Utility Project Profile Database

SCE - Using Machine Learning to Monetize Risk in Pole Loading Program

  

ANALYTICS AREAS

 

  1. Which of the following analytics areas apply to your project (see definitions at end of document)? Place an “X” in the appropriate boxes.

 

 

Implementation underway

 

 

Project completed

Future

 

Meter data analytics

_______

 

 

_______

_______

Credit and collections

_______

 

 

_______

_______

Call center optimization

_______

 

 

_______

_______

Fraud detection

_______

 

 

_______

_______

Campaign management

_______

 

 

_______

_______

Customer segmentation

_______

 

 

_______

_______

Pricing optimization

_______

 

 

_______

_______

Demand response programs

_______

 

 

_______

_______

Energy efficiency programs

_______

 

 

_______

_______

Distributed generation management programs, including electric vehicles

_______

 

 

_______

_______

Transformer management

_______

 

 

_______

_______

Substation equipment management

_______

 

 

_______

_______

Overall transmission & distribution management

_______

 

 

____x___

_______

Outage management

_______

 

 

_______

_______

System modeling

_______

 

 

____x___

_______

Power quality optimization

_______

 

 

_______

_______

Advanced distribution management

_______

 

 

_______

_______

Analytics for real-time network operations

_______

 

 

_______

_______

 

Other: 

 

 

PROJECT OVERVIEW

 

  1. In two or three sentences or a short paragraph, please describe this project (include business drivers, how project aligns with utility’s corporate strategy, project goals, results achieved/desired outcomes).

 

SCE wanted to minimize the cost of pole maintenance on over-loaded poles while simultaneously maintaining reliability standards. Over the course of 7 years, data was gathered on distribution poles to use for the project. Using machine learning techniques, SCE can develop an optimized maintenance schedule for overloaded poles based on a given level of risk and cost associated with risk.

BUSINESS VALUE

 

  1. Describe actual and/or anticipated QUALITATIVE (e.g. better customer service) and QUANTITATIVE (e.g. ROI, cost savings, revenue increases) benefits of this project.
    • Note: Any and all quantification of benefits helpful including estimates, ranges, percentage increases or decreases in costs, resources, etc.

 

Qualitative impacts are improved O&M performance and lineman safety through proposed optimized maintenance program. Potential quantitative impacts listed in chart below. 

 



BUSINESS PROCESS IMPACTS

 

  1. List the key business processes impacted (or expected to be impacted) and describe the nature of those impacts.

 

The goal is really to optimize pole maintenance to save on O&M costs, improve safety, and meet reliability standards (or better, to improve upon them). Maintenance scheduling is the main business process potentially impacted, in that SCE will move from scheduled pole maintenance to optimized pole maintenance.

 

TECHNOLOGY

 

  1. Identify the PRIMARY TECHNOLOGY COMPONENTS associated with this project including the technologies used (or expected to be used) for: 1) data collection/integration 2) data analysis 3) data presentation and 4) data storage.
    • Notes: Identify the solution providers and products by name if possible

 

For the pole loading model, the Gradient Boosting Machine algorithm produced the best results tested against actual failure data. This algorithm has been used for many years in predictive analytics and is included in all of the most widely used analytics platforms including R, Python, and SAS. The model was designed to rank the poles from most likely to pass to least likely to pass as well as the chance that the model is wrong.

 

  1. Identify the PRIMARY DATA INPUTS/ELEMENTS utilized and their sources.



There were 85,756 observations, of which there were 8,890 failures. The model was trained on 68,604 observations; 17,152 were held out to validate the model.


PEOPLE

 

  1. Identify the following individuals by job title (as applicable):

 

  • Business-side Project Leader: _Statistical Data Scientist________________

 

  • Technology-side Project Leader: _____________

 

  • Executive Sponsor: Director, Performance & Reliability

 

 

  1. Identify and describe the role of external providers (e.g. strategy, consultant, design, development, implementation, ongoing operations).

 

 

 

  1. Describe the investments required to implement and provide ongoing support for this initiative.

 

  • Internal resources (identify type of internal resources, e.g. job titles and quantities of each (e.g. # of full time/# of part time, total FTEs):

 

 

 

  • Technology investments (estimate/provide ranges for one time and recurring costs if possible or indicate that existing technologies were leveraged):

 

 

 

 




BEST PRACTICIES/LESSONS LEARNED

 

  1. What advice would you share with those either planning or implementing a similar initiative (e.g. best practices, lessons learned, what to do, what not to do)?

 

 

0 comments
57 views

Permalink