基于自组织IT2FNN的城市固废焚烧过程炉膛温度预测控制
Furnace Temperature Prediction Control Based on Self-organizing IT2FNN for Municipal Solid Waste Incineration Process
-
摘要: 针对与城市固废焚烧(municipal solid waste incineration, MSWI)过程的稳定运行、燃烧效率、污染排放直接相关的炉膛温度难以有效控制的问题, 提出基于自组织区间二型模糊神经网络的炉膛温度模型预测控制(self-organizing interval type-2 fuzzy neural network model predictive control, SOIT2FNN-MPC)方法。首先, 采用IT2FNN建立预测模型以洞悉炉膛温度动态变化; 其次, 提出参数在线学习和结构在线自组织算法对模型进行动态优化以改善预测效果; 然后, 设计基于梯度下降法的滚动优化策略在线求解最优控制律; 最后, 对IT2FNN预测模型的收敛性和控制系统的闭环稳定性进行证明。基于北京某MSWI厂的实际过程数据仿真验证了所构建模型和控制器的有效性。Abstract: Furnace temperature determines the operational stability, combustion efficiency and pollutant emission of municipal solid waste incineration (MSWI) process. Effectively controlling the furnace temperature remains a central challenge. To address the above problems, this paper proposes self-organizing interval type-2 fuzzy neural network model predictive control (SOIT2FNN-MPC). First, IT2FNN served as a prediction model to predict the dynamic change of the furnace temperature. Second, an online learning algorithm of parameters and an online self-organizing algorithm of structure were introduced to dynamically optimize the model to improve the prediction effect. Then, a rolling optimization strategy based on gradient descent method was designed to solve the optimal control law online by optimizing the objective function. Finally, the convergence of the IT2FNN prediction model and the closed-loop stability of the control system were proven. The effectiveness of the constructed model and controller was verified through simulation based on actual process data from an MSWI plant in Beijing.
下载: