燃气锅炉主汽压的RBF神经网络滑模控制策略设计

Design of a Sliding Mode Control Strategy for the Main Steam Pressure of Gas-fired Boilers Based on Neural Network

针对燃气锅炉主汽压控制的非线性、干扰大和模型参数易变的特点,常规比例积分微分(PID)控制方法难以取得满意的控制效果。利用滑模控制对系统参数变化和扰动不灵敏的优点,提出一种基于神经网络和滑模控制相结合的主汽压优化控制策略。采用径向基函数(RBF)神经网络调节滑模控制器的切换增益以降低其在平衡点的抖振,并通过在系统中设计干扰观测器实现对扰动的补偿。结果表明:与常规滑模控制和常规PID控制相比,不同工况下本文提出的控制策略超调量最多减少6.62%,调节时间最多减少57.45 s,且系统抖振小,跟随性能、抗干扰能力和鲁棒性能良好;工程实例表明,采用所提控制策略后,系统抖振比常规滑模控制降低33%,主汽压波动范围在-0.04~0.1 MPa,控制系统抗干扰能力得到显著提高。

To solve the problem that the conventional PID control method is difficult to achieve satisfactory control effects during the main steam pressure control of gas-fired boilers due to the characteristics of non-linearity, large interference and variable model parameters of the control system, an optimized control strategy was proposed based on neural network and sliding mode control, by taking the advantages of the sliding mode controller that is not sensitive to the parameter changes and disturbance of the system. The RBF neural network was used to adjust the switching gain of the sliding mode controller to reduce its chattering at the equilibrium point, while the disturbance compensation was realized by adding a disturbance observer to the system. Results show that compared with the conventional sliding mode control and the conventional PID control, the overshoot of the control strategy proposed under different working conditions could be reduced by up to 6.62%, and the adjustment time could be reduced by up to 57.45 s, with small chattering, good following performance, high anti-interference capability and strong robustness of the control system. Application results indicate that after adopting the strategy proposed, the system chattering has been reduced by 33%, compared with the conventional sliding mode control. The fluctuation range of the main steam pressure is within-0.04-0.1 MPa, and the anti-interference ability of the control system has been significantly improved.