半導體廠DUPS( 柴油不斷電系統 )失效率、可靠度及風險的探討

Abstract

半導體製造廠，因製程的需求，對於供電品質要求甚高，任何電力供應事故包含電壓驟降與停電，會造成生產中斷、晶圓(WAFER) 須重工（REWORK)或損壞、機台當機或故障損壞等重大損失。為了供電穩定與提升電力品質，安裝不斷電系統，在市電即使發生電力壓降或是停電時，能確保生產不中斷及機台不當機，是非常必要的。而隨著生產規模逐漸擴大，風險也會因產值變大而變高。如何提昇不斷電的供電系統的可靠度，是非常重要的事。 本研究的系統以安裝柴油不斷電系統(Diesel Uninterruptible Power Supply，簡稱D-UPS)多機併聯供電方式來因應生產負載逐漸增加，並採N+1台備機方式建置與市電併聯供電。本文首先用統計的手法來測定市電的供電可靠度及單台DUPS運轉的失效機率及可靠度做為基本的探討。經由幾年的市電供電對本系統的事故及維修停機記錄統計，來決定設備的失效機率及分析整體妥善率。多台系統以並聯系統來計算整體DUPS的系統可靠度。透過失誤模式及影響分析(Failure Mode Effect and Analysis FMEA )、分析可靠度因子，經由改變使用的零件及增設工程的方法來增加系統可靠度。最後探討負載需量及裝置容量的可靠度關係來進行風險辨識及分析，在產能逐漸增加中做為維持或降低風險的指標。 在本研究案中由定期保養及緊急維修制度來維持良好的可靠度，並藉由改變碳刷材質，單機失效機率降低0.09%及增設濾網確實能達到增加MTBF(Mean Time Between Failure)的時間，停機保養時間間隔由每三個月改善至半年，降低維修成本，設備妥善率由2007年92.19%，逐漸提升至2012年的98.62%，提高供電系統的可靠度，降低生產損失的風險。 關鍵字:DUPS、可靠度、失效率、風險、妥善率、MTBF

Due to process requirements , high quality is essanial to power supply for semiconductor manufacturing. Any power supply accident, including voltage sags and power failure, may cause production disruptions and damage to machine ,to enhance the quality and stability of the diesel-powered uninterruptible power supply ( referred to as D-UPS),were installed to inhibit of power drop occurred for the Taiwan Power Company, or power failure, and ensure uninterrupted production with the production scale . Gradually expanded, Due to plant load demand, how to build a reliable UPS is very important. In first case, the statistical method was adopted to determine the stand-alone operation as the basic reliability and failure probability of the records through the years, Statistical analysis of the maintenance shutdown of every single service, was alse performed to determine the probability of failure of the operation. More than one system were calculated so as to understand the overall parallel system DUPS the system reliability. Through then analysis of the availability factor, we changed the parts and utilized additional engineering approach to increase the system reliability. Finally, we explore the relationship between the risk estimated, the capacity through the knowledge of amount of load required and the reliability of installed capacity to maintain or gradually increase in risk reduction as a basis. Present case utilized the regular maintenance and emergency maintenance system to maintain good reliability.By changing the carbon brush material, every DUPS unit failure probability reduced to 0.09% and providing additional strainer could indeed achieve the increase of MTBF of time.Downtime maintenance time was estended form three months to half a year,leading to the reduce of maintenance cost.Equipment to proper rate increased from 92.19% in 2007 to 98.62% in 2012.This improved the reliability of power supply system and reduced the risk of loss.