EFFECT OF PIPE ROUTING ON TURBINE FLOW METER READINGS AND SUBSEQUENT DISCREPANCIES IN CONDENSATE LOADING AT BANGORA GAS FIELD

International Journal of Chemical & Petrochemical Technology (IJCPT) ISSN 77-4807 Vol., Issue, Jun 01, 41-54 TJPRC Pvt. Ltd. EFFECT OF PIPE ROUTING ON TURBINE FLOW METER READINGS AND SUBSEQUENT DISCREPANCIES IN CONDENSATE LOADING AT BANGORA GAS FIELD MOHAMMAD FARZID HASAN Production Specialist, Tullow Bangladesh Limited, Dhaka-11, Bangladesh ABSTRACT In this Q1 of 01 there are eighty one producing wells of natural gas in Bangladesh among which four belong to the Bangora Gas Field located at Muradnagar, Comilla. The production rate from this field rounds about 105 MMscfd of natural gas along with 0 bbl condensate per day. On an average five to six tankers are loaded everyday at Bangora Central Processing Facility for exporting the produced condensate. This loading purpose is served with two centrifugal pumps (P-640A and P-640B) which are supposed to be used alternately. Although the pumps are of the same model, P- 640B loads each tanker with about 148 Liter of extra condensate according to the level change of the storage tanks (S-710 and S-715). Besides, loading from S-715 causes somewhat overloading. Finding out the core cause(s) of this event of overloading and then making out the most feasible solution(s) were the main objectives of this study. Based upon physical observation and mathematical analysis, the additional loads induced by the bends and the extra length of pipe present in the flow path regarding P-640B and S-715 have been found responsible for the concerned inconsistency. According to the theory of pump characteristics, it has been shown that the operating point of P-640B has been shifted from the desired zone due to those extra loads resulting in a flow rate lower than that delivered by P-640A. Then the relation between these lower flow rates and erroneous readings given by the flow meter has been revealed utilizing the working mechanism of the pertinent turbine flow meter. Finally some plausible ways out have been suggested and upon taking initiatives accordingly the company has already been able to enhance its export rate of condensate by 10 Liter per month (about 0.86% of monthly export) even after maintaining the same production rate as before. KEYWORDS: Bangora Gas Field, Centrifugal Pump, Condensate Loading, Frictional Energy Losses, Pipe Routing, Turbine Flow Meter NOMENCLATURE Symbol Meaning Unit A Cross-sectional area of the flow line ft D Pipe diameter ft D e Equivalent diameter of a flow path ft f Fanning friction factor Dimensionless f p Pulse frequency (number of pulses per unit time) pulses/min. g Acceleration due to gravity ft/s g c A constant force-mass conversion factor (lb m.ft)/(s.lb f ) H Height of a condensate column ft h A Energy added to the fluid with a mechanical device such as a pump ft h L Energy losses from the system due to friction in pipes or minor ft losses due to valves, fittings and bends K K-factor of a turbine flow meter pulses/l

4 Mohammad Farzid Hasan L Length of a flow path (e.g. a pipe) ft N Impeller speed (revolution per unit time) of a pump rpm P L Horse power required (i.e. lost) to compensate for the frictional hp loss P Pressure under which the fluid exists lb f /ft Q Volumetric flow rate ft /s or L/min. or US GPM Re Reynolds number Dimensionless t Time required to load each tanker min. v Fluid velocity through a pipe ft/s z Elevation head (potential energy per pound of fluid) ft p Pressure drop of the fluid due to frictional head loss lb f /ft Equivalent roughness of pipe γ Specific weight (amount of weight per unit volume of a substance) lb f /ft η Pump efficiency % ω INTRODUCTION Density of condensate lb m /ft Kinematic viscosity of condensate Angular velocity (revolution per unit time) of the rotor of a turbine flow meter ft ft /s or cst Upon declaration as a commercial discovery in December 00, the Bangora Gas Field commenced its production of natural gas in May 006 with a single well at a rate of 50 MMscfd. Afterwards three more wells were put on stream and gradually the production rate crept up to 105 MMscfd within the next two years. Since the day of first gas at export, the field has been undergoing different development and facility upgrading projects of which the commissioning of the Hydrocarbon Dew Point (HCDP) unit in Q1 of 010 is the latest so far. After putting this state of the art technology for separating liquids from natural gas online on March 17, 010 the rate of condensate production has been enhanced by 00% comparing to the previous glycol dehydration system and since then about 0 bbl condensate has been being produced per day. As there is no fractionation facility in Bangora, the produced condensate needs to be exported regularly. Hence condensate loading is a daily routine operation at Bangora Central Processing Facility for which there are two centrifugal pumps tagged as P-640A And P-640B. The pumps are supposed to be operated on a days on / days off basis, that is, either pump would be used for two consecutive days and then kept in rest for the next two days when the other one would be used for loading condensate. Each tanker is expected to be loaded with 9000 L condensate which is maintained according to the reading displayed by a turbine flow meter (FQI-640) located downstream of the pumps. Although both the pumps are of the same model in every design aspect according to the vendor, they show the following discrepancies (based on the condensate loading data of five random days summarized in appendix A) while functioning in practice Table 1: Distinguishing Features of P-640A and P-640B Distinguishing Features on an Average to Load Each Tanker P-640A P-640B Loading time (min.) 7 8 Level change of the storage tank (%) 7.45 7.57 Total amount of condensate delivered (L) 900.58 9147.59 Volumetric flow rate or loading rate (L/min.) 186.08 114.45 rpm

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 4 Beyond the pump functions, another incongruity has been observed with the storage tanks (S-710 and S-715) which are listed bellow Table : Distinguishing Features of Different Pump/Storage Tank Combinations Distinguishing Features on an Average to Load Each Tanker S-710 S-715 Level change of the storage tank in case of P-640A (%) 7.8 7.5 Level change of the storage tank in case of P-640B (%) 7.51 7.6 Total amount of condensate delivered in case of P-640A (L) 8917.99 9087.17 Total amount of condensate delivered in case of P-640B (L) 9075.08 908.01 Observation says that the amount of condensate loaded to each tanker varies significantly with the combination of pumps and storage tanks. But ultimately it is evident that, on an average, the amount of condensate loaded with P-640A remains within the expected range while P-640B loads about 148 L extra condensate to each of the tankers for which the company is not being paid. Figure 1: Existing Layout of the Condensate Loading System. (Figure is not in Scale) Although both of the pumps and both of the storage tanks are identical with their own basics, the condensate flow paths (from storage tank to tanker) for all the pump/tank combinations are not alike. In figure-1, six points (1 to 6) have been identified which can be termed as the points of inconsistency, because they are the only differences that the pipe routings have and consequently supposed to be responsible for the inconsistent loading phenomena.

44 Mohammad Farzid Hasan The effects of disparities in condensate flow paths upon the overall loading system have been analyzed here by approaching the theory of centrifugal pump characteristics and the working principle of a turbine flow meter. After having the root cause found out, some suggestions have also been made in order to getting over with the problem. ANALYSIS OF CENTRIFUGAL PUMP CHARACTERISTICS AND FLOW DYNAMICS The manner in which a centrifugal pump operates depends not only on the pump performance characteristics, but also on the characteristics of the system in which it has to operate. For the pumps under consideration, figure- shows the pump operating characteristics (head versus volumetric flow rate) for a selected speed of operation, usually close to the speed that gives optimum efficiency. The figure also shows the system characteristic curve, that is, the pumping head required for a particular system plotted as a function of the volume flow rate. The point of intersection of these two curves refers to the true operating point of the pump in this system which determines how much flow is actually delivered into the system. The pump will automatically seek this operating point after it is energized [1]. The pump characteristic curve depends solely on the design of the pump, while the system curve represents a particular system associated with that pump. Therefore, the pump characteristic curve is fixed for a particular pump while different systems designed with different loads would yield different system curves for that specific pump. The operating point shown in figure- advises the pump to be operated in such a system that it yields a flow rate of 51 US GPM for achieving its maximum efficiency []. But in practice, the average flow rates for P-640A, P-640B and a combination of P-640B & S-715 have been found to be 9.75, 0.07 and 96.65 US GPM respectively. So it is obvious that the operating point has been shifted to somewhat left from that determined by the manufacture. The observed flow rates imply that the system curve for P-640A must be steeper than that referring to 51 US GPM and the curve for P- 640B would be steeper than that for P-640A while that for the combination of P-640B & S-715 is the steepest of all concerned here. This scenario is well depicted by figure- where the practical system curves along with their corresponding operating points have been interpolated on the plot provided by the manufacturer. Figure : Shift of Pump Operating Points Due to Steeper System Curves. (Courtesy: Grundfos Canada)

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 45 The extent of steepness of the system curve depends on the loads existing in the system. At a particular flow rate, the total head developed by a pump would be higher for a system associated with greater loads than that for a system with fewer loads. That means, the system in which P-640B operates contains some extra loads those are absent in the flow system regarding P-640A. Thus figure- provides an opportunity of cross checking the phenomena of increasing loads and decreasing flow rates. If the decrease of flow rates, which has been observed practically, is considered to be true then the system curves need to be steeper that correspond to the existence of greater loads. On the other hand, if the existence of greater loads, which have been identified in figure-1, is considered true then also the system curves need to be steeper (as more head would be developed in order to undertaking more loads) that correspond to the decreased flow rates. Therefore it can be incurred that both the phenomena must be true simultaneously. Points, 4, 5 and 6 (figure-1) have been identified as the sources of the extra loads undertaken by P-640B since a portion of the total energy that the pump adds to the fluid is lost there due to frictional effects. The fluid delivered by P- 640B would have a total energy less than that delivered by P-640A by an amount equal to what is lost at these points. It can be explained by the general energy equation [] that can be obtained simply by balancing the respective mechanical-energy terms between sections 1 and in figure-1.1 on the basis of Bernoulli s theorem of energy conservation p1 v1 z1 h g A h L p z v g (1) Where, p z v g h A h L Pressure head (energy per unit weight stored in the fluid by virtue of the pressure under which the fluid exists) Elevation head (potential energy per pound of fluid) Velocity head (kinetic energy per pound of fluid) that refers to the flow rate Energy added to the fluid with a mechanical device such as a pump Energy losses from the system due to friction in pipes or minor losses due to valves, fittings and bends In the above equation, subscripts 1 and refer to sections 1 and in figure 1 respectively. Although p 1 z 1 v 1 g is equal for both the pumps, both of h A and h L are different for them. Since the operating points of the pumps have been shifted to different extents, they would be operating with different efficiencies and the empirical trend of efficiency curve confirms that P-640B would be less efficient at its operation than P-640A. So h A added by P-640B must be less than that added by P-640A. On the other hand, h L must be greater for P-640B as the bends denoted by points, 4, 5 and 6 do not exist in the flow path associated with P-640A. As a result, 640B must be less than that for P-640A. p z v g for P- This means that the fluid delivered by P-640B has less energy than that delivered by P-640A. Since this energy corresponds to the driving force that causes the fluid to flow, so logically, P-640B should deliver condensate at a lower rate than that of P-640A and this is what happens in practice. During loading from S-715, the condensate has to travel an extra length of pipe (4.85 ft) between points 1 and (figure-1) where lays an additional frictional head loss that eventually increases the value of h L in equation (1). Therefore,

46 Mohammad Farzid Hasan p the combination of P-640B & S-715 yields the least value of z v g resulting in the lowest of all the flow rates encountered here. The frictional energy losses considered here have been calculated in detail and presented in appendix B. EFFECT OF FLOW DYNAMICS ON TURBINE FLOW METER READINGS The working principle of a turbine flow meter which is available in literature [4] says that fluid flow through the meter impinges upon the turbine blades which are free to rotate about an axis along the center line of the turbine housing. The angular velocity (rotational speed) of the turbine rotor is directly proportional to the volumetric flow rate of the fluid and the relation can be expressed mathematically as follows ω Q ω K 1 Q ω Q K 1 () Where, ω Angular velocity (revolution per unit time) of the rotor Q Volumetric flow rate of the fluid flowing through the meter K 1 Proportionality constant (a function of geometrical dimensions, flow velocity and kinematic viscosity) There are several ways to detect the rotational speed of the rotor. The most common detection methods are mechanical detection and magnetic detection. Mechanical detection of the rotor speed is measured by transferring the rotor speed through the rotor axis and via gears to a mechanical counter. During magnetic detection an impulse is measured by disrupting a magnetic field every time a designated point on the rotor, for example the rotor blades, passes a measuring point. Thus the change in magnetic flux induces a voltage in the pickup and the frequency of the sinusoidal alteration corresponds to the revolution of the rotor per unit time. After amplifying and transforming the pickup signal a square wave pulse signal is available. The number of pulses, either mechanical or magnetic, per unit time depends on the rotor speed in direct proportion which can be expressed mathematically as follows f p ω f p K ω ω f p () K Where, f p Pulse frequency (number of pulses per unit time) K Proportionality constant that depends upon the transfer mechanism of mechanical dynamics into electromagnetic signals Combining equations () and (), Taking K 1 K K, Q f p K K 1 Q K f p (4) K in equation (4) is commercially known as K-factor which is a characteristic parameter for calibrating the respective turbine meter in respective situation. Dimensional analysis of equation (4) yields the dimension of K to be the

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 47 number of pulses detected per unit volume of fluid passed through the meter. Therefore, for convenience, if the units of Q and f p are taken as L/min. and pulses/min., then the unit of K would be pulses/l. During a typical calibration the volume of a tank determined as accurate as ± 0.01 % is filled with a constant flow passing through the turbine flow meter. The output pulses of the turbine flow meter are added electronically and calculated for a volume unit to receive the K-factor in pulses per liter. Strictly speaking this K-factor applies only for a certain flow rate. For the application of turbine flow meters, however, it is necessary to know the linear measuring range, i. e., the range with a constant K-factor. This range is determined by successively repeating the filling process at constant frequency intervals and different flow rates. These individual measurements will result in the calibration curve from which the average or mean K-factor can be drawn. The mean K-factor applies for the entire measuring range [5]. Figure : Typical Calibration Curve (K-Factor vs. Flow Rate Plot) for a Turbine Flow Meter (Courtesy: Küppers Elektromechanik GmbH) When the flow rate varies within the range of constant K-factor, the pulse frequency varies proportionately according to equation (4) and the meter gives the accurate reading. But if the flow rate alters in a range in which K-factor varies significantly, the meter would give an erroneous reading as it does not encounter the variation of K-factor. If the K- factor can be adjusted for the changed flow rate, the meter should give accurate readings again. For example, the condensate loading system can be considered where liquid condensate is pumped from the storage tank to the tanker with a certain flow rate. The amount of condensate loaded, which is nothing but the flow rate (Q) multiplied by time (t) for which the flow continues, that is, the pump is kept running, is measured with the aid of a turbine flow meter located down stream of the pump. Now if the flow rate (Q) is reduced, pulse frequency (f p ) would be reduced in a direct proportion as per equation (4). So the pump would be required to run for some more time to load the fixed amount of condensate. This phenomenon can be demonstrated simply using equation (4) Qt f p t (5) K Where, t Time required to load each tanker So, Qt amount (volume) of condensate loaded to each tanker Therefore, t must be increased to keep Qt fixed when Q is decreased. But the extents of this increase and decrease may not be proportional if K deviates from its constant range. For instance, if K is reduced with a reduction in Q (as well as f p ) according to the pertinent calibration curve (K-factor vs. Q curve), which often happens naturally beyond the linear range of K (figure-), slight increase of t would result in a fixed amount of condensate, which is equal to (f p /K)t according to equation (5), as reduction of Q or f p has already been made up to some extent by the reduction of K. But if K is not

48 Mohammad Farzid Hasan reduced, although it needs to be, then increase of t should be large enough to compensate for the reduction of f p to full extent. This means that the pump would be required to run for some extra unnecessary time resulting in an over amount of condensate loaded to the tanker. PROSPECTIVE SOLUTIONS Solution-1 The operating point regarding P-640B can be shifted to a suitable extent by changing the pump characteristic curve either by changing its speed (rpm) of operation or by replacing it with a new one having different performance characteristics [1]. An estimated increase of the impeller speed or motor horse power could serve the purpose concerned here. For example, the following equation [] can be used to determine the necessary rpm of P-640B: Q1 N 1 (6) Q N Here Q and N denote the flow rate and the impeller speed (rpm) while suffixes 1 and refer to their present and required values respectively. Since equal flow rates are expected from both the pumps, putting the value of the average flow rate delivered by P-640A for Q in the above equation would yield the required impeller speed for P-640B in terms of N. Solution- The volumetric flow rate of condensate has been observed to vary from 96.65 to 9.75 US GPM. So if the existing flow meter is replaced with one having the range of constant K-factor within 50 ~ 50 US GPM, appropriate readings should be obtained in either case. Solution- Condensate can be loaded on the basis of level change of the storage tanks. Certain extent of tank level would be lowered on delivering a certain amount of condensate. So installing a level transmitter in the pump operating zone can solve the problem. Solution-4 If a new design of condensate loading system is considered, each of the flow paths for either pump and either storage tank must be made identical (as shown in figure-4). If so is done then one appropriate calibration of the flow meter would suffice either combination of the pumps and the storage tanks.

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 49 Figure 4: Symmetrical Orientation of Either Pump/Storage Tank Combination. (Figure is not in Scale) CONCLUSIONS It is evident from physical observation and mathematical analysis, unlike that of P-640A the motor of P-640B has to undertake some extra loads induced by the respective flow path of condensate. As a result, the operating point of P-640B shifts in a manner that its delivery rate becomes lower than that of P-640A which ultimately makes the flow meter reading erroneous enough to result in a significant donation of money in terms of condensate. Having notified the entire scenario, the authority has halted the operation of P-640B immediately on an ad hoc basis and requested the maintenance department to find out a suitable flow meter for serving the purpose accurately. Moreover, the management is going to consider the modified design of pipe routing in its next phase resource expansion and facility upgrading project in near future. However, it goes without saying that in the following month after isolating P- 640B the company could export 10 Liter of extra condensate, provided that the production rate was the same as before.

50 Mohammad Farzid Hasan REFERENCES 1. Franzini, J. B., and Finnemore, E. J., Fluid Mechanics with Engineering Applications, International Edition, The McGraw-Hill Companies, Inc., 1997. 9614075 CR 90-1 Centrifugal Pumps, User Manual, Grundfos Canada, 005. Mott, R. L., Applied Fluid Mechanics, 4 th ed., Prentice Hall, New Jersey, 1994 4. Stoltenkamp, P. W., Dynamics of turbine flow meters, PhD Thesis, Eindhoven University of Technology, 007 5. HM Series Turbine Flow Meters, User Manual, Küppers Elektromechanik GmbH, 005 6. Peters, M. S., and Timmerhaus, K. D., Plant Design and Economics for Chemical Engineers, 4 th ed., McGraw- Hill, Inc., Singapore, 1991 APPENDICES Appendix A. Summary of Condensate Loading Data Table A1: Summary of Condensate Loading Data of Five Random Days when P-640A was used Date 6/6/011 10/6/011 16/06/011 0/09/011 0/10/011 Storage Tank S-710 7.7 S-715 7.58 S-710 7.8 S-715 7.45 S-710 7.4 S-715 7.59 S-710 7.8 S-715 7.5 S-710 7.5 S-715 7.49 Average Level Change for Loading Each Tanker (%) 7.47 7.4 7.5 7.45 7.4 7.45 Average Amount of Condensate Delivered to Each Tanker (L) 7.45% 900.58 7.6bbl 1% 159L 1bbl Since the average time needed to load each tanker is 7 min. for P-640A, the respective loading rate 900.58L 7 min. 186.08 L/min. 9.75 US GPM Table A: Summary of Condensate Loading Data of Five Random Days when P-640B was used Date Storage Tank Average Level Change for Loading Each Tanker (%) Average Amount of Condensate Delivered to Each Tanker (L) 1/09/011 /09/011 5/09/011 6/09/011 9/09/011 S-710 7.5 S-715 7.68 S-710 7.51 S-715 7.59 S-710 7.51 S-715 7.54 S-710 7.54 S-715 7.71 S-710 7.48 S-715 7.6 7.6 7.55 7.5 7.6 7.54 7.57 7.57% 9147.59 7.6bbl 1% 159L 1bbl

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 51 Since the average time needed to load each tanker is 8 min. for P-640B, the respective loading rate 9147.59L 8 min. 114.45 L/min. 0.07 US GPM It should be noted that each of the storage tanks has a capacity of 760 bbl. Appendix B. Detailed Calculation of Frictional Energy Losses When P-640B is used for loading the tankers with condensate, the frictional head losses at points, 4, 5 and 6 (figure-1) can be calculated as follows [6]: Basis: Loading each tanker with 9147.59 L condensate in 8 min. Volumetric flow rate, Q 114.45 L/min. Referring to points and 4, 9147. 59L 1000cm 8min. 1L 0.67 ft /s 1min. (0.08) ft 60s 1cm Cross-sectional area of the flow line, A D 4 [D Pipe diameter 6 in. 0.5 ft]. 1416 0. 5 4 ft 0.196 ft Q 0. 67 Velocity of condensate flow, v A 0. 196.49 ft/s Kinematic viscosity of condensate (according to the resulting data of condensate analysis conducted at PMRE department, BUET), 0.5 cst 0.5cm /s (0.08) ft [1 St 100 cst 1cm /s] 100 1cm 5.79 10 6 ft /s Reynolds number, Re D v 0. 5. 49 6 5. 7910 1879.54 Taking equivalent roughness ( ) for new pipes of cast iron as 0.00085 ft, Relative roughness, /D 0.0017 So from Moody diagram, Fanning friction factor, f 0.006

5 Mohammad Farzid Hasan Frictional head loss at point (for a standard tee used as elbow, entering branch), h L f v g D L e 0. 006 (.49) 90. 17 [For a standard tee (used as elbow, entering branch), L e /D 90] 0.95 ft Frictional head loss at point 4 (for a 90 elbow of standard radius), h L4 f v g D L e 0. 006 (.49) 0.140 ft [For a 90 elbow of standard radius, L e /D ]. 17 Referring to points 5 and 6, Cross-sectional area of the flow line, A D 4. 1416 0. 4 ft [D Pipe diameter 4 in. 0. ft] 0.086 ft Q 0. 67 Velocity of condensate flow, v A 0. 086 7.814 ft/s Reynolds number, Re D v 0. 7. 814 6 5. 7910 47986.50 Taking equivalent roughness ( ) for new pipes of cast iron as 0.00085 ft, Relative roughness, /D 0.006 So from Moody diagram, Fanning friction factor, f 0.006 Frictional head loss at point 5 (for a 90 elbow of standard radius), f v L h L5 g D e 0. 006 (7.814). 17 [For a 90 elbow of standard radius, L e /D ] 0.79 ft Frictional head loss at point 6 (for a standard tee used as elbow, entering branch), f v L h L6 g D e 0. 006 (7.814) 90. 17 [For a standard tee (used as elbow, entering branch), L e /D 90].050 ft Total frictional head loss due to the four bends, (h L ) bends h L + h L4 + h L5 + h L6 (0.95 + 0.140 + 0.79 +.050) ft.14 ft

Effect of Pipe Routing on Turbine Flow Meter Readings and Subsequent Discrepancies in Condensate Loading at Bangora Gas Field 5 Density of condensate (according to the resulting data of condensate analysis conducted at PMRE department, BUET), 0.7916g cm.lb 1cm m 49.5 lb m /ft 1000g (0.08) ft So, the pressure exerted by a 1 ft column of condensate H lb m lb f 1 ft 49.5 1 ft lb m 49.5 lb f /ft 49.5lb f /ft Pressure drop due to frictional head loss at bends, ( p) bends. 14ft 16.55 lb f /ft 1ft (condensate) g g c Horse power required (i.e. lost) to compensate for the total frictional head loss due to bends, (P L ) bends ( p) bends 550η 16. 55lb ft 0.40 hp f Q 0.67ft s s (hp) 550ft.lb f 1 0.50 [η Pump efficiency 50%] It should be noted that since the pump is not running at the rated operating point, its efficiency must be lower than 66.8% that was determined by the manufacturer. That is why it has been assumed to be 50% here. Figure-1 shows that in the case of S-715, the liquid (condensate) experiences a 90 standard elbow at point 1 and a standard tee (used as elbow, entering run) at point in its flow path for which the values of L e /D are and 60 respectively, that is, 9 in total. But in the case of S-710, the tee at point is used as an elbow with flow through the branch for which L e /D 90. So the combined frictional effects at these points can be considered to be identical for both the tanks. However, during loading from S-715, the condensate has to travel extra 4.85 ft of pipe length between points 1 and before reaching the pump and the regarding frictional loss, when P-640B is used, can be calculated as follows: Basis: Loading each tanker with 908.01 L condensate in 8.0 min. 908. 01L 1000cm 1min. (0.08) ft Volumetric flow rate, Q 8. 0min. 1L 60s 1cm 0.660 ft /s Q 0. 660 Velocity of condensate flow, v A 0. 196.70 ft/s Reynolds number, Re D v 0. 5. 7 6 5. 7910 109.64 Taking equivalent roughness ( ) for new pipes of cast iron as 0.00085 ft, Relative roughness, /D 0.0017 So from Moody diagram, Fanning friction factor, f 0.0065

54 Mohammad Farzid Hasan Frictional head loss due to the extra length of pipe, (h L ) pipe f v g D L 0.0065 (.7) 4.85.17 0.5 0.8 ft Pressure drop due to frictional head loss along the extra length of pipe, ( p) pipe 49.5lb f /ft 0. 8ft 11.5 lb f /ft 1ft (condensate) Horse power required to (i.e. lost) compensate for the frictional head loss due to the extra length of pipe, (P L ) pipe ( p) pipe 550η 11.5lb ft f Q 0.660ft s (hp) 1 [η Pump efficiency 50%] s 550ft.lb 0.50 f 0.0 hp Therefore, when P-640B is used for loading condensate from S-715, the total frictional head loss due to the bends and the extra length of pipe, (h L ) total (h L ) bends + (h L ) pipe (.14 + 0.8) ft.54 ft And the total horse power required (i.e. lost) to compensate for this total head loss due to friction, (P L ) total (P L ) bends + (P L ) pipe (0.40 + 0.0) hp 0.4 hp