The study of energy consumption across various building clusters offers a path to discerning intricate patterns and establishing energy efficiency metrics. However, these analyses have mostly been limited to small, controlled settings, leaving a vast potential for broader application in energy efficiency management and classification untapped. This research leverages machine learning models to determine gas consumption patterns and energy efficiency characteristics of buildings in real-life settings, based on a range of parameters including insulation properties and year of construction. The developed system was applied to a comprehensive dataset comprising eight distinct clusters of buildings, with a total of nearly 10,000 hourly gas consumption. To supplement the analysis, additional data was gathered concerning the building features. The findings indicate that the average gas consumption varies significantly across clusters, with dependencies shown for the age of the building, insulation characteristics, and building orientation The developed framework proved to be suitable for gaining insights into average gas consumption and usage patterns at a building level, non-intrusively and on a large scale. The additional data provided comparative insights between different building groups. The developed system can be easily expanded for other building characteristics and could be used to drive tailored feedback on energy efficiency improvements within buildings. This research paves the way for a more comprehensive approach to building energy efficiency, one that goes beyond the traditional parameters to include a broader set of variables such as building usage, occupant behavior,and heating system efficiency

The primary objective of this research is to develop an energy management system for the Co-creation center (CCC) that maximizes the use of passive energy sources while maintaining indoor thermal comfort. Passive energy sources have the potential to significantly reduce the energy consumption of the building. However, to achieve optimal energy savings, it is necessary to integrate multiple passive energy sources and develop a control strategy that can manage them effectively.

Model Predictive Control (MPC) strategies have been extensively researched in the literature as a means of optimizing energy consumption in buildings. However, most studies only consider a single passive energy source or energy distribution in multiple zones. There is limited research on the optimal management of multiple passive energy sources. To address this gap, this thesis investigates the use of an MPC strategy to optimize the operation of multiple passive energy sources in a building. Specifically, the research focuses on four solar blinds, a Phase Change Material (PCM) battery, sky windows, heat recuperation, natural ventilation, and an active energy source.

Grey-box modeling is used to model the building, and the model is calibrated using experimental data. The MPC problem is then set up to minimize energy supplied by Heat pump (HP) while ensuring indoor thermal comfort during occupied periods. An adaptive comfort model is used as a criterion to satisfy during occupied periods. The proposed MPC control is then implemented in the building.

The results show that the proposed MPC outperforms the rule-based controller in terms of energy consumption and maintaining thermal comfort. The research further provides insights into the potential of MPC strategy to increase the energy flexibility of buildings. The final parts of this research focused on varying the PCM temperatures and using a more flexible thermal comfort model and studying its effects on the energy demand of the building. The findings could be used to inform the design of energy-efficient buildings and the development of smart energy management systems

The built environment is responsible for a substantial amount of energy use and greenhouse gas emissions. To improve energy efficiency and reduce environmental impacts, buildings can be renovated or replaced with more energy-efficient alternatives. Although new buildings often cause less environmental impacts from operating energy use, the higher material requirements cause additional environmental impacts. Housing corporations own more than 25% of residences in the Netherlands and need to decide between different energy efficiency improvement methods for their building stock. Therefore, the proposed thesis has the following research question: What are optimal building renovation or replacement solutions for housing corporations to improve energy efficiency and reduce environmental impacts in the context of the Dutch climate goals? Environmental impacts of materials and energy have been calculated according to the MPG+ method, which follows the life cycle assessment approach. The results show that extensively renovated porch flats and new buildings lead to similar amounts of environmental impacts if both use a collective heat pump and are insulated according to nearly-zero energy building standards. The comparison between the environmental impacts of extensive renovations and building replacement depends on the expected building life span after the intervention, the quantity of solar panels, and the use of sustainable materials. Renovations including lower energy efficiency levels or fossil energy sources cause more environmental impacts. The MPG+ method aligns with the Dutch policy context but lacks transparency, completeness, and data certainty. The alternative scenarios in this study are compared per square meter, so the overall environmental impacts may increase if the relative apartment size per resident is greater in the new building. Considering the Climate Agreement, the housing crisis, and many uncertainties, extensive renovations are recommended as a no-regret solution for housing corporations.

Especially in colder climates, like The Netherlands, a large quantity of non-renewable energy is used for space heating by burning natural gas. To meet climate goals, natural gas should be replaced with renewable sources. Reducing the total energy demand for space heating and lowering the design water temperature in a heating system enable more renewable energy production alternatives for space heating. So high temperature heating (HTH) systems need to be replaced with low temperature (LTH) alternatives to get dwellings through the energy transition.

Despite the indisputable advantages of LTH, it also comes with risks. One risk is that heating elements need to become exceedingly large to ensure sufficient heating capacities. Insufficient capacities may adversely affect indoor thermal comfort. To mitigate this effect, extra measures should be taken in the façade, ground floor and roof to reduce infiltration and increase insulation. In order to introduce these measures a renovation should be carried out. This study will focus on façade renovations.

For the assessment of comfort levels when changing to LTH, this research presents a case study into a typical Dutch terraced dwelling. This case study was carried out through simulations in TRNSYS 17. This is a validated program simulating energy flows in transient systems. Although TRNSYS 17 has built-in comfort calculation options, these were not deemed sufficiently transparent in their workings for this study. Furthermore, calculations of view-factors, mean radiant temperatures (MRT) and the predicted mean vote (PMV) showed deviations from standards NEN-EN ISO 7726 and 7730. Therefore a workaround is presented which calculates comfort levels closer in line with standards.

Two façade options were investigated for 1 HTH and 3 LTH systems. The HTH model was calibrated to match measured air and surface temperatures of a test dwelling. LTH was then simulated while keeping the current radiators, which showed that diminished capacities significantly reduce comfort and façade renovations cannot realistically mitigate this. This option is not possible when changing to LTH in dwellings in general. Then a system with specialized LTH radiators was investigated, with capacities determined through ISSO 51. This still showed a high peak of discomfort at current insulation levels. Comfort levels were improved significantly after a renovation to match the new ‘Target Values’ (‘Streefwaarden’) for insulation and an ‘excellent’ qualification for infiltration reduction. Finally, underfloor heating was investigated, resulting in enhanced comfort levels both with current and updated insulations levels. In all cases, a renovation has a positive effect on comfort levels and reduces heat transfers.

According to the European Commission, buildings are responsible for approximately 40% of the EU energy consumption and 36% of the CO2 emissions, resulting in the largest energy consumer in Europe. Especially, the reduction of energy consumption for heating of residential dwellings has gained more and more interest in the last decades. Although some reduction has been achieved by implementing passive or active strategies and improving the thermal performance of new or renovated dwellings, the decrease is found to be lower than predicted by the energy performance models. One possible reason is that the existing thermal comfort models, the PMV and the adaptive thermal comfort model, are not appropriate for all different types of buildings and climates, resulting in inaccurate predictions of thermal comfort and leading to amounts of energy use different than the actual ones. In addition, the occupants’ behaviour, which is determined by their thermal comfort preferences might play a significant role on the energy use which needs to be explored. Therefore, the main topic of the current thesis is to explore the relationship that different parameters such as dwelling and installation characteristics, indoor and outdoor climate parameters, household and physiological characteristics as well as occupants’ behaviour have with thermal comfort and energy use.

A measurement campaign took place within the context of the OPSCHALER project and data related to thermal comfort perception, indoor climate and energy consumption was collected for 96 dwellings in the Netherlands during periods ranging from two to twelve months over a one-year period (2017-2018). The gathered data included both quantitative and subjective data. The quantitative data included the indoor climate data (air temperature, relative humidity, CO2, presence), the gas use and the electricity use collected through smart meters and sensors. The subjective data related to comfort was gathered by a comfort mobile application where users recorded their thermal sensation and other comfort related parameters (thermal preference, clothing, metabolic activities, actions), while questionnaires and inspections did also take place. This data was analysed using different types of graphical representation (bar charts, pie charts, stacked bar plots, scatterplots) as well as statistical tests such as chi-square tests, regression analysis, Kruskal-Wallis and correlation tests. Due to limited or unavailable data only 78 dwellings were included in the analysis, while the reliability and generalisation of the results entails some uncertainty due to the number of dwellings and data points available.

All in all, it was found that the thermal sensation is correlated with the four comfort related factors (thermal preference, clothing, metabolic activity, action), the indoor air temperature as well as the energy labels and the ventilation types, while only age was resulted in not being associated with thermal sensation. Indoor air temperature was found to be related with the energy label of the dwellings, but not with their ventilation type. In addition, the results indicated that the heating energy use is related with the outdoor air temperature as well as with the indoor-outdoor air temperature difference, but not with the indoor air temperature. Energy labels and ventilation types were found to be related with the heating energy use, while thermal sensation, actions and income did not show any correlation with it.

Based on the findings it is concluded that the adaptive thermal comfort model is more accurate and closer to the reality of a residential environment in comparison with the steady-state PMV model, where no adaptations take place. However, it is highly recommended that instead of improving the existing thermal comfort models, personalised comfort models could be developed for each single person separately by gathering data using sensors. This would help to trace all different parameters that influence the thermal comfort and the energy use as well and lead to more accurate prediction and lower actual energy use.

The outbreak of the COVID-19 virus has increased the interest in respiratory droplet dispersion in the built environment. The dispersion of these droplets is inseparable from transmission risk. Sneezing, coughing and breathing can be modelled in many ways. In this study several models of increasing complexity are proposed and compared.
Three different models are distinguished. The first is a simple ballistic model, where the trajectories of particles of different sizes are simulated, given an initial velocity and particle diameter. This simulation gives a first insight in the movement of particles from Newton’s second law.
Model 2 works by releasing particles in a turbulent jet. Mean properties of turbulent jets have been derived from experiments and computer modelling. By means of particle tracking, the dispersion of three different sizes of particles are monitored. The turbulent behaviour of the jet is captured by applying a stochastic discontinuous random walk model (DRW) in Model 2. The smallest particles (10 μm) are airborne, while the large (100 μm) particles fall out of the turbulent jet and settle on the ground. Medium size particles (50 μm) have the widest range of dispersion. The effect of evaporation has been studied in this research as well. Both relative humidity and relative velocity (the velocity between the surrounding air and the particle) influence the evaporation speed. The medium size particles are affected the most by evaporation. Depending on the ambient conditions, the medium size particles either behave more like the large particles or as the small particles.
The third model contains the effect of an ambient crossflow, representing ventilation flows. Trajectories of bent jets have been calculated. Eulerian solutions for dilution have been found for such bent jets. No mean properties in literature have been found for bent jets and hence the Langrangian particle tracking model could not be used.
Based on the results of model 2 we conclude that ventilation flows will have a tremendous impact on the dispersion of respiratory droplets, in particular the particles with initial diameter 75 - 100 μm. These droplets become airborne and will follow the ventilation flow. In dryer air droplets evaporate faster and will become airborne more quickly.
Future research could be done to create a CFD model to generate analytical formulas for jets in ambient crossflows. More detailed room conditions like the stratification of air temperature can be taken into account.

Infiltration is the uncontrolled flow of air into the building through the envelope, which contributes to a significant amount of heat loss and is estimated to be around 30% of the heating demand. It is therefore essential to estimate the amount of infiltration, to design a suitable heating and cooling system in a building. Empirical relations which are used to determine infiltration, generally overestimate the infiltration rate, since the effects of weather, mechanical ventilation and building geometry are not taken into account. Large Eddy Simulation (LES) is used to determine the pressure distribution around a building along with an airflow network model to determine the infiltration flow rate. Validation studies of the LES model show that the pressure data at the front and back faces of a surface-mounted cube are accurately captured, whereas, the pressure at the top and side faces slightly deviate from reference data with an error of 16%-20%. This occurs due to a shift in the recirculation region at the top towards the leeward part of the cube surface. The reattachment of the flow is not observed at the top and side surfaces, which explains the deviations in pressure. The infiltration rate is analysed in the Matrix VII building using the in-house model, which combines the LES model and the airflow network model. It is then compared with other relevant standards and models like the NTA 8800 and an energy balance model. The comparison is made in terms of the cross-correlation between the models and the heat loss associated with infiltration. The in-house model shows good cross-correlation with the NTA 8800 model, with a Pearson correlation coefficient of 0.8454. The building standards which employ empirical relations, overestimate the heat loss by a factor of 10 compared to the in-house model. The in-house model is then used to develop a database of infiltration values for different building heights and airtightness values. The specific infiltration rate is observed to decrease in buildings taller than 120 m due to the minimal changes in velocity higher up the atmospheric boundary layer. The database is observed to provide values of infiltration with reasonable accuracy, compared to the existing database in the ISSO standard.

Sustainable development is the key solution to mitigate the impact of the 21st century’s major challenges such as the depletion of our natural resources, emissions that accelerate climate change together with the large waste generation. One of the largest contributor to the aforementioned issues is the existing building stock, in particular the office sector. Upgrading the sustainability performance of the existing building stock is not only in the interest of regulators or society. Real estate professionals are interested in the sustainability performance of their property, portfolio or occupied property. Due to the complexity, it is not feasible for market participants to verify the sustainability performance of a property. As a consequence, voluntary certification schemes have emerged to provide an objective evaluation. However, at the same time, appraisers are paradoxically expected to provide an objective estimation/opinion on the potential added value of sustainability while their duty according to governances is to derive the market value by analysing available data. The inability of current valuation practices for incorporating a thorough sustainability assessment resulted in decision-makers to focus solely on energy-related costs and justify their investment decisions based on energy-related benefits. For this reason, many investment decisions are believed to be profitable for users instead of owners. However, previous research has shown that property owners of sustainable offices might achieve positive effects in relation to an enhanced image, corporate social responsibility, less risky profile for investors and improved marketability. In the effort to shed light on these far-reaching tentacles of sustainability, this thesis sets forth the base for developing an indicator-based sustainability assessment framework for the appraisal of offices. The proposed framework integrates sustainability and the value it holds for various stakeholders by conceptualizing the relationship between sustainability indicators and different value systems as well as the economic parameters that are adjustable by an appraiser in the income capitalization method. In this way, an appraiser is able to provide an objective estimation of the possible added value of sustainability. Starting from an extensive literature review of different certification schemes, a pre-selection of sustainability indicators was made. Through conducting 8 semi-structured interviews with sustainability experts, the first selection of indicators could be determined together with their goals, measurement and criteria. It is remarkable that according to the participants, actively monitoring the consumption and other management-related indicators are defining the sustainability performance of an office as well, next to the physical characteristics. The online survey was distributed among sustainability experts with a questionnaire designed with the constant sum method and were requested to allocate points over the indicators as well as the categories to determine the relevance. Ultimately, the selection resulted in a set of 34 indicators. Combining the findings from the interviews, data from the survey and the literature review resulted in a framework which was discussed in an expert interview with an appraiser. The findings of the expert interview indicate that there is still a huge gap between practice and theory. The nature of the current valuation techniques hampers the integration of sustainability by focussing mainly on the rental agreement. In order to integrate sustainability aspects into the appraisal, a shift from only looking at the rental agreement to looking at the current use and the fitness for use would be needed. Future research could focus on validating and expanding the linkage between the sustainability indicators and the economic adjustable parameters and reducing the subjectivity of input for the indicators.

Buildings are a central element in our everyday private and social life, meeting some of our most basic needs such as shelter, sanitation and easy access to energy. However, these comforts do not come without a cost: European buildings alone consume 40% of the total final energy and are responsible for 36% of all CO2 emission. In Italy, 60% of the residential building stock was built before the first national law on energy efficiency and in fact some 25% of it can be considered as being highly energy inefficient. In 30 years, many of these buildings will still be standing and this threatens to lock-in the Italian residential sector on its historical emission pathways....

The residential sector is responsible for over 55% of the natural gas consumption in the Netherlands. In the climate accord of Paris, the Netherlands came to an agreement with the rest of the world leaders to limit the overall temperature rise by reducing the consumption of and by switching away from carbon-based fuels. The gas mining induced earthquakes in the northern part of the Netherlands increases the pressure on Dutch society to reduce natural gas consumption. The residential sector can play a role in reducing the consump- tion of natural gas in accordance with the Paris accord and to mitigate gas mining induced earthquakes. The amount of natural gas consumed per household is dependent on the behavioural aspects of residents, which are the biggest cause of uncertainty in estimating natural gas consumption. Current natural gas consumption based calculations are based upon dwelling characteristics and are not adjusted to individual behavioural as- pects. The inside temperature in a dwelling is seen as the primal indicator of residential heat consumption. The behavioural aspects of residents in the form of thermostat interaction are analysed in this thesis. The potential saving in the residential sector is addressed by including thermostat practices of residents in the estimation of potential savings.
The goal of this thesis is to identify thermostat practice in dwellings by the use of disaggregated energy con- sumption data and estimate the impact of adjusting individual thermostat practices on the natural gas con- sumption in the residential sector. Disaggregated energy consumption data is seen as detailed individual household consumption data. To reach the goal the following research question is answered:
What insights in thermostat practices that influence natural gas consumption of individual house- holds can be identified by a combined analysis of electricity and thermostat use?
Practice theory is used to understand the underlying mechanisms at play in household interaction with their thermostat. Disaggregated consumption data is used to gain insights in thermostat practices of individual dwellings. The thermostat practices of households are used to group specific practices and indemnify poten- tial savings in the residential sector.
The smart meter/ thermostat Toon is used to gather individual thermostat interactions and gas and electricity consumption data. Grouping of individual households with the use of clustering on the basis of thermostat settings is used to determine similar thermostat practices. Households with similar thermostat practices are grouped together, with the use of unsupervised classification in the form of hierarchal clustering. Similar thermostat practices groups are used to shape potential thermostat adjustments and assess the impact of these adjustments.
Thermostat practices of households are evaluated with the use of occupancy detection to identify potentials savings. A connection between thermostat practices and household occupancy is made with the use of elec- tricity consumption data. Individual electricity consumption data of households is used to determine the occupancy in a dwelling, by detecting moments of relative high consumption. Residential occupancy is de- tected with model ensemble of a Hidden Markov Model and a Rolling Mean Model to determine an overall occupancy schedule. The combined analysis of occupancy and thermostat practices is used to determine potential savings and determine the extent of these savings.
The generated insight in saving possibilities by a combination of thermostat practices and residential occu- pancy is used to develop and estimate the impact of 4 different thermostat practice adjustments. The esti- mations are based upon 3 different calculation methods: relative shift, heat demand and temperature and gas regression model, to gain a comprehensive understanding of the impact of each of the saving options. The potential savings are expressed in condition based lowering of the thermostat settings inside individ- ual dwellings. The relative shift method calculates the potential saving by relating gas consumption to the difference between in and outside temperature. Heat demand calculations are based upon the number of non-active heating hours for each of the saving options. In the regression method, a linear regression model for every individual household is build, to estimate the gas consumption on the basis of the difference be- tween the in and outside temperature.
More than half of the households in the sample group heats their home during daytime while the occupancy of these dwellings is detected at around 50%. The other half of the population has a clear heating pattern of
i
ii
morning and evening heating. For each of the detected thermostat practices, adjusting thermostat settings result in a potential gas saving. There is a factor 2.5 difference in potential saving between the different ther- mostat practices. Residents are able to save from 2% to 5% depending on their thermostat practice, resulting in an overall average saving of 34 euro per year. The largest saving potential of lowering the thermostat to 15 degrees overnight. Thermostat practices of households have a bigger impact on the gas consumption than currently used household characteristics. Natural gas consumption of households with similar thermostat practices have shown disparate consumption due to dwelling specific characteristics. Household specifics as thermostat practices and thermal inertia have shown to impact natural gas consumption. The detected ther- mal inertia of dwellings with the use of disaggregated consumption data is not in line with estimated values based upon physical characteristics of dwellings. The potential savings based upon physical characteristics of dwellings over estimates the potential saving in the residential sector. The impact of thermostat practices and thermal inertia of dwellings determined with the use of disaggregated consumption data is substantial. Cur- rent energy consumption studies and energy labelling are based upon physical characteristics of dwellings. The mismatch between estimated energy consumption and measured consumption indicates that estimat- ing residential gas consumption for individual households on the basis household characteristics alone is redundant. By including actual residential consumption data in energy labelling and shifting policy towards adapting household behaviour future policy measures can be improved.
The overall potential savings in the residential sector by adjusting thermostat settings are relatively small but can help to reach the climate goals. The results are based upon occupancy detection in a dwelling by analysing dis-aggregated consumption data. By improving the occupancy detection, the confidence in the results and number of included households can be improved. Expanding the size and representativeness of the sample group improves the applicability and confidence in the results. The personal preferences of residents in both the adoption adjustments and the actual impact on their thermal comfort is unknown. Fu- ture research is needed in order to close the gap between actual savings and potential savings by thermostat adjustments and to determine the impact on thermal comfort.

European policy makers have established a set of 20-20-20 targets for the year 2020 in the energy sector, goals based on security of supply, competitive markets and sustainability. The 3 key objectives include a 20% reduction of greenhouse gases (GHG) and CO2 compared to 1990 levels (and up to 80% by 2050), a 20% improvement in energy efficiency and a 20% share of basis energy consumption from renewable sources.
On its own, the building sector is responsible for 1/3 of the world’s energy consumption, of which 60% is due to heating and cooling. In the European Union, this proportion achieves 40% of the total energy consumption, of which heating, cooling and domestic hot water account for around 70%. Fossil fuels represent 75% of the primary energy supply for heating and cooling needs whereas renewable energies account for only 18% followed by nuclear energy with 7%.

A brief energy context being stated, this work focused on the assessment and the optimization of a heating, ventilation and air conditioning (HVAC) system, coupled with a building integrated photovoltaic and thermal (BIPV/T) system and a simple BIPV system. The objective was to reduce the overall energy consumption of a very well insulated dwelling aiming to make it a net zero energy building (NZEB). This HVAC system incorporated a dual-tank water-to-water heat pump (WWHP), used for space heating and cooling, using the BIPV/T system as a heat source thanks to preheated air. The dwelling is a so-called Deep Performance Dwelling (DPD) developed by the Solar Decathlon team from Canada: TeamMTL. This DPD was used as case study all along this Master thesis.

On the one hand, this study aimed at the modelling of both integrated PV systems which were developed firstly with the MATLAB software and then with the TRNSYS software. One of the objectives was to highlight the impact of the thermal recovery from the BIPV/T system on its PV modules efficiency compared to the BIPV system. Furthermore, the same initial and boundary conditions were used to ensure a truthfulness on-site energy production from both BIPV/T and BIPV systems as well as for the validation of these systems, later integrated to the complete DPD modelling.
On the other hand, sensitivity analyses were conducted to investigate the effects of various parameters on the NZEB objective and the on-site energy fraction (OEF) covered by the energy produced by the DPD. These parameters included the efficiency of the PV modules, their tilt angle, the materials used for both floors and walls as well as the size and the control strategy used of both electrical and thermal storage.

First, even if the comparison of both MATLAB and TRNSYS modellings led to a similar net increase in the BIPV/T average electrical efficiency of + 0.0014 [-] compare to the BIPV one, the TRNSYS one achieved an 11.0% lower yearly electrical generation due to an intrinsic factor used in TRNSYS called Incidence Angle Modifier (IAM). The impact of this factor affected all the more the thermal yield, namely a reduction by 28.6% for the annual thermal production along with a 32.0% lower thermal efficiency of the BIPV/T system compared to the one modelled in MATLAB. However, given that TRNSYS was required to model the transient behaviours of the DPD, not conceivable with MATLAB due to the time constraint of this Master thesis, it was logically concluded to use the TRNSYS modelling of the integrated PV systems.

Then, to have a good representativeness of the performances of the DPD, 4 cities with very different climates were chosen. For each location, the DPD was categorized as a nearly zero energy building (nZEB). Then, the simulation results emphasized that an optimal tilt angle adapted to every location, concrete-based floors as well as high efficiency PV modules allowed for reaching the NZEB objective in 3 cities, the 4th one facing low annual irradiation.

Moreover, sensitivity analyses were carried out in order to improve the energy performance of the DPD. Based on the interpretation of the results concerning both on-site energy indices and energy outputs, a better configuration than the one of the as-built DPD was suggested for each city. Indeed, the sensitivity analyses indicated that the bigger the electrical storage, the better the electrical OEF index. Then, a 300-litre hot water tank as well as a 150-litre cold water tank led to a better overall OEF index and a lower energy consumption compared to the initial configuration. These results were obtained by applying the initial control strategy relying on price benefits of the grid based on bidirectional power flows between the batteries and the grid. However, a grid-tied configuration with no more batteries-grid connection and not taking anymore advantage of incentive prices of the grid stood out an even better overall electrical OEF index, making a higher use of the on-site power generation.

Finally, the proportion between the on-site PV production and the amount of energy supplied by the grid to meet the demand of the DPD were calculated. Besides, the impact of the solar house in terms of CO2 emissions as well as its advantages for the grid were alluded.

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.

This thesis deals with determination of Resistance (R) value and Heat Capacity of building envelopes in the commonly found residential architecture in Holland i.e. pitched roof houses. The aim of this thesis is to find an easy to implement novel method which gives reasonably accurate results. In the absence of actual sensor data, the data-sets are obtained by modelling the buildings in EnergyPlus, with the help of Design Builder which provides the specifications of the buildings and its energy systems. Data-sets obtained via EnergyPlus provide the heat demand of the buildings. These data-sets, which are considered to be an approximation of actual data that would be available in future, is fed into MATLAB to perform the inverse modelling using Genetic Algorithms (GA) which estimates the unknown parameters by fitting the energy demand curve. Genetic Algorithms are known to converge to the global optimum unlike other regression techniques and parameter estimation methods. The objective function of GA is derived from the most accurate thermal network model. The equation is an energy balance of the room for the indoor temperature node which constitutes transmission losses, ventilation losses, solar gain and internal heat gain.The working and behaviour of Genetic Algorithms with varying optimisation parameters are thoroughly studied to make a comprehensive report on the scope of the Algorithm.