Intelligent Water Processes (Invapro)

The water treatment sector supplies drinking water to the population and ensures sustainable circulation of wastewater, and is thus crucial for society. The sector is facing several and significant challenges related to increased requirements for quality, safety and the environment. In particular, the capacity of existing facilities are challenged by increasing influent due to increasing population, demographic changes and climate change. Furthermore, many facilities are old, in need of maintenance and new constructions. The INVAPRO project consists of the public sector partners Veas, Trondheim municipality, and Bergen Vann KF, and research partner SINTEF Digital. The objective of INVAPRO is to meet the increasing needs and challenges by moving the Norwegian water treatment sector towards more optimal, efficient and environmentally friendly operation. The project will exploit machine learning (ML) and available operational data for increased process knowledge and control. In particular, today's process control is mainly based on linear assumptions, while in reality processes is non-linear and complex. To increase accuracy and autonomy, Invapro will develop data-driven algorithms and non-linear control for daily operations of water treatment processes. The effects of these will be experimentally verified in live deployments at the partners' facilities. In addition, Invapro aims to facilitate innovation within both the water utilities and the supplier industries by developing not only technologies, but also enable new services, business models and ways of collaboration. Status per 30.09.2020: Work has begun on the process description for the water treatment section at the Veas plant. This work involves describing the connections in the process, identifying and retrieving available data, conducting data analysis, finding connections and identifying potential utility in discoveries. Based on this, proposals for real experiments in the facility ("use-case") will be prepared. A "use-case" can be part of the process which, for example, is particularly challenging to manage optimally and / or where unnecessary use of input factors is particularly cost-driving. The same analyses will be performed at drinking water treatment plants in Bergen and wastewater treatment plants in Trondheim. This work will start in 2021. Status as of 30 Nov 2021: "Use-case" for Veas has been identified. Data has been collected and data analysis has started. Training in machine learning has been completed at Veas. An initial meeting has been held to establish a process description to identify the "use-case" at the drinking water treatment plant in Bergen. Similar work for Trondheim will start in 2022

Water treatment is crucial for a well-functioning and sustainable society. In spite of its importance, the water treatment sector is facing several and significant challenges related to increased requirements for quality, safety and the environment combined with increasing influent and large investment needs for existing facilities. Thus, there is a need for both knowledge-building and innovation to ensure more efficient and optimal operation so 1) the increasing demands can be met, 2) the existing facilities can be exploited to a maximal extent so investments and maintenance can be postponed and/or dispersed over a longer period of time, and 3) the use of chemicals and effluent discharge are minimized. The objective of Invapro is to move the Norwegian water treatment sector towards more optimal, efficient and environmentally friendly operation by exploiting machine learning (ML) and available operational data for increased process knowledge and control of these highly complex and non-linear processes. Previous activities among the project partners confirm that there is a large potential for knowledge and value creation in applying these methods to water treatment processes. Critical R&D challenges in Invapro include integration of existing domain knowledge with data-driven methods to ensure safe, robust and explainable application of data-driven models, and evaluation of benefits and business opportunities enabled by this technology. In particular, Invapro will establish methods, analyses and concrete measures for improved operations of water treatment processes that are of interest both to utility owners and the supplier industry. The project results are expected to improve the competence level in the field among the public sector partners and enhance operation for Norwegian water treatment industry through increased autonomy and knowledge.