Machine Learning Method Based Industrial Risk Analysis and Prediction

Abstract

IoT-based technologies are gaining popularity around the world. People all across the world gained skill in industrial automation after the industrial revolution. Machines, on the other hand, have a higher level of precision than human-made products. As a result, machines and robots are increasingly taking human jobs. We've seen machines that run solely on human written rigid logic in the past. As a result, mistakes were inevitable. There used to be a variety of machine failures all the time. We discovered a major solution to get out of this problem that works not only like efficient robots but also thinks like people. We can comprehend the primary up-gradation strategies by studying human evolution. Humans have a strong desire to learn new things. In our system, machine will also learn based on the situation that have been made by any occurrence. Our raspberry pi-based system helps to make proper analysis over the machines. We take voltage, current, gas value, 24/7 camera observation and temperate, humidity values as input parameter. Machine learning module matches/compares these real time sensors data with training data (which is used to train the system). As a result, The machine learning module provides some statistics graphs or plotting of sensor data. Machine performance can analyze by observing these graphs. Using input values we can not only easily make an observation of current performance but also make proper prediction what’s going to happen with that machine over the years. Also, determine the efficiency and predict the possibility of upcoming threats or risks