配煤过程基于神经网络和数学模型的专家控制(英文)

Expert Control Based on Neural Networks and Mathematical Models for the Coal Blending Process

介绍一种专家控制策略 ,它解决配煤过程控制中精确地决定和跟踪每种煤的目标配比这两个关键问题 .首先基于统计数据和经验知识 ,建立神经网络 ,数学模型和规则模型 .然后提出一种决定目标配比的专家推理方法 ,它结合了神经网络 ,数学模型和规则模型 .最后使用一种分布式PI控制方案 ,跟踪决定的目标配比 .这种专家控制策略由一个专家控制器和一个分布式控制器实现 .

This paper describes an expert control strategy to solve two key problems in the control of the coal blending process, which are to accurately determine the target percentage of each type of coal to be blended, and to track them. First, neural networks, mathematical models and rule models are constructed based on statistical data and empirical knowledge. Then, an expert reasoning method combining the neural network, mathematical models and rule models is proposed for determining the target percentages. Finally, a distributed PI control scheme is used for tracking the target percentages. The proposed strategy is implemented by an expert controller and a distributed controller. The results of actual runs show the validity of the strategy.