Condition - based maintenance employing system errors

This work supports the introduction of condition-based maintenance in a capital goods company in the poultry processing industry. In the absence of condition monitoring information of the system, capital goods companies who offer maintenance services to their customers need to search for other sources of information to enhance proactive maintenance activities. With this purpose, we explored prognosis-based maintenance decision-making using event data to represent the condition of the system. We proposed a framework to support maintenance decisions through early failure detection offering immediate action based on a long-term maintenance policy that minimizes maintenance costs. The framework uses as inputs system errors and then, identifies failure patterns, followed by defining failure predictions and finally deciding on actions to determine a Condition-based maintenance policy. The proposed framework was tested in a real capital good company in poultry processing industry and proved that maintenance costs can be significantly reduced by introducing condition-based maintenance as part of its service strategy.