可靠度工程技術在馬達動力系統設計開發上之應用

Abstract

由於電動車與油電混和車的發展，馬達動力系統逐漸被大量使用於車輛載具上。而針對使用於車輛載具的馬達動力系統而言，人們除需關注馬達本身的效率外，也須考量驅動控制、製造、可靠度、散熱等各方面的問題。其中，馬達動力系統開發時，會因製造品質、材料特性與零組件裝配等問題，影響馬達動力系統的可靠度，因此，探討如何將可靠度工程技術應用於馬達動力系統的設計與開發成為了一個重要議題。 本研究以國內製造的一顆100 kW三相鼠籠式感應馬達為範例，執行該馬達之失效模式與影響性分析(Failure Mode and Effect Analysis, FMEA)，而後依據FMEA結果進行故障樹分析(Fault Tree Analysis, FTA)，探討馬達失效與其零組件故障的因果關係，接著導入各零組件的失效率，以延伸隨機派翠網(Extended Stochastic Petri Net, ESPN)進行馬達壽命評估。本研究分別使用MIL-HDBK-217F N2與NSWC-11兩套可靠度預估方法，估算各零組件的失效率(Failure Rate)與平均失效時間(Mean Time to Failure, MTTF)；而後再推估馬達系統之壽命分布與可靠度；最後並進行敏感度分析，探討各零組件故障對系統失效的影響，以篩選出關鍵零組件。研究結果顯示，所探討感應馬達壽命可依指數分布嵌合，其平均失效時間為9.86年；而依據FMEA關鍵性分析矩陣以及敏感度分析結果可知，馬達系統中，風險較高的零組件依序為定子繞組、軸承與轉子繞組。

The reliability issues of an electrical motor and its components are studied in this thesis, and a domestically manufactured 100 kW three-phase squirrel-cage induction motor is taken as an illustrative example. First, the failure mode and effect analysis (FMEA) of the induction motor is carried out. A fault tree analysis (FTA) is performed as well. Both are aimed to understand the potential failure causes of the induction motor. Afterwards, simulation based on the concept of extended stochastic petri nets (ESPN) is carried out to find the reliability of the motor system from knowing individual failure rates of its components. Finally, sensitivity study of individual components with respect to the system in terms of reliability is carried out. In numerical calculation, reliability prediction methods documented in the US MIL-HDBK-217F N2 and NSWC-11 are adopted for finding failure rates of the components and their mean times to failure (MTTFs). The result shows that life distribution of the studied motor system is best fitted by exponential distribution and the MTTF of the motor is 9.86 year. Based on both the criticality matrix constructed from FMEA and sensitivity analysis, it is identified that components bearing high risk are stator winding, bearings, and rotor winding in sequence.