Artificial intelligence for optimisation and demand side response in built environment

In recent years, data analytics and machine learning have become important for generating insights and creating competitive advantages across many industries. However, despite recent developments in analytics and machine learning technologies, as well as accessible tools to perform the implementation of such technologies, built environment is still a largely unexplored field, where many engineering operations remain manual. Recent advances in building management systems and data engineering have provided vast amount of data streams of all types of sensors around the built environment: comfort variables, assets, security installations, meteorological measurements, electricity demand, etc. This creates a wide range of opportunities to explore data and extract real value for various operational purposes. The aim of this Thesis is to develop a series of tools related for control and optimisation of built environment, from demand side response events prediction to building operations management. Equipped with the results of this work, building managers will be more prepared to respond to energy demand events, organise energy resources more efficiently and to acquire a proactive approach to system failures and system errors tractability. The pilot results of the thesis have been successfully implemented in industrial applications.