Merlin Gerin LV Breaker

Yes, it could be your breaker. Those can be serviced and tested by major manufacturers of that kind of breaker. Eaton, formerly Cuttler Hammer will probably work with you as they can test anyone's breaker. Unfortunately, you must send it in to them to be tested.

Now, before you go to all of that trouble, consider the effects of inductive and capacitive coupling. One very common mistake occurs when the electrician is too lazy to cut the extra cable that he pulled. Frequently, it gets wadded up into a messy coil or inductor. Or, if you have very long cable lengths (more than 50 Meters) then you are creating a long capacitor. During startup, there may be some high frequency components as contacts make and break (sometimes called noise) and the combined effect of "sparking" and "bad wiring practices" will give you exactly what you are experiencing.

Eventually, you will be able to find places where the transient at startup melts the insulation. Don't wait for that to show up. Go find the problem now.

Thank you for your comment. I am curious about the capacitance effect on the tripping of my breaker. For one thing, I have tested my cables and found it okay (Insulation resistance in Gigohms), no addition of load since commissioning, I am figuring out where this capacitance could have multiplied, how to measure them and eventually to reduce them.

As you may remember, capacitive coupling occurs when you have a large conductive surface area next to another large surface area, very close to each other but separated by an insulator. This could happen with long cables between the breaker and the motor.

You may not be able to measure it directly. What you have to consider is that the current through the insulation (capacitor barrier) is a function of frequency. That is to say that the higher the frequency, the more current you can pass. And the frequency that I am talking about is not your primary frequency of 50 ro 60 Hz

Maintenance on big breakers includes dissassembly and cleaning or replacement of the contact surfaces. Old breakers accumulate dirt and oxides both of which may cause arcing of the contacts. An arc contains many high frequency components and this is when the capacitive coupling becomes important. The arcing inside the breaker will induce high frequency current and voltage changes which could be made worse due to parallel paths provided by capacitive coupling.

Even the bounce of contacts in a contactor can cause high frequency burst and under the right conditions some unexpected current flows will bypass filters and cause problems. You can see the effects on an oscilloscope but it will do no good to try to quantify or measure this because each arc or bounce of the contacts is unique.

Another example of this kind of problem is a fluorescent lamp. When it is off, you can measure several megaohm or gigahoms of resistence from one end to the other depending on your equipment. Clearly the resistance is caused by the fact that the gas inside acts as an insulator until a plasma is formed. To establish the plasma, you must start the lamp. But now you can not use your ohm meter to measure the resistance inside the lamp. You can measure the voltage across it and the current through it and then calculate the resistance. With the right ballast, this change of several megaohms to a couple hundred ohms occurs very quickly (sometimes in microseconds). And without the ballast, you can explode the lamp reaching currents about 100 times the normal current in a few milliseconds. I've actually done this. No need to measure anything. All I had to do was set up the right conditions.

Some questions to better understand your problem :

1- Do you have other loads connected to the same breaker ? (the 3 pumps look like a very small load for a 2500 A breaker !). If so, what is the current of these loads and are they running when the tripping occurs ?

2- How long has this breaker been in service ? Is this random tripping something new ? In this case, did something change in your installation ?

3- Do you have a 24 V aux supply connected to the trip unit ?

Thank you for your post. Yes there are loads connected, they are not running when the trip occurs. before these loads run(chiller loads), the pumps has to run first, this is when the trip occurs, the starting of any of these three pumps. This breaker had been in service for 18 years and then this tripping is occuring 2 years now. There were no additional equipment added ever since. Yes the supply for the tripping unit of the breaker is 24 V. dc. coming from a converter that's not a part of the breaker.

This is a random tripping, because there are times that it will not trip and all the loads are runned without problem. But when these loads has to be shut off during cold season, the pumps has to be shut off also. But when its time to run again these chiller loads, the 3 pumps has to run first. Then during starting of any of these pumps, we experience tripping. I am suspecting the LV main breaker as the problem, because this is the one tripping, and not even the pump breaker trips.

If I understood well the first unexpected trip occurred after 16 years of good running. Unfortunately at this point, even if we pay a lot of attention to the electronic components selection, it is possible that some components have drifted out of their original specification and are the cause of the problem.
After this period of time (and moreover in a tough environment, but this might not be the case here) Schneider recommends to upgrade the installation. A good way to give a new life to your installation would be to use the Masterpact "Ecofit" offer which allows easy and quick retrofitting.
The local Schneider Electric offices will be very pleased to support you on this matter.

As a short-term solution, you could try to increase the ground fault pick-up value (if acceptable) and/or the ground fault time delay. The tripping may be caused by the sub-transient inrush current, that lasts for less than 20 ms.

Thanks Pierre for comment, I might agree that replacing the breaker of the same type may not be feasible, instead, to be upgraded with some new type of breaker. I would like to consult our local scneider then.