Ranking of sensitive positions based on statistical parameters and cross correlation analysis

Abstract

Condition Based Monitoring of a machine refers to analysis of the health status of the machine and its components. For this purpose, acoustic data is acquired from various positions on the machine. Acquiring data from large number of sensor positions is not economically viable. It is preferable to have an effective monitoring system that is faster in data acquisition without compromising on the robustness of the system. In fact taking data from too many positions would directly affect its reliability due to various kinds of noises. Therefore there is a need to locate sensitive positions on a machine. These sensitive positions are expected to exhibit appropriate fault characteristics in a much better way as compared to other sensor positions. This paper presents a novel method for ranking sensitive positions based on statistical parameters. While selecting the required number of sensitive positions, cross-correlation among the positions is taken into consideration to avoid redundancy. Furthermore, a standalone application for implementing the same has been developed on Android platform. The scheme and application can be used for many other applications as well, where data is acquired from multiple sensors.