Since the creation of DMA leakage calculations, analysts have needed a method of understanding whether the numbers they are calculating are reliable.
This is typically undertaken in the UK using a Per Capita Consumption (PCC) test, where daily leakage is calculated using a minimum night flow water balance (Minimum night flow – night use allowances), and validated against the total daily flow into the area. This validation results in an estimate of the total volume of water being used by household customers, which can be compared against an expected range of individual customer consumption.
When areas have a low number of household properties or an erratic flow profile, this method can be unpredictable so a simpler daily balance approach is often used instead.
At the start of AMP7 the use of DMA operability tests was included in Ofwat’s reporting guidance for leakage, and expectations set that water companies should be achieving an average of 90% operability across all DMAs to give confidence to their reported leakage numbers. As we move into AMP8 this continues to be a requirement on UK water companies for reported leakage.
PCC estimates can vary considerably between individual DMAs, based on factors such as property type, affluence and occupancy, as well as increasing during hot and dry weather. This means that water companies have to use quite wide ranges of allowable PCC to achieve an acceptable level of DMA operability. Typically, in the UK this can include ranges between 50 and 500 litres/person/day to absorb the ‘noise’ of variations at DMA level.
With such a wide range of PCC used for operability assessments, and with analysts treating the outputs in a very binary way (i.e. operable or non-operable), it is difficult to identify what an expected PCC for a DMA should be. This makes it difficult to understand when this starts to drift, suggesting a problem with the integrity of the area.
In the example shown below, Paradigm was used to analyse a DMA which had high levels of unaccounted for water. Investigations, based on the three days on the left of the graph, identified a logger scaling error, which was resolved to give a much closer match between the actual net flow and the modelled profile (shown on the right of the graph).
Even though there was a scaling factor error apparent on the graph, the DMA would still have been operable both before and after this issue was resolved, with PCC values of 245 l/d and 146 l/d falling comfortably within the 50 to 500 l/prop/d range.
With DMA operability being the first indicator of whether calculated leakage can be trusted for an area, this could easily have been taken as a green light for leakage targeting in this DMA, resulting in valuable targeting resources being deployed into an area where they couldn’t deliver any benefit.
Analysis using Paradigm suggests that up to a third of DMAs in UK Water companies could have underlying problems masked by poor operability tests. This could form a significant barrier for water companies looking to further drive down leakage on their water networks over the coming years.
Water companies could explore how to use operability numbers more effectively, but now that operability has been made part of the reporting requirements for water companies, any attempt to make the range of acceptable PCC values tighter to capture DMAs with anomalies would likely result in a reduction in overall operability. There would be a risk of being perceived, at least from a reporting perspective, as demonstrating a reduction in performance regardless of any improvement in approach that resulted, forming a regulatory risk to the water company.
The technical barriers to adjusting operability ranges on a DMA by DMA basis are also considerable. This would require additional calculation methods in leakage reporting software and an approach to estimating expected PCC at DMA level based on property characteristics and historical performance. Systems would also need the ability to adjust on a seasonal basis to account for hot weather, which may be difficult for analysts to fully interpret and understand, further reducing overall confidence in this measure.
DMA operability has secured its place in leakage reporting, now forming part of our regulatory requirements and giving an understanding of overall confidence in water companies’ night use allowances and data availability. However, we should be mindful when targeting individual DMAs that operability alone can’t give full confidence in either the data available, or the resulting calculated DMA leakage.
With the availability of additional data sets in the UK, such as customer smart meter readings, small area monitors and open-source census data, water companies are well placed to create 15 minute or hourly profiles of expected flows for DMAs, allowing an analysis of flow data using the full shape of the net flow. By not just relying on minimum night flow and total daily flow, these full profiles allow analysts to understand what should be happening in DMAs and properly interpret what is going on, promoting a culture of understanding issues not just of trusting binary calculated metrics.
At SME Water our pioneering Paradigm Club Project has been helping water companies to do just that, creating modelled profiles for DMAs and improving analyst understanding of how they should be interpreting flow data.
Interested in hearing more? Reach out to one of our team by contacting us today.
