Online data processing for proactive UK water distribution network operation

Abstract

Abstract. Operational benefits and efficiencies generated using prevalent water industry methods and techniques are becoming more difficult to achieve; as demonstrated by English and Welsh water companies' static position with regards the economic level of leakage. Water companies are often unaware of network incidents such as burst pipes or low pressure events until they are reported by customers; and therefore use reactive strategies to manage the effects of these events. It is apparent that new approaches need to be identified and applied to promote proactive network management if potential operational productivity and standards of service improvements are to be realised. This paper describes how measured flow and pressure data from instrumentation deployed in a UK water distribution network was automatically gathered, checked, analysed and presented using recently developed techniques to generate apposite information about network performance. The work demonstrated that these technologies can provide early warning, and hence additional time to that previously available, thereby creating opportunity to proactively manage a network; for example to minimise the negative impact on standards of customer service caused by unplanned events such as burst pipes.Each method, applied individually, demonstrated improvement on current industry processes. Combined application resulted in further improvements; including quicker and more localised burst main location. Future possibilities are explored, from which a vision of seamless integration between such technologies emerges to enable proactive management of distribution network events.

A semi-automated approach to validation and error diagnostics of water network data

“…Type 1 contains data with no relevance for the operation of the system, because the integrity of the data is damaged or missing. This type of anomaly has also been described as 'dirty data' (Mounce, Boxall, and Machell 2010;Machell et al 2014).…”

“…This is of growing importance, as utilities are drawing inferences at an increasing rate from data-driven applications and the collected data itself as sensors and data transmission become cheaper. Combining all data sources can be used proactively to improve network operations in a utility (Machell et al 2014). Andrews et al (2017) found that it is common for many utilities to struggle with the validation of their own data.…”

Urban Hydroinformatics: Past, Present and Future

“…To make sense of this increasing amount of information, research and practice have made significant progress towards better analytics, including but not limited to those: (i) Capable of extracting valuable information from the data (from smart alerts to customized advice for water users); (ii) performing better stochastic simulations to improve the ability to produce longer timeseries (based on observations) for long-term scenario development and stress-testing; (iii) performing advanced optimisation to identify better solutions in this information richer environment; and (iv) providing novel ways of visualizing and understanding the decision tradeoffs within complex decision spaces. Examples of these new analytics, include AI/ML analytics for proactive management of water distribution systems (including burst detection) demonstrated in UK case studies [40,41], asset deterioration assessment [42], as well as the use of deep learning techniques for defining novel control strategies that are more robust against cyber-attacks of water distribution systems [43]. Examples also include recent work on using smart meter readings to parametrise residential water demand models [44] as well as the methods and tools developed to investigate the properties of these timeseries at fine timescales [29].…”

Flowrate Time Series Processing in Engineering Tools for Water Distribution Networks

“…Regardless of the acquisition and transmission settings, the obtained raw flowrate time series may contain measurement errors, such as missing, repetitive or even false readings (Mounce et al., 2010; Xenochristou et al., 2020). These errors can have their origin in sensor or logger malfunctioning, faulty transmission system due to battery failure, inadequate acquisition range (e.g., above or below meter range or bidirectional flow) and data storage limitations (Loureiro et al., 2016; Machell et al., 2014; Xu et al., 2020). These measurement errors, typically known as outliers or anomalous values (Kirstein et al., 2019), should be detected and corrected before they can be used in engineering applications (e.g., hydraulic modeling, calibration, leak detection) (Romano et al., 2014).…”