Improved neural network models for coordinated controller design of supercritical coal-fired power generating unit

A supercritical coal-fired power generating unit is a typical multi-variable strong coupling system with large time-delay, slow time-variant and nonlinear characteristics, so it is of vital significance to study its operation characteristic by means of modeling method, and improve the coordinated control quality with model-based advanced intelligent control strategy. In this paper, an improved Elman neural network that was used to build a nonlinear mathematical model of the load and main steam pressure characteristics for a 600MW supercritical coal-fired power generating unit was established. The improved Elman model has time-delay inputs and time-delay outputs feedback. The training data is the operation data over wide-range load-changing conditions for a 600MW supercritical boiler unit. The off-Line and on-Line verification tests for load-changing conditions showed that the improved Elman model with time-delay inputs and outputs feedback can fit the complex non-linear, dynamic characteristics between three inputs(fuel, feed-water flow, turbine governing valve opening) and two outputs (unit's load, main steam pressure) with high precision and strong generalization ability. Compared with the original Elman model, the improved model is favorable for coordinated controllers' design with simple structure, high precision and strong generalization ability. It can meet the engineering application requirements to be used as a prediction model to build an intelligent controller for supercritical coal-fired power generating unit coordinated control.