Skip to Main Content
The hydraulic turbine system is a complex and nonlinear controlled object and it is difficult to obtain its dynamic model via accurate mathematic theory. In this paper, a bayesian inferring solution is proposed for the inverse model structure learning method of hydraulic turbine system. And in the training of the bayesian inferring model, off-line training procedure includes the identification of the parameters in the called threshold matrix D optimized by evolutionary algorithms(EAs). The sliding window driven method is used to sustain the scale of the bayesian inferring model when on-line prediction applications. The presented bayesian inferring model was applied to identify the inverse model of the hydraulic turbine system. It is shown from the simulation results that the given bayesian inferring model provides a attractive method for the inverse dynamic characteristic modeling of the hydraulic turbine system.