Running CATALYST at the data Center in Pont Saint Martin

Diego Arnone, our project co-ordinator and senior researcher in his day job at Engineering Ingegneria Informatica spa explains the concept of “energy flexibility” and practical information at one of our pilot sites at Pont Saint Martin in Italy.

It’s time for validation! After several months of developments and many tests already performed in all the CATALYST Pilot data centers, it’s now time to evaluate what kind of benefits the software developed so far can give to a data center, the surrounding area and the environment. In Pont Saint Martin, in Valle D’Aosta Region, the Engineering Group, one of the biggest IT player in Italy and coordinator of the CATALYST H2020 project, owns and manages one of its 4 data centers (the other three are in Turin, Milan and Vicenza).


This 1 MW colocation Data Center has already put in place many actions to improve its own energy efficiency. The Company has updated the cooling system by adopting a state-of-the-art geothermic system that exploits the water at 13 degrees in the underlying waterbed. As a result, savings of 20% of electricity are now possible as well as an increase in the refrigerating capacity. Thanks to this change, the Data Center has achieved a Power Usage Effectiveness of less than 1.5, which is a valid international standard for measurement of energy efficiency. But since the digitisation of our lives and activities is dramatically increasing DCs’ electricity consumption, the Data Center in Pont Saint Martin has been actively participating in several research initiatives in order to prepare itself to a more sustainable future. CATALYST is one of the running research projects that the Data Center is involved in. Once that the deployment of the CATALYST software is complete, the Data Center Manager will be able to assess what kind of flexibility the data center is able to provide to other stakeholders, like the Distribution System Operator or energy prosumers in the neighbourhood. The power adsorption of such a data center is basically flat, close to a constant 1.1 MW. But all the tests performed so far prove that the power adsorption curve can be changed, without affecting the daily operation, by exploiting the thermal inertia of the server rooms and the intrinsic redundancy of the existing equipment (e.g. refrigerating units, UPS, back-up generators). There is room for flexibility that the Data Center could sell to a Grid Operator: it’s a win-win situation because the data center can benefit of new revenue streams and the Grid Operator can benefit from the provided flexibility that could allow it, even in an aggregated form with other users, to ensure a better balance between electricity demand and supply.

As example, the following picture shows how the power adsorption curve of the data center changes during one of the tests performed in Pont Saint Martin.

If the Grid Operator requires a reduction of power consumption in a certain time slot to solve a congestion problem, the Data Center could switch off the refrigerating units for a while and monitor the temperature of the manifold water return till it is close to threshold over which the back-up cooling system is triggered for avoiding possible malfunction of the refrigerating units. In the timeslot when refrigerating units are off, the power adsorption drops by 17% (almost 180kW under the baseline). Today, this reduction, properly aggregated with other reductions provided by other users, could be sold on the Italian Ancillary Service Marketplace (MSD: Mercato dei Servizi di Dispacciamento) whose unique buyer is the Transmission System Operator, but in a future envisaged by CATALYST, it could be sold on a local electric flexibility marketplace whose unique buyer is the local DSO. The CATALYST software suggests to the data center manager when this reduction is possible and convenient.

But can the DC provide more flexibility than 190kW?

The figure below shows what the DC can do when it works on islanded mode, i.e., when it is disconnected from the Power Grid. Thanks to its back-up generators, the DC can reduce its power absorption by 100%: a flexibility of 1.1MW could be an important resource for the local DSO and, today, could be sold even without an aggregator.

However, using diesel generators is not the greenest choice that a DC can make. That’s the reason why CATALYST aims to simply align the provision of this flexibility service to the maintenance plan of the diesel generators that are periodically activated to check if the back-up systems work properly.

So, maintenance plans, technical data sheets of the equipment, monitoring data, already gathered by the existing monitoring and control infrastructure, are stored in the CATALYST data base, instantiation of the data model defined in the research project, and periodically processed to assess the available flexibility: in few words, the software is able to suggest the DC Manager to sell or buy thermal energy, electric energy or to sell flexibility to the Grid Operator. And a Data Center is a perfect candidate to provide different flexibilities on different carriers because it is a hub at the crossroad of at least three networks: electricity network, district heating network and data network. Processing IT workload means consuming electricity and producing heat: so, the three commodities are intrinsically linked to each other. Indeed, CATALYST also considers IT workload migration as a leverage to provide flexibility, since moving workload from a DC to another federated and trusted DC is like moving the energy consumption necessary to process the same workload.

This is one of the functionalities that the CATALYST software already implements but whose activation depends on the type of data center: for a data center like the PSM’s one, IT workload migration is not a viable option. But this is the focus of a future article.

At a glance

  • No: 768739
  • Acronym: CATALYST
  • Title: Converting data centres in energy flexibility ecosystems
  • Starting date: October 1, 2017
  • Duration in months: 36
  • Call identifier: H2020-EE-2017-RIA-IA
  • Topic: EE-20-2017; Bringing to market more energy efficient and integrated data centres