7 steps to finding the right pipeline risk tool | Rezatec

Not all geospatial AI is equal: 7 steps to finding the right pipeline risk tool


With water companies coming under pressure to cut losses they are turning to new digital tools that can make leak prevention proactive. But with different products available, the key to success is choosing the right one.

Were any reminder necessary, the unprecedented heatwave that has hammered the USA, Europe and beyond this summer has brought the issue of water conservation into particularly sharp focus. With rivers running down to a trickle, dry reservoirs and hosepipe bans in place, the pressure is piling on water companies to do more to tackle leaks and non-revenue water losses.

At the end of October around a quarter of the USA was in severe to extreme drought status and nearly half was in moderate to extreme drought. Looking at individual states, in Texas, about 27% of the state is under the most severe category of drought and well over half is under “extreme drought” conditions. Florida too is facing a significant drought now, but in 2006 faced a drought lasting 124 weeks.

It’s not just the environmental costs and failing crops associated with wasted water that impact utilities and consumers. There are significant economic implications associated, too. CNN recently reported that in Coalinga, California, new restrictions forced the city to turn to the open market to buy in water. Amid surging water prices, they spent roughly $1.1 million dollars buying in water that would under normal conditions cost around $114,000.

In Italy wasted water runs to around 40 per cent, meaning almost half of all the treated water in the distribution network is simply leaked away. The scale is such that these system losses are estimated to cost Italian consumers around EUR four billion every year. Indeed, Italian water losses are so high that regulatory authorities have imposed new requirements on utilities to ensure that costs can be curbed.

Increased regulatory oversight is just one of the commercial risks water companies face when it comes to tackling leaks. With extreme drought conditions also affecting every continent, water companies the world over are looking for more effective ways to address water losses.

Switching to a proactive approach to water loss
Given the growing demand to improve performance, many water utilities and distribution network companies are searching for more efficient alternatives to conventional leak detection. Instead of a reactive approach in which failures are detected once they occur and crews are dispatched to a repair, they are instead embracing new digital technologies that can help them identify the likelihood of pipeline failure even before a problem has occurred. Often, this means a fundamental change in strategy, switching leak detection and repair to a more proactive methodology in which failures are predicted and prevented ahead of time.

Adopting these kinds of proactive tactics leads to a number of key benefits. They support long-term planning for network upgrades and with it the ability to carefully optimise capital expenditure. This in turn reduces many of the risks associated with investment in the distribution network by allowing a directed focus on priority areas. There are operational benefits, too. Identifying high risk areas means resources that are deployed for inspection, detection and maintenance can also be prioritised, even when operators are faced with a large and complex network. In both cases, data-led decisions can be supported with robust plans that build confidence. Furthermore, by reducing non-revenue water losses, operational expenditures can be curbed without necessarily increasing investment in detection and prevention. Being able to more effectively plan network upgrades also reduces the considerable disruption associated with leaks when they occur and thus the financial impact of such interventions can also be minimised. All these factors inevitably improve profitability. In order to adopt such a strategy, many operators are considering satellite data and AI products in the expectation they can deliver the necessary insights into their networks.

Not all satellite and AI solutions are the same
There are now multiple technology providers offering solutions that use AI by combining many data feeds, some of which include satellite data. Other sources of geospatial data can also be applied to developing mapping or other visualisations that support improved leak detection.

There are also different approaches to developing tools for decision-making using this data. The core objective is to derive data-led outputs that add value. Nonetheless, while some of the tools available use artificial intelligence and advanced analytics, it is vital that the outcomes give a true assessment of network condition and the ability to act on that information.

However, while almost all service providers offer a slightly different approach, unfortunately not all stand up to real world scrutiny or meet expectations in terms of promised performance.

There can be many reasons for these discrepancies and water operators need to understand the three broad categories of product and their mix of data and analytics:

  • Satellite data only
  • AI only
  • Satellite, geospatial data and AI

Satellite data only
Some pure satellite data providers rely on a limited range of frequencies for their observations, using only the L-band data equivalent to UHF and low frequency radar for instance. This neglects regions of the EM spectrum like optical and infrared that can potentially provide valuable additional data and improve accuracy. Others use a wider range of frequencies but don’t provide any substantial analysis. It is that analysis which supports the real-world value from interpretation of the data to deliver effective outcomes.

AI only
Pure AI providers use smarter analysis but rely on limited data sets that don’t provide either a complete picture or the maximum bandwidth, ignoring potentially important characteristics that can be used to derive key conclusions. For instance, some use geospatial data on characteristics associated with soils, terrain and pipeline qualities such as material, age and diameter, but don’t access satellite information that can shed light on ground heave and vegetation vigour – itself an accurate sign for leakage.

Satellite, geospatial data and AI
Providers offering a product that fuses satellite with multiple geospatial data feeds and AI use data on soils, pipeline attributes and slope angle, among others. They take advantage of important additional data opportunities, such as satellite-derived ground motion and information on vegetation vigour, moisture and potential root intrusion. This combination of the diverse data sets, plus the most advanced analytics model, results in greater accuracy: finding up to 84% potential failure in the top 20% of the network.

Choosing the right pipeline risk tool: 7 key questions
Given the wealth of products available offering a wide range of water loss prevention solutions as well as the range of qualities and characteristics they measure, it can be difficult for water companies to navigate through the market and assess products on their merits.
However, a simple step by step guide can help with both product assessment and ensure that the product meets user needs and delivers the desired outcomes.

1. Does it proactively prioritise your network?
Perhaps the most important characteristic when moving to a proactive approach is to be certain that the product will enable proactive assessment of the distribution system. The key here is the use of AI technology to predict failures across the entire network. Secondly, more accurate predictions depend on more data, so multiple data feeds are another vital parameter. So, does the solution use AI technology, does it use multiple data feeds, and does it continue to improve the model over time as more data becomes available to zone in on what matters?

2. Is it just a leak detection tool?
Proactive detection means assessing a network to determine both the risk of leaks and where those leaks will likely occur in future. It is not a better way of locating failures that have already taken place. While some claim to be proactive if they only detect leaks, they are not always. It is important to ask, does it do more than detect leaks? And can I gain new insights into potential failure and address the issue before it takes place?

3. Does it analyse both the likelihood of failure and consequence of failure?
Consideration of leaks also requires an understanding of the consequences of failure if it is to add real value in terms of capital investment and operational expenditure. There is a world of difference between a minor leak in a sparsely populated rural location and a major, disruptive leak on a main city thoroughfare outside a hospital. To gain maximum value from the proactive approach the tool must reveal both the likelihood and consequences of failure.

4. Does it measure ground motion and vegetation (moisture)?
Many solutions use multiple sources of data but not all include the key information that can be a major factor in determining the risk of pipe failure. Ground motion or vegetation growth (a sign of seepage) is a critical factor in failure. This kind of information can only be effectively produced across a network using earth observation satellite data and so it is important to determine that satellite data as well as other geospatial information is used as an input.

5. How accurate is the model?
It seems obvious, but the outcomes of even the most comprehensive data stream can only be of value if the conclusions are accurate. While modelling precision can be difficult to assess, it is possible to eliminate those which make impossible claims. No satellites can record data from beneath ground level, for instance, and so any which claim they can, may be immediately discounted. Look out for claims that can’t work in reality. The key to a successful partner is often found in advanced analytics and the interpretation of the data.

6. Is the vendor credible?
For all their benefits, digital developments also present additional risks as it can be harder to assess quality of such tools at face value. It important to assess the credibility of a vendor and the product by attempting to determine if they have long-term, happy customers who are able to provide a reference. Ask who their customers are and how many have progressed their technology roll out beyond an initial pilot?

7. Is the offering a commercial product?
The digital world is littered with infamous examples in which beta testing takes place after roll out and during commercial customer deployment. Asking if the product has a history of real-world experience, and builds on meeting the challenges and needs of multiple water utilities can avoid future problems. It is also worth asking if the product can be easily scaled to cover everything from the smallest to the largest networks so that it can grow to meet your future needs.

Parameters such as the diameter and age of a pipeline, the material it is made of – ductile iron or polyethylene for instance – geospatial information like the terrain and soil characteristics such as pH and temperature, and satellite-derived environmental data like ground displacement and vegetation vigour can all have a big impact on assessing the likelihood of pipeline failure. Many products which claim to offer an effective pipeline risk service reach their assessments without the use of a full suite of data. Some providers only use satellite data for example, others don’t use satellite data at all.

However, it is only when that full range of data is intelligently allied with an understanding of how those parameters have affected previous failure rates, that vital and accurate insights into the risk of future failures can be provided. Further coupling that knowledge with the consequences of any particular failure delivers a wealth of actionable information. It is information that can rapidly deliver a significant return on investment in terms of both capital and operational expenditure by targeting resources on those areas most at risk of failure and with the greatest consequences were such a failure to occur.

The benefits are clear, what is not always so obvious is how to choose the right pipeline risk monitoring solution that can deliver those advantages. By asking seven simple questions, water utilities can make the right decision first time and realise the benefits of a proactive approach to pipeline failure.


Find out more about the technology at rezatec.com https://www.rezatec.com/solutions/water-utilities/pipeline-risk/ 

See how we can help

Take a look at geospatial AI in action. Explore how satellite data can provide a view of your entire asset base across vast areas. Discover how analytics drill down to a level of detail you never knew existed.