Figure 1. The basic configuration of the DVR. Table 1. The states of three phase three level inverter at each arm. Figure 3. Three level three phase inverter. Figure 2. Three level single phase inverter. The DVR injects the compensating voltage (the difference between the sag and pre-sag voltage) through the injecting transformer as shown in Figure 5. Figure 7. The DVR connected to the medium voltage. e At the medium voltage as shown in Figures 6 and 7 the DVR is connected to 11kVr.m.s radial distribution feeder that supplies the load's transformer e At low voltage as shown in the Figure 8 the DVR is connected to the secondary winding of the load's transformer which is 400V Figure 8. The DVR connected to the low voltage. Research & Reviews: Research and Review: Journal of Engineering anc Technology Figure 9. The electrical power circuit with the DVR. Figure 10. The phaser diagram for the power circuit. VR ~ “supply “Load From Figure 10, compensation voltage (Vinj) of the DVR can be calculated as following below: Figure 11. The basic configuration of the fuzzy logic controller. Xesearch & Reviews: Research and Review: Journal of Engineering and fechnology The input and output membership function for the first, second and third controller is as shown in figures below (Figures 12-22), Figure 12. The input membership function for the error. Figure 13. The input membership function for the change of error. Figure 14. The output membership function. Research & Reviews: Research and Review: Journal of Engineering ane Technology Figure 15. The input membership function for the error. Figure 16. The input membership functions for the change of error. Figure 17. The output membership function. Figure 18. The input membership function for the error. Figure 20. The output membership function. Figure 19. The input membership function for the change of error. Fast controller third controller Figure 21. (a) The rules set for the first controller (b) The rules set for the Second controller (c) The rules set for the third controller. DYNAMIC VOLTAGE RESTORER BASED ON FUZZY NEURAL CONTROLLER Research & Reviews: Research and Review: Journal of Engineering anc Technology Figure 24. The perceptron. Perceptrons (MLP) as shown in Figures 24 and 25. Each hidden layer uses the same activation functior Research & Reviews: Research and Review: Journal of Engineering anc Technology ANFIS Architecture is an adaptive network that uses supervised learning algorithm the function of ANFIS is similar to the Takagi-Sugeno fuzzy inference system. Figure 29. Flow chart for ANFIS algorithm. The flow chart for ANFIS algorithm is as shown in Figure 29. Flow Chart for ANFIS Algorithm Train Data and ANFIS Editor for the Fuzzy Neural Controllers The train data of ANFIS is shown in Figures 27-29 and the editors of ANFIS are shown in Figures 30-35. ANFIS optimize PI gave us better results than PI in both linear and nonlinear loads but did not give us better result than Fuzzy logic where Fuzzy logic was far more efficient than (ANFIS optimize Pl). Therefore we choose ANFIS optimiz fuzzy logic whose called (fuzzy neural) where we has obtained better results than fuzzy logic when we added the 2" an 3'¢ harmonics to the sag and swell Figure 28. Figure 32. Train data for the third ANFIS. Figure 31. Train data for the second ANFIS. Figure 30. Train data for the first ANFIS. Research & Reviews: Research and Review: Journal of Engineering anc Technology Figure 34. ANFIS editor of the second controller. Figure 33. ANFIS editor of the first controller. Research & Reviews: Research and Review: Journal of Engineering an Technology Figure 36. The Simulink model of the test system with linear load. Figure 38. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (c) Load voltage with DVR based on FN controller. Figure 39. (a) Load voltage without DVR; (b) Load voltage with DVR on FL controller; (¢) Load voltage with DVR based on FN controller. Double line to ground fault: As shown in Figure 34 DLGF (unsymmetrical sag) is simulated. It is started from 0.8s and it is kept until 0.95s. DVR based on Fuzzy logic and Fuzzy neural is restored the load voltage to the nominal value Figure 39(a-c). Figure 40. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (¢) Load voltage with DVR based on FN controller. Three phase fault: In Figure 40(a-c) three phases to ground fault (symmetrical sag) is simulated. It is started from 0.8s and it is kept until 0.95s. The load voltage is kept at the nominal value with the help of the DVR based on Fuzzy logic anc Fuzzy neural. Figure 41. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (¢) Load voltage with DVR based on FL controller. Figure 42. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (c) Load voltage with DVR based on FN controller. Swell: As shown in Figure 42(a-c) the three phases will be swelled to 150% for 0.15s from 0.8s to 0.95s. As can be seen the load voltage is kept at the nominal value with the help of the DVR based on Fuzzy logic controller and Fuzzy neural controller. Figure 43. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (c) Load voltage with DVR based on FN controller. Sag with 2nd order harmonic and 3rd order harmonic: As shown in Figure 43(a-c) three phase sag with presence 2" and 3 harmonics is simulated. It is initiated from 0.8s to 0.95s. As can be seen, the load voltage kept at pre-sag voltage and without harmonic distortion with the help of the DVR based on Fuzzy neural. DVR based on Fuzzy logic was incapable to restore the load voltage to the pre-sag value and kept the THD within acceptable limits. Figure 44. (a) Load voltage without DVR; (b) Load voltage with DVR based on FL controller; (¢) Load voltage with DVR based on FN controller. Research & Reviews: Research and Review: Journal of Engineering and Technology Figure 45. The Simulink model of the test system with non-linear load. Table 2. Total Harmonic distortion for load voltage without DVR and with DVR based on fuzzy logic controller and fuzzy neutré controller.
Connect with 287M+ leading minds in your field
Discover breakthrough research and expand your academic network
