Calibrating subjective data biases and model predictive uncertainties in machine learning-based thermal perception predictions

Real-time thermal comfort evaluation is not only essential in constructing the control module of Heating, Ventilation and Air Conditioning (HVAC) systems in residential buildings but also rather critical in energy conservation. In the transient thermal environment, current thermal comfort is not stable and varies from time to time. Therefore, if we only evaluate the current thermal sensation that will cause prediction error. Existing thermal comfort models mainly focus on the evaluation of current thermal comfort. However, research on the evaluation of thermal sensation variation trend is vacant. Furthermore, since individual differences play an important role in thermal comfort evaluation, physiological indices should be considered. To solve this problem, in this paper, the authors exclusively propose the concept of relative thermal sensation which accounts for the thermal sensation variation trend and give its real-time evaluation method by analysis of skin/clothes temperatures of ten local body segments using machine learning algorithms. By incorporating the relative thermal sensation model with an ordinary thermal comfort model, a novel complex thermal comfort model is derived, which has the ability to predict the current thermal comfort and the thermal sensation variation trend simultaneously and provides an early warning mechanism for thermal discomfort.

Reinforcement learning (RL) for controlling building heating, ventilation, and air conditioning (HVAC) systems has shown promise in delivering energy-efficient, energy-flexible and resilient buildings. However, training RL agents in real-time in buildings is generally considered time and cost prohibitive. RL agents can be trained offline in simulated environments, reducing the time to performance once online, but complex models for simulation can be difficult to generate. What level of complexity is required in an offline training model to produce acceptable online results is unknown. This work seeks to determine how simplified models can be used to train RL agents for building HVAC control. The work uses the EnergyPlus simulation environment paired with HVAC systems modeled in Modelica to act as the real-world analogue environment. Simplified reduced-order models of the building were developed using the lumped capacitance method. The results show that an RL agent trained using Deep Q Learning on a simplified building model can be transferable to a more complex building simulation environment with comparable performance. In virtually all cases, the RL agents were able to deliver improved performance over the baseline proportional controller. Online learning gave mixed results in terms of enhancing the policy transfer process. The implementation in this paper was more suitable for improving the performance of poor policy transfer rather than fine-tuning already well-performing policies.

Energy savings and occupant thermal comfort are the two most important factors in controlling heating ventilation and air conditioning (HVAC) operation in buildings. Typically, it is found that thermal comfort is not always met in buildings. Hence, there is still an opportunity to improve indoor thermal comfort, and at the same time save energy by controlling HVAC set points. The objective of this paper is to propose a method to obtain energy savings by adjusting HVAC set points based on occupant comfort measured using Predicted Mean Vote (PMV) and occupancy information. The idea is to calculate hourly PMV values based on real-time occupancy information, indoor temperature set points and humidity in a building. Then, a new set of temperature set points that can maintain occupant comfort, i.e., PMV = 0, is derived. To evaluate the effectiveness of the proposed method, a building model is developed in eQUEST using the information from a real-world building located in Alexandria, VA. Research findings indicate that HVAC electrical consumption savings of 14.58% is achieved when the proposed set point adjustment method is implemented as compared to that of the base case. To study the impact of adding occupancy information on HVAC energy savings, another scenario is simulated where HVAC set point is increased when the building is unoccupied, e.g., during lunchtime or holidays. Research findings indicate that additional HVAC electrical consumption savings of 8.79% is achieved when taking into account occupancy information in HVAC control.

Heating, ventilation, and air conditioning (HVAC) systems in buildings are an emerging application area for model predictive control (MPC) due to the significant cost benefits that can be achieved via load shifting in modern electricity markets. In this paper, we discuss some of the opportunities and challenges associated with applying MPC to commercial HVAC systems. After defining the control problem, a decomposition of the centralized MPC is presented and demonstrated for an example system. Recent work at the Stanford University campus is also highlighted to show these ideas in practice, and an outlook for the field is given.

Performance Assessment of Desiccant air Conditioning System in an Institutional Building in Subtropical Climate

Achieving nearly zero-energy buildings (nZEB) is one of the main objectives defined by the European Union for achieving carbon neutrality in buildings. nZEBs are heavily reliant on distributed renewable generation energy sources, which create new challenges associated with their inherent intermittency. To achieve nZEB levels, demand management plays an essential role to balance supply and demand. Since up to two-thirds of the total consumed energy in buildings is dispended for Heating, Ventilation and Air Conditioning (HVAC) operations, intelligent control of HVAC loads is of utmost importance. The present work aims to offer a solution to improve a building microgrids’ flexibility by shifting thermal loads and taking advantage of room thermal inertia. Innovation is present in using the internet of things to link several decentralized local microcontrollers with the microgrid and in the applicability of different control algorithms, such as the pre-emptive heating/cooling of a room. The developed solution relies on smart thermostats, which can be integrated into a building management system, or in a microgrid, and are capable of fulfilling the occupants’ need for comfort while complementing the building with needed power flexibility. The equipment is capable of controlling several HVAC systems to guarantee thermal and air quality comfort, as well as coordinate with a building/microgrid operator to reduce energy costs by shifting thermal loads and enacting demand control strategies. The smart thermostat uses an algorithm to calculate room inertia and to pre-emptively heat/cool a room to the desired temperature, avoiding peak hours, taking advantage of variable tariffs for electricity, or periods of solar generation surplus. The smart thermostat was integrated into a university campus microgrid and tested in live classrooms. Since the work was developed during the COVID-19 pandemic, special attention was given to the air quality features. Results show that smart HVAC control is a viable way to provide occupant comfort, as well as contribute to the integration of renewable generation and increase energy efficiency in buildings and microgrids.

<p>We investigate the use of cross-validation (CV) techniques to estimate the accuracy of equivalent-source (also known as equivalent-layer) models for interpolation and processing of potential-field data. Our preliminary results indicate that some common CV algorithms (e.g., random permutations and k-folds) tend to overestimate the accuracy. We have found that blocked CV methods, where the data are split along spatial blocks instead of randomly, provide more conservative and realistic accuracy estimates. Beyond evaluating an equivalent-source model's performance, cross-validation can be used to automatically determine configuration parameters, like source depth and amount of regularization, that maximize prediction accuracy and avoid over-fitting.</p><p>Widely used in gravity and magnetic data processing, the equivalent-source technique consists of a linear model (usually point sources) used to predict the observed field at arbitrary locations. Upward-continuation, interpolation, gradient calculations, leveling, and reduction-to-the-pole can be performed simultaneously by using the model to make predictions (i.e., forward modelling). Likewise, the use of linear models to make predictions is the backbone of many machine learning (ML) applications. The predictive performance of ML models is usually evaluated through cross-validation, in which the data are split (usually randomly) into a training set and a validation set. Models are fit on the training set and their predictions are evaluated using the validation set using a goodness-of-fit metric, like the mean square error or the R² coefficient of determination. Many cross-validation methods exist in the literature, varying in how the data are split and how this process is repeated. Prior research from the statistical modelling of ecological data suggests that prediction accuracy is usually overestimated by traditional CV methods when the data are spatially auto-correlated. This issue can be mitigated by splitting the data along spatial blocks rather than randomly. We conducted experiments on synthetic gravity data to investigate the use of traditional and blocked CV methods in equivalent-source interpolation. We found that the overestimation problem also occurs and that more conservative accuracy estimates are obtained when applying blocked versions of random permutations and k-fold. Further studies need to be conducted to generalize these findings to upward-continuation, reduction-to-the-pole, and derivative calculation.</p><p>Open-source software implementations of the equivalent-source and blocked cross-validation (in progress) methods are available in the Python libraries Harmonica and Verde, which are part of the Fatiando a Terra project (www.fatiando.org).</p>

Indoor climate experience, migration, and thermal comfort expectation in buildings

The increasing world population has led to a significant increase of energy consumption in the building sector in recent years. Reasons are manifold. They mainly include increased thermal comfort requirements for the building occupants. Higher thermal comfort implies higher needs for the buildings' heating, ventilation and air-conditioning (HVAC) systems at their peak levels. This operation eventually entails the risk of load peaks, which increases risks for electric black-outs. Demand side management (DSM) is an established concept for buildings, which includes the optimal shifting of potential loads to time periods such that peak loads can be prevented effectively. The concept thus provides that electric and thermal storage potentials are exploited adequately. Further development should enforce the use of renewable energy sources, which offers another possibility to efficiently reduce the risks of potential load peaks. Mathematical modeling of HVAC systems is an established approach to understand the system behavior and to trigger plans and strategies on how to tackle the HVAC system controls in order to recognize potentials and consequently perform efficient load shifting in an automated way. Model based control concepts endorse the systematic and transparent approach towards automated load shifting. This paper presents the methodology for the design of a model based predictive control concept for load shifting in the non-residential building sector. The methodology includes the mathematical modeling of the HVAC systems and the loads of a passive house standard non-residential building. The concept for the model based predictive controller is developed upon a simplification of the plant model.

Field study of mixed-mode office buildings in Southern Brazil using an adaptive thermal comfort framework

Modeling and optimization for energy efficient smart buildings are interesting and promising research areas. According to Paris Protocol signed in 2015, energy efficient, smart and green buildings are imperative concerns. Heating, ventilation and air-conditioning (HVAC) or air-conditioning and mechanical ventilation (ACMV) systems, consume around 40% of the total energy, and the systems also directly impact on the environmental conditions, especially the indoor environmental conditions, such as air temperature, air humidity, air velocity, air quality, etc. In this thesis, the main objective is to systematically optimize the ACMV systems to operate efficiently and maintain indoor environmental conditions as comfortable and healthy as possible for occupants. The thesis is organized into the following systematic three-phase methodology to enhance ACMV systems' energy efficiency and indoor occupants' thermal comfort in smart buildings:

In part 1, a literature review was done to summarise and introduce the theoretical background knowledge of thermal comfort and passive cooling technology. The adaptive thermal comfort was explained because it is applicable to a free-running building which is the studied object of this research. The basic theory and design standards of adaptive thermal comfort were reviewed. A brief overview of passive cooling techniques was given. The techniques were then reviewed based on their relationships with urban morphology, building shape, layout, opening and “elements”. The study started with a Chinese vernacular building (chapter 4) because these always use the passive way to achieve a comfortable living environment under the limitations of technology at that time. Firstly, the spatial design strategies for passive cooling of a Chinese vernacular house were investigated in a field survey. The design of modern rural houses under free-running conditions compared with the Chinese vernacular house. It was found that the modern rural house did not achieve a satisfactory thermal summer environment under free-running conditions, while the vernacular house did. Furthermore, the vernacular house was deeply analysed by field measurements and dynamic thermal simulations. It was found that the particular spatial design of the vernacular house has its own building microclimate, which is important for the occupants’ thermal summer comfort. The concept of building microclimate was identified. In this study, the scale of “building microclimate” refers to a type of microclimate, involving the indoor space and the spaces around the indoor spaces of a particular building. It is the extension of the indoor climate. The spatial scale is smaller than the urban fabric. It rarely covers an area more than several hundred meters wide, but is bigger than an indoor space alone. It is limited to one particular building, whether a small house or a big stadium. The building microclimate is mainly defined by the spatial and the thermo-physical properties. Similar to the influence of urban morphology on urban microclimate, the spatial configuration influences the building microclimate significantly. To have a particular microclimate at the building scale, some key factors of spatial configuration such as spatial diversity, spatial arrangement and boundary conditions between spaces should be identified. The spatial design of modern house is different from the vernacular house due to the evolution of people’s lifestyle over a long period. Can a modern house have a good building microclimate? To answer this question, the spatial design and thermal environment of a modern house were analysed through field survey and simulation. It was found that a modern house can also have its own microclimate and that the microclimate of this particular building can provide considerable thermal comfort for the occupants in summer under local climate conditions. Adaptive actions, for example movement, can explain why occupants can achieve thermal comfort in a building microclimate with diverse spaces. To find the relationship between the occupants’ spatial perception and thermal perception, a questionnaire was put forward. It was found that the spatial openness of a particular space significantly affects the occupants’ visual perception, wind speed perception and thermal perception. It was revealed that the occupants’ spatial perception and thermal perception are associated. The strongest correlation is between spatial openness and visual perception and wind speed perception. That means spatial boundary conditions can strongly influence occupants’ comfort perception, and subsequently influence the occupants’ spatial choice and movement in a particular thermal environment, given the opportunity, as Humphreys (1997) pointed out: when people are free to choose their location, it helps if there is plenty of thermal variety, giving them the opportunity to choose the places they like. The fundamental assumption of the adaptive approach is expressed by the adaptive principle: “if a change occurs such as to produce discomfort, people react in ways which tend to restore their comfort”. Nicol et al. (2012) proposed that there are at least five basic types of adaptive actions. One important adaptive action is selecting a different thermal environment. Occupant movement in a particular building microclimate is significant for thermal comfort. Occupants can change their location for different activities. Movement is possible between buildings, between rooms, around rooms, out of the sun and into the breeze, and so on (Nicol et al., 2012).

Heating, Ventilation, and Air Conditioning (HVAC) systems are among the largest energy consumers in buildings, and their control strategies play a crucial role in achieving energy efficiency, indoor comfort, and sustainability. This review examines the main HVAC control strategies, ranging from traditional methods to Internet of Things (IoT) based and Artificial Intelligence (AI) driven approaches. In addition, occupancy-based control strategies and energy-efficient designs are critically examined for their potential to enhance system responsiveness and reduce operational costs. More than 100 recent research studies were systematically analyzed and compared across key parameters such as the objective of the study, the used control technique, strength, weakness, opportunity, threat, software/tool used, building type (residential, commercial, or industrial), control scale (single zone/multi-zone), and specific HVAC part controlled (e.g., AHU, VAV). To enable consistent comparison, a Multi-Objective Robustness Criterion (MORC) is used to assess how different control strategies balance energy consumption, thermal comfort, and Indoor Air Quality (IAQ) under changing conditions and how each control technique overcomes the limitations of the previous ones. In addition, the review also introduces a case analysis based on three U.S. cities, Charlotte, Greensboro, and Raleigh, to demonstrate the impact of outdoor climate variations and occupancy pattern on HVAC energy consumption. The results show that even a one-degree change in outdoor temperature can cause a considerable increase in annual energy consumption. Moreover, occupancy-based HVAC control can save more than $7,000 annually in the studied educational building. These findings highlight the strong impact of location-specific temperature differences on HVAC performance and emphasize the need for adaptive, climate-aware control strategies, while also pointing to future directions for smarter and more energy-efficient HVAC systems.

In recent years, the energy conservation demand attracted much attention due to the depletion of energy resource and environmental impact by increasing energy consumption. In particular, heating, ventilation and air-conditioning (HVAC) systems in buildings is responsible for significant portion of global energy demand. Heat or energy recovery is one of the key energy-efficient technologies, which reveals to overcome the increase of energy consumption in building without reducing the indoor air quality. However, in the conventional heat recovery system, only the sensible heat was recovered, but the latent heat was ignored. In this work, a novel energy recovery ventilation (ERV) model is developed with semi-permeable membrane, and the performance of sensible and latent energy subject to tropical climate conditions is investigated by both numerical and experimental methods. The 3D ERV model is comprehensively studied first by CFD simulation for analysis of important parameters, such as velocity, temperature, humidity of supply and exhaust airflow. The CFD simulation results show that the sensible and latent effectiveness could be gained at 75% and 65% respectively. Moreover, preliminary experiment is also conducted to validate the simulation results for the impacts of ERV on energy consumption. The dynamic model of HVAC is then constructed and developed in Simulink. The model predictive control strategy for the control of temperature, humidity and CO2 level will be implemented in this model for optimization of ERV control integrated into the whole HVAC system to achieve much more energy saving.

Increasing demand on heating, ventilation, and air-conditioning (HVAC) systems and their importance, as the respiratory system of buildings, in developing and spreading various microbial contaminations and diseases with their huge global energy consumption share have forced researchers, industries, and policymakers to focus on improving the sustainability of HVAC systems. Understanding and considering various parameters related to the sustainability of new and existing HVAC systems as the respiratory system of buildings are vital to providing healthy, energy-efficient, and economical options for various building types. However, the greatest opportunities for improving the sustainability of HVAC systems exist at the design stage of new facilities and the retrofitting of existing equipment. Considering the high available percentage of existing HVAC systems globally reveals the importance of their retrofitting. The attempt has been made to gather all important parameters that affect decision-making to select the optimum HVAC system development considerations among the various opportunities that are available for sustainability improvement.