Editorial

Need for Innovation

Transformers are some of the most expensive devices installed in the electric power system that play a critical role in its operation. Since their invention they have been protected with a limited number of functions - differential and overcurrent, as well as some based on gas - sudden pressure or Buchholz relays.

Even with the introduction of microprocessor based protection devices this did not change a lot. It is true that now we do not need interposing transformers to compensate for the connection of the transformer or the CT ratio and to provide a zero sequence trap. But this is more or less where the innovation stopped. This is probably due to the same old fact - we keep doing things the way we have always done them.

Yes, the traditional transformer protection tools have served us well for more than a century, but we should not forget that the technology has changed so much, especially in the last twenty five years, that we need to get out of our comfortable box and think what we can do better.

What has been happening for all these years is that some company makes a transformer that is purchased by a utility. The protection engineers then select the relays that will be used to protect the transformer based on the protection philosophy and the list of approved devices. And even if the Intelligent Electronic Device (IED) used for protecting the transformer has some advanced functions developed by the manufacturer, most of the time they are not used because "we have always done it that way". For example transformer overload protection in the past was done using an overcurrent relay, set for example at 1.4 times the nominal current. That was the available technology. However, it is well known that the transformer, under emergency conditions, can be overloaded for some time at twice its normal rating without being damaged. That is why now many relays have thermal overload protection. Are you using it?

Even though protection engineers are very concerned with reliability and redundancy, most substations have a single battery. This is justified by saying that if the battery fails, remote backup can be used to clear a fault. However, under some conditions, a low side transformer fault may not be seen by the distance relays at the remote substations that are supposed to clear this fault, resulting in the loss of the transformer. A very inexpensive way to solve this problem is to use on the high side of the transformer a self-powered relay that will detect and clear this fault even if the substation battery fails. Are you using one?

So, I hope it is clear that the first thing we need to do is to find out what devices and functions are already available that can help us improve the protection of these expensive transformers. An example is using thermal overload protection instead of overcurrent for such condition.

But we can do better than that. We should stop thinking about protection as an isolated function. We should instead start working with the transformer designers from the very beginning. We should ask them what can go wrong and how we can detect it, what kind of sensors we need to install inside of the transformer, so that we can observe the small changes that will indicate a potential failure before it happens. And then based on all this data we can develop a combined condition monitoring/protection system. For example, instead of using thermal overload protection that uses a model to predict the overheating of the transformer windings, we should directly measure the temperature at specific points inside the transformer and use this measurement to make a decision if there is a possibility to damage the transformer.

Another example of what we can improve is the inrush detection. Traditionally this has been done using second harmonic detection. Gap detection has been a more advanced method looking for an expected characteristic in the shape of the current waveform. But since every transformer is different, we can use the recording capabilities of microprocessor based relays to learn what the inrush looks like and develop an advanced method for inrush detection using artificial neural networks.

With all the talk about the Smart Grid it is time to start thinking and developing smart machines, such as smart transformers – with built in sensors, condition monitoring and even protection functions. This will require a change in thinking and INNOVATION.

“Innovation distinguishes between a leader and a follower.”

Steve Jobs

Ad: Dependable Communications for Critical Infrastructure
Ad: GE Multilin's 3 series - Fast, accurate & flexible protection
Ad: Protection made simple. HighPROTEC the 4th generation of power protection