EPRI: Electric Power Research Institute

Smart Grid Resource Center

EPRI Distribution Modernization Demonstration (DMD) Data Mining Initiative

Data Mining Initiative

Proposal: The Distribution Data Mining Initiative is intended to validate solutions to some of the key data challenges faced by distribution system and power delivery service providers. Its purpose is to leverage collaboration between EPRI, electric service providers, and data solutions providers, including both academia and data analytics companies. DMD members will provide real power system datasets and supporting information to the data repository. The enhanced partnerships established as part of this initiative will foster a better understanding of industry needs, capture leading data analytic practices, transfer knowledge from industry experts, and accelerate ideas (solutions) to the market.

Sponsoring Members:
Numerous DMD Project Members

Background:
Electric service providers across the world are finding new ways to leverage smart grid investments and big data technologies to provide improved power system management and visibility. Simultaneously, there are many solution providers (with various levels of experience in data analytics) offering applications that may yield new and unique insights from utility datasets.

The challenges for electric service providers are the amount of time, effort, and financial investment required to evaluate the numerous data analytics applications and to determine how valuable the new insights are. Additionally, electric service providers have difficulty in obtaining the right data sets (from existing silos) in order to validate the performance of the applications.

To begin to address these challenges, one service provider (Duke Energy) created a data repository they referred to as a “Sandbox.” This dataset/model combined data elements from a variety of systems that were useful to analyze and identify new value opportunities from Smart Grid sensors and other data sets. The initial effort by Duke Energy successfully identified opportunities to implement data solutions and the insights and value that would be gained from these implementations.

The EPRI initiative builds upon the Duke Energy work by expanding the research into data analytics to an international group of service providers with a larger number of datasets spanning months of time.

In order to accomplish the initiative’s objectives, the work will:

  • Develop and maintain a data repository where the analysts can get the data sets of interest.
  • Prioritize a wide variety of data driven value cases (or use cases) where the mining initiative would provide insight and/or benefits.
  • Define the internal and external data sets necessary to populate the data repository and support each use case.
  • Determine suitable data size ranges to adequately evaluate emerging big data technologies.
  • Document to the extent practicable, any data ingestion, semantic or other challenges, associated with the data sets used in the initiative.
  • Develop a consolidation of the most valuable use cases per utility business unit, and describe the implementation requirements to accomplish the use cases.
  • Estimate the (application value) of attaining the various use case insights.

Overall, the goal of this effort is to learn what can be done with existing data, identify insights from the data that were previously unknown or perhaps not even conceptualized yet, and to become more versatile with big data analytics strategy and activities.

Current DMD members:

  • Ameren Services Company
  • American Electric Power Service Corporation (AEP)
  • Arizona Public Service Company (APS)
  • Bonneville Power Administration (BPA)
  • CenterPoint Energy
  • Central Hudson Gas & Electric Corporation
  • City Public Service San Antonio
  • Consolidated Edison Company of New York, Inc.
  • Consumers Energy
  • DTE Energy
  • Electric Power Board of Chattanooga (EPB)
  • FirstEnergy Service Company
  • Hydro One Networks, Inc.
  • Hydro-Quebec
  • LG&E and KU Energy LLC
  • Long Island Power Authority (liP A)
  • New York Power Authority (NYPA)
  • Pepco Holdings, Inc.
  • Sacramento Municipal Utility District (SMUD)
  • Salt River Project Agricultural Improvement and Power District (SRP)
  • Seattle City light
  • Southern Company
  • Tennessee Valley Authority (TVA)
  • Tennessee Valley Public Power Association (TVPPA}

Data Analytics Cases:

Assets Awareness
Predictive Health Index for Distribution Service Transformers 278 KB
Predictive Health Index for Non-communicating Reclosers 293 KB
Predictive Health Index for Distribution Line Regulators 247 KB
Distribution Capacitor Bank Problem Detection 250 KB
Load and DER Awareness
Customer-owned PV Forecasting 236 KB
Detection of Electric Vehicle Load Signatures 360 KB
Identifying Load Abnormalities 235 KB
Load and DER Signature Recognition 231 KB
Outage Awareness
Sequence of Outage Events Replay 248 KB
Leveraging AMI Meter Flags to Analyze Momentaries and Voltage Sags 233 KB
Dynamic Momentary Outage Detection and Calculator 229 KB
System Awareness and Grid Optimization
Optimal Placement of Automated Distribution Switches 233 KB
Optimal Sizing and Placement of Distribution Capacitor Banks in Conjunction witr Controllable Smart Inverters 235 KB
Load Balancing Using SCADA and AMI Data 233 KB
Selecting Bellwetrer Meters for Grid Optimization Programs 234 KB
Virtual Monitoring of Distribution Lines 294 KB
Development of Electrical Load Model Utilizing SCADA and AMI Data 235 KB
Near Real-time Measurement and Verification (M&V) for Grid Optimization Programs 253 KB
Visualization of Distribution Network Voltage Excursions 366 KB
Foundational
Validating Meter to Service Transformer Association 249 KB
Verifying Correct Phasing and Measurement Excursions of SCADA Devices 232 KB
Meter Connectivity (Phase ID) 278 KB
Correcting Meter to Protective Device Hierarchy 263 KB
Impedance Calculation of Secondary Conductors 235 KB
Imagery Related Categories (New Area trat is Being Explored)
Leveraging Imagery to Assess Wooden Distribution Pole and Cross-arm Health 232 KB