Research On Power System Static Security Analysis Considering Uncertainty

The number of renewable energy sources(RESs)(e.g.,wind power and photovoltaics)increases every year.Concurrently,distributed generations and electric vehicles are also increasing annually.This growth considerably increases the operational uncertainty of power systems,making safe and stable operation difficult.Power system static security analysis is a basic tool for ensuring stable and reliable power system operation.However,in a situation where the uncertainty on both sides of the power system is significantly enhanced,traditional static security analysis methods based on deterministic models fail to meet application requirements.Aimed at attenuating the shortcomings of existing static security analysis methods,this study examines static security analysis methods taking uncertainty into account.The methods researched include the probability modeling method of uncertain factors,efficient probabilistic power flow algorithm,preventive control method,and corrective control method.Based on the theoretical study of these methods and with the support of the National High Technology Development and Research Project(863 project),an on-line preventive and correction control auxiliary decision-making system was developed.This research has theoretical significance and practical value for application ensuring reliable and stable operation of power systems and promoting safe renewable energy consumption.A probability model for describing the uncertainty from power sources and loads is an essential basis for a static security analysis that considers the influence of uncertainty.The uncertainty stems primarily from a forecasting error of the bus power injection,and its probability distribution is influenced by many factors,including the type of power injection,forecasting method,forecasting scale,etc.There is presently no accurate unified probability model for describing such forecasting errors.To solve this problem,this study proposes a unified probability distribution model based on the Johnson system.The model uses the first four moments of random variables and a function transformation to characterize the probability distribution of arbitrary variables.Theoretically,the Johnson system has the advantage of not being constrained by the range of skewness and kurtosis.Thus,it can uniformly describe the probability distribution of the forecasting error from different bus power injection types and different forecasting scales.The effectiveness of the Johnson system was validated using actual wind power forecasting error data and load forecasting error data.The unified probability distribution model provides a solid model foundation for studying probabilistic power flow algorithms,the preventive control method,and the corrective control method.Existing probabilistic power flow algorithms struggle to meet the accuracy and efficiency requirements of on-line static safety analysis for large-scale power systems because large-scale power systems have many correlated non-normal random input variables.Addressing this problem,this study proposes an efficient probabilistic power flow algorithm based on enhanced high-dimensional model representing(HDMR).The hierarchical structure of HDMR facilitates consideration of the cooperative effects of random input variables and effectively improves the estimation accuracy of the first four moments of the random output variables.Applying the Johnson system with the moment results from HDMR,the probability distribution of the random output variables was obtained accurately.To improve the efficiency of the moment estimation procedure,two efficiency improvement methods(principal component analysis and identification of significant cooperative effects)are proposed.The simulation results show that,compared with traditional algorithms,the proposed algorithm is less sensitive to the correlation of random input variables and the degree of volatility,achieves superior performance in terms of accuracy and efficiency,and has potential for on-line applications.Preventive control is a crucial measure for improving the global steady state operational security of the system.However,current preventive control methods do not consider the influence of various uncertainty factors.Furthermore,it is difficult to solve the preventive control model efficiently because of the large scale of the constraints.This study proposes a preventive control method based on chance-constraint programming to solve this problem.This approach considers the influence of power transmission equipment failure probabilities and combines the equipment failure probabilities possessing the same consequences to form contingency probabilities.Applying the theory of total probability,the uncertainty of the RESs,and loads,the contingency probabilities were effectively merged to form an overarching probability distribution of the branch power flow,with which a preventive control model based on chance-constrained programming was constructed to limit the overloading probability of branch power flow.The proposed method effectively considers the influence of uncertainty factors from sources,networks,and loads of a power system,with improved overall safety and robustness of the system operation state.The scale of constraints is reduced considerably via the theory of total probability,which significantly improves computational efficiency and creates the potential for on-line calculation.Corrective control is an essential measure for increasing safety for a single contingency.However,existing corrective control methods generally disregard the influence of the uncertainty from RESs and loads,which may lead to insufficient robustness of the corrective control strategy.Consequently,this study proposes a static safety corrective control method that considers the uncertainty of the RESs and loads.A chance-constrained programming model based on the AC power flow model was established to minimize the cost of corrective control.The problem of solving chance-constrained programming with nonlinear constraints was surmounted via iterative calculation of deterministic AC optimal power flow and probabilistic power flow using the proposed efficient probabilistic power flow algorithm to improve the calculation efficiency.To ensure the reliability of the solution,the deterministic optimal power flow problem was relaxed by applying the quadratic convex approach.Using a sensitivity screening method,a collection of effective control variables was obtained,effectively reducing the number of operations by system operators in practical use.The simulation results show that compared with the traditional corrective control methods,the proposed corrective control method is significantly safer and more robust,and meets efficiency and accuracy requirements.Based on this theoretical study,and using the big data analysis platform and the calculation engine of the 863 project,an on-line static safety preventive and corrective control auxiliary decision-making system suitable for large-scale urban power systems was developed.The demonstration results show that the developed system meets the timeliness requirements for on-line applications.It provides multi-view operational control auxiliary decision-making support for operators and effectively improves power system operation safety.