面向可持续制造的生产与维护及能源计划的耦合建模与优化研究

This project takes account of the energy supply-side and proposes to study the three interrelated aspects of production scheduling, maintenance policy and energy planning arising from the operational management of sustainable manufacturing systems. Three problems are investigated including the problem of minimizing the system’s expected total cost, the peak load management and the emergency management. The constraint between production and maintenance is analyzed from the viewpoint of availability and reliability. The expectation formula of the electricity cost under the differential price mode is derived. For the uncertainty of the machines’ random breakdowns, considering the impact of the real time constraint of peak power, a proactive approach is proposed based on the combining of the initial plan and reactive rules. Then, the algorithm based on the division of decision variables is devised to solve the stochastic mathematical model with the objective of robustness. For the uncertainty of emergency, a dynamic responsive mode is constructed, as well as the approximation model at each decision node. Then, the iterative algorithm based on the two-stage rolling approach is devised to solve the model. All the studied modeling and solving methods not only extend current theory of operational management, but also provide core support for the scientific decision-making process of the company in the energy area.

本项目以面向可持续制造的车间生产运作管理为背景，考虑能源供给侧的结构特点，联合决策生产调度、设备维护、能源计划三个相互耦合的系统层面，提出了车间运作综合期望成本最小化、峰值电量管理以及停/错峰要求下的应急管理三个相互关联的理论问题。首先构建生产与维护在可用度及可靠度方面的制约关系，并推导差异性电价模式下系统运行的能耗成本表达式，为耦合数学模型的建立与优化提供研究基础。针对设备随机故障的内部不确定性，考虑电量峰值约束的实时性对系统的影响，建立初始基础计划与反应调度相结合的前摄优化框架，并设计基于变量分离的嵌套式搜索算法以求解该鲁棒性随机数学模型。针对突发事件的外部不确定性，建立动态响应模式和决策节点上的近似数学模型，并设计基于两阶段滚动优化决策逻辑的搜索算法。以上研究的建模和算法都是对现有生产运作理论的扩展与延伸，同时也为企业在能源时代下车间运作的科学决策提供核心支撑。

本课题关注双碳背景下的能源供给侧新特点，聚焦于制造企业的绿色化转型、智能化决策，重点分析了分时电价等电费模式对企业用电成本的影响、峰值电量阈值对企业运作计划的制约、突发临时限电对企业生产的扰乱等三个方面的情况。对于给定设施设备和既有硬件条件下的车间运作管理，事实上涵盖了制定生产、维护、能源三个层面的计划方案；与传统的各层面分属于不同领域而独立研究有所不同，本课题重点面向车间整体环境，充分考虑了各层面内部的特性以及彼此之间的关联关系，研究其系统建模与优化决策的基本理论与求解方法。通过对相应问题建立数学模型并设计精确算法/启发式算法进行求解，数值实验结果表明所得耦合方案，实现了总体运营成本的最优性、计划方案的可行性与系统运行的稳定性。相关研究成果不仅发表于国内外一流学术期刊之中，而且对于企业生产实践具有直接的决策参考价值，可以帮助企业更好地适应国家能源的新政策，助力我国双碳战略目标的早日实现。