面向可持续制造的机械加工车间能效意识控制与优化研究

In China, the power required by mechanical machining industry is both enormous and inefficient. To achieve energy-saving in the industry, this project aims at improving energy efficiency in mechanical machining shops from machine-level, system-level, and shop-level perspective. From the machine-level perspective, as the machine conditions are variable along the time, the time-varying machine parameters and real-time operational information of machines will be determined by analyzing the energy signatures of machines and employing machine learning methods. From the system-level perspective, the Markov model of mechanical machining systems will be established, several energy conscious control strategies for gathering scattered energy-saving opportunities will be proposed according to the real-time operational information, then the energy-saving efficiency of these strategies will be proved by simulation. From the shop-level perspective, the workload of the mechanical machining systems with the strategies employed will be evaluated using the state space equations, the temperature will be predicted using the convolution neural networks, then the energy model of the mechanical machining shop considering the heating, venting, and air-conditioning (HVAC) system will be established to improving the energy efficiency of the entire shop. Finally, in order to improve the feasibility and effectiveness of the research, an energy management prototype software will be developed for mechanical machining industry. The project will advance the theories of sustainable manufacturing in mechanical machining shops, and provide a technical support for sustainable manufacturing in the industry.

我国机械加工行业的能耗问题非常严重，能源节约前景很大。为此，本项目以机械加工车间为研究对象，从设备、系统及车间等三个方面开展面向可持续制造的机械加工车间能效意识运行控制与优化研究：1）考虑设备参数和性能时变的特点，分析能量信号蕴涵的变动因素，应用机器学习辨识设备时变参数和实时运行状态信息；2）在此基础上，建立机械加工系统马尔科夫模型，根据实时运行状态提出能效运行控制策略，验证对零散分布的设备节能机会的聚集效果，通过将设备变迁至低能耗状态降低能耗；3）应用状态空间方程评估受控运行的机械加工系统的工作负荷，应用卷积神经网络预测温度，构建考虑暖通系统响应滞后性的机械加工车间能量耦合模型，综合优化车间能效；4）研发面向机械加工行业的节能服务原型软件，验证本项目理论研究的可行性和有效性。本项目研究成果将有助于提升机械加工车间可持续制造的基础理论，并为可持续制造理论研究在机械加工行业的应用提供技术支持

“可持续制造”是可持续发展对制造业的必然要求。可持续制造不仅强调生产成本、效率的经济指标，同时关注能量消耗、环境排放等绿色指标。本项目以机械加工车间为研究对象，从设备、系统及车间等三个方面开展面向可持续制造的机械加工车间运行控制与优化研究，实现车间经济指标和绿色指标的协同优化。在设备层面，搭建设备功率监控软硬件模块采集机床运行过程能量数据，分析机床运行过程的状态变化及相应的能量特性。考虑设备参数和性能时变的特点，应用基于滑动窗口的分段线性表示方法在线辨识加工周期。在系统层面，基于设备运行状态模型和可靠性模型建立机械加工系统的能效运行控制模型，采用自适应模糊控制方法实现基于实时信息的设备状态变迁，以降低能耗；建立机械加工系统能效意识调度和加工参数集成优化模型，设计三种能效运行控制策略，通过理论和实验验证其有效性；建立生产线系统生产计划与设备故障的集成优化模型，通过仿真验证模型的可行性，进一步分析不同结构配置的影响。在车间层面，基于暖通空调工况数据，采用基因表达式编程方法建立暖通空调设备能耗模型，并使用平均绝对百分比误差、均方误差、拟合优度等三个指标进行效果评估。考虑暖通空调系统与机械加工系统的能量交换关系，利用辐射时间序列方法建立制造环境热量模型。建立暖通空调系统运行优化模型，以能耗为优化目标，考虑设备启停、舒适性、热交换平衡等约束条件，运用遗传算法求解。本项目研究成果丰富和发展了机械加工车间可持续制造的理论和技术，有助于提升我国制造企业竞争力和可持续发展能力。本项目在国内外期刊上发表论文6篇，授权国家发明专利2项，获计算机软件著作权登记2项。