A few weeks ago now, our MD (Steve Sommerville) gave a short presentation in Distributed Energy Resource (DER) Forum in relation to energisation of transformers on large generation sites with multiple transformers. The current P28 standard does not have a provision of how to consider these sites, and requires that all transformers are considered using these worst case assumptions. This often results in developers unnecessarily spending a large amount of money on Point On Wave switching controllers that are not needed, and unnecessarily delaying projects.
The theory if transformer energisation is fairly well understood and the subject of many papers. In essence, it is a probabilistic approach which depends on the residual flux in the transformer, the closing angle of the breaker and the network fault level.
The ENA P28 standard provides a way of assessing transformer energisation, based on a number of conservative assumptions, such as high residual flux, minimum fault level, zero crossing and a range of network voltages. It is conservative process, but overall it is a fairly robust method that most of the industry are reasonably happy with.
However, on larger sites where the are multiple distribution transformers for the Solar / BESS / Wind, problems can be encountered. This occurs where an individual transformer end up as a Category 2 energisation (Infrequent). Whilst this may not sound difficult, consider the case where the site contain a number of transformers.
ENA P28 allows allows 4 Category 2 events per calendar month. Each event consisting of 4 RVCs, separated by 10 minutes. So if a site contains 16 transformers it can take 4 days to energise, if the P28 rules are followed – which is unrealistic from a practical level. If a site contains more than this, then real problems are encountered.
Why is the P28 standard wrong? As noted earlier, transformer energisation is inherently probabilistic. Taking a combination of worse case factors for a single transformer is reasonable, but if there is a large group of transformers, then the worst case assumptions become increasingly unrealistic – the results is then a study and analysis that bears little resemblance to reality, and ends up costing developers £££. Aurora are in the process of encouraging the ENA and DNOs to reconsider these requirements, and add in some clarity to allow a ore holistic & probabilistic approach for large sites.