FUZZY LOGIC BASED CONTROL ON OSCILLATORY POWER SWING IN A SYSTEM.
ABSTRACT: In this paper, a comparative research work is done to solve a major problem of power swing. Power swing is seriously an abnormal operation status in power system. When this situation occurs in power system current, voltage and power fluctuate periodically which may seriously damage generators, motors and other devices. At the time of symmetrical and unsymmetrical faults situation becomes more troublesome. Unnecessary switching and its sustained presence in system for some cycles are equally undesirable. So a detailed analysis of its occurrence and multi level algorithmic approach is required to eliminate the very phenomenon in system. In this research work fuzzy logic is being utilized to solve this important problem and its effectiveness is also being explored. System is implemented in MATLAB/SIMULINK.
Index Terms: Power swing, Fuzzy logic, and Symmetrical and Unsymmetrical faults.
Power swing is a seriously abnormal operation status in a power system. When this situation occurs in a power system, the current, voltage and transmission power in the network will fluctuate periodically which may seriously damage the generators, motors and other devices. If the power swing can be controlled in time using required protection, the system may be stable again. If the control measures fail, the system will be out of step. Since distance relays are prone to interpret a power swing as a three-phase fault, they should be blocked during the power swing to prevent undesired trips. On the other hand, if any fault occurs during a power swing, they should be fast and reliably unblocked. Although unblocking the relay is straightforward in the case of asymmetrical faults by using the zero-sequence and/or negative-sequence component of current, detecting symmetrical faults during a power swing is still a challenge. There are two basic types of swings in power system.
One is sustained and one is for a short interval of time. If swing occurred in the system for a short interval of time than it is expected that our protection system is intelligent enough to avoid unnecessary tripping. But at the same time if it's a sustained oscillatory swing in system then it must be isolated by ensuring necessary but limited tripping so that other healthy part of system must be safe from this problem. Normally it is seen that modern distance relaying is capable enough to distinguish between then and hence the need of Power Swing Blocking function is there and the very function should be unblocked at the time of fault and let the relay to trip circuit breaker. Because we know that repeating tripping would be harmful for devices like circuit breaker and transformer and also it would be considered non ideal at the end of consumer.
But at the same time at the time of symmetrical and unsymmetrical faults necessary tripping should be there and if power swing is unstable and persisting in system for more time then it is mandatory to do necessary tripping. Several techniques are there which take into account variations of different parameters at the time of swing and take decision on basis of those parameters. One common technique is to detect the impedance of the system at that point also there are other techniques which detect power swing based on the changing of oscillation of current. There is also another method to determine power swing based upon voltage and its changing rate . The variation of active power is also one of the criteria but ultimate goal of all methodologies is same that is to detect the power swing and to suggest proper actions to avoid unnecessary tripping of distance relays and also system stability and security should not be at stake.
Some advanced techniques based on wavelet transform and S-transform and some new parameters to determine power swing are described in references [4-8]. There is always need to mature directional relaying because it is cost effective and technically sound solution of the very phenomenon as local back up of distance protection. Fuzzy logic can solve these complex problems more efficiently by taking an optimized decision relative to severity of a disturbance. The fuzzy system described in references - is taken into account in order to solve problems related to the power swing phenomenon. Artificial intelligence techniques sometimes alone and sometimes in conjunction with several stochastic ways of optimization are proving quite helpful in order to solve problems related to power systems.
We know that trend is shifting toward expert systems which are more rule based and all those events which could happen and effects of those events or abnormalities are being realized and modeled in these systems so that automotive decision could be taken without any manual interference.
In order to increase the strength of technique parameters on which decision is being made are increased and more rules are defined so that prompt and reliable decision of tripping is being taken. So in this proposed research we will utilize fuzzy logic to solve this important problem and will compare results with existing techniques. Moreover effectiveness of the proposed technique will be explored which would justify that multiple parameters will increase the intelligence of device provided that its rules are being defined after careful and detailed simulation. Lower and upper cutoffs marking are real milestone to be achieved.
2. PROPOSED METHODOLOGY
By doing detailed stability study of power system from different standard texts it is seen that some sequence parameters of the voltage and current vary in a special way during power swing and faulty conditions, so if behavior of those parameters is analyzed under different abnormal conditions then a good decision could be taken which is ensuring the reliability of power system as well as necessary tripping could be done. We know that at time of ordinary power swing there is function in modern distance relays known as power swing blocking function which restricts the relay to operate in order to ensure the reliability of system but when asymmetrical or symmetrical faults occur in power system and specially at the time of swing then there is need that fault should be distinguished from the swing and let the relay to operate.
There are three parameters which would be utilized in the intelligent decision taking process. One parameter is angle between pre fault and post fault current, second parameter is angle difference between fault voltage and current and third parameter is ratio of negative to positive sequence of current. Through vigorous simulation behavior of all these three parameters are observed and based upon these parameters decision is being taken.
Sequential procedure of our all process can be indicated in term of following steps.
1. Modeling of the generator, transformer and transmission lines.
2. Interconnection of all power system elements and running of system in normal condition.
3. Observing behavior of three parameters in normal condition.
4. Realization of power swing in the system by adding load in the system after some time.
5. Observing response of above three parameters and measuring of cut off values.
6. Introducing symmetrical faults in system separately at various positions in the system.
7. Again measurement and analysis of decision parameters.
8. Introducing unsymmetrical faults (with fault resistance and without fault resistance) at various positions in the transmission lines network.
9. Measurement and analysis of decision parameters.
10. Simultaneous introduction of swing and symmetrical faults in the system.
11. Simultaneous introduction of swing and un-symmetrical faults in the system.
12. Measurement and analysis of decision parameters under all these condition.
13. Defining low, medium and high level ranges of all decision parameters.
14. Analysis of those parameters which are showing expected variations.
15. Analysis of those parameters which are showing uncertain variations.
16. Identification of critical parameter/parameters which demands immediate tripping signal for breaker and verily identification of those parameters which show limited variation and hence does not require tripping action.
17. Definition of those parameters as membership functions in fuzzy inference system.
18. Designing of three 3 by 3 matrices relating variations of three parameters with respect to one another during all normal and abnormal conditions of system.
19. Defining 27 rules in order to do fuzzification of the three parameters based upon logic defined in above three matrices.
20. Observing system during simulation and synchronizing FIS with power system.
21. Refresh the system continuously with small refresh rate so that continuous status of system and hence the parameters would be observed on rule viewer side of fuzzy inference system.
22. Non-Tripping action is observed in case of three parameters are in normal range or limited variations as per definition of 27 rules thus ensuring reliability of the system.
23. Tripping action is observed even if only parameter is observed which is violating limits defined in above rules.
24. Tripping action is observed by various rules in case of severe abnormalities thus ensuring safety.
3. SYSTEM OF INTEREST
In this research work, three phase power system is simulated with two line transmission network as shown on Figure 1.
Generator is three phase synchronous machine with steam turbine and 600 MVA generating capacity and 22KV generating voltage with operating frequency 50 Hz. Details of the system are given in the APPENDIX. Transformer is of same power handling power capability and it is step up transformer of this ratio 22KV/400KV and Y/Y connection. Sending end of the system is three phase synchronous machine and receiving end is infinite bus. Four breakers are there which would be managed by FIS whether to generate tripping signal or not as per situation of swing and during symmetrical or asymmetrical faults. Two transmission lines are there which are modeled in MATLAB using distributed parameters.
Flow chart of overall strategy is shown on Figure 2. First of all behavior of system is observed in normal condition, system voltage is shown in Figure 3.
Current of the system is shown in Figure 4.
We can see that current is not of very higher values as we are measuring the currents on 400KV transmission network where current is stepped down due to step up of voltage. Figure 5 shows the ratio of negative to positive sequence of current and in Figure 6 above curve is angle between pre fault and post fault current, and down curve angle difference between fault voltage and current under normal condition.
It is clear from above curves that under normal condition ratio parameter is zero and angle difference show really flat variations, so these parameters easily depict normal condition of the system. Now we need to analyze the behavior of system in case of power swing.
We can see from the Figure 7 and Figure 8 power swing is there in the system as voltages is decreased and current of system is increased. Power swing is introduced in the system by adding load in the system from 1.2 s to 1.8 s.
Now need is to investigate the variation of decision making parameters during the phenomenon of power swing. We will see that as compare to normal condition significant variations would be there which would help us to identify power swing. From Figure 9 it is seen that ratio parameter is gone to 0.8 during power swing, same parameter was zero under normal condition. Similarly it can be observed that upper curve in Figure 10 is showing significant high values of angle which would help us to take clear decision. Also we can see that lower curve of angle difference between fault voltage and current during power swing show uncertain variations. That is why fuzzy logic technique is being applied and three parameters are being considered so that if one parameter is failed during power swing, symmetrical faults, asymmetrical faults and simultaneous occurrence of above events in the system then suitable decision could be taken based upon other parameters which would be accurate one.
System behavior is analyzed under all symmetrical and asymmetrical faults (line to ground, double line to ground) on different positions separately and along with simultaneous power swing as well. Based upon those variations three ranges small, medium and large are defined for each membership function and based upon them 27 rules are defined which will avoid unnecessary tripping and hence ensuring reliability of system and also necessary tripping would be there when variations would be out of limit.
I1(angle between pre fault and post fault current)= small range(0 to 60 degree), medium range(61 to 110 degree), large range(111 degree and above).
I2(Ratio of negative to positive sequence of current)= small range(0 to 0.14), medium range(0.15 to 0.8), large range(0.8 and above).
I3(angle difference between fault voltage and current)= small range(0 to 65 degree), medium range(66 to 130 degree), large range(131 and above).
Based upon these rules three matrices of 3 by 3 are formed which are shown in Table 1, and these 27 rules are embedded in Fuzzy Inference System. We can see that second parameter is really critical one, after vigorous simulation it is seen that tripping action is necessary even if it is lying in medium range. But as far as parameter one and three are concerned we can avoid tripping even if its lying in medium range but in case of large variations of any of three parameters tripping in necessarily done.
Table 1: Rules embedded in Fuzzy Inference System
Small###No Trip###No Trip###Trip
Small###No Trip###No Trip###Trip
Medium###No Trip###No Trip###Trip
Figure 11 shows implementation of these rules in MATLAB FIS file. Also we can see from Figure 12 and 13 that no tripping is there if values of membership function are in normal range, last signal is output signal showing tripping or non tripping, if its shaded it means its non tripping and if its shaded it means tripping signal is there. Figure 13 is showing tripping signal is coming from five different sides because out of 27 rules 5 are violated in case of this abnormal condition of system.
Power system stability problems are really critical to deal in case of faulty conditions and that of swing. Condition of the system immediately after clearance of faults, addition of load or rejection of load need to be addressed in order to ensure protection of costly power system equipment and continuity at the end of user. After vigorous simulation, results show that at the time of power swing sequence components of current and voltage and angles(pre fault and post fault) associated with them shows such variations which are quite different as compare to normal condition of power system. Also sometimes some parameters show uncertain variations but at that time fuzzy logic algorithm helps as we are not dependent only on one parameter to take the tripping decision.
There are three parameters and most of the times behavior of all three and certainly two out of three are showing variations as per prediction and helping us to take good decision which is justifiable from both sides that is to ensure continuity to end consumer as well as protection of expensive power system equipment. In this research work this is clear that that maturity of any devise is increased by observing the parameters which show salient variations at the time of abnormal conditions as compare to normal condition of system and also in spite of depending on only one parameters we should increase the numbers of parameters based upon which we are going to take decision of tripping or non tripping. In this way our relaying mechanism would be more efficient and hence reliability, protection and continuity of system would be ensured.
Parameters of 400KV system are as follows:
Generator: 600 MVA, 22 kV, 50 Hz, inertia constant 4.4 MW/MVA. Xd=1.65 p.u. Xd'=0.25 p.u. Xd"=0.2 p.u. Xq=1.59 p.u. Xq'=0.46 p.u. Xq" =0.2 p.u. Xl = 0.14 p.u Tdo' =4.5 s Tdo"=0.04s Tqo'=0.67 s Tqo"=0.09
Inertia coefficient H(s)=0.8788, Rotor type= round
Transformer: 600 MVA, 22/400 kV, 50 Hz, Y/Y. Transmission lines:
Three phase PI section lines
Line length of above line = 130 km
Line length of bottom line = 280 km;
Z1=0.12+ 0.88 /km;
Z0=0.308+ 1.28 /km;
C1=10.87 n F/km;
C0=7.68 n F/km;
 J. Mooney and N. Fischer, "Application guidelines for power swing detection on transmission systems", Proceedings of Power Systems Conference on Advanced Metering, Protection, Control, Communication and Distributed Resources, 2006. PS'06, pp. 159-168, 2006.
 X. Lin, Z. Li, S. Ke, and Y. Gao, "Theoretical fundamentals and implementation of novel self-adaptive distance protection resistant to power swings", Power Delivery, IEEE Transactions on, Vol. 25, No. 3, pp. 1372-1383, 2010.
 X. Lin, Y. Gao, and P. Liu, "A novel scheme to identify symmetrical faults occurring during power swings", Power Delivery, IEEE Transactions on, Vol. 23, No. 1, pp. 73-78, 2008.
 S. M. Brahma, "Distance relay with out-of-step blocking function using wavelet transform", Power Delivery, IEEE Transactions on, Vol. 22, No. 3, pp. 1360-1366, 2007.
 M. Afzali and A. Esmaeilian, "A novel algorithm to identify power swing based on superimposed measurements", Proceedings of 11th International Conference on Environment and Electrical Engineering, pp. 1109-1113, 2012.
 N. Z. Mohamad, A. F. Abidin, W. N. W. A. Munim, and M. Yahya, "Symmetrical fault detection technique during power swing based on S-transform", Proceedings of 8th International IEEE Conference on Signal Processing and its Applications, pp. 7-10, 2012.
 B. Mahamedi and J. G. Zhu, "A novel approach to detect symmetrical faults occurring during power swings by using frequency components of Instantaneous three-phase active power", Power Delivery, IEEE Transactions on, Vol. 27, No. 3, pp. 1368-1376, 2012.
 J. G. Rao and A. K. Pradhan, "Differential power-based symmetrical fault detection during power swing", Power Delivery, IEEE Transactions on, Vol. 27, No. 3, pp. 1557-1564, 2012.
 B. Das and J. V. Reddy, "Fuzzy-logic-based fault classification scheme for digital distance protection", Power Delivery, IEEE Transactions on, Vol. 20, No. 2, pp. 609-616, 2005.
 A. Pradhan, A. Routray, S. Pati, and D. Pradhan, "Wavelet fuzzy combined approach for fault classification of a series-compensated transmission line", Power Delivery, IEEE Transactions on, Vol. 19, No. 4, pp. 1612-1618, 2004.
|Printer friendly Cite/link Email Feedback|
|Date:||Oct 31, 2016|
|Previous Article:||PHYTOCHEMICAL SCREENING AND ANTIOXIDANT POTENTIAL OF SELECTED SPECIES OF LAMIACEAE.|
|Next Article:||BIOGEOGRAPHY OF RUST FUNGI AND THEIR HOSTS IN PAKISTAN.|