Skip to main content
This project ended in Jul 2023 and is now closed.

Smart Meter Innovations and Test Network (SMITN)

Network operators
  • National Grid Electricity Distribution
Funding mechanismNetwork Innovation Allowance (NIA)
DurationMar 2022 - Jul 2023
Project expenditure£915k
Research areaOptimised Assets and Practises
  • South West
  • South Wales
  • West Midlands
  • East Midlands
  • May 2023

    Early work on the closedown report has started. The first draft of the closedown report is under review. Preparing for Loughborough to present a SMITN poster…


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



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.



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.