The Pilot site 3: Segrate – Hospital and research campus, as use case of the RENergetic project, will see the installation of machine learning software for big data mining to maximize the level of energy autarky of a local energy system (Energy Island). The Energy Island of Pilot 3 includes the IRCCS Ospedale San Raffaele (OSR), the OSR Cogeneration Plant and Segrate municipality (Milano 2). At a smaller scale, the pilot will initially involve the Ospedale San Raffaele energy ecosystem to scale, in a second phase, all actions at the Segrate municipality level.
The IRCCS Ospedale San Raffaele is a university hospital and scientific institute that covers a total area of more than 290,000 square meters, eleven buildings hosting the hospital departments, clinical laboratories, research laboratories, the university, student residences, and the underground shopping centre. The OSR Cogeneration Plant in Vimodrone supplies hot water, superheated water, freezing water and electricity to IRCCS. Hot water is provided also to Segrate municipality.
The researchers of the Center for Advanced Technology In Health And Well-Being (link here), in collaboration with the Energy Manager and the Energy Technicians of both the hospital and the Cogenerator plant, run an energy audit to define parameters and measures on energy demands, estimate consumption, basic needs and requirements to securing and maximizing the level of energy autarky and reducing energy waste in the Hospital area. Currently, the share of electricity supplied to each single building is not directly monitored, but it is possible to know the thermal and cooling energy supplied to the buildings attested to each substation. For this reason, and as a first step, the researchers planned to install new necessary sensors.
Specific buildings have been selected to monitor consumption in the island based on two factors: the higher quantity of electricity and thermal energy requests, and the greater variation of consumption registered in time, excluding the hospital and clinic areas that need fixed supplies of 24 hours a day. The result of these preliminary investigations allowed researchers to identify the buildings involved in the experiment: the University, the administrative offices, and research laboratories; draw an energy map and a list of meters to implement in these areas to measure all necessary data.
Meanwhile the AI algorithm have been drafted and applied to the historical data processing and procurement, this allowed the data scientist of the Research Center, Daniele Baranzini, to understand what factors can help to improve the prediction of gap between supply and demand, and to identify the root causes of wasted energy. After few simulations with different variables, the algorithm provided an interesting result that could grant a conspicuous energy saving based on the fine regulation of the hot water flow.
The field verification will provide realistic data about the production plant for the management and forecasting of demand peaks, and will allow the energy managers to schedule the production on the basis of economic convenience and energy efficiency, and consequently to program the production priority, the use, purchase and sale of energy carriers, according to the cost of energy carriers and incentives related to production.