Predictive Maintainance and Machine Monitoring

Predictive maintenance which is an old age problem has been gaining the attention of late due to the popularity of IoT and applications and deep learning. Over the past few years, deep learning solutions have produced a state of the art results in various domains. Our main agenda is to provide predi

Project Title

Predictive Maintainance and Machine Monitoring

Project Area of Specialization

Artificial Intelligence

Project Summary

Predictive maintenance which is an old age problem has been gaining the attention of late due to the popularity of IoT and applications and deep learning. Over the past few years, deep learning solutions have produced a state of the art results in various domains. Our main agenda is to provide predictive maintenance of machines in the industry in a way that both maintenance and repair activities can be predicted. It leverages deep learning solutions by exploring different systems.
When a manufacturing process stops for an unplanned event (e.g., a motor failure) it accumulates downtime. While downtime is most often associated with equipment failures(breakdowns), it actually encompasses any unplanned event that causes your manufacturing process to stop. For example, downtime can be triggered by material issues, a shortage of operators, or unscheduled maintenance. The unifying element is that although production is scheduled, the process is not running due to an unplanned stop.
This project focuses on machine parameters that optimize and implement fault analysis by enabling a predictive analysis of data set to receive from a particular system.
Our solution consists of two different parts of software and hardware. For hardware, we are going to use Raspberry Pi and Power Analyzer. On the software, side Datasets would be collected, analyzed by the use of Python as a programming language. After applying certain algorithms our predicted result will be shown on a user-friendly web application.

“If you automate a process that has errors, all you’ve done is automate the generation of those errors.” ? W.L.W. Borowiecki

Project Objectives

To Collect Data: We will collect data from different systems which will have different parameters in different conditions and make a model according to the data collected by us from different systems To Detect Anomaly in System: By training our model to a machine we will install our hardware to the machine which will analyze different parameters from the system and predict the anomaly which will be appearing due to the change in the parameter’s. To Classify the Anomaly: Our main goal is to find the anomaly. After finding out the anomaly we will we working on the occurrence of the anomaly that is on which factors anomaly appears and damage the physical part of the system. To Display the result on Web App: After the development of the final model we will be working on the web based application which will remotely show you the health of the machine and anomaly if it is going to happen

Project Implementation Method

The advances in manufacturing technology and competition in the market necessitate the continuous availability of machinery for production. This has created a need for effective and efficient maintenance practices resulting in improved plant performance. Often, manufacturing equipment is utilized without a planned maintenance approach. Such a strategy frequently results in unplanned downtime, owing to unexpected failures. The emergence of Industry 4.0 and smart systems is leading to predictive maintenance (PdM) strategies that can decrease the cost of downtime and increase the availability (utilization rate) of manufacturing equipment. The thought process of our project is to start by selecting different machines to find out which parameters affect their performance over time and collecting different sort of data/parameters from various choices of sensors would give us the idea of where an anomaly is produced while the machine is operating. Our hardware would be installed on different machines and would be connected through the internet which we will be gathering data time by time and simultaneously the data would be studied and analyzed. Useful data would be extracted and a basic model would be designed and tested. We will mount our hardware on the machine by practicing our model on a machine to evaluate various parameters from the system and anticipate the anomaly that will occur due to the change in the parameters. Finding the anomaly is our key priority. After discovering the anomaly, we can work on the event of the anomaly on which anomaly of variables occurs which damages the physical part of the system. After all the data is gathered and the basic model is tested. The final model will be made, implemented, and reviewed to make it more reliable on hardware.  (web-app part remaining now only)

Benefits of the Project

It is an era of IoT 4.0 where industries are also shifting towards digitalization, there is a must need to automate the whole process of fault detection and machine monitoring.
In industries, downtime is the period when a machine is not in production. According to Analysts, unplanned downtime can cost a company as much as $260,000 an hour. We will predict the fault in the machine with the help of previous machine data and after applying some machine learning algorithms we will predict the faults in the machine. This eliminates unnecessary time spent on looking for the cause of the problem.

These are the beneficiaries which will help the industries in their maintenance task:
Reduce maintenance costs
Avoid unexpected downtime
Forecast failures
Real-time data insights and analytics.
Remote access to machine stats
Increase the productivity of the plant
User-friendly web app application
Low-cost implementation of the project
Quick to Deploy

Most Importantly by collecting sensor data of important parameters from a machine we will have access to the dataset which is the new gold nowadays.

Technical Details of Final Deliverable

Our final product consists of:
Raspberrypi with sensors
Power Analyzer
Web Application
Valuable data of machine

The main goal of the project:
To collect Data
To Detect Fault in system
To Classify the fault
To Disply Result on Webapp

The raspberry pi and power analyzer are used to collect the data(will send to the server)  which will be preprocessed and after applying machine learning algorithms the output will appear on the HMI with real-time analytics and stats of the machine at that moment. The application will be available remotely.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Manufacturing

Other Industries

Others

Core Technology

Artificial Intelligence(AI)

Other Technologies

Internet of Things (IoT)

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com