动态环境下多电机驱动带式输送机自学习优化控制

Multi-motor-driven belt conveyor is a kind of high energy-consuming equipment which has been widely used in coal mining, chemical industry and bulk terminals. It is of terrible working environment, complex work-condition, and uncertain mechanical parameters. The existing control systems for belt conveyors cannot interact autonomously with dynamic environment and thus fail to achieve self-learning and adaption, which is likely to cause energy waste and security risks. This project will take the high-order and multi-time-scale dynamic characteristics and parameter uncertainties of multi-motor-driven belt conveyors into consideration. The self-learning optimal control problem for multi-motor- driven belt conveyors under dynamic environment will be studied by using the singular perturbation theory, adaptive dynamic programming method, parametric model reduction technology and receding horizon optimization principle of model predictive control. This project will propose self-learning optimal coordinated control methods for the multi-motor-driven system, self-learning dynamic optimization methods for the speed set-point and stability analysis methods for the whole closed-loop control system of the multi-motor-driven belt conveyors. In addition, this project will also develop a hardware-in-the-loop simulation platform and a physical experiment platform, through which the feasibility and effectiveness of the proposed results will be illustrated. This research can fundamentally resolve the key difficulties of self-learning optimal control for multi-motor-driven belt conveyors under dynamic environment, improve the energy-saving effectiveness and safety, and provide a theoretical framework to solve the self-learning optimal control problem of general multi-time-scale complex systems.

多电机驱动带式输送机是广泛应用于煤炭、化工、码头等行业的高耗能装备，其工作环境恶劣、运行工况复杂、机械参数不确定。现有带式输送机控制系统不能与动态环境自主交互，无法实现自学习和自适应，容易造成能源浪费和安全隐患。本项目将综合考虑多电机驱动带式输送机的高阶次多时间尺度动态特性和参数不确定性，利用奇异摄动理论、自适应动态规划方法、参数化模型降阶技术和模型预测控制滚动优化原理研究动态环境下多电机驱动带式输送机自学习优化控制问题。通过研究，拟提出多电机驱动系统自学习优化协调控制方法、带速设定值自学习动态优化方法和整体闭环控制系统稳定性分析方法；研制半实物仿真平台和物理实验平台，验证所得方法的可行性和有效性。研究成果将从根本上解决动态环境下多电机驱动带式输送机自学习优化控制面临的关键难题，显著提高带式输送机的节能效果和安全性，并为研究一般性多时间尺度复杂系统自学习优化控制问题提供理论与方法支撑。

广泛应用于煤炭、化工、码头等行业的多电机驱动带式输送机，工作环境恶劣、运行工况复杂、机械参数不确定。现有带式输送机控制系统不能与动态环境自主交互，无法实现自学习和自适应，容易造成能源浪费和安全隐患。本项目综合考虑了多电机驱动带式输送机的高阶次多时间尺度动态特性和参数不确定性，利用奇异摄动理论、自适应动态规划方法、参数化模型降阶技术和模型预测控制滚动优化原理研究了动态环境下多电机驱动带式输送机自学习优化控制问题。通过项目组的合作研究，建立了刚柔耦合多电机驱动系统的数学模型，提出了自学习优化协调控制方法；建立了带速设定值优化模型，提出了动态优化方法；确立了整体闭环控制系统稳定性分析框架和方法；研制了带式输送机半实物仿真实验平台和刚柔耦合多电机物理实验平台，验证了所得方法的可行性和有效性。相关研究成果共发表了20篇学术论文，授权国家发明专利9项，南非发明专利1项，获得软件著作权5项，培养研究生10人，相关成果荣获教育部高等学校科学研究优秀成果奖自然科学二等奖1项和中国自动化学会自然科学二等奖1项。本项目解决了动态环境下多电机驱动带式输送机自学习优化控制面临的关键难题，仿真实验表明能够提高带式输送机的节能效果和安全性。