Condition based (re)investment in transformer populations
Abstract Slowly but surely, condition-based maintenance is coming to be understood worldwide as the number-one choice for optimising the reliability and cost-effective service of transformers....
byGeorg Daemisch
Abstract
Slowly but surely, condition-based maintenance is coming to be understood worldwide as the number-one choice for optimising the reliability and cost-effective service of transformers. Up to now, however, it has remained difficult to understand the actual condition of a transformer based only on traditional methods like water-in-oil or dissolved gas analysis (DGA) data. Also, widely used furan analysis for evaluating the paper degradation usually provides less than clear results. Many users try to work with any of the “standard methods” based on IEC or IEEE, or any other generalised method. Considering the problem of significantly different transformer designs and service data, this attempt will rarely be accurate and is usually doomed to fail. To this day, it remains in the realm of long-experienced specialists to truly understand the complete and complex data, and to also understand the defects and weakness of normally available values, and to finally come up with a reliable result which can be validly used as a basis for further maintenance decisions.
Keywords: transformer, condition based maintenance, life-cycle costs
1. Introduction
Very often in the discussions with users, it is found that the traditional preventive maintenance idea is widely used and is still accepted as a common practice at their organisations. In many cases, the maintenance teams feel under pressure to do any maintenance actions so they can prove that something was done and that the failure could not possibly have occurred due to lack of correct maintenance.
There are vibrant discussions in web forums like on LinkedIn, about quantities which indicate and influence the ageing and show the actual condition. These discussions show clearly the complexity of such procedures. The experience also shows that limit and “typical” values cannot be generalised. Taking into account the differences in design, service, ambient conditions and a number of other factors the same values can tell a completely different story in different cases. Limit values are generally only valid in a certain population.
2. Condition assessment
The biggest challenge remaining is to truly understand the actual condition of a transformer. Nowadays, a number of measurement systems and processes are being offered with expectations to provide the user with all the necessary data required for evaluating the actual condition of their transformer.
1. The first step is to evaluate the data critically in order to discern the reliable data from unreliable data.
2. Based on reliable data, further decisions can be made whether or not additional measurements are interesting for the final diagnosis purpose.
3. Based on diagnosis, an action plan may be worked out based on the complete technical and financial conditions.
3. Sampling and measurements
Our experience tells us that the following is needed:
– Do we have a comprehensive DGA and oil quality history? This is highly important in order to understand the trending up to now. So the actual condition can be understood as a continuous process and evaluated accordingly.
Making well based-decisions requires the following preliminary processes.
A reference oil sampling should always go with a more thorough discussion with the staff to obtain a valid understanding of the individual condition in the power/industrial plant substation and, naturally, the transformers located in the substation! Based on these facts, the first condition assessment can be made, and based on that assessment, further investigations may follow; or in cases where the data is sufficient, a final report may be prepared.
Reference oil sampling should be done with reliable state-of-the-art equipment.
4. Report and results
The first report has to contain at least the following:
1. actual health condition
2. actual ageing condition (remaining substance)
3. load capability assessment
4. risk assessment
5. technical risk
6. financial risk