Multi-Objective Adaptive Robust Voltage/VAR Control for High-PV Penetrated Distribution Networks

Significance 

Global trends championing for the development and incorporation of renewable energy into existing power grids are on the rise. Specifically, literature has it that the use of photovoltaic (PV) systems has significantly proliferated in active distribution networks, bringing clean, cost-effective and sustainable energy thus reducing the dependence on fossil fuels. Unfortunately, it has been noted that excessive injection of PV power into the distribution feeders leads to a voltage rise challenge. In fact, this is congruent with the fact that PV power generation is highly uncertain and stochastic. Consequently, it impairs power quality by large voltage deviations hence causing operating constraint violations. To overcome this shortfall, voltage/VAR control (VVC) has been employed to regulate bus voltages and reduce network power loss. Owing to the multi-objective nature of the problem, multi-stage coordination of the multiple VVC devices has been developed in various literatures. Nonetheless, a thorough review of such studies reveals that the various VVC devices employed fail to robustly coordinate considering the multiple objectives and uncertainties.

In general, the multi-objective VVC problems are not efficiently solved to obtain an accurate Pareto front. Actually, a robust optimization method for multi-objective programming (MOP) problems has not been adeptly studied. On this account, Australian researchers from the University of New South Wales: Dr. Cuo Zhang, Professor Zhao Yang Dong and Dr. Rui Zhang, in collaboration with Dr. Yan Xu at Nanyang Technological University in Singapore, designed a new multi-objective adaptive robust (MOAR) VVC approach to address the deficiency of the existing works. Their work has been published in the research journal, IEEE Transactions on Smart Grid.

In their approach, a new load-weighted voltage deviation index (LVDI) was proposed to quantify the network voltage deviation status. The researchers also proposed a multi-objective adaptive VVC framework where multiple devices respond in multiple timescales, so as to minimize both voltage deviation and network power loss. In addition, a multi-objective adaptive robust optimization (MOARO) method was developed to fully address uncertainties in the VVC problem and different solution algorithms are developed to obtain robust Pareto solutionsLastly, a comprehensive case study was performed to verify effectiveness of the LVDI, compare performances of the different MOP based solution algorithms and verify operating robustness of the proposed MOAR-VVC method.

The authors found that the adaptive weighted-sum algorithm could generate an accurate and smooth Pareto front with fairly distributed solutions. In addition, the adaptive weighted-sum algorithm could detect insensitive regions and provide boundary solutions. Further, the normal boundary intersection algorithm could also provide an accurate Pareto front with evenly distributed solutions and the “knee” solution which indicated a fair trade-off between the objectives. Moreover, feasibility check results showed that there was no infeasible case occurring for the proposed MOAR-VVC method. Technically, this was seen to be an indication that the day-ahead scheduling decision was robustly optimized with full consideration of the uncertainties and that the intraday inverter dispatch could efficiently track uncertainty realization to keep operating constraints.

In summary, the study proposed a MOAR-VVC approach which could robustly coordinate multiple devices in different time-scales to minimize voltage deviation and network power loss simultaneously under uncertainties. The proposed approach was tested on a modified 123-bus distribution network with a comprehensive case study. Firstly, the case study verified the effectiveness of the proposed LVDI. Secondly, different MOP based solution algorithms were utilized to obtain Pareto fronts and compared. Thirdly, the operating robustness was validated. In a statement to Advances in Engineering, Professor Zhao Yang Dong explained their simulation results demonstrated high effectiveness of the LVDI, high efficiency of the solution algorithms and full operating robustness of the proposed MOAR-VVC method against any uncertainty realization.

Reference

Cuo Zhang, Yan Xu, Zhao Yang Dong, Rui Zhang. Multi-Objective Adaptive Robust Voltage/VAR Control for High-PV Penetrated Distribution Networks. IEEE Transactions on Smart Grid, Volume: 11 – Issue: 6, 2020.

Go To IEEE Transactions on Smart Grid

Check Also

Light-Speed Encryption: Unlocking the Future with Spatially Incoherent Diffractive Neural Networks - Advances in Engineering

Light-Speed Encryption: Unlocking the Future with Spatially Incoherent Diffractive Neural Networks