Tips for identifying defects with rolling element bearings with vibration analysis

Introduction

I recently shared a post about tips for identifying defects in rolling element bearings, and it received a great response. However, I felt limited by the format of LinkedIn posts, especially when I wanted to include images and delve deeper into the topic. That's when I discovered the articles and newsletters feature on LinkedIn. Excited by the opportunity to explore topics in more detail, I decided to give it a try. In this article I'll be expanding on the tips I previously shared and providing more insights into this important aspect of condition monitoring.

This newsletter will serve as a platform for me to document and pass on the tips, tricks, tools, and techniques I've learnt throughout my career.

We live in an era where sharing knowledge has never been easier. I'm hopeful that the information I share will be useful to others in one way or another.

Set a baseline

This is easier said than done sometimes, as you may not know the condition of the bearing when you started collecting data. If you can, start with a baseline that you know is good. If you don’t know the bearing condition to set the baseline, consider if you have identical machines in the plant you can use for comparison.

Correct Measurement Parameters

For detecting rolling element bearings early, we rely on the acceleration spectrum, time waveform, and an envelope spectrum or similar. When a defect starts within a bearing, they normally generate high-frequency signals. Acceleration captures these high-frequency signals more effectively than velocity, making it easier to detect early-stage defects. The envelope spectrum, or demodulated spectrum, goes one step further by extracting the modulation information from the vibration signal making it easier to identifiy the low amplitude defects. I hope to go in more detail about setting up measurment parameters in future articles.

Additional Technologies

Vibration analysis is a versatile tool, but utilising more than one technology will provide further detail on the extent of problems. For early-stage identification of bearing defects, ultrasound can be a great additional technology to help with diagnosing a fault. It can also provide an audio reference as to what is happening at the bearing.

Bearing Information

Having information on a bearing can be fantastic it can provide real clarity and confirmation of a potential problem. If only getting the information was easy.

If you working with older machinery this can be a problem. With new machines, bearing information should ideally be requested at the project phase to ensure all information is entered into a CMMS system.

Some manufacturers display their bearing information on a name plate. However some do not. If you have the bearing information then you are in luck, you can use this to help towards your analysis. Many analysis softwares will allow you to enter the bearing information, which will enable it to provide guidelines as where to expect the defect peaks to appear.

For machines where you do not have the bearing information, experience and knowledge will be needed. We can use information in the vibration data to help understand what type of defect we are looking at. If the peaks have sidebands, this can help us identify what the problem might be. We can also use a process of elimination to help us diagnose a problem when we don't have the bearing information.

Reliable and Repeatable Data

To ensure that we collect good, reliable data and trends that we can rely on, with whatever machines we are monitoring, we must ensure that the operating and process parameters are the same for a repeatable result. This might be the load or running speed, etc.

Share with everyone!

Everyday defects are spotted, saving machines from failure. There's much to gain from others' successes, and I believe in the importance of using platforms to share these stories. Its a great opportunity to help others learn!

Example Bearing Defect from a Fan Bearing

This fan bearing was replaced and the below image shows the extent of the defect. I've also included the vibration data before and after as a comparison.

Collaborate and ask for support

I think at times there can be some trepidation to asking for help or advice. Especially in vibration analysis, where there can be pressures from managers or colleagues, that once you’ve done your training you should just know what you’re doing. If only vibration analysis was so simple.

Had I known all the people I do today and the resources available, it would have been a lot easier and quicker for me to learn. So, if you are unsure, there are lots of people on LinkedIn and Mobius Connect who would be more than happy to support with analysis—including me!

?