Validation of the accuracy of practical diagnostic tests for power equipment

SUMMARY Diagnostic Techniques are increasingly employed by utilities as part of integrated programmes to manage their infrastructure assets. These are sophisticated techniques being applied to complicated and diverse networks of equipment. Consequently there are many concerns that these techniques a) are not sufficiently accurate and b) could aggravate the system by, at the very least, depriving other areas of vitally short resources. Thus there is a compelling need to develop and deploy simple and robust analytical techniques that can address these problems. These evaluation approaches would then identify the effective programmes worthy of full support, while minimising the resources deployed on ineffective approaches that are ineffective. It is not the intention of this paper to dwell on the well known issues associated with either the diagnostic techniques themselves or their detailed interpretation. Instead this paper focuses on a number of the methods we have developed to assess how well diagnostic information on cable systems relates to the performance of a specific system. Primarily this means comparing the predictions from the diagnostic information with actual service data both before and after the diagnosis. The paper looks at three main approaches: • Direct Comparison - do assets identified as “bad” fail in service and do the “good” not fail? • Performance Ranking - consideration of the whole continuum of performance (not just “good” and “bad”) as measured by diagnostic data and correlation/validation with service experience. • Diagnostic Outcome Mapping – how the failures in service are affected by selection, testing & maintenance actions. The issues and experience are discussed using examples such as underground cables, poles, and transformers. The techniques described are particularly suited to the diverse nature of practical diagnostics and network architectures. In addition, the economics of diagnostic programmes are qualitatively discussed from the perspective of diagnostic accuracy and system quality disbursement.