We recently PM’d a 37 yr old oil filled transformer(PF, hot collar, Ir, TTR, Winding Resistance, Excitation current, etc) and it tested good… but was loud as can be…! My Lead tech called the manufacturer and they stated it was good to return to service.
I’ve been troubled by how loud it was( the reason we were called in) and have come across the concept of Magentostriction?!
I am asking how we measure the core and when does it become a problem worth noting? I haven’t found much useful info online.
Unless the noise is directly related to the connected load, I’m not sure there is much you can do about the sound, as this is usually the result of a physical change in the transformer as it ages.
A transformer is magnetically excited by an alternating voltage and current so that it becomes extended and contracted twice during a full cycle of magnetization. This change in dimension is independent of the direction of magnetic flux, occurring at twice the line frequency.
Over time, the transformer core adhesive starts to break apart and the laminated layers separate from each other slightly. The vibration of these layers is the humming noise you can hear and once the adhesive starts to break, the sound gets louder.
Sweep Frequency Response Analysis (SFRA) is a method to evaluate the mechanical integrity of core, windings and clamping structures within power transformers by measuring their electrical transfer functions over a wide frequency range.
If the noise is related to the core, it would probably need to be re-laminated. The SFRA test would provide a good visual indication. It is best performed during acceptance as a baseline and then used as a comparison for future tests.
Excessive magnetostriction can lead to mechanical stress and potential damage to the transformer. The load should also be investigated as it could have an effect over time.
You may find this article about Transformer Noise helpful:
@cptviggi - Something to follow up with regarding @SecondGen 's response. SFRA is a great test. But if you don’t have a prior test to compare to, it may be hard to identify if anything has shifted over time. All transformers have a similar curve shape as a “Big Picture” overview. But as with any test, having “sister” transformers or prior test results are really what is needed here. Maybe you will see something in the curve shape that shows an issue. You could compare A&C phases since they will be similar in curve shapes. B phase is going to be a little different.
Maybe one day we will see SFRA as a mandatory test in large power distribution transformers. I know I told my boss 12 years ago that this test has some major benefits. He had a response that was very logical - “I need to be able to sell the testing. If I can’t sell it, it doesn’t justify spending $20,000 on the test equipment”.
Short story - I was working at a WMATA site and a brand new transformer had faulted. The transformer was shipped back to the manufacturer to be investigated. I went with WMATA to the factory and sat across from all of these corporate CEO’s and the President of the company - had to be 20 of us in the room. The manufacturer traced an issue during the transport of the transformer. The identified a specific pot hole on the highway that the tractor trailer hit that caused 3G’s of force that may have been the culprit. Everyone looked at me and asked, what can we do to prevent this in the future? My first response was that they should look at the sensors on the truck for any impacts. 2nd thing was to perform SFRA at the factory and again on site. I believe they are now performing SFRA for every transformer that goes to WMATA as well as having a 3rd party perform SFRA once it’s installed.
So I think eventually, SFRA may be a mandatory NETA test.