Jung-Ho Lewe’s research while affiliated with Georgia Institute of Technology and other places

This study proposes a deep learning model utilizing the BACnet (Building Automation and Control Network) protocol for the real-time detection of mechanical faults and security vulnerabilities in building automation systems. Integrating various machine learning algorithms and outlier detection techniques, this model is capable of monitoring and learning anomaly patterns in real-time. The primary aim of this paper is to enhance the reliability and efficiency of buildings and industrial facilities, offering solutions applicable across diverse industries such as manufacturing, energy management, and smart grids. Our findings reveal that the developed algorithm detects mechanical faults and security vulnerabilities with an accuracy of 96%, indicating its potential to significantly improve the safety and efficiency of building automation systems. However, the full validation of the algorithm’s performance in various conditions and environments remains a challenge, and future research will explore methodologies to address these issues and further enhance performance. This research is expected to play a vital role in numerous fields, including productivity improvement, data security, and the prevention of human casualties.

Organizations continuously develop Digital Twins across a wide number of applications and industries. While this represents a testimony to the benefits and opportunities Digital Twins provide (AIAA 2022), their development, maintenance and evolution still face major challenges (Bordeleau et al. 2020). Most past and current efforts focusing on the development of Digital Twins have relied on ad‐hoc approaches, where most of the efforts start with building models without properly framing the problem (Martin 2019, Lu et al. 2020). More importantly it has led to the development of models that provide a solution to the wrong problem and consequently fail to address the core questions and needs of the stakeholders (Martin 2019)]. The development and implementation of Digital Twins also lack standardization (Shao 2021), instead relying on bespoke methods and technologies (Niederer et al. 2021), which in turns leads to a lack of consistency in their description and implementation, as well as limited interoperability across applications, tools and disciplines (Niederer et al. 2021, Piroumian 2021). The development of Digital Twins has been plagued by a lack of scalable approaches, leading to implementations that are highly specialized and require considerable resources in terms of subject matter expertise (Niederer et al. 2021). The development of standardized methodologies has been identified as a means to unleash the full potential of Digital Twins (Shao 2021, Piroumian 2021) and increase their adoption across a wider range of disciplines and applications (Niederer et al. 2021). To that end, this paper presents a domain agnostic and descriptive reference model, which builds on recognized industry practices and guidelines, to support the planning, description, and analysis of Digital Twins. It also introduces real‐world examples of aerospace Digital Twin use cases and discusses how the generic reference model supports the various use case applications. Finally, this paper briefly discusses recommendations and next steps on how to realize value from Digital Twins more broadly.

View Video Presentation: https://doi.org/10.2514/6.2021-3363.vid Net positive water and energy buildings aim to generate or harvest more water and energy than they consume. Buildings pursuing this goal typically do so through the generation of energy using photovoltaic (PV) arrays and through the collection of rainwater using the building’s rooftop and a large cistern for storage. As PV systems are limited by their intermittency in power generation, energy storage systems (ESS) are being investigated to augment the availability of their generated power. Many ESS have been investigated, but little previous work has looked into the potential use of rainwater harvesting cisterns as passive thermal storage reservoirs for thermal-electric energy exchange. This work performed a simple investigatory study into this idea using actual data from an academic building located on Georgia Tech’s Atlanta campus. An energy and water building performance model was coupled with a cistern thermal energy storage model. The sensitivity of the building’s PV array area, cistern volume, maximum cistern temperature, and thermal-electric conversion efficiency are presented. The addition of a high specific heat material to the cistern is also evaluated and the impact of this addition is presented. Preliminary results suggest that an academic building may save up to 58% of its supplementary electrical demand by adopting a passive cistern thermal energy storage system.

District cooling systems are growing in prevalence as they offer building fleets significant cost and energy savings compared to traditional systems. Their analysis requires the integration of hydraulics and thermodynamics to capture their overall behavior. This work aimed to develop an integrated model for simultaneously representing the hydraulic and thermodynamic performance of a chilled water network in consideration of the heat exchange occurring within each of the buildings in its fleet. Real data from the Georgia Institute of Technology was used to validate the model across key buildings and applied to the entire chilled water network, and a procedure for calibrating the model was introduced. This approach made the developed model both scalable and agile. As a simple demonstration of the model’s capabilities, a plant-centric optimization method was compared to a more holistic approach in which the optimum was selected in consideration of both the chiller plant and its network of buildings. The finding challenges the traditional heuristic by which these systems are operated and opens up opportunities for achieving higher energy efficiency.