Implementation of Condition-based and Predictive-based Maintenance using Vibration Analysis

Abstract

Companies frequently neglect the need of creating a robust maintenance system because it is perceived to be expensive and time-consuming. In contrast, the reality is that maintenance reduces costs and delays, offering organisations a competitive advantage in the long run. Condition-based Maintenance (CBM) and Predictive based Maintenance (PM) are two effective maintenance methods. While the CBM monitors the current condition, the PM will use the CBM results to generate a future prediction for a machine; hence, both are complementary. Three engine pumps were used in a case study at a chilli sauce factory in West Jakarta, Indonesia. For this rotating machine, vibration analysis is the most effective method of measurement. An accelerometer is used to collect vibration data. To determine the current state of the engine, the Root Mean Square of the data was calculated and compared to the ISO 10816 standard. The Fast Fourier Transformation is used in the engine's damage analysis to group each vibration to its frequency. The implementation of CBM and PM at a low cost demonstrates that technology-enabled maintenance is feasible for Small-Medium Enterprises. The internet will be used to collect data in the future, and machine learning will be used to improve prediction.