Are You Working on Use Cases in the Enclosed Categories?

  • Are You Working on Use Cases in the Enclosed Categories?

    Posted by Leslie Cook (Adm) on July 13, 2021 at 9:57 am

    …If yes, we want to hear from you! 

    Hello UAI Members, 

    Your fellow members are asking if you are working on use cases in the following categories or if you have more information to share on these hot topics. If you are working on use cases in the following categories or have information to share, please let us know by replying to this post or reaching out to me directly at lcook@utilityanalytics.com. Thanks so much for helping each other!

    • Substation asset analytics
    • Asset Optimization 
    • Asset Life for Vehicles
    • Transmission Smart Grid Journey
    • Condition-based maintenance for substation equipment
    • Integrating drone data with historians & analytics around drone data
    • IR data integration into asset health analytics
    • Tableau Mapping
    • Best practices on and how to tune modeling
    • Various Architectures
    • Share data science tools you are using and how you are using them: Azure, Azure Data Brick vs Machine Learning Studio vs Containers – how are members actually using Azure?
    • What models are you doing in the public cloud? 
    • Productionalization of models
    • Machine learning operations 
    • Team Orgs
    • How AMI has enhanced the operations of utilities – hearing from those experienced in developing an enterprise system and using the AMI to support it
    • Artificial Intelligence and Machine Learning
    • Customer Analytics
    • I would like to find out how other utilities are getting HR to provide a wage that allows them to retain people with desired skillsets
    • Unfortunately, safety is a smaller data set. Telematics from driving is the only exception. How can one work with smaller data sets?
    • Project tracking, Displaying trends, Navigating data analytics in a non-analytical role
    • Visualizations
    • User engagement
    • Data Engineering

    I look forward to seeing your replies and input on this post. Once I hear back, we can get some separate posts going on the individual topics and/or use cases. 

    Thanks so much! 

         @Leslie

    ——————————
    Leslie Cook
    Membership & Digital Engagement Manager
    Utility Analytics Institute (UAI)
    719-203-8650, lcook@utilityanalytics.com
    ——————————

    Jason Pegg replied 3 years, 4 months ago 3 Members · 3 Replies
  • 3 Replies
  • Leslie Cook (Adm)

    Member
    July 13, 2021 at 10:02 am

    Hello UAI Members,

    I noticed we have an unanswered question on Ask Your Network, so I want to reply to it to bring it back up to the top. Does anyone have any insight for @Hemanya on the following question?

    I was wondering what kind of algorithms are being used to detect anomalous customer usage behavior given that there is NO SMART METER data and the usage is captured monthly. Is there a way to capture events like zero usage, broken meter, bypass, etc. by using monthly usage data?

    Thanks in advance for providing your replies!

         Leslie​

    ——————————
    Leslie Cook
    Membership & Digital Engagement Manager
    Utility Analytics Institute (UAI)
    719-203-8650, lcook@utilityanalytics.com
    ——————————

  • Jason Pegg

    Member
    July 14, 2021 at 1:57 am

    Great question.  Full disclosure: I haven’t personally USED these solutions, but I’m actively looking into them (and will be talking about them a bit in this month’s Analytics Architecture and Technology meeting).  Amazon has a couple tools that may apply: Lookout for Equipment and Lookout for Metrics.  Both of them take in and train on historical time-series data (which would include monthly reads) and can then monitor new data coming in to find anomalies.  Each has strengths and weaknesses for different use cases, but you might look into these as possible (active) monitoring solutions.  At the very least, they may automate some of the searching for past, and alerting on new, anomalies.

    Note that I am not an Amazon employee, nor do I receive anything from Amazon for any recommendations.  I’m just a regular person working for a utility that happens to have selected AWS as its primary cloud provider.  :-)  Hope this helps!

    ——————————
    Jason Pegg
    Domain Architect
    Avista Corp.
    Spokane WA
    509.495.4731
    ——————————
    ——————————————-
    Original Message:
    Sent: 06-23-2021 12:36
    From: Hemanya Tyagi
    Subject: Anomalous Usage Detection

    Hi Everyone,

    I was wondering what kind of algorithms are being used to detect anomalous customer usage behavior given that there is NO SMART METER data and the usage is captured monthly. Is there a way to capture events like zero usage, broken meter, bypass, etc. by using monthly usage data?

    ——————————
    Hemanya Tyagi
    Senior Data Scientist
    National Grid
    7654098029
    ——————————

  • Eddie Fee

    Member
    July 14, 2021 at 8:14 am

    While we (OUC) have been an ‘AMI shop’ for several years now, we had some prior success in using the monthly usage data to identify issues like you have listed.  For example, while it was pretty ‘manual’ in nature, we would analyze by area (usually by traditional meter reading route) by showing a ​quick visualization in Excel of each premises electric usage over the past 18 months, along with an average daily usage value over the past 90 days.  Because this was done by meter reading route, it allowed an analyst to quickly ‘scroll’ through the route in the Excel format to find anomalies, basically neighbor-by-neighbor, that may require field investigation, and lead to discovery of theft/tamper or faulty meters.  We also had the advantage of being a water utility, whereby we could also show on the graph their monthly and average daily water usage…and had even extended that to calculate a water-to-electric usage ratio.  Incorporating the premise water usage into the Excel visualizations and calculations improved the hit rate of actual revenue issues (theft/tamper or meter issues).  I would guess that incorporating premise gas usage would provide similar benefits (assuming your electric & gas territories align).

    ——————————
    Eddie Fee
    Director, Metering & Field Ops
    Orlando Utilities Commission
    Orlando FL
    407.434.2262
    ——————————
    ——————————————-
    Original Message:
    Sent: 06-23-2021 12:36
    From: Hemanya Tyagi
    Subject: Anomalous Usage Detection

    Hi Everyone,

    I was wondering what kind of algorithms are being used to detect anomalous customer usage behavior given that there is NO SMART METER data and the usage is captured monthly. Is there a way to capture events like zero usage, broken meter, bypass, etc. by using monthly usage data?

    ——————————
    Hemanya Tyagi
    Senior Data Scientist
    National Grid
    7654098029
    ——————————

Log in to reply.