1Spatial’s response to the Leakage Routemap
1Spatial’s response to the Leakage Routemap
There’s so much information to digest in Water UK’s 113 page “A Leakage Routemap to 2050 report” published in March. At 1Spatial, we paid particular attention to the data and geographical elements of the document.
A Global View
The world has a circumference of 40,075kms. The UK has a water pipe network of 347,000kms – enough to go round the world over 8.5 times. That’s a lot of network to manage, maintain and reduce leakage in.
The different approaches taken worldwide fascinated us! How exciting it was to read the approaches taken in different countries to model the world differently. Japan have to consider the geographical phenomena such as earthquakes. The Netherlands have to model a different geography that is typically flat leading to lower operating pressure and therefore less stress on the network. Their pipes are also located either in verges or pavements. And in Germany, factoring in that the soil type and the water chemistry means it is is already favourable towards fewer leaks. Demographic changes of recent years with the Covid pandemic have also altered water companies leakage models. As well as transient and seasonal variations with Ramadan and students winter and summer breaks (and the changing location of the student at those times).
Key Trends
The key trend that we took from the report is in the quantification of background leaks. Infrastructure asset condition and performance can depend on the pipe’s age, its material and the deterioration of fixtures and fittings, or how the pipes are laid. Soil type and environmental factors as well as external pressures from ground movements (such as proximity to transportation networks or weather patterns) all have an effect on background leakage based on the factors identified in the above paragraph as well as history of bursts/repairs and characteristics of the affected pipes.
The report highlights the need for data and asset focused improvement, and key to this is collecting and exploiting more data from the network. This includes smart metering and using sensors and advanced analytics to identify the optimum pressure and identifying pipes most at risk, needing replacement. Vast volumes of data and computing power are needed to support new ways of analysing data to identify trends and inform proactive maintenance of the network. Decades of historical data overlayed with new data from an increasing volume of sensors can be analysed using combined artificial intelligence techniques (machine learning and rules-based approaches) to pinpoint pipes with the highest leak probability.
The key drivers water companies are working towards are their Public Interest Commitment (PIC) in 2019 to triple the pace of leakage reduction during 2020 – 2030 (from 1990 – 2020) saving a billion litres of water / day, and the national infrastructure challenge (NIC) to halve leakage by 2050 from 2018 levels.
1Spatials Approach to Leakage
1Spatial’s SaaS Leakage Solution combines both advanced spatial and non-spatial rules-based logic and machine learning (ML) techniques to offer enhanced leakage location prediction capabilities. It has been designed in collaboration with UK utility SMEs to pinpoint pipes with high leakage probabilities, focusing where incident detection teams should start investigations. We produce a "heat map", directing the most effective resources more efficiently, and prioritising the finding and fixing of the biggest leaks. Our solution calculates leakage propensity scores for each of the pipe assets within the processed area. This propensity score helps identify “high-risk” assets allowing leakage detection engineers to prioritise their search locations. It can also be utilised for preventative repair and maintenance programmes.
Our solution provides the following data insights that have been derived from the model:
- A heat map classified into High, Medium and Low burst propensity.
- Predictive Leak Detection for each pipe in the network in terms of burst probability.
Additional insights that can also be derived including:
- Factors that are influencing the leakage breakouts.
- Pipe replacement and risk analysis.
- Asset / Sensors location planning.
Reducing leakage can be helped by using data to identify the optimum pressure in the network, trying to predict assets most at risk based on the asset’s infrastructure, reducing the leakage awareness using tools such as sensors and trusting the data quality to make informed decisions around preventative management and leakage repair.
We look forward to innovating and collaborating with the water industry in understanding how 1Spatial can help organisations to identify areas of background leakage using data to meet the PIC and NIC targets.