FIELD RELIABILITY VERSUS PREDICTED RELIABILITY - AN ANALYSIS OF ROOT CAUSES FOR THE DIFFERENCE

The significant difference between predicted reliability and actual reliability of fielded systems-what we have termed the "reliability delta"--has been a long-standing problem in reliability engineering. An extensive literature review was conducted to identify root causes contributing to this difference. In addition, SAS analysis of data collected in a 1988 study yielded interesting results. This research asked 38 reliability experts from the Department of Defense and Industry to identify and rank order factors which potentially contributed to the reliability delta. Total Quality Management (TQM) tools were used to analyze the relationships among and impact of key variances on the reliability prediction process and explain apparent differences between the DoD and industry perspectives. The literature review identified eleven factors as possible causes of the reliability delta; however only six were noted as being significant by the experts. The six factors identified were: problems with data collection, assumptions underlying and use of prediction technqiues, lack of understanding of the operational environment, problems with manufacturing processes, short-term management focus, and design-related problems. Each of these factors was evaluated from both DoD and industry perspectives. Although the same six factors were identified by both groups, the rank order of the importance of these factors was virtually reversed between DoD and industry personnel. The contrasting priorities may in part explain the focus of management efforts of one group on factors which are considered less important by the other. The final section highlights some of the logistics impacts of the reliability delta.