Advances in Digitalization and Machine Learning for Integrated Building-Transportation Energy Systems

By Peter D. Lund

Advances in Digitalization and Machine Learning for Integrated Building-Transportation Energy Systems examines the combined impact of buildings and transportation systems on energy demand and use. With a strong focus on AI and machine learning approaches, the book comprehensively discusses each part of the energy life cycle, considering source, grid, demand, storage, and usage.

Opening with an introduction to smart buildings and intelligent transportation systems, the book presents the fundamentals of AI and its application in renewable energy sources, alongside the latest technological advances. Other topics presented include building occupants’ behavior and vehicle driving schedule with demand prediction and analysis, hybrid energy storages in buildings with AI, smart grid with energy digitalization, and prosumer-based P2P energy trading. The book concludes with discussions on blockchain technologies, IoT in smart grid operation, and the application of big data and cloud computing in integrated smart building-transportation energy systems.

A smart and flexible energy system is essential for reaching Net Zero whilst keeping energy bills affordable. This title provides critical information to students, researchers and engineers wanting to understand, design, and implement flexible energy systems to meet the rising demand in electricity.

• Introduces spatiotemporal energy sharing with new energy vehicles and human-machine interactions
• Discusses the potential for electrification and hydrogenation in integrated building-transportation systems for sustainable development
• Highlights key topics related to traditional energy consumers, including peer-to-peer energy trading and cost-benefit business models

• Language: English
• Publisher: Elsevier Science
• Release date: Nov 21, 2023
• ISBN: 9780443131783

Book preview

Chapter 1

Smart buildings and intelligent transportations with artificial intelligence and digitalization technology

Deng Pan¹ and Yuekuan Zhou¹,²,³,⁴,    ¹Sustainable Energy and Environment Thrust, Function Hub, The Hong Kong University of Science and Technology (Guangzhou), Nansha, Guangdong, P.R. China,    ²Department of Mechanical and Aerospace Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong SAR, P.R. China,    ³HKUST Shenzhen-Hong Kong Collaborative Innovation Research Institute, Futian, Shenzhen, P.R. China,    ⁴Division of Emerging Interdisciplinary Areas, The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong SAR, P.R. China

Abstract

This chapter provides a holistic overview of smart buildings and intelligent transportations with artificial intelligence and digitalization technology.Achieving carbon neutrality and sustainability calls for smart buildings and intelligent transportation systems. Digitalization is now becoming a new paradigm in the buildings and transportation sectors. To achieve the carbon peaking and carbon neutrality, smart buildings with energy integration under e-mobility frameworks play a key role in renewable penetration and grid independence. A comprehensive literature is reviewed and discussed on the smart buildings and intelligent transportations with artificial intelligence and digitalization technology In both academic and industrial circles, the application of artificial intelligence and digitalization technology is regarded as a competitive technique for smart buildings integrated with intelligent transportations.

Keywords

Smart buildings; intelligent transportations; artificial intelligence; digitalization technology

1.1 Introduction

Due to low energy efficiency, buildings currently account for over 30% of global energy consumption and one-third of carbon emissions [1]. Buildings and transportation contribute to a significant proportion of total energy consumption. Achieving carbon neutrality and sustainability calls for smart buildings and intelligent transportation systems. Digitalization is now becoming a new paradigm in the architecture and transportation industries [2]. To achieve carbon peaking and carbon neutrality, smart buildings with energy integration under e-mobility frameworks play a significant role in renewable penetration and grid independence. Furthermore, artificial intelligence (AI) and digitalization technology can promote a high-efficiency transition in smart buildings and transportation. In both academic and industrial circles, the application of AI and digitalization technology is regarded as a competitive technique for smart buildings integrated with intelligent transportations. This chapter provides a holistic overview of smart buildings and intelligent transportations with AI and digitalization technology.

1.2 Smart buildings

Buildings play a key role in energy consumption in cities. To improve energy efficiency, smart buildings utilize digitalization technology [3], which includes six basic characteristics: automation, multifunctionality, adaptability, interactivity, efficiency, and intelligence [4]. Early smart building research only focused on the automated operation of building services systems, such as heat pumps, chillers, lighting, escalators, and so on [5]. Nowadays, with the deployment of renewable energy supply onsite and the role change from energy consumer to energy prosumer, peer-to-peer (P2P) energy sharing has attracted widespread interest among occupants and building owners. The increasing demand for sustainable and green buildings has directed the industry toward the development of smart buildings [6]. Energy savings generally sacrifice thermal comfort conditions. Quantifying building energy performance through the development and utilization of key performance characteristics is essential to achieving smart building goals [4]. The learning and application ability of buildings is considered one of the key performance indicators of smart buildings, which can improve energy performance over time based on accumulated experience (e.g., training data) [7]. Many new technologies are used in smart buildings, which mainly focuses on HVAC (heating, ventilation, and air conditioning) systems, energy flexible buildings, and P2P energy sharing.

1.2.1 Heating, ventilation, and air conditioning systems

HVAC systems account for almost 31% of the energy consumption in buildings [8]. HVAC systems are key areas to improve energy efficiency and reduce carbon emissions of smart buildings. Most studies mainly focus on energy management [3,9] and smart controls for HVAC systems. In terms of control strategies for HVAC systems of smart buildings, advanced control strategies include model predictive control and adaptive control methods [10]. A new HVAC control system with a multistep predictive deep reinforcement learning algorithm for smart buildings is proposed to save energy costs of the HVAC system while ensuring occupants’ thermal comfort [11]. A two-stage conditional value-at-risk model is developed to optimize the day-ahead dispatch of smart buildings with HVAC systems [3]. Li et al. [12] proposed an event-driven multiagent-distributed optimal control strategy based on cyber-physical systems for HVAC. Chen et al. [13] presented a transfer learning model to design the HVAC control system and natural ventilation in a new smart building, which can reduce costs in control system design and commissioning. Chaouch et al. [9] proposed a smart method with fuzzy logic and machine-to-machine communication to control HVAC systems in smart buildings. Results show that the energy management system (EMS) can save 16% of annual energy consumption on average.

1.2.2 Energy flexibility

The energy flexibility of smart buildings refers to their capability in managing onsite renewable energy and energy demand in accordance with climate conditions, building demand, and grid requirements [14–16]. A new hierarchical optimization framework is designed to increase the joint flexibility of smart building communities [17]. A smart building is designed so that each power user has a flexible contract power to decrease electricity costs [18]. Energy flexibility is designed by thermal photovoltaics integrated with thermal storage systems for smart buildings to achieve smart energy management [19].

More and more smart appliances are being utilized in smart buildings. Thus the power load of smart buildings is more complicated. To precisely adjust the building power loads, a regulation structure using the load control method is developed [20]. Smart appliances can be dispatched spontaneously and harmonized with the EMS in a smart building, and the usage patterns of some flexible appliances (such as washing machines) can be used to shift loads by adjusting the operating time [21]. A multiobjective mixed-integer model using the weighted addition method is proposed for smart home dispatch, which can reduce 28.76% of operational costs and 11.37% of carbon emissions [22].

1.2.3 Peer-to-peer energy sharing

From the perspective of energy, prosumers in buildings with an onsite renewable power supply can be self-sufficient [23]. The energy sector is transforming from centralized energy systems to decentralized energy systems, which promotes the generation of P2P energy-sharing modes [24]. Renewables are often unstable and produce much energy surplus, which can be shared within the smart community [25]. Blockchain is being utilized to stimulate P2P energy trading due to its transparency, security, and high efficiency [26]. P2P energy sharing plays a key role in the renewable penetration ratio and energy flexibility of smart buildings. P2P energy sharing will provide additional energy flexibility and maximize profits for smart buildings by optimizing energy usage. Trading algorithms have been applied in the P2P energy trading market of smart energy communities to improve economic performance. Park et al. [27] designed a novel P2P trading mechanism to guarantee both flexibility and stability of energy supply within a smart community. Alam et al. [28] evaluated the energy costs’ impact of P2P energy sharing among smart homes and proposed a near-optimal optimization algorithm to eliminate the unfair distribution of costs. Qiu et al. [29] proposed federated reinforcement learning to improve the economic and environmental performance of smart buildings when combined with P2P energy/carbon trading systems. Results show that it can reduce 5.87% of total energy costs and 8.02% of carbon emissions cost. Cutsem et al. [30] proposed a decentralized framework to manage the energy sharing of intelligent buildings using blockchain technology, which can be utilized within 100 intelligent buildings for a district community. Zhou et al. [31] adopted a P2P smart community energy pool and a user-dominated demand side response to improve usage efficiency and reduce the energy costs of an intelligent community.

1.3 Intelligent transportations

Transportation, which plays a key role in human society, is also one of the main sources of CO2 emissions. However, the decarbonization potential in transportation is promising. With the increase in urban population, transportations face significant challenges, such as traffic congestion, safety, accidents, pollution issues, and energy consumption [32,33]. To solve these issues, new energy vehicles and energy savings in transportation are significant to decarbonize. Intelligent transportation systems have been rapidly developing in recent years, especially in fuel energy savings and decarbonization. Intelligent transportation systems have become one of the most important ways of decreasing traffic congestion, carbon emissions, and energy consumption [34]. Conventional intelligent transportation systems, include smart traffic lights, smart junction management, and intelligent traffic paths [35]. A smart EMS based on the road power demand model is utilized to decrease fuel consumption for transportations [33]. Zhao et al. [36] investigated the impact of intelligent transportation on carbon emissions in China. Results show that a 1% growth in intelligent transportation in a province can reduce the carbon emissions of the local and neighboring provinces by 0.1572% and 0.3535%, respectively. A multiscale carbon emissions computing platform is proposed for intelligent transportation to provide green transportation guidelines [37]. The impact of smart transportation systems on energy savings and decarbonization is investigated. Cui et al. [38] investigated the impact of different pathway guidance strategies on carbon emissions of smart transportations. Results show that a smoother path and a strategy with traffic data of the overall path will contribute to the decarbonization of intelligent transportation. Smart transportation systems use AI and digitalization technology to improve energy efficiency and decrease traffic jams, accidents, and emissions. Dynamic pricing mechanisms have a significant effect on smart transportation systems. An inappropriate dynamic pricing mechanism may result in more serious traffic stagnation, environmental pollution, and larger energy consumption [39]. The results of energy conservation and decarbonization are becoming primary rationales for investments in smart transportation [40].

Studies of intelligent transportation are also focusing on reducing traffic congestion. Smart transportation systems can plan highly effective real-time routes to reduce traffic congestion though smart communications [41]. Intelligent transportation utilizes ways of accessibility and emergency response for traffic control to solve traffic problems [42]. Rawashdeh et al. [43] established a communication framework to deal with intrusion threats to smart sensors of intelligent transportation systems. Traffic prediction based on machine learning can help plan pathways and control traffic congestion [44].

The digitization of the intelligent transportation system calls for the security and reliability of identity management and authentication [45]. The management and sensing systems of smart transportation may suffer from cyber-attacks [46]. Shari et al. [47] proposed a secure data dissemination project based on blockchain for intelligent transportation to achieve the security needs of reliability, privacy, and liability together.

Moreover, vehicular middleware and heuristic methods are utilized for intelligent transportation systems to achieve high mobility in transport and low postponement of the system [48]. Millimeter wave technologies are used to transfer large amounts of data efficiently for smart transportation [49]. Trading travel credit schemes for traffic jams, income protection, and Pareto-improving methods for peak-hour transit demand management are used for demand management strategies for intelligent transportation [50]. Nanogenerators-based self-powered sensors may be used in place of many sensors of intelligent transportation, which can accelerate the process to achieve complete smart transportation [51]. Intelligent transportation planning helps to build the optimal path to reduce traffic jams based on information, techniques, and models [52]. Apps for intelligent transportation can guide traffic with AI methods. AI with big data is used to construct a smart intelligence system. Traffic jams and path optimization selection utilize machine learning algorithms for prediction [53]. A smart transport system for the intelligent internet of vehicular network traffic is presented with tree-based machine learning models [54].

1.4 Artificial intelligence and digitalization technology

With the rapid development of information technology, AI and digitalization technology have been extensively used in all professions and trades, bringing innovation for smart buildings and intelligent transportations [55]. AI is considered a vital point for economic growth worldwide. Most countries invest a great deal of funds to develop AI and digitalization technology for the advancement of human society. The application of AI and digitalization technology in the building industry brings a new era for the development of design and construction companies. AI has been widely utilized in building service systems, mainly in demand forecasting and intelligent controls. Machine learning is used for solar radiation prediction [56]. The reinforcement learning approach can be used for smart charging/discharging on PCM (phase change material) storage to achieve more energy savings [57]. Anooj et al. [58] proposed a machine learning approach with a diagnosis for a PCM thermal management system to forecast liquid fraction. Zhou et al. [59] used supervised machine learning and heuristic optimization algorithms to optimize the design and operation of PCM coupled with a renewable system.

Machine learning for suitable AI applications in BIM models and combined optimization of topological rule inference will be the future trends [60]. The adoption of deep learning (DL) against the background of digital twins can drive the development of intelligent cities [61]. To accomplish more complex jobs in the modern world, intelligent industrial products need to be established with greater flexibility and adaptability. A digital twin provides for flexibility and adaptability [62]. Virtual sensing technologies play important roles in smart services in lifecycle construction and digitalization [63]. Virtual sensors are used in the HVAC systems of smart buildings for energy consumption monitoring, and they have also been broadly utilized in intelligent transportations for driving behavior and tracking. AI is applied to detect and diagnose building services systems using a labeled time series [64].

AI and digital twins are meaningful for the energy industry, producing profound effects on sustainable energy and the environment. The utilization of AI in smart buildings is an effective way of improving environmental and energy performance. AI and digitalization provide new chances for the transition from the conventional energy industries and producing new technologies that can improve the economic resilience of the energy sector [65]. AI technologies (such as big data, IoT, blockchain, cloud data storage, machine learning, and DL) help control energy data centers and P2P energy sharing, which have been increasingly utilized in the performance prediction of nonlinear energy systems [66]. Electricity system operators and utilities are widely applying AI techniques to cover daily operations and onsite service operations, which can optimize the operation of the power networks and enhance the flexibility, reliability, and efficiency of the grid [67]. A 3D indicator is proposed for AI utilization in the energy industry in terms of maturity level, supervisory risks, and potential benefits [68]. AI-generated renewables can learn from bioinspired lessons and provide smart energy systems for the carbon neutrality transition