Probabilistic Power Flow Calculation For Power Systems Based On Improved Vine-Copula And Quasi-Monte Carlo Simulation

Power flow calculation is the foundation of state analysis and system planning for power systems,which ensures the safe and stable operation of power systems.However,in recent years,with the rapid development of new energy power generation,a large number of distributed power sources,such as wind power generation and photovoltaic power generation,have been incorporated into the power system.Due to the fact that the output of wind power and photovoltaic power is directly determined by the wind speed and illumination of their location,distributed power generation itself has strong volatility and uncertainty.The large-scale grid connection of distributed power generation will inevitably bring enormous challenges to the power system,and will also make the power flow calculation problem more complex.At the same time,many uncertain factors also make traditional deterministic power flow calculation methods no longer suitable for modern power systems.In order to solve the above problems,the concept of probabilistic load flow emerged as the times require.PLF computing can well deal with uncertain factors in power systems.Since it was proposed,it has received extensive attention from researchers at home and abroad,and has made a significant breakthrough.Due to the fact that distributed power sources of the same type in the same region or even adjacent regions often have a certain correlation relationship due to the impact of environment and climate,in order to keep the algorithm results closer to the actual situation,the correlation between random variables should also be taken into account when performing PLF calculations.Therefore,in view of the current problems faced by PLF calculation,such as slow convergence speed,low computational efficiency,and imperfect correlation model establishment,this paper proposes a power system PLF calculation method based on improved Vine-Copula theory and Quasi Monte Carlo method.This method first uses Vine-Copula theory to establish a correlation model between random variables.Compared to the establishment of correlation models using traditional PLF calculation methods,Vine-Copula method has good flexibility in establishing correlations between high-dimensional random variables,and its model fitting effect is also good.Secondly,the QMC method is used to sample data samples.The low difference sequence generated by the QMC method is superior to the pseudo random numbers in traditional sampling methods due to its good uniformity,so the QMC method has a good convergence speed.The PLF calculation method proposed in this paper takes into account both computational efficiency and accuracy,as well as the correlation between input random variables in power systems.Finally,simulation verification was conducted on an IEEE30 node system,and the comparison proved that the PLF calculation method proposed in this paper has advantages in correlation modeling and computational efficiency.Through the research in Chapter 3,it was found that the Kendall’s correlation coefficients between the same type of distributed power sources in the same or adjacent areas are relatively close.Therefore,when establishing a correlation model,the D-vine Copula structure can be directly selected to construct a correlation model between the output of distributed power sources.Therefore,based on the research in Chapter 3,a PLF calculation method based on D-vine Copula theory and randomized QMC is proposed.This method improves the QMC method proposed in Chapter 3 by using the Random Shift method to randomly shift the LDS generated by the QMC method.By randomly shifting the LDS,not only can the calculation error be estimated by calculating the confidence interval,but also the computational efficiency of the PLF algorithm can be improved.Finally,simulation verification was conducted on an IEEE118 node system to demonstrate the effectiveness of the proposed PLF calculation method.The simulation results show that the proposed method in this paper can accurately describe the correlation between variables compared to traditional PLF calculation methods,and has high computational efficiency and accuracy.