Research Status & Deliverables

Task: Power Transformer Population performance based on industry wide data
Task Description: Collect & Analyze Industry wide Data (in-service and failure records, maintenance work orders, bushing power factor and other test data) to better understand transformer population performance (e.g., hazard rates, oil leak rates, tap changer problems etc.) and inform other research (e.g., spare strategies)
Update:

• To date, received in-service and failure records from over 35 utilities. In cases where the data is not up to date we are reaching out to utilities. Careful curation and processing underway. Processed data is used in analysis to better understand failure and replacement rates by family, make, model, application, and age.
• In addition, we have also collected and processed maintenance records (work orders) and historical bushing power factor records from four utilities.
  • Maintenance records (work orders) analysis provides insights to better understand maintenance issues affecting transformer performance - incurred O&M, leak rates, tap changers, cooling system etc.
  • Bushing power factor data analysis provides insights to draw meaningful and statistically defensible inferences regarding bushing performance.

Task: Power Transformer Expert System Software (PTX)
Task Description: Develop and deliver new version of the software.
Update:

• Work underway to –
  • Continued Main tank assessment indices improvement with continuous validation based on blind testing using data set of in-service & failed transformers, correlating the paper degradation index with degree of polymerization test results & improve moisture assessment
  • Improve software with increased flexibility in import from databases, improve reporting & visualization
  • Test online DGA monitor data analysis algorithm with
  • Field data to ensure minimal false positives &
  • Field data overlaid with lab generated fault data to ensure timely discernment of fault

Task: Circuit breaker Population performance based on industry wide data
Task Description: Collect & Analyze Industry wide Data to better understand current performance, project future performance, and inform additional research:

• In-service & removal records to better understand replacement rates by family, make, model, application & age
• Maintenance records (preventive & corrective) to develop new insights & analytics that utilities can use to improve existing maintenance programs, selection of new breakers & deciding which ones to repair/replace/refurbish
  Update:
• Guidelines to aid consistent data collection - reviewed and made necessary updates to the technical content
• To date received raw data for over 100,000 transmission circuit breakers from 22 utilities. Careful data curation and processing is underway. Data from 12 utility fleets is being used to develop new metrics and analytics for:
  • Maintenance program development; task and timing selection
  • Benchmarking – comparison among utilities and breaker makes and models
  • Replacement decision support
  • Specification and selection of new circuit breakers

Task: Circuit Breaker Replacement Ranking Software (CBRR) development.
Task Description: Develop and deliver new version of the software.
Update:

• Efforts underway at seven new utilities to implement the CBRR tool using their fleet information
  • Applying methodology on live as well as dead tank circuit breaker fleets
• Investigating how to incorporate new inputs in the ranking process
• Incorporating industry wide circuit breaker data analysis results
• Incorporating lessons learned from framework implementation at various utilities

Task: Advanced Data Science, Artificial Intelligence & Statistical Analytics
Task Description: Evaluate the application of Artificial Intelligence techniques for Natural Language Processing to extract meaningful information from descriptive records (e.g. maintenance work orders, outage records, field test reports etc.)
Update:

• Gathered free form records (unstructured data) from member utilities (demographic, historical maintenance and outage records and field inspection reports)
  • Circuit breakers (22 utilities)
  • Capacitor banks (4 utilities)
  • Transformers (4 utilities)
  • Disconnect switches (one utility)
  • Substation batteries (4 utilities)
• Investigated nine different Natural Language Processing algorithms to assess whether they can be applied to:
  • Organize and curate data
  • Classify maintenance and outage records into meaningful categories
  • Extract actionable information such as dominant issues and trends, non-apparent patterns, and relationships
• Initiated new efforts to apply NLP algorithms to improve information retrieval from:
  • Root cause & failure analysis reports
  • Historical test reports in .pdf or .doc format
• Initiated new efforts to develop Dictionary of Terms and Acronyms to facilitate consistent extraction of information from typical utility text records

Task: Balance of substation asset performance analysis
Task Description: Collect and analyze readily available data for balance of substation assets (relays, batteries, capacitor banks, disconnect switches, instrument transformers, ground grid etc.)
Update:

• Received and analyzed substation battery and capacitor bank data from four utilities
• Resulting insights help better understand:
  • Maintenance trends and rates
  • Performance characteristics of certain populations
  • Influence of factors such as age on asset population performance

Planned Deliverables for 2023

2022 Deliverables