Enabling condition based maintenance in a precious metal processing plant

In precious metal processing plants, suitable maintenance systems should be deployed to provide the necessary plant operating conditions, ensure plant availability and provide critical fume extraction and filtration systems for health and safety. A key focus for this work was use of sensors to monitor and predict abnormal equipment behaviour and examining how the data obtained could be processed into information and better inform the condition monitoring systems. The requirement for this research was driven by escalating maintenance costs in an industrial case study in South Africa. Historical records on equipment failure and was used in Pareto analysis to define the critical assets and failure modes for equipment that dominated the escalating maintenance costs for the industrial case study. Vibration data from mounted sensors was then collected the identified critical assets, analysed and used to infer condition of critical assets. Statistical tools based on process capability index values for processes in control and out of control were defined for tracking system deterioration and enabling predictive maintenance. The results show that sensors and thresholds based on process capability index can be used in predictive maintenance to alert maintenance teams to attend to vacuum pumps and fans pre-failure and hence improve plant availability and operation and reduces cost.