Fast Calculation Method For Voltage Stability Modulus Based On Neural Network

With the continuous expansion of grid,the transmission situation with heavy load,long-distance and high-voltage has become increasingly prominent,forcing the power system often runs on the edge of voltage stability.There are many voltage collapse accidents in power system of the world.Therefore,it is very important to evaluate the voltage stability margin of current system quickly and accurately.The traditional continuous power flow calculation method cannot meet the requirements of speed and accuracy.And the introduction of neural network as an auxiliary tool for voltage stability on-line rapid assessment provides a practical solution.In this paper,the AC power system can be equivalent to the real-varying circuit based on the comprehensive dynamic equivalent circuit analysis method;and the critical point criterion of the system voltage stability is deduced;and thus the normal impedance modulus margin index is proposed to evaluate the system static voltage stability.At the critical point of the voltage stability,it is proved that the normal impedance modulus margin index is equivalent to the Thevenin impedance modulus margin,which can well reflect the current voltage stability margin of the system.Considering that the distributed generation(DG)has a certain influence on the voltage stability of the power system,and DGs can be regarded as the"negative" load and the comprehensive dynamic analysis method based on the real variable can be extended to the voltage stability analysis of DGs system.The simulation results show that the proposed method is an effective and practical method to analyze the static voltage stability of power system;and the normal impedance modal margin has a better linear relationship than the Thevenin impedance modulus margin and is suitable for the online rapid prediction of voltage stability based on neural networks.The paper proposes a fast calculation method of voltage stability margin index based on neural network.The calculation problem of the system voltage stability margin index is transformed into the online prediction problem of the neural network by introducing the neural network as an auxiliary tool;then the voltage stability assessment model based on BP neural network optimized by the intelligent algorithm is established.In this model,the neural network is trained with the load normal impedance modulus as the sample value;The nonlinear mapping relationship between the active power(reactive power)and the normal impedance modulus margin of load node was established.Meanwhile,the weights and thresholds of the neural network are optimized by the artificial intelligence algorithm,which greatly improves the prediction precision.Simulation results of IEEE standard system shows that the calculation speed of the normal impedance modulus margin is accelerated greatly compared with the traditional power flow calculation;and it is more convenient to realize online prediction on system voltage stability;thus,the quickness and accuracy of the established model is verified.