Conference Paper

A Bayesian Network Model for Predicting the Cooling Load of Educational Facilities

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

Abstract

In the U.S., educational facilities consume a large amount of energy. Model predictive control schemes can improve the energy efficiency of educational facilities. Accurate and fast prediction of the cooling load is essential to performances of model predictive control schemes. Although many methods for the cooling load prediction were proposed, they are not suitable for educational facilities due to the lack of an efficient way to reflect the impact of internal activities on the cooling load. After analyzing the characteristics of cooling load of educational facilities, we proposed to use the day type instead of the day of the week as the input for the prediction. Then we constructed a Bayesian Network model based on that. To evaluate how the proposed inputs enhance the cooling load prediction, we also implemented the other Bayesian Network model with inputs recommended by the literature. To assess performances of those models, we performed a case study in which on-site measured cooling load and meteorological data was used for the training and testing. The results show that the Bayesian Network models can capture the trend of cooling load even with a limited size of training data. Replacing the day of the week by the day type can significantly improve the accuracy of cooling load prediction for educational facilities.

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.

... Xiao et al. (2016) employed a Bayesian Network model in the fault detection for the air handling unit systems. In the previous study, we also developed a Bayesian Network model for optimizing the condenser water set point (Huang, Malara, et al. 2016) and performed a preliminary study, which aims to extend the model to load prediction purposes (Huang, Zuo, et al. 2016b). While significant, these applications were of limited scope and application of the Bayesian Network models. ...
... For the other two parent nodes (outdoor dry bulb temperature and outdoor wet bulb temperate), we discretized the temperature into 2 degree increments (2 o C) that spanned the full range of these temperature data: for the outdoor dry bulb temperature, the discrete sections are [0, 2),…, [40, ∞); for the outdoor wet bulb temperature, the split sections are [0, 2), …, [30, ∞). We chose 2 o C increment because it is the best according to our sensitivity analysis (see details in Huang, Zuo, et al. 2016b). ...
Article
Full-text available
Cooling load prediction is indispensable to many building energy saving strategies. In this paper, we proposed a new method for predicting the cooling load of commercial buildings. The proposed approach employs a Bayesian Network model to relate the cooling load to outdoor weather conditions and internal building activities. The proposed method is computationally efficient and implementable for use in real buildings, as it does not involve sophisticated mathematical theories. In this paper, we described the proposed method and demonstrated its use via a case study. In this case study, we considered three candidate models for cooling load prediction and they are the proposed Bayesian Network model, a Support Vector Machine model, and an Artificial Neural Network model. We trained the three models with fourteen different training data datasets, each of which had varying amounts and quality of data that were sampled on-site. The prediction results for a testing week shows that the Bayesian Network model achieves similar accuracy as the Support Vector Machine model but better accuracy than the Artificial Neural Network model. Notable in this comparison is that the training process of the Bayesian Network model is fifty-eight times faster than that of the Artificial Neural Network model. The results also suggest that all three models will have much larger prediction deviations if the testing data points are not covered by the training dataset for the studied case (The maximum absolute deviation of the predictions that are not covered by the training dataset can be up to seven times larger than that of the predictions covered by the training dataset). In addition, we also found the uncertainties in the weather forecast significantly affected the accuracy of the cooling load prediction for the studied case and the Support Vector Machine model was more sensitive to those uncertainties than the other two models.
... Huang et al. [13] proposed a cooling tower control strategy for legacy chiller plants composed of multiple chillers, multiple cooling towers, primary chilled water pumps of constant speed, and condenser water pumps of constant speed. As a model predictive control method, this strategy adopted a novel model to predict chiller power and cooling tower power based on the predicted T wet and system cooling load [14] . The set point of T cwr is optimized to minimize the total power of chillers and cooling towers. ...
Article
In the domain of optimal control for building HVAC systems, the performance of model-based control has been widely investigated and validated. However, the performance of model-based control highly depends on an accurate system performance model and sufficient sensors, which are difficult to obtain for certain buildings. To tackle this problem, a model-free optimal control method based on reinforcement learning is proposed to control the building cooling water system. In the proposed method, the wet bulb temperature and system cooling load are taken as the states, the frequencies of fans and pumps are the actions, and the reward is the system COP (i.e., the comprehensive COP of chillers, cooling water pumps, and cooling towers). The proposed method is based on Q-learning. Validated with the measured data from a real central chilled water system, a three-month measured data-based simulation is conducted under the supervision of four types of controllers: basic controller, local feedback controller, model-based controller, and the proposed model-free controller. Compared with the basic controller, the model-free controller can conserve 11% of the system energy in the first applied cooling season, which is greater than that of the local feedback controller (7%) but less than that of the model-based controller (14%). Moreover, the energy saving rate of the model-free controller could reach 12% in the second applied cooling season, after which the energy saving rate gets stabilized. Although the energy conservation performance of the model-free controller is inferior to that of the model-based controller, the model-free controller requires less a priori knowledge and sensors, which makes it promising for application in buildings for which the lack of accurate system performance models or sensors is an obstacle. Moreover, the results suggest that for a central chilled water system with a designed peak cooling load close to 2000 kW, three months of learning during the cooling season is sufficient to develop a good model-free controller with an acceptable performance.
... In addition, ̇ can be estimated using the load prediction model shown in [15] and ( ) can be obtained from the weather forecast service. ...
Article
Full-text available
Achieving the optimal control of cooling towers is critical to the energy-efficient operation of current or legacy chiller plants. Although many promising control methods have been proposed, limitations in their applications exist for legacy chiller plants. For example, some methods require the change of the plant's overall control structure, which can be difficult to legacy chiller plants; some methods are too complicated and computationally intensive to implement in old building control systems. To address the above issues, we develop an operational support system. This system employs a model predictive control scheme to optimize the condenser water set point and can be applied in chiller plants without changes in the control structure. To further facilitate the implementation, we propose to increase the optimization accuracy by selecting a better starting point. The results from a case study with a real legacy chiller plant in Washington D.C. show that the proposed operational support system can achieve up to around 9.67% annual energy consumption savings for chillers and cooling towers. The results also show the proposed starting point selection method can achieve a better accuracy and a faster computational speed than commonly used methods. In addition, we find that we can select a lower optimization frequency for the studied case since the impact of the optimization frequency on the energy savings is not significant while a lower optimization frequency does reduce the computational demand to a great extent.
Article
Probabilistic graphical models (PGMs) can effectively deal with the problems of energy consumption and occupancy prediction, fault detection and diagnosis, reliability analysis, and optimization in energy systems. Compared with the black-box models, PGMs show advantages in model interpretability, scalability and reliability. They have great potential to realize the true artificial intelligence in energy systems of the next generation. This paper intends to provide a comprehensive review of the PGM-based approaches published in the last decades. It reveals the advantages, limitations and potential future research directions of the PGM-based approaches for energy systems. Two types of PGMs are summarized in this review, including static models (SPGMs) and dynamic models (DPGMs). SPGMs can conduct probabilistic inference based on incomplete, uncertain or even conflicting information. SPGM-based approaches are proposed to deal with various management tasks in energy systems. They show outstanding performance in fault detection and diagnosis of energy systems. DPGMs can represent a dynamic and stochastic process by describing how its state changes with time. DPGM-based approaches have high accuracy in predicting the energy consumption, occupancy and failures of energy systems. In the future, a unified framework is suggested to fuse the knowledge-driven and data-driven PGMs for achieving better performances. Universal PGM-based approaches are needed that can be adapted to various energy systems. Hybrid algorithms would outperform the basic PGMs by integrating advanced techniques such as deep learning and first-order logic.
Chapter
Occupancy behaviour plays an important role in energy consumption in buildings. Currently, the shallow understanding of occupancy has led to a considerable performance gap between predicted and measured energy use. This chapter presents an approach to estimate the occupancy based on blind system identification (BSI), and a prediction model of electricity consumption by an air-conditioning system is developed and reported based on an artificial neural network with the BSI estimation of the number of occupants as an input. This starts from the identification of indoor CO2 dynamics derived from the mass-conservation law and venting levels. The unknown parameters, including the occupancy and model parameters, are estimated by using a frequentist maximum-likelihood algorithm and Bayesian estimation. The second phase is to establish the prediction model of the electricity consumption of the air-conditioning system by using a feed-forward neural network (FFNN) and extreme learning machine (ELM), as well as ensemble models. To analyse some aspects of the benchmark test for identifying the effect of structure parameters and input-selection alternatives, three studies are conducted on (1) the effect of predictor selection based on principal component analysis, (2) the effect of the estimated occupancy as the supplementary input, and (3) the effect of the neural network ensemble. The result shows that the occupancy number, as the input, is able to improve the accuracy in predicting energy consumption using a neural-network model.
Article
Full-text available
Net-zero energy communities (NZECs) are critical to assuring the sustainability and resilience of modernized power systems. System modeling helps overcome technical challenges in designing and operating NZECs. In this paper, we present an open-source NZEC virtual testbed in Modelica based on a real NZEC in Florida, USA. This testbed consists of two sets of models: (1) higher-fidelity physics-based models that consider the interaction between subsystems of the studied NZEC and capture fast dynamics, and (2) lower-fidelity data-driven models that require fewer resources to establish and/or run. All models are validated against measurements from this real NZEC. In addition, this testbed includes a simulation framework that streamlines the processes for simulation and thus allows the use of developed models to form a virtual testbed. To demonstrate the usage of the virtual testbed, a case study is conducted where a building-to-grid integration control is evaluated via simulation. The evaluation results suggest that the tested control significantly smooths the power draw of the studied community and does not sacrifice thermal comfort to a great extent.
Preprint
Full-text available
Net-zero energy communities (NZECs) are critical to assuring the sustainability and resilience of modernized power systems. System modeling helps overcome technical challenges in designing and operating NZECs. In this paper, we present an open-source NZEC virtual testbed in Modelica based on a real NZEC in Florida, USA. This testbed consists of two sets of models: (1) higher-fidelity physics-based models that consider the interaction between subsystems of the studied NZEC and capture fast dynamics, and (2) lower-fidelity data-driven models that require fewer resources to establish and/or run. All models are validated against measurements from this real NZEC. In addition, this testbed includes a simulation framework that streamlines the processes for simulation and thus allows the use of developed models to form a virtual testbed. To demonstrate the usage of the virtual testbed, a case study is conducted where a building-to-grid integration control is evaluated via simulation. The evaluation results suggest that the tested control significantly smooths the power draw of the studied community and does not sacrifice thermal comfort to a great extent.
Article
Full-text available
Cooling Load based Control (CLC) for the chiller sequencing is a commonly used control strategy for multiple-chiller plants. To improve the energy efficiency of these chiller plants, researchers proposed various CLC optimization approaches, which can be divided into two groups: studies to optimize the load distribution and studies to identify the optimal number of operating chillers. However, both groups have their own deficiencies and do not consider the impact of each other. This paper aims to improve the CLC by proposing three new approaches. The first optimizes the load distribution by adjusting the critical points for the chiller staging, which is easier to be implemented than existing approaches. In addition, by considering the impact of the load distribution on the cooling tower energy consumption and the pump energy consumption, this approach can achieve a better energy saving. The second optimizes the number of operating chillers by modulating the critical points and the condenser water set point in order to achieve the minimal energy consumption of the entire chiller plant that may not be guaranteed by existing approaches. The third combines the first two approaches to provide a holistic solution. The proposed three approaches were evaluated via a case study. The results show that the total energy consumption saving for the studied chiller plant is 0.5%, 5.3% and 5.6% by the three approaches, respectively. An energy saving of 4.9–11.8% can be achieved for the chillers at the cost of more energy consumption by the cooling towers (increases of 5.8–43.8%). The pumps’ energy saving varies from −8.6% to 2.0%, depending on the approach.
Conference Paper
Full-text available
This paper describes a case study of the model-based condenser water temperature setpoint control for a chiller plant with multiple chillers and cooling towers. The plant was modeled using the Modelica Buildings library. Both the supervisor control and local feedback loop controls were modeled to more realistically represent the real system. The total energy consumptions and computing times under different optimization strategies were compared. The result shows that the daily optimization can provide similar energy saving with significantly less computing time compared to the hourly optimization.
Article
Full-text available
Many of the popular building energy simulation programs around the world are reaching maturity — some use simulation methods (and even code) that originated in the 1960s. For more than two decades, the US government supported development of two hourly building energy simulation programs, BLAST and DOE-2. Designed in the days of mainframe computers, expanding their capabilities further has become difficult, time-consuming, and expensive. At the same time, the 30 years have seen significant advances in analysis and computational methods and power — providing an opportunity for significant improvement in these tools.In 1996, a US federal agency began developing a new building energy simulation tool, EnergyPlus, building on development experience with two existing programs: DOE-2 and BLAST. EnergyPlus includes a number of innovative simulation features — such as variable time steps, user-configurable modular systems that are integrated with a heat and mass balance-based zone simulation — and input and output data structures tailored to facilitate third party module and interface development. Other planned simulation capabilities include multizone airflow, and electric power and solar thermal and photovoltaic simulation. Beta testing of EnergyPlus began in late 1999 and the first release is scheduled for early 2001.
Article
Full-text available
General regression neural networks (GRNN) were designed and trained to investigate the feasibility of using this technology to optimize HVAC thermal energy storage in public buildings as well as office buildings. State of the art building simulation software, ESP-r, was used to generate a database covering the years 1997–2001. The software was used to calculate hourly cooling loads for three office buildings using climate records in Kuwait. The cooling load data for 1997–2000 was used for training and testing the neural networks (NN), while robustness of the trained NN was tested by applying them to a “production” data set (2001 data) that the networks have never “seen” before.Three buildings of various densities of occupancy and orientational characteristics were investigated. Parametric studies were performed to determine optimum GRNN design parameters that best predict cooling load profiles for each building. External hourly temperature readings for a 24 h period were used as network inputs, and the hourly cooling load for the next day is the output. The performance of the NN analysis was evaluated using a statistical indicator (the coefficient of multiple determination) and by statistical analysis of the error patterns, including confidence intervals of regression lines, as well as by examination of the error patterns.The results show that a properly designed NN is a powerful instrument for optimizing thermal energy storage in buildings based only on external temperature records.
Article
This paper presents the development of a data driven probabilistic graphic model to predict building energy performance. A directed graphical model, namely, a Bayesian Networks (BNs) model is created. Each node in the BNs represents a random variable such as outside air temperature and energy end use. The links between the nodes are probabilistic dependencies among these corresponding variables. These dependencies are statistically learned and/or estimated by using measured data and augmented by domain expert knowledge. BNs models became popular models in the last decade and only recently received attention for HVAC (Heating, Ventilation and Air-conditioning) applications. A case study of using a BNs model to predict HVAC hot water energy consumption in an office building is presented. The energy estimation results meet with the criteria recommended by ASHRAE Guideline 14. This case study also shown that the discretized Bayesian Network model is sensitive to the discretization policy (i.e., bin size selection) employed. The applicability of a BNs model becomes questionable outside the range in which the model is learned.
Article
Energy conservation and indoor environment concerns have motivated extensive research on various aspects of control of Heating, Ventilating and Air-Conditioning (HVAC) and building systems. The study on optimal operation as well as modeling of HVAC and building systems is one of the fastest growing fields that contribute to saving energy and improving indoor environment of buildings. The reasonable operation adjustment is one of the main methods to improve the energy efficiency. Cooling load prediction is the foundation of the optimization operation. This paper is devoted to the development of a comprehensive modeling of cooling load for a large building with ice-storage systems in Beijing, China. The models describe the dynamics of cooling load, outdoor climate parameters and indoor parameters as one multi-variable nonlinear system in a way that is useful for prediction analysis. The cooling load data collected is from June to September, and then the method of similarity for both longitudinal and transverse waveforms is used to judge whether there is abnormal data. The optimal parameter setting in the proposed model is studied. Principle Component Analysis (PCA) method was applied to select input parameters. A load prediction model has been constructed based on BP neural networks. Taking account of the generalization ability of neural networks, this paper has chosen the bayesian regularization algorithm, which can get better fitting effect than other training algorithms, to train the neural networks. Then, the BP neural network model is used for the summer hourly cooling load prediction of the business building. Evaluation of the prediction accuracy of the proposed models is based on the root mean square error (RMSE). The results show that the prediction model can accurately predict the future hourly load of 1 week and 1 day, with the prediction error at 1.60% and 1.18% respectively. The analysis shows that this model is suitable for the practical engineering application and can provide a basis for optimal operation of air conditioning control systems of large public buildings.
Article
It is not feasible to apply the traditional prediction method to predict hourly building cooling load at the urban energy planning stage because of the limited building information and complexity of energy prediction. This paper presents a simplified prediction model: Hourly Cooling Load Factor Method (HCLFM) that can provide quick and fair estimate of building cooling load for large-scale urban energy planning. The paper introduces the assumptions and principles of the proposed method, as well as discussing the implication and limitation of the approach. As a verification and demonstration, the method is applied to an office building in Beijing. The predicted results show that the dynamical trend of the cooling load is reasonable. The study further analyzes the potential causes of prediction errors and the significance of various cooling load influence factors.
Article
Building cooling load prediction is one of the key factors in the success of energy-saving measures. Many computational models available in the industry have been developed from either forward or inverse modeling approaches. However, these models usually require extensive computer resources and lengthy computation. This paper discusses the use of the multi-layer perceptron (MLP) model, one of the artificial neural network (ANN) models widely adopted in engineering applications, to estimate the cooling load of a building. The training samples used include weather data obtained from the Hong Kong Observatory and building-related data acquired from an existing prestigious commercial building in Hong Kong that houses a mega complex and operates 24 h a day. The paper also discusses the practical difficulties encountered in acquiring building-related data. In contrast to other studies that use ANN models to predict building cooling load, this paper includes the building occupancy rate as one of the input parameters used to determine building cooling load. The results demonstrate that the building occupancy rate plays a critical role in building cooling load prediction and significantly improves predictive accuracy.
Article
Cooling load prediction is important and essential for many building energy efficient controls, such as morning start control of chiller plant. However, most of the existing methods are either too complicated or of unsatisfactory performance for online applications. A simplified online cooling load prediction method is therefore developed in this study. The method firstly selects a reference day for each day according to load profile similarity. The load profile of the reference day is taken as the initial prediction result of the cooling load. Secondly, the most correlated weather data is identified and its hourly predictions are used to calibrate the initial load prediction result based on the reference day. Lastly, the accuracy of the calibrated load prediction is enhanced using the prediction errors of the previous 2 h. The developed load prediction method is validated in the case studies using the weather data purchased from the Hong Kong Observatory and the historical data from a super high-rise building in Hong Kong. The load prediction method is of low computation load and satisfactory accuracy and it can be used for online application of building load prediction.
Article
The DOE-2 building energy analysis program was designed to assist engineers and architects in the performance of design studies of whole-building energy use under actual weather conditions. Program development was guided by several objectives: (1) that the description of the building entered by the user be readily understood by non-computer scientists; (2) that the calculations be based upon well-established algorithms; (3) that the program permit the simulation of commonly available heating, ventilating, and air-conditioning (HVAC) equipment; (4) that the costs of running the program be minimal; and (5) that the predicted energy use of a building be acceptably close to measured values. These objectives have been met. We present here an overview of the DOE-21D version of the program. An annotated example of DOE-2 input and output is shown in the Appendix. 9 refs., 12 figs., 21 tabs.
Article
This paper stresses the importance of a proper estimate of ground reflectivity in building energy simulation, particularly in the presence of snow. Several factors influencing ground reflectivity are reviewed. Two models for estimating ground reflectivity in the presence of snow are developed. The first one is based on the number of days with snow depth greater than 5cm, and is appropriate for use with ‘typical’ meteorological years. The second makes use of actual records of snow depth, and takes into account snow accumulation, ageing and melting to calculate reflectivity values. The algorithms were implemented in ESP-r and tested by simulating a passive solar house for six Canadian locations. The sensible heating load calculated by the simulations is reduced by up to 10.9% on a yearly basis, and 23.3% on a monthly basis, when the ground reflectivity takes into account the presence of snow.
Article
This paper describes TRNSYS, a computer program designed specifically to connect component models in a specified manner, solve the simultaneous equations of the system model, and display the results. Solar energy system components are described by individual FORTRAN subroutines. These subroutines comprise a growing library of equipment models available to the user for system simulation.
Article
A great amount of world energy demand is connected to the built environment. Electricity use in the commercial buildings, accounts for about one-third of the total energy consumption in Turkey and fully air-conditioned office buildings are important commercial electricity end-users since the mid-1990s. In the presented paper, the interactions between different conditions, control strategies and heating/cooling loads in office buildings in the four major climatic zones in Turkey – hot summer and cold winter, mild, hot summer and warm winter, hot and humid summer and warm winter – through building energy simulation program has been evaluated. The simulation results are compared with the values obtained from site measurements done in an office building located in Istanbul. The site-recorded data and simulation results are compared and analyzed. This verified model was used as a means to examine some energy conservation opportunities on annual cooling, heating and total building load at four major cities which were selected as a representative of the four climatic regions in Turkey. The effect of the parameters like the climatic conditions (location), insulation and thermal mass, aspect ratio, color of external surfaces, shading, window systems including window area and glazing system, ventilation rates and different outdoor air control strategies on annual building energy requirements is examined and the results are presented for each city.
Article
A Bayesian Network approach has been developed that can compare different building designs by estimating the effects of the thermal indoor environment on the mental performance of office workers. A part of this network is based on the compilation of subjective thermal sensation data and the associated objective thermal measurements from 12,000 office occupants from different parts of the world. A Performance Index (Π) is introduced that can be used to compare directly the different building designs and furthermore to assess the total economic consequences of the indoor climate with a specific building design. In this paper, focus will be on the effects of temperature on mental performance and not on other indoor climate factors. A total economic comparison of six different building designs, four located in northern Europe and two in Los Angeles, USA, was performed. The results indicate that investments in improved indoor thermal conditions can be justified economically in most cases. The Bayesian Network provides a reliable platform using probabilities for modelling the complexity while estimating the effect of indoor climate factors on human beings, due to the different ways in which humans are affected by the indoor climate.
Article
In this paper, support vector machine (SVM) is used to predict hourly building cooling load. The hourly building cooling load prediction model based on SVM has been established, and applied to an office building in Guangzhou, China. The simulation results demonstrate that the SVM method can achieve better accuracy and generalization than the traditional back-propagation (BP) neural network model, and it is effective for building cooling load prediction.
Article
Predicting the next-24-hour load in a building is essential for the optimal control of heating, ventilating and air-conditioning (HVAC) systems that use thermal/cool storage technology and also for cost and energy reduction of the non-storage systems. To fully integrate the advantages of several models and improve the accuracy of forecasting load, the application of the combined forecasting method to hourly load forecasting is presented in this paper. The method of Analytic Hierarchy Process (AHP) is employed to deduce the weights of each model. A case study shows that the combined forecasting model based on AHP may be better than the individual ones in predicting the building's hourly load for the future hours.
Article
Based on building energy and indoor environment simulations, this study uses a recently developed method relying on Bayesian Network theory to estimate and compare the consequences for occupant performance and energy consumption of applying temperature criteria set according to the adaptive model of thermal comfort and the more conventional PMV model. Simulations were carried out for an example building with two configurations (with and without mechanical cooling) located in tropical, subtropical, and temperate climate regions. Even though indoor temperatures differed significantly between building configurations, especially in the tropical climate, the estimated performance differed only modestly between configurations. However, energy consumption was always lower in buildings without mechanical cooling, particularly so in the tropical climate.The findings indicate that determining acceptable indoor thermal environments with the adaptive comfort model may result in significant energy savings and at the same time will not have large consequences for the mental performance of occupants.
Article
Cooling load is a heat value of cold water used for air conditioning in a district heating and cooling system. Cooling load prediction in a district heating and cooling system is one of the key techniques for smooth and economical operation. In this article, cooling load prediction in such a district heating and cooling system is considered. Unfortunately, since actual cooling load data usually involve measurement noises, outliers, and missing data for several reasons, a prediction method considering the effect of the outliers and missing data is desirable. In this article, a new prediction method using a simplified robust filter to improve a numerical stability problem of a robust filter and a three-layered neural network, is proposed. Applications of the proposed method and some other methods to actual cooling load data in a district heating and cooling system involving outliers and missing data show the usefulness of the proposed method.
Discretizing Continuous Attributes While Learning Bayesian Networks
  • N Friedman
  • M Goldszmidt
Friedman, N., Goldszmidt, M. 1996. Discretizing Continuous Attributes While Learning Bayesian Networks. In Proc. of International Conference on Machine Learning, Bari, Italy.
Buildings Energy Data Book
  • J D Kelso
Kelso, J. D. 2012. 2011 Buildings Energy Data Book, D&R International, Ltd.
Model Predictive Control for the Operation of Building Cooling Systems
  • Y Ma
Ma, Y., et al. 2010. Model Predictive Control for the Operation of Building Cooling Systems. IEEE Transactions on Control System Technology 20(3): 796 -803.
Optimal Control of Chiller Plants using Bayesian Network
  • A C L Malara
Malara, A. C. L., et al. 2015. Optimal Control of Chiller Plants using Bayesian Network. In Proc. of The 14th International Conference of the IBPSA Hyderabad. National Climatic Data Center. Quality Controlled Local Climatological Data. Retrieved May 14, 2015, from http://www.ncdc.noaa.gov/dataaccess/land-based-station-data/land-baseddatasets/quality-controlled-localclimatological-data-qclcd.
Automatically Calibrating a Probabilistic Graphical Model of Building Energy Consumption
  • D Tarlow
Tarlow, D., et al. 2009. Automatically Calibrating a Probabilistic Graphical Model of Building Energy Consumption. In Proc. of The 11th Conference of IBPSA, Glasgow, Scotland.