGrattan, and J. V. Mullin, 1996: ALOFT-PC smoke plume trajectory model for personal computers, Proc. of 19 th Technical Seminar, June 12-14, Calgary, Alberta, Canada, Environmental Protection Service, Wood, E., S. Herndon, R. Miake-Lye, D. Nelson, and M. Seeley, 2008: Aircraft and Airport- Related Hazardous Air Pollutants: Research Needs and Analysis, Airport Cooperative Research Program (ACRP) Report 7, Transportation Research Board, Washington, DC Wright, S.J., 1984: Buoyant jets in density-stratified crossflow, J. Hydraulic Eng., v. 110, no. 5, Wygnanskii, I., and H. Fiedler, 1969: Some measurements in self-preserving jet, J. Fluid Mech., v. 38, Xiao, J., J.R.. Travisand, and W. Breitung, 2009: Non-Boussinesq Integral Model for Horizontal Turbulent Buoyant Round Jets, Science and Technology of Nuclear Installations, Article ID , 26

27 Appendix A: Glossary of Terms Aircraft Control: An ability to direct the movements of an aircraft with particular reference to changes in attitude and speed. Aircraft Stability: The property of an aircraft to maintain its attitude or to resist displacement, and if displaced, to develop forces and moments tending to restore the original condition. Assessment: Estimation of the size/scope of risk, or quality of a system or procedure. Cause: Events that result in a hazard or failure. Causes can occur by themselves or in combinations. Eddy Dissipation Rate (EDR): An aircraft-independent, universal, objective measure of the rate at which turbulent energy dissipates in the atmosphere. Effect: The effect is a description of the potential outcome or harm of the hazard if it occurs in the defined system state. Euler angles: The yaw, pitch and roll angles describe angular orientation of the aircraft s body with respect to the inertial frame Expert analysis: Techniques for clarifying uncertainties by gathering relevant information from a representative number of the subject matter experts, and analyzing the information with an adequate formalism, for example with fuzzy-logic-based methods that mimic human s process of making decisions. Gust: Abrupt, short blast of wind; the extreme magnitude of the instantaneous velocity. Hazard: Any real or potential condition that can cause injury, illness, or death to people; damage to or loss of a system, equipment, or property; or damage to the environment. A hazard is a condition that is a prerequisite to an accident or incident. Intermittency: Random temporal-spatial alteration of two clearly distinguishable states, like periodic and chaotic phases in dynamical systems, strong dissipation/ no dissipation regions in fully developed turbulent flows (internal intermittency), or turbulent/ non-turbulent regions on the far boundaries of turbulent shear flows (external intermittency). Likelihood: An expression of how often an event is expected to occur. Severity must be considered in the determination of likelihood. Likelihood is determined by how often the resulting harm can be expected to occur at the worst credible severity. Mitigation: Actions taken to reduce the risk of a hazard s effects. Plume: Thermal updraft generally associated with exhaust from the smoke stacks of power generating facilities, industrial production facilities, or other systems that have an ability to release large amounts of pressurized or otherwise unstable air. Can be visible or invisible in the air and disperse at various velocities and directions for a given facility output and atmospheric conditions. Process: A set of interrelated or interacting activities which transforms inputs into outputs. Requirement: In engineering, requirements document the need of what a particular product or service should be or do and used as an input into the design stage of product development. 27

28 Risk: The composite of predicted severity and likelihood of the potential effect of a hazard in the worst credible system state. Safety: Freedom from unacceptable risk. Severity: The measure of how bad the results of an event are predicted to be. Severity is determined by the worst credible outcome. Upset: A loss of airplane control in flight that happens when the pitch and/ or bank angles exceed specified limits. Wing stall: A condition where angle of attack exceeds its critical value which causes decrease in the lift. Worst Credible Outcome: The most unfavorable, yet believable and possible, condition given the system state. 28

29 Appendix B: Reports on the Plume-Related Incidents It seems constructive to begin with the National Transportation Safety Board (NTSB) report on the helicopter accident on August 9, 1989 at 11:15 Pacific daytime (NTSB Accident/ Incident Number LAX89LA270). A Bell Helicopter 206B lost engine power and autorotated to a parking lot striking several automobiles at Pyro Pacific Company s Mount Poso Cogeneration plant located near Bakersfield, California. One certified commercial pilot and one passenger received serious injuries. The flight purpose was to film the power generating plant. Visual meteorological conditions prevailed at the time of flight and the helicopter orbited the plant three times. On the third pass helicopter flew slowly over the plant s exhaust chimney. When at altitude about 20 ft directly above the chimney, the engine lost power. The pilot entered autorotation and the helicopter descended towards a parking lot. The helicopter was severely damaged and debris from wreckage was thrown outward damaging several vehicles. The plant was operated at the time of accident. The exhaust contained 70.3% of nitrogen, 20.9% of carbon dioxide, 5.2% of water vapor, 3.6% of oxygen,11.8 parts per million (PPM) of sulfur dioxide, 47.3 PPM of nitrogen oxides, and 47.5 PPM of carbon monoxide. The plume was invisible and its temperature was 350º Fahrenheit. According to the Allison Gas turbine Engine Company, the helicopter s engine Allison 250C20 is only certified up to 120º Fahrenheit. Two features of the incident are important for the present study. The helicopter has stayed just above the stack in the oxygen-depleted air with temperature of 350º Fahrenheit for a relatively long time before the engine lost power. As indicated by several witnesses, the helicopter approached the exhaust stack very, very slowly and hovered just above the top. It illustrates that for the engine to stall, it should be exposed to high air temperature and/or oxygen-depleted air for long enough time. No noticeable effects of efflux chemicals on the pilot and cameraman were reported in spite of extremely high concentration and relatively long time of exposure. It illustrates that chemical impact of plume overflight on aircrew could be negligibly small due to a short time of exposure, low concentration of hazardous gases away from the orifice, and a slow injection rate of an ambient air into an aircraft cabin. Below is an excerpt from another report on the plume-induced incident. In this and the following excerpts we have highlighted the statements that are especially important for the present study. December 18, 2008 Attention: Ms. Johnson Aviation Safety Hotline Program Office Reference: MGW ILS Rwy 18/Severe Turbulence Dear Ms. Johnson, On 18 December 2008, United Express flight 6922 operated by Colgan Air from CKB- MGW-IAD experienced severe turbulence during approach into MGW. The flight was on the ILS approach to runway 18, inside the Final Approach Fix, when the flight entered severe turbulence. The flight immediately executed a missed approach and diverted to the final destination, IAD, landing without any further incidence. The airplane was grounded for a severe 29

30 turbulence inspection. During the approach the airplane was in IMC conditions winds calm 100 overcast temperature 1 Celsius and surface visibility 2 miles. This was the second identical incident within the last two months. After reviewing the ILS 18 Rwy MGW approach plate we focused on the obstacle between the FAF and the runway. The obstacle stands at 1577 MSL. We called the MGW control tower to investigate the obstacle and we were told it is the smokestack from a power plant. We were also told by the tower that when the temperature is just right and the surface winds are calm the smoke creates turbulence during the final approach in to MGW. The tower also told us that FAA check flight was not happy during the checking events for the approach. According to my information this condition is not being reported to the flight crews. Our crews in this event reported uncontrolled flight, left engine ignition lights were activated, engine oil pressure lights illuminated, and all 3 axis trim circuit breakers tripped. Sincerely, Dean Bandavanis Director Operations The incident has occurred with the SAAB 340 Turboprop aircraft at altitude of about 1,700 ft above ground level (AGL) and about 1,150 ft above the smoke stack. One should note that severe turbulence was encountered at altitude of 1,700 ft AGL, much above 1,000 ft AGL recommended in AFS-420 (2006). The altitude of aircraft above the stack of about 1,150 ft greatly exceeded the maximum ROC of 350 ft in the final segment. Below are excerpts from several reports on turbulence encounters above the Blythe Power plant located in the close proximity to the Blythe airport (BLH), California. Mr. Joe Sheble from Sheble s Flight Service has reported his concerns with the Blythe turbulence on 02/19/2004. As a pilot who performs check rides for the FAA on student and commercial pilots on Instrument Landing System (ILS) approaches to various airports, he has experienced turbulence three times when flying over the Blythe plant while utilizing the ILS approach. He was flying either a Cessna 172 or a Beachcraft Traveler. He was about 300 feet above ground level (AGL) when flying over the plant. Some pilots fly 200 feet AGL over the plant, and Mr. Sheble believes the turbulence is enough to cause pilot trainees to do something stupid. A couple of pilots have told him that they have experienced turbulence as well. He believes that two thirds of the flights to Blythe Airport are done using visual flight rules (VFR) and many pilots do not see the power plant. He has also experienced even greater turbulence when flying downwind over a coal-fired power plant located about one mile from the Loflin Bullhead Airport in Arizona. The plant has one stack which is over 200 feet tall. His elevation when passing over the facility was 800 to 1000 feet AGL. Mr. James S. Adams reported his conversation with Mr. Eric Nordberg on 08/02/2004. I talked to Mr. Nordberg about his experience with turbulence from the Blythe power plant cooling towers. He and a co-pilot were flying a Lear jet (1800 lb. airplane) on an Instrument Landing System approach to Blythe airport s Runway 26 early (6:30 7) morning on May 4, They did not see any plumes and were about 550 feet above ground level with airspeed of 124 knots when they passed over the plant. The wind was calm with good visibility. They experienced moderate to severe turbulence which caused the plane to veer from side to side with considerable shaking. 30

31 They were surprised but able to regain control of the plane. It was not an emergency situation but it was an uncomfortable experience. I advised him that we had reports from several other pilots who have experienced the same thing and we were investigating the situation. On 06/09/2004 Mr. James S. Adams reported a conversation with Mr. Luis Magana from Sheble Aviation. Mr. Magana is a pilot and flying instructor who has been using Blythe Airport for several years. On the morning of May 4, 2004, he was aboard a two-engine Beechcraft airplane piloted by a student. They were on final approach to Runway 26 and saw the Blythe power plant in front of them. No plume was visible. Their elevation was approximately 550 feet above ground level and the airspeed was 110 miles per hour. As they flew over the cooling towers, they encountered significant turbulence which knocked the plane on its side or about 50 to 60 degrees off center. The student pilot was startled but was able to level the plane and proceed with the approach. After they landed, Luis discussed the incident with the student pilot and he considers it a good example of being prepared for the unexpected. He is very worried about new and inexperienced pilots in smaller planes such as a single engine Cessna 150 or 172 encountering similar turbulence. The smaller plane could be inverted and sent into a downward spiral, possibly crashing into or near the power plant. He also told me that a high percentage of the pilots that use the Blythe Airport are student pilots. The reported incidents allow drawing the following conclusions. The plume-related incidents are quite frequent in the airports that have exhaust stacks underneath of, or very close to typically used arrival/ departure flight routes. A majority of plume-related incidents remains unreported unless the incident creates quite a serious problem like severe turbulence or a pilot is asked to provide a report. Quite often pilots do not identify a plume as the cause for air turbulence because they do not see the stack or do not attribute turbulence to the exhaust plume. Invisible plumes are the most hazardous because the impact on aircraft is unexpected and takes PIC by surprise. The physical impact of exhaust plumes on aircraft is especially hazardous for new and inexperienced pilots. Calm winds and low air temperature intensify the impact of exhaust plumes on aircraft. The aircraft is exposed to the plume-induced high air temperature and/or oxygen-depleted air for too short time for affecting noticeably the engine performance at typical flight conditions. Impact of exhaust effluent on aircrew is negligibly small due to a short time of exposure and low concentration of hazardous gases away from the orifice at typical conditions. 31

32 Appendix C: Existing Approaches to the Problem Two conceptually different approaches to dealing with impact of exhaust plumes on aviation safety are considered and discussed in this appendix. AFS-420 (2006) Study In accordance with the FAA SRM methodology, the hazard-associated risk was assessed as a composite of predicted severity and likelihood of the potential effect or outcome of the hazard in the worst credible system state. The subject matter experts (SME) team identified the following plume-related hazards: H1-2006: High efflux temperature or velocity from industrial facilities (power plant exhaust plumes) may cause air disturbances that would have the potential to cause airframe damage and/or negatively affect the stability of aircraft in-flight. H2-2006: Exhaust plumes from industrial facilities (power plant, gas or coal fired furnaces, etc.) could results in restricted visibilities with high level of water vapor, icing, and engine/aircraft contaminants that would have a detrimental effect on aircraft/aircrew performance. These individually or cumulatively could possibly result in substantial aircraft damage, and/or loss of aircraft and crew as well as damage to ground facilities. The hazards H and H characterize dynamic and chemical impacts of the plume on aircraft. One can see that Specific Task 1 and Specific Tasks 2 4 for this project (see Introduction) are directly related to the potential hazards H and H2-2006, respectively. Task 5 includes risk analysis and assessment for the identified hazards as well as recommendations on treatment/ mitigation of the risk, if required. Specific Tasks 1 5 emphasize an inherent connection of the present scientific analysis to the AFS-420 (2006) study. The SME team agreed at the brainstorming sessions that the power plant exhaust plumes do not present an immediate or critical increase in human mental or physical workload resulting in any commensurate decrease in performance. This statement was based on a presumption that the pilot is prepared to see and avoid the plumes in which case his/her mental and/or physical resources would not be as task-overloaded as to preclude a safe maneuver out of and away from the condition. A severity of the hazards H and H was defined as respectively the major and the minor. Evaluation of the hazard s likelihood was based on a search of two databases: the National Aeronautics and Space Administration (NASA) Aviation Safety Reporting System (ASRS) and the FAA National Aviation Safety Data Analysis Center (NASDAC) Accident/Incident Data System (AIDS). As stated in the AFS-420 (2006), over 671,006 NASA ASRS pilot reports gathered over 30-years period indicated zero overflight incidents with exhaust plumes from facilities such as power plants. The NASDAC AIDS database of approximately 150,000 records indicated no accidents and one possible but not confirmed helicopter incident in Using the FAA annual surveys on the number of hours flying by GA pilots, the SME team defined the target level of safety and assessed the probability of occurrence for plume-related incident/ accident to be much below that level. The likelihood of incident/ accident caused by overflight of an exhaust plume was therefore evaluated as extremely remote. The expert team combined severity of the hazards and their likelihood into a risk matrix and determined the risk of aircraft overflight of industrial exhaust plumes as low which means 32

33 acceptable without restriction or limitation; hazards are not required to be actively managed but are to be documented. As emphasized in the Introduction, the AFS-420 (2006) study has been as thorough as possible at limited time and resources. It produced an assessment of a cumulative risk level for plume overflight that does not depend explicitly on exhaust parameters, plume characteristics, atmospheric conditions, dynamic and chemical impact of effluents on aircraft and aircrew, and the pilot reaction to the impact although is supposed to incorporate all these factors implicitly. It was however presumed in brainstorming sessions that the pilot can see and avoid power plant exhaust plumes. The presumption works well for visible plumes but cannot be applied to invisible plumes unless the pilot is well familiar with the area. Physical impact of a plume on an aircraft greatly depends on atmospheric conditions and even knowing about the plume pilots might underestimate its severity and fly over. Encounter of severe turbulence by the United Express commuter aircraft and the helicopter accident in California (Appendix B) demonstrate that severity of the dynamic hazard H could reach the hazardous level. Brainstorming is attractively simple and effective technique for evaluating cumulative severity of the hazards. However the technique provides a qualitative evaluation and is accurate only in the undoubtedly extreme situations such as no safety effect or catastrophic. When severity could gradually change from one extremity to another, like from hazardous just near an orifice of high-velocity/ temperature plume to no safety effect sufficiently far from the orifice, a qualitative evaluation could be inaccurate, hence cannot be used in the present study. Similarly to evaluating the hazard severity, the most straightforward and effective approach to assessing the probability of occurrence/ likelihood was chosen in AFS-410 (2006) study: statistical analysis of available databases. The approach is very reliable and provides accurate results if (but only if!) the databases are comprehensive, adequate, and representative. It seems that the NASA ASRS and the FAA NASDAC AIDS databases do not satisfy those requirements. Neither the NASA ASRS nor the FAA NASDAC AIDS database contains plume-related pilot reports or the NSTB report on the helicopter accident in California on 08/09/1989 that are listed in Appendix B. Encounter of severe turbulence by the United Express airplane on approach to MGW on 12/18/2008 happened recently and has not been considered either. Moreover, GA pilots often do not report turbulence encounters and even more significant problems like a roll up to 60º unless it caused an incident/ accident or they were specifically asked to do so (Appendix B). Hence none of official databases could be sufficiently comprehensive for analyzing the probability of occurrence for such relatively rare event as the plume overflight. In the USA there are 599 airports certificated to serve commercial air carrier aircraft and about as many used only by the GA aircraft. However only a few of the airports have exhaust stacks near designated departure/ arrival flight routes, probably no more than twenty. To be adequate to the considered problem, statistical evaluation of the probability of occurrence should be executed only for the airports with exhaust plumes in the close proximity. It seems unlikely that there exists a database containing all information that is necessary for the evaluation. Even if a database on airports with exhaust stacks nearby existed, it would unlikely be representative for conclusive analysis. Indeed, the number of flights from each one of 33

34 such airports might be insufficient for reliable statistical evaluation. As mentioned above, the pilot reports might be scarce which would dramatically affect statistics of rare events. In addition, the unfamiliar with the area pilots might fly over a plume, experience strong impact but do not attribute it to the plume they know nothing about. Therefore statistical analysis of available databases may provide inaccurate results and cannot be used in the present study. The above considerations illustrate neither evaluation of the hazard s severity with brainstorming nor estimation of the hazard s likelihood with available databases is an adequate approach to the problem in hand. The Australian CASA AC (0) Other than AFS-420 (2006) report, the only known official regulation on exhaust plumes is the Australian Civil Aviation Safety Authority (CASA) Advisory Circular (AC) (0) Guidelines for Plume Rise Assessment. The AC utilizes conceptually different approach to the problem from that in AFS-420 (2006). Instead of considering an overall severity and a cumulative risk level, the AC prescribes quantitative thresholds to specific flow characteristics for discriminating severity of gas efflux impact on aircraft and aircrew between the hazardous and non-hazardous. Applying the thresholds to a plume-induced aerodynamic field, one separates the field into hazardous and non-hazardous regions, and the region s location and size depend explicitly on the exhaust parameters and atmospheric conditions. Probability of occurrence for the regions is naturally estimated with the climatology data for the stack location. More specifically, the CASA AC (0) defines vertical velocity from gas efflux as the potential hazard that may cause airframe damage and/or affect the handling characteristics of an aircraft especially during initial take-off climb or approach to land with flaps extended and/or gear down. The document provides guidelines for assessing the potential hazard to aircraft operations of industrial facilities located within 15-km distance from airport. Following the guidelines, one could execute a study of potential hazards to aviation for existing or proposed exhaust stack. Using climatology for the stack location, one can estimate size, location, and probability of occurrence of hazardous regions for a given time of a day, season, specific atmospheric conditions, etc. The results could be used by aviation authorities for evaluating the impact of a smoke stack on a navigable airspace and/or by local authorities for issuing or denying a permit for construction/ alteration of the stack. An example of plume study in accordance with the CASA AC (0) guidelines can be found in Katestone (2007, 2010). The CASA AC (0) specifies only one threshold for quantifying hazardous impact of the exhaust on aircraft, the maximum magnitude of a vertical gust. More specifically, the AC states: 4.1 Airworthiness authorities have established that a vertical gust in excess of 4.3 m/s may cause airframe damage to an aircraft at critical stages of flight, e.g. when approaching to land with flaps extended. 34

35 4.2 CASA therefore requires that an exhaust plume which has an average vertical velocity exceeding the limiting value of 4.3 m/s at the aerodrome OLS 5 or at least at 360 feet AGL to be assessed as a potential hazard to aircraft operations. As far as we were able to find out, the limiting value of 4.3 m/s for the vertical gust does not have any physical basis and it is too conservative. The FAA FAR Part requires light aircraft to be able to structurally handle ±25 fps gust normal to the flight path for takeoff, approach and landing configurations (25 fps = 7.62 m/s); see Appendix E-3 for more details. The structural limit is much higher for commuter aircraft (±66 fps = ±20.1 m/s). It is noteworthy that even naturally occurring vertical velocities can reach 8 m/s in strong convective conditions (Spillane and Hess 1988). The paragraphs 4.1 and 4.2 contradict one another. The 4.1 limits a gust that is a peak value of the instantaneous vertical velocity. The 4.2 applies the same limit of 4.3 m/s to the temporally averaged vertical velocity which is about twice smaller than gusts at the plume centerline where relative turbulence intensity is typically about 30% (e.g., Hossain and Rodi 1982). The confusion increases further when temporally and spatially averaged velocity is used for evaluating the plume-related hazards because the cross-section mean velocity is about twice smaller than that at the plume centerline like in Katestone (2007, 2010). Moreover, specifying the peak or average magnitude of the gust without indicating its spatial scale along the flight route or time of exposure is physically inaccurate. It is obvious that the impact of a vertical gust on aircraft depends not only on the magnitude of its velocity but also on the gust width and/or duration. Furthermore, the impact is obviously dependent on the aircraft type, configuration and flight conditions; Section 2.4 and Appendix G. Although the basic approach to quantitative evaluation of the impact of exhaust plumes on aviation safety in the CASA AC (0) is constructive, a single limit on the vertical gust magnitude of 4.3 m/s seems fully inaccurate and cannot be used in the present study. 5 OLS - Obstacle Limitation Surfaces are boundaries of airspace surrounding the airport which is protected by Commonwealth Legislation (Australia) from intrusion or interference by buildings, trees, cranes, smoke, lights, air turbulence and any effect which could affect aviation safety. 35

36 Appendix D: Aerodynamic Criteria for Potentially Hazardous Air Disturbances If there were an established set of criteria for quantifying the plume impact on different types of flying through aircraft, it would reduce the problem to modeling exhaust plumes for practically realistic conditions and merely applying the criteria for separating non-hazardous and hazardous regions. As shown in Appendix C, the only available quantitative criterion, the CASA AC (0) limit on the vertical gust, is inadequate to the problem in hand. The lack of established criteria means that each potential aerodynamic hazard of an exhaust plume has to be considered separately. As found in AFS-420 (2006), those are windshear, turbulence, high temperature, and high concentration of chemical contaminants and water vapor. Existing criteria for quantifying the impact of windshear and turbulence are considered in this appendix. The impact of high temperature and chemical contaminants is considered in Appendix F and the effect of condensation clouds is discussed in Appendix E-9. Windshear The only known criterion for quantifying windshear thread is the NASA windshear hazard index (Bowles 1990, Frost and Bowles 1984, Proctor et al 2000). The index, typically referred to as the F-factor, incorporates observable atmospheric parameters, and it scales with aircraft flight performance in such a way as to predict impending flight path deterioration. FAA FAR Part requires all turbine-powered airplanes to be equipped with the airborne windshear detection and/or warning system. Although the F-factor is not specifically named in the FAR Part , it is imbedded into signal processing algorithms of the windshear systems. The concept employed in the derivation of the F-factor is the total aircraft energy and its rate of change. The total aircraft energy is a sum of the air-related kinetic energy (proportional to the square of airspeed) and the internal potential energy (proportional to the altitude above the ground). The local value of F-factor is defined as follows (Bowles 1990): U x w F = & g (D.1) V Here U x is the component of atmospheric wind directed horizontally along the flight path (positive for tail wind), U & x is the shear term the rate of change of U x experienced by aircraft, w is the updraft velocity, g is the gravitational acceleration, and V a is the aircraft airspeed. The F-factor is a non-dimensional parameter quantifying the impact of inhomogeneous wind field on the rate of change in the aircraft s total energy. Positive values of F correspond to windshear that decreases the energy state of an aircraft by decreasing airspeed (down to the stall airspeed), altitude (down to the ground), or both. The larger positive values of F are, the more hazardous the impact is. On the contrary, negative values of F act to increase the energy state and hence correspond to the non-hazardous windshear. a 6 FAR Part Operating requirements: Domestic, flag, and supplemental operations, Section Lowaltitude windshear system equipment requirements 36

37 One can find with Equation (D.1) that local magnitudes of F-factor are negative throughout vertical plume at all realistic efflux parameters and aircraft airspeed. To scale F-factor with aircraft flight performance, the local values (3.1) are averaged over a characteristic length. A one-km length scale has been adopted which corresponds to approximately 13-sec time scale at a typical approach speed of 75 m/s (Proctor et al 2000 and references therein). If one uses 35 m/s as a typical approach speed for light GA aircraft like Piper PA-28 Cherokee or Cessna 182 T, a 13 sec time scale corresponds to about 450 m. That is much larger than typical width of the exhaust plume not far from an orifice where the plume velocity is large enough for inducing noticeable windshear. Therefore, the plume-induced windshear is never a hazard if one applies the F-factor as the impact criterion. However, it does not really mean that plume-induced windshear cannot be hazardous especially for light GA airplanes. The F-factor has been developed for detecting and avoiding atmospheric hazards like wind fronts and especially microburst. Large-scale atmospheric events are relatively smooth hence it requires a time interval for a pilot to recognize a threat and react to it, and another interval for an engine to spool up to maximum thrust. Those intervals were considered of 5 sec each for jet transport aircraft (Lewis et al 1994) meaning that the horizontal extent of windshear exceeds at least 75 m/s 10 s = 750 m. Windshear scale in exhaust plumes is relatively small; usually no larger than 20 orifice diameters which correspond to about 120 m for a typical diameter of 6 m (e.g., URS 2006 and Katestone 2007) hence the respective time for the impact is about 3.5 sec at airspeed of 35 m/s. It means that physical impact of an exhaust plume differs conceptually from that of a microburst. If light GA airplane hits invisible vertical plume, it could be thrown up, down and/or rolled faster than a pilot could react. If such unexpected impact is strong enough, it could cause an incident/ accident especially for inexperienced pilot. It is noteworthy that the rate of change of U x experienced by aircraft in a plume is exactly opposite to that in a microburst. When crossing a microburst, an airplane first experiences head and then tail wind while in a plume this sequence is inversed. One can therefore conclude that F-factor is inadequate criterion for quantifying hazardous impact of windshear in vertical exhaust plumes. The ICAO Turbulence Index The International Civil Aviation Organization (ICAO) developed procedures for measuring and reporting turbulence intensity that has been accepted by FAA (FAA 2006). The ICAO Standards 7 define turbulence intensity in terms of the Eddy Dissipation Rate (EDR); see Heuwinkel (2005) and FAA (2006). EDR is an aircraft-independent, universal, objective measure of the rate at which turbulent energy dissipates in the atmosphere. EDR metric is a function of the atmospheric state, not of the characteristics of the aircraft measuring EDR a Cessna 172 should derive the same EDR turbulence metric value as a Boeing 747 when flying in the same turbulent air. Hence a given EDR level may mean a different impact on different aircraft, or even the same aircraft at a differing operating condition. A physical basis for choosing EDR is outlined in Heuwinkel (2005) and rigorous analysis can be found in Cornman et al (1995). ICAO documentation requires measuring both the average and the peak EDR values over 15-min interval for generating the turbulence index. If less than 15 min. of EDR are available, index 28 is assigned to the null case; Table D1. 7 ICAO, 2001: 14 th Edition of ICAO Annex 3 Standards and Recommended Practices "Meteorological Service for International Air Navigation", Appendix 4, section 2.6 Turbulence 37

38 Turbulence is defined as light for the index between 1 and 5, moderate for the index between 6 and 14, and severe for the index between 15 and 27. At any realistic conditions, the horizontal width of stack-induced turbulent domain does not exceed 150 orifice diameters, or about 900 m for a typical diameter of 6 m. Light GA aircraft at landing speed 35 m/s would traverse such domain for less than half a minute hence the ICAO turbulence index cannot be estimated ( null case). Similarly to the F-factor for windshear, EDR was chosen for characterizing atmospheric turbulence, mainly en route. The horizontal extent of atmospheric turbulent domains is of the order of tens of kilometers and 15-min interval is natural and realistic. Furthermore, the integral scale of atmospheric turbulence at altitude above 2.5 km exceeds 1 km which is much larger than an aircraft size. A size of aircraft of less than 100 m corresponds to that of small-scale turbulent eddies thus the inertial sub-range of the turbulence spectrum was utilized for relating EDR as the measure of turbulence severity to the aircraft vertical acceleration (Cornman 1995). This is not the case for plume-induced turbulence where the turbulent integral scale is of the order of or even smaller than a size of aircraft. Peak EDR value of turbulence Average EDR value of turbulence EDR (m 2/3 s -1 ) Null EDR (m 2/3 s -1 ) < >0.8 < > Null 28 Table D1: The ICAO turbulence index - reproduced from FAA (2006) At any realistic conditions, the horizontal width of stack-induced turbulent domain does not exceed 150 orifice diameters, or about 900 m for a typical diameter of 6 m. Light GA aircraft at landing speed 35 m/s would traverse such domain for less than half a minute hence the ICAO turbulence index cannot be estimated ( null case). Similarly to the F-factor for windshear, EDR was chosen for characterizing atmospheric turbulence, mainly en route. The horizontal extent of atmospheric turbulent domains is of the order of tens of kilometers and 15-min interval is natural and realistic. Furthermore, the integral scale of atmospheric turbulence at altitude above 2.5 km exceeds 1 km which is much larger than an aircraft size. A size of aircraft of less than 100 m corresponds to that of small-scale turbulent eddies thus the inertial sub-range of the turbulence spectrum was utilized for relating EDR as the measure of turbulence severity to the aircraft vertical acceleration (Cornman 1995). This is not the case for plume-induced turbulence where the turbulent integral scale is of the order of or even smaller than a size of aircraft. Therefore, neither the ICAO turbulence indices in Table D1 nor the EDR as a physical parameter are adequate criteria for quantifying the impact of plume-induced turbulence on aviation safety. 38

39 Measures of Turbulence Intensity In the Jeppesen Sanderson Training Products, Lester (1993, 1995) presents the quantitative ranges for the turbulence intensity which are often referred to in plume studies; e.g., Katestone (2010). The measures are given in Table D2 where the airspeed fluctuation is defined as the maximum variation from the sustained airspeed and vertical acceleration is a peak deviation from normal acceleration of 1.0g (g is the Earth gravitational acceleration). The sustained airspeed is related to the sustained wind and the latter is defined as the wind speed averaged over one two minutes for the low level turbulence. Typical time of plume overflight does not exceed sec which makes applicability of the airspeed fluctuations in the table at least questionable. Moreover, the landing airspeed of a light aircraft like Cessna 172 is about 65 kts. Fluctuations of 14.9 kts imply variations in the airspeed from about 50 kts to about 80 kts which is certainly much above light turbulence. Airspeed Vertical Derived Gust Fluctuations (kts) Acceleration (g) (ft per min) Light Moderate Severe 25 or higher Extreme or higher 3000 or higher Table D2: Quantitative measures of turbulence intensity. Values may be negative or positive. The table is reproduced from Lester (1995) The derived gust velocity in Table D2 is defined as an index of turbulence intensity; a theoretically determined, nearly aircraft independent, estimate of the vertical gust velocity (Lester, 1993, p. A-7). The definition seems unsubstantiated and even confusing. The FAA FAR Part defines the derived gust velocity as the limiting, airframe-damaging gust value which depends on the maximum acceptable load factor. The relation between the load factor and the derived gust velocity is given in the part and the magnitudes of the derived gust velocity for the Part 23 certificated aircraft are given in the FAA FAR Parts and In particular, the FAA FAR Part states that for the takeoff, approach or landing with flaps fully extended, the Part 23 certificated airplanes must withstand positive or negative vertical gust of 25 ft/s = 1,500 ft/min. Such gusts could cause airframe damage and should be identified as measures of severe / extreme rather than moderate turbulence. It follows from the above considerations that quantitative measures of turbulence intensity in Table D2 could be reasonable for characterizing extended areas of large-scale atmospheric turbulence for cruising commercial jets. The measures however are certainly inadequate for characterizing the impact on an aircraft at the takeoff, approach or landing of highly localized, small-scale turbulence induced by exhaust plumes. 39

40 Appendix E: Limited Expert Analysis As explained in Section 2.2, the developed methodology for assessing the risk of plume overflight requires information that could only be provided by the subject matter experts. Ambiguous classifications, non-established quantitative limits, and concepts allowing different interpretations are typical issues that could be resolved only by utilizing empirical wisdom of the highly skilled specialists in the field. At the performed round of gathering information, seven experts have been interviewed. Four of them are from the FAA Small Airplane Directorate, Kansas City, MO: Ross Schaller, Aerospace Engineer, Regulations and Policies Wes Ryan, Manager, Programs & Procedures, Aircraft Certification Service Peter L. Rouse, Aviation Safety Engineer, Aircraft Certification Service Robert Stegeman, Aerospace Engineer, Regulations and Policies Three more experts have been interviewed in Oklahoma City, OK: Alan Jones, Senior Research Analyst, Flight Standards Service, AFS-420 Lt Col Jim Rose, Military Program Manager, Flight Standards Service, AFS-420 Wayne D. Fetty Jr., Division Manager, Standards-Air Force Instrument Procedures Center, HQ Air Force Flight Standards Agency Specific issues of questioning the experts and their inputs are presented below. E-1 Frequency of stack overflight The frequency was supposed to be included into an estimate for the impact likelihood. The experts however indicated that there are at least three scenarios requiring different potential ways for estimating the frequency of stack overflight: the stack is located underneath the routinely used flight route(s), close to the commonly used routes, and far away from the routes. One cannot provide rational frequency estimates for the second and especially the third scenario. The expert consensus was to presume that there always is a possibility of the stack overflight at low altitude unless it is prohibited by the FAA regulations. The helicopter 1989 incident clearly demonstrates the point; Appendix B. Conclusion: presume that stack overflight at low altitude is always a real possibility unless it is prohibited by the FAA regulations. Suggestion: consider the impact likelihood to be equal to the total probability of the impact occurrence which is determined by climatology in the stack location and presumptions adopted for estimating the impact s severity. E-2 Definition of upset for light GA airplanes Upset is one of the major causes for loss of airplane control in flight. It is exactly identified and thoroughly studied for large, swept-wing commercial jet airplanes (e.g., Sumwalt, 2003; and Boeing, 1998). The follows four conditions are generally used for describing upset of large airplanes: Pitch attitude more than 25 degrees nose up, 40

41 Pitch attitude more than 10 degrees nose down, Bank angle more than 45 degrees, and Flight within these parameters at airspeeds inappropriate for the conditions. However the conditions for upset of light GA airplanes are not uniquely identified or documented. The experts were asked to share their thoughts on whether the upset conditions for light airplanes could be considered the same as those for large commercial jets. The consensus was that the upset conditions for light airplanes differ from those for large commercial jets, depend strongly on aircraft type and cannot be universally defined. It was suggested to consider severity of each specific impact on aircraft and aircrew (like pitch or bank angle) rather than use more generic and vaguely identified upset. Conclusion: for light GA aircraft, upset is an inadequately defined cause for the loss of aircraft stability and control. Suggestion: instead of upset, specific physical impacts on aircraft and aircrew like changes in pitch and bank angle should be considered as the causes for the loss of aircraft stability and control. E-3 Aircraft tolerance to vertical gusts The FAA FAR Section and requires aircraft certificated under the Parts 23 and 25 to be able to structurally handle ±25 fps gust normal to the flight path for takeoff, approach and landing configurations. It is stated in Section and that the vertical gust could be positive or negative and its shape is defined as follows: U de 2π s wgust ( s) = 1 cos (E.1) 2 L g Here w gust is the distribution of the vertical velocity across the gust, s is the distance into gust, L g is the gust width, and U de is the derived velocity referred to in subparagraph (1) of Section (c). The FAA FAR Section provides the same shape of the wind gust for the aircraft certified under the Part 25. For takeoff, approach and landing configurations U de = 25 fps and L g = 25 c for aircraft certified under both the Parts 23 and 25; hereafter c is the mean geometric chord of wing. It is specified in Sections and that certificated airplanes must be designed to withstand loads on each lifting surface resulting from the indicated gusts although neither section indicates possible damage that could be caused by the gust at w gust > U de and L g 25 chords. This issue was discussed with the experts who indicated that the damage could range from light permanent deformation of lifting surfaces to bended and broken control surfaces, especially flaps at landing configuration. The experts also emphasized that a single gust could cause airframe damage if its intensity exceeds the specified parameters. 8 The FAA FAR Part 23 Airworthiness standards: Normal, utility, acrobatic, and commuter category airplanes; Section High lift devices 9 The FAA FAR Part 25 Airworthiness standards: Transport category airplanes; Section High lift devices 10 Section Flight envelope 11 Section Gust [and turbulence] loads; Section Gust load factors 41

42 Conclusion: a single positive or negative vertical gust could cause significant damage to aircraft ranging from light permanent deformation of lifting surfaces to bended or broken control surfaces if its intensity exceeds that specified by Equation (E.1) Suggestion: the worst credible scenario, bended or broken control surfaces should be considered for assessing severity of the airframe damage E-4 Impact of high temperature and oxygen-depleted exhaust on airplane The experts were asked to address three questions related to a potential impact of high temperature and oxygen-depleted air in exhaust plumes. 1. What are the minimum air temperature and corresponding time of exposure that might cause structural damage of a light aircraft? 2. What are the magnitude of air temperature and time of exposure that could cause the engine stall for a light aircraft? 3. Could oxygen-depleted plume exhaust cause the engine stall for a light aircraft? The experts definitely stated that times of aircraft exposure to high temperature in exhaust plumes are as short as no structural damage could happen. It is well-known that temperature diffusion is much faster than that of vertical velocity and the temperature excess becomes negligible at ft above the stack orifice at the worst credible scenario; Section 2.5. The width of high-temperature region at such height does not exceed 150 ft which corresponds to no longer than sec of aircraft exposure. Such short exposure is certainly insufficient for causing any structural damage. None of experts was able to recall any FAA regulation that indicates possible damage to airframe by high air temperature. Short exposure time prevents engine stall due to high temperature or oxygen-depleted air as well. In the worst case, a light blip in the engine power could occur although it would not affect PIC actions or aircraft characteristics; see Senta (2010) for rigorous quantitative assessment. The helicopter 1989 accident (Appendix B) illustrates the expert s opinion. Its engine lost power after the helicopter hovered about 20 ft above the stack orifice for several seconds. The exhaust contained only 3.6% of oxygen and its temperature was 350 F while the engine was certified up to 120 F. Even after the engine stall, the pilot was able to land helicopter on autorotation and avoid casualties. If a light aircraft flew 20 ft above the stack where the plume width cannot exceed 30 ft, its exposure time would not exceed 0.3 sec thus the engine would not be affected significantly. The experts however emphasized that incidents/ accidents similar to that with helicopter must be prevented by prohibiting any flights too close to stack orifice in the FAA regulations (which is not currently the case). Conclusion: high temperature and oxygen-depleted air in exhaust plumes could not cause structural damage or engine stall for a light aircraft due to too short exposure time. Suggestion: indicate in the study conclusions that flying too close to stack orifice should be prohibited by the FAA regulations E-5 Airplane types, configurations, and flight conditions for modeling Quantitative physical impact of the plume-induced air disturbances on aircraft was evaluated by modeling dynamics of aircraft flight through disturbances of variable magnitude and scale. 42

43 The experts were asked to recommend airplane types for the initial modeling. Among suggested light GA airplane like Cessna 172, mid-size GA airplane like Gulfstream G550 and/ or small commuter airplane like Saab 340, the experts suggested executing initial modeling for Cessna 172 which could represent the worst credible case. Among possible flight conditions like level flight, landing, or departure, the experts recommended the final approach for the initial modeling. The experts also recommended using the DOD MIL-STD-1797B 12 standard for characterizing vertical gusts and analyzing aircraft dynamics in turbulence whenever the standard is applicable. Recommendation: as the starting point, analyze dynamics of Cessna 172 aircraft on the final approach flying through disturbances of variable magnitude and scale. E-6 Severity of selected physical impacts on airplane and aircrew Experts have been asked to rate a severity of several selected physical impacts on Cessna 172 airplane and its aircrew on the final approach at the altitude of about 1,000 ft AGL using the table E1. The table is reproduced from AFS-420 (2006) with definitions for severity from the SMS Manual, Version 2.1. Hazard Severity Classification - No effect on flight crew - No effect on safety - Inconvenience to occupants - Slight increase in flight crew workload - Slight reduction in safety margin or functional capabilities - Physical discomfort of occupants - Significant increase in flight crew workload - Significant reduction in safety margin or functional capabilities - Physical distress possibility including injuries - Excessive workload of flight crew - Large reduction in safety margin or functional capabilities - Serious or fatal injury to small number of occupants or cabin crew Outcome would result in: - Hull loss - Multiple fatalities Rating 5 (minimal severity) 4 (minor severity) 3 (major severity) 2 (hazardous severity) 1 (catastrophic severity) Table E1: Qualitative classification for the impact severity - the table is reproduced from AFS-420 (2006) with insignificant modifications If expert cannot provide a rating, he was asked for a blank space. Because only five experts were asked to provide the ratings, no weighted averaging or other sophisticated statistical procedure was applied but rather a median of the expert-provided estimates was selected as a true value. The expert-provided ratings for the impacts and the medians are as follows: Airframe damage: 2, 2, 3, 2. 1; median 2 Engine stall: 3, 2, 2, 3, 3; median 3 Wing stall: 3, 2, 2, 4, 2; median 2 12 The Department of Defense (DOD) standard MIL-STD-1797B: Flying Qualities of Piloted Aircraft, 15 Feb

44 Upset: 4, 3, 4, 1, 3; median 3 As shown above, the engine stall could not be caused by exhaust plume unless helicopter hovers just above the orifice for several seconds and this impact is excluded from further consideration. Upset will not be considered either because it is replaced by more specific impacts on aircraft and aircrew like abrupt changes in the pitch and roll angles. Conclusion: two potential plume-induced impacts on aircraft and aircrew, airframe damage and wing stall, will be considered as those of hazardous severity E-7 Quantitative characterization of other physical impacts Five experienced pilots were asked to provide their estimates for the magnitude of sudden physical impacts on Cessna 172 airplane and its aircrew on the final approach at the altitude of about 1,000 ft AGL that would affect a GA pilot with an average training and experience with specified severity using the below table (g is the Earth gravitational acceleration). As with the above described ratings, experts were asked for blank space when they cannot provide an estimate. Again, a median of the expert-provided estimates was selected as a true value. The expert s estimates are presented in the table and the medians are highlighted. Physical impact on aircraft Classification of the impact severity - (4) Slight increase in flight crew workload - Slight reduction in safety margin or functional capabilities - (3) Significant increase in flight crew workload - Significant reduction in safety margin or functional capabilities - (2) Excessive workload of flight crew - Large reduction in safety margin or functional capabilities The erroneous PIC actions that could result in incident/ accident Plunge, feet 150, 20, 40, 20, 50 50, 40, 100, 30, , 100, 150, 40, , 50 Ascend, feet 50, 30, , 40, , 50, , 75 Change in the pitch angle, degrees 5, 5, 5, 10, 5 10, 10, 10, 20, 10 20, 15, 15, 30, 15 45, 20, 40 Change in the bank angle, degrees 20, 10, 10, 30, 10 45, 30, 30, 40, 25 90, 60, 45, 50, , 60, 60 Amplitude of abrupt changes in the acceleration ±1g, ±0.2g, 0.5g, +2g or -1.5g, +1.0g or -0.3g, 1g, More than +2g or -1.5g, +1.5g or -0.5g, 1.5g, Table E2: The expert-provided estimates for the magnitude of sudden physical impacts that would affect a GA pilot with an average training and experience with indicated severity The highlighted estimates in Table E2 could be interpreted as quantitative limits for severity of potentially hazardous plume-induced physical impacts on Cessna 172 airplane and its aircrew on the final approach at the altitude of about 1,000 ft AGL. The results are summarized in Table 1. 44

45 Although some of the experts tried to guess sudden impacts that may cause erroneous pilot-incommand (PIC) actions (see Table E2), the consensus was that such magnitudes cannot be reasonably estimated. PIC actions depend on his/ her training, experience, alertness at given instant of time, etc. It was suggested to identify abrupt changes in a flight path, the Euler angles, and acceleration as potentially dangerous (instead of upset), and exclude erroneous actions of PIC from the list of analyzed impacts. Conclusion: table E3 identifies potentially hazardous plume-induced physical impacts of aircraft and aircrew and provides preliminary quantitative limits for their severity which could be applied to assessing the risk of plume overflight. E-8 Pilot response time Experts were asked to provide their estimate for the total reaction time (a sum of detection and response times) of a GA pilot with average training and experience flying Cessna 172 airplane on the final approach at the altitude of about 1,000 ft AGL to sudden changes in the aircraft behavior. The changes were supposed to be strong enough to require the pilot s counteraction. The expert s estimates are presented in the below table and the median values are highlighted. Sudden changes in Plunge or Change in the pitch Fluctuations in the the aircraft behavior abrupt ascend or bank angle vertical acceleration Total reaction time, sec 1.2, 1.0, 5.0, 3.0, , 0.5, 5.0, 3.0, , 1.5, 5.0, 3.0, 2.0 Table E3: The expert-provided estimates for the reaction time of GA pilot with average training and experience to sudden physical impacts The pilot s reaction time is not included directly into modeling of aircraft dynamics. It however defines time interval during which the impact could still be considered abrupt, unexpected and thus the maximum scale of air disturbance to be simulated. It follows from Table E3 that the total reaction time of 2.5 sec is a reasonably conservative estimate. Conclusion: 2.5 sec could be considered as a typical reaction time of a GA pilot with average training and experience to sudden physical impacts E-9 Obscuration effect of plume induced condensation clouds The AOSC-defined Specific Project Task 4 is: Examine the obscuration effects of plume induced condensation clouds. It is noteworthy that there are several well-tested condensation models. Such models provide reliable estimates for expected frequencies of visible cooling tower plumes as well as the potential for incidences of local fogging in the vicinity of the power plant sites when visible cooling tower plumes may occur at ground level. As an example, one can mention the Seasonal Annual Cooling Tower Impact (SACTI) model used in the URS (2006) study for the proposed electrical generating plant at NCAS Miramar, San Diego, California. SACTI also calculates local icing effects that would occur when moisture plumes intersect the ground or other surfaces during freezing ambient conditions. Therefore modeling of condensation clouds induced by industrial plumes, including fogging and icing effects, is a straightforward task that can be executed with available models. The expert s discussion was focused on another aspect of the issue. It was indicated that reduced visibility does not affect airplane operations under the Instrument Flight Rules (IFR). It was also indicated that flying at low visibility under the Visual Flight Rules (VFR) is strictly regulated. In particular, the FAA CFR, Title 14, Basic VFR weather minimums, states: Except as 45

46 provided in paragraph (b) of this section and , no person may operate an aircraft under VFR when the flight visibility is less, or at a distance from clouds that is less, than that prescribed for the corresponding altitude and class of airspace in the following table (Table E4). Airspace Flight visibility Distance from clouds Class B 3 statute miles Clear of Clouds. Class C Class D Class E: Less than 10,000 feet MSL 3 statute miles 3 statute miles 3 statute miles 500 feet below. 1,000 feet above. 2,000 feet horizontal. 500 feet below. 1,000 feet above. 2,000 feet horizontal. 500 feet below. 1,000 feet above. 2,000 feet horizontal Table E4: Limitations on flying under VFR at low visibility (only part of a table is reproduced from ) The regulation imposes clear and strict limitations on flying at low visibility under VFR. If PIC obeys the requirements in Table E4, he/ she will not be endangered by the effects of plume condensation cloud. Therefore the clouds do not affect aviation safety which is the object of the present study. It was noted that the condensation clouds could affect an ability to fly near the plume. The requirement to stay clear of clouds may impact traffic patterns, visual acquisition of landing environment, etc. and cause pilots to implement diversions and other cloud-avoidance maneuvers. However the maneuvers should still satisfy the regulations in hence the PIC safety is not compromised. The experts agreed that the plume-induced degraded visibility should not be treated as a safety issue although it might affect the airport efficiency. Conclusion: plume induced condensation clouds do not affect aviation safety 46

47 Appendix F: Safety Effects of Exhaust Effluent Specific Project Tasks 2 and 3 require examining the potential health impact of exhaust effluent on aircrew and passengers of aircraft traveling through a plume. The requirements of the Environmental Protection Agency (EPA) and/or the Occupational Health and Safety Administration (OSHA) should be considered in the examination process. Pollutant Averaging Time Primary Standard Particulate Matter -- Diameter 10 µm (PM-10) Particulate Matter Diameter 2.5 µm (PM-2.5) Ozone (O 3 ) Secondary Standard Annual arithmetic mean Chronic: 50 µg/m 3 Same 24-hour arithmetic mean, 99 th percentile, average over 3 years Acute: 150 µg/m 3 Annual arithmetic mean Chronic: 15 µg/m 3 Same 24-hour arithmetic mean, 98 th percentile, average over 3 years 8-hour, 3-year average of the annual 4 th highest daily maximum 8-hour concentration Acute: 65 µg/m ppm Same Sulfur Dioxide (SO 2 ) Annual arithmetic mean Chronic: 80 µg/m 3 24-hour, not to be exceeded more than once per year 3-hour, not to be exceeded more than once per year Acute: 365 µg/m 3 Acute: 1300 µg/m 3 Nitrogen Dioxide (NO 2 ). Annual arithmetic mean 1300 µg/m 3 Same Carbon Monoxide (CO) 8-hour, not to be exceeded more than once per year 1-hour, not to be exceeded more than once per year Chronic: 10 mg/m 3 Acute: 40 mg/m 3 Table F1: The NAAQS concentration limits for criteria pollutants The EPA established the National Ambient Air Quality Standards (NAAQS) that apply for outdoor air throughout the country. Primary standards are designed to protect human health, with an adequate margin of safety, including sensitive populations such as children, the elderly, and individuals suffering from respiratory disease. Secondary standards are designed to protect public welfare from any known or anticipated adverse effects of a pollutant (e.g. building facades, crops, and domestic animals). 47

48 NAAQS requires the EPA to set standards on six criteria air contaminants: ozone (O 3 ), particulate matter (PM 10, coarse particles: 2.5µm to 10 µm in size; and PM 2.5, fine particles: 2.5 µm in size or less), carbon monoxide (CO), sulfur dioxide (SO 2 ), nitrogen oxides (NO x ), and lead (Pb). The standards are listed in the EPA Title 40 CFR Part 50 and each standard has its own criteria for how many times it may be exceeded, in some cases using a three year average. NAAQS establishes concentration limits for criteria pollutants, i.e. those pollutants with a wellestablished dose/response relationship. An example of the NAAQS criteria is given in Table F1 where acute refers to concentrations that may lead to immediate death or illness, and chronic refers to concentrations where the health effects may materialize in the longer term. Table F1 shows that the averaging time for estimating the impact of air pollutants and the acceptable exposure time indicated in the EPA regulations are sufficiently large. In particular, the minimum interval of one hour is specified for high-concentration carbon monoxide. A travel time of aircraft through an exhaust effluent can be estimated by considering the maximum possible plume diameter of 2,000 ft (that could be reached at the distance of more than 5,000 ft from the stack orifice) and the aircraft speed of 100 ft/s (the minimum landing airspeed of Cessna 172). That gives 20 sec as the maximum imaginable aircraft exposure time to the effluent. No EPA or OSHA requirement was found for so short exposure time. The EPA standards were however applied as guidance for very thorough and extremely conservative study of a potential health impact of exhaust effluent; Mariposa (2010). It was found that the exposure time of aircrew and passengers to exhaust effluent is too short to exceed US government regulations including those of the EPA, OSHA, and Office for Environmental Health Hazard Assessment, National Institute for Occupational Safety and Health and the Agency for Toxic Substances and Decease Registry. Table F1 also illustrates that the EPA and OSHA requirements are very stringent. They are designed to protect human health including sensitive populations such as children, the elderly, and individuals suffering from respiratory disease who are exposed to the exhaust effluent for days, weeks and years. One should note that no exhaust stack can be built if it does not satisfy the EPA standards and thorough study of air contamination is the mandatory part of the approval process; e.g., URS (2006), AECOM (2009), and Mariposa (2010). Therefore the initial concentration of air contaminants in the exhaust plumes should not be too dangerous to satisfy the EPA requirements. It is indicated in Section 2.5 that concentration of air contaminants decreases very fast with distance from the orifice and it becomes negligibly small when the dynamic hazards are still an issue; see thorough quantitative estimates in Mariposa (2010). Described in Appendix B incident with the helicopter that hovered about 20 ft above the stack orifice for several seconds before the engine stall has shown no noticeable effects of efflux chemicals on the pilot and cameraman in spite of extremely high concentration and relatively long time of exposure. It further illustrates that chemical impact of plume overflight on aircrew could be expected negligibly small. As indicated by the experts, the impact of high temperature and oxygen-depleted air in exhaust plumes could not cause structural damage or engine stall for a light aircraft due to too short exposure time; Appendix E-4. The above mentioned helicopter incident also supports this statement: the helicopter hovered for several seconds in the stack exhaust containing only 3.6% of oxygen with the temperature of 350 F before the engine stall. Such extremes cannot be expected in exhaust plumes if pilot stays outside the region with the unacceptable risk due to 48

49 dynamic impacts; Section 2.6. Thorough quantitative evaluation has found that no significant effect on the engine performance can be expected at the plume overflight; Senta (2010). The last noteworthy issue is that air contamination in crowded airports due to idle/ taxi exhaust from aircraft engines often exceeds significantly the EPA standards; Wood et al (2008). Pilots are typically exposed to contaminated air in airports for much longer than 20 sec. hence the much stronger impact on their health could be expected than that of exhaust plumes. The above considerations could be summarized as follows. No EPA or OSHA requirement was found for the exposure time of 20 sec or less which is the maximum expected duration of aircraft travel through exhaust effluent The exposure time of aircraft, aircrew and passengers to exhaust effluent is too short for any noticeable effects on the human health or aircraft performance Concentration of the air contaminants decreases with distance from the orifice much faster than the amplitude of vertical gusts thus the expected contamination impacts are much weaker than the dynamic ones Effect of exhaust effluent from smoke plumes on a health of aircrew and passengers is expected much lower than that of a contaminated air in crowded airports One can conclude that no adverse impacts of plume effluent neither on a health of aircrew or passengers nor on the aircraft performance are to be expected. 49

50 Appendix G: Modeling the Aircraft Dynamics As explained in Section 2.6, quantitative evaluation of the impact severity was accomplished in two steps. In the first step, the magnitude of each potentially hazardous plume impact defined in Section 2.3 was evaluated to get spatial distributions (maps) for the impact magnitude that would be experienced by aircraft flying through the plume at variable conditions. To accomplish this step, one needed to relate the impact magnitude to characteristics of the modeled aerodynamic plume field. The relations were established using the appropriate FAR s and analyzing aircraft dynamics. Two impacts were considered in the performed study: the airframe-damaging vertical gusts and the aircraft vertical acceleration. Aerodynamic characteristics that can cause the first impact were established using the appropriate FAA regulations. As explained in Appendix E-3, the FAA FAR Sections and require certificated under the Parts 23 and 25 aircraft to be able to structurally handle ±25 fps gust normal to the flight path for takeoff, approach and landing configurations. The vertical gust could be positive or negative and its shape is defined by Equation (E.1). The worst credible scenario, bended or broken control surfaces, were recommended by the experts to consider when the gust velocity exceeds ±25 fps and its width is equal or smaller than 25 mean geometric chords of wing; Appendix E-3. Thus the FAA FAR Sections , , and define quantitative limits for the airframe-damaging turbulent gusts. Praskovsky (1982, 1983) found that a typical shape of turbulent gusts with high amplitudes is very close to sinusoidal as in Equation (E.1) in fully developed turbulent flows. Exhaust plumes from smoke stacks are characterized by very high Reynolds number and are certainly fully developed above about five orifice diameters. Praskovsky (1982, 1983) results also allowed relating the width of turbulent gusts with given amplitude to the standard deviation of turbulent fluctuation and the turbulent integral scale. There are no FAA regulations relating the aircraft vertical acceleration to characteristics of aerodynamic field. Such relations were established by analyzing dynamics of aircraft flying through vertical gusts. A simple analytical expression for the load factor on an aircraft flying through a discrete vertical gust was used for the analysis (Schmidt, 1998): t U () de λ nt = ωsinωt+ e cosωt ωsinωt 2 (G.1) 2g 1 + ( ωt) λ λ Here n is the incremental vertical load factor at the center of gravity due to the gust described by Equation (E.1), t = s / V a is penetration time into the gust, ω = 2πV a / L g is the gust frequency, V a is the airplane airspeed, λ = 2W a / (c Lα ρ S a ), c Lα, W a and S a are the aircraft lift curve slope, weight and wing reference area. One should note that the load factor on an aircraft is equal to the vertical acceleration experienced by aircrew and passengers. Therefore, Equation (G.1) relates the magnitude of incremental changes in the vertical acceleration to the same gust characteristics as for airframe damage: the gust amplitude U de and its width L g. It follows from Equation (G.1) that the acceleration is linearly proportional to the gust amplitude U de thus modeling with only the unity amplitude was needed. Two examples of modeling the vertical acceleration of Cessna 172 aircraft on the final approach as it flies across discrete vertical gusts are presented in Figure G1. The gust shape is given by Equation (E.1) and the modeling was executed at the unity amplitude U de = 1 and variable width L g from 10 m to 50 m. 50

51 Distribution of the vertical velocity across the gust is also shown in Figure G1. Cessna 172 aircraft on the final approach was modeled at the airspeed V a = 70 knots, weight W a = 2,000 lb, wing area S a = 172 sq ft and the lift curve slope c Lα = 4.6. The maximum vertical acceleration experienced by Cessna 172 aircraft flying through discrete vertical gusts with the unity amplitude and variable width is illustrated in Figure G2. The modeling results relate the magnitude of the impact on aircraft, the maximum vertical acceleration to quantitative characteristics of the affecting vertical gusts U de and L g. Maximum acceleration Maximum acceleration Figure G1: Vertical acceleration of Cessna 172 in g-units along discrete vertical gusts with the unity amplitude (top), and velocity distribution in the gusts (bottom) Left the gust width L g = 10 m, right - L g = 50 m The vertical gusts with the amplitude U de and width L g were related to the standard deviation of turbulent fluctuation in the vertical velocity and the turbulent integral scale in the plume-induced aerodynamic field. Equation (G.1) clearly indicates that the magnitude of the vertical acceleration depends on the aircraft and flight conditions through V a, c Lα, W a and S a. Figure G2: The maximum vertical acceleration for Cessna 172 aircraft at the unity gust amplitude It should be noted that simplified Equation (G.1) provides preliminary estimates for the aircraft vertical acceleration although those are sufficiently accurate for addressing the goals of the performed scientific analysis. The magnitudes of two impacts, the airframe damage and the aircraft vertical acceleration were found to be defined by the same aerodynamic characteristics: the amplitude of the turbulent vertical gusts and the turbulent integral scale. 51

52 Appendix H: Modeling the Plume-Induced Aerodynamic Field The dynamics of turbulent buoyant jets (forced plumes) is rather complex task involving a large number of governing parameters. Exact simulation of plumes from smoke stacks or cooling towers at realistic geometry, exhaust velocity and temperature and ambient conditions is virtually impossible, even with the most powerful computers. One thus needs a model as a simplified picture of reality. A model never contains all details of the real systems but could contain and describe accurately all features of the system that are relevant to the considered problem. Models are widely used for predicting the system behavior and identifying the best solutions for managing specific problems. Modeling of the plume-induced aerodynamic field is a natural choice of an adequate tool for analyzing the impact of exhaust plumes on aviation safety. A vast diversity of semi-empirical and empirical models for thermal and smoke plumes and turbulent buoyant jets has been developed around the world during the last century. The models differ by their complexity, flow parameters they describe, reliability, universality, etc. The key element of an effective modeling is to choose an appropriate one that provides all relevant characteristics of the system at the required scale and matches the complexity of the task. The following criteria were applied for selecting an adequate model for the present project: Suitability: It should provide all mean flow characteristics at special scales of the order of meters and temporal scales of the order of seconds Computational efficiency: A typical run time of a model on a personal computer should not exceed several minutes Universality: A model has to be able to describe a single jet and merging jets issuing from multiple sources over realistic range of stack and exhaust parameters and atmospheric conditions Flexibility: It should allow easy modification and enhancement Reliability: A model should be physically substantiated, tested under a wide range of modeling conditions, and verified with available high-quality experimental data Simplicity: It should have a relatively simple mathematical formulation A diverse body of potentially suitable models could be separated into two major groups: atmospheric dispersion models and aerodynamic models of turbulent buoyant jets and plumes. The atmospheric dispersion models (e.g., Sykes et al, 1997; Karachandani et al, 2000; and Godden et al, 1983) simulate dispersion of the air contaminants and are used to estimate or predict the downwind concentration of air pollutants emitted from sources such as industrial plants. The U.S Environmental Protection Agency (EPA) has developed or accepted different models ranging from relatively simple Gaussian plume dispersion models such as ISC3 (EPA, 1995) to more advanced and complicated ones like CTDMPLUS (EPA, 1989), ALOFT-PC (Walton et al, 1996), AERMOD (EPA 2004), and CALPUFF (Scire et al, 1997). An expanded list and brief description of atmospheric dispersion models used in the US and abroad could be found online 13 and comprehensive review and physical analysis of the models is given in, e.g., NIWAR (2004). Although a plume model is always embedded into atmospheric dispersion model and often provides estimation for the plume trajectory, mean velocity and temperature, the 13 E.g., and 52

53 dynamic characteristics are of secondary interest for the pollution problem. The key element of the models is in accurate description of pollutant s diffusion and its transport by atmospheric motions. The only physically relevant dynamic characteristics of a plume are its width and height of rise at varying atmospheric conditions. The majority of atmospheric dispersion models employ the classic Briggs approximations for the plume rise (Briggs 1965, 1975). The plume models used for modeling dispersion of pollutants in the atmosphere do not provide accurate description of plume-induced small-scale motions that are critical for the considered problem. For this reason, none of such plume models is adequate for analyzing the plume impact on aviation safety. On the contrary, aerodynamic models of turbulent buoyant jets and plumes are focused on details of small-scale dynamics. The models describe flows in jets and plumes issuing into unbounded stratified ambient environment under stagnant or steady ambient conditions. Such turbulent shear flows have been extensively studied in wind tunnel experiments. An abundance of experimental data makes modeling of turbulent buoyant jets and plumes especially attractive for scientists. There is a large body of models for turbulent jets and plumes with wide range of complexity. The simplest models provide algebraic expressions for averaged plume characteristics such as the mean velocity, temperature and concentration based on an analysis of equations of motion and experimental data; e.g., List (1982a) and references therein. Quite complicated models involving solution of Reynolds equations with multi-parametric turbulent closures in the form of partial differential equations represent another extreme; e.g., El Hayek (2004), Hossain and Rodi (1982), Muppidi et al (2007) and references therein. Although such models could describe averaged plume characteristics and parameters of turbulence, their numerical solution is too complicated for the present task. The integral models combine advantages of both approaches and provide a compromise between universality, reliability, accuracy and efficiency on the one hand and simplicity on the other; e.g. Morton et al (1956), Fan et al (1969), Pantokratoras (1998); Best et al (2003); Jirka (2004), Xiao et al (2009) and references therein. None of the existing integral models describes turbulence but the model s flexibility allows adding equations for any chosen turbulence characteristics. Analysis of models for turbulent buoyant jets and plumes has lead to the conclusion that integral methods are the most suitable for analyzing the plume impact on aircraft and hence on aviation safety. All integral models use the following simplifications. Boundary-layer approximation presumes that the rate of changes of plume parameters along the centerline is much smaller than that in the transverse direction. It greatly simplifies the Reynolds equation by neglecting the longitudinal derivatives. Specific profiles of plume characteristics in the transverse direction are presumed based on an analysis of equations of motion and experimental data. It reduces partial differential equations to ordinary differential equations. The turbulence closure is embedded into the model and is pronounced through empirical functions and constants. It allows easy tuning of integral models to experimental data at variable ambient conditions. From a large variety of integral models, the model by Jirka (2004, 2006) was chosen for describing the average characteristics of the plume-induced aerodynamic field in the present study. The model has been developed for predicting velocity, temperature, width, trajectory, and concentration of pollutants from single or multiple exhaust stacks. It is capable to handle rather 53

54 an arbitrary but steady vertical wind profile and covers a wide range of atmospheric conditions. The model is formulated for the conservation of mass, momentum, buoyancy, and scalar quantities in the turbulent jet flows. It employs an entrainment closure approach that distinguishes between separate contributions of transverse shear and azimuthal shear mechanisms, and contains a quadratic law turbulent drag force mechanism that follows from detailed experimental investigations of the dynamic of transverse jets into cross-flow. The model formulation for multiport buoyant jets includes several significant three-dimensional effects related to local merging processes from multiple jets. The model is validated in several stages. First, comparison with basic experimental data for five asymptotic, self-similar stages of buoyant jet flows such as the pure jet, the pure plume, the pure wake, the advected line puff and advected thermal, support the choice and magnitude of the turbulent closure coefficients contained in the entrainment formulation. Second, comparison with many types of non-equilibrium flows support the proposed transition function within the entrainment relationship, and also the role of the drag force in the jet dynamics. Third, a number of spatial limits of applicability have been determined beyond which the integral method becomes invalid due to its parabolic formulation. One terminology-related issue is noteworthy. In the contents of the current study, plume is defined as thermal updraft generally associated with exhaust from the smoke stacks of power generating facilities, industrial production facilities, or other systems that have an ability to release large amounts of pressurized or otherwise unstable air; AFS-420 (2006). Studies of smoke plumes began decades ago and were mainly motivated by the air quality and pollution problem. The issues of major interest in these studies were the plume rise and dispersion of pollutants by ambient winds. The plume dynamics in general and its rise in particular were mainly associated with buoyancy forces; see classic papers by Briggs (1965, 1975). In these studies and in aerodynamics in general, the plume is considered to be driven by density gradient at the orifice that produces the weight deficiency and pressure gradient. The latter accelerates the heated fluid vertically in the convection process; e.g., List (1982a, b). Completely different type of an aerodynamic flow is a turbulent jet that is driven by pressure drop at the orifice. The buoyant jets are defined as aerodynamic flows that are driven by both pressure drop and density gradient thus combining the dominant features of plumes and jets. Importance of the difference and reason for this comment is that dynamics of jets and plumes (in a narrow sense of the term plume ) are completely different. In particular, the axial vertical velocity at calm winds decays in jets as 1/z and in pure plumes as 1/z 1/3 (here z is a height above the stack). This study is centered on aviation safety that could be affected by pressurized exhaust air with typically very high exhaust velocity and relatively moderate temperature. Dynamics of such high-velocity plumes is that of buoyant jets rather than pure plumes. However, majority of existing analyses of plume impact on aviation safety utilize equations for buoyancy-driven plumes; e.g., Katestone (2007) and (2010). The major reason for selecting the Jirka (2004, 2006) model is that it describes reliably both buoyant jets and pure plumes. The chosen model is mathematically expressed as a system of ordinary differential equations. A computer code in MATLAB has been developed for solving the equations using the standard forth-order Runge - Kutta algorithm. The model and the codes were thoroughly tested by comparing simulations with experimental data for round jets and plumes (Chen and Rodi, 1980; Dimotakis et al, 1983; Wright, 1984) as well as with modeling by other techniques (Hossain and Rodi, 1982). 54

55 As was indicated in Section 2.2, the minimum set of turbulent parameters that are needed for estimating intensity, scales and probability of occurrence of the plume-induced vertical gusts includes standard deviation of the turbulent fluctuations in the vertical velocity, the turbulent integral scale of the fluctuations in the horizontal (along-the-flight-route) direction and the intermittency factor. None of the existing plume models provides these characteristics hence new turbulence model was needed to be developed. The developed model needed to be consistent with the integral model by Jirka (2004, 2006) that is chosen for modeling the mean flow. The new model has been developed that satisfies the above requirements. It is based on the field model for buoyant flows presented in Hossain and Rodi (1982). (Field models utilize partial differential equations for considered variables.) There were several reasons for choosing the Hossain and Rodi model as the starting point. First, it employs the most reliable and well-tested two parametric k ε field model. Hereafter k is the turbulent kinetic energy and ε is the rate of its dissipation. Second, the standard k ε model is specifically modified for describing reliably a broad range of buoyant flows. Third, the model has been thoroughly tested by comparing with a diverse body of available experimental data and its application to vertical buoyant jets and plumes has been carefully analyzed by Hossain and Rodi. Finally, and the most important, one can estimate the standard deviation of turbulent fluctuations and the turbulent integral scale with k and ε. To derive an integral model for k and ε from the field one, the transverse profiles of these characteristics were approximated in the Gaussian form similar to those for the mean characteristics in Jirka (2004, 2006). The profiles were substituted in the partial differential equations and integrated across the transverse cross-sections. These time-consuming although straightforward operations resulted in two ordinary differential equations for the values of k and ε on the plume centerline. Inevitable for any turbulence model empirical constants were partly taken from Hossain and Rodi (1982). In addition to Hossain and Rodi, experimental data on turbulent characteristics in round jets from the following sources and references therein have been used for estimating empirical constants in the model: Becker et al (1967), Wygnanskii and Fiedler (1969), Hinze (1975), Antoine et al (2001), Antonia and Zhao (2001), Argawal and Prasad (2003), and Burattini et al (2005). Figure H1: The amplitude of the plume-induced turbulent vertical gusts, ft/s. Left calm winds; middle - the ambient wind speed 2 knots; right 4 knots As an illustration, aerodynamic fields for exhaust stacks of three power plants were modeled at neutral stratification which represents the worst case scenario. The modeling parameters are listed in Table 2. An example of the mean velocity and temperature fields for the Fort Martin 55

56 Power Station at calm winds and neutral stratification is presented in Figure 3. As indicated in Section 2.5, an adequate turbulence modeling is necessary for evaluating potential plume impact on aviation safety including the amplitude of turbulent vertical gusts, characteristic scale of such gusts in the along-the-flight-track (horizontal) direction and the external intermittency factor. The developed turbulence model provides these characteristics which are illustrated below for the Fort Martin Power Station. Figure H1 (left) demonstrates that the total amplitude of the plume-induced vertical gusts (that is, a sum of the mean and turbulent vertical velocities) on the plume centerline is about twice larger that the mean vertical velocity in Figure 3. One can also see from comparison of two those figures that the width of a region with significant turbulent gusts is more than twice larger than that of significant mean velocities. It means that potentially hazardous impact of plume-induced turbulence on aircraft is more significant than that of the mean velocities. Figure H2: Spatial distributions of the turbulent integral scale in the horizontal direction, ft Left calm winds; middle - wind speed 2 knots; right 4 knots The utmost importance of plume-induced turbulence for analyzing the impact of plumes on aviation safety is further demonstrated in Figure H2. As explained in Appendix G, the hazardous impact is always associated with high-amplitude vertical gusts of relatively small length of about 25 mean geometric chords of wing or less. The conventionally considered plume radius is defined as that where the mean velocity drops to of its centerline value and is quite large, about 0.16 heights above the stack; e.g., Katestone (2007). On the other hand, the most intensive gusts are determined by turbulence (Figure 5) and their width is defined by the integral scale of the vertical turbulence velocity in the horizontal direction. The latter is much smaller than the plume radius and defines potentially hazardous vertical turbulent gusts with a characteristic size of an aircraft. The amplitude of vertical gusts as in Figure H1 and the turbulent integral scale as in Figure H2 determine severity of two considered hazardous plume impacts on aircraft and aircrew; Figure 5 in Section 2.6. The above statements are further demonstrated in Figure H3 showing dependence on the height of flow characteristics at the plume centerline at calm winds. All parameters in this figure are normalized by their values at the stack orifice. One can see in the top screen that temperature excess decays much faster than the centerline vertical velocity illustrating again that temperature does not affect significantly aircraft or aircrew. The mean velocity decays faster than the amplitude of turbulent gusts. Therefore, turbulent gusts could still affect aircraft when the mean velocity becomes very small. Another aspect of plume-induced turbulence is that the turbulent integral scale is much smaller than the plume radius. The utmost importance of this feature to the 56

57 problem in hand is illustrated in the bottom screen showing characteristic gradients for the mean and turbulent velocities. The load on aircraft and the vertical acceleration are mainly determined by the gradients in the vertical velocity rather than by velocities per se; Appendix G. One can see that turbulent gradients are about order in magnitude larger than those produced by mean velocities (the ordinate in this screen is in logarithmic scale). Hence turbulence plays the dominant role in the potentially hazardous impact of vertical plumes on aircraft and aircrew. Height AGL, ft Figure H3: Dependence of flow characteristics at the plume centerline on the height Top: blue mean vertical velocity, yellow mean temperature, red amplitude of the vertical turbulent gusts, turquoise - ratio of the turbulent integral scale to the plume radius Bottom: blue mean velocity gradient, red velocity gradient in the turbulent gusts The external intermittency factor is presented in Figure H4. It characterizes the probability of occurrence of the vertical turbulent gusts with the amplitude in Figure H1 in a given location. As explained in Section 2.4, this probability is necessary for evaluating the impact s likelihood. Figure H4: Spatial distributions of the external intermittency factor Left calm winds; middle - wind speed 2 knots; right 4 knots The presented results provide quantitative support to the statements in Section 2.5 that were illustrated qualitatively in Figure 4. The amplitude of turbulent vertical gusts is at least twice larger than the mean velocity 57

58 The width of strong turbulent gusts is about three times smaller than the plume radius Turbulent gusts have the largest velocity gradients and their impact on aircraft and aircrew is significantly more hazardous than that of mean velocities A plume area with strong turbulent gusts is more than twice wider than that with considerable mean velocities The statements substantiate the major finding of the accomplished analysis: the plume-induced turbulence is the only real cause of its potentially hazardous impact on aviation safety. Severity of the plume impact on aircraft and aircrew is determined by the turbulence intensity and integral scale, and likelihood of the impact depends on the intermittency factor. Thus comprehensive turbulence modeling is a vital component of a representative study of the impact of vertical plumes on aviation safety. 58

59 Appendix I: Evaluating the Plume-Induced Risk Following the FAA SRM process, the risk was assessed as the composite of predicted severity and likelihood of the potentially hazardous plume impacts in the worst credible system state; Section 2.1. The process is illustrated in Figure 6 which shows the risk level for the Cessna 172 vertical acceleration. It follows from Table 1 that severity differs for different impacts while the likelihood is the same for the airframe-damaging gusts and the vertical acceleration. Severity of these two impacts for the Fort Martin Power Station at variable ambient winds and neutral stratification is illustrated in Figures I1 and I2. Figure I1: Severity of the plume-induced airframe-damaging vertical gusts at variable winds. Red hazardous, white minimal ; left ambient wind 2 kts, center - 4 kts, right - 6 kts As explained in Appendix E-3, the vertical gust could cause airframe damage of aircraft certificated under the Parts 23 and 25 for takeoff, approach and landing configurations when the gust amplitude exceeds 25 fps and the width is equal to or smaller than 25 mean geometric chords of wing (the chord is 4.9 ft for Cessna 172). The criteria for the impact were applied as follows: the impact was hazardous when an amplitude of the vertical gusts (defined as a sum of the mean and turbulent vertical velocities) exceeded 25 fps and the turbulent integral scale was equal or less than 25 mean geometric chords of wing. Severity of the airframe-damaging vertical gusts is illustrated in Figure I1 and it could be either hazardous or minimal (Appendix E-6 and Table 1) hence the maps are binary red and white colors only. Figure I2: Severity of the plume-induced vertical acceleration for Cessna 172 aircraft at variable winds. Green minor, yellow major, red hazardous Left ambient wind 2 knots, center - 4 knots, right - 6 knots On the contrary, severity of the aircraft vertical acceleration could be from the minimal to hazardous dependent on the acceleration magnitude; Appendix E-7 and Table 1. Evaluation of 59

60 the vertical acceleration severity for Cessna 172 aircraft on the final approach at an ambient wind 1 knot is described in Section 2.4 and illustrated in Figure 5. Figure I2 illustrates severity of the vertical acceleration for Cessna 172 at higher wind speeds. It is seen in Figures I1 and I2 that the height of regions with considerable severity for both impacts decreases systematically as an ambient wind speed increases. The width of the regions grows with the ambient winds up to 5 knots and then decreases. The detrimental impact on aviation safety of the quite typical considered plumes at the ambient wind speed above 5 knots is insignificant compare to that at the speed below 5 knots. It means that only low-speed ambient winds might be considered in plume studies. Figures I1 and I2 also demonstrate that spatial dimensions of the plume-induced region with hazardous airframe-damaging gusts are much smaller than those of the region with the vertical acceleration of hazardous severity. Therefore the aircraft vertical acceleration is the dominant of the two considered impacts and it is used for mapping the plume-induced risk level. Figure I3: Hourly temperatures versus wind speeds As explained in Section 2.4, the impact likelihood (that is, the total probability of the impact p o in Table 4) is defined in the presented prototype as a product of the intermittency factor, the presumption-dependent probabilities and the climatology-defined probability of considered weather conditions. Few examples of modeled distributions of the intermittency factor are shown in Figure H4. Figure I4: Likelihood of the plume-induced impacts at variable ambient wind speed White extremely improbable, green extremely remote, yellow remote Left - calm winds, center ambient wind 2 knots, right - 4 knots 60

61 A scatter plot of hourly wind speeds and temperatures for the year 2005 at the nearest to the Fort martin Power Station location with measurements at heights up to several thousand feet is shown in Figure I3. As demonstrated above, the detrimental impact of the considered plume on aviation safety is significant at the ambient wind speed below 5 knots hence only data for the speed below 8 knots is shown in Figure I3. The presumption-dependent probabilities of and are introduced in Section 2.4. The probability of modeled ambient conditions was estimated with the data in Figure I3 for an ambient temperatures 22 F; Table 2. The impact likelihood for the same ambient wind speeds as in Figure H4 is presented in Figure I4. As illustrated in Figure 6, the risk was estimated as a composite of the impact s severity and likelihood combined in accordance with Table 3. More risk maps for the vertical acceleration of Cessna 172 aircraft on the final approach at variable ambient winds are presented in Figure I5. Figure I5: Risk level for the vertical acceleration of Cessna 172 aircraft at variable winds Green low risk, yellow medium risk, red high risk Left - calm winds, center ambient wind 2 knots, right - 4 knots It should be noted that an ambient wind was always modeled at fixed direction from the left in all figures. Symmetric plume is considered in the present study and the left-half risk levels in the central and right screens in Figure I5 as well as in Figure 6 are merely symmetric images of the right halves. Figure I6: Composite map of the plume-induced risk for Cessna 172 aircraft on final approach Green low risk, yellow medium risk, red high risk The composite map for plume-induced risk for aircraft and aircrew at all considered ambient winds is presented in Figure I6. One can see that exhaust stack of the Fort Martin Power Station near the Morgantown airport creates the high-risk region for Cessna 172 aircraft on the final approach with a height of 1,810 ft AGL and a width of 330 ft. 61