Power transformers are high value capital items which require close attention. Their management, through offline testing and online condition monitoring, must be merged into an overall asset management approach that provides value to the organisation in both the short and long term.
Many utilities are faced with a fleet of power transformers, many loaded to near nameplate levels, where the present condition of the units is unclear and failure rates exceed industry norms; many units have already passed their original expected design life.
Power transformers are typically large capital items with a long lead time for delivery and may have a significant impact on system reliability if they are unavailable for service in an unplanned manner.
Owner/operators balance the predicted power transformer population performance with the need for investment. Consequently, it is critical that there is a strong link between subject matter experts in transformer condition and those who plan for capital investment in order to efficiently invest.
Linking condition to investment requires an analysis of risk. But what exactly is risk?
According to international standard developments on asset management, risk is the combination of two things – the likelihood of an event occurring and the consequence of the occurrence.
Power transformers may fail for a variety of reasons, and vital information about their viability is generated through monitoring, text and inspection. Particular families of transformers may be suspected of having particular weaknesses of design or construction which lead to failure.
As transformers age, there is an expectation that their reliability will fail – accumulation of mechanical damage from successive through faults and monotonic degradation of paper insulation, providing the mechanisms for ultimate failure. However, the expected ‘end-of-life’ element of the bathtub curve has been predicted but has not yet actually arrived. The management of in-service transformers requires an understanding of these issues and prudence in the analysis of data and subsequent actions.
There are many older transformers in service, but the age distribution does not demonstrate those which are most likely to fail, or those which have the highest impact if they do fail. Figure 1 shows the age profile for a population of transformers owned by a single utility. Age is a useful indicator of where failures may be most likely to occur, but we need to look at each unit individually.
|Figure 1: Age Profile of distribution transformers belonging to a single utility|
The underlying questions related to power transformers were really crystallised at Doble’s Life of a Transformer seminar in February of this year. The two questions which apply to both populations and to individual transformers are: What is the likely mode of failure? And what is the time to failure?
Answering these questions requires both operational data on current situations and long-term performance data for the whole fleet. These can be supplemented through additional data and knowledge brought in through benchmarking, peer review and discussion with colleagues in the same field.
Data, decisions and action
Data is the commodity which underpins decision making related to power transformers, whether they are in good condition or whether they need maintenance, refurbishment or replacement.
The availability of data on both their quality and timeliness is crucial to good decision making. It is the value of the decision which governs the investment in support data. Power transformer failure may have serious consequences in terms of the replacement cost of the transformer, the impacts on reliability and availability statistics, and in terms of, for example, safety or in collateral damage.
Asset owners collect asset data to understand the nature of the asset groups they have – by manufacturer, by design, by location, by impact or criticality.
Data may be used in several ways:
- By itself and with reference to its own history, as with the operating time on a breaker, looking for variation against ‘expected’ values.
- In synthesis/combination with other available data – transformer load and top oil temperature, noting that a correlation may be time delayed.
- In aggregation with other individual units: Does one unit stand out? Does one unit have an operating time within specification, but which is slower than others in the same family? Is the relationship between two parameters not one of cause and effect, but one of two effects deriving from a common cause?
Condition data are collected as part of an iterative process to identify actions – maintain or replace. Condition and operational data may require both intervention in the short term and longer term investment. The stages of the overall iterative process may be summarised as:
- Collect data – condition, operation, family/design, industry etc.;
- Analyse data – looking for anomaly and outliers;
- Identify anomaly – based on individual asset data, which may combine condition and operation data or aggregate data across several assets or asset types;
- Diagnose anomalous data – find possible causes;
- Identify a prognosis for anomalous data – what is likely to happen and with what consequences in what timescales;
- Plan for intervention – replace a failed unit, repair/refurbish a unit that is failing ‘gracefully’ or plan for a longer term investment. One action plan may be to not physically intervene, but plan for failure;
- Monitor the situation – some decisions may be ‘fuzzy’ or ill-defined, continue analysis and review timeliness and appropriateness of intervention;
- Iterate – further data collection, and checking and review of ‘anomaly detection’.
Several actions are covered by ‘intervention’, one of which is ‘do nothing’. That is really a misnomer because one of the requirements is to check that doing nothing is still an appropriate action as time moves on.
Decisions regarding power transformers are often required in ‘real time’, with an immediate response needed after a system event. Below we present three monitoring and survey approaches which support good decision making in a tactical situation.
Dissolved Gas Analysis (DGA) testing is a well understood test for determination of transformer condition. There are many standards available for interpretation of the results.
Regular samples, particularly for larger or more critical units, yield a regular view on transformer condition, with a good DGA programme giving early indication of failure in up to 50 per cent of incipient failures. Data from annual sampling, though relatively sparse, is thus effective as an asset management tool. Figure 2 gives an indication of key DGA levels for a transformer which subsequently failed.
|Figure 2: Key gas evolution over time measured by DGA in the laboratory|
Although most gas levels had been stable for some time, the hydrogen had been showing an increasing trend and the final failure brought a dramatic increase in most dissolved gas parameters.
By itself, regular DGA is a useful asset management tool in assisting with the identification of suspect units. An online DGA monitor gives further information, bridging the ‘silence between samples’ which can mask rapid deterioration.
From the data in Figure 2 it can be seen that an online monitor, such as a Doble Delphi device, may have been able to give early warning of the failure if it had been applied and there was a ‘graceful’ element to the deterioration. Of course, if the failure was sudden and catastrophic there may have been no ability to act. It is interesting to note that DGA for transformers covers the possibilities of either regular or occasional ad hoc sampling, and continuous online monitoring. Both require their own individual asset management approaches.
The application of online Partial Discharge (PD) monitoring of power transformers in service is one of the most promising technologies to detect and localise defects in the coil insulation. With the results of these diagnostic methods and the consideration of the network operation and management, a condition-based predictive maintenance and replacement planning is feasible.
Figure 3 indicates phase resolved PD in a transformer which was a critical component of a transmission system. This ‘baseline’ provides a clear visual indication of the state of PD in the system. A subsequent measurement set gave the results, also shown in Figure 3. The characteristics of the phase resolved PD are very different – there is a lot more PD across the whole phase range, indicating a change in the nature of the PD itself.
|Figure 3: Comparison of an initial and final phase resolved PD signature|
In this case, the change in PD gave early indication of a change in the transformer which could have led to failure. The data in Figure 3 were collected in the four hours before noon on a single day, and then subsequently in the eight hours after noon on the same day. What caused the change in PD character? Subsequent investigation showed significantly deteriorated insulation within the transformer windings, which would have continued to a failure of the unit if left in service.
Over the last decades there has been an increase in interest in monitoring the performance of bushings, known as Bushing Leakage Current monitoring, both in cases where the bushings are on transmission and generator transformers, and in cases where failure modes are suspected for particular bushing types. To extract the value of monitoring, an understanding of the measurement being made is critical, and the effect of power system and ambient conditions must be considered.
In this case, a wye-configured 370 MVA, 345/129 kV autotransformer was manufactured by 1986. Due to the transformer being a critical link for an independent power producer an online monitoring system, a Doble IDD was installed. At the time of commissioning, all the high-voltage winding and neutral bushings were GE Type U, and the low-voltage winding bushings consisted of two Lapp POC and one ABB Type O+C.
As part of the commissioning process, offline capacitance and dissipation factor measurements were performed on all of the bushings and found to be acceptable, i.e. less than 0.5 per cent.
Over time, degradation was detected by the IDD in the form of increasing power factor and capacitance, as demonstrated in Figure 4. As the insulation degrades, there would be an increase in leakage current through the insulation, eventually resulting in catastrophic failure.
|Figure 4: Degradation is detected as an increasing power factor and capacitance|
A decision was taken to remove the bushing from service and replace it. Offline testing confirmed the online indication of insulation degradation. The degradation in the insulation was found during a tear down, after the removal of the bushing.
While it is impossible to predict when the bushing would have failed, both the online and offline measurements indicated that the right decision was made to remove the bushing from service. The subsequent dissection of the bushing substantiated the decision to remove the bushing from service, because it revealed significant deterioration of the core insulation.
Planning is all-important
It is important to plan ahead for transformer failure and consequences. Both short-term response to issues and longer-term capital programmes need to be used to support the safe operation of the entire fleet.
The data generated by DGA, PD and through bushing monitoring may be voluminous and generated quite frequently, but the data are required in terms of both the importance and the frequency of the decisions being made.
It is a challenge to identify appropriate monitoring solutions and testing regimes that are both efficient and cost effective. There is a way to view the deployment process in terms of a double control loop, which links the short term ‘tactical response’ and longer term ‘strategic response’ (see Figure 5). But the double control loop does not reflect the dynamic nature of the situation completely. There is a spectrum of data gathering and timescales, and there is a spectrum of responses. But the double loop helps focus on short-term activities and longer term goals.
|Figure 5: Long and short-term loops for transformers|
The role of asset management is to understand the loops and how they enable us to move from one type of response to another. At some point, the response time in the outer loop is no longer adequate to manage an asset’s health. At such a point, more monitoring and testing may be necessary, but it is also essential to plan for deeper intervention: refurbishment and replacement.
Doble ARMS is a suite of tools that bring together risk and criticality information into a single visual display, as shown in Figure 6. Based on a geographic interface with drill down capabilities, DobleARMS manages both the long-term ‘watch list’ transformers targeted for replacement and also the short-term items, which need sudden and unplanned attention in a single tool, to provide owners and operators with situational awareness and up-to-date information.
|Figure 6: Doble ARMs takes an in-depth look at the whole system – the stations, the assets and the issues|
Supported by real-time data and Doble’s extensive database of results, the system also permits the entry of consequence or criticality data, should an asset fail or be unavailable. This provides an invaluable tool in managing risk because it is often the case that the equipment failure costs are very small in comparison to the contingency costs of safety, environmental and loss of service penalties.
A balancing act
Power transformer asset management is a balance of short-term intervention and long-term intervention. The intervention applied to individual transformers may change over time. They may move from longer term loops to shorter loops, for example, as planned maintenance becomes inappropriate and asset replacement is required.
The decisions being made with relation to those transformers are based on available data, and the value of the decision justifies the amount, frequency and quality of data collected. Online data generation, such as DGA, PD and bushing monitoring provide support and all-important insight into transformer condition.
The role of the asset manager in identifying the appropriate tests and monitoring for a particular situation is related to health, criticality and budget. Thus a risk-based approach is engendered through generation of appropriate data for on-going and future decisions.
The authors are all from Doble Engineering and are: Kenneth Elkinson, an apparatus analytics engineer; Matthew Lawrence a solutions manager for SFRA and Circuit Breaker Diagnostics, focusing on diagnostic testing solutions; Gregory Topjian, a solutions manager for Partial Discharge; and Tony McGrail, a solutions manager for On-Line Diagnostics, providing condition, criticality and risk analysis for utility companies.
For more information, visit www.doble.com