A machine learning approach to preventive maintenance in industrial machines.

Abstract

A study of how machine learning can be used to study patterns of machine failure and predict future possible failures. The project focuses on temperature and vibrations as the key factors of machine failure. A study by US Air Force Avionics Integrity Program shows that temperature contributes to 50% of machine failure and vibration 20%. In the project we implemented a prototype to measure the different variables used in the model. We used an LH35 sensor to measure the temperature of the machine and an mpu6050 accelerometer to measure the vibrations. We further studied the data collected using a machine learning model and trained it to make predictions using KNN algorithm. The model went through model tuning so as to ensure a high accuracy of the model’s future predictions. This report consists of four chapters as described below;

Chapter One: Introduction; this chapter includes the background of the research project, problem statement, general and specific objectives, justification and scope.

Chapter Two: Literature Review; this chapter contains a study and analysis of previous work done in line with the topic of the research project. It further includes descriptions of the different tools and equipment that were used in the design and implementation of the research project.

Chapter Three: Methodology; this chapter includes all the steps taken to implement the specific objectives of the research project.

Chapter Four: Challenges, Recommendations and Conclusion; this chapter includes the challenges encountered in the course of this project. It also has the recommendations that should be fulfilled for the project to attain a large impact together with the conclusion.