Fuzzy identification of the steam multivariable temperature system based on improved GK clustering algorithm

Boiler steam temperature system shows non-linear and time-varying, so the accurate modeling of steam temperature system is particularly important. A kind of method of fuzzy identification based on improved GK clustering algorithm (λ-sectional set fuzzy weighted GK clustering) is proposed in connection with the traditional Fuzzy clustering algorithm´s defects such as low precision and slow search speed. By analyzing the correlation of input and output as weighted coefficient of fuzzy clustering algorithm, it is employed to cluster the input data of sample space. A more appropriate division of the input data is achieved, at the same time the sectional set fuzzy GK clustering is proposed to identify the model structure off line to improve searching rate, the method confirms the premise parameter by improved fuzzy partitions clustering algorithm and the consequence parameters is decided by LS algorithm. In this paper, the simulation of the temperature control TITO system of the boiler can illustrate that the method is accurate and effective.