Abstract:

It is reported that most of the industrial energy consumption is attributed to the manufacturing process. The application of renewable energy resources in the manufacturing plant can contribute to alleviating the environmental pollution problems caused by the massive use of fossil fuels. It focuses on the micro-grid system including the manufacturing plant， renewable energy generation system and power storage system. Apart from the power generated by the renewable resources， additional power from the main electric grid is necessary in order to satisfy the power demand of manufacturing line at any time. A mixed integer linear programming model integrating production scheduling， maintenance planning and energy controlling is established to minimize the total power cost of manufacturing plant under the real-time electricity price. The power generated by the renewable resources can be discharged to the manufacturing plant or stored in the storage system who has a limited maximum storage level. The price per unit of power drawn from the main grid depends on the time period. And， the unit price during the peak hours is higher than that during the valley hours. Based on the classification of decision variables， a two-layer algorithm combining the genetic algorithm， heuristic rules and sub-problem exact algorithm is designed to solve the model. In the out layer， the variables related to the production and maintenances are searched in the meta-heuristic according to the evaluation results obtained from the inner layer. In the inner layer， the variables related to the energy are optimized using the minimum cost network flow algorithm. Compared with CPLEX， the effectiveness and the efficiency of designed GA are validated in the numerical experiments. Since CPLEX cannot get the optimal solutions for the large-sized problems， an effective algorithm based on relaxation method is designed to get the lower bound for the problem. The gap between the lower bound and the solution obtained by GA is smaller than 10%， which also validates the effectiveness of designed GA. Compared with the traditional independent decision-making policies， the total power cost can be reduced by 40 percent using our model. The optimal operational plan shows that production should be arranged to the period with low electricity price since production is a kind of activity consuming energy heavily. However， it is inevitable to process the jobs during the high-price hours considering the non-preemption of jobs and the constraint of production deadline. Therefore， the power generated by the renewable energy system should be proactively stored in the storage system and then discharged to the manufacturing plant during the corresponding period. Thus， the existence of renewable resources can help the company decrease the behavior of electricity purchasing during the high-price hours effectively， which will improve its competiveness in the fierce market.

Key words: production planning, preventive maintenance, real-time price, renewable resource