Clock Icon - Technology Webflow Template
min read

GRIDSIGHT’S INSIGHTS Q1 & Q2 2023 | Network Capacity

We discovered the most actionable data, network and customer insights in Q1 & Q2 2023. It’s all about network capacity.

At Gridsight, our mission is to accelerate the grid's transition to renewable energy. As we increase the number of electricity distributors we're working with across Australia and New Zealand, we're increasingly discovering new insights that many in the industry are often not yet aware of. 

Gridsight’s Insights is our quarterly highlights reel of what we’ve discovered while working with the most innovative and forward thinking electricity distributors in Australia and New Zealand.

Below you’ll find the most actionable data, network and customer insights that we discovered during Q1 & Q2 of 2023. Over the past six months we've uncovered some key insights centred on Network Capacity. We hope that you find this report useful. If you would like to learn more about any of the insights covered here, please reach out to us - we would love the opportunity to discuss them further with you.

Here is the downloadable PDF version.


Network and Transformer Loadings

Using downstream NEM12 consumption data, we can estimate distribution transformer (TX) loading and understand load and generation thermal constraint headroom.

The estimations can be produced per timestep and inherently account for known factors affecting network net loading such as seasonality, weather conditions and type of day (weekday / weekend) as shown below in Figure 1.

Figure 1: Transformer Load Estimation Timeseries and Summary Statistics

Using the confidence interval for each estimation point along with an overall categorical confidence flag, users can now make well-informed decisions about the estimated loading on each distribution transformer.

The connections, planning and asset management teams can now readily leverage the temporally aggregated transformer loading summaries, the yearly load bandwidth graphs, and the load duration curves for assessing network thermal capacity and load profiling. These aggregations enable per substation minimum and maximum loading to be identified, both for all time and specific date ranges, and can identify the specific timestamp when maximum or minimum loadings have been estimated. An example is shown below in Figures 2 and 3.

Figure 2: Transformer Loading Bandwidth Chart
Figure 3: Load Duration Curve

The developed per transformer analytics can also be combined across the fleet of transformers to enable assessment of overall average and peak asset utilisations. These can then be aggregated to assess overall transformer fleet utilisation. Figure 4 below provides an example of a sample Gridsight managed dataset.

Figure 4: Fleet Level Distribution Transformer Capacity Report

Figure 5 below shows the results of the capacity investigation, which included 27,000 transformers. This graph shows the relative utilisation bins in 25% increments (e.g. a value of 0-25 contains all transformers with a relative load between 0 and 25%) and the yaxis the proportion of transformers that fall within this bin. The cumulative proportion provides the relative proportion of transformers below a given upper bin (e.g. a cumulative proportion of 0.75 in the 51-75% bin indicates that 75% of transformers have a utilisation less than or equal to 75%).

Figure 5: Overall Transformer Fleet Relative Utilisation

Analysis of the above graph shows that approximately nearly 3/4 of transformers have at least 25% of their capacity remaining and only 13% have a maximum utilisation over 100%. This indicates that in general there is sufficient thermal capacity in the majority of distribution transformers to support some level of electrification and network expansion.


Calculating Hosting Capacity with Low Meter Penetration

Validation against transformer monitor data have shown accurate estimations of net power at the distribution transformer with consumption penetrations as low as 10%.

Figure 6: Typical Meter Penetration & Loading Distribution
Figure 7: Effect of Meter Penetration on Power Estimation Error

The development of Gridsight algorithms used validation with transformer monitors to provide a ground truth comparison. Figure 7 provides an overall summary of the observed accuracy vs meter penetration trend. It is evident that accuracy improves as meter penetration increases, with the trend following a relatively linear relationship. The decreasing accuracy points at the 50 and 80% penetration bins are a function of smaller sample sizes in these bins coupled with the inherent variability amongst the distribution substations comprising these bins. In general, useful accuracies can be achieved with meter penetrations as low as 10-20%.

Example comparison timeseries can be seen below in Figure 8, with meter penetration increasing from 10 to 50%. Note the legend on the graphs showing transformer monitor data in blue, estimate values as a dashed orange line, and the confidence range as an orange shaded area around the estimate. The width of the confidence range indicates the confidence of the estimate, and for good estimates, the transformer monitor trace should exist within a tight confidence range. The purple trace indicates the meter sum, with the difference between the sum and estimate traces indicating the magnitude of the estimated load.

Gridsight continues to work with its extensive data and customer base to improve transformer load estimations. This constant development ensures that workflows related to connections, planning and hosting capacity are more accurate and reliable as networks become increasingly embedded with decentralised and renewable resources.

Figure 8: 10% Meter Penetration Transformer Monitor Accuracy Comparison Example
Figure 9: 30% Meter Penetration Transformer Monitor Accuracy Comparison Example
Figure 10: 50% Meter Penetration Transformer Monitor Accuracy Comparison Example



Gridsight-powered load estimations are delivering customer benefits by:

  • Enabling faster, more accurate and more consistent response times to customer connection enquiries by having load and generation headroom readily available.
  • Proactively and holistically identifying constrained areas of the network allowing for the planning of network upgrades to support electrification and EV uptake. This enables consumers to easily electrify without DNSP enforced constraints. For example, Figure 5 shows that for a large network sample studied by Gridsight, nearly 75% of transformers have at least 25% of their capacity remaining. Such areas can support electrification and network expansion without requiring manual intensive analysis. If a substation is reaching capacity, it can be identified and solutions such as load balancing or network switching can be employed to proactively increase capacity. This enables customers to connect additional load without delays or costs associated with capital-based network upgrades.
  • Paving the way for more general per substation Hosting Capacity assessment enabling increased generation and export limits thereby maximising consumer solar installation potential.


Gridsight helps electrical utilities transition to a decentralised grid by generating actionable, AI-powered network insights. These insights enable utilities to dramatically reduce the network augmentation required to safely and efficiently support more residential solar, batteries and electric vehicles. Based on CEO Brendan Banfields PhD research, Gridsight was founded in 2020 to accelerate the transition to renewables.

Brendan Banfield

Co-Founder, CEO

Passion for research and innovation related to integrating renewable energy technologies into distribution networks using data.