SMRP 18 – What is Machine Learning with Philip Garcia
Reliability is a very wide term and has numerous applications in different industries. With the new technologies, concepts, and reliability solutions based on iIOT, Cloud servers, and distributed computing, the reliability programs do everything on any kind of machines. Machine Learning plays a huge role in doing these wonderful things. Machine learning can be used to do enhanced condition-bases monitoring. There are a number of variables that we need to take care of if we want to prevent failures ahead of time and increase the uptime of the assets. This is a very generic application of Machine Learning.
In this episode, we covered:
- What is Machine Learning?
- Where do organizations start with Machine Learning?
- What types of learning models are there?
- Do we need to go buy more sensors?
- And much more!
There are different techniques that are used for data evaluation in Machine Learning such as Tree Model, Neural Networks, and all the tools that can be used to serve the purpose. The selection of tools starts from building a problem case. There are numerous instruments out there that can be used for getting different specific results. The organizations can just begin by starting a specific business case and then build a metrics around it. This model would serve as the baseline for solving a problem and then you can improve and build on it with time. You just need to define a process and then gather data based on it.
Machine learning is not just about sensors and data points. You need to evaluate the current data and data points that you have in place along with the sensors. Once you have done that, you can look for more sensors or improving your data gathering process. After the sensors are in place, you can look at the history of failures. The sensors can only send you an alert when there’s a vulnerable pump or bearing. You are the one who has to make a decision on how and when to fix it.
Once you have followed up on the failure, you need to change the practice that goes around. You need to be able to trend that and make a learning mechanism that will help you in the future. Using technologies like Machine Learning and iIOT is amazing and it has revolutionized the maintenance side of things but the organizations should always take care of the fundamentals first before they bring in such tools. They should have good data collection and analysis strategy. Then they should be able to trend it so that everyone knows the value of that data.
The technology is out there and it really works. The results have shown the organizations the benefits of having a smart tool around to improve their reliability programs. The real challenge lies in making that proven technology for your organizational assets. But the organization need to be able to understand the problem that they have. They need to be able to prioritize their business cases and then bring the solutions to help with that. The Machine Learning just helps you prevent failures before they occur or at the very least increase the meantime between multiple failures. If the organizations can keep it simple, it works wonders for your maintenance and reliability programs.
Rooted In Reliability podcast is a proud member of Reliability.fm network. We encourage you to please rate and review this podcast on iTunes and Stitcher. It ensures the podcast stays relevant and is easy to find by like-minded professionals. It is only with your ratings and reviews that the Rooted In Reliability podcast can continue to grow. Thank you for providing the small but critical support for the Rooted In Reliability podcast!