Condition monitoring of a process filter applying wireless vibration analysis

This paper presents a novel wireless vibration-based method for monitoring the degree of feed filter clogging. In process industry, these filters are applied to prevent impurities entering the process. During operation, the filters gradually become clogged, decreasing the feed flow and, in the worst case, preventing it. The cleaning of the filter should therefore be carried out predictively in order to avoid equipment damage and unnecessary process downtime. The degree of clogging is estimated by first calculating the time domain indices from low frequency accelerometer samples and then taking the median of the processed values. Nine different statistical quantities are compared based on the estimation accuracy and criteria for operating in resource-constrained environments with particular focus on energy efficiency. The initial results show that the method is able to detect the degree of clogging, and the approach may be applicable to filter clogging monitoring.