Smart Meter Innovations and Test Network (SMITN)

Objective(s)

The objectives of the project are;

1) To determine a representative test network of selected distribution substations and validate the key features of this network by carrying out surveys.

2) To capture smart meter data using new aggregation groups

3) To develop algorithms using smart meter data for the following use cases.

a. Customer to Phase connection prediction

b. Customer to Feeder connection prediction

c. Low Carbon Technology identification of potential locations and types

d. Provision of LV feeder and Distribution substation planning profiles for use in network planning.

4) To apply the algorithms for data relating to the test network

5) To assess the performance of the algorithms and where possible identify the factors that affect accuracy

6) To capture the learning from the project and disseminate this to interested parties

Problem(s)

The phase to which a single phase customer is connected has historically not been captured and where given has often been shown to be incorrect. Similarly the feeder to which a customer is connected has not always been captured correctly. With the expected increases in load at LV from heat pumps and Electric Vehicle charge points, there is a risk that the degree of unbalance on a network could become significant and that assuming a balanced network would not be a reasonable planning assumption. There are still issues with low carbon technology installations not being registered with Distribution Network Operators (DNOs) and without correct records the accuracy of network planning will be affected. At the same time, most LV networks are not monitored and the profiles used for planning do not incorporate the data that is becoming available from smart meters. This project aims to solve the problems of missing or incorrect data for LV networks by applying algorithms to smart meter data in novel ways.

Method(s)

The project will resolve the issue of missing or incorrect data for LV networks via a series of work packages.

The first work package will determine the test area to be used including substations of different types intended to reflect the variety of our service areas. It also includes researching the algorithms available using smart meter data to address the issues, select those that are appropriate and propose amendments to take into account the data available. The research into available algorithms will include a workshop with other DNOs to avoid duplicating the work that has already taken place in this area such as the work from Scottish Power’s NCEWS project (Network Constraint Early Warning System) which has been reflected in their NAVI platform.

In order to be certain that the data for the test network is correct, a phase survey will confirm the phase to which customers are associated. This work package also covers determining and setting up the data processing architecture to be used during the project. Data processing to evaluate the algorithms will take place in a custom built environment provided with suitable data from National Grid Electricity Distribution (NGED) systems, survey results, monitoring data and smart meter data. The data processing environment is then validated by GHD to ensure that the algorithms have been set up correctly. Smart meter load data will be aggregated to avoid privacy issues but where possible these aggregation groups will involve smaller sections of LV feeders than previously used. Similarly voltage data will be captured at MPAN level to determine whether this can provide valuable insights.

Work Packages 2-4 apply the algorithms to the defined test network for phase identification, low carbon technology detection and to create estimates for feeder and distribution substation load that can be compared to data from site monitoring. This will reflect the methods used for settlement of non-half hourly metered customers as well as incorporating the available smart meter data. Each work package will create a report capturing the learning.

Work Package 5 includes the development of the feeder finder tool so that it can be used on another survey to confirm the feeders to which customers are associated without requiring access within customers’ premises.

Once developed and tested this feeder finder tool will be used to carry out a survey of part of the test network to allow the feeder association algorithm to be evaluated.

Finally Work Package 6 involves gathering the learning together from the separate work packages to create a final report before delivering a dissemination event.

• Scope

  The scope of the project involves investigating and applying algorithms that use smart meter data for the following use cases.

  1) Customer to Phase connection prediction

  2) Customer to Feeder connection prediction

  3) Low Carbon Technology identification of potential locations and types

  4) Provision of LV feeder and Distribution substation planning profiles for use in network planning.

  If the project is successful, it will lead to improvements in data quality and availability which will then allow for;

  1) More accurate planning by modelling unbalanced three phased networks.

  2) Reduced losses through better phase selection for new connections

  3) Reduced fuse operations from imbalanced phases

  4) Enhanced data available for LV Fault Location

  5) Improved IIS reporting

  6) Reduction of sites that require monitoring and more effective targeting of

  7) Reduced costs to confirm customer phase

  8) Identifying and fixing data errors ahead of need

  9) Enabling LV self-serve facilities

  10) Improved use of our data by third parties

  The financial benefits from these improved outcomes is approximately £740k per annum within NGED’s licence areas made up of;

  • Savings from reduced losses when connections are made to existing networks estimated assuming 20,00 connections each year to existing unbalanced networks and a 10% reduction of losses resulting from reduced imbalance. (approx. 11k)

  • A reduction in fuse operations by improved management of phase imbalance and altering fuses where this is seen via modelling to be a high risk resulting in approx. £79 k per annum saving based on 100 events a year at £790 each.

  • A reduction in excavating at inaccurately identified fault locations due to incorrect customer connectivity saving approx. £340k per annum assuming that this is avoided 200 times a year at a cost of £1.7k per event

  • A reduction in the cost of monitoring the network by having better data to better select locations where monitoring is required and being able to omit some locations where monitoring is not required.

• Success Criteria

  The success of the project will result in;

  • A set of phase validated networks and a subset where customer-to-feeder association has also been validated that can be used to test future innovations beyond this project.

  • An understanding of how well the smart meter data can be used to support phase identification and what factors affect the accuracy of the algorithms used.

  • An understanding of whether smart meter data can be used to support the identification of unregistered LCT and what factors affect the accuracy of the algorithms used.

  • An understanding of whether smart meter data can be used to support the validation of customer-to-LV feeder associations and what factors affect the accuracy of the algorithms used.

  • An understanding of how well aggregated smart meter data can be combined with other data available to DNOs to create estimated LV feeder load profiles and what factors affect the accuracy of these estimates.

  • An understanding of the practical issues for using the newly developed feature for validating customer-to-LV feeder association without requiring connection to the customer’s wiring and confirmation that the new feature is fit-for-purpose

  • An increased insight of the phase imbalance on the test networks.

  • A view of the value of disaggregated data from smart meters rather than the default position of aggregation at the LV feeder level.

• Potential for New Learning

  The knowledge dissemination activity will cover the learning from the practical implementation of the algorithms to identify the phase, feeder for a given MPAN and any unregistered LCT for an MPAN or small group of MPANs.

  This will outline how well they worked and what factors affect the success rate such as smart meter coverage.

  The learning from the feeder and substation level profile estimation methods will identify whether the additional complexity of replicating the settlement process results in improved results compared to simpler methods. It will also detail the accuracy of the profiles and how they were adapted for use in Connect LV.

  The learning will also include that from the development of the Feeder Finder, how it was validated to ensure it was suitable for use and practical issues that were encountered during the survey.

   

  Learning will be disseminated during the project externally via;

  • One or two SMITN Webinars

  • LCNI

  • Innovation Showcase

• Documents and Links

  NIA Project Registration and PEA Document

  Annual Progress Report 2022 

  Early Learning Workshop Slides

  Combination Load Profiles 

  MPAN to Feeder Association Analysis Report

  Phase Identification Report

  Voltage Validation of Unregistered Equipment Report

  Dissemination webinar slides

  Project Closedown Report

   

• Monthly updates

  • {{update.title}}

  Show more