Article
To read the full-text of this research, you can request a copy directly from the authors.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... -Data collection is fundamental for occupant behaviour modeling. Often when modeling occupants, they are static and are based on subsequently fixed schedules, which have been demonstrated to deviate notably from those specified in standard guidelines such as from ASHRAE suggesting that a one-size-fits all approach does is not ideal with the variety in buildings even within specific categories such as offices [13] [14]. A way to address this has been through probability distributions based on simulated or observed data. ...
... This should, according to prior similar studies, allow for predictions on the scale of one day, with the possibility to expand this to multiple days or beyond. For example, the study by Liang et al. used a timestep of 5 minutes for recording their sensor data [14]. ...
... As previously noted, it is important to establish a proper selection of model and data sources in implementing successful occupancy prediction systems based on sensory data such as that from IoT devices [35]. In the case of preparing data for the training of any AI context, the data will need to be pre-processed [14]. When first installed, the sensor data is likely prone to errors, imprecision, and inaccuracies, among other factors. ...
... Czynnikami decydującymi o energochłonności budynku są przede wszystkim [66,67,68,58,69,70,31,71,72,73]: ...
... Na etapie eksploatacji budynków biurowych, zachowanie użytkowników może decydować nawet o ponad 40% zużycia energii [68], a zatem ma kluczowe znaczenie dla rzeczywistej efektywności energetycznej całego budynku. Aby zredukować wpływ działań użytkowników na energetyczne osiągi budynków, poddaje się automatyzacji systemy, które ludzie tradycyjnie sterowali w sposób ręczny dostosowując warunki mikroklimatyczne we wnętrzu (np. ...
... Niektórzy pracownicy wielokrotnie, w sposób nieregularny opuszczają swoje stanowiska pracy w ciągu dnia. Schematy obecności poszczególnych pracowników przy stanowiskach mają więc silnie stochastyczny charakter i są bardzo zróżnicowane [68,268,269]. W przyszłości, różnice te będą się prawdopodobnie pogłębiały, ze względu na wprowadzanie bardziej elastycznego stylu pracy, który daje pracownikom większą swobodę wyboru jej czasu i miejsca (możliwość pracy poza biurem) [48,270]. Także stanowiska pracy coraz częściej nie są przypisane konkretnym pracownikom, lecz mogą być wykorzystywane w sposób elastyczny przez wiele osób 9 . ...
Thesis
Full-text available
W pracy zawarto opisy analiz, badań symulacyjnych, ankietowych i doświadczalnych w zakresie możliwości poprawy efektywności energetycznej oświetlenia, przy użyciu nowoczesnej instalacji elektrycznej stosowanej w budownictwie biurowym typu smart. Scharakteryzowano różne metody sterowania oświetleniem, przeanalizowano ich energooszczędność, a także zwrócono uwagę na aspekt akceptacji przez ludzi systemów automatycznego sterowania. Wykazano, że indywidualne preferencje oświetleniowe ludzi są zróżnicowane, a w biurach znacząca liczba osób preferuje natężenie oświetlenia niższe niż rekomendowane. Zaproponowano metodę lokalnego, spersonalizowanego sterowania oświetleniem w przestrzeni biurowej typu open space, która wykorzystując matrycę niezależnie sterowanych lamp LED oświetlenia ogólnego, łączy regulację indywidualną według osobistych preferencji oraz sterowanie zależne od zajętości poszczególnych stanowisk. Optymalizację tego typu problemu sterowania przeprowadzono na przykładzie wybranego pomieszczenia doświadczalnego w oparciu o dwa modele matematyczne systemu oświetleniowego: liniowy i wykorzystujący sztuczne sieci neuronowe. Rezultaty wykazały, że proponowana metoda sterowania umożliwia poprawę efektywności energetycznej pracy systemu oświetleniowego oraz warunków oświetleniowych wytwarzanych na stanowiskach pracy, względem tradycyjnej instalacji oświetleniowej.
... Actual data acquisition becomes a primary step in studying the occupant's impacts. The most commonly used techniques are sensors, smartphones, and cameras [13,14,21,29,[65][66][67]. At the same time, a significant amount of research has been carried out to develop modeling methods to explore the energy impact from the realistic occupant's data [11,13,20,26,28,[68][69][70][71][72][73][74][75]. The frequent applied modeling methods are agent-based modeling (ABM) [76,77], data mining [78,79], stochastic modeling [80,81], and statistical analysis [82,83]. ...
... Data mining is the process to explore if there is a consistent relationship among variables based on large databases. Existing research usually utilized the data mining approach to learning the long-term occupant behavior from the energy consumption of different end-user groups, such as electricity usage data [13,26,[68][69][70]. Stochastic modeling methods are usually adopted due to the random feature of occupant behavior. ...
... • Produce a time series of the state of presence that includes the duration of presence and absence, arrival and departure time, and typical random movement; • The model does not simulate the displacement of occupants from one zone to another. Occupancy schedule and partners [69] 2016 Office Case study, data mining, cluster analysis, decision tree ...
Article
Three main categories of occupant behaviors were summarized for building through this literature review, including the occupancy, interactions and behavioral efficiency. The results of the review confirmed that the actual occupancy and the interactions with buildings are the key influencing factors determining the energy building consumption. Behavioral efficiency has been identified as an efficient and economical method compared with retrofitting technologies. But, categorizing and quantifying the behavioral inputs as well as the validations are needed to be improved in the future. Window opening behavior has been seldom taken into account when calculates the energy impact because window opening behavior is rare in most of the centrally air-conditioned buildings. Also, the energy impacts of genders are not as much as engineering significance in energy consumption, but they are important for understanding the variation of the personal and other environmental control. The results are important to identify the key factors and address the determining considerations from those factors, in order to avoid unnecessary and redundant data collection. That is also significant to the energy simulation and the development of software packages in future work.
... The role of residents is a major factor in building energy consumption. Based on the results of research conducted by Liang, Hong, et al. (2016) in an office building in Philadelphia, the USA shows that how the building's energy consumption is based on four distinctive patterns of user presence through cluster analysis. Besides the occupant presence, to predict energy use patterns (e.g., electricity, gas, and water), user behavior (e.g., opening and closing windows, turning on and off lights) is assessed based on historical data [16]. ...
... Based on the results of research conducted by Liang, Hong, et al. (2016) in an office building in Philadelphia, the USA shows that how the building's energy consumption is based on four distinctive patterns of user presence through cluster analysis. Besides the occupant presence, to predict energy use patterns (e.g., electricity, gas, and water), user behavior (e.g., opening and closing windows, turning on and off lights) is assessed based on historical data [16]. ...
... Based on its effect on humans, noise is divided into irritating (less loud intensity, such as snoring), masking (sound that obscures clear hearing, and indirectly affects occupational health and safety), and damaging or injurious (sound that exceeds the Threshold Value). It damages or reduces hearing function [16] [74]. Noise above 80 dB causes restlessness, malaise, tiredness of hearing, stomach pain, and circulatory problems. ...
Article
Full-text available
An office is a center of operations where various activities take place, depending on the business conducted. Office conditions determine the credibility and synergy of business entity, as well as the officers' performance and productivity. This study formulates an interior design concept based on the Post Occupancy Evaluation (POE), with Mixed Methods Research (MMR) approach, which emphasizes the empirical meaning based on relevant field data. The research object is the Haleyora Powerindo (HPI) office building in Jakarta. MMR approach ensures the significance level of the work environment quality in the room and the workplace, as well as the officers' performance. The first step was analyzing the spatial conditions in the field based on standardization, which was conducted according to the satisfaction survey of officers. The results show that the work environment quality is high and responds positively. According to the workplace quality, satisfaction is mainly related to workspace area and quality, comfort and style of furniture, circulation, and accessibility to the work table, effectiveness and efficiency of workplace layout, ICT implementation, and HVAC and maintenance. Social space and entertainment facilities provide flexibility in interacting between officers of different divisions that were not in the previous office. Based on the results of multiple regression calculations, the in-door environment (IEQ) and the officer workspace (WQ) quality have a significant effect on officer performance (OP). These are 3 main factors critical in conducting systematic evaluations to obtain results from office space design. This study is expected to be a strategic reference for planning and designing a similar spatial atmosphere in different locations or cases
... These platforms provide users with integrating mechanisms such as issue 23 reporting, pull requests, commenting and reviewing support . 24 During software development, all the produced data are stored in software reposito-25 ries. Version Control and Source Management System (SCMS) repositories include all the 26 changes of all the open source software projects. ...
... There are some typical clustering models [22] [23] [24]: connectivity models (e.g., 117 hierarchical clustering), centroid models (e.g., k-means clustering and k-medoid clustering), 118 distribution models (e.g., multivariate normal distributions), density models (e.g., DBSCAN 119 and OPTICS), subspace models (e.g., Biclustering also known as Co-clustering or two-mode-120 clustering), graph-based models (e.g., HCS clustering algorithm). In ML and statistics areas, classification is a type of supervised learning method that 123 aims at identifying the categories of instances. ...
Preprint
Version Control and Source Code Management Systems, such as GitHub, contain large amount ofunstructured historical information of software projects. Recent studies have introduced Natural Language Processing (NLP) to help software engineers retrieve information from very large collection of unstructured data. In this study, we have extended our previous study by increasing our datasets and ML and clustering techniques. Method: We have followed a complex methodology made up of various steps. Starting from the raw commit messages we have employed NLP techniques to build a structured database. We have extracted their main features and used as input of different clustering algorithms. Once labelled each entry, we have applied supervised machine learning techniques to build a prediction and classification model. Results: We have developed a machine learning-based model to automatically classify commit messages of a software project. Our model exploits a ground-truth dataset which includes commit messages obtained from various GitHub projects belonging to the HEP context. Conclusions: The contribution of this paper is two-fold: it proposes a ground-truth database; it provides a machine learning prediction model. They automatically identify the more change-proneness areas of code. Our model has obtained a very high average precision, recall and F1-score.
... In this context, the aim of this paper is to develop the most typical indoor temperature profiles that can be used as baseline energy models developed to estimate the energy consumption of homes in Albania, as well as to closely estimate energy savings for future energy retrofitting programs. Hierarchical cluster analysis is used in this study to develop the most typical indoor temperature profiles, as one of the most used statistical analysis methods used previously in the development of occupancy, heating, and indoor temperature profiles [26][27][28][29][30][31]. Belaïd, Ranjbar, and Massié [32] have also used Ascending Hierarchical Classification to define French housing stock profiles based on 1400 dwelling representatives of the French residential sector, aiming to explore the cost-effectiveness of the energy efficiency implementation measures in the residential sector. ...
... Cluster analysis a statistical analysis method used in previous studies to determine occupancy and heating profiles [26][27][28][29][30], as well as indoor temperature profiles in living rooms [31]. There are two main cluster analysis methods, k-means and hierarchical. ...
Article
Full-text available
Oversimplifying occupant behaviour using static and standard schedules has been identified as a limitation of building energy simulation tools. This paper describes the use of hierarchical cluster analysis to establish the most typical indoor temperature profiles of Albanian dwellings based on monitored indoor temperatures in winter and summer, along with building and occupant surveys undertaken in 49 randomly selected dwellings in Tirana. Three statistically different profiles were developed for each summer and winter, indicating that homes are used in different ways, as well as revealing possible comfort requirements. Furthermore, statistical analysis was undertaken to determine the strength of the association between the clusters and contextual factors related to the building, household, and occupancy. A statistically significant association was found between the presence of children and the clusters in winter, suggesting that families with dependents use more energy. Building-related factors including building type, building age, and wall insulation were found to be statistically significantly associated with clusters in summer. These profiles could provide more accurate outcomes of energy consumption of Albanian homes and energy savings from retrofits. They could also facilitate the development of low-energy strategies and policies for specific households.
... Energy use during an occupant's absence has a considerable impact on energy-saving potential [117]. Data-driven based approach is a promising method to extract the occupancy schedules and predict the number of occupants [118]. Panchabikesan et al. [89] developed a datadriven framework to extract the occupant activity and presence schedule from several apartments using shape-based clustering and change-point detection techniques. ...
... As mentioned earlier in this section, plug load, lighting load, setpoint schedules, and DHW usage are the other occupant-related inputs that need to be defined while performing the energy simulation. Assigning the appliances and lighting load schedules by occupancy is crucial as they are highly correlated [118]. ...
Article
Occupant-related inputs are significant parameters that influence energy simulation accuracy at both the building and urban levels. In most previous research works, fixed occupant schedules were used in urban-scale building energy modeling. The main reason is the lack of data availability to model the dynamic occupancy schedules. In recent years, urban data sets and modern estimation and detection techniques were introduced to increase the availability of occupant-related data sets. Yet, using these data to model the detailed occupancy at the urban scale is challenging and not much explored. Also, it is unclear how detailed the input regarding the occupancy should be. Addressing these research questions is the main objective of this study. This paper presents a comprehensive review of the occupant behavior (OB) modeling approaches, occupant-related input parameters with particular focus on the occupancy schedule, lighting, appliances use schedule, temperature set-point schedule, and domestic hot water usage for urban building energy modeling (UBEM). Strategies to consider the occupancy sub-models as co-simulation connecting to urban building energy simulation are discussed. Some potential datasets that could be used to derive occupant-related inputs at the urban scale are presented and highlighted. Further, the correlation between occupants’ activity level, plug, and lighting loads is discussed in detail. Finally, the limitations and challenges of the occupant-related data connecting to building energy modeling are discussed, along with the research gaps and future directions of occupant-centric UBEM.
... DM techniques were developed to learn and predict occupancy in three main formats in previous studies: binary occupancy (i.e., occupied or unoccupied) [31] , numerical values (i.e., occupant numbers) [12] , and continuous occupancy (i.e., the probability distribution of occupancy) [29] . ML is an important principle embodying DM [32] , allowing computers to learn from historical data and predict target values. Two major ML types are used frequently in building engineering research areas: supervised and unsupervised learning algorithms [33] . ...
... Two major ML types are used frequently in building engineering research areas: supervised and unsupervised learning algorithms [33] . Supervised learning is a traditional learning method with training data and target labels [32] , and It can be divided into two categories: classification and regression. Classification is used to predict the data categories (e.g., fruit breed prediction), while regression is utilized to predict continuous value based on previously observed data (e.g., housing price prediction and height estimation). ...
Article
A reliable occupancy prediction model plays a critical role in improving the performance of energy simulation and occupant-centric building operations. In general, occupancy and occupant activities differ by season, and it is important to account for the dynamic nature of occupancy in simulations and to propose energy-efficient strategies. The present work aims to develop a data mining-based framework, including feature selection and the establishment of seasonal-customized occupancy prediction (SCOP) models to predict the occupancy in buildings considering different seasons. In the proposed framework, the recursive feature elimination with cross-validation (RFECV) feature selection was first implemented to select the optimal variables concerning the highest prediction accuracy. Later, six machine learning (ML) algorithms were considered to establish four SCOP models to predict occupancy presence, and their prediction performances were compared in terms of prediction accuracy and computational cost. To evaluate the effectiveness of the developed data mining framework, it was applied to an apartment in Lyon, France. The results show that the RFECV process reduced the computational time while improving the ML models’ prediction performances. Additionally, the SCOP models could achieve higher prediction accuracy than the conventional prediction model measured by performance evaluation metrics of F-1 score and area under the curve. Among the considered ML models, the gradient-boosting decision tree, random forest, and artificial neural network showed better performances, achieving more than 85% accuracy in Summer, Fall, and Winter, and over 80% in Spring. The essence of the framework is valuable for developing strategies for building energy consumption estimation and higher-resolution occupancy level prediction, which are easily influenced by seasons.
... For the existing work, a special attention is given to the smart meters and to the advanced metering infrastructure using the electrical energy consumption using data mining and analysis, as from them the consumption data generated [5]. This infrastructure is still under development but many regulators and governments put the focus especially to the smart meter deployment. ...
Article
Full-text available
In this study paper, the feasibility of constructing a complete smart system for anticipating electrical power consumption is created, as electricity's market share is expected to expand over the future decades. Smart grids and smart meters will help utility companies and their customers soon. New services and businesses in energy management need software development and data analytics skills. New services and enterprises are competitive. The project's electricity consumers are categorized by their hourly power usage percentage. This classification was done using data mining (five algorithms in specific) and data analysis theory. This division aims to help each group minimize energy use and expenditures, encourage energy-saving activities, and promote consumer involvement by giving tailored guidance. The intended segmentation is done through an iterative process using a computer classification computation, post-analysis, and data mining with visualization and statistical methodologies. This is an open access article under the CC BY-SA license.
... They used motion detector sensors to test their model. Liang et al. [35] worked on the problem of occupancy pattern learning and occupancy schedule prediction in office buildings. Their hybrid approach first recognizes the occupants' presence patterns using cluster analysis and then learns the schedule rules using the decision tree. ...
Article
Full-text available
Indoor occupancy prediction can play a vital role in the energy-efficient operation of building engineering systems and maintaining satisfactory indoor climate conditions at the lowest possible energy use by operating these systems on the basis of occupancy data. Many methods have been proposed to predict occupancy in residential buildings according to different data types, e.g., digital cameras, motion sensors, and indoor climate sensors. Among these proposed methods, those with indoor climate data as input have received significant interest due to their less intrusive and cost-effective approach. This paper proposes a deep learning method called CNN-XGBoost to predict occupancy using indoor climate data and compares the performance of the proposed method with a range of supervised and unsupervised machine learning algorithms plus artificial neural network algorithms. The comparison is performed using mean absolute error, confusion matrix, and F1 score. Indoor climate data used in this work are CO2, relative humidity, and temperature measured by sensors for 13 days in December 2021. We used inexpensive sensors in different rooms of a residential building with a balanced mechanical ventilation system located in northwest Copenhagen, Denmark. The proposed algorithm consists of two parts: a convolutional neural network that learns the features of the input data and a scalable end-to-end tree-boosting classifier. The result indicates that CNN-XGBoost outperforms other algorithms in predicting occupancy levels in all rooms of the test building. In this experiment, we achieved the highest accuracy in occupancy detection using inexpensive indoor climate sensors in a mechanically ventilated residential building with minimum privacy invasion.
... The latter approach is the development of models to predict occupant behaviour and schedule prediction based on the actual behaviour occupancy data obtained from different kinds of sensors [150]. The primary goal of the studies measuring the occupancy pattern in libraries using occupancy sensors is to provide students with more efficient library use. ...
Thesis
Full-text available
Millions of people migrate every year, aiming to settle either permanently or temporarily in new places. People from countries with different intensities of daylight might have various perceptions and expectations towards the climatic and indoor conditions. It might result from a previously accustomed lighting environment and other associated factors, namely individual cultural background. It is important because study outcomes may be utilised by architects and lighting professionals on how to design buildings and interior spaces depending on occupants’ perceptions and expectations to increase occupants’ satisfaction. This knowledge also can be used to save energy because the efficient use of daylight can reduce the energy consumption of both HVAC and illumination systems. To date, few studies have explored the relationship between cultural background and daylight perception; however, they mostly focused on glare sensitivity rather than daylight intensity. Cross-cultural studies aiming to investigate lighting preferences in interior environments are also rare. What is not yet known is the importance of cultural background and its impact on daylight perception, expectation, and satisfaction. Therefore, the development of a methodology for assessing daylight perception and its application in the context of cultural background are the main objectives of this research project, in order to investigate the impact of cultural background on daylight perception. In this thesis, three subjective evaluation methods were used to assess participants' daylight perceptions: subjective ratings, seat preference, and daylight boundary line drawings, and the perceived daylight availability obtained through these methods were compared to measured daylight availability. It was demonstrated that perceived daylight availability obtained through seat preference and subjective statement methods corresponds to some extent with actual daylight availability (p<0.01 and p =0.002, respectively). The findings obtained from both students’ seat selections and occupancy data from motion sensors in the library also highlighted the importance of daylight availability in the seat selection of students in the libraries. However, the lit area drawn by participants representing the perceived daylight conditions as part of the daylight boundary line method varied extensively from person to person regardless of actual daylight measurements. In other respects, a systematic review was conducted to create a conceptual framework of cultural background in the lit environment, and factors thought to be influencing daylight perception in the cultural context had been defined in four ways. These were ethnicity and/or physiological properties of individual eyes, the residential area, the previous luminance environment and sociocultural background. Finally, the developed methodology based on the previous findings was applied to understand if individuals perceive daylight conditions differently due to their cultural backgrounds. Although some findings proved that culture might be an important factor in daylight perception, the study results did not provide strong evidence of a cultural background influence on daylight perception. However, the number of participants in this study (N=193) was limited, and this unique topic requires additional research with larger sample size.
... Clustering method, as an unsupervised learning, was adopted to find representative energy consumption patterns in the seven classrooms and to simplify the complicated aspects of energy usage. In this study, the kmeans and self-organizing map (SOM) algorithms were chosen as the frequently-used clustering methods [19,20,[23][24][25]29,31,[35][36][37][38][39][40][41][42]. The IBM SPSS Statistics program was used for the k-means clustering, and a toolbox in the MATLAB environment was used for the SOM clustering. ...
Article
Existing building energy ratings are typically derived with the annual average energy consumption of the buildings. This approach may be appropriate for formulating community-level energy strategy at the macro level, but it cannot be directly linked to occupant behavior for energy savings at the micro level. In light of this, this study aimed to propose a novel process model for developing a scalable room-level energy benchmark using real-time bigdata, which focused on identifying representative energy usage patterns and encouraging occupant behavior change for energy savings. When creating a scalable room-level energy benchmark, three views were taken into account: (i) space unit as perceived by occupants, for which space-specific energy usage datasets were classified based on space attributes; (ii) time unit to which occupants can respond simultaneously, for which hourly energy usage datasets were used; and (iii) equipment unit to which occupants can precisely respond, for which energy usage datasets by different types of electrical installation and appliance were utilized. Based on the scalable room-level energy benchmark, the main findings can be summarized: (i) five representative energy usage patterns were identified using k-means clustering method; (ii) the year-round distributions of the five representative patterns were investigated by month and weekday; and (iii) the annual average variance (or uncertainty) of the room-level scalable energy benchmark was determined to be 19.6%. By providing spatio-temporal information on energy usage patterns in real time, it is expected that occupant behavior change can be voluntarily encouraged to save energy in buildings using the proposed approach.
... To investigate the correlation between occupant behaviour and building energy consumption, we performed methods used for statistical analysis that has been widely employed in several domains owing to their relatively high prediction accuracy [3,45]. In some rooms, the AC opening rate was significantly low throughout the year. ...
Article
Over the past two decades, the electricity consumption of residential air conditioning (AC) has increased by 13.4 times. Moreover, carbon emissions from building operations remain high. Occupant behaviour related to AC operation is the most significant factor affecting the energy consumption of residential ACs. Hence, in this study, occupant behaviour towards using air conditioning in high-rise residential buildings was analysed to provide suggestions for increasing the AC usage efficiency. Annual data on the AC power consumption, indoor temperature, and humidity of typical high-rise residential buildings in hot summer and cold winter zones in China were collected. Five target parameters—AC opening rate, AC standby rate, start–stop time, start–stop temperature, and stable temperature—were statistically analysed. The results reveal that the AC start–stop time follows a normal distribution. The start and stop times of the AC during the cooling period are 22:32 and 7:05 the next day, respectively, whereas during the heating periods these values were 20:42 and 23:02 or 7:56 the next day, respectively. The standard deviation of the time distribution of AC starting or stopping at night is larger than that during the day. The AC start-up temperature followed an exponential distribution. The indoor temperature during the cooling and heating periods is more than 29.42 °C and less than 11.08 °C, respectively. The opening rate of the room AC is 80%.
... On the other hand, few studies have explicitly considered periods of vacancy and its impact on energy use in the buildings [25][26][27][28]. 1 Some of the detailed studies of occupancy indirectly reveal vacancy information when they report zero Wi-Fi counts or no activity from occupancy sensors. Although a number of studies have documented occupancy patterns in diverse commercial buildings revealing significant periods of zero occupancy (weekends and holiday periods were ignored), the number of vacant hours and duration were not treated with similar detail [15,26,[29][30][31][32][33][34][35][36]. Nonetheless, commercial buildings are vacant a significant number of hours each year and vacant long enough to justify special energy-saving strategies such as in the present study. ...
Article
Full-text available
Recent events have forced building managers to examine energy use during vacant periods and revealed miscellaneous electrical loads (MELs) as an opportunity for savings. This paper addresses a key step in unlocking these savings, specifically the reliable identification of when a building is vacant. A Vacancy Inference Engine (VIE), using sensor fusion, was developed to identify vacant periods based on outputs from common sensors, historical building vacancy patterns, and expert knowledge. The VIE calculates the confidence that a building is vacant, allowing building managers to balance the capture of energy savings with the possibility of complaints due to powering down MELs. The VIE has the advantage over logistic regression and other models in that it does not require a full set of ground truth for the training process. The VIE successfully predicted vacancy in an office building using input data streams of instantaneous electricity demand, indoor carbon dioxide concentrations, and the number of active Wi-Fi connections. The VIE’s ability to predict vacancy was compared to that of logistic regression using a metric based on the Complaint Opportunity Rate and found to be nearly identical (0.94 versus 0.95, respectively).
... Although researchers have worked on predicting the occupancy effect (e.g., the number of occupants present, interaction trends with building systems) [30,[39][40][41][42][43][44][45][46][47][48][49], the accuracy is still unsatisfactory. Some relevant factors are still undermined, and part of them are closely related to the type of occupants in the building, affecting comfort perception and more frequently triggering their interaction with the building. ...
Article
Full-text available
Visual comfort can affect building occupants’ behaviour, productivity and health. It is highly dependent on the occupant and how they perform a task indoors. In that regard, an occupant centred approach is more suitable for evaluating the lighting perception of the indoor environment. Nevertheless, the process of rating and estimating the visual comfort makes a limited distinction between physiological differences (e.g., ageing eye, light sensitivity), field of view, and personal preferences, which have been proven to influence the occupants’ lighting needs to complete their tasks. Such features were not considered while establishing the visually comfortable conditions; perhaps due to the challenge of coupling the assumptions made during building design to the performance indicators monitored during building operation. This work focuses on reviewing literature findings on how the common design approach deviates from real building performance, particularly failing to prevent visual disturbances that can trigger the inefficient operation of building systems. Additionally, it is highlighted that redesigned visual comfort assessment methods and metrics are required to bridge the gap between the lighting environment ratings computed and surveyed. One possibility is to consider such physiological features that induce lighting experiences. Finally, it was deducted that it is important to target the occupants’ eye response to calibrate limit thresholds, propose occupant profiling, and that it is convenient to continuously monitor the occupants’ perception of indoor lighting conditions.
... To answer the research questions, a data mining approach was followed. Data mining is a relatively recent field of computer science and consists of discovering patterns from large and unstructured datasets, involving the areas of pattern recognition, machine learning, artificial intelligence, cloud architecture, and data visualization (Liang, Hong, and Shen 2016). Some methodologies exist to solve data mining problems, such as: CRISP-DM (CRoss Industry Standard Process for Data Mining), SEMMA (Sample, Explore, Modify, Model, and Assess), and KDD (Knowledge Discovery in Databases) (KDnuggets™ n.d.). ...
Article
This research proposes an investigation on the problem of detecting and predicting occupancy by using solely readily available electricity consumption data, obtained from smart metres. The following research questions are defined: (1) Is it possible to predict occupancy by using solely electricity consumption data?; (2) Is it possible to use a single classification model to monitor occupancy in multiple households? The findings show that an occupancy detection accuracy of up to 92% can be achieved by using solely electricity consumption data. The problem of generalizing the classification model, i.e. using a single classification model to monitor occupancy in multiple households, is also addressed. It is found that an occupancy detection accuracy of up to 83% is achievable in this case. Regarding occupancy prediction, occupancy in multiple households with an accuracy of up to 75% is obtained, by using solely electricity consumption data. For both occupancy monitoring and prediction, it is found that households with a low level of occupancy can benefit more from these systems.
... productivity, mental health, and overall well-being are as important as energy reduction in the design and operation of intelligent and sustainable buildings. Various studies have established that occupant behavioral change has significant influence on energy consumption [2]- [4]. ...
Conference Paper
Smart buildings focus on providing optimal comfort for the occupant with reduced energy consumption. Better occupant prediction and behavior analysis can significantly reduce building energy usage. Human being is an important parameter in the building control process and his comfort is paramount. Therefore, occupant modeling is critical in improving building efficiency while maintaining indoor comfort. Although, there are many different algorithms developed for occupancy modeling, the Markov chain, and its derivative models are extensively used because of their simplicity, flexibility, and prediction efficiency. In this context, this paper proposes a state-of-the-art review focused on the Markov chain and its derivative models for occupant modeling.
... Another approach is presented by Shih [6], where programmable pan-tilt-zoom cameras detect and track occupants of the room. Liang et al. [7] create a situation where sensors that track people entering and exiting a building are being used. However, with apartment or office buildings, the direct method poses some problems; there are significant privacy concerns, as these systems do not predict occupancy; they detect occupants at that moment in time. ...
Conference Paper
Full-text available
Occupancy detection is crucial when trying to lower the emissions that a building produces. Some buildings are equipped with motion sensors or cameras to find how many occupants are in a room. However, this is not entirely accurate as people could be stationary in situations like sitting at a desk or watching television. Using environmental sensors, we can determine if a room is occupied even if the occupants are not moving. When occupants are inside a room, they give off extra CO2 or increase the room's temperature. We can find the small differences in the environmental values used to accurately predict a room's occupancy levels. We use relatively inexpensive IoT sensors that almost every building's HVAC system should have in the near future. We apply K-means clustering with success to predict occupancy levels. Our algorithms can be used in smart thermostats to automatically adjust the room's heat depending on how many occupants are in a room.
... In general, forecast methods can be classified into four categories: (1) There are also models that introduce hybrid concepts. Liang et al. [96] However, several gaps in the research need to be carefully defined and addressed. 586 ...
Article
Full-text available
Indoor environment construction for occupants has high energy consumption; as such, occupancy plays a noteworthy role in the complete life cycle phase of buildings, including design, operation, and retrofitting. In the past few years, building occupancy, which is considered the basis of occupant behavior, has attracted increasing attention from researchers. There are increasing requirements for buildings to be both comfortable and energy efficient; with the development of detection methods and analyzing algorithms, occupancy prediction has become a topic of interest for building automation and energy conservation. Therefore, this article reviews the literature regarding future building occupancy predictions (forecasting). This review is distinguished from occupancy simulation and detection research and focuses on the research purpose, physical routine, and complete methodology of occupancy forecasting. First, the research purposes, including the application field and detailed requirements for occupancy forecasting, are summarized and analyzed. Next, an overall methodology of occupancy forecasting, including data acquisition, modeling techniques, and evaluation, is discussed in terms of issues affecting prediction performance. Finally, the current challenges and perspectives of occupancy forecasting are highlighted, considering the insights of natural characteristics, on-site implementation, valid dataset sharing, and research techniques. Overall, accurate and robust future occupancy predictions will help to improve building system operations and energy conservation.
... Liang et al. [126] used data mining technique for occupant prediction in offices. Occupancy schedules pattern is observed using cluster analysis, learning the behavior rules using decision trees and prediction is done based on the scheduled rules. ...
Article
Full-text available
Energy use in buildings is increasing to provide optimal comfort for the occupants. People spend 90% of their life in buildings. Therefore, indoor environment quality and comfort management have crucial role in maintaining occupants health and productivity. Reducing the energy consumption for optimal comfort management is important to minimize the CO2 emission and global warming in the building sector. Literature states that by analyzing the occupant behavior and by predicting number of occupants, it is possible to control and reduce the energy consumption. A detailed comprehensive review is carried out in this paper to analyze the existing methodologies to model occupant behavior and prediction with respect to comfort and energy consumption considering developments, challenges in modeling occupant behavior and prediction maintaining the comfort.
... The difference between energy consumption in resulting clusters was referred to different occupant behavior and actions. Furthermore, occupancy scheduling [13,14] and indoor air quality indexing [15] have also been investigated using the occupant related data set. ...
Article
In this study, a data mining-based framework is developed to provide practical insight into occupants’ energy behavior in residential buildings. Occupant Activity Indicator (OAI) and Residential Energy Intensity Indicator (REII) are introduced as two new indicators. The proposed data-mining framework aims to evaluate the energy-related behavior of the buildings’ residents based on the difference between the target REII and actual REII. The dissimilarity, which is found between the target and the actual REII, is used to identify when occupants behave wasteful/efficient in different zones of the building. The practicality of the proposed methodology is demonstrated by applying it to a three-bedroom apartment. As a result, the time and zones for potential energy savings in the considered building are identified. The extracted knowledge identifies the zones that occupants need to be more cautious about their energy consumption. Besides, location and time for the best and the worst energy-related behavior by the building's occupants are defined. Furthermore, the variations of occupants' energy-related behavior in the apartment are identified by the time of day, day of the week, and months.
... In most investigations to date, just domestic energy consumption data have been utilized for data mining studies of households' standard behavior patterns [16]. This method is intended to overcome the weaknesses of the previously mentioned techniques, especially when managing enormous data streams, by suggesting reliable occupant behavior models with the great potential for a quick examination and better replication [30,59,96]. In the opposite sense to the data mining approaches, agent-based modeling (ABM), a simulation-based approach that is usually built on real buildings, has been initiated within the occupant-centered virtual environment. ...
Article
Full-text available
Energy consumption in buildings depends on several physical factors, including its physical characteristics, various building services systems/appliances used, and the outdoor environment. However, the occupants’ behavior that determines and regulates the building energy conservation also plays a critical role in the buildings’ energy performance. Compared to physical factors, there are relatively fewer studies on occupants’ behavior. This paper reports a systematic review analysis on occupant behavior and different modeling approaches using the Scopus and Science Direct databases. The comprehensive review study focuses on the current understanding of occupant behavior, existing behavior modeling approaches and their limitations, and key influential parameters on building energy conservation. Finally, the study identifies six significant research gaps for future development: occupant-centered space layout deployment; occupant behavior must be understood in the context of developing or low-income economies; there are higher numbers of quantitative occupant behavior studies than qualitative; the extensive use of survey or secondary data and the lack of real data used in model validation; behavior studies are required for diverse categories building; building information modeling (BIM) integration with existing occupant behavior modeling/simulation. These checklists of the gaps are beneficial for researchers to accomplish the future research in the built environment.
... Cluster analysis is one of the most common data mining methods used widely for the identification of energy-related behavioural patterns in buildings (D'Oca & Hong, 2015;Liang et al., 2016;Pan et al., 2017;Ren et al., 2015). There are two broad categories in which traditional clustering methods fall: relocation (such as k-means and expectationmaximization) and hierarchical. ...
Article
The understanding of energy-related occupant behaviour and its better reproduction in building energy analysis has recently become a primary field of interest. The combination of computing ease and the availability of big streams of high resolution building performance data enhance the ability to study this behaviour. In this context, data mining methods are increasingly employed. The aim of this study is to propose a data mining methodology for the characterization of heating behaviour in residential buildings. The methodology consists of two multivariate statistical analysis methods, namely Principal Component Analysis (PCA) followed by cluster analysis. The methods were applied on monitored gas consumption data of five dwellings in Italy. Findings support literature indicating that people heat their homes in different ways. It was found that households do not always follow a different heating schedule on weekends, have very different temperature preferences and operate the heating system at different hours during the day. In fact, some households may change heating practices over the heating season. The highlight of the proposed methodology is the insightful and simple way that PCA can extract succinct information about the heating behaviour of the user.
... While centroids of those clusters were used to create occupancy profiles, which could be used as an input to the building performance simulation. A similar study, but on a larger scale of the whole building, was done by Liang et al. [23] where k-means clustering was used to find the occupancy patterns, while profiles were created using the Decision Tree supervised method. Authors have managed to create the occupancy profiles and predict their occurrence using time of day, weekday and season. ...
Article
Full-text available
Climate change and technological development are pushing buildings to become more sophisticated. The installation of modern building automation systems, smart meters, and IoT devices is increasing the amount of available building operational data. The common term for this kind of building is a smart building but producing large amounts of raw data does not automatically offer intelligence that would offer new insights to the building’s operation. Smart meters are mainly used only for tracking the energy or water consumption in the building. On the other hand, building occupancy is usually not monitored in the building at all, even though it is one of the main influencing factors of consumption and indoor climate parameters. This paper is bringing the true smart building closer to practice by using machine learning methods with sub-metered electricity and water consumptions to predict the building occupancy. In the first approach, the number of occupants was predicted in an office floor using a supervised data mining method Random Forest. The model performed the best with the use of all predictors available, while from individual predictors, the sub-metered electricity used for office equipment showed the best performance. Since the supervised approach requires the continuous long-term collection of ground truth reference data (between one to three months, by this study), an unsupervised data mining method k-means clustering was tested in the second approach. With the unsupervised method, this study was able to predict the level of occupancy in a day as zero, medium, or high in a case study office floor using the equipment electricity consumption.
... Yoshino et al. [2] reported that in the past, most of the research works were focused on understanding the influence of the physical factors (weather, building envelope, and equipment), and more studies must be performed on studying the influence of human factors (occupant behavior, operation & maintenance). In recent years, the data mining (DM) and data-driven tools are well-acknowledged for pattern recognition [18,25], for exploring the role of occupant behaviour in achieving energy savings in residential [12,26] and institutional buildings [27], occupancy prediction [18,28] and energy demand modelling [29]. Yu et al. [25] used the end-use load data to explore occupant behavior's influence on building energy use. ...
Article
Occupant's schedules and their energy-use behavior are substantial inputs for building energy simulations and energy management in buildings. In practice, most of the research studies consider default occupant schedules from the standards. The temporal variations associated with occupancy is often missed out, leading to uncertainties in simulation results. This study aims to address two research problems in terms of occupancy: 1) upon the availability of the data, how to systematically extract the different occupant schedules, 2) when the occupancy data is not available, what are the other commonly logged parameters (such as plug load, lighting energy consumption, indoor carbon dioxide (CO2) concentration, and indoor relative humidity data) that shall be used to represent the occupancy in buildings. Regarding the first objective, a generic data-driven framework with the combination of shape-based clustering and change-point detection method is proposed to extract the distinct occupancy in residential buildings in terms of occupant activity schedule and presence probability. To demonstrate the outcomes of the framework, it was applied to the dataset collected from eight apartments located in Lyon, France. The results show the existence of different occupant patterns in buildings with respect to day of the week and season of the year. To achieve the second objective, linear and logistic regression models were developed to represent the occupant activity level and occupant presence/absence state, respectively. The linear regression model results show that among the examined variables, the lighting, and plug load consumption data along with the hour of the day show better prediction results in terms of adjusted R2 and mean absolute percentage error. For the occupant presence/absence state, the logistic regression model developed using CO2 concentration and plug load energy consumption dataset shows better results in misclassification error, confusion matrix, and receiver operating characteristic curve.
... Currently, the energy consumption of the building sector mainly comes from commercial and residential buildings. Large-scale commercial buildings are reported to have a high energy consumption, which can be up to 300 kW h/m 2 , 5 to 15 times of that in residential buildings (Liang et al., 2016). In the commercial sector, healthcare buildings are particularly energy-intensive due to their constant need for power supply and strict requirements for air quality and disease control (García-Sanz-Calcedo et al., 2019; Bawaneh * Corresponding author. ...
Article
Full-text available
Healthcare buildings are characterized by complex energy systems and high energy usage, therefore serving as the key areas for achieving energy conservation goals in the building sector. An accurate load prediction of hospital energy consumption is of paramount importance to a successful healthcare building energy management. In this study, eight machine learning models of single learning and ensemble learning were developed for predicting healthcare facilities’ energy consumption. To validate the performance of the proposed model, an experiment was conducted on a general hospital in Shanghai, China. It was found that the two ensemble models, Extreme Gradient Boosting (XGBoost) model and Random Forest (RF) model, outperformed single models in daily electrical load prediction. A further comparison between models trained with daily and weekly temporal resolution electrical data shows that it is more likely to achieve higher accuracy with finer time granularity. Through feature importance analysis, the most influential features under the daily and weekly electrical load prediction were identified. Based on the prediction results, it is expected that hospital facility managers will be able to conveniently assess the expected energy usage of their hospitals with the machine learning models.
... Besides energy consumption analysis, analysis of the impact of occupants' behavior has been largely overlooked in building energy performance analysis [8]. Examples include a study that analyzed oneyear observed data of an office building in Philadelphia to discover the occupancy schedule patterns and extrapolate the occupancy schedule [9], and a framework combining statistical analysis with 2 data-mining techniques, cluster analysis and association rules mining to identify valid window opening/closing operational patterns in the measured data of 16 offices in a natural ventilated office building in VOLUME 7, 2019 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ ...
Article
Full-text available
With the development of modern information technologies and more frequent utilization of information systems to operation and maintenance (O&M) management, a great amount of O&M data are collected nowadays. However, because of the large volume and poor quality, as well as a lack of effective data analysis techniques, these data are rarely analyzed and translated into useful knowledge for O&M decisions. This study presents a data model, which is named as datacube with multi-dimensional and unrestrained characteristics, for these data to better support data mining algorithms. The model organizes all the different data in both relational database and in the memories and is able to support analysis-requirements-oriented data extractions. Based on this datacube, an O&M data mining approach is proposed with procedures of data preparation, data clustering and data mining. The proposed datacube-based data mining approach was applied to the Kunming Chang Shui international airport terminal. More than 7 years on-site repairing data were used for data mining and the outcomes verified the model and the approach to be feasible and valuable for improving O&M management.
... Another work from [7] demonstrated a case study on office buildings using de- 35 cision tree C4.5 and k-means clustering with different distance measures. The inclusion of features related to season, daylight saving time, and weekday allowed them to reach 80% of accuracy. ...
Article
Full-text available
Heating, ventilation and air conditioning systems represent considerable potential for energy savings, which can be realized through intelligent occupancy-centered control strategies. In this work, both supervised and unsupervised algorithms to forecast occupancy are proposed with the highest accuracies of 98.3% and 97.6% respectively. Building on their output, a rule-based air conditioning scheduling technique is developed. As an example, a potential of 15.4% of energy savings is calculated using a dataset collected in a mid-size (4000 m²) building in Portugal.
Article
EMF has a variety of biological impacts and has an impact on the metabolic process in the human body. Antenna towers, anechoic chambers, and other sources can all produce this. Some of the human populations live very close to the EMF-emitting antenna towers. We can make humans aware of the EMF radiation and protect from diseases if there is a proper method to anticipate the EMF radiation of antennas installed in different places. For the study of telecom data and EMF emission, many machine learning and deep learning techniques have been developed in recent years. Predictive analytics played a bigger part in this. For prediction, it comprises advanced statistics, modeling and more machine learning methodologies. However, the appropriate hyper parameters must be established for the model’s effective prediction, but this cannot be guaranteed in a dynamic environment where the data is always changing. The learning model’s performance improves when these parameters are optimized. The goal of this study is to use the Telecom dataset to create a novel hybrid deep learning model for forecasting the trend of EMF radiations. The patterns were first discovered using Artificial Neural Networks (ANN) and Multilayer Perceptron (MLP) combined with the Particle Swarm Optimization method (PSO). Later to boost its performance the hybrid approach (MLP-RFD-PSO) was developed and 98.8% accuracy was achieved.
Article
Occupancy patterns are key elements in improving our understanding of building energy usage and implementing demand-oriented strategies to reduce energy consumption. Occupancy pattern is changed over time driving by multiple factors, but there are lacks of research to reveal its temporal change. We analyzed occupancy pattern of residential building based on time-use survey (TUS) data, and compared data from the years 2008 and 2017, to reveal temporal changes in occupancy patterns and to explore the influencing factors of behavioral change. In this study, occupancy patterns on workdays and weekends were distinguished and categorized. Clustering analysis, function fitting, and correlation analysis were used to identify typical occupancy profile of residential building in 2008. Occupancy duration, influencing factors, and distribution curves were compared between 2008 and 2017 to reveal changes and to clarify personnel-related factors affecting occupancy patterns. The length of time Chinese residents spent in their homes was revealed to have increased from 18.35 to 20.31 h on workdays and decreased from 20.73 to 20.01 h on weekends. Further, residents shifted from the living room to the bedroom, in their homes. The age of family members was positively correlated with the length of time spent at home. Also, education level and employment status were inversely proportional to the length of time spent at home. These results deepen our understanding of occupancy patterns and are of great significance for analyzing future trends and their impacts on energy consumption in residential buildings.
Article
Building occupancy patterns are an important factor in considering the energy efficiency of buildings and a key input for building performance modelling. More specifically, the energy consumption associated with heating, cooling, lighting, and plug load usage depends on the number of occupants in a building. Identifying occupancy patterns and profiles in buildings is a key factor for the optimisation of building operating systems and can potentially reduce the performance gap between the planning stage and the actual energy usage. This study aims to identify the patterns and profiles of the occupants in a selected case study building in England. In this study, occupancy data were collected over 12 months at five minutes intervals. A sensor was used to obtain high accuracy occupancy data compared to previous studies that encountered uncertainties in data collection. A set of clustering analyses was carried out to identify occupancy patterns and profiles in the building. The results of this study identified three different occupancy patterns and profiles as well as four drivers that influenced the occupants in the case study building: the beginning of the academic term, the examination period, the weekday/weekends, and the vacation driver.
Article
The short-term prediction of air-conditioning (AC) energy consumption is a crucial part of building operation optimization and demand-response strategies. However, the AC energy consumption in some type of buildings, can be highly stochastic due to arbitrary occupants' behavior of AC usage, which make it had to predict. Currently, very few prediction models considered the impacts of stochasticity related to occupants’ behavior of AC, which leads to less accuracy. To predict the AC energy consumption under the stochastic energy use scenario, this study proposes a behavior-orientated prediction method for short-term AC energy consumption with a case study on teaching building blocks. Firstly, cluster analysis is employed to identify typical patterns to quantify stochastic AC energy use. Then, the AC usage rate is predicted based on weighted k-Nearest-Neighbors method, which can provide a precise prediction of stochastic AC usage rate. Based on these, Randomforest method is used to develop a basic prediction model of AC energy consumption. The importance of each variable is also evaluated. Finally, an enhanced part of prediction is implemented by support vector machine to achieve higher accuracy under stochastic AC energy use scenarios. The results show that both proposed models can accurately predict the AC energy consumption, while the proposed enhanced model can reap significant accuracy improvement under the more stochastic scenario. This study provides a new approach to predict AC energy consumption in buildings with stochastic AC energy use, which can match the real-time amount with good accuracy. The proposed model can provide an accuracy reference for the demand response strategies and optimization of AC systems operation.
Article
Fusion's light steel frame system was chosen as a viable lightweight steel solution that would equal the pace of construction of a timber frame, eliminate fire hazards during and after construction, reduce waste, provide greater durability, versatility, safety, quicker construction timeframes, cost-effectiveness, and space maximisation, and minimise building shrinkage and movement for sustainability, thermal, reliable, and acoustic construction by simplifying construction with recycled materials, according to industry recommendations. This research provides information and recommendations for using light steel frames in residential construction, such as single-family homes and apartments. Galvanized cold-formed steel sections that may be fitted as prefabricated panels are the main structural elements of light steel framing systems. This research focuses on design elements that are governed by the building code (England and Wales). It also looks into the construction and characteristics of light steel frames, as well as how they interact with other materials and components.
Article
Owing to the convenient installation, low energy consumption, wide applicability, and flexible operation of room air conditioner (RAC), high application potential can be achieved in terms of load dispatch or demand response by aggregating numerous single RACs’ loads. The diverse behaviors of single occupants using RACs caused by different individual demands is one of the main factors leading to deviations in short-term (e.g., hourly) energy consumption prediction. In addition, owing to the lack of suitable quantitative indicators and subjective data, previous studies are rarely focused on the group behaviors of RACs, which is more appropriate for obtaining specific regional energy consumption. Aiming at optimizing the regional power grid load fluctuations in residential building, this paper proposes a framework for predicting short-term energy consumption based on RAC group characteristics by randomly selecting approximately 2000 samples from a cloud data platform. First, this study constructs multi-dimensional group features and based on the clustering method identifies three typical groups of RACs: i) Cluster a: continuous operation in summer and high energy consumption behavior; ii) Cluster b: annual intermittent operation and high energy consumption behavior; iii) Cluster c: mainly continuous operation in summer and energy-saving behavior. Second, the study generates five different aggregation sizes of RACs and integrates them with feature selection and random forest algorithms, to predict the hourly energy consumption of a residential RAC group. High precision achievement is shown by indicator indexes, such as the coefficient of determination (R²) reaches 90% and the relative mean absolute percentage error (MAPE) is basically within 15%. Finally, by adjusting the group behavior of the load aggregator scenarios of the above three RAC groups in this method, peak shaving capacity of approximately 35% is realized on a typical summer day.
Article
Over the past two decades, machine learning (ML) has elicited increasing attention in building energy management (BEM) research. However, the boundary of the ML-BEM research has not been clearly defined, and no thorough review of ML applications in BEM during the whole building life-cycle has been published. This study aims to address this gap by reviewing the ML-BEM papers to ascertain the status of this research area and identify future research directions. An integrated framework of ML-BEM, composed of four layers and a series of driving factors, is proposed. Then, based on the hype cycle model, this paper analyzes the current development status of ML-BEM and tries to predict its future development trend. Finally, five research directions are discussed: (1) the behavioral impact on BEM, (2) the integration management of renewable energy, (3) security concerns of ML-BEM, (4) extension to other building life-cycle phases, and (5) the focus on fault detection and diagnosis. The findings of this study are believed to provide useful references for future research on ML-BEM.
Article
Full-text available
The use of digital technologies such as Internet of Things (IoT) and smart meters induces a huge data stack in facility management (FM). However, the use of data analysis techniques has remained limited to converting available data into information within activities performed in FM. In this context, business intelligence and analytics (BI&A) techniques can provide a promising opportunity to elaborate facility performance and discover measurable new FM key performance indicators (KPIs) since existing KPIs are too crude to discover actual performance of facilities. Beside this, there is no comprehensive study that covers BI&A activities and their importance level for healthcare FM. Therefore, this study aims to identify healthcare FM KPIs and their importance levels for the Turkish healthcare FM industry with the use of the AHP integrated PROMETHEE method. As a result of the study, ninety-eight healthcare FM KPIs, which are categorized under six categories, were found. The comparison of the findings with the literature review showed that there are some similarities and differences between countries’ FM healthcare ranks. Within this context, differences between countries can be related to the consideration of limited FM KPIs in the existing studies. Therefore, the proposed FM KPIs under this study are very comprehensive and detailed to measure and discover healthcare FM performance. This study can help professionals perform more detailed building performance analyses in FM. Additionally, findings from this study will pave the way for new developments in FM software and effective use of available data to enable lean FM processes in healthcare facilities.
Article
Accurate occupancy information can help in optimizing the operation of building systems. To obtain this information, previous studies suggested using WiFi connection counts due to their strong correlation with occupancy counts. However, validating this correlation and investigating its variation have remained limited due to challenges regarding the collection of ground-truth data. Therefore, many studies suggested a single (fixed) value as the conversion factor of WiFi connection counts to actual occupancy counts based on short-term ground-truth data. This study addressed this gap by proposing a method for investigating the correlation between WiFi connection counts and actual building occupancy over a longer duration using continuous ground-truth data collected from camera-based occupancy counters. The proposed method focused on (i) identifying the influential features on this correlation and their effectiveness, as well as (ii) developing models to estimate real-time occupancy counts and to forecast day-ahead occupancy counts. To validate the proposed method, it was applied in a library building in Montreal, Canada with data collected between January and March 2020. Results showed time-related features including Hour of the day and Day of the week, as well as occupancy level influenced the correlation between Wifi and occupancy counts. Furthermore, the proposed models successfully estimated real-time occupancy counts with an average accuracy (R²) of 0.96 for weekdays and 0.98 for weekends, while day-ahead occupancy forecasting models had an average accuracy (R²) of 0.92 for weekdays and 0.82 for weekends. These findings provided a proof-of-concept for the proposed methodology, demonstrated the potential of using WiFi connection count for estimating/forecasting occupancy counts, and most importantly highlighted the important considerations that need to be addressed when using WiFi connection counts as a proxy for occupancy to optimize building operation.
Article
Occupancy plays a vital role in optimizing the operation of building service systems. This study proposed a novel model for predicting short-term building occupancy. In the model, the autocorrelation function (ACF) and partial autocorrelation function (PACF) are used to determine the most relevant occupancy status. Then the multi-order transition probability (MOTP) is established and integrated with the deep forest (DF) for occupancy prediction. To assess the effectiveness of the proposed MOTP-DF model, a validation experiment was conducted in an office room to compare its prediction performance with conventional methods, including Markov chain, decision tree (DT), and support vector machine (SVM) models. The results show that the proposed model can track occupancy change with higher accuracy and fewer fluctuations. Moreover, it improves the prediction accuracy by 6.3-10%, 4.6-8.3%, and 4.8-8.3% over the Markov chain, DT, and SVM models, respectively. A further evaluation indicates that MOTP can quantitatively incorporate occupancy’s status interdependency and stochastic characteristics into its prediction. Compared with the DT and SVM models, the DF model with deep and ensemble structures could benefit more from the integration with MOTP due to its higher robustness. This study also found that a proper selection of transition probability orders, forest algorithms, and corresponding maximum tree depths can further enhance the prediction accuracy of the proposed model.
Article
Full-text available
Building occupancy is the basis for building energy simulations, operations, and management. With the increasing need for energy conservation and the occupant-centric service of building energy systems, occupancy forecasting has become an essential input for simulations. These applications include model predictive control and demand response, with the potential to optimize the use of renewable energy sources. Based on recent research, occupancy forecasting tends to be based on occupancy data of the previous time step or continuous lagged dependent variables. Previous analysis demonstrated that building occupancy is innately temporal and sequential with seasonal features, which may be instructive for forecast research. This study proposes a temporal-sequential (TS) analysis and machine learning integrated approach for occupancy forecasting. Using hourly occupant data from 16 different buildings, we demonstrate that the proposed temporal-sequential analysis using a 1-week seasonal period with an artificial neural network structure (TS-week-ANN) outperforms other baseline methods.
Article
There has been an increasing interest in addressing and discovering the factors influencing the households’ load profiles instead of their end-use energy demand. The rationale behind this tendency is to provide households with load shifting recommendations and flatten the load profiles by making use of the knowledge obtained from these temporal and contextual determinants. Methodologies connecting households’ activities and presence to load profiles are often under-investigated, and the flexibility in the presence and activity routines of households throughout a long period is ignored. In this study, a data-driven framework is developed to extract households’ daily occupancy patterns throughout a year, determine the regular high- and low-energy consumption periods, and discover influencing activity factors of energy consumption within the obtained periods. The purpose of this study is to provide households with customized load-shifting and energy-saving suggestions based on their specific traits and routines. The results suggest that the distribution of occupancy patterns between seasons and weekdays varies considerably among different households. It is further recognized that days with similar occupancy patterns can have nearly similar peak timings in different apartments. The developed framework is generic and can be generalized to different households with different presence and activity routines.
Article
Adapting building systems' operation to occupancy variations can provide significant energy savings, but this is typically constrained by the unavailability of occupancy information. Although several technologies have been introduced to integrate real-time occupancy information in heating, ventilation and air-conditioning (HVAC) systems operations (e.g., CO2 monitoring), the logistical and cost issues associated with deploying these technologies remained a key issue. Previous studies proposed using WiFi counts as a proxy for occupancy in building operations and showed a strong correlation between occupancy and WiFi counts in several building types. However, the difficulty of integrating real-time WiFi traffic data in building automation systems hinders wide-scale deployment of this approach. To this end, this study proposes a framework for extracting occupancy indicators from WiFi traffic data. The proposed framework utilizes several machine learning algorithms and statistical analysis methods to predict patterns of building occupancy as well as to identify peak occupancy time and earliest/latest arrival and departure times. To validate the proposed framework, it was implemented in a case-study using data collected from an academic building in Montreal, Canada between January and March 2020. Results revealed that the proposed models could successfully predict weekly building occupancy patterns, with an average accuracy (R²D) of 0.98 for weekdays and 0.81 for weekends. Furthermore, the analysis identified peak occupancy timing, as well as arrival and departure times variations between different zones. These findings provided a proof-of-concept for the proposed framework and demonstrated its potential to provide actionable information to modify the sequences of operation of building systems based on buildings’ unique occupancy patterns.
Article
Association rule mining has been widely utilized for extracting hidden operation patterns from huge volumes of building operational data. It usually generates a large number of association rules. However, most of them cannot reveal useful patterns. It is still unclear about how to develop effective methods to extract useful association rules. The major barrier is that characteristics of association rules mined from building operational data are still unknown. To reveal the characteristics, this study analyzes 101,787 association rules mined from the historical operational data of a typical chiller plant. The association rules are classified into useful ones and useless ones based on domain knowledge. Three insights are revealed successfully. Firstly, the useful association rules are related to component/sensor faults, control strategies, abnormal operation patterns, and normal operation patterns. And the useless association rules result from lacking physical meanings, insufficient information, and transient operation patterns. Secondly, the useful association rules only account for a very small proportion (4.64%) of all the association rules. Thirdly, three widely-used statistic indexes, i.e., support, confidence and lift, are ineffective in distinguishing between the useful and useless association rules actually in this field. Based on the three insights, five future research directions are proposed.
Article
Implementing machine-learning models in real applications is crucial to achieving intelligent building control and high energy efficiency. Over the past few decades, numerous studies have attempted to explore the application of machine-learning models to building energy efficiency. However, these studies have focused on analyzing the technical feasibility and superiority of machine learning algorithms for fitting building energy-related data and have not considered methods of implementing machine learning technology in building energy efficiency applications. Therefore, this review aims to summarize the current practical issues involved in applying machine-learning models to building energy efficiency by systematically analyzing existing research findings and limitations. The paper first reviews the application status of machine learning-based building energy efficiency research by analyzing the model implementation process and summarizing the main uses of the technology in the overall building energy management life cycle. The paper then elaborates on the causes of, influences on, and potential solutions for practical issues found in the implementation and promotion of machine learning-based building energy efficiency measures. Finally, this paper discusses valuable future machine learning-based building energy efficiency research directions with regard to technology opportunity discovery, data governance, feature engineering, generalizability test, technology diffusion, and knowledge sharing. This paper will provide building researchers and practitioners with a better understanding of the current application statuses of and potential research directions for machine learning models in building energy efficiency.
Article
On the basis of convolution neural network, deep learning algorithm can make the convolution layer convolute the input image to complete the hierarchical expression of feature information, which makes pattern recognition more simple and accurate. Now, in the theory of multimodal discourse analysis, the nonverbal features in communication are studied as a symbol system similar to language. In this paper, the author analyzes the deep learning complexity and multimodal target recognition application in English education system. Multimodal teaching gradually has its practical significance in the process of rich teaching resources. The large-scale application of multimedia technology in college English classroom is conducive to the construction of a real language environment. The simulation results show that the multi-layer and one-dimensional convolution structure of the product neural network can effectively complete many natural language problems, including the tagging of lexical and semantic roles, and thus effectively improve the accuracy of natural language processing. Multimodal teaching mode helps to memorize vocabulary images more deeply. 84% of students think that multi-modal teaching mode is closer to life. Meanwhile, multimedia teaching display is more acceptable. College English teachers should renew their teaching concepts and adapt themselves to the new teaching mode.
Article
Building performance optimization is a powerful technique that assists designers in identifying optimal design solutions. However, occupant-related uncertainty is considered a major challenge that can influence the credibility of the selected optimal design. This paper proposes an occupant-centric method for window and shading design optimization in office buildings that assists designers in evaluating and handling the impact of occupant-related assumptions. The method consists of three major steps: generating occupant scenarios, conducting occupant-centric mathematical optimization, and deriving rules for selecting optimal design parameters using decision trees. An office building model with numerous possible window-related parameters was used to demonstrate the proposed optimization process. Sixty-four different occupant scenarios were considered. The results of the optimization process demonstrated that occupant-related assumptions are crucial in determining optimal window and shading design, as different occupant scenarios led to different optimal solutions. The study findings also provided insights to building designers regarding classifying and selecting window and shading design parameters under varying occupant scenarios.
Article
The building sector is responsible for 32% of global energy consumption and 19% of all energy-related greenhouse gas emissions. The urgent requirement for energy conservation and greenhouse gas emission reduction in the building sector has been recognised at the highest level of governments around the world. One potential solution, which has yet to be critically considered, is the application of Building Information Modeling (BIM) to overcome building energy performance gap (BEPG), defined as the discrepancy between the designed and actual energy consumption in buildings. This study performs a systematic and comprehensive literature review to identify the specific causes of the BEPG, and then analyses the application of BIM for addressing the BEPG. A life-cycle BIM engaged framework was developed, including the function of “information exchange”, “design review”, “energy-related quality control”, “life-cycle commissioning”, and “real-time operation and maintenance management”. It is expected that the proposed framework will assist researchers and practitioners better understand application of BIM to systematically improve building energy performance.
Thesis
Full-text available
Occupancy detection plays an important role in many smart buildings such as reducing building energy usage by controlling heating, ventilation and air conditioning (HVAC) systems, monitoring systems and managing lighting systems, tracking people in hospitals for medical issues, advertising to people in malls, and to search and rescue missions. The global positioning system (GPS) is used most widely as a localization system but highly inaccurate for indoor applications. The indoor environment is difficult to handle because along with the loss of signals, privacy is a major concern. Indoor tracking has many aspects in common with sensor localization in Wireless Sensor Networks (WSN). The contribution of this work is the demonstration of a non-intrusive approach to detect an occupancy in a building using wireless sensor networks to detect energy from cell phones in a secure facility and perform indoor localization based on the minimum mean square error (MMSE). To estimate the occupancy, the detected cellular signals information such as signal amplitude, frequency, power, and detection time is sent to a fusion server, matched the detected signals by time and channel information, performed localization to estimate a location, and finally estimated the occupancy of rooms in a building from the estimated locations.
Article
Owing to economic and technical reasons, campus buildings in south China were not equipped with air-conditioners (AC) for a long time. With the improvement in teaching conditions in south China, ACs are gradually being installed in teaching buildings, leading to soaring energy consumption. The teaching buildings constitute a large part of the total built-up areas on a campus. Due to the stochastic occupancy in teaching buildings, AC energy use is complicated, which is hard to describe quantitatively. Owing to high coupling relationships among the AC usage, indoor temperature, and energy consumption, it is hard to formulate any strategy on energy-savings management. Based on the data collected from an energy monitoring platform at a typical university in south China, typical patterns were proposed using data mining (DM) approaches. There were 6 AC usage patterns, 4 indoor temperature patterns, and 4 energy consumption patterns, all of which could represent complicated AC energy use. To propose precise energy-savings strategies for random AC usage, the coupling relationships among these components were revealed by multiple machine learning (ML) methods, including the decision tree, AdaBoost, and RandomForest. After that, the energy-saving control rules were formulated. As for short-time AC usage, “Turning off as leaving” is an effective way to save energy. The scales of classrooms need to be considered for usage with medium usage hours, while AC set temperature is a critical control parameter for long-time AC usage. These results provide support for more accurate simulation of energy consumption and efficient energy-saving management in teaching buildings.
Article
Full-text available
Overtime is a common phenomenon around the world. Overtime drives both internal heat gains from occupants, lighting and plug-loads, and HVAC operation during overtime periods. Overtime leads to longer occupancy hours and extended operation of building services systems beyond normal working hours, thus overtime impacts total building energy use. Current literature lacks methods to model overtime occupancy because overtime is stochastic in nature and varies by individual occupants and by time. To address this gap in the literature, this study aims to develop a new stochastic model based on the statistical analysis of measured overtime occupancy data from an office building. A binomial distribution is used to represent the total number of occupants working overtime, while an exponential distribution is used to represent the duration of overtime periods. The overtime model is used to generate overtime occupancy schedules as an input to the energy model of a second office building. The measured and simulated cooling energy use during the overtime period is compared in order to validate the overtime model. A hybrid approach to energy model calibration is proposed and tested, which combines ASHRAE Guideline 14 for the calibration of the energy model during normal working hours, and a proposed KS test for the calibration of the energy model during overtime. The developed stochastic overtime model and the hybrid calibration approach can be used in building energy simulations to improve the accuracy of results, and better understand the characteristics of overtime in office buildings.
Article
Full-text available
While most studies focus on energy savings during occupied hours, this paper shows the shocking quantities of energy wasted during non-occupied hours in commercial buildings. At least five detailed energy audits were carried out in the hot and dry climates of Botswana and South Africa. The work shows that more energy is used during non-working hours (56%) than during working hours (44%). This arises largely from occupants’ behaviour of leaving lights and equipment on at the end of the day, and partly due to poor zoning and controls. There is a crying need for building occupants to learn to switch off what they do not use. The golden rule is: “If you don’t need it, don’t use it!” This is the simplest and cheapest lesson with amongst the biggest savings. Apart from the above, the work also contributes to several other fields of scientific research: it helps in development of benchmarks from sub-hourly field data; it contributes apportionment of energy amongst sub-systems of HVAC, lighting and office equipment; it provides a picture from cooling dominated climates, which normally differs from the largely researched heating dominated climates; it contributes to development of diversity profiles necessary for improvement of simulation accuracy.
Conference Paper
Full-text available
Inhabitants' behaviour is a significant factor that influences energy consumption and has been previously incorporated as static activity profiles within simulation for energy control & management. In this paper an agent-based approach to simulate reactive/deliberative group behaviour has been proposed and implemented. It takes into account perceptual, psychological (cognitive), social behavioural elements and domestic context to generate reactive/deliberative behavioural profiles. The Brahms language is used to implement the proposed approach to learn behavioural patterns for energy control and management strategies.
Article
Full-text available
Article
Full-text available
Informational interventions were employed to promote two behaviors relevant for efficient heating of individual offices in a large office building. In two successive winter seasons, interventions were applied during 4-week periods. Short-term effects were assessed weekly, and long-term effects were assessed 1 year after each of the two intervention periods. Improvements were observed in each intervention period, with partial behavior maintenance 1 year later. The changes observed in the individual offices across conditions are suggestive of the program's capacity to correct relapses in earlier proenvironmental behavior.
Article
Full-text available
Behavior of the household members and home characteristics are the major determinants of residential energy use. This study is focussed on energy-related behavioral patterns. It is based on self-reported behaviors of 145 households in Vlaardingen, The Netherlands.Energy-related behavior may not be a separate type of behavior but a contingency for other types of household behaviors, such as recreation, child care, and household chores. Two important aspects of energy contingency are home temperature and ventilation. Based on these two components, we distinguish five clusters of behavior or behavioral patterns:conservers, spenders, cool, warm, and average. The energy use of these clusters differs considerably. Consercers use less energy, while spenders use more energy than theaverage group. The coot and the warm cluster use less energy than the average group.For energy policy, the differences between these behavioral patterns (clusters) are relevant, whereas each cluster is different on sociodemographic and attitudinal variables. This requires different strategies for changing and maintaining energy-related behaviors.
Book
The proliferation of massive data sets brings with it a series of special computational challenges. This "data avalanche" arises in a wide range of scientific and commercial applications. With advances in computer and information technologies, many of these challenges are beginning to be addressed by diverse inter-disciplinary groups, that indude computer scientists, mathematicians, statisticians and engineers, working in dose cooperation with application domain experts. High profile applications indude astrophysics, bio-technology, demographics, finance, geographi­ cal information systems, government, medicine, telecommunications, the environment and the internet. John R. Tucker of the Board on Mathe­ matical Seiences has stated: "My interest in this problern (Massive Data Sets) isthat I see it as the rnost irnportant cross-cutting problern for the rnathernatical sciences in practical problern solving for the next decade, because it is so pervasive. " The Handbook of Massive Data Sets is comprised of articles writ­ ten by experts on selected topics that deal with some major aspect of massive data sets. It contains chapters on information retrieval both in the internet and in the traditional sense, web crawlers, massive graphs, string processing, data compression, dustering methods, wavelets, op­ timization, external memory algorithms and data structures, the US national duster project, high performance computing, data warehouses, data cubes, semi-structured data, data squashing, data quality, billing in the large, fraud detection, and data processing in astrophysics, air pollution, biomolecular data, earth observation and the environment.
Article
The technology for building knowledge-based systems by inductive inference from examples has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in a variety of systems, and it describes one such system, ID3, in detail. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete. A reported shortcoming of the basic algorithm is discussed and two means of overcoming it are compared. The paper concludes with illustrations of current research directions.
Chapter
Classification, in its widest sense, is needed for the development of language, which consists of words that help to recognize and discusses the different types of events, objects and people. The classification has two main aims. The first is the prediction-separating diseases that require different treatments. The second is to provide a basis for research into aetiology- the causes of different types of disease. This chapter aims to provide objective and stable classifications. Objective in the sense that the analysis of the same set of organisms by the same sequence of numerical methods produces the same classification; stable in that the classification remains the same under a wide variety of additions of organisms or of new characteristics describing them. It describes briefly a number of applications of cluster analysis reported in some of these disciplines. cluster analysis
Article
Energy modeling is globally used during the design phase to estimate future building energy performance. Predictions obtained from common energy estimation software typically deviate from actual energy consumption levels. This discrepancy can mainly be attributed to the misrepresentation of the role that building occupants play in the energy estimation equation. Although occupants might have different and varying energy use characteristics over time, current energy estimation tools assume they are constant. This paper proposes a new agent-based approach to commercial building energy modeling by accounting for the diverse and dynamic energy consumption patterns among occupants, in addition to the potential changes in their energy use behavior attributable to their interactions with the building environment and with each other. The impact of an active modeling of occupancy is then illustrated in a case study of an office in a university building, where more than 25% variation in the predicted energy consumption was obtained when using the proposed method versus a traditional commonly used method with static occupancy parameters. DOI: 10.1061/(ASCE)CP.1943-5487.0000158. (C) 2012 American Society of Civil Engineers.
Article
Lighting consumes about 20% to 40% of the total electricity use in large office buildings in China. Commonly in building simulations, static time schedules for typical weekdays, weekends and holidays are assumed to represent the dynamics of lighting energy use in buildings. This approach does not address the stochastic nature of lighting energy use, which can be influenced by occupant behavior in buildings. This study analyzes the main characteristics of lighting energy use over various timescales, based on the statistical analysis of measured lighting energy use data from 15 large office buildings in Beijing and Hong Kong. It was found that in these large office buildings, the 24-hourly variation in lighting energy use was mainly driven by the schedules of the building occupants. Outdoor illuminance levels had little impact on lighting energy use due to the lack of automatic daylighting controls (an effective retrofit measure to reduce lighting energy use) and the relatively small perimeter area exposed to natural daylight. A stochastic lighting energy use model for large office buildings was further developed to represent diverse occupant activities, at six different time periods throughout a day, and also the annual distribution of lighting power across these periods. The model was verified using measured lighting energy use from the 15 buildings. The developed stochastic lighting model can generate more accurate lighting schedules for use in building energy simulations, improving the simulation accuracy of lighting energy use in real buildings.
Article
Two improvements were made to the Sensor-Utility-Network method of occupancy estimation found in the literature, which casts the occupancy estimation/prediction problem as a convex program with an objective function based on various data sources that contribute to occupancy information. The improvements were the inclusion of a scheduling term in the objective function and a mechanism for updating the historical data automatically. The schedule term weighted the difference between projected estimates and scheduled occupancy based on zone usage such that zones with stricter adherence to schedules have more influence. The historical data were updated through an exponentially weighted average of estimates at the same time on past similar days. The estimates for a given time on a given day were calculated as weighted averages of all estimates for that time, including estimates from past and future times on that day. After implementing the SUN algorithm in Matlab, it was tested using data from an agent-based occupancy model. The improvements were added and resulted in a one third decrease in estimation error.
Article
There is no comprehensive consensus about the way people operate blinds or the motivating factors that influence their decisions. However, there is a substantial body of research, encompassing data from more than 50 buildings worldwide, that offers guidance. Patterns of blind use can affect the energy consumption of buildings substantially and therefore energy and daylighting simulations that do not account properly for these blind use patterns will have greater error. An outline for a new more nuanced model for simulation-based blind control is offered. Specifically, this paper reviews the energy implications of blind use, blind occlusion and rate of change from field studies, specific quantitative measures influencing user blind control, investigations into user acceptance of automated blind control, and finally conclusions and knowledge gaps are summarized and suggested next steps for research are provided. Addressing these gaps is critical for the continued progress of the energy efficiency movement.
Article
The technology for building knowledge-based systems by inductive inference from examples has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in a variety of systems, and it describes one such system, ID3, in detail. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete. A reported shortcoming of the basic algorithm is discussed and two means of overcoming it are compared. The paper concludes with illustrations of current research directions.
Article
This paper examines the statistical properties of occupancy in single person offices of a large office building in San Francisco. A probabilistic model to predict and simulate occupancy in single person offices is proposed. It is found that vacancy intervals are exponentially distributed and that the coefficient of the exponential distribution for a single office could be treated as a constant over the day. Occupancy intervals are more complex than vacancy intervals. The distribution of occupancy intervals is time varying. Variations among different offices are examined. The implications of the findings on thermal and air quality control are discussed.
Article
The space heating energy needed during the winter heating season in Seattle Washington, USA, was monitored over a 15-year period, 1987–2002. Two single family residence houses were constructed to building code standards in force at the time of construction and two to standards calling for envelopes with improved thermal resistance. Space conditioning energy needs are strongly affected by occupant behavior, but generally simulations ignore the behavior of the occupants in estimating the energy needed for heating and cooling. The data suggest that estimates of energy savings can be based upon envelope thermal resistance for moderate occupant behavior. For such behavior, the space heating is well characterized by the daily average difference between house average space temperature and outside air temperature. Characterizing in terms of indoor temperature, outdoor air temperature, wind speed, and insolation gives a slightly better representation but requires more information than is usually available. However, vigorous conservation tactics, which produce a thermal response that is highly transient, can lead to substantially different energy needs. No correlation could be established from the measured space heating when aggressive conservation made use of thermostat setback at every opportunity.
Article
This paper reports the development of a building energy demand predictive model based on the decision tree method. This method is able to classify and predict categorical variables: its competitive advantage over other widely used modeling techniques, such as regression method and ANN method, lies in the ability to generate accurate predictive models with interpretable flowchart-like tree structures that enable users to quickly extract useful information. To demonstrate its applicability, the method is applied to estimate residential building energy performance indexes by modeling building energy use intensity (EUI) levels. The results demonstrate that the use of decision tree method can classify and predict building energy demand levels accurately (93% for training data and 92% for test data), identify and rank significant factors of building EUI automatically. The method can provide the combination of significant factors as well as the threshold values that will lead to high building energy performance. Moreover, the average EUI value of data records in each classified data subsets can be used for reference when performing prediction. One crucial benefit is improving building energy performance and reducing energy consumption. Another advantage of this methodology is that it can be utilized by users without requiring much computation knowledge.
Article
Energy use in buildings is closely linked to their operational and space utilization characteristics and the behavior of their occupants. The occupant has influence due to his presence and activities in the building and due to his control actions that aim to improve indoor environmental conditions (thermal, air quality, light, noise). Due to the demand for sustainable buildings more passive buildings will be built. Consequently the weight of the user behavior on the energy balance of a building increases. In the design phase, e.g., via building performance simulation, this effect has only recently been recognized. In this study the effect of user behavior on building performance has been evaluated further to assess requirements for design solutions to arrive at buildings that are more robust to the influence of user behavior. In the research two recent developments with respect to user presence and user interactions in a building have been combined. The results indicate that for specific buildings user behavior should be assessed in more detail, to allow the building design to be optimized for the actual user and its peculiarities. A guideline supports the decision process.
Keeping up with the Quants: Your Guide to Understanding and Using Analytics
  • T H Davenport
  • J Kim
T.H. Davenport, J. Kim, Keeping up with the Quants: Your Guide to Understanding and Using Analytics, Harvard Business Review Press, 2013.
  • C M Bishop
C.M. Bishop, Pattern Recognition and Machine Learning, Springer, 2006, pp. 12e33.
Annual Energy Review, DOE/EIA - 0384
  • Eia
EIA. (2010, 09.03). Annual Energy Review, DOE/EIA e 0384, 2010, Retrieved on 09.03.10 from. http://www.eia.doe.gov/aer/pdf/aer.pdf Available: http:// www.eia.doe.gov/aer/pdf/aer.pdf.