To identify transformers in the fleet which warrant attention for maintenance, refurbishment, or replacement, asset managers often use transformer assessment indices, or health scores to rank transformers within the fleet. However, there are no industry standardized methods for health scoring and methods used vary. Effective health scoring is essential so that the asset managers rank which transformers are high risk, having the most urgent needs, versus transformers which are low risk.
Many traditional methods used to generate health indices tend to focus on laboratory oil analysis: dissolved gas analysis (DGA) and oil quality tests. Oil analysis can provide valuable insight into developing fault conditions in the transformer main tank or the load tap changer (LTC). However, some failure modes in the Main Tank and LTC may be difficult to diagnose specifically from oil quality and DGA tests alone. Also, transformers can fail from defects in components outside of the Main Tank or LTC.
Transformer bushings, and cooling system problems can also cause failure, and these are not properly evaluated using health indices that focus solely on oil analysis. This article presents a practical method for calculating transformer asset health indices called Condition Index which evaluates condition on all major transformer components including Main Tank (Active Part), LTC, bushings, and cooling system. The method is holistic and combines data from oil analysis along with inspection data, electrical tests, and transformer online monitoring, integrating these into a single health score metric.
Holistic Assessment
Parameters utilized in the Condition Index are shown in Table 1. With this comprehensive scope of parameters considered in the asset health indices evaluation, more accurate and detailed health assessment is possible.
With this comprehensive scope parameters considered in the asset health indices evaluation, more accurate and detailed health assessment is possible.
Condition Index Health Score Method
Methods for calculating transformer assessment scores vary among utilities and in industry. Asset Managers use different methods, with some developed in-house by subject matter experts, and others developed by external 3rd Party services. Although there are no standardized approaches across industry, some concepts and strategies are described in the Cigre Guide 761 – Condition Assessment of power transformers[1].
The Condition Index method in this article is categorized as a hybrid scoring method by Cigre 761. Hybrid methods use two components to evaluate transformer health. The first component, called the Condition Group, indicates the ‘Worst-case’ condition of any component: Active Part, Cooling System, On Load Tap Changer, and Bushings. The second component is called the ‘Overall Health Index’ and is a weighted sum of all the transformer diagnostic parameters indicating the overall health.
Condition Group is an integer value between 1-5, with 1 indicating ‘Good’ condition, and 5 ‘Short Term Risk’ indicating equipment is at high risk for failure. Overall Health Index is a decimal value from 0.00 to 0.99, with 0.00 being the best overall health and 0.99 being the worst overall health. The two components are combined in a single quantitative metric ranging from 1.00-5.99:
Condition Index Format:
X.YY
Where,
X – Condition Group of the transformer (worst case method)
YY – Overall health index (weight scoring method)
Condition Group and Overall Health Index scores are assessed based on thresholds applicable to components and parameters listed in Table 1. For brevity and proprietary reasons, thresholds which determine the Condition Group and are not shared.
Qualitative descriptors for Condition Groups are provided in Table 2, these are based on information from Cigre Guide 227 – Life management techniques for power transformers[2].
An example of an evaluation for Condition Index is shown in Table 3. In this example the Bushings component is ‘Worst-Case’, yielding a Condition Group score of 4, and the weighted sum of all parameters yields an ‘Overall Health’ score of 0.284. The final Condition Index score is rounded to 4.28.
Benefits for Fleet Ranking
A key benefit of the Condition Index health assessment indices is for use in fleet ranking transformers. Because the Condition Group value represents the ‘Worst Case’ component, if any single component has a serious defect this immediately drives the asset health score to higher risk in the fleet regardless of the Overall Health.
The metric also enables transformers to be ranked even when not all transformers in the fleet have the same data available for assessment. Some transformers may be equipped with online monitoring whereas others are not. Or some transformers may have recent electrical tests while others have not been tested for many years. The Condition Group determined by the worst-case component assures that based on the information available the asset is ranked roughly where it belongs relative to other transformers in the fleet. Beyond this coarse ranking from Condition Group the Overall Health Index then allows for more fine-tuned ranking within the list.
Figure 1 shows a set of assessed transformers with Condition Indices ranging from Short-Term to Long-Term Risk. The Location and Company names fictitious to maintain customer anonymity. The transformer with the highest Condition Index score of 5.12 has a relatively low Overall Health Index of 0.12 (close to good) but is elevated to the top of the list because of one component exceeding the Short-Term Risk Threshold. Multiple transformers are operating at Condition Group 4 but with Overall Health scores ranging from 0.47 to 0.10.
The Condition Index health assessment indices allow the asset manager to see which transformer has the most significant health risks, while also allowing for more detailed ranking and prioritization among transformers in similar condition.
The condition index Health assessment Indices allow the Asset manager to see Which transformer has The most significant Health risks, while Also allowing for more Detailed ranking and Prioritization among Transformers in similar Condition.
Benefits for Evaluating Risk
The format of the Condition Index metric lends itself well to evaluating Risk. Because the metric is a continuous parameter ranging from 1.00 to 5.99, it is possible to translate the Condition Index metric to estimate probabilities of failure (PoF). Risk is commonly calculated as follows:
$ Risk = $ Cost of Event × Probability of Failure
The consequences of a transformer failure and associated costs is not the same for all transformers. These consequences depend on multiple factors including the number of customers impacted, cost of equipment repair/damage to other equipment in the substation if the transformer fails under explosion and/or fire. Table 4 illustrates an example of how the Condition Index score can be translated to estimate PoF. The PoF ranges are suggested based on estimated failure rates from Cigre 248 – Economics for Transformer Management[3]. The actual failure rate can be interpolated within the range based on the Condition Index value. Note, the PoF figures from Cigre 248 are intended only to illustrate the application of this method; actual PoF rates in transformers can differ widely based on region, application, and operational history.
Table 3: Estimation of PoF based on Condition Index Health Assessment Indices
Conclusion
transformer reliability, for asset managers it is important to leverage transformer data and information as best they can to maximize availability. The Condition Index method for evaluating transformer asset health indices has advantages over traditional approaches that focus only on oil condition due to the Condition Index holistic assessment. Data from online monitoring and electrical tests can be ingested along with DGA and oil quality to evaluate health. Additionally, the hybrid scoring approach allows for better distributed ranking of assets and adapts well to raising risk score on worst case condition, as well as handling ranking where not all assets have all the same parameters available. The Condition Index metric can also be translated to estimate the PoF for evaluating Risk, when managers want to understand the Risk in terms of consequences and dollars.
References
[1] CIGRE WG A2.49, “Condition assessment of power transformers,” Technical Brochure, 2019.
[2] CIGRE WG A2.18, “Life management techniques for power transformer, 2003.
[3] CIGRE WG A2.20, “Guide on economics of transformer management”, June 2004.
Anatoliy Mudryk received his engineer degree in Electrical and Mechanical Engineering from the technical University of Zaporizhzhya in 1996. During his professional life, Anatoliy occupied leading engineering positions in ZTZ-Service, GE Kelman and currently is the Transformer Expert in Camlin Energy.
The main areas of activity include transformers design, diagnostics, and reliability. Anatoliy is member of CIGRE and IEEE transformers committee WGs.
Dr. Marco Tozzi is Asset Performance Director in Camlin Energy. He is currently involved in providing condition assessment of power transformers, supporting customers during the purchase of new transformers as well as developing advanced solutions to help utilities to convert data into decisions.
Marco has a background on DGA and Partial Discharge (PD) analysis and modelling, authoring patents on PD sensors and de-noising algorithms; IEEE senior member, active in CIGRE Working Groups, author of more than 50 technical and conference papers.
Dr. Nathan Jacob is Senior Asset Specialist at Camlin Energy. In this role Nathan provides expertise and technical guidance in areas related to condition monitoring, data analytics, and condition assessment for power transformers, generators, and other power system equipment.
Nathan has 19 years of experience in the power industry. He was formerly Senior Test Engineer at Manitoba Hydro’s High Voltage Test Facility and prior to that was Field Services Insulation Systems Test Engineer. Nathan is active in multiple technical working groups within the IEEE, namely IEEE Transformers Committee and the IEEE Electric Machinery Committee.
