The power industry is always changing, whether it be the technology of the equipment through to the context of business decisions and asset management. The growth in renewables, particularly remote or offshore sites, has led to a focus on reliability as the costs of intervention are usually punitive; sending out a team to inspect and maintain a windfarm transformer is costly. Consequently, operators look to reduce such activity, whether related to maintenance/inspection of the primary assets, or to manage the ancillary equipment, such as protection devices and condition monitoring.
Condition monitoring is a powerful tool which can provide great value in supporting decisions related to transformer health: can we detect and diagnose any failure modes presently in operation, and if so, how long do we have? Traditional approaches to transformer monitoring have concentrated on dissolved gas analyses (DGA) for the transformer oil and bushings, as the latter are a significant cause of failures across the industry; additional monitoring through partial discharge (PD) is also of great diagnostic value, but interpretation of such data is far more complex.
We will share practical cases from a range of applications and look at lessons learned: including the need for data with appropriate accuracy/precision for the decision to be made, but also the need for reliable monitors on unstaffed sites.
For online dissolved gas analysis, examples include:
- ‘Detector’ DGA devices indicating a change in status, but only at a ‘low’ level; a contextual analysis led to laboratory samples being taken in a risk managed manner, subsequent electrical testing and internal inspection, which, ultimately, saved the transformer from catastrophic failure
- Multigas DGA devices which may be applied to a variety of oil types but which require a lab sample for confirmation of findings and diagnoses, including a case in the USA where the transformer failed while confirmatory samples were being taken
- Multigas DGA devices used to make diagnostic decisions in ‘real time’ without requiring subsequent samples
For online bushing monitoring, examples include:
- The pros and cons of sum current analyses for detection/diagnostics of bushing issues
- Demonstration of the need for tracking data from raw measurements through to the derived values, to ensure good diagnostics
- Relative and true power factor cases where complementary approaches aid in diagnostics and tracking of deterioration over time
- Unexpected data where a bushing deteriorates but then seems to improve, while at another location four bushings seem to go bad simultaneously
- The role of partial discharge (PD) as a detector and diagnostic for bushings
In addition some PD cases which demonstrate the need for expert support while interpreting data and supporting decisions.
Condition monitoring can be likened to a chain saw – when well used you get the job done quickly and efficiently; when misused, you may end up hurting yourself and the environment. The cases presented here are real cases, with real consequences, which ‘played out’ in real time. The successful cases are those where an agreed action plan was in place at monitor installation, and where the plan was subsequently followed.