Introduction:

Sandia National Laboratories is currently developing a predictive maintenance system (WinR-PdM) that makes use of equipment maintenance records to focus attention on key failures that lead to equipment unavailability and downtime.  Sensors are then employed to monitor important indicators of equipment performance.  Real-time sensor data is combined with information from ongoing equipment maintenance to constantly update equipment failure predictions and to identify the most likely causes of failure.  Commercialization of WinR-PdM is planned when development is complete.

Benefits:

• WinR_PdM can
• reduce equipment down time
• providing a "heads up" on pending equipment failures
• quantitative basis for improved maintenance strategies
• cost effective spares inventories,

Capabilities:

When completed, WinR_PdM will provide early notification of many pending equipment failures thereby reducing equipment downtime and increasing availability and productivity.  Other WinR_PdM modules in development will provide the following capabilities:

• Maintenance records analysis to better understand equipment performance and costs as well as identifying major reliability problem areas,
• Spares inventory optimization to ensure the most effective spares inventory in terms of lowest downtime for the lowest cost, and
• Equipment simulation for optimizing predictive and preventive maintenance strategies.
• WinR_PdM can be linked to a Computerized Maintenance Management System (CMMS) to streamline the collection of  failure, repair, and inspection information.

Applications:

As part of the development process, WinR_PdM is being installed on a flexible manufacturing system (FMS) at Allied Signal Federal Manufacturing and Technology center in Kansas City.  The FMS consists of 6 milling machines, 2 wire-guided robots, 2 coordinate measuring machines, a wash station, and a central coolant distribution system.  When complete, WinR_PdM will monitor all 6 milling machines, the coolant system, and the wash station. 

Figure 1. Predictive Maintenance System Architecture.