模糊神经网络预测控制在配料系统中的应用

Application of Fuzzy Neural Network Predictive Control in Material Proportioning System

针对水泥生产过程中皮带配料系统的惯性、滞后、非线性及现场干扰频繁等特点,设计一种模糊神经网络预测控制算法,将模糊控制、神经网络与预测控制相结合,增强算法的自学习、跟踪与抗干扰能力,神经网络预测模型有效地补偿了传统预测控制基于线性模型的局限性。将该控制算法用于皮带配料控制系统中,仿真结果表明,物料流量控制效果优于传统的PID控制,配料精度有了明显的提高。

In accordance with the technique features of material proportioning belt system in the cement production, e.g., inertia, time lag, non-linearity and frequent disturbance in work field, a fuzzy neural network predictive controller based on neural network prediction model is designed. By combining fuzzy control, neural network and predictive control, it can enhance self-studying, tracking and anti-interference capabilities of the algorithm, and the neural network can compensate with the limitation of conventional predictive control that based on linear model. With this algorithm the weigh belt is controlled, and the simulation experimental curves show that the control effect of the material flow is effective, and the precision of ingredient proportion has evidently improved.