The Collection and Analysis of Single Sensor Surface Vibration Data to Estimate Operating Conditions in Pilot-Scale and Production-Scale AG/SAG Mills

"Improved measurement tools for monitoring and control of AG/SAG mill performance have a high level of interest in industry. Historically, acoustic emission monitoring has been widely used on SAG/AG/ball mills and hydrocyclones to provide new information on the conditions inside these devices affecting their performance. The current work was focused on a new technique using surface vibration monitoring of a number of pilot-scale and production-scale AG/SAG mills to yield insights into the events occurring in the mills. The development of relationships between signal analysis outputs and mill operating conditions can be used for soft-sensor measurement of mill process variables. The technique used an array of accelerometers mounted on the mill lifter bolts linked to a fast data acquisition system which monitored the surface waves generated during operation. Details of the system are presented. Mill operating variables were adjusted systematically to provide a wide range of operating conditions for surface vibration signal collection. The experimental approach is discussed.The signals acquired during the experiments were analysed using a range of techniques. The results show that strong relationships exist between measured vibration power (both total power and within certain frequency bands) and mill load level/power and mill speed. A new method for extraction of power spectral density and band power information on features of interest in a mill rotation cycle is also presented.These results show that single sensor analysis information from the surface vibration system can be used as soft sensor inputs for monitoring process variables such as mill load that until now, have not been reliably measured. System outputs also show considerable promise in the area of condition monitoring. An example of the signal analysis output is presented showing elevated vibration levels in an area with higher resultant wear. Eventually, it is considered that outputs from this system for process variable and wear condition monitoring will be key new inputs to an online AG/SAG mill monitoring and control system."