San Raffaele Hospital

Segrate Neighborhood

Description of the pilot site

The pilot consists of two entities: 

  • HSR: San Raffaele Hospital (HSR) in Milano (Italy) is a world renowned highly specialized multi-disciplinary medical centre with more than 50 clinical specialties with a yearly 51,000 patient admissions, 72,000 emergency room encounters and delivery over 1.5 million outpatient medical services;
  • Segrate: is a Municipality in Milan metropolitan area, close to Milan on East. It is 17,49 sqkm wide, it has 35.234 inhabitants with a density about 2.000 inhabitants per sqkm. It developed mainly as residential settlement with some tertiary excellences: Mondadori (designed by Niemeyer), Fininvest, IBM, Microsoft (now moved).

 

HSR and SEGRATE buildings are powered by the HSR Cogenerator plant.

  • The total multi-vector energy source (heat, cooling and electricity) for HSR and Segrate (heat for Milano 2) is supplied 100% by the HSR owned Cogenerator plant located proximal and external to the HSR premises as shown in the Figure.
Co-generator capacity palnt and multi-vector energy sourcing in MWh/yr

Pilot's goals within the RENergetic project

The overall Pilot 3 focuses on testable and viable solutions searching for both AI enhanced energy island efficiency, energy waste containment and autarky equipped with replicable social/organisational capacity to fit real municipality contexts.

Main goals:

  • Full scale experimentations on dedicated scenarios of implementation, called Epics (see below);
  • Right-fitting EI energy-demand – response cycles from HSR to scaled-up SEGRATE solutions;
  • Increase HSR/SEGRATE heat and electricity energy efficiency methods;
  • Creation of a tool to support the work of technical managers and energy decisions;
  • Increase Project adoption levels via social and organisational awareness campaigns;
  • Optimise energy demand by deployment of AI tools and Machine Learning methods to forecasting intelligence, anomaly detection, classification and predictive/prescriptive models to augment precision energy (Figure below).
Some application for AI-based energy management in Pilot 3

Epics to be implemented within the RENergetic project

  • Heat Supply Optimization: the concept for decreasing local CO Emissions as one objective of RENergetic via increasing energy efficiency here is to be used as a combination of forecasting heat supply (from the local CHP plant) and forecasting heat demand on building and even on room level using AI prediction modelling. Thus, increased reliability, precision and granularity of prediction in a second step is to be turned into recommendations to HSR energy management how (when and to which degree) to reduce heat provided from the power plant to the HSR buildings. To this end, building (room) demands will be aggregated and fed back to plant management enabling the reduction of heat production at certain times and thus reducing CO2 emissions accordingly;
  • EV Demand Response: HSR has 5 private CS (Charging Stations) for HSR personnel and 5 public CS. Especially the private CS should be used to supply information about potential occupation for the subsequent 24 hours to the affected staff members, asking them to refrain from charging in order to reduce or avoid power peaks. Again, here the ultimate goal is not to increase REN usage but to reduce CO2 emissions due to an optimized alignment of power production and supply at the CHP. Also, regarding automated EV DR, again aiming at power peak management, the access to CS service provider has been granted in 2023. Therefore, smart scheduling algorithms (Reinforcement Learning architectures) will be tested via in-field experimentations in mid 2023 onwards;
  • Electricity Supply Optimization: this epic mirrors the heat supply optimization one: using data science and AI-based forecasting to model distributions of buildings electricity consumptions and optimization of production accordingly. Thus, increasing energy efficiency concerning the electricity needed is to provide a specific level of beneficial energy services;
  • Forecasting: AI models and predictive capacity for future heat/electricity demand-response optimization. Forecasters implemented in Kubeflow software will deliver added value to all other epics;
  • Interactive Platform: use to communicate with energy managers and make all-purpose building users informed and more aware of the current energy consumption, shares and real consumption in the total EI buildings and areas. The Pilot 3 focuses on setting up the interactive platform displaying energy demand structures and energy supplies in the same visualizations;
  • Energy Reduction Campaign: Dedicated workshops and meetings in HSR will drive a social solution campaign to increase social organizational awareness about the energy topic and the RENergetic approach in the hospital. The Segrate part of the Pilot also is linked to such a social campaign especially in the neighborhood “Milano 2”, which acts as a replicator in RENergetic. In addition, here, energy focus groups, with private inhabitants, but mainly with political decision makers, are being set up discussing ways to advance the self-sufficiency and energy reduction in Milano 2.

San Raffaele Hospital Segrate Neighborhood

Milan - Italy