HAVC systems account for the majority of energy use in buildings. Therefore, research efforts have been made to develop control strategies to improve energy efficiency during the day and peak time. Studies have traditionally emphasized energy optimization while treating occupant experience using temperature constraints or standard generic metrics of comfort. A well-known strategy, in this category, is the use of a penalizing term when the temperature in an environment is deviated from a pre-defined setpoint. However, in reality, individual differences lead to diverse preferences and sensitivities to indoor thermal environments. Prior studies have not systematically evaluated the impact of such differences on user experience when using advanced control strategies for demand response. Accordingly, in this study, we have proposed a novel occupant-centric control framework (PICO: Personalization-Integrated Co-Optimization) that seeks to minimize energy cost (using dynamic pricing) with a penalizing term that is informed by probabilistic personalized comfort models of the occupants. We hypothesized that such integration results in increased efficiency (during the day and peak time) and improved user experience. Through a comprehensive uncertainty quantification analysis (to account for diversity in occupants' preferences, sensitivities, and number of occupants), we evaluated this framework by comparing it against three commonly used control strategies with varying levels of emphasis on user experience. Our analysis using numerous realizations of the framework operation for different combinations of simulated occupants showed that the proposed framework can adapt to different scenarios and improve the efficiency of operations. Summarizing the energy saving and user comfort experience metrics in an energy productivity measure (that quantifies the comfort gain per unit of energy use) we demonstrated that PICO increases peak time productivity up to 18.37% across various realizations. Moreover, the framework was demonstrated to be more consistent in providing an improved user experience reflected in a considerable reduction of standard deviations for thermal comfort experience, specifically for one occupant scenarios.

The effective control of heating, ventilation, and air conditioning (HVAC) systems can reduce peak energy demand and balance energy usage throughout the day. Co-optimizing energy cost and comfort is one of the main paradigms in HVAC systems' control. This paper presents a feasibility study of a human-centered co-optimization-based controller. Two methods were utilized to factor in comfort in optimization formulations including deviation from a temperature setpoint and the integration of personalized thermal comfort models. We compared the effectiveness of these two approaches in achieving comfort during peak times by hypothesizing that the integration of personalized comfort measures will result in an improved comfort experience for users during demand response. The controllers' performance was evaluated using a simulated benchmark small office and various combinations of comfort profiles for single occupancy scenarios. Our results indicate that integrating personalized comfort profiles into an optimization-based controller has the potential to improve users' comfort experience by considerably decreasing the variance of comfort experience compared to the conventional controller.

HVAC systems account for the majority of energy consumption in buildings. Efficient control of HVAC systems can reduce energy consumption and enhance occupants’ comfort. In the existing literature, energy-comfort or cost-comfort co-optimization frameworks commonly involve manual tuning of the balancing coefficient between energy and comfort through parameter tuning by an expert. Nevertheless, achieving the optimal balance between energy usage and occupant comfort remains challenging. This limitation restricts the generalizability of different formulations across various scenarios or testing on different environments. In this paper, we propose an implicit evolutionary Reinforcement Learning (RL) approach to learn and adapt the trade-off parameter of an energy-comfort optimization formulation. We have developed a predictive comfort-energy co-optimization formulation for controlling the setpoint of a building. The RL agent utilizes a novel guidance-induced random search method to learn the energy-comfort trade-off coefficient and guide the optimization formulation. The reward function of the RL model is energy productivity (comfort over energy consumption). To evaluate the feasibility of our proposed approach, we conducted experiments on a real-world testbed - i.e., an apartment unit. Our feasibility study shows that the proposed approach can learn an optimal control parameter and reduce energy consumption by 24.3% while decreasing comfort by only 1% compared to the baseline.

HVAC systems account for majority of energy consumption in buildings and play a vital role in energy efficiency and occupants' comfort. Efficient control of HVAC systems could reduce energy consumption while maintaining occupants' comfort at an acceptable level. Predictive control strategies that leverage the thermal capacity of buildings have been shown to be an effective approach in decreasing the energy consumption of buildings. One of the conventional methods in representing comfort in the formulation of predictive controllers is to consider a fixed temperature range as a constraint. However, this method does not account for differences in occupants' thermal preferences. Therefore, in this paper, we have compared the performance of two model-predictive controllers in terms of energy consumption and thermal satisfaction: the first one is a conventional controller constrained by a fixed temperature range and the second proposed controller is constrained by information from personal comfort profiles. The controllers were formulated as optimization problems using multivariate regression for predictive modeling and genetic algorithm for optimization. To represent human thermal preferences, probabilistic comfort profiles of occupants were developed by utilizing real-world thermal votes. The performance of these controllers was evaluated in a residential building through EnergyPlus simulations for different multi-occupancy scenarios of one, two, and four occupants. The proposed MPC controller improves thermal satisfaction by 15% while increasing energy consumption by 4% on average.

HVAC systems account for 50% of buildings' energy use and could play a critical role in energy management in buildings both for energy-saving, and demand-side management to reduce peak energy use. According to contextual demands of occupants, efficient control of HVAC systems could result in peak energy saving and decreased energy costs in demand response programs. Accordingly, in this paper, we have introduced an agent-based model consisting of three agents: human agent, thermostat agent, and utility agent. In this model, the thermostat agent receives the real-time electricity price from the utility agent and aggregates thermal comfort profiles of the occupants from human agents. By considering these inputs, the thermostat agent employs a predictive model of a house and calculates the next set point of the HVAC system on energy cost and occupants' comfort. Then, the thermostat agent signals the suggested set point to the thermostat of the building at each time step. For evaluating the proposed controller's performance, the electricity price profile from ERCOT, which supplies the state of Texas, is used as a signal from the utility agent. Realistic thermal comfort data were used to simulate the thermal preference of occupants represented as human agents. The evaluation was carried out in a co-simulation using a Python and EnergyPlus model of a residential unit. Our results show that the proposed controller reduces the peak energy by 5.5% to 10% and increases occupants' thermal satisfaction up to 12%. The main contribution of this paper is developing an agent-based model that humans, as the main stakeholders of the buildings, play a role in controlling HVAC systems for peak reduction and energy saving.

Resilient transportation networks are a critical component of urban societies during and after a disaster, necessary for emergency services, rescue operations, and access to major population and activity centers. Previous studies observed that beside physical damages to transportation infrastructure, post-disaster unusual traffic patterns may lead to gridlock, congestion, and failures in transportation networks. However, little is known about analyzing the impacts of post-disaster unusual human mobility and traffic patterns on the performance of transportation networks. The objective of this study is to empirically and statistically examine two hypotheses: 1) post-disaster unusual traffic patterns perturb the topology of transportation networks; and 2) perturbed topological features of the transportation network affect the network performance. Historical records of taxi GPS traces in New York City were used to examine the two hypotheses on the New York City road network after Hurricane Sandy in 2012. The results of the statistical process control using the exponentially weighted moving average control chart show that post-disaster unusual traffic patterns perturbed the topology of the New York City transportation network and significantly shifted the network betweenness index from its usual variations. The outcomes of the Granger causality test confirm the second hypothesis and indicate that the deviated betweenness index resulted from the perturbed network topology is statistically associated with the network closeness index indicating that the perturbed topological features of the network affect its performance. The outcomes of this study will help decision makers empirically analyze impacts of post-disaster unusual traffic patterns on performance and resiliency of transportation networks.

The United States has more than 615,000 bridges. US national bridge inspection standards developed by the Federal Highway Administration (FHWA) require routine inspections of these bridges every 24 months regardless of bridge characteristics such as age, average daily traffic (ADT), and current deterioration condition of a bridge. Previous studies reported that this routine inspection process is considerably costly and inefficient. If the future condition of a bridge can be predicted accurately, costly routine inspections with uniform intervals can be avoided. The objective of this study is to create a forecasting model that predicts future bridge deterioration conditions based on the bridge characteristics. Historical data of more than 28,000 bridges in the state of Ohio from 1992 to 2017 were used to create an ordinal regression model to statistically examine effects of bridge characteristics on variations in bridge condition and predict future bridge conditions. The outcomes of this study indicate that bridge characteristics such as age, ADT, deck area, structural material, deck material, structure system, maximum length of span, and current condition of the bridge are statistically significant variables that explain variations in bridge deterioration. The results of the forecasting process show that the created ordinal regression model can statistically predict future bridge conditions precisely. This study will help bridge owners and transportation agencies accurately model and predict bridge deterioration and assign inspection and maintenance resources efficiently. The efficient inspection process, customized based on predicted deterioration condition, can result in investing the millions of dollars currently funding unnecessary inspections into much-needed infrastructure development projects. (C) 2019 American Society of Civil Engineers.