基于单神经元Smith预估的过热汽温控制

针对典型大时滞大惯性、参数变化的电站锅炉过热汽温受控对象 ,设计了基于单神经元Smith预估过热汽温控制系统。在仿真实验的基础上 ,对单神经元Smith预估控制和最优常规PID控制进行了比较和分析。仿真结果表明 ,单神经元Smith预估控制能够充分利用神经元自学习、自适应的能力以及Smith预估补偿控制的优点 ,使系统的控制品质提高 ,而且具有更强的鲁棒性和自适应性

Main Stream Temperature Control Based on Single Neuron with Smith Predictor

Considering the controlled object of main stream temperature with large time delay and varying parameters in utility boilers, a single neuron controller with Smith predictor is presented. Based on the simulation, the comparison is made between single neuron controllers and optimal general PID controllers. The results indicate the proposed control system has advantages of the self-learning ability, the self adaptability and the compensation control so that the control quality, the robustness and the adaptability of the control system are superior to the general PID control.