Water utilities are at a difficult juncture in business. In the UK, they have invested over $170 billion in the last10 years in a bid to maintain and improve assets. In the US, capital expenditure on clean and waste water infrastructure will exceed $683billion over the next decade. These costs are unsustainable. To add to the cost challenges, traditional solutions have begun to fall short of the expectations.
The situation calls for the adoption of next generation solutions that go above and beyond traditional solutions.
Situation on the ground
The problem that water utilities find most frustrating is with their ageing below-ground assets. There is a lack of visibility into their performance due to opaque operating conditions. Failure to fix problems related to assets can result in non-compliance and hefty penalties.
In addition, water utilities are under tremendous pressure from growing customer expectations. Customers are used to services like Amazon and Uber, where everything happens in real time and they have complete control over the service. Utilities have to match the experience.
Finally, the unforeseen impact of climate shifts and the ineffective management of a distributed field workforce are sending costs spiraling. These are monstrous challenges. But they also present the perfect opportunity to examine technological advances and solutions that can transform the face of water utilities.
Some water utilities have already made remarkable progress in their efforts to leverage technology. For example, the water utility inJapan’s Fukuoka city uses an unsupervised class of Artificial Neural Networks (ANN) named self-organizing maps (SOM). Sensors attached to distribution pipes are continuously monitored, helping regulate pressure through remote motor valve operations. These sensors also help reduce the dependence on operators with experience in managing distribution. Utilities like the one in Fukuoka are discovering just how useful these systems can be in the event of leaks and damage to pipes that require immediate and precise responses. The systems also help determine precise flow to consumers based on real-time demand.
An US-based water company replaced primary in-pipe inspection method that uses manual CCTV analysis with artificial intelligence (AI) and neural networking to monitor below-ground assets. The water utility leveraged open source neural networks and a code framework to hone its AI and Machine Learning (ML) applications in the field of pipe inspection for automated cracks, structural deformations, blockages, debris etc. This enabled proactive maintenance, wherein more pipes could be inspected in a shorter span of time, maximizing budgeting allocation, and ensuring overall system stability.
The number of utilities that deploy technologies like ML and AI, to forge a path into a real-time insights-driven future, are bound to grow.
Traditional systems have major gaps in addressing asset management KPIs. Fortunately, the proliferation of sensors, real-time monitoring devices and video/image feeds is changing this. Advanced asset-centric technologies, aimed at generating water insights, are taking the guesswork out of leak detection, pressure management, work order distribution, etc. Technologies like the Internet of Things (IoT), ML, AI, Deep Learning, data and analytics make a difference by targeting asset management KPIs.
The question before smart water utilities is: Should they invest the next few million dollars in identifying and servicing leaks after they have impacted consumers or in insights and predictive systems that help minimize leaks, flooding and incidents before they impact customers? The answer is self-evident.
Smart utility built on the right technology
A predictive system/insights tool, specifically built for water utilities, will allow utilities to stay one turn ahead of issues at each stage of clean and waste water supply, distribution and management. The tool will use data leveraged using extraction and analytical processes, algorithms, models and best practices that are specific to water utilities. The insights, decisions and recommendations will be provided to business users and applications across a variety of devices. (See Figure 1 for the reference architecture of the insights tool)