A Novel Machine Health Monitoring System

The health monitoring of a machine is an assessment wherein the significant change in performance parameters is observed and analyzed with an intention to predict the failure. In this context, many types of analyzers are available but the overall cost is very high and may not be affordable to all. Hence this research work presents a low-cost solution for the development of a novel machine health monitoring system that is reliable, accurate and easy to handle. An open-source platform Arduino Mega 2560 along with the ADXL335 accelerometer is integrated with MATLAB to store & display the acquired data. The program is designed to convert time domain response to frequency domain response using classic Fast Fourier Transform. The system exhibits time domain and frequency domain response to interpret the state of a machine component. The experimental assessment of the developed system with respect to the standard commercially available analyzer is carried out and showed equivalent competency. This system can be used as a monitoring tool in small scale industries & for laboratory work in engineering colleges.