Water Benefits From Predictive Maintenance
EP Editorial Staff | August 8, 2024
Predictive maintenance is cutting maintenance costs and improving reliability in water-treatment facilities.
By Krishna Prashanth, ABB
Thirty years ago, the United Nations warned that, unless drastic action was taken, the world would face a water deficit of 40% by 2030. With a world response that has fallen short of targets, water scarcity is now an actual, rather than a theoretical, problem: 2.2 billion people are estimated to lack access to safely treated water.
As the Earth’s population swells from 7.7 billion in 2023 to an expected 9.7 billion by 2050, creeping urbanization and growing industrialization will mean that demand for fresh water will surge by 20% to 30%, making water scarcity even more acute.
Meanwhile, 80% of wastewater (171.3 billion m3/yr.) is lost into the wider environment without being recovered or reused. Pollution, contamination, and ageing infrastructure compound the problem. While water-utility budgets are limited, the pressure is on to find new ways to effectively manage water. Predictive maintenance is emerging as a keystone solution in the fight to address the problem.
Predictive maintenance in water utilities is a proactive approach that enables operators to anticipate equipment failure and infrastructure breakdown before it occurs. Analytics and machine learning algorithms are used to analyze real-time data from sensors installed in water-treatment plants, pipelines, and distribution systems. These help identify patterns, trends, and anomalies that could indicate potential issues or failures, providing a cue for corrective action before the equipment malfunctions or fails completely.
Using this approach, water utilities can minimize downtime, cut operational costs associated with reactive and scheduled maintenance, and ensure sustainable performance. Furthermore, predictive maintenance can help utilities lighten their environmental impact.
In the past, the water industry has generally reserved close analysis and detailed risk assessment for critical operations, effluent discharge monitoring, and safety elements – disciplines that apply across all industries. For the majority of other operations, the mantras ‘fit and forget’ and ‘run to fail’ were considered to be the standard approach to industrial efficiency. When one takes human and animal safety into consideration and at a time where every drop counts with crops under threat and drinking water is at a premium, this approach is no longer feasible. As a result, scrutiny is being applied across all operations throughout the water treatment and distribution process, notably leakage, waste, infrastructure aging, and the integrity of distribution networks.
Modeling risks using data
Precursors of failure in water networks can be described in several ways: devices that have aging characteristics, such as the build-up of material on the electrodes of a magnetic flow meter which reduces connectivity or the reduced light that an aging LED screen emits. Then there are the random failures, that prompt the question, Are they genuinely random?
Predictive maintenance begins as operators consider the signs of deterioration, aging, and decay, building up a library of failure signals. These signals may depend on a single parameter or on multiple related parameters, although it is worth noting that the precise interaction of these parameters may be unclear. The greater the volume and quality of data from multiple diverse parameters, the more precisely the interaction of those factors can be understood and the more accurately models can assess the likelihood of some kind of malfunction or breakdown.
Adoption and implementation of predictive-maintenance strategies requires buy-in by operators who need to assess available data sources, infrastructure, and staff readiness.
New analytics capabilities
Predictive maintenance would not be possible without the technologies and techniques that are driving the move toward Industry 4.0 and the IIIoT. Sensors and transmitters that use Wi-Fi, Bluetooth, 4G or 5G and Narrowband Internet of Things (NB-IoT – a low-power wide-area radio technology specifically designed to connect remote spaces such as underground car parks and basements to the internet) are not only opening up new realms of data that can feed into risk models, but also providing new analytics capabilities. The combination of new diagnostic features and advances in digital technologies, including sensing and communications, and analysis using edge and cloud computing, mean that operators can identify antecedents of device malfunction, component failure, and infrastructure breakdown far earlier than ever before, and respond with timely interventions that cause minimal disruption and downtime.
These expanded capabilities are enabling water companies to move from preventive maintenance strategies, where all devices were maintained to a schedule whether they needed maintaining or not, to strategies based on predicted changes in device behavior which are more finely attuned to the real-world situation in the field and are less expensive and disruptive to implement.
Predictive challenges
That’s not to say, however, that predictive maintenance is an instant panacea for water conservation and future supply guarantees. Adoption and implementation of predictive-maintenance strategies requires buy-in by operators who need to assess available data sources, infrastructure, and staff readiness. They need to select the right equipment and sensors; and they need to call upon expertise in data analytics and machine learning to build predictive models and algorithms.
While new techniques for implementing predictive maintenance for water treatment and supply planning are becoming mainstream, they nevertheless require a degree of expertise to evaluate the best solution for a specific end use. Many companies lack the in-house resources to take such decisions internally.
Questions of data
By far the greatest challenge to the implementation of predictive maintenance are the volumes of data that are required for such strategies to work effectively. While water utilities often hold considerable banks of data, the data is sometimes incomplete or inaccurate, and can be provided in formats that are unusable or inaccessible. Even clean and accessible data needs to be shareable if it is to realize its full value.
Unfortunately, the remote nature of water-treatment plants and the traditional methods of collecting and handling data have often led to data silos within the water industry. For predictive maintenance to be viable, key questions, such as Where and how do you store and process the data?, For how long?, At what cost? and, Who owns the data?, must be addressed to step outside those silos and bring all relevant elements together.
There are many reasons why operators may choose to pause and give the matter careful thought before leaping in. The investment in time, money and expertise required to implement predictive maintenance may, for example, be an immediate stumbling block if it is beyond the resources of some water utilities. In addition, a few operators, resistant to change, may prefer traditional methods. Even among operators who have decades of experience in water-systems management, the application of IIoT technologies and techniques to legacy practices can be daunting.
Data-gathering solutions
Thankfully many suppliers of data-gathering technologies offer solutions that make this transition as painless as possible. One solution is the development of asset-management systems that link operational and engineering data to create an environment in which data can be collected and shared easily throughout an organization. This approach opens opportunities for fleet-wide predictive maintenance, where data from numerous devices of different makes across multiple locations can be collected and analyzed to identify where intervention is required.
Predictive maintenance is becoming an essential tool for water utilities striving to improve their asset reliability while avoiding service disruption. As technology advances, predictive maintenance is likely to become more effective, more accessible, and more readily accepted, while its benefits will be more widely understood. It may not be the silver bullet that solves the world’s water problems, but it will save money, increase efficiency, and, most important, conserve and perhaps expand the availability of our most precious resource, thus contributing toward the UN’s goals of sustainable development.
Krishna Prashanth is Global Product Line Manager, Electromagnetic Flowmeters and Segment Leader for Water and Wastewater, BL Instrumentation, and ABB Measurement and Analytics. He is on a journey to design a world for sustainable living and is passionate about water. Prashanth is based in the UK with the ABB Stonehouse factory.
View Comments