Marco Savino Piscitelli Ph.D. dissertation
Piscitelli Marco Savino successfully defended his Ph.D. thesis entitled "Enhancing energy management in buildings through data analytics technologies". We congratulate him for this great achievement!
Abstract
Advanced metering infrastructures are enabling the collection of large amounts of buildingrelated data that are leading to a profound transformation of the energy management paradigm in buildings and energy grids. Building-related data are full of hidden knowledge that can enable significant energy savings when a proper knowledge discovery process is performed. To this purpose advanced Energy Management and Information Systems (EMIS) based on the application of powerful and novel data analytics techniques can be employed. The focus of this dissertation is on the specific segment of EMIS technologies called Decision Support Systems (DSS). DSS include Energy Information Systems (EIS) and Fault Detection and Diagnostic (FDD) systems and can be classified as enabling tools in the building energy management process. Differently from advanced control systems, DSS provide feedbacks to human users (e.g., energy manager, building owner, energy service company) assisting them in improving building performance during operation. The installation of such systems is characterized by a low investment cost and a high energy saving potential making them strategic technologies in the building sector. However, their penetration in the market is still not satisfactory. In this dissertation four advanced and innovative data analytics based DSS tools (three EIS tools and one FDD tool) at both meter and system level are proposed with the aim of overcoming three main barriers that today thwart the full exploitation of such systems: (i) low level of user engagement, (ii) inadequate detail of the analysis and information provided, (iii) insufficient level of interpretability of the results obtained. For each scale of the analysis considered a novel methodological framework is employed for addressing the main tasks typically required to advanced EIS and FDD systems.
(...)
All the developed tools leveraged on time series analytics and automated rule extraction techniques with the aim of maximizing the amount of information extracted from building data while maintaining a high level of feedback interpretability. The results obtained demonstrated the added value of data analytics in the process of building energy management and its effectiveness in extracting hidden, useful and actionable knowledge at different scales of analysis. Findings and outcomes of the present research study are discussed providing a robust reasoning about the optimal design of data analytics processes according to specific mining purposes. Eventually, a wide overview on the lessons learned throughout this research study is proposed for clearly outlining the future application opportunities, and barriers of data analytics technologies in the energy and building sector.