Computing in Civil Engineering 2019: Smart Cities, Sustainability, and Resilience

Collecting and analyzing psychological and environmental data to build an occupant behavior model and pairing them with targeted interventions can result in increased occupant energy efficiency, productivity, and well-being. Specifically, lighting has been shown to drastically influence these factors in occupants; however, lighting’s effects on occupant emotion and occupancy as time series data have not been explored. As a result, in this paper, we explore the effects of lighting on occupants’ emotion and occupancy over time using a system of LED lights, webcams, and Raspberry Pi’s. We gauge occupant emotion and occupancy changes by alternating the brightness and color of a light strip located on their work space, and analyze their emotional state through the webcam image feeds. The results of this study show, on average, the color red is associated with the highest occupancy ratios, the color yellow with the highest engagement, and the color blue with the highest valence.

Predicting remaining useful life (RUL) of critical machines in heavy maintenance industries, such as oil refineries and upgrading, is crucial to deliver robust and cost-effective plans for resource procurement and maintenance scheduling in a challenging and constrained work environment. Statistical methods have been extensively utilized based on condition monitoring (CM) and sensor data. In this paper, we investigate the use of statistical methods in RUL prediction and conduct the taxonomy of those methods to provide better understanding of their potential application in the context of asset maintenance management for an industrial plant. Then the application of statistical learning methods for data-driven RUL prediction is illustrated with sensor data collected from a running plant. The research deliverable is intended to provide data-driven RUL predictions as part of plant asset management system and shed light on decisions directly pertaining to industrial plant maintenance management.

The objective of this paper is to model and examine the impacts of different levels of infrastructure service losses caused by disasters on the households’ well-being residing in a community. An agent-based simulation model was developed to capture complex mechanisms underlying households’ tolerance for the service outages, including household characteristics (e.g., sociodemographic, social capital, resources, and previous disaster experience), physical infrastructure attributes, and extreme disruptive events. The rules governing these mechanisms were determined using empirical survey data collected from the residents of Harris County affected by Hurricane Harvey as well as the existing models for power outages and service restoration times. The analysis results highlighted the spatial diffusion of service risks among households living in affected areas in disasters. The proposed simulation model will provide utility agencies with an analytical tool for prioritization of infrastructure service restoration actions to effectively mitigate the societal impacts of service losses.

Sewer networks’ performance includes several uncertainties, which increases the corresponding risks of the system. To improve performance of sewer pipeline networks, a useful method is to inspect the operation of the pipes regularly. Since random inspection of pipes is greatly expensive, the pipeline network decision-makers tend to have a predictive model to anticipate the condition of these systems. In this research, a model was developed to predict the performance of pipes, using multiple regression method. This model utilizes historical data as input including the failures, and hydraulic data. The results of this model revealed that the predictive model was more sensitive to the age of the pipeline system as well as the type of pipes. This model assists engineers and decision-makers to identify the pipe segments which are more inclined to failure and helps them to develop a comprehensive risk management system for sewer pipeline systems.

Achieving communities resilient to natural disasters has been a big deal for the decision-makers. It takes different time for the communities to recover from similar disasters, due to their different pre-disaster and post-disaster conditions. Many complicated social, economic, environmental, etc. conditions and their dynamic relationship and interaction should be considered to understand how they cause delays. This research aims to study how the barriers to disaster recovery affect the timeliness of recovery process and what is their relationships. Thus, the barriers were identified, and through a social network analysis (SNA), the relationships and interactions among them was modeled and analyzed. Among different barriers, damage to major infrastructure systems was recognized as the main cause of disaster recovery delays. Implementing the findings of this study, decision-makers of post-disaster recovery could plan to achieve a more resilient community by identifying the relationships among different preventive indicators and planning to avoid them.

Different aspects of disaster recovery have been studied in recent years through which a variety of barriers relative to different perspectives were determined. However, the barriers causing delays in the process of post-disaster recovery have not been studied in detail. Some of these barriers are related to pre-disaster conditions, and some relate to post-disaster situation. In this study, post-disaster recovery barriers causing delay have been studied comprehensively. Based on the existing literature, the barriers were identified, categorized, ranked, and weighted. The 33 identified barriers are presented in economic, social, environment, policy and politics, infrastructure, and transportation categories. Furthermore, social and economic categories had the highest weight, 35% and 25% respectively, among all the categories. The found preventives in this study can be used by researchers for later studies. In addition, the decision-makers can use the results for making proper disaster recovery planning.

Delamination assessment of the bridge deck plays a vital role for bridge health monitoring. Thermography as one of the nondestructive technologies for delamination detection has the advantage of efficient data acquisition. But there are challenges on the interpretation of data for accurate delamination shape profiling. Due to the environmental variation and the irregular presence of delamination size and depth, conventional processing methods based on temperature contrast fall short in accurate segmentation of delamination. Inspired by the recent development of deep learning architecture for image segmentation, the convolutional neural network (CNN) based framework was investigated for the applicability of delamination segmentation under variations in temperature contrast and shape diffusion. The models were developed based on dense convolutional network (DenseNet) and trained on thermal images collected for mimicked delamination in concrete slabs with different depths under experimental setup. The results suggested satisfactory performance of accurate profiling the delamination shapes.

Understanding occupants’ thermal comfort is essential for the effective operation of heating, ventilation, and air conditioning (HVAC) systems. Existing studies of the “human-in-the-loop” HVAC control generally suffer from: (1) excessive reliance on cumbersome human feedback; and (2) intrusiveness caused by conventional data collection methods. To address these limitations, this paper investigates the low-cost thermal camera as a non-intrusive approach to assess thermal comfort in real time using facial skin temperature. The framework developed can automatically detect occupants, extract facial regions, measure skin temperature, and interpret thermal comfort with minimal interruption or participation of occupants. The framework is validated using the facial skin temperature collected from twelve occupants. Personal comfort models trained from different machine learning algorithms are compared and results show that random forest model can achieve an accuracy of 85% and also suggest that the skin temperature of ears, nose, and cheeks are most indicative of thermal comfort.

Generation of digital models of existing industrial facilities is labor intensive and expensive. The use of state-of-the-art deep learning algorithms can assist to reduce the modelling time and cost. However large databases of labelled, laser-scanned industrial facilities do not exist to date, henceforth training of deep learning models is not possible. Our paper solves this problem by proposing a new benchmark dataset, which consists of five labelled industrial plants. The labelling schema that we followed for the generation of this dataset is based on the frequency of appearance of industrial object types. We labelled the ten most frequent industrial object shapes as identified in previous work. We present CLOI (channels, L-shapes, circular sections, I-shapes): a richly annotated large-scale repository of shapes represented by labelled point clusters. CLOI has more than 140 million hand labelled points and serves as the foundation for researchers who are interested in automated modelling of industrial assets using deep learning algorithms.

This paper presents a critical review of the state-of-the-art data-driven machine learning methods utilized for building energy forecast. Specifically, it offers a look into the advantages and disadvantages of four widely adopted machine learning methods: artificial neural networks, support vector machines, genetic algorithms, and decision trees. Based on the performance of these methods explored in previous studies, recommendations of application are provided for different categories such as building type (e.g., residential), forecasting method (e.g., long-term), and building energy (e.g., electricity). Some of the main identified research gaps include the lack of studies dedicated to long-term energy forecasts and inability to successfully incorporate occupant behavior into the models. This review also highlights the potential and prospects of hybrid models as avenues of growth in the domain of building energy forecast. Further research efforts in these areas of study can reap future benefits by promoting energy conservation thereby reducing the ecological footprint.

Over the past decade, construction companies have been using unmanned aerial vehicles (UAVs) extensively as a tool for data collection. While UAVs are being used on construction job sites on daily basis, there is no comprehensive quantitative method to measure the risks of flying UAVs over a specific area on a job site. Despite some work in quantifying direct risk of UAV flights, indirect risks of flying UAVs have been completely ignored. This paper presents a model for measuring indirect risks of UAV flights that are primarily associated with proximity between construction workers and UAVs. The proposed model is applied to a case study, using Monte Carlo simulation, and the results show the applicability of the model to a construction job site environment.

Crowdsourced data acquired through pedestrians’ physiological response [e.g., electrodermal activity (EDA), gait patterns, brain activity, and heart rate] offer a unique opportunity for diagnosing and maintaining built environment features in a neighborhood. However, using crowdsourced raw physiological responses obtained from naturalistic ambulatory settings cannot adequately capture local prominent patterns, since various technical challenges (e.g., motion artifacts, electrode contact noise) and compounding factors (e.g., heightened physiology because of movement) make it difficult to identify meaningful fine-grain signal fluctuations. Motivated by this, this study proposes a method to assess adverse built environment features that cause physical discomfort and/or emotional distress to pedestrians, by using saliency detection analysis on multiple physiological responses (i.e., EDA, heart rate, and gait patterns). Each physiological response data is segmented with a bottom-up segmentation approach in an unsupervised way. A physiological saliency cue (PSC) is proposed and used to compute the distinctiveness of physiological responses over the segment of interest in comparison to the remaining ones, and collective PSC of a physical point of interest (POI) is measured across participants. The results using physiological response data acquired from wearable devices indicates that the proposed saliency detection analysis is effective in capturing prominent local patterns. The outcome of this research will provide a foundation of using physiological response data to evaluate built environment features and eventually will promote neighborhood walkability.

An ideal lighting controller should balance between occupant comfort and energy efficiency. In a prior study, we introduced LightLearn, an occupant centered lighting controller based on reinforcement learning and demonstrated its control operation under real conditions in a university building. The controller learned individual occupant preferences and indoor environmental conditions, and adapted its control parameters accordingly by determining personalized set-points. In this paper, we implement a simulation framework to comprehensively evaluate LightLearn. This framework consists of Python and EnergyPlus, and we simulated the performance of LightLearn under various indoor lighting conditions and occupant behaviors. These changes vary the interactions between LightLearn and human input, which effects the performance of the control system. Despite such variations, we show how LightLearn adapts to any indoor light condition and occupant behavior, and determines the optimal policy for balancing occupant comfort and energy efficiency.

There has been an increasing demand for data-driven and machine learning-based bridge deterioration prediction approaches for supporting enhanced bridge maintenance decision making. Bridge inspection reports, which contain a wealth of information about bridge conditions, open opportunities for data analytics to better understand and predict bridge deterioration. However, learning from the reports is challenging, because they usually contain multiple—even ambiguous, uncertain, and conflicting—information about the same bridge element, its deficiencies, and its deficiency measurements. Learning from such data negatively affects the generalizability and the separability of machine learning models, which compromises the performance of data-driven prediction. To address this challenge, this paper proposes a hybrid information fusion method. The method includes two main components: named entity normalization for fusing concepts in ambiguous surface forms into a canonical form, and data fusion for fusing numerical deficiency measurements containing uncertainties and conflicts into a unified and consistent representation. This paper focuses on analyzing the data fusion requirements, and presenting the proposed data fusion method and its evaluation results. The results indicate that the proposed method can adequately address the fusion requirements.

Over the past years, a growing trend of utilizing camera-equipped drones for periodical building facade inspection has emerged. Building façade anomalies, such as cracks and erosion, can be detected through analyzing drone-captured video, photographs, and infrared images. Such anomalies are known to have an impact on various building performance aspects, e.g., thermal, energy, moisture control issues. Current research efforts mainly focus on computational image processing methods to recognize certain types of facade anomalies. However, there is a lack of research on mapping detected anomalies into the building model and managing lifecycle inspection information. This paper aims to propose a systematic process for detecting and managing building façade anomalies based on drone-collected images. An overall data structure, data flow, and related processing techniques within this systematic process are defined. The proposed systematic process will support the facade anomaly detection and building maintenance decision-making.

Dual water distribution systems, which deliver potable and non-potable water separately, have been proposed as a technological infrastructure solution to enhance the sustainability and resilience of urban water supply system by improving performance and decreasing energy consumption. The objective of this study was to evaluate the long-term performance of dual versus existing singular water distribution systems through developing a simulation model that captures long-term dynamics of water distribution infrastructure systems. The model integrates utility agency’s renewal decision-making processes with the physical infrastructure degradation to simulate the long-term transformation of pipeline infrastructure. Accordingly, various system performance measures, including network average condition, break frequency, leakage, energy loss, service reliability, and life-cycle costs, were simulated over a 50-year horizon for the city of Fort Collins, CO. The analysis results enabled a quantitative comparison between dual and singular water distribution systems in terms of different long-term performance measures.

The ability to automatically classify activities performed by various equipment in real-time can assist project managers in reliable decision-making and project control. Such an endeavor requires the identification of individual sequential work-motions (e.g. excavator swinging empty) performed by equipment, which then combine to a specific activity (e.g. excavator loading truck). Towards this end, this paper describes an automated activity recognition framework for construction equipment that uses multiple inertial measurement units (IMU) attached to the equipment’s articulated implements. Initial data were collected using multiple IMUs attached at strategic locations on an excavator. Collected data were then segmented, labelled, and used as inputs in the supervised machine learning classifier. The result demonstrates the ability of the proposed framework to obtain real-time insight into the operation performance of construction equipment using low-cost sensors that are already available on the equipment, but until now only used for pose estimation and automated machine guidance.

Improving building water efficiency is crucial in meeting the long-term water needs of urban dwellers and achieving global sustainability targets. Although a body of research efforts has been conducted on modeling urban water use, there is still a limited understanding of the impacts of various factors on water consumption in different geographic regions, and of the interdependencies of building water and energy usage. Towards addressing these knowledge gaps, this paper proposes a machine learning-based model to predict the water consumption of buildings based on their physical characteristics and energy consumption levels. Building water consumption data from New York City, Boston, and Philadelphia were used. A support vector regression (SVR) algorithm was used to build the prediction model. The paper discusses the proposed model and its performance results, identifies the features that affect building water consumption and their importance patterns, and analyzes the impacts of the identified factors on water consumption in different cities.

The evaluation of the road surface roughness condition plays an important role for the assessment of the pavement condition. To measure the international roughness index (IRI), new emerging surveying techniques can be used in addition to the profilometer. Mobile laser scanner (MLS) data were acquired on a stretch of road and interpolated on a grid to build a 3D model representing the whole road surface. To evaluate the roughness, both dynamic and geometric IRIs were computed from (digital elevation model) DEM-derived elevation attributes. The computed values showed a similar trend and a high correlation with that derived from the profilometer; the differences between the corresponding values were generally less than 5%. To evaluate the effects of vibrations on human comfort, the built DEM was integrated into a dynamic simulation software user interface to compute the effects on the drivers of the vertical accelerations, considered another indicator of the pavement performance.

The adoption of smart meters in residential households provides electricity consumption data with high temporal resolution. Considering the wealth of the generated information, data analytics methods can be employed to segment the households based on the timing and magnitude of consumption. Specifically, the resultant load shapes reveal the lifestyle usage of consumers, which brings opportunities for the management of demand or the integration of renewables. Prior efforts have focused on the segmentation based on the whole-house consumption data. However, the discussed conclusions did not involve a deeper analysis of appliances-specific usage and the potential of energy reduction at critical times. To this end, this paper explores the contribution of major appliance types and the associated time-of-use patterns for the formation of common daily load shapes. The objective was to explore if there is any specific relationship between the type of appliances used and the load shapes. A preliminary case study evaluation was conducted on a real-world energy consumption dataset. The data includes both aggregate-level and appliance specific power time series with 15-minute resolution from 60 households for two weeks. Clustering on the aggregate building-level data was employed to capture the representative load shapes. Using the appliance-level data for each load shape, the appliance contribution in terms of consumption amount at different time intervals were investigated. The findings could help to improve the understanding of the drivers of households’ daily consumption patterns in addition to investigating a more informed recommender system for direct load control.

The main aim of this paper is to measure land displacements, as a result of fluid overextraction, and evaluate the impact of measured deformation on the groundwater quality in the Inglewood region, Los Angeles. Therefore, persistent scatterer interferometry (PSInSAR) technique was applied on a set of Sentinel-1 radar images to monitor the subsidence rate in the interested area between June 2017 and May 2018. Then, groundwater data were used to assess the variations in water parameters affecting its quality. Changes in the concentration of the studied parameters including total dissolved solids (TDS), biochemical oxygen demand (BOD), arsenic, and lead were investigated. The PSInSAR outputs demonstrated a high-rate of subsidence in the Inglewood field (30 mm/year). The results showed that uplift as a result of water injection lead to significant increase in TDS, arsenic, and lead, diminishing the water quality and exceeding quality limits and standards. Water injection does not enhance water quality, only preventing the subsidence in this area.

We present an automatic technique for structural damping estimation, based on traffic-induced vibrations of a long-span suspension bridge. First, an identification algorithm is applied to the acceleration data of the Lysefjord Bridge to capture traffic load events, i.e. when vehicles are crossing the bridge. Thereafter, a time domain decomposition is adopted to compute the modal displacement response triggered by the traffic excitation. The damping ratio and the eigenfrequency are then estimated by fitting a single degree-of-freedom system to the modal response. The damping estimates exhibit a lower dispersion and bias than the corresponding results based on a covariance driven stochastic subspace identification algorithm. Estimating damping ratios using traffic-induced decays removes the problems of frequency resolution for short-length records and may, therefore, complement other ambient vibration identification techniques.

Widespread destruction from large-scale disasters results in lengthy and costly recovery, with societal and economic impacts persisting many years after a hazard event. Thus, reconstruction planning plays a key role in rapid post-disaster recovery. Optimal recovery strategies that maximize socioeconomic benefits must consider limited federal, state, and local resources in planning sequential and organized reconstruction—which is the ultimate goal of the post-disaster recovery process. This paper presents a new post-disaster recovery prioritization model that evaluates optimal recovery priorities of damaged facilities and affected regions through a resource allocation analysis. This proposed model is designed for quantifying and optimizing the social and economic benefits of a recovery strategy and resource allocation for reconstruction and repairing of homes and other damaged critical facilities. Multiple sources of data including historical disaster data, GIS, and Federal Emergency Management Agency damage and facility data are referenced in developing and validating the proposed model.

Active shooter incidents present an increasing American homeland security threat to public safety and human life. Several municipal law enforcement agencies have released building design guidelines intended to offer increased resilience and resistance to potential attacks. However, these design recommendations mainly focus on terrorist attacks, prioritizing the enhancement of building security, whereas their impact on safety during active shooter incidents, and corresponding human-building interactions (HBIs) that influence the outcomes (response performance), remain unclear. To respond to this research gap, virtual reality, with its ability to manipulate environmental variables and scenarios while providing safe non-invasive environments, could be a promising method to conduct human-subject studies in the context of active shooter incidents. In this paper, we identify the requirements for developing virtual environments that represent active shooter incidents in buildings to study HBIs and their impacts on the response performance. Key components constituting virtual environments were considered and presented. These include: (1) what types of buildings should be modeled in virtual environments; (2) how to select protective building design recommendations for active shooter incidents and model them in virtual environments; (3) what types of adversary and crowd behavior should be modeled; and (4) what types of interactions among participants, buildings, adversaries, and crowds should be included in virtual environments. Findings on the above key components were summarized to provide recommendations for future research directions.

Visual inspection of concrete structures is an important task; however, it is labor-intensive work. Automatically detection of deterioration from digital pictures has been extensively researched. Current deep learning approaches have improved the detection compared with conventional approaches. Most deep learning approaches use publicly available deep learning models, such as faster R-CNN and single shot multibox detector (SSD). Although the results of deterioration detection methods have been compared with the conventional approach, the deep learning models themselves have not been compared. In this research, using the same datasets, seven deep learning models (YOLOv3, RetinaNet-50, RetinaNet-101, RetinaNet-152, SSD512, SSD300, and faster R-CNN), were compared for detecting five types of deterioration (cracks, exposed reinforcing bars, free lime c-type, free lime d-type, and free lime e-type). YOLOv3 showed the highest mean average precision (mAP) of 91.1%, whereas the other models had mAPs of approximately 80%.

Forecasting activities of construction workers and equipment is immensely desirable for construction safety management, as it can enable capturing near-miss of accidents and right-time intervention of collision, struck-by, trespassing, and improper use of tools. In this paper, an activity forecasting framework for construction safety management is presented, and an application to forecast workers and equipment’s motion trajectory from previously observed motion is implemented. A long short-term memory (LSTM) encoder-decoder network is developed to forecast future locations with mixture density network (MDN) used to model uncertainty in predictions. Two contextual cues are proved to help activity forecasting: (1) worker/equipment placement and distances; (2) object type attributes. A joint training schema is employed to forecast target’s locations at different future times. The proposed framework can handle very long sequences (around 2000 steps) well and accurately predict future locations in maximum 40 frames (2 seconds). Our experiment results show that the proposed model significantly outperforms conventional time-series analysis models. In 1080p high-definition videos the final model achieves average localization error in 10, 20, 40 future frames with 7.30, 12.71, and 24.22 pixels, respectively.

Building energy models have under predicted actual electricity consumption by about 30%. These models are typically based on physical attributes of an actual building and simulate its operation at static indoor conditions. The attributes of building designs and systems may not represent accurate properties of the building because the data becomes available only as architectural and engineering design progresses. The lack of building data details creates uncertainties in the model. Internet of things (IoT) data enable us to obtain reliable building operational data. This research addresses a major challenge for using IoT data during a building energy prediction and operation. IoT data currently pose two major challenges: 1) IoT data has been generated in existing buildings, but it is often unused, and 2) uncertainties in building energy predictions have resulted in unreliability of the simulation results. The research objective is to present a methodology of IoT-initiated building energy modeling and monitoring to increase the reliability of energy predictions and energy management. The uncertainty-integrated energy modeling enables for reliable energy planning prior to building construction, and for resilient monitoring during post construction and facility operation.

In bridge management practices, detecting damage and taking proper maintenance actions in a timely manner are significant issues. Due to the limited professional manpower and budget, providing a guide concerning potential problematic conditions is important to support on-site bridge inspections. The aim of this study was to estimate the number and severity of damage occurrences on a bridge deck using the Korean bridge management system (KOBMS). In this research, we considered identification, structural, inspection, and environmental factors and developed a deep neural network (DNN) model using 15,309 data, and we determined 36 influencing factors. The DNN model successfully predicted the number of damage occurrences on bridge decks and their severity with about 94.68% accuracy, confirmed by inserting external environmental data and span information. The findings emphasized the benefit of using machine learning algorithms when analyzing bridge conditions, and it showed potential for application to network-level decision making for preventive maintenance.

Automated interpretation of closed-circuit television (CCTV) inspection videos has the potential to improve the speed, accuracy, and consistency of sewer pipeline condition assessment. Previous approaches have focused on defect classification in images, with little focus on defect localization (i.e., calculating location of defects relative to the pipe). Recent studies have shown that deep-learning based object detection models can be used to classify and localize operational defects, such as roots and deposits; however, the detection of structural defects, such as pipe fractures is challenging, given their fine silhouettes in images. This paper presents a two-step defect detection framework that uses a 5-layered convolutional neural network (CNN) for classification followed by the you-only-look-once (YOLO) model for detection of pipe fractures. The framework was trained using 1,800 images and yielded a 0.71 average precision (AP) score in detecting fractures, when tested on 300 images. The proposed framework can also achieve a significant reduction in time taken to process CCTV videos. Ongoing research aims to validate the framework on videos from a variety of pipes and extend the framework to defect additional defect categories.

Cognitive buildings, being aware of their own performance, actively learn and adapt to the environment, enhancing comfort and safety of occupants. While the integration of Internet of things (IoT) technology, machine learning, reasoning, and augmented reality into cognitive building concepts has made substantial progress, the integration of structural health monitoring (SHM) and control systems into cognitive buildings has received little attention. In this paper, a semantic model for describing SHM and control systems associated with cognitive buildings is proposed. The semantic model, formally describing SHM and control systems components as well as interrelations between the components, serves as a basis to document, to manage, and to optimize cognitive buildings. Through the semantic model, a laboratory test structure, mimicking a cognitive building, and a prototype SHM and control system, which is installed on the laboratory test structure, are described using building information modeling (BIM). Both the cognitive building and the SHM and control system are physically implemented in the laboratory for validation tests. The results clearly demonstrate the ability of the semantic model to formally describe SHM and control systems associated with cognitive buildings.

Roadway construction workers have to work in close proximity to construction equipment as well as high-speed traffic, exposing them to an elevated risk of collisions. This research aims to develop a solution to reduce the risk of collisions at roadway work zones. To this end, an automated hazard detection and prevention system is designed to detect potential threats and provide warning of potential collisions. A hazard detection algorithm is developed to identify potential proximity hazards between workers and connected/automated vehicles (CAV) and/or construction equipment. A real-time work zone map is developed that displays the real-time locations of workers, equipment, and CAVs. To assure the accuracy of hazard detection, the algorithm accommodates various parameters including variant safety distance for workers-on-foot, vehicles and equipment, direction of movement, distance from border, shape of road, etc. The designed system is developed and evaluated through various experiments on the Smart Road at Virginia Tech.

This study seeks to evaluate the applicability of heat flux sensors, as a proxy for contextual thermal comfort representation in an environment, for the human-in-the-loop (HITL) heating, ventilation, and air conditioning (HVAC) operations. In accounting for personalized thermal comfort inference, the predicted mean vote (PMV) model is not often applicable and data-driven methods using features from ambient environmental conditions have demonstrated moderate accuracies (60–75%). Therefore, the use of physiological sensing for improved performance has gained attention in recent years. Skin temperature has been widely used as an alternative indicator of human heat dissipation adjustment, which could be leveraged for personalized thermal comfort inference. However, considering the heat dissipation adjustments through skin as an adaptation mechanism in an environment, in this study, we have investigated the hypothesis that heat flux sensors could be an effective alternative for enabling feedback from human body to the control system. To this end, we designed and conducted an experimental study on 10 human subjects, wearing a thin-film heat flux sensor on their wrist, under transient thermal conditions (from 20 to 30°C) to investigate the association between human thermal perception and the corresponding heat flux. Through this experiment, the correlations between heat flux, air temperature, relative humidity, and thermal preferences were analyzed. The high correlation factors (0.90 on average) between heat flux and thermal preferences suggest that heat flux sensors have a high potential to be used as an effective sensing modality in the HITL HVAC operation.

With the networks of sophisticated sensors and devices, building systems have the potential to serve as the infrastructure that provides essential data for the Internet of things (IoT)-enabled smart city paradigm. However, current building systems lack inter-system connectivity or exposure to the larger networks of IoT devices. In this paper, we propose a scalable data acquisition framework for the smart built environment—smart buildings, smart communities, and smart cities—that enables the utilization of the data housed in separate building systems for innovative IoT use cases, by understanding IoT stakeholders’ common data needs from buildings and identifying the overlaps between the data protocols used by different building systems. An architecture of IoT-enabled smart cities based on this data acquisition framework is also demonstrated.

A large amount of resources are spent on constructing new facilities and maintaining the existing ones. The total cost of facility ownership can be minimized by focusing on reducing the facilities life-cycle costs (LCCs) rather than the initial design and construction costs. With the developments of machine learning in predictive analytics and the utilizing building systems that provide ubiquitous sensing and metering devices, new opportunities have emerged for architecture, engineering, construction, and operation (AECO) professionals to obtain a deeper level of knowledge on buildings’ LCCs. This paper provides a state-of-the-art overview of the various machine learning applications in the facility LCC analysis field. This paper aims to present current machine learning for LCC research developments, analyze research trends, and identify promising future research directions.

Raw point clouds of build structures do not provide semantic nor component level information. Automatic bridge information model generation from point cloud data would remedy this issue. This paper describes, applies, and evaluates a strategy for top-down partitioning of reinforced concrete bridges into structurally distinct component regions for girder, box girder, and slab bridges. The first phase is a user-guided denoising and coordinate frame specification process. The second phase utilizes the reference frame and geometric heuristics to identify the individual spans and whether the bridge contains AASHTO girders or not. It identifies cutting surfaces to apply for partitioning the bridge. The third phase identifies the substructure, superstructure, and deck components. This process either merges partitioned regions or further partitions regions to isolate bridge components with different functions. Isolated components include the girders/boxes/slabs, the piers, the abutments, the diaphragms, the road, and the parapets. The final coarse bridge model has a level of detail from 100 to 200. Application to four bridges demonstrates the low processing time and high partitioning accuracy, and provides visualizations of the IFC bridge model output.

Humans have a greater influence on energy consumption in residential buildings than other types of buildings. Although existing studies focus on how energy consumption is affected by building technologies and occupant demographics, few studies have incorporated the impact of occupant energy use patterns. The goal of this study is to identify the features that affect energy consumption in residential buildings and to measure their predictive performance. The researchers examined the impact of occupants’ energy use behaviors and the energy use patterns of home appliances on home energy consumption. The patterns reflect on a combination of appliances, their use times and frequencies, and the configurations set by users. Data from the Residential Energy Consumption Survey (RECS) are analyzed to select features for prediction, using multiple machine learning algorithms including support vector machine (SVM) and random forest. The results provide a list of features that efficiently predict energy consumption in residential buildings. The selected 32 features achieve 98% of the prediction performance of that from the entire 271 features. This list of effective features can be used to improve the effectiveness of energy saving programs and to educate occupants about their energy use patterns. The relationship between occupants’ behavior patterns and energy use patterns revealed from this study provides the groundwork for researchers to further explore the prediction of occupant behavior from energy consumption.

UAVs have been widely used in visual inspections of buildings, bridges, and other structures. In either outdoor autonomous or semi-autonomous flights missions strong GPS signal is vital for UAV to locate its own positions. However, strong GPS signal is not always available, and it can degrade or fully loss underneath large structures or close to power lines, which can cause serious control issues or even UAV crashes. Such limitations highly restricted the applications of UAV as a routine inspection tool in various domains. In this paper a vision-model-based real-time self-positioning method is proposed to support autonomous aerial inspection without the need of GPS support. Compared to other localization methods that requires additional onboard sensors, the proposed method uses a single camera to continuously estimate the inflight poses of UAV. Each step of the proposed method is discussed in detail, and its performance is tested through an indoor test case.

UAVs showed great efficiency on scanning bridge decks’ surface by taking a single shot or through stitching a couple of overlaid still images. If potential surface deficits are identified through aerial images, subsequent ground inspections can be scheduled. This two-phase inspection procedure showed great potentials on increasing field inspection productivity. Since aerial and ground inspection images are taken at different scales, a tool to properly fuse these multi-scale images is needed for improving the current bridge deck condition monitoring practice. In response to this need a data fusion platform is introduced in this study. Using this proposed platform multi-scale images taken by different inspection devices can be fused through geo-referencing. As part of the platform, a web-based user interface is developed to organize and visualize those images with inspection notes under users’ queries. For illustration purpose, a case study involving multi-scale optical and infrared images from UAV and ground inspector, and its implementation using the proposed platform is presented.

In this study, a hierarchical density-based spatial clustering of applications with noise algorithm (HDBSCAN) is implemented for the identification of clusters of pipe failures in drinking water distribution systems. The clustering method is adapted to incorporate the space-time distances between failure events to capture spatiotemporal clusters in the failure data. The proposed approach is demonstrated on a dataset of pipe failures retrieved from the maintenance records of a real-life, large-scale water utility. The clustering approach was able to capture clusters of varying spatial and temporal spread. Overall, this study targets identifying patterns of pipe breaks, which can improve the understanding of the significant factors that promote the deterioration of the pipe infrastructure.

Public opinion is a valuable tool for evaluating transportation services and planning. With the popularity of Web 2.0, there is an increasing number of people utilizing social media platforms (e.g. Twitter) to express their opinions. Therefore, social media is an inexpensive and abundant source of information that can be collected and utilized to enhance transportation services. Sentiment analysis, as a common social media data (SMD) processing technique, has been widely used to measure sentiment of user’s opinions on specific topics. However, previous research often neglected the influence of user habit in sentiment value analysis. For example, if one user utilizes social media to express only negative opinions, then their tweets expressing opinions on transportation service would also only be negative, which could cause interference with the evaluation result. The purpose of this research is to investigate the influence of user habit when conducting sentiment analysis on SMD in transportation field. The hypothesis is that user habits will have a positive correlation on sentiment value of selected tweets. To test this hypothesis, a case study is presented that collected tweets posted in Miami-Dade County from May 27^th, 2017, to May 27^th, 2018. The AFINN sentiment analysis method will be used to calculate the sentiment value of all of the tweets. Tweets relevant with personal opinions on transportation services (POTS) were extracted automatically using Naïve Bayes algorithm. All selected POTS tweets were given a revised sentiment value based on tweet itself, and a user average sentiment value based on this user’s historical tweets. The correlation analysis results of this study can test the correctness of hypothesis that user habits have great influence on sentiment value of selected tweets. Significantly, this innovative sentiment analysis method can make the SMD more reliable and valuable for transportation planning and decision-making process.

Nuclear power plant accidents such as those at Fukushima and Chernobyl release a significant amount of hazardous radiation. It takes a long time to recover from such an accident and the difficulty of the task is very high due to the many risks. Restoration work using unmanned vehicles protects workers from the risk of secondary accidents such as radiation exposure and building collapses. Unmanned tasks in high-risk environments should be performed in accordance with carefully planned and optimized work scenarios that prioritize specific tasks. This study examined the non-radiological and radiological risks, accessibility, task difficulty, and efficiency that determine work priorities in such cases and developed a methodology to assess the risks associated with specific unmanned restoration tasks. The proposed risk assessment methodology is based on a risk matrix that can be used to derive work priorities for each zone at a post-disaster nuclear site.

Autonomous and connected vehicles (ACVs) will play a significant role in near future, and some autonomous vehicles are already driven on the road. The more autonomous vehicles are introduced, the more transportation systems will be affected. The literature has been quick to report the effects of ACVs based on simulation models, but it is uncertain what the effects might be in the future due to ACVs. This paper conjectures that existing simulations may not be enough to predict the effects of ACVs. The objective of this paper is to perform a critical literature review of the predicted impacts of ACVs with an emphasis on the foundations of the simulations used to reach conclusions. This paper reviews factors which are important in existing transportation systems and factors which should be incorporated into simulating the future transportation systems. The results of this study contribute to guiding future simulation applications for implementing ACVs.

The use of unmanned aerial vehicles (UAV) are gaining popularity for bridge inspection due to its capabilities to efficiently reach an inaccessible place, collect imagery, and identify damage. Bridges inspected manually are found to be time consuming, and involve safety risks for the inspector on hard-to-reach places. Therefore, it is necessary to advance the current bridge inspection technology using UAV. The main goal of the paper is to inspect different types of damage of an in-service glued-laminated arch bridge with concrete decking using a commercially available UAV and evaluate the UAV’s ability for collection good quality bridge inspection data. As part of the inspection, the arch timber stringer bridge located near city of Keystone in the state of South Dakota was selected. Preliminary inspection was performed over the bridge to identify major damage locations using UAV, and then a detailed inspection was conducted for damage specific to the bridge components such as stringers, abutments, and underneath deck. High-quality digital images on the components with different degrees of damage/deterioration were captured using UAV. Then, these images were used to identify the types of damage on each component: cracks on concrete abutment, exposed rebar and efflorescence on concrete deck, and checks and stains on glulam stringers. The damage identified by UAV was compared to those from the inspection report completed by South Dakota Department of Transportation (SDDOT). Key findings showed that the UAV-enabled inspection efficiently discovered detailed damage for individual bridge components, and provided results almost identical to those gained from the inspection report.

Streaming data can provide a timely understanding of the state of infrastructure networks to enable real-time monitoring and control. However, erroneous data is also inevitable and, if not identified and isolated effectively, may results in erroneous decisions and adverse consequences. This study leverages intra- and inter-similarity structures from monitoring stations for data validation in infrastructure networks. First, validation rules are developed to estimate similarity of new data from an individual sensing station to its routine patterns and flag suspicious data streams. Second, an unsupervised learning model is applied to identify clusters of stations that exhibit similar streaming data and test for cross-similarity to estimate the likelihood of the flagged data stream being invalid. The proposed model is demonstrated using steaming data collected from a real water distribution system. Preliminary results have revealed how a sensor’s readings diverge from its historical patterns when a sensor fault occurs, and have discovered the presence of clusters of sensor stations with significant cross-similarity for boosting data validation performance.

The impact of a snowstorm on the safety and mobility of roadway transportation depends mainly on the storm’s level of severity. Defining storms’ severity, though, is challenging due to the high number of weather parameters needed to describe these events and the non-linear relationships among these parameters. Finding patterns among snowstorms can conceivably simplify this process and help practitioners better analyze and prepare for such events, even when the severity is not explicitly quantified. Therefore, this study interrogated historical data to assess and compare clustering methods and to identify patterns manifesting in snowstorms to lay the necessary foundations for building a more reliable and objective winter severity index. The research team selected three hierarchical clustering methods that differentiated similar groups of snowstorms among more than 2,000 events dated between 2006–2016 in Nebraska. The team then evaluated the performance of these methods using the Calinski-Harabasz index. A range of clustering scenarios were reviewed visually using principal component analysis to determine the optimal number of clusters. The results indicate that while some districts can be described by as few as three clusters, others can experience up to six different clusters of snowstorms. The use of PCA and visualization in this context can facilitate a better understanding of these high-dimensional data, and the findings of this study can help agencies better comprehend snowstorms and prepare for them, which can help communities to maintain the safety and mobility of their drivers.

The concept of smart cities has been developed with advances in information and communication technologies (ICT). Aging infrastructure, declining budgets, changing population, sustainability, and investments are some of the current issues in urban environments. In many different types of the context of smart city solutions, transportation and logistics form the main part of that solution. Hence, there is a need for leveraging the digital footprint and using the data along with the decision tools to get the best performance of the infrastructure assets. The decisions between all options of transportation are currently based on current time and traffic conditions, and other factors such as sustainability, and costs are not considered. This paper aims to create decision support tools to make city planning and operations more efficient and improve the quality of life of citizens along multiple dimensions. The proposed method tries to investigate various factors of each transportation option including but not limited to the cost, times to be taken, and fuel consumption.

Model-based data interpretation has the potential to increase knowledge of structural behavior and support asset management. Models are usually conservative and contain many parameters and sources of systematic uncertainty, which need to be taken into account for accurate model updating. However, interpreting measurements using physics-based models is computationally expensive. Supplementing physics-based models with inexpensive surrogate models might facilitate practical implementation of data interpretation. In this paper, development of regularized linear-regression surrogate models for simulating structural behavior of a full-scale bridge and their use in error-domain model falsification for structural identification is presented. In this methodology, uncertainties from systematic sources and surrogate model error are considered explicitly during model updating. Results are verified with knowledge of parameters used to simulate measurements on a full-scale bridge. Use of simple regularized linear-regression models helps achieve accurate knowledge of updated structural behavior, which can then be used for making better asset management decisions.

Traditional service area analysis has been utilizing circular buffers or road network data and travel distance to create ‘reachable’ areas within set threshold time. The method is limited in terms of reflecting the real world with dynamic traffic conditions. In this study the authors propose data-driven service area analysis system with real taxi GPS data, which consists of dynamic fleet service area analysis for creating service area map with real traffic condition. Taxi GPS location data, collected in Seoul for more than 2 years, is used to create data-driven service area of each vehicles in dynamic time range from 5 minutes to 30 minutes. The process is conducted on Hadoop distributed computing system due to large data size (412 GB) and computation. Proposed data-driven dynamic fleet control system would allow fleet control of multiple vehicles for wider space coverage resulting in global optimization. The suggested system can be implemented in public safety fields which is a key issue in smart cities, where safety should be maintained by implementing various methods. For example, the system can be applied to police patrol which must cover wide area simultaneously with other patrol cars to maintain safety and to respond to crime within golden time. Accurate analysis of service area within threshold time, will allow more sophisticated patrol guide, thus maximizing the efficiency. In terms of private fields, upcoming autonomous vehicles could benefit from such system. By implementing the suggested method, autonomous vehicles will be able to calculate their own service areas and will provide optimized service in terms of car-sharing and driving.

Unmanned aircraft systems (UAS) can be used to collect visual assets such as 2D geo-referenced and infrared images. These 2D images can be converted into 3D point clouds through the UAS photogrammetry process. Such point clouds can then be utilized to detect obstructions around airports. This research describes the process of data collection, using UAS at airports to analyze point cloud data and document the potential implications of UAS-based airport inspections. Two field tests were conducted and three different UAS platforms were deployed. The Point Cloud Library (PCL) was employed to detect and measure trees around runways. The main contribution of this study is the provision of an overview of how UAS and visual assets can be used for airport inspections, in order to improve work efficiency and reduce human error.

In Japan, road markings are greatly damaged by snow-removal operations. In order to effectively maintain and manage the markings within limited budget and personnel, it is necessary to evaluate the damages frequently and quantitatively judge the repainting using a low-cost system. The purpose of this study is to develop a diagnostic system for road markings considering both visibility in the daytime and nighttime by using the road surface image photographed by the in-vehicle camera. First, the stripping ratio of road marking was estimated by image features. To evaluate the nighttime visibility, the reflectivity measured by equipment was estimated from the relationship with stripping ratio. Second, we conducted a questionnaire survey and clarified the relation between stripping ratio and daytime visibility. Finally, based on the above results, we have proposed the road marking condition index (RMCI) and the repainting criterion considering the daytime or nighttime visibility.

Finite element (FE) modeling is widely used to simulate the behavior of actual structures under various loading conditions. However, the simulation results from an FE model are usually different from an as-built structure. To improve the accuracy of the FE model, selected structural parameters are to be updated, which is known as FE model updating. This paper discusses the updating of an FE model based on a steel pedestrian bridge on Georgia Tech campus. By setting structural parameters as optimization variables, FE model updating is formulated as optimization problems to minimize the difference between experimental and simulated frequency-domain modal properties. Two modal property difference formulations are presented, one using modal assurance criterion (MAC) values and the other using direct difference between mode shapes. For each updating formulation, derivative Jacobian of the objective function is derived in detail. The MATLAB code and data are available on GitHub.

Because of frequent extreme weather conditions, accelerated disaster reconnaissance has become extremely important. In particular, surveying traffic sign damage and conditions has become essential for determining and prioritizing necessary repair/replacement. Under the research project sponsored by the National Academy of Sciences NCHRP-IDEA program, a conventional sign detection algorithm based on color, shape, and texture has been developed to process the images of signs. The developed algorithm has been enhanced by digital image processing and deep-learning methods, such as convolutional neural networks (CNN), generative adversarial networks (GAN), and region-based convolutional neural networks (RCNN). In this paper, a method is designed to process the images obtained using unmanned aerial vehicles (UAV), employing a model UAV, to further develop of our current research. The preliminary test shows that it is promising to use a UAV and machine learning to develop an expedited infrastructure condition evaluation following natural disasters because of its automatic and non-contact nature. The preliminary outcomes show the detection rates have satisfying FN rates and very low FP rates. Besides, the designed algorithm and data-gathering method provides a real-time and on-site computation capability that reduces the quantity of data to be stored by filtering out unnecessary data instantly.

Currently, sustainable development is one of the most critical challenges of the society, and rail transit causes multiple impacts on the environment, such as pollutant emissions, noise, and land use. Furthermore, sustainability research in the field of rail transport requires deep integration of multiple factors, including safety, punctuality, energy efficiency, and riding comfort. The present paper proposes a multi-agent system (MAS), in which trains can directly exchange information with other trains and make decisions. Compared with the centralized control approach, this MAS approach can reduce the computational load of the central controller and improve the efficiency of information exchange. Furthermore, a hybrid communication and decision-making system within the MAS enables each train to seek a sustainable driving strategy. The feasibility of the proposed MAS is tested by a case study, and the results show a good potential of this kind of system. The presented information is expected to promote the thinking of how to realize fully automatic train control approaches in the future.

Congestion impacts urban mobility, fuel consumption, and air pollution, so it is a critical societal issue. In this study, we have explored the relationship of transportation diversity—the availability of transportation modes (richness) and their distribution in a community (evenness)—on the traffic jam characteristics represented by congestion duration. In doing so, we have analyzed heterogonous data sources through a case study of the Washington D.C. metropolitan area. The participatory-sensing-based data, collected through Waze, a community-based navigation application, are utilized to examine the spatiotemporal patterns of congestion duration over a period of three months. Transportation diversity is measured at the scale of the zip code tabulation areas. Comparison of zip code tabulation areas regarding their transportation diversity demonstrates that areas with higher congestion duration time generally have limited diversity in their transportation system. The spatiotemporal distribution of these metrics could be used for performance-based planning and prioritizing resource allocation to the most critical areas.

The objective of this paper is to model and characterize the percolation dynamics in road networks during a major fluvial flooding event. First, a road system is modelled as planar graph, then, using the level of co-location interdependency with flood control infrastructure as a proxy to the flood vulnerability of the road networks, it estimated the extent of disruptions each neighborhood road network experienced during a flooding event. Second, percolation mechanism in the road network during the flood is captured by assigning different removal probabilities to nodes in road network according to a Bayesian rule. Finally, temporal changes in road network robustness were obtained for random and weighted-adjusted node-removal scenarios. The proposed method was applied to road flooding in a super neighborhood in Houston during hurricane Harvey. The result shows that, network percolation due to fluvial flooding, which is modelled with the proposed Bayes rule based node-removal scheme, causes the decrease in the road network connectivity at varying rate.

The sustainable success of public infrastructure projects is becoming a driving force of the advancing urbanization boom in developing countries globally. Making informative decisions from sustainable infrastructure designs has been drawing considerable attention in the research community of infrastructure sustainability (IS). However, integrating knowledge of decision-makers (DMs), who are multidisciplinary professionals with over differentiated expertise, encounter uncertainties in representing decision information due to their various preferences. This study, therefore, has developed an integrated multidisciplinary decision-making (MDM) methodology leveraged by a five-step modeling technique for ranking analysis of sustainable infrastructure design alternatives. It comparatively combines different preferences of multidisciplinary DMs in representing decision information among intuitionistic fuzzy, linguistic, and mixed decision environments. The integrated MDM methodology is then validated by an illustrative example, which is assumed to simulate the collaborative ranking analysis for sustainable infrastructure designs. The validation findings demonstrate the capability of this proposed integrated MDM methodology in deriving objective and consistent ranking results in a multidisciplinary group level of detail. This research contributes to a reasoning methodology of managing representation uncertainties of decision information from multidisciplinary DMs with various preferences for sustainable infrastructure developments.

Disruption of transportation is one of the largest contributors to social and economic losses due to disasters. An essential aspect of regional resilience is the ability of transportation networks to maintain functionality under stress. In the U.S., transportation networks in seismic regions are deemed especially vulnerable because of the poor current condition of a significant portion of their bridges. This vulnerability necessitates a quantified investigation of transportation network perturbation due to anticipated major seismic events as well as the recovery process from such perturbations. This information is crucial for regional planning and policy-making towards enhanced resilience in the long run, and for emergency response during a disaster. In this study, data and modeling needs are identified to advance this research area beyond the boundaries drawn by trade-offs of scale and granularity. A preliminary framework is also presented.

Understanding the states of spatiotemporal flux in cities is critical for risk-informed decision making in disaster situations, resulting in crucial influences on physical, environmental, and social systems of cities. This paper presents a new framework to better understand dynamic spatiotemporal fluctuations related to the vulnerability of cities. We first leverage crowdsourced visual data to automatically identify and localize potential risks associated with vulnerable objects in cities before/during/after disasters. Then we bring multimodal sensing-based reality information into a 3D city model and an interactive computer-aided virtual environment (CAVE), which facilitates to interact the spatial information of vulnerable objects at the intersection of reality-virtuality of cities. For evaluation, case studies were conducted on Houston, TX. The resulting digital twin city reflecting multimodal sensing-based vulnerability information of cities has the potential to be used as a basis for simulating what-if scenarios for risk-informed decision making in disaster situations with enhanced reliability of analytics.

Understanding the spatial and temporal distribution of energy consumption in cities is critical to facilitate the identification of potential energy saving opportunities and planning of new renewable and integrated district energy systems. Previous work analyzing urban building energy usage has been largely limited to either modeling of individual buildings at granular temporal scales (i.e., hourly or less) or an entire stock of urban buildings at the yearly temporal scale. While such analyses are valuable, their lack of both spatial and temporal granular modeling limits their applicability in planning and design of integrated district energy systems. This paper proposes a new urban building energy model that produces hourly demand profiles for the building stock of New York City (NYC) using only open publicly available data. First, we utilize a machine learning model to predict annual energy consumption of NYC’s entire building stock from a subset of buildings that have publicly available annual energy usage data. We validate this part of the model using city-wide electricity data from New York Independent System Operator (NYISO). Results show that random forests have the best building-level prediction accuracy with a mean log squared error of 0.293. Next, we apply a novel optimization algorithm to construct temporal granular hourly profiles using the Department of Energy's commercial and residential simulation building reference sets, and the predicted annual energy values from the random forests model. Results indicate that we are able to achieve an error rate of ~10% (MAPE) in comparison to the overall hourly electricity profile of NYC. Moreover, we found that our iterative approach demonstrates that error rates diminish as buildings are added to the aggregated profile, which underscores the merits of applying our proposed method to model the entire building stock of a city rather than an individual building. In the end, our proposed method takes the first step of large-scale spatial and highly granular temporal characterization of urban building energy usage.

The major difficulties associated with vibration monitoring of in-service systems lie in the environmental and operational effects which exert a strong influence on the dynamic properties of such systems. Within this context, the problem of response prediction for condition diagnostics and control application becomes an intricate task which needs to be based on estimation methods able to account for system variability. This paper proposes a two-step Bayesian framework for input-state and parameter estimation of time-varying systems. The proposed approach employs a time-series model in order to firstly track the evolution of modal properties and subsequently deliver them to a modally reduced-order model for joint input-state estimation on the basis of a limited number of vibration measurements. The latter is carried out using an output-only Bayesian filter and the entire framework is demonstrated on a simulated case study.

Prolonged drier seasons increase the frequency of wildfires in southwestern U.S. and have become a major concern for infrastructure decision-makers. Changes in watershed triggering excess runoff and sediments in post-wildfire rainfall can impact dams, reservoirs, and floodplains. Excess runoff and sedimentation could result in a dam failure and downstream flooding. This study proposes a framework to identify and quantify post-wildfire flood impacts on earth dams and their reservoirs by incorporating post-wildfire rainfall outcomes, attributes of earth dam and reservoir, and floodplain occupancy. Additionally, the framework uses multi-criteria and Pareto optimization to estimate the critical wildfire affected area of a watershed. The proposed framework is demonstrated on an earth dam in the southwestern U.S. The results suggest that when wildfire spreads over 21% of watershed area, a post-wildfire rainfall can have significant impact on the earth dam. Such knowledge can aid decision-makers to evaluate mitigation responses to reduce impact on stakeholders.

Normative messaging interventions have proven to be a cost-effective strategy for promoting pro-environmental behaviors. The effectiveness of normative messages is partially determined by how personally relevant the comparison groups are as well as the lag of feedback. Using readily available energy use data has created opportunities to generate highly personalized reference groups based on households’ behavioral patterns. Unfortunately, it is not well understood how data granularity (e.g., minute, hour) affects the performance of behavioral reference group categorization. This is important because different levels of data granularity can produce conflicting results in terms of group similarity and vary in computational time. Therefore, this research aims to evaluate the performance of clustering methods across different levels of temporal granularity of energy use data. A clustering analysis is conducted using one-year of energy use data from 3,000 households in Holland, Michigan. The clustering results show that behavioral reference groups become the most similar when representing households’ energy use behaviors at a six-hour interval. Computationally, less granular data (i.e., six and twelve hours) takes less time than highly granular data which increases exponentially with more households. Considering the enormous scale that normative messaging interventions need to be applied at, using less granular data (six-hour intervals) will permit interveners to maximize the effectiveness of highly personalized normative feedback messages while minimizing computation burdens.

Building a disaster resilient community is a shared responsibility among multi-sector stakeholders, including government officials, private sectors, non-government organizations (NGOs), and community residents. Different stakeholders have different concerns and needs, and they make different decisions regarding the priorities of implementing disaster resilience strategies given limited resources. These differences are deeply rooted in their value systems, which are defined as ranked systems of things that are of importance, merit, and utility to the stakeholders. The differences in their value system cause conflicts and disputes for disaster-related decision making, which may result in longer decision-making time and huge financial losses. Therefore, there is a need to understand and systemically integrate the value systems of multi-sector stakeholders to support more effective and collaborative decision making in disaster resilience. To address such need, this paper proposes a stakeholder value aggregation (SVA) model to systematically and quantitatively aggregate divergent stakeholder value systems by leveraging a reinforcement learning-based method. The proposed SVA model was illustrated using a case study that focuses on aggregating the value systems of responsible and impacted stakeholders. The paper also discusses our proposed framework for a cloud-based decision support system that allows for automatic aggregation of divergent stakeholder value systems. The model, together with its potential application of cloud-based decision support system will facilitate more collaborative and efficient decision making by identifying the collective value priorities on disaster resilience practices.

This paper presents a sustainable disaster recovery decision-making framework that meets the needs of the stakeholders in rebuilding damaged infrastructure while simultaneously decreasing the built environment’s social, economic, and environmental vulnerabilities. Through utilizing agent-based modeling (ABM), the needs and preferences of the multi-sector stakeholders are integrated into the objectives of the modeled recovery agency. The proposed framework integrates well-established vulnerability indicators into the recovery agency’s objective function to better guide the redevelopment process. Through utilizing an individual learning module, the model optimizes the post-disaster recovery strategies to decrease the vulnerabilities of the built environment throughout the recovery phase while increasing the overall recovery rate. The proposed model was tested on the post-Katrina recovery in Mississippi. In comparison to the actual recovery data and simulation scenarios, the proposed model provided higher recovery rates and elevated the built environment vulnerability status to achieve long-term sustainable disaster recovery.

Operations as well as health and safety aspects during the maintenance of off-shore wind farms are connected within a complex process that is highly dependent on environmental conditions and the chosen maintenance strategy. Health and safety restrictions must be carefully balanced with economic performance parameters and local environmental conditions. To explore this balance, we developed a simulation model for different operation and maintenance strategies that provides information about accessibility of wind turbines, technical availability of the wind farm, and the mean time to repair. The model represents the maintenance process using discrete events steered by stochastic environmental and process related input parameters. It is built upon the abundance on existing maintenance simulation models for off-shore wind farm and can therefore be considered as stable enough to explore the specific effects of health and safety modeling aspects. The model implements health and safety aspects using a rule-based decision system that defines certain thresholds beyond which maintenance processes will not be carried out. These thresholds are based on existing health and safety standards for off-shore windfarm maintenance routines. To illustrate the model, we implemented it for a fictional offshore wind farm. Initial simulation experiments show that the model allows to quickly understand the effects of different choices on several key performance indicators.

This paper proposed and tested a multilayer framework for modeling network dynamics of inter-organizational coordination in resilience planning among interdependent infrastructure sectors. Each layer in the network represents one infrastructure sector such as flood control, transportation, and emergency response. Coordination probability was introduced to approximate the inconsistent coordination between organizations, based on which the intra-layer or inter-layer link removal was conducted and inter-organizational coordination efficiency within and across infrastructure sectors was hereby unveiled. To test the proposed framework, a multilayer collaboration network of 35 organizations from five infrastructure sectors in Harris County, Texas, was mapped based on a survey of Hurricane Harvey. The analysis results showed that before Hurricane Harvey, coordination among flood control, transportation, and infrastructure development sectors lacked essential integration to foster robust resilience plans. The proposed framework enables an assessment of coordination efficiency among organizations involving in resilience planning and provides an indicator for urban resilience measurement.

Current crisis informatics research using social media to identify human location and behavior has neglected to identify which populations may be prevented from using social media during a disaster. Unfortunately, most crisis informatics focuses on either individual postings or large-scale analyses of signal changes and is potentially overlooking or misrepresenting vulnerable populations. To assess this, we utilize scaled spatial nets and disaggregated population and vulnerability index data in order to identify relationships between social media signal deviation and vulnerable populations within areas that experienced substantial infrastructural damage from Hurricane Harvey. Many studies have identified a strong positive correlation between areas experiencing high amounts of hurricane damage and increased Twitter activity; however, we find that highly damaged areas with more elderly people, disabled people, and people without access to vehicles instead have a significant negative correlation with Twitter activity during disaster. In defining this relationship, we show that some vulnerable populations have decreased instead of increased social media visibility during disasters. These findings identify demographic inequalities in the scalar representation of data from humans-as-sensors in disaster.

Remote sensing is gaining in popularity for structural condition assessment, and in current practice, engineers are collecting large amounts of data throughout a system’s life-cycle. These data must be interpreted to characterize defects, and to help engineers estimate a structure’s remaining life. The challenge is that remotely-sensed defects are not typically quantified in a manner that robustly conveys life-cycle dynamics. We propose that the geometric concept of the convex hull can help address this limitation. The convex hull is the smallest convex polygon that surrounds all points in a set, regardless of dimensionality, and is a unique representation of the set. Describing a defect in terms of its parameterized hull enables consistent temporal tracking for predictive purposes, while implicitly reducing data dimensionality and complexity as well. Furthermore, the concept of the convex hull can be extended to high-dimensional spaces, supporting the fusion of multiple sensors and data types. In this study, 2D point clouds analogous to information derived from point clouds were generated over simulated life-cycles. The evolutions of point cloud hull parameterizations were modeled as stochastic dynamical processes via vectorized autoregression and compared against ground truth. The results indicate that this convex hull approach provides consistent and accurate representations of defect evolution across a range of defect topologies and is reasonably robust the noisy measurements, however assumptions regarding the underlying dynamical process play a significant the role in predictive accuracy. Longer term, the results of this work will support finite element model updating for predictive analysis of structural capacity.

To enhance the overall resilience of interconnected CIs as a system-of-systems (SoS), it is important to model the complex inter- and intra-dependencies using an integrated framework. Network flow model is one important approach in studying infrastructure interdependencies. To address some limitations in existing studies of infrastructure network flow models, this study proposed an integrated flow-based model that captures the heterogeneity of nodes, different levels of restoration of disrupted nodes and arcs, and both intra- and inter-dependencies in an infrastructure SoS. The effectiveness of the proposed model is tested in a case study consisting of intra- and inter-dependent electricity, coal, oil, and gas infrastructure in the state of Connecticut.

In the U.S., the building sector consumes 70% of electricity and lays massive pressure on national grids. To avoid electricity blackouts, demand response programs incentivize end-consumers for reducing their electricity demand during peak hours. Therefore, it is essential for grid operators to understand the electricity shaving capacity of buildings. However, previous studies either simplify buildings as black-boxes—resulting in low accuracy in estimations, or represent buildings with detailed information—resulting in over-parameterized models. In this study, the authors provided a computational framework to identify principal factors that dictate peak shaving capacities of buildings. In total, fifteen buildings during twelve DR events were used as testbeds as validation. The results showed that the day-in-the-week and the quantity of relevant equipment are part of the principal factors behind peak capacity determination. With this framework, practitioners can represent buildings beyond black-boxes with less complexity and promising accuracy of peak shaving capacity determination.

The transportation system in urban areas plays a critical role in evacuation and resource supply during disasters. Transportation diversity, i.e., the availability and distribution of transportation modes across a community, impacts the mode complementarity, which can affect population movements and connectivity among spatial units—especially during extreme events when a failure in one mode can be compensated by other modes. However, few quantitative efforts have explored the impact of physical infrastructure configurations, and more specifically transportation diversity, on urban mobility and connectivity in an actual disaster context. In this paper, we measure transportation diversity with two factors—functional richness and functional evenness translated from the concept of diversity in ecological systems. Disaster resilience is regarded as a dynamic process pre, during, and post disaster in distinct communities and is tracked and quantified based on geospatial Twitter data of population cross-ZIP code tabulation areas movements with an ecology-inspired method—Fisher information—over time. Specifically, we compared the resilience of different neighborhoods in the city of Houston in Hurricane Harvey and found that neighborhoods with a higher level of transportation diversity demonstrated higher resiliency in their mobility during and after the hurricane. This study contributes to understanding how the availability and distribution of transportation infrastructure can impact the disaster resilience of human-infrastructure systems in urban areas. The results can improve transportation design and future urban planning, especially in the context of climate change and natural hazards.

Operationalizing resilience principles, in particular when interdependency is considered, remains a major challenge. Using two interconnected infrastructure systems in an urban district of Hong Kong as case study, this paper conducts a precursor vulnerability analysis and investigates their resilience and interdependency in respond to weather-related events. The study incorporates relevant infrastructure information using the ArcGIS platform and integrates the physics-driven vulnerability analysis methods from the two distinct knowledge domains. Cascading effects are also delineated using an empirical nominal damage function. The research findings show that the incapacity of vulnerable components in one infrastructure system could degrade the performance of sub-systems of other infrastructures, and even lead to their network-level resilience performance deterioration. This research combines the multiple approaches for quantitative infrastructure vulnerability analysis and lays technical foundation for subsequent restoration and adaptability analysis of infrastructure resilience, and provides an interactive and integrative environment for resilience analysis from multidisciplinary perspectives.

Hospitals play a crucial role in the mitigation and recovery of earthquake-hit regions, which are expected to be resilient enough so as to provide medical care that is badly needed in the aftermath of earthquakes. Nevertheless, there lack appropriate quantitative assessment approaches of hospital earthquake resilience, which makes it challenging for devising and benchmarking resilience enhancement measures. In this paper, a quantitative assessment approach of hospital resilience to earthquakes considering the physical damage and the recovery of hospital functionality is proposed. A system dynamics (SD) model is developed to simulate the hospital functionality hence enabling quantification of hospital resilience to earthquakes. The efficacy of the proposed approach is validated in a case study. The proposed approach can provide a tool for hospitals to quantitatively assess their resilience to earthquakes so as to propose targeted resilience enhancement measures for future earthquakes.