Detection of Adverse Operating Conditions in Dense Medium Cyclones by Acoustic Emission Monitoring

A proof-of-concept non-intrusive contact Acoustic Emission (AE) monitoring and soft-sensor analysis system has been developed and tested for passive monitoring of adverse operating conditions in coal washery Dense Medium Cyclones (DMCs). AE sensors externally mounted at feed, mid-body, underflow and overflow locations on an operating DMC are used to monitor local process conditions. Robust soft-sensor analysis methods based on low frequency power envelope analysis have been developed to detect adverse operating conditions, specifically incipient surging, discharge overload, clay ball build-up and feed rate balance for parallel DMCs. Specific adverse operating conditions in a DMC were typically induced, AE data logged, and post-analysis in association with plant variables used to formulate alarms for the imminent on-set of the condition. 

The current post-analysis system has been successfully developed and tested for detection of adverse operating conditions on several industrial DMCs in a series of short-term plant trials at West Cliff and Bengalla washeries in NSW. The most significant outcome of the project is the demonstrated ability of a low-cost, non-intrusive multiple sensor contact AE monitoring and analysis system to robustly detect a variety of incipient and actual adverse operating conditions.

Specific findings of the project of significance to plant operators include:

• Discharge overload can be detected from analysis of the overflow and underflow AE power envelopes. Incipient overload can be detected several minutes prior to detection on discharge screens.
• Surging can be detected from analysis of discharge AE power envelopes. Incipient surging can be detected from the overflow power envelope over ten minutes prior to detection on discharge screens.
• Clay ball build-up and flushing can be detected from analysis of the underflow AE power envelope. Partial blockage of a discharge is characterised by low levels of local AE power.
• Feed balance to parallel DMCs can be assessed from the ratio of AE envelope power detected by two identically mounted sensors. 

The next step is to fast track the development, manufacture and test in a four month plant trial a commercial prototype AE analyser with alarm capability for detecting the imminent onset of adverse operating conditions including incipient surging, spigot and vortex finder overload, clay ball build-up and discharge blockage. Such an analyser has the potential to provide advisory alarms to significantly improve plant recovery, capacity and run-time of existing equipment. CSIRO considers that this work can be completed within 12 months of commencement. At completion of the work analysers based on the prototype design will be available for purchase by the Australian coal industry. CSIRO Minerals has already established a strategy to supply AE analysers to the coal industry. An analyser based on AE signals will be relatively low-cost and easy to install in an operating washery.