An investigation into energy optimization techniques for buildings was initiated that led to the development of a Toolbox with several functions for analysis, optimization and prediction techniques for thermal energy demand of a school building. The HHS, or The Hague University of Applied Sciences in Delft, was the most sustainable building for the years 2011-2012. Naturally, the building also incorporated features and capabilities which can help an engineer to study methods of making a smart building, better.
Using sensor driven data from stored databases in the building, optimization and analysis tools have been developed for the building, at the room level. These analyses are automated into the toolbox for any given room of the building, with minor changes. The goal is to help an expert analyze the room in a quick and efficient manner.
Using the indoor/outdoor climate data, occupancy related profiles, and internal heat loads, the model can also generate predictive patterns and determine the explanatory power of each of these variables on the thermal energy demand of a room. To do this, the Toolbox is designed with two predictive modeling techniques, unique in their own ways. The first being a Multivariate Linear Regression model, that allows for estimation of thermal demean based on a linear thermal balance equation of the room. This is followed by the use of Artificial Neural Networks, to dive deep into the intricacies of the complex data of a room, especially in the case of a highly controlled indoor climate of a room. The goal here was to understand the predictive capacity of these techniques over a) real time data, and b) over the data belonging to a room and not the entire building.
Finally, looking outwards to optimizing energy demands of buildings, this Toolbox, aims at estimating quick wins that can be gathered from a smart building, to reduce energy demand further and tend the building towards nearly zero energy in the future.