Case Example 3: Measurement of machine performance degradation using a neural network model

Abstract

Machines degrade as a result of aging and wear, which decreases performance reliability and increases the potential for faults and failures. Equipment reliability and maintenance drastically affect the three key elements of competitiveness — quality, cost, and product lead time. Well-maintained machines hold tolerances better, help reduce scrap and rework, and raise consistency and quality of the part. They increase uptime and yields of good parts, thereby cutting total production costs, and also can shorten lead times by reducing down-time and the need for retooling [1]. The recent rush to embrace computer-integrated manufacturing (CIM) has further increased the use of relatively unknown and untested technology. Today, many factories are still performing maintenance on equipment in a reactive, or breakdown, mode. This is due to traditional process monitoring systems that can only detect machine or process faults when they occur. Reactive maintenance is expensive because of extensive unplanned down-time and damage to machinery [2–7]. In high-performance systems one often cannot tolerate significant degradation in performance during normal system operation. This has led many US manufacturers to look to suppliers for smarter equipment that will ease the need for strong technical support for equipment.