Abstract
Nowadays, there is a huge range of different monitoring facilities for power transformers available on the market.
That is why we are increasingly faced with the dilemma of how best to improve the reliability of transformer service and how best to use the data for that purpose.
A typical marketing ruse used by manufactures of monitoring systems is to lead the users to believe that now all their problems are solved. But is it really a solution if your doctor tells you that you are ill and that there is no therapy for your illness? Also, in many cases of transformer failure the finger of blame is pointed at the staff, who typically get accused of “lack of maintenance”.
While this is a very typical scenario, nobody ever told these poor people what they should have done. In most cases, somebody will tell them to go ahead and install more monitoring devices. But will more monitoring devices safely rule out future failures? What do we get from monitoring? At first glance: “Data”! BUT what to do with the data? Which data ought to trigger a reaction, and what kind of reaction? Is it safe to put our trust in the limit values? Which data ought to trigger which action, and at which priority? What should we do if any of the values is reached or even exceeded?Have any contingency plans been established, what should be done or what procedure should be observed if a certain condition is reached?
To really get the most out of the data, it helps to rely on long-term experience and to think outside the box: it is the unconventional solutions which allow us to classify the data, to work with them, and to use reliable reference data as a diagnostic tool to design an adequate therapy.
Monitoring :
Formulating and defining the goals
What good are measurements and monitoring activities if the operator does not know what to do with the measurement results? In that case, the whole monitoring effort will be a waste of money, time, and material.
Since every time a transformer fails, the finger of blame is first pointed at the operator, who gets accused of either not enough or faulty maintenance. Then what happens is that monitoring measures are suggested as a solution to the problem, which are typically sold to the client as “maintenance”.
As a general rule, however, monitoring is not the same thing as maintenance, and it certainly does not solve any problems!
Here is where the system first starts to look illogical. As a general rule, monitoring is (or should be) the basis of a condition-based maintenance system. What this means, of course, that the goal of condition-based maintenance should be kept in mind even when first selecting a monitoring system. When selecting a system, it is therefore a good idea to set a specific goal as to what should be monitored, and for what purposes.
It is also crucial, when selecting a monitoring system, to plan for the appropriate measures to be taken if the monitoring system outputs certain critical data. Indiscriminately collecting and storing data really won’t do much good as even critical developments are not necessarily recognized.
Georg Daemisch
