考慮單機台故障率與健康狀態的預防性維修週期與生產排程之決策研究

Abstract

在工業4.0與物聯網的環境下，數據的取得變得更加便利，透過各種不同感測器的監控，工廠對產線上的機台狀況也一覽無疑，當能夠有效監控機台健康狀態與不同的健康狀態對加工時間，以及機台故障率的影響，在物件加工的排程以及維修保養的規劃上，也能更加有效率，且有效的降低成本。 本研究之探討目標為考量機台健康狀態之單機台的預防性維修與生產排程的問題，當機台健康狀態分為完美、普通，故障時，以及物件加工時間隨著機台健康狀態而改變時，如何規劃預防性維修之週期，以及提出一個整數規劃模型以描述不同的加工順序，造成不同的維修成本的情況。本研究使用MAPLE 17.0與LINGO 11.0進行求解，隨著物件數的增加，問題的複雜度上升，導致欲求解之問題的求解時間過長，導致效率不彰。本研究將使用兩階段演算法對此模型進行求解，以期能夠迅速的得到品質不錯的可行解。透過不同的參數設定，好比加工物件數、機台狀態轉換率與維修保養成本等等，發現兩階段演算法在問題規模較大時，仍保持不錯的求解品質。

According to Industry 4.0 and Internet of things, it is easier to get data from machines through sensors which are equipped on machines. For factories, they can get information about machine health immediately. When factories understand the relationships between machine health and machine failure rate of processing time, they can do scheduling and arrange preventive maintenance effectively. In this research, we devote to single machine scheduling with consideration of preventive maintenance and machine health. We separate machine status into perfect, normal and breakdown. When machine status changes, the processing time of jobs will increase and the machine failure rate will increase. We try to figure out how to arrange preventive maintenance interval and sequence jobs with consideration of machine health. In the research, we use MAPLE 17.0 and LINGO 11.0 to solve our model; however, when number of jobs increases, the problem becomes complex and it takes a long to get the solution. We apply two-step heuristic method to solve our model, and hoping to get the solution with high quality. Through different parameter settings, such as machine failure rate, number of jobs and repair and maintenance cost, we find out that two-step heuristic can get solution with high quality in big problem.