Predictive Maintenance: IoT, Cloud Computing, and Machine Learning Approach

The title of this story is the title of my master thesis!
So I started a series of stories to share my experiences through this thesis. and I will write my day to day activities from scratch to the final stage of my thesis.

What I’m going to do in simple words

Generally, in this project, I’m going to utilize the abilities of computer engineering to empower heavy industries in the aspect of maintenance of their assets.

There are some industries, right? like oil, gas, iron, steel, food and etc.
In every industry, there are plenty of assets (called machinery) like pipes, motors, pumps, fans and etc.

Each of these assets needs some maintenance, some are normal and periodically like the oil change for a motor or cleaning inside of an oil pipe, in the other hand, some are not periodically or normal, for example when a motor stops working because of changing in voltage or when a pipe breaks because of an accident. So they need maintenance!

Among all the assets of industries, some are much more important and are considered as the hurt of an industry. Pipes are not the hurt obviously so which ones are? Motors, rotors, pumps are! This kind of assets which usually contain a rotating part are very important in the industry and they are called rotating machinery. They need to maintain very carefully since the productivity of the industry is based on these machines and they are complicated.

So, every industry pays a lot of attention to these machines. All the industries have a “maintenance program”. This program can be super simple or complicated, more complicated more accurate. In a simple scenario, operators inside an industry monitor the machines and log some reports by reading the gauges, measuring the temperature or vibration of the machines and thousand of other ways, they then evaluate the reports using software and their minds and experience to check if the machine is working properly or if it’s about to break and if it is, they report it before something bad happens.

So, my project is going to help this maintenance process by utilizing the abilities of computer engineering in this ways:
* Using IoT sensors instead of the operators to collect data from machines
* Using the cloud to receive, collect, process and display the data that comes from sensors
* Using machine learning to let the computer detect and classify the faults of machines instead of operators.

However, each of the technologies that I’m going to use to empower the maintenance in the industries can be super helpful, combining them would move the maintenance to the next generation.

Now that you know what I’m going to do, I’ll dive into details in the next stories.