Future FM Research Overview

Imperial Building

The research project aims to answer the question "How can novel data gathering and analysis strategies help facilities managers deliver future-proofed energy efficiency improvements across the non-domestic buildings sector?" The project will develop and demonstrate novel adaptive methods both to improve the energy performance of existing buildings and to ensure that these gains are preserved in the face of technological and societal change.


 Case Study Buildings

The methods developed from this project will be demonstrated for buildings in the following (1) Educational, (2) Commercial, and (3) Retail. Demonstrations will first be made in virtual test beds that mimic the operations including abilities of the building energy management systems. The current case study buildings are described below.


Office Building: Imperial Estates

Imperial blue building•Imperial College administrative hub

•Centralised heating, cooling, and ventilation systems

•Uses thermal energy from combined heat and power system and campus heating loop

•Staff access tracked by ID card at 8 building entrances

•Digital meeting schedules for conference rooms




Supermarket: Sainsbury’s



 •Large standard floor plan supermarket

 •Biomass boiler serving 3 air handlers

 • Over 50 refrigeration cabinets using CO2 refrigerant
 •No additional space cooling required

WP 0: Virtual Test Bed Development


The thermal, moisture, and HVAC system response as well as the data management and communications capabilities of case study buildings will be simulated in a virtual test bed. These virtual replicas of the case study buildings will be used a part of our hybrid physics based and statistically based fault detection as well as to quantify the benefits of advanced building control algorithms.

Work flow


WP 1: Data Mining and Machine Learning for Future FM

WP 1 dataThe aim of this work package is to use building data streams currently monitored and logged to understand the variability of occupancy and the changing operational status of the HVAC System. More specifically the building occupancy will be inferred from data streams such as security access logs, administrative meeting schedules, and electricity consumption. In addition, maintenance logs paired with historical BMS data will be used to determine fault behaviour of HVAC system components which will be used for future fault detection and diagnosis as well as an adaptive model of HVAC system performance.


WP 1 mining data


WP 2: System Design through Global Sensitivity Analysis

WP2 picture

Global Sensitivity Analysis (GSA) is a method that simultaneously varies the input parameters of a model to determine their effect on the desired output. This method will be used in the current effort to determine which parameters have the largest impact on building energy usage allowing for the development of a reduced-order, or simplified representation of the building’s dynamics. GSA will also be used to determine the effect of sensor accuracy on the resulting energy consumption.





WP 3: Smarter Control Algorithms for Facilities Management

Wp3 algorithmsThe predictive and adaptive models developed in Work Package 1 as well as the reduced order models from Work Package 2 will be used to develop a supervisory model predictive control (MPC) algorithm. The MPC will allow the buildings to minimize energy consumption and maintain thermal comfort while considering changing patterns of occupancy and HVAC system performance.



WP 4: New Business Plans for Building Management Services


This work package aims to devise a virtuous cycle in which building stakeholders push towards smart building performance. Researchers will review and evaluate financial models in the non-domestic sector that can alleviate the capital burden and provide added value to building owners and operators, while at the same time offering a sustainable business model to energy service providers by focusing on the impacts it can have on building owners, building managers, tenant, and FM teams.