A Review of Present-Day Transformer Loading Practices at Duke Energy

Presented by:

David Wallach

Duke Energy Corporation

Principal Electrical Engineer

TechCon 2022

Many transformer loading practices in North America are based on IEEE C57.91, IEEE Guide for Loading Mineral-Oil-Immersed Transformers, and Step-Voltage Regulators. A Guide allows the end-user more freedom to implement their own program to fit their own needs. Aspects of Duke Energy’s transformer loading program are reviewed and discussed.


Duke Energy is a result of mergers of multiple Investor-Owned Utilities over the last twenty or so years who each had their own legacy transformer loading practices. While each practice was based on the IEEE C57.91 IEEE Guide for Loading Mineral Oil-Immersed Transformers and Step-Voltage Regulators, the implementations differed. The IEEE Guide’s scope begins with “This guide provides recommendations…”

Duke Energy has documented and implemented transformer rating methodology for NERC defined Bulk Electric System (BES) transformers. We are gradually harmonizing our transformer rating processes over time. We must balance remaining in compliance with our existing regional documented process and resulting ratings until we are ready to “flip the switch” to an updated harmonized process. We are focusing on BES ratings. Our non-BES transformers also have ratings but the methodology between regions has not been harmonized. Practices for non-BES vary from the use of multipliers to calculated results.

This presentation will focus on the BES transformer rating approach.

Input Data

IEEE C57.91 discusses two calculation methods known as “Clause 7” and “Annex G”. Clause 7 is a simplified model that requires less information to implement. Table 1 shows some of the differences.

Clause 7 and Annex G
Annex G and Clause 7
Table 1

Until recently, test reports were not required to report bottom oil rise. There are ways to estimate inputs for older transformers with test reports that contain less information to utilize Annex G, however, the Clause 7 calculation method tends to see more use in the industry. Most rating calculations at Duke Energy use the Clause 7 method for simplicity and input data that is not available on most test reports.

Operating Conditions

Duke Energy uses a common Ambient Temperature Cycle across all regions (Figure 1). We have found one ambient temperature cycle shape adequately describes our temperature cycle in all seasons and all regions. We shift the daily peak temperature up or down as needed for the given calculation.

Ambient Temperate

The load cycle can vary between summer and winter as well as regional load differences that are influenced by the grid topography (e.g. location of generation sources, how retail substations are served, etc.). We implement these curves in Figure 2-4 by region. Winter curves are applied when the daily peak ambient temperature is 320 F (00 C) or below, else the summer curve is applied. The flat load curve is used in some locations near generators or some direct-serve customers.

Region Load Curves
Region 3 Load Curves
Region 4 Load Curves

Rating Durations and Peak Temperatures

A Team with inputs from System Planning, Operators, and Asset Management developed a rating methodology using a seasonal load cycle with peak ambient temperatures of 105, 100, 95, 90, 80, 70, 60, 50, 40, 32, 20, 10, and 0 degrees Fahrenheit. The Team also chose rating durations to include continuous, 1 year, 12 hours, 2 hours, and 15 minutes.


  • Continuous – available peak load ability of a transformer for a given daily peak ambient temperature and load cycle curve which can be repeated indefinitely.
  • 1 year – available peak load ability of a transformer for a given daily peak ambient temperature and load cycle curve which can be repeated for up to one year. The intent of this rating is for usage on transformers that may have reached planning maximum capacity sooner than expected and a project must be developed to increase capacity.
  • 12 hours – available contingency load ability of a transformer for a given daily peak temperature using the continuous rating as a pre-load to allow a transformer to carry a larger load for up to 12 hours to allow for switching of load, installation of a mobile transformer, etc.
  • 2 hours – available contingency load ability of a transformer for a given daily peak temperature using the continuous rating as a pre-load to allow a transformer to carry a larger load for up to 2 hours to allow for switching of load.
  • 15 minutes – available contingency load ability of a transformer for a given daily peak temperature using the continuous rating as a pre-load to allow a transformer to carry a larger load for up to 15 minutes to allow for switching of load.

We found the contingency onset times for our 12 hours or shorter durations are influenced by the onset time. We performed a sensitivity analysis with our load curves and adjust the onset time per region (per load curve) to obtain the worst-case condition resulting in the more conservative rating using Figure 5. When a contingency is only a small portion of the daily load cycle, we do not ‘maintain load shape’ to calculate one flat operating capability during the contingency to communicate as a limit during the contingency.

Onset Times

Calculation Approaches

There are different approaches available to develop a transformer rating.

  • Ultimate temperatures reached during user-defined load level for a certain duration
  • Load (MVA) above nameplate at which user-defined temperature limits are reached (duration also a variable)
  • Loss of life consumed during the cycle

Duke Energy uses the second approach defining the ultimate top oil and hot spot temperature limit for each loading scenario while keeping the loss of life consumed within reason under the third bullet with the selected temperature limits considered.

Figure 6 are the top oil temperature limits and Figure 7 are the hot spot temperature limits we use.

Hot Spot Limits

Duke Energy recognizes the ‘hard limits’ of Table 9 in IEEE C57.91-2011.

Calculations and Path Rating

Duke Energy uses EPRI’s PTLOAD v6.2 software tool to implement the calculations of IEEE C57.91. We model each transformer then use the ‘batch mode’ where we set up 65 discrete rating scenario to provide to planners and operators. All five ratings at the 13 peak temperatures with applicable load curves are calculated. One feature we use in the batch mode is for contingencies 12 hours or less to use the continuous rating (above nameplate) as the preload and load prior condition to the contingency onset.

Once the transformer is modeled and the results for the 65 rating scenarios are created, we drop the two PTLOAD files on a server. Processes run several times per day and the files are processed into our in-house developed software tool, Facilities Ratings Database. This is where our system planners build a path and develop the ratings for the complete facility. Once a facility rating is developed, it is published and is available to our Energy Control Center and System Operators.

Calculation and Path Rating

Restrictions and Limitations

Load Tap Changers

IEEE C57.131-1995, “IEEE Standard Requirements for Load Tap Changers” section 6.1.2 states that when LTCs comply with the definition of maximum rated through current (1.2 times maximum rated through current continuously and temperature rise of contacts does not exceed 20 deg C rise above fluid surrounding the contacts) can be loaded in accordance with applicable ANSI or IEEE loading guide. In this case, the applicable loading guide is IEEE C57.91-1995. The operation manual for the LTC typically states whether the LTC complies with IEEE C57.131. If the LTC does not comply with IEEE C57.131-1995 requirements or compliance cannot be verified, then we cap the transformer rating at the LTC manufacturer’s continuous rating.

Internal Leads, Connections, and DETC

Our present-day transformer specification specifies that the transformer hot spot shall be in the coil and not elsewhere. This rule results in some level of assurance that an internal lead will not be a limiting factor in transformer loading. Transformers purchased before the 1995-time frame before our purchase specification addressed this requirement, and the hot spot may be in a lead.

The rating process cannot take this into account due to a lack of internal design details, so the rating development process recognizes this as an acceptable risk.

It is assumed that the manufacturer has a proven method for secure, low resistance connections between leads or between leads and windings. The process of developing ratings above nameplate cannot consider manufacturing-induced concerns such as loose connections, poor quality brazed connections, etc., and is, therefore, a recognized acceptable risk. We found DGA analysis to be valuable for the detection of suspect connections when loading the above nameplate.

Like internal leads, we assume that the DETC will not be a limiting factor when determining transformer ratings. The present-day transformer specification method specifies a minimum DETC current rating in our performance requirements document (specific design detail document that goes with our general procurement specification) to provide operating margin for reasonable assurance that the DETC will not be a limiting factor.

Stray Flux

Duke Power hired Bill McNutt in 1994 as an industry consultant to provide advice to the developing transformer rating process. Mr. McNutt suggested that Transformers designs older than ~1975 and greater than 150 MVA may lack finite element analysis and could restrict some rating scenarios. Duke Energy limits ratings of this category of transformers to 125% of top nameplate rating when with a top rating is 150 MVA and above and manufactured prior to January 1, 1980.


A new transformer is expected to have <0.5% moisture in cellulose. Moisture levels can increase if the oil preservation system allows the oil to meet the atmosphere. Modern oil preservation systems:

  • conservator systems use an air cell and desiccant breather, or
  • inert air system uses a Nitrogen gas space replenished by an on-board bottle, or
  • a sealed tank system has a Nitrogen blanket gas space and remains sealed from the atmosphere, so it doesn’t need routine replenishment.

Duke Power ‘loved’ conservator tanks on transformers of all sizes, even before air cells, so the region that was Duke Power has many transformers with high moisture levels that we operate with. Moisture also generated from the aging of the paper insulation in a transformer even when sealed from the environment. The risk comes with loading a transformer with high moisture where bubble evolution can cause transformer failure. Duke Energy has a couple of documented failures due to operating transformers with high levels of moisture with a problem that developed with the cooling systems (e.g. fans not operating). Please see Annex A of C57.91 or review TechCon 2013 Paper “Catastrophic Power Transformer Failure Due to Bubble Evolution – A Case History”.

Other Concerns

Transformers can develop operating concerns as found by routine dissolved gas analysis (DGA) indicating a developing fault, a bushing type that has issues that can be exacerbated by operating at higher temperatures, compromised cooling such as a pump failure, or other reasons. We have TechCon North America 2022 – Phoenix, AZ methods in place to remove the calculated rating until the condition is fixed or the transformer is replaced.


A transformer does not have “one” overload rating but has permissible ratings depending on specifically defined conditions such as:

  • Load shape/pattern
  • Duration
  • Environmental Factors (ambient temperature and cycle)
  • Design Characteristics
  • Transformer Condition

It is up to the end-user to define each necessary condition, calculate each rating, and present them in a usable format for use by system planners and operators.


  1. IEEE dStandards Association C57.91-2011. “IEEE Guide for Loading Mineral Oil-Immersed Transformers and Step-Voltage Regulators”, 2011. IEEE 3 Park Avenue, New York, NY.
  2. Electric Power Research Institute (EPRI) (2007). PTLOAD v6.2.72 [Computer Software].

Join our email list

We use cookies to give you the best online experience. By using this website you agree with our cookie policy.