Experimental investigation of vibration signal of an industrial tubular ball mill: Monitoring and diagnosing

Abstract To achieve improvements in the production capacity and energy efficiency of an industrial tubular ball mill, a novel method, monitoring an unmeasured parameter (level of coal powder) and diagnosis of the operating modes of the mill, was proposed. Two accelerometers were installed on bearing housing to pick up the vibration transferred from the mill shaft. A system was designed to record the vibration signals and transfer them into energy amplitudes, by using a wavelet packet approach, based on the KINGVIEW program, as well as Visual C++ environment. Based on these, a series of experiments was conducted in a 250 MW power plant to investigate the vibration characteristics corresponding to the effects of different levels without air particle removal, milling times, and varying levels in practical working conditions. The experimental results show that the operating modes of the mill, such as mill over-load, stable case, etc., can be diagnosed by proper interpretation of these vibration characteristics; and also the unmeasured parameters, i.e., level of coal powder filling the mill, can be monitored on line. Finally, methods for diagnosing operating modes and a model for on line monitoring of varying level by using a non-linear partial least square (NPLS) algorithm were presented. Combining the method and the model, we believe that the performance of ball mill can be improved.