Water has been referred to as the oil of the 21st century-- and, for reasons good and bad, that may prove to be a very apt metaphor. Water has always been a major driver of growth, and this is increasingly the case. But it is the scarcity of water that is coming to define major changes in economic activity.
That’s where a company like Olea Edge Analytics comes in. This week they announced they had raised an additional $35 million in Series C funding to accelerate their development of artificial intelligence-based solutions for water management. Olea works with water utilities to help them better deliver water to homes and businesses, in spite of growing demand and aging infrastructure. Olea’s technology utilizes sensors to track critical water resources and provide a level of visibility to water management that empowers them to make smarter decisions about supply, billing, and conservation.
Water delivery systems are very complex and there are many opportunities for errors that could lead to under-billing, or over-supplying and subsequently cause lost revenue and/or wasted water. It is easy to see how machine learning could serve a tremendously useful role here. But to do so, there would need to be sufficient data inputs. When it comes to water utilities, the available data definitely leaves something to be desired. In many cases, the data being collected on water systems is limited and periodic. For example, utilities may receive only periodic meter readings, and receive pressure and other readings from only some of the many tens-of-thousands of pipes and other infrastructure that makeup the water system.
Olea helps address this problem already, in part, through the use of sensors. By applying sensors at critical points throughout the water system, they significantly increase the amount of continuous available data streams. This critical and voluminous data can be used to train a model to detect anomalies and/or to predict when potential failures may occur. This works by teaching the computer to understand what typical or healthy operation looks like, so it can spot potential deviations from those operations. A model could also be trained on historic data and incidents to find patterns in usage, flow, pressure and other variables that indicate a problem is likely to occur-- and, perhaps even more importantly, identify potential problem spots within the system.
This kind of approach can be significantly more cost effective and faster than non-AI based solutions. Water systems can be vast, and while collecting more data on their performance can be very helpful, digesting all of that data rapidly to spot subtle patterns that may point to problems is incredibly difficult for humans to manage on their own. Additionally, merely programming computers to alert human operators of specific red flags would mean having to pre-conceive every potential anomalous activity and collection of activities that might give rise to a problem. Machine learning provides a way to monitor a near-endless number of possible problems at scale, allowing human resources to be targeted to where they are most needed.
This is especially important in the infrastructure arena, because the cost of addressing a potential problem can be huge-- and the savings/revenue from addressing true problems can be, as well. Without smart solutions, water utilities would often either need to forgo major amounts of lost revenue because of faulty systems, or spend millions of dollars on inspections and repairs of major sections of their infrastructure that are not actually targeted to where true problems exist. If you’re going to tear up a sidewalk or disrupt water delivery to a city block, you want a high degree of confidence that the approach is as surgical as possible. Machine learning can help pinpoint those problems in a way that offers that confidence, and with this new round of funding, Olea is likely to continue developing technology that helps cities and utilities make smarter decisions about this incredibly important resource.