Study For Modeling And Single Neuron Controlling Of Hydro-turbine Governing System

Hydro-turbine governing system, with responsibility for regulating the frequency to meet the requisition of power system, is one of the most important control systems that influence operating quality and security of power system. So study the hydro-turbine governing system, seek the ideal control method, have very important meanings.On the basis of analysing the principle of hydro-turbine governing system, this paper adopts the module modeling method, divides hydro-turbine governing system into several subsystems like penstock, hydro-turbine, generator, governor, models them separately by mechanism, thus the whole system math model is established. As for the francis turbine, this paper generalizes the characteristics of complete characteristic curve nonlinear model and internal character analysis nonlinear model. Utilize the turbine basic equations with considering some nonliner factors like the power loss, this paper establishs the analytic nonlinear model of hydro-turbine which has specific physics meaning and just needs few parameters that easily getted. Use this model, the hydro generator unit start-up simulation is processed, the result validates the availability of the established model.Hydro-turbine governing system is a nonlinear and time mutative system that is usually conducted following analysis and calculations:â‘ calculation of regulation guarantee under transiting process of acute fluctuation;â‘¡dynamic analysis under small fluctuation. On the condition of small fluctuation, the nonlinear model could be treated as linear model. This paper establishes the linear model of hydro-turbine and deducts its transfer coefficients. Then the optimized operation parameters are gained in the frequency disturbance and load rejection simulation. Through comparing dynamically, this paper analyses the impact on its dynamic performance if some governing system parameter changes.Presently, most microcomputer governors adopt PID control law. The traditional PID control law cannot deal with the conflicts of stability, accuracy and speediness. So parameters auto-tuning on-line must be adopted if stable operations under all conditions wanted to be guaranteed. This paper adopts the Single neuron controlling algorithm which has the function of self-learning and self-adaptive. Then a single neuron adaptive controller is designed. The simulation results indicate that the single neuron adaptive controller not only has a high controlling accuracy, but also has the characteristic of fast...