Static Voltage Stability Probabilistic Assessment Of Wind Farms Integration System

With the constant adjustment of the global energy structure, the scale of new energy power generation has developed rapidly. As the most mature generating electricity technology by new energy, wind power has grown at an unprecedented speed, and exceeded the nuclear power. However, the inherent defects of wind power, such as randomness and lack of controllability, would cause some undesired consequences of static voltage stability in power system and this e ffect also became more and more obvious with the increase of the penetration of wind power. In order to comprehensively analyze the influence of wind farms integration, this thesis focuses on probabilistic power flow and probabilistic assessment for power system voltage stability with wind farms. The main work is as following:Because of the inherent defect in EM(Expectation Maximization) algorithm that is adopted to model wind farms, the fitting precision of WGMD(Weighted Gaussian Mixture Distribution) is reduced. According to this problem, an improved probabilistic modeling method based on DAEM(Deterministic Annealing Expectation Maximization) algorithm is proposed in this thesis. In making maximum likelihood estimation of parameters for wind farm mode ling, the DAEM algorithm can avoid the shortcoming that EM algorithm would easily converge to local optimization. The samples from the modeling of wind farm outputs are calculated by measured data of active power from wind farms. It is shown that the propo sed method can distinctly improve the calculation accuracy of modeling method.Then, this thesis presents an improved Markov Chain Monte Carlo(MCMC) simulation method. At present, Gibbs sampling that is widely used in MCMC simulation method suffers from complicated sampling iterations when gets accurate results from probabilistic load flow. According to the defect, an improved MCMC method based on Slice sampling is proposed in this thesis and is integrated into probabilistic load flow for wind farms integr ation system. It is shown that the proposed method can improve the calculation accuracy of MCMC method. Additionally, the Markov Chain generated by Slice sampling can reach a stationary distribution more quickly and stably than Gibbs sampling with the same iterations.Finally, according to the deficiency of probabilistic assessment for power system voltage stability with wind farms, an improved probability assessment for power system voltage stability with wind farms on L index is proposed. The L index is adopted to assess the voltage stability of PQ buses, including wind farms integration buses, and the sensitivity of L index to bus injection power is firstly derived. Combined with L index, the improved probabilistic assessment method is adopted to compute the risk of voltage instability, which could assess the voltage stability of wind farms integration system and identify weak buses. The results from IEEE 14-bus system and IEEE 39-bus system show that this method can present more accurate information for compensation measures of improving the static voltage stability of wind farms integration system.