The Research Of Voltage-Reactive Power Control Equipment In A Substation

In this paper the complexity of voltage control in power system is discussed through the analysis of relationship between power system balanced states of voltage and reactive power. A mathematical modal is established by means of the theory of reactive power/voltage control in a substation. It proposed a new control strategy because of the research of disadvantages resulting from the traditional nine-zone diagram methods. The control strategy is based on nine-zone diagram and combined with artificial neural network which make a decision about the action of transformer onload tap changers and the cut and input of capacitor. It can reduce blindness of action at border areas of nine-zone diagram thus the times of action can be decreased. To solve the utilization of the neural network well, the problems existing in the BP neural network used in control strategy are investigated from a few different points of view such as initial weights, numbers of neurons in the hidden layer and transfer function. A training method using gradient descent with momentum and adaptive lr backpropagation is proposed based on traditional BP algorithm. Simulation results for a practical example show that the BP neural network is more accurate.A method based on optimal curves of bus voltage and reactive power for computing control range setting of voltage and reactive power in a substation is proposed in this paper. The nine-zone diagram using the method and combining with the neural network can be taken effect in reactive power/voltage control in a substation.A high capability voltage-reactive power integrative control equipment based on industrial computer and Windows operation system is presented in this paper. It solve the difficulties resulting from the limitation of software and hardware of old devices and the disadvantages of traditional nine-zone diagram control methods.