Study On Equipment Anomaly Detection Based On Visualization Technology

ABSTRACT:In this paper, we research on the limited local resources problem which exists in the online monitoring system of the work field. Visualization technology is used for the Equipment anomaly detection. We need to solve two problems in order to save local resources and improve monitoring efficiency. Firstly, anomaly real-time data which can be made further diagnosis to solve the limited local resources problem must be detected by using visualization technology. Secondly, real-time monitoring data should be made visualization diagnosis. Thus we can express the statistical characteristics of the system condition as images which can be observed by the operator or upload to control center.The motion status of rolling bearing is proposed in this paper. Based on visualization technology, we research on the the Equipment anomaly detection in three different methods.At first, the time series bitmaps technology was applied in anomaly detection for rolling bearings at the base of waveform test data. From the research, the time series bitmaps which reflect the rolling normal and abnormal conditions are founded greatly differences. Anomaly detection of mechanical systems can be done well by using the technology of Time bitmaps. And also, a new analysis way for the waveform is prompted which is at the base of bitmap techniques of time series.Secondly, from the perspective of Characteristic quantities, the radar chart is used as a visual method for anomaly detection of rolling bearings. And develop a "reference for normal state by radar chart" as a mechanical system whether the status is normal. And realize anomaly detection of rolling bearing. Based on above points, the concept of "state circle" is proposed. It can make the radar chart more identifiable and the detection of the abnormal status of rolling bearing better.Finally, starting from the mechanical system status, the anomaly detection of rolling bearing is done by using the recurrence plot technology which is at the base of phase space reconstruction. From the research, it is founded that recurrence plot is very sensitive to the impact properties:The graphical model and the impact of abnormal waveforms are corresponded, and the abnormal state of rolling bearings can be detected. The results of this paper have much prospect which can be applied in many engineering areas for example, the online monitoring of mobile equipment in the mountain, the wearable monitor which is used for human health, and monitoring for the military mobile vehicles.