| 摘要: | 本研究發展自動化產線的排程系統。目標是提供工作站發生異常時的即時排程，以及針對工作站主控台中的可程式邏輯控制器(Programmable Logic Controller, PLC)設計整合方式。本研究先設計自動化產線排程系統之架構，包含訂單模組、排程模組、派工模組、工作站模組，異常監控模組、回報模組。然後，設計它們之間的連結方式，包含資料、格式與交換方式。根據實際自動化產線的問題，發展排程模組之核心。設計派工模組使PLC專心執行排程任務。工作站模組除了執行PLC的程式外也會記錄與回報實際製程的資料、機台狀態與異常。當工作站主控台偵測異常發生並發出通知時，觸發排程系統的重排機制。此時，異常模組比對現有和排程的資料，提供排程模組待處理的任務以產生新的結果。
在驗證實作上，本研究應用Google OR-Tools作為排程核心，以SQL資料表作為模組間的溝通介面，將實際工作站與排程系統整合。驗證測試包括正常情況、全線停工以及單個工作站故障的情形的排程。從實作和預期的比對可以發現此實驗有些微的誤差，來自於實際的工時的設定差異。整體而言，此驗證結果可證明此系統的設計理念，包含實現工作站發生異常後的即刻重新排程以及降低PLC與排程系統的整合工作量。
;This study develops a scheduling system for an automation production line. The objective is to provide real-time scheduling when an abnormality occurs in a workstation and to design an integration approach for the Programmable Logic Controller (PLC) as the main console of the workstation. In this study, we first design the architecture of the automated production line scheduling system, including order module, scheduling module, dispatch module, workstation module, exception monitoring module, and report module. Then, the connection methods between them are designed, including data, format and exchange methods. According to the actual problems of the automation production line, the core of the scheduling module is developed. The dispatch module is designed to make the PLC to focus on the scheduling task. The workstation module will not only execute the PLC program but also record and report the actual process data, machine status and anomaly. When the workstation console detects the anomaly and sends a notification, it triggers the rescheduling mechanism of the scheduling system. At this time, the anomaly module compares the existing and scheduled data and provides the scheduling module with the unresolved tasks to generate new results.
For verification, this study applies Google OR-Tools as the core of scheduling and uses SQL data table as the communication interfaces among modules to integrate the actual workstations with the scheduling system. The validation test includes the scheduling of normal conditions, full-line downtime, and single workstation failure scenarios. From the comparison between the actual and the expected results, it can be found that there are some small errors in this experiment due to the difference in the actual working hours settings.
In overall, the validation results proved the design concept of the system, including the realization of immediate rescheduling after a workstation occurred abnormality and the reduction of the integration workload of PLC and scheduling system. |
