退火炉加热过程工况分类模型与多模态模糊自适应板温控制

According to the demand for high precise control of strip temperature in annealing furnace heating process with fluctuating and fast-changing operating modes, multi-mode classification model and multi-modal fuzzy adaptive control method for heating process of annealing furnace are studied..Firstly, to solve the problem of multi-mode classification for annealing furnace heating process, the extraction method of key characteristic parameters in the multiple operating modes and self-organizing fuzzy clustering algorithm are studied. Based on the optimization of fuzzy subsets for operating modes, a multi-mode classification model for steady modes is established. Then, the mode forecasting based on dynamic Bayesian network model is made to establish a migration model. Secondly, aiming at the control of strip temperature, the method to mine fuzzy control rules based on data-driven is used to design a fuzzy controller with variable domain and parameters self-tuning for steady modes. The fuzzy controller based on the gas flow prediction model is designed for operating modes migration process. In addition, a strategy of coordination and online update for the multi-modal fuzzy controller is applied to achieve the bumpless switching of multi-modal fuzzy control and fuzzy rules updating online..Finally, the simulation and experiment system is set up to verify the correctness and effectiveness of theory technology, and do the industrial applications. In this research, the new design methods of data-driven intelligent control are provided to satisfy the requirements of the industrial process under complex operating modes, which is represented by annealing furnace heating process. Therefore, the research of this project is of great academic significance and application value.

本项目面向工况复杂变化的退火炉加热过程板温高精度控制的需求，研究其工况分类模型与多模态模糊自适应板温控制方法。针对退火炉加热过程工况分类问题，研究复杂工况模式关键特征属性参数提取方法与工况模式自组织聚类算法，通过工况模糊子集的优化，建立稳态工况模型；基于动态贝叶斯网络模型进行工况预测，建立工况迁移模型。针对板温控制，通过数据驱动型模糊控制规则的挖掘，设计稳态工况下变论域参数自整定模糊控制器；基于煤气流量预测模型，设计工况迁移过程模糊控制器；研究多模态模糊控制协调与更新策略，实现变工况情况下多模态模糊控制无扰切换与控制规则的更新。建立仿真与实验系统，验证理论技术的正确性和有效性，并进行工业应用。本项目的研究，将为以退火炉加热过程为代表的复杂工况环境下的工业过程提供一种数据驱动的智能控制新方法，具有重要的研究意义和应用价值。

本项目面向工况复杂变化的退火炉加热过程板温高精度控制需求，结合机理分析与数据分析，建立了基于模糊子集的退火炉加热过程工况分类模型，设计了面向多工况的变论域参数自整定模糊控制器，建立了面向工况迁移过程的自适应板温控制参考模型，制定了多模态模糊控制协调与更新策略，提出一种面向退火炉加热过程的多模态模糊自适应板温控制系统结构，实现了退火炉加热过程板温高精度控制。针对工况分类与识别问题，项目组建立了基于模糊子集的退火炉加热过程工况分类模型，挖掘出退火炉加热过程各工况特征；制定了基于多尺度特征融合的退火炉加热过程多模态识别策略；建立了基于长短时记忆和门控循环单元的工况识别模型，实现了工况的实时精确识别。针对各工况下板温控制问题，项目组提出了基于异常数据检测的模糊规则快速挖掘方法，设计了面向未知论域的变论域参数自整定模糊控制器。针对工况迁移过程板温控制问题，项目组建立了基于遗传算法和支持向量机的煤气流量预测模型和基于AdaBoost的煤气流量预测模型，根据工况的变化直接计算所需煤气流量。针对多模态模糊控制协调问题，项目组提出了基于所需模糊规则决策的模糊输出计算方法，设计了一种改进型等价输入干扰估计器。针对多模态模糊控制更新问题，项目组建立了一种退火炉加热过程板温动态支持向量回归模型，提出了一种动态子空间建模方法，还提出了具有扰动补偿的非参数模型自适应控制方法，并设计了一种启发式模糊逻辑系统。最后，为了验证提出方法的有效性，项目组构建了退火炉加热过程仿真模型，搭建了退火炉加热过程仿真实验平台，开发了退火炉加热过程智能控制与优化系统，进行了工业应用分析，仿真与应用结果表明，本项目提出的方法具有重要的研究意义与应用价值。