Despite the global trend around the digitalization of processes and the implementation of technology from the Industry 4.0, the Information and Communication Technologies (ICT) solutions and classical automation approaches using on-off and PI/PID controllers are still the most common ones in the wastewater treatment sector. However, there are other multivariable control solutions, like Model Predictive Control (MPC), that have shown excellent results in other industries, unlocking process optimization when considering multiple inputs and constraints. Moreover, the use of Digital Twins (DT) has received an increasing attention in the last years.

The main objective of the HADES (Herramienta de Apoyo a la Decisión para optimización de la operación de EDARs) project is the demonstration at lab-scale of a Decision Support System (HADES DSS) for Wastewater Treatment Plants (WWTP), able to use data in near-real time, predictions and calibrated models -working as DT- for optimizing the operation according to multiple criteria and to protect it from undesirable events, and that could be used remotely.

An initial phase will be carried out based on the gathering of information from a WWTP, including design and construction (volume of the reactor, geometry, number of diffusers, etc.), historic routine, sensoring and analytical campaign data. The information will be used in later phases for the modelling of the WWTP DT (using common commercial and specialized platforms as GPS-X and ANSYS Fluent) and two Artificial Intelligence (AI) tools (a knowledge-based system, KBS and a case-based reasoning system, CBRS), which will provide reasoning and self-learning capabilities to the HADES DSS. In the next phase, the HADES DSS will be complemented with an optimizer system, using Model Predictive Control (MPC), that will be used to find the optimum operation strategy. This risk of undesired side effects of the strategy will be assessed with a heuristic model, and the actuation plan will be compared with historical similar cases that will include previous success/failure information to avoid the systematic repetition of mistakes and provide an automatic learning tool. Finally, before the demonstration of the HADES DSS, it will be necessary to develop the overall software that controls every subsystem of the solution.

The successful demonstration of the HADES DSS will allow the later development of tools based on this solution, which will suggest the WWTP managers the optimum process strategy using almost-real time data from the plant and will provide additional information as the DT contained in the tool can be used as virtual macro-sensors. 

The HADES DSS can be implemented in both existing and new WWTPs. This DSS could provide ACCIONA with a competitive advantage within the wastewater treatment sector and will help to consolidate its position as a leader in the sector.