Area-wide dynamic line ratings based on weather measurements

Transcription

1 SESSION B2-106 Area-wide dynamic line ratings based on weather measurements R. PUFFER 1), M. SCHMALE 2), B. RUSEK 3), C. NEUMANN 3), M. SCHEUFEN 1) 1) RWTH Aachen University, Aachen, Germany, 2) TenneT TSO GmbH, Bayreuth, Germany, 3) Amprion GmbH, Dortmund, Germany SUMMARY The load of overhead lines has increased due to raised transmission of electrical energy within Europe as well as in the consequence of a growing feed-in from regenerative energy resources. The transmission capacities of overhead lines (OHL) are limited and already in many cases the bottlenecks which restrict the power flows. Dynamic rating is a measure to increase the ampacity of overhead lines depending on the actual weather situation. In the first section this paper focuses on the description and evaluation of different methods and tools for OHL ampacity determination. The conductor temperature is measured directly with a sensor as well as determined indirectly using a weather station close to the line and commercially available weather data. Different scenarios where the measured wind speed is higher than the wind at the line due to shadowing effects are taken into account. The results show that the conductor temperature determination from weather data may be done with a maximal adequacy of about 5 K in the conductor temperature range up to 50 C. This fact means that the dynamic line rating is applicable but some safety margins have to be implemented. It is also shown that investigations on the thermal behaviour at low wind speeds seem to be necessary. In the second section the focus is on the implementation of a dynamic rating system into the 380 kv transmission grid and the necessary measures to increase the ampacity up to 3,150 A. This dynamic rating system uses the ambient temperature and the wind speed to determine the dynamic ampacity of the line. Retrofitting in substations such as replacing circuit breakers and current transformers as well as measures at overhead lines such as inspecting joints and raising towers had to be taken. Dynamic aspects of the transmission system are discussed. The protection concept was revised in order to allow a quick location of faults even in the case of a single technical component failure. The operational experience using the dynamic rating system proves that it fits very well with the existing control centre s techlogy. The distinct increase in dynamic current rating compared to static rating verifies the technical and ecomical capability. KEYWORDS Ampacity, monitoring system, up rate, overhead line, thermal limit, meteorological data, real time thermal rating, dynamic rating puffer@ifht.rwth-aachen.de 1

2 1. Introduction To enable the ongoing connection of regenerative energy sources in a timely manner, there is a need for flexible measures to better utilise the existing transmission grid. One of those measures is dynamic rating of overhead lines. Based on the thermal balance of the conductor it can be determined that the meteorological conditions (wind speed, angle of oncoming flow of wind and ambient air temperature) affect the temperature of the line conductors considerably. With certain combinations of these meteorological factors, the influence results in a cooling effect, thus much more heat caused by the current flow can be generated without exceeding the maximum operating temperature. This means, that an overhead line can be loaded more intensively (above minal current). This effect can also be used for a temporary increase of the line capacity (dynamic line rating). Thus it is possible to increase the ampacity of the line and at the same time to maintain the maximum allowable values for the conductor temperature and the clearance to ground or objects. This paper is divided into two main parts. The first section focuses on the methods for determination of the conductor ampacity. For this purpose, the three methods (one direct and two indirect) will be compared in terms of adequacy. The second section focuses on the measures necessary to enable a retrofitting of existing 380 kv overhead lines in order to prepare them for dynamic rating. The most significant among these measures were the introduction of dynamic rating, essential actions in the overhead line sections like rising towers and inspection and replacement of primary technical equipment in the substations. Further the ampacity increases of OHL operated with dynamic rating are presented. 2. Comparison of methods for ampacity determination 2.1 Principal methods and tools For dynamic line rating, the conductor temperature can be used. The interesting questions are how to acquire the temperature, which adequacy is necessary and which monitoring system is the most suitable from the ecomical point of view. In order to answer these questions, a pilot installation of two monitoring systems at Amprion GmbH (German TSO) was established (figure 1). The first system measures directly the temperature of the conductor by means of two passive surface acoustic wave (SAW) sensors SAW 1 and SAW 2 located in two conductors of one circuit. The second system (ls - local station) bases on an indirect method using a thermal model for the calculation of conductor temperature under consideration of the meteorological conditions. In this case, a weather station is placed in the vicinity of the conductor. The cost of the pilot systems and the technical effort was quite high. Taking into account the actual costs Fig. 1 Pilot installation of monitoring systems with SAW sensors and local weather station of the measuring devices, the effort for the installation and the necessary line outages the installation of such systems throughout the network is t ecomical. The application of weather stations similar to (ls) does t require de-energising of the line and the installation cost expenditure is significantly lower. However, the necessity to equip all lines with weather stations still remains. The simplest solution is to use already available weather stations, which are located in the vicinity of the monitored line, and to project the measured meteorological conditions on the line. Therefore, the pilot project was extended by an additional indirect method (rs - remote station) which calculates the conductor temperature using the commercially available meteorological data. The aims of the project are to define the adequacy of conductor temperature determination from the meteorological data, provided by the local and remote weather station and to find the most suitable system from ecomical point of view. puffer@ifht.rwth-aachen.de 2

3 2.2 The thermal model of the conductor For the indirect method the CIGRE-thermal model of conductor [2] is used. The principle of the model is well kwn; the thermal balance of the line, i. e. introduced heat and dissipated heat, must be equal at any time (equation 1). P ( T ) + P + P + P = P ( T ) + P ( T ) + P (1) j c m c s r c where : P j - current losses, P m magnetisation losses, P s -solar radiation, P c - corona losses; P k - convection, P n rain, P r radiation Kwing the weather conditions and some constant conductor parameters, the ampacity at which the conductor reaches its maximum permissible temperature (in this case 80 C) which also corresponds to the maximum permissible conductor sag can be calculated according to equation 1 (see also figure 2). The calculation of conductor temperature based on the load current and the weather conditions is quite complex, because of k c n Fig. 2 Principle of calculation of maximal conductor temperature and conductor kwn load current numerous dependencies on the conductor temperature itself. The direct derivation of conductor temperature (with a few simplifications) is possible [7]. This equation may also be solved by an iterative approach (figure 2). The temperature of the conductor (T c ) will be varied until both sides of equation (1) are almost equal (error < 0.1 C). 2.3 Statistical approach The investigations cover almost 2 year of measurements almost 50,000 data sets with 15 min resolution. The most suitable approach for analysis of such amount of data is a statistical approach [8]. Hence in this work, the relative frequency of different parameters [11] will be applied. From this statistical evaluation given in figure 3 it can be derived for both sensors SAW 1 and SAW 2 that the conductor temperatures were lower than 10 C in 30% of time period considered and for 10% of time period considered higher than 30 C. This shows that the conductor temperatures are far below the permissible limit of 80 C and that there is a potential for increasing the ampacity which can be utilised by overhead line monitoring. 2.4 Shadowing scenarios The weather conditions which are measured by local and remote weather stations give information about the condition in this particular location [8][9]. Since the overhead lines may be many kilometres long, the weather conditions along the line will t be the same and particularly affected by different wind speeds and directions. Additionally, a shadowing effect has to be taken into account, if the line passes valleys or forest aisles [10]. Therefore, four scenarios with varying wind speed and direction will be considered to simulate shadowing effects (table 1). The principle of application of shadowing effect is shown in figure 2. scenarition wind speed wind direc- S1 v = 0.6 m/s φ = 90 S2 v = v real /2 & min(v) = 0.6 m/s φ = φ real & min(φ) = 30 S3 v = v real /2 & φ = 30 min(v) = 0.6 m/s S4 restrictions φ = φ real & min(φ) = 30 Tab. 1: Different shadowing scenarios, where v real is the measured wind speed an v is the wind speed used for ampacity calculation 2.5 Measured conductor temperatures and possible ampacities In the considered period of time a variety of different weather and load conditions have been measured. The conductor temperatures varied between -10 and 50 C and the load current between 0 and 100% of conductor minal current (figure 3). That means that the complete current range was covered, but the maximum conductor temperature of 80 C was t reached. puffer@ifht.rwth-aachen.de 3

4 Fig. 3: Measured conductor temperature (A) and currents of the monitored overhead line (B) 2.6 Discussion of results The two sensors of the conductor temperature measuring system (SAW 1 and SAW 2) are mounted on two conductors of the same circuit, but at different heights. Due to the phase arrangement of the circuit on the tower, sensor SAW 2 on conductor 2 is mounted about 3 meters higher than sensor SAW 1 on conductor 1. The wind profile and herewith the cooling effect is depending on the height above ground. Thus the cooling of the higher located conductor 2 should be more intensive than the cooling of conductor 1. The resulting conductor temperatures should differ. This dependency on wind conditions can be seen in figure 4, curve (1). The positive deviations from T c _saw1 show that for 90% of considered period the temperature of the lower conductor is few degrees higher. Since the temperature Tc_saw1 is higher than Tc_saw2, the conductor sag of conductor 1 will be larger than of conductor 2. Therefore, the temperature of the conductor 1 (Tc_saw1) will be chosen as reference value. The temperature of the conductor (T c ) is calculated using CIGRE-model and different shadowing scenarios. The weather conditions are measured by local weather station (ls) and remote weather station (rs). The difference between calculated conductor temperatures (M2 and M3) and directly measured conductor temperature (M1) is shown in figure 4. The cumulative frequency distribution of these deviations for exemplary study cases is also shown. Fig. 4: Frequency of differences between measured and calculated conductor temperature for different shadowing scenarios depending on the location of weather station (ls-local, rs-remote) The curves in figure 4 can be analysed as follows: Positive deviations mean that measured conductor temperature (Tc_saw1) is higher than the calculated temperature. This condition implies a risk to exceed the maximum allowed conductor temperature. puffer@ifht.rwth-aachen.de 4

5 Negative deviations are allowed, since the conductor temperature is in the permissible range. In this case, the maximal ampacity will t be reached in any case. Scenario S1 (only temperature, curve (8): This scenario gives the best adequacy because only 1% of deviations is higher than 5 C. However, it does t utilise the maximum of conductor ampacity, because 80% of deviations are negative. Interesting is, that the use of ambient temperature and standard rmative weather conditions (v = 0.6 m/s, φ = 90 ) still produces positive deviations in 20% of data points. Hence, it can be concluded that either there are worse weather conditions than those in standard or the models are t good eugh at those small wind speeds. This fact requires further investigations. Scenario S2, (half wind speed, φ = φ real & min(φ)=30, curves (6) & (7)): This scenario applies the real wind direction in the model, in contrary to scenario S3 at which a constant angle of wind direction is used. It is well to be seen in particular, if the curves for local station (5) and (7) are compared. The positive deviations are almost equal to the use of temperature only (scenario S1). The method provides a very good safety margin. Hence, the distribution curve for a local station is very narrow and the scenario has a good safety margin, this location of the weather station and this shadowing scenario give the best representation of the conductor temperature. Scenario S3 (half wind speed, φ = 30, curves (4) and (5)): For the local station there are about 50% of positive deviations and for the remote station slightly less than 40%. However, only about 5% of deviations are higher than 5 C. On the negative side of axis the calculations according to data from remote station exhibits much more inadequacy than for data coming from local station. Scenario S4 (wind speed with restrictions, curves (2) and (3)): Almost 80% of deviations are positive. However, only 10% of deviations are higher than 5 C. Here, the calculated temperature is lower than the measured one. The direct use of wind speed with restrictions for calculation of conductor temperature is apparently t adequate eugh. The local weather station is located at a similar height above ground as the conductor; therefore it should provide comparable results. The reason for this effect which mainly appears at low wind speeds (see figure 5A) may occur due to inadequacy in weather measurements, the conductor models and/or dynamic thermal behaviour of conductor. Moreover, the sensors measuring the temperature directly are calibrated for wind speeds in range of 4 m/s and may produce small deviations at low wind speeds. According to discussion above, the scenario S2 is the most accurate for ampacity determination. But the results also show that there is method which can determine the conductor temperature very exactly. In case of ls_s2 (local weather, half wind speed, and taking into account wind directions) a deviation of 5 C may be reached. It has to be mentioned here, that this deviation refers only to the conductor temperatures up to 50 C. Above this value, observations have been made. puffer@ifht.rwth-aachen.de 5

6 Fig 5: Deviations of theoretically calculated conductor temperature to the measured conductor temperature dependent of the wind speed (A), ambient temperature (B) and the season of the year (C) The dependency of the deviations between measured conductor temperature and calculated temperature using the local and the remote weather conditions at scenario 2 is shown in figure 5. At higher wind speeds the deviations become smaller (figure 5A). The dependency on the ambient temperature (figure 5B) cant be detected at the first look. The dependency on the season of the year (figure 5C) is also difficult to recognise. Obviously the largest deviations occur within few days (peaks in figure 5C) and at ambient temperatures of about 10 C (peak in figure 5B). For clarification of these details further investigations are needed. 3. Area-wide dynamic rating of 380 kv overhead lines An important rth to south connection in the German grid of TenneT TSO GmbH is located between the cities of Hamburg and Frankfurt (see figure 6). To adjust the grid to the new boundary conditions it is necessary to build new lines as well as to improve the efficiency of the existing infrastructure. The policy is to fully utilise the existing infrastructure first and to have new lines erected only for further demand. An overview of the necessary measures to be taken in order to increase the transmission capability of this part of the grid is given in figure 7. In the 380 kv transmission grid it is essential to keep up the stability of the system. By raising the maximum allowable ampacities the development of a new protection concept became necessary. Also a concept to maintain system safety for higher utilisation was required. As a result facilities providing reactive power had to be installed and changes to the protection system were carried out. Netherlands TenneT Denmark Energienet.dk Amprion Frankfurt EnBW Sweden Svenska Kraftnät Hamburg 50Hertz Transmission Czech Republic CEPS Austria TIWAG APG Fig. 6: Retrofitted connection between Hamburg and Frankfurt puffer@ifht.rwth-aachen.de 6

7 check design temperature increase design temp. special inspection replace check joints replace check substation check protection and dynamic aspects replace adapt, replace everything done? magnetic field threshold implementation of ampacity algorithm adapt max. ampacity overhead line dynamic rating applicable weather station and location documentation Fig. 7: Workflow of necessary measures for dynamic rating implementation 3.1 Dynamic aspects of the transmission system The transmission capacity is increased by raising the ampacity of the 380 kv bottleneck circuits to 3,150 A. With an increasing utilisation of the currently existing connections without adding new transport capabilities from the rth to the south, the risk of loosing the generator stability (phase angle stability) increases as well (see figure 8). The stability limits mainly depend on the grid impedances (inductive and capacitive). Also the protection system is affected by high minal currents, in regard to selectivity and reliability. It is more difficult to safely detect and isolate faults. High minal currents make it hard to differentiate between operational and fault currents without the chance of incorrect tripping. P P DLR U1 U = sinδ X 2 P2 reduction of transmission angle increasing transmission capacity with expansiof transmission system P 1 without expansiof transmission system static boundary 0 45 stable instable Übertragungswinkel transmission angle Fig. 8: transmission angle as a function of the transmission capacity (DLR Dynamic Line Rating) Where P 1 is the feeding power and P 2 is the load power, P DLR is the increased feeding power using dynamic line rating, U 1 and U 2 are the voltages at the beginning and the end of the line, and δ is the angle between current and voltage To improve the transport capacity of the grid the system boundaries which arise from system stability and grid protection have to be taken into account during the grid planning process. Analysing the possibilities of enhancing the capacity of the main transmission corridors from rth to south system many studies were conducted. As a part of the EWIS study (European Wind Integration Study) the results were verified in an European context, considering coordinated robust scenarios (European market model for power plant operation) and measures (e. g. the grids at the borders to the Czech Republic, Poland, Austria, Benelux or Denmark). puffer@ifht.rwth-aachen.de 7

8 3.2 Protection of the transmission system The protection concept was revised in order to allow a quick location of faults even in the case of a single technical component failure. The analysis of the calibration principles of the protection system revealed that raising the boundary protection current and accordingly a less sensitive tripping setting on its own was t sufficient to meet the increased requirements. In addition the influence of large transmission angles on the measuring precision of the distance relays in the event of electric ark faults and compensation currents in healthy lines at a single line fault are factors of consideration. 3.3 Realization of necessary measures The ampacity of a high voltage circuit can be limited by the primary equipment in the substation bays and by the overhead lines. The equipment used in the substation bays can generally be retrofitted to match the ampacity that can be achieved using the dynamic rating of the line. In order to determine presently available ampacity, the weather data has been evaluated. This evaluation shows that in about 80 % of the year a weather dependent ampacity of 3,150 A is possible (see figure 9). 5,000 weather dependent ampacity [A] 4,000 3,000 2,000 4,000 A 3,150 A four bundle conductor ACSR 240/40: 2,580 A 2,500 A 2,000 A 1, time [%] Fig. 9: Available ampacity with dynamic rating of the connection between Hamburg and Frankfurt 3.4 Measures in substations In the substations the following primary equipment has been reviewed and replaced as needed: circuit breaker current transformer isolating switch voltage transformer switch bay lines For the circuit breaker, current transformer, isolating switch and for some voltage transformers (special types) the rated currents were checked. For equipment with thermal ratings below 3,150 A, the device had to be replaced. The switch bay lines and bus bars were checked in regard to their short circuit capability. Insufficiently dimensioned components have been replaced. The above mentioned measures took a great amount of planning and coordination because of the limited operability of the grid during the intervention and the aim t to cause any impact on customers. puffer@ifht.rwth-aachen.de 8

9 3.5 Measures at overhead lines Since the usable ampacity of an overhead line depends among other things on the specific design temperature, generally the ampacity can be increased by raising the clearance. If the specific design temperature of lines is lower than the rated conductor temperature an ampacity increase can usually be reached by the elevation of selected towers. However, the elevation of a 380 kv line tower yields a couple of challenges. For example, in order to elevate a 10 ton tower, special cranes are needed which have to approach the tower if necessary over provisional tracks (figure 10). Fig. 10: Pictures of inserting additional body sections in a 380 kv tower near Stade The joints in the line section were completely inspected on site. Only intact connection components remained in the line whereas conspicuous components were replaced. The condition of joints was determined using infra-red thermography. Raising the ampacity of an overhead line raises the maximum value of the magnetic field as well. Calculations of the magnetic field were carried out in order to assure the compliance with legal specifications. 3.6 Implementation of dynamic rating into the operating system In order to improve the ampacity of 380 kv lines a dynamic rating system has been developed and implemented into the existing operating system of TenneT TSO in Germany. This monitoring system uses the ambient temperature and the wind speed to determine the dynamic current rating of the line. It is directly linked to the control centre and delivers a dynamic current rating. There is need for any direct measurement of the conductor temperature to determine the dynamic current rating. The operational experience using the dynamic rating system for the operation of 380 kv lines proves that it interposes very well with the existing control centre s techlogy. The distinct increase in dynamic current rating compared to static rating verifies the technical and ecomical capability. puffer@ifht.rwth-aachen.de 9

10 4. Conclusion Many overhead line monitoring systems (direct and indirect temperature determination methods) are presently available on the market. The determination of conductor temperature using weather data is riddled with many inadequacies. Hence, the conductor temperature determination from weather data may be done with a maximal adequacy of about 5 K in the conductor temperature range up to 50 C. The findings of this paper show that the accuracy of indirect monitoring systems is comparable with direct monitoring systems considering adequate safety margins. Moreover, investigations on the thermal behaviour at low wind speeds seem to be necessary as well as better understanding at what time the largest deviations to the measured conductor temperature occur. A dynamic rating system using weather measured in substations to calculate the ampacity was installed on 800 km of 380 kv overhead lines. Different retrofitting measures like checking the joints, changing substation equipment, raising towers and verifying the stability criteria of the grid had to be done. The operational experience using the dynamic rating system delivers a distinct increase in dynamic current rating which verifies the technical and ecomical capability of the system. BIBLIOGRAPHY [1] Dräger, H.-J.; Hussels, D.; Puffer, R.; Development and Implementation of a Monitoring- System to Increase the Capacity of Overhead Lines, Paper B2-101, Cigré Session 2008 [2] Cigré, Technical Brochure 207; The Thermal Behaviour of Overhead conductors, August 2002 [3] H. Kühn, F. Martin, M. Schmale, R. Puffer, W. Winter; Mehr Energie von Nord nach Süd, ew 3/2011 and 4/2011 [4] M. Schmale, R. Puffer; Freileitungen sicher betreiben und Reserven nutzen, netzpraxis, Jg. 49 (2010), Heft 11 [5] Blumenroth, F., Löbl, H., Großmann, S., Puffer, R. Hussels, D.: Ageing of high current joints in power transmission ans distribution systems, Cired 2007, paper 537 [6] Schmale, M.; Puffer, R.; Dräger, H.-J.; Experience with a Dynamic Rating System to Increase the Ampacity of 380-kV Overhead Lines, Cigré Symposium Bologna 2011 [7] Temiva, R.; Hinrichsen, V.; Freese, J.; Hudasch, M.; Bebensee, R.; Neumann, C.: Betriebserfahrungen mit passiven funkabfragbaren OFW-Sensoren zur Messung der Temperatur von Freileitungsleiterseilen und Trennschalterkontakten, ETG Tagung, 2006 [8] Lange, M.; Focken, U.: Studie zur Abschätzung der Netzkapazität in Mitteldeutschland in Wetterlagen mit hoher Windeinspeisung, Studie energy & meteo systems, Oldenburg, 2008; gefördert vom BMU [9] Roman, H. Dangrieß, G.; Darendorf, S. Struck, T.: Freileitungsmonitoring im Hochspannungsnetz - Theoretisches Potenzial. ew Jg. 108 (2009), Heft 12, S [10] Roman, H. Dangrieß, G.;Darendorf, S. Struck, T.: Freileitungsmonitoring im Hochspannungsnetz - Reales Potenzial. ew Jg. 108 (2009), Heft 13, S [11] Neumann, C.; Rusek, B.; Puffer, R.: Weather dependent loading of overhead lines based on statistical consideration of weather data (in German). ETG Congress, Duesseldorf, 27. to 28. October 2009, Report 3-7 puffer@ifht.rwth-aachen.de 10