Probabilistic Transient Stability Assessment Of Power System Considering Wind Power Uncertainties And Correlations

With the maturity of wind power generation technology,more and more wind farms are integrated into the power grid,increasing the penetration level of wind power.The intermittent,volatile,uncertain characteristics of wind power and the correlation between adjacent wind farms increase the uncertainties in the power system,which have a great impact on the transient stability.Therefore,the probabilistic transient stability assessment has been more and more essential.Based on the practical dynamic security region,the solution of the probability transient stability index can be simplified from the n-dimension integration of multi-variable probability density function to a simple analytical calculation.However,considering the non-normal and non-independent distribution of the wind farm output,the existing convolution method and point estimation method can not solve the problem while meeting the requirements of calculation accuracy and speed for high wind power penetration level.Therefore,this thesis is aimed to quickly and accurately evaluate the probabilistic transient stability based on the probabilistic transient stability index derived from practical dynamic security region for any wind power penetration level considering the wind power uncertainties and correlations.Based on the probabilistic transient stability index derived from practical dynamic security region,this thesis proposes an improved multi-point estimation method combining inverse Nataf transformation.The correlation between non independent random variables was modeled by inverse Nataf transformation.Then the transient stability probability was calculated by the multi-point estimation method based on Gauss-Hermite integral.Specifically,the hyper-plane expression of practical dynamic security region boundary was used to determine the contribution of non-normally each distributed random variable and an improved particle swarm optimization algorithm was used to obtain the optimal sampling number of each non-normally distributed random variable that meets the accuracy of transient stability probability solution in order to avoid the blindness and redundancy of the existing multi-point estimation method in selecting the number of mining points.The results of the case study on New England 10-machine 39-bus system showed that: the proposed method has the advantages of high calculation accuracy and calculation speed for any wind power penetration level and wind power correlation.When the wind power penetration rate is low,compared with traditional three-point estimation method,the calculation time of the proposed method is 56% shorter and the maximum calculation error can be reduced by 52.1%;the proposed method can achieve the same calculation accuracy as the Gauss-Hermite integration nine-point estimation method while shortening the calculation time by 25%.When the wind power penetration rate is high,the maximum calculation error of the proposed method can be almost 1/8 of the traditional three-point estimation method;the calculation time and maximum calculation error can be reduced by 34% and 43.5%,respectively,compared with the Gauss-Hermite integration nine-point estimation method.As a useful supplement to the traditional deterministic assessment index,the transient stability probability index further reveals the operation condition of the power system;wind power correlation and wind power penetration level have significant influence on the transient stability probability of the system,and the critical lines that are more sensitive to the wind power correlation and wind power penetration level can be screened according to the degree to which they are influenced.