This project intends to develop an integrated, real-time software solution, useful for a large class of water management utilities that is flexible enough to integrate other utilities‘ operational tools as required at specific locations.


The EFFINET project will develop innovative, energy-aware water network control algorithms based on MPC techniques for the operational management of urban water networks, controlling pumping and valve actuation in real-time to cope with varying energy-prices and water demand in an economically profitable and risk-averse way. The decision support software module developed within EFFINET will be based on stochastic model predictive control (MPC) techniques, taking into account the information coming from the monitoring and demand management modules to determine the current state of the network and to derive stochastic models of water demand. Such information is integrated with stochastic models of energy prices and with dynamical models of water flows and storage, operating constraints on pumps, valves, and on water quality, and the economic and risk figures to be optimized.

Dynamic and stochastic models are combined to create scenario trees describing the possible evolutions of the water network, customer demand, and energy prices, and used to formulate a stochastic optimal control problem that can be solved by parallelizable convex optimization algorithms. Therefore, the proposed algorithms not only will account for uncertainty, but it will actually interact and learn from its environment as well as adapt its decisions to changing exogenous conditions. Such a predictive capability is an important distinguishing feature of EFFINET, as predicting the future explicitly and reliably in the control algorithm is a good way to anticipate countermoves and to therefore avoid myopic actions leading to management inefficiency.


In this project, intelligent data processing methodologies and decision support algorithms will be developed for:

  • Quantitative network monitoring, with special emphasis in leak detection and isolation, combining direct (simulation) and inverse (parameter estimation) approaches.
  • Qualitative network monitoring, for detecting changes in several quality indicators, to be identified in the first phase of the study, for locating the area where the abnormal quality indices are, in real time. Those methods will also combine direct and inverse approaches, and improve the existing methods.
  • Integrating quantitative and qualitative approaches with nodal demand forecasting approaches, to estimate the demand in each network node (that typically is not measured) as well as with DMA high level monitoring indices and sensor data validation/reconstruction schemes.
  • Optimally placing sensors, for quantitative and qualitative network monitoring that enhances the performances regarding detecting and locating leaks and abnormal quality behaviors, respectively.
  • Sensor data validation/reconstruction to guarantee that the sensor measurements used for the network monitoring are free from outliers or sensor faults, which may otherwise affect dramatically the fault detection and isolation algorithms.
  • The network monitoring module of the EFFINET software will include an innovative fault detection framework with learning algorithms for approximating key correlations between measured variables during operations. This will allow handling highly unstructured environments, which goes beyond the current state-of-the-art in this area.


Recent advances in smart metering allow collecting timely information about water individual customers’ consumption. The possibility to get information using AMR facilitates the interaction with consumers and the use of existing instruments. The main innovation provided by EFFINET is related to the following aspects:

  • As real-time information about consumption becomes available, its use in demand forecasting will improve the total efficiency of service. The information will also allow designing new tariffs models based on actual consumption.
  • The information about real-time consumption will facilitate the definition of different profiles of consumers, that will be useful for leak detection.
  • Most importantly, the project will design and develop an application for communication of consumption patterns to consumers, in order to influence a change in demand and therefore water savings.
  • Give information to the consumers about their water consumption for decreasing total volume of water consumed. It is possible, using AMR, to have information about consumption in every household several times during the day. For this reason, the inclusion of the total amount of water consumption in the bill and comparing it with the average consumption in the city is the informative goal.
  • Use of smart phone application to involve consumers to give feedback, relies on using interactive applications not only to give consumption information, but also to get the feedback for the consumers, such as surveys or designing incentives (e.g. giving some specific merchandising) for having their opinion. In case of using different tariffs only will be implemented in some specific households receiving compensation for their contribution.