Anna Whitmore (Mott MacDonald), Kenny Williamson (Watercare Services Ltd.), Glenys Rule (Watercare Services Ltd.), Owen Flanagan (Mott MacDonald), Evan Harwin (Mott MacDonald)
Poor quality plant operating data is a common issue for water utilities and limits operators’ ability to understand and improve plant performance. Emerging technology such as soft sensors and digital twins can improve confidence by supplementing and enhancing operating plant data.
As well as enabling operators to better understand plant performance, digital twins and soft sensors can also assist with the development of control strategies to reduce operational expenditure and greenhouse gas (GHG) emissions. Watercare forecast a greater than $50M spend on wastewater asset operating costs for FY2022. The adoption of digital technology therefore has the potential to realise significant savings.
The Rosedale Wastewater Treatment Plant (WWTP) consists of primary sedimentation, biological treatment through the Modified Ludzack-Ettinger (MLE) process, clarification, UV disinfection, and anaerobic digestion for solids stabilisation.
A digital twin of Rosedale WWTP was developed in 2020 to provide real-time insights into plant performance and scenario testing, to support commissioning of the fourth MLE reactor. The solution combines biological modelling, machine learning, predictive rainfall and scenario analysis. This paper describes the development methodology and functionality, including the data, technology and analytics that were used.
Watercare’s 2022 Rosedale Innovation programme comprises two workstreams which build on the existing digital twin solution; soft sensors for instrument failure detection and opex and greenhouse gas baselining and prediction.
The soft sensors workstream is focused on the development of soft sensors to proactively identify instrumentation drift and failure events. The soft sensors use a combination of machine learning and biological models to predict instrumentation drift and alert operators to irregular operation. This paper describes the development methodology including data exploration and modelling techniques.
The objective of the opex and GHG workstream is to develop real-time operational insights to enable the baselining, monitoring, and control of GHG and opex. This paper outlines the different data sources and calculations and describes how they were codified and surfaced through the digital twin. This workstream also includesanalysing process emissions. This involved installing nitrous oxide (N2O) monitors on MLE4 and calibrating the digital twin to act as a soft sensor for N2O emissions. The challenge of unreliable data and instrumentation failure is not unique to Rosedale WWTP. The application of emerging digital technology can enhance operational plant data through the creation of new data sources. This data provides new insights into plant operations and enables operators to consider operational strategies to reduce opex and emissions. This paper addresses how the Rosedale technology can be scaled to other treatment works to improve data quality, optimise opex and GHG emissions and improve instrumentation reliability.