摘要
回转窑的生产过程是一个复杂的物理化学反应过程,具有大惯性、纯滞后、非线性等特点。工艺过程复杂多变,难以得到精确的数学模型,常规控制算法难以满足控制要求。本文提出一种利用神经网络作为预测模型,遗传算法作为滚动优化策略的预测控制算法。将这种算法用于回转窑温度控制系的仿真研究表明,该控制方案具有较强的鲁棒性和自适应能力,明显优于传统的PID控制。
The production process of rotary kiln is a complicated physicochemical process of reaction and has the characteristics of big inertia, pure time-delay and nonlinearity. Exact mathematic model is hard to be obtained, and conventional control can not meet the request of production. A robust generalized predictive control based on neural network model with genetic algorithm is proposed in this paper. Simulation results show that this method has perfect control performance, strong robustness and adaptive ability, which is significantly superior to the traditional PID control.
出处
《武汉工业学院学报》
CAS
2007年第3期80-82,共3页
Journal of Wuhan Polytechnic University
关键词
回转窑
神经网络
预测控制
鲁棒
rotary kiln
neural network
predictive control
robust
