Differential Evolution-Based Identification of the Nonlinear Kaplan Turbine Model

Abstract

In this paper, various new models of double-regulated Kaplan turbines are proposed, whose parameters are determined on the basis of field measurement data acquired during normal operating conditions. The model being identified is an extension of the nonlinear single-regulated turbine model obtained through an approximation function that defines the relationship between the wicket gate opening and runner blades angle. To determine the unknown parameters of the approximation function, a stochastic search algorithm called differential evolution (DE) is used. This paper focuses on the progressive development of the DE algorithm-based methods that are applied to determine different forms of the approximation functions, until an optimal form is attained that ensures the best possible agreement between the measured and calculated responses.