Study On Boiler Superheated Steam Temperature Control Based On Neural Network Inverse Models

Boiler superheated steam temperature (SST) is one of the important parameters closely related to the safe and economic operation of a coal-fired power unit. Superheated steam temperature either too high or too low will pose a threat on the operation safety and economy. At present, cascade PID control methods are usually used for SST control. Because the boiler superheater system is relatively complex, with large delay, large inertia and higher nonlinearity, especially when the load changes, the original PID controllers are often no longer with good control quality and thus need resetting. PID parameters're-setting is often a time-consuming and labor-intensive work. With the development of computer technology in recent years, the neural network technology has got a rapid development and widely used in power plant modelling and control, providing a new idea for SST optimization control.By understanding the characteristics of the superheater system, this paper studies the neural network inverse control scheme deeply based on neural network inverse system modelling. Firstly, the paper discusses the structure of BP neural network and its training methods. Then the basic concept of neural network inverse control, neural network inverse controller's design principle and method are introduced. The characteristics of superheater system and the influence factors of superheated steam temperature are analyzed. On this basis, the neural network inverse system models for the superheaters of a 300MW sub-critical boiler unit and a 600MW supercritical boiler unit are built, trained and validated separately with MATLAB neural network toolbox. The control schemes are determined and NN inverse controllers are designed and then programmed with MATLAB language. Control tests are made by communicating real-time with the full-scope power plant simulators. It is shown by tests that,compared to the original cascade PID control, the neural network inverse control scheme can significantly reduce the overshoot and the stabilization time, thus significantly improve the quality of superheated steam temperature control.