The Research Of Model Identification Based On Real-time Data In A Power Plant

As one of the main branches of the automatic control science, system identification has been applied in many fields. Based on system identification technique, under the real-time data, taking the power plant thermal process for research object, this paper is aimed to study the thermodynamic system in thermal power plant from the angle of entropy theory, analyze the major influence factor of system identification and point out the problems of concern. Aiming at the effect of coal feeding on drum pressure, the classical identification method, least square method, genetic algorithm and neural network are used to identify the coal feeding-drum pressure model, and comparison was made to identify the advantages and disadvantages. At last, using the Levenberg-marquardt neural network, the multi-input multi-output system, whose inputs are coal feeding, water feeding and desuperheater spray and outputs are main steam temperature and main steam pressure, is identified with a fine result.