| With the fast development of smart distribution network and government’s firm support on clean energy,distributed generations such as wind power generation and photovoltaic generation and new loads such as electric vehicles have been growing rapidly.These new sources and loads are affected by natural resource conditions,customer behavior patterns,exterior environments and geographic locations,and thus exhibits strong stochastic and correlated features.Their extensive integration into distribution networks would bring about massive uncertainties and correlations,affecting the security and reliability of distribution systems.Neglecting these factors might lead to deviations in the evaluation of system operating status and reliability.On the other hand,failures of electrical equipment including distributed generators and distribution lines would directly sabotage system balance,causing severe consequences to power supply reliability.The reliability indices and operating status of electrical equipment are influenced by exterior operating conditions and device damage states,and also carry uncertainties and correlations.Therefore,it is of necessity to investigate distribution system reliability assessment methods considering multiple uncertainties and correlations.This thesis takes into consideration multiple uncertain and correlated factors in distribution networks,and conducts analyzations and modelling of these factors.To incorporate uncertainty of correlation coefficients of stochastic variables,a triangular fuzzy coefficient matrix is built;based on the uncertainty of electrical equipment reliability parameter caused by system operating status,a distribution line failure probability-line power flow conditional dependence model is proposed;to depict correlations between distributed generation failures,a Copula model of generator operating status with discrete margins is established.The above models are applied to distribution network reliability assessment.Power flow analysis and calculation is an essential part of distribution network reliability assessment.To incorporate the fuzzy correlations of uncertain variables,this thesis proposes a distribution network fuzzy probabilistic power flow calculation approach based on an improved threepoint estimate method.The proposed approach handles fuzzy variables arithmetically,generates fuzzy correlated samples though fuzzy Nataf transformation,calculate fuzzy moments of output variables,and derive the fuzzy probabilistic distribution and fuzzy membership function of system power flows variables analytically through a modified Cornish-Fisher expansion.Conventional power system reliability evaluation methods cannot deal with the correlation between discrete variables such as electrical equipment operating status and the uncertainty of equipment reliability parameters.To incorporate correlations between distribution generation failures and the time varying trends of line failure probability against power flow,a distribution system reliability assessment method based on an improved Markov Chain Monte Carlo simulation method is proposed.The proposed method derives variable full-conditional distributions,and samples the operating status of distributed generators and distribution lines through the blocking Gibbs sampling approach.The method further adopts the fuzzy probabilistic power calculation method proposed above to perform system supply status determination,and calculate system reliability indices accordingly.Two IEEE standard test systems and an actual distribution system in East China are taken as test systems to perform system fuzzy probabilistic power flow calculation and reliability assessment using the proposed methods,and the effectiveness and practicability of the proposed methods are validated.Further,the case studies prove that the multiple uncertain and correlated factors considered in this thesis would exert multiscale influences on distribution network operation and reliability. |
