Optimal operational strategy for waste heat recovery from an IT servers’ room in a uni- versity campus

Authors: Alejandri Jose Di Cicco, Marilena Scorzetti, Rossano Scoccia and Marcello Aprile

Abstract: Waste heat recovery in data centres can be an effective strategy to reduce both the overall energy consumption of the server room and the space heating in nearby buildings. Considering the complexity of the interactions between cooling, heat recovery, and building heating systems, this study aims to evaluate the operational integration of data-centre waste heat recovery under realistic demand constraints. This is assessed using a Mixed- Integer Linear Programming (MILP) optimization framework focusing on a case study that includes a university building from the Politecnico di Milano campus and an IT server room, considering a water-to-water heat pump for waste heat recovery. Results show that waste heat recovery reduces the Power Usage Effectiveness (PUE) from 1.59 to 1.50 – 1.46 and produces an Energy Reuse Factor (ERF) from 14.5% to 21%. Annual financial savings range from 34k to 49 k€, while equivalent CO₂ emissions are reduced by 18%-26%. Evaluation of different extended heating schedules with early start results in the best overall performance, highlighting the importance of heating load profile optimization for better waste heat utilization.

Key words: Data Centre, Waste Heat Recovery, MILP modelling