Search

Lithium-ion Batteries (LiB) have a wide range of applications in daily life. However, as they get used over time, battery degradation becomes inevitable, which can lead to a drop in performance and a reduction in the battery's cycle life. The State of Health (SoH) is widely regarded as the health indicator for the battery pack. In Electric Vehicle (EV) applications, the EV user defines the lower limit of SoH when they experience that the battery no longer supports the EV; at that point, the battery is said to be translated from first life to second life. The SoH estimations of Second Life Batteries (SLB) have plenty of uncertainties, such as the availability of battery's previous history, non-uniform degradation in the EV application, variations in chemistry, and charging protocols defined by vehicle manufacturers, making the SoH estimation of SLB a challenging task. This paper discusses the equipment, timelines, computational complexity, health indicators, and list of parameters that need to be considered for the SoH estimation of SLB. The SoH estimation methods are classified into direct and indirect techniques. Direct assessment techniques involve cyclic ageing experiments followed by dismantling the battery for microscopic studies performed by previous researchers that were explained. Indirect assessment techniques include physical and chemical based approach, electrical, and Artificial Intelligence (AI)-based methods that estimate SoH indirectly through incremental, differential approaches and other parameters such as Integrated Voltage (IV) and Probability Density Function (PDF). Health indicator identifications play a vital role in indirect assessment methods to gain critical insights regarding battery degradation. The challenges involved in SoH estimation are categorized into equipment requirements, parameters, SoH accuracy and efforts required to compute SoH, which are discussed. Of all the SoH estimation methods, comparison of such methods in First Life Batteries (FLB) and SLB perspectives are discussed. To estimate the SoH of SLB, this paper explains all aspects, such as computational methods, filtering data, data sampling frequency, and the need for a specific algorithm to post-process the battery test data. Equipment availability and timelines are interrelated with the cost incurred in the SoH estimation of SLB. The efficacy and practicality of SoH estimation methods that are proposed for SLB is discussed. Overall, this paper provides necessary insights into the parameters required for SoH estimation and the computational and experimental methods that can be considered for estimating the SoH of SLB while some of the methods are applicable to FLB as well. • Review of State of Health (SoH) estimation methods for lithium-ion battery pack translating from first life to second life. • Critical analysis of equipment's and test protocols subjected to cyclic ageing. • Classification of SoH estimation methods in the form of physical and chemical based approach, electrical and Artificial Intelligence (AI) based techniques. • Listed the parameters acquired from battery to be considered in second life SoH estimation. • Presented the challenges associated with SoH methods for Second Life Batteries (SLB). [ABSTRACT FROM AUTHOR]

The motivation of this paper is to investigate the resistance of countries' energy research and development (RD&D) expenditures to random shocks. The analysis includes the five countries (France, Germany, Japan, the United States, and the United Kingdom) that are the biggest investors in RD&D in fossil fuels, renewables, energy efficiency, and nuclear energy. Thus, economies will be able to take precautions against global shocks in producing sustainable energy policies. To achieve this aim, this paper uses the artificial neural network approach, which uses a distributed computing model that can store and generalize data after a learning period. Empirical results widely depend on countries and distinct energy technology policy fields. The key findings provide evidence that RD&D expenditures, except for renewables, do not tend to mean revert. The economies of Japan, Germany, and the United States should make more renewable investments to reduce the resistance of renewable energy sources to shocks and thus produce policies for both environmental sustainability and energy security. The economies of France and the United Kingdom can ensure energy security by continuing their sustainable energy policies. • Applied SDG policies focused on energy technology RD&D are presented. • The stochastic characteristics of energy technology investments are examined. • Carbon-reducing policies are recommended for the top five economies. • Energy technologies are perused through artificial intelligence approach. [ABSTRACT FROM AUTHOR]

With the increasing installation of battery energy storage systems, the safety of high-energy-density battery systems has become a growing concern. Developing reliable battery fault diagnosis and fault warning algorithms is essential to ensure the safety of battery systems. After years of development, traditional fault diagnosis techniques based on three-dimensional information of voltage, current and temperature have gradually encountered bottlenecks. It is necessary to adopt a proactive approach by using mulitidimensional information to advance fault diagnosis techniques. This involves integrating advanced sensing technologies, collecting multidimensional data and uncovering subtle changes in battery behavior. This paper delves into the mechanisms and evolutionary paths of battery faults, with a specific focus on the multidimensional observable signals associated with different faults for enhanced safety strategy. Furthermore, the paper provides a comprehensive overview of potential applications of different sensors for multidimensional measurement in battery fault diagnosis. It also explores the future trends and research directions of the next generation of battery fault diagnosis techniques driven by multidimensional data collection and artificial intelligence algorithms. [Display omitted] • Enhanced safety through proactive, multidimensional fault diagnosis techniques. • Integration of advanced sensing tech for precise multidimensional data collection. • Uncovering subtle battery behavior changes for improved fault detection. • Specific focus on multidimensional signals to enhance safety strategies. • Future trends in battery fault diagnosis driven by AI and multidimensional data. [ABSTRACT FROM AUTHOR]

Timely monitoring of abnormal vibration of machinery in highly harsh environments is essential to ensure the safe operation of mechanical systems. This paper uses a self-charging power module to harvest energy to provide sustainable power for monitoring multidirectional ultra-low frequency mechanical vibrations. The power generation unit consists mainly of a spherical electromagnetic triboelectric hybrid nanogenerator (SETE-HNG). It is equipped with advanced sensing functions to support deep learning to identify signals with different directions of vibration, different frequencies, and different amplitudes of vibration, thereby enhancing the high-precision perception function. The accuracy of the prediction results is as high as 98.6622%, 100%, and 99.3333%, respectively. A Power Management Circuitry (PMC) has been meticulously crafted to maximize the utilization of vibrational energy. It efficiently charges a 40 mAh lithium polymer battery to 3.3 V in just 26 min, all without the requirement of an external power source. This advancement facilitates self-powered Global Positioning System (GPS) tracking of vibrational signals. Moreover, the stored energy is harnessed to energize a microcontroller and a low-power Bluetooth module. This enables real-time monitoring of mechanical vibrations via a mobile phone. The design presented in this paper is a testament to the potential of self-powered multidirectional mechanical vibration monitoring, contributing significantly to the safety and efficiency of mechanical systems. The design in this paper enables self-powered monitoring of abnormal vibrations in mechanical systems in extremely harsh environments, thus ensuring the safe operation of equipment. [Display omitted] • Large-scale vibration energy harvesting provides an innovative and efficient method. • SETE-HNG has multidirectional ultra-low frequency mechanical vibration monitoring and energy harvesting functions. • A power management circuit is designed to store energy efficiently. [ABSTRACT FROM AUTHOR]

Thermal networks require thermal energy storage (TES) provisions for balancing thermal energy sources with variable consumer demand. Harvesting ice is an economical option for latent heat TES systems in cooling networks given the wide availability of the storage medium. This paper presents an artificial intelligence (AI) based model to monitor the state-of-charge (SoC) and the outlet temperature of the heat transfer fluid (T o) of an ice tank under fluctuating operating conditions. The AI model is a non-linear autoregressive network with exogenous inputs (NARX) that was trained and tested with datasets obtained from experimental measurements of a practical ice tank and a physics-based model of the tank. The NARX model was sensitised with physics-informed attributes to recognise different heating and cooling zones. The model exhibits a high accuracy in predicting the operating conditions of the ice tank when benchmarked against both experimental measurements of a practical tank and outputs from the physics-based model. For instance, it achieves R 2 values of 0.9943 and 0.9842 for SoC and T o , with root mean square errors of 1.73% for SoC and 0.3161°C for T o. The NARX model is 86% faster than its physics-based counterpart and its implementation requires limited computational resources—making it suitable as a standalone estimator for the TES operation and the accelerated simulation of energy systems containing latent heat TES units. Furthermore, given the limited availability of NARX models in open-source libraries, the presented NARX model and relevant datasets have been made available alongside this paper to contribute to open-science in energy research and the broader AI community. [Display omitted] • The state of a TES unit is continuously quantified with a non-intrusive NARX model. • The NARX model trained on physics-based data replicates the real system's response. • Heating zone-wise segregation of prediction tasks boosts the model's accuracy. • The NARX model is 86% faster than a physics-based model built for fast computation. • Limited public availability of a NARX model is addressed by code and data sharing. [ABSTRACT FROM AUTHOR]

• Estimated battery SOH using CNN model based on the raw data. • A high tolerance to the partial charging condition. • Introduced transfer learning method to reduce the offline training data. • Satisfied generality to different types of batteries. State of health (SOH) estimation is essential to the health diagnostic of lithium-ion battery. The data-driven approach with charging feature extraction is promising for online SOH estimation and has been widely explored over years. However, their deployment can be barriered by the lack of complete charging data in real-world applications. Motivated by this, this paper proposes an artificial intelligence-based SOH estimator using the transient phase between constant current (CC) and constant voltage (CV) charging, which is easily obtained in real-world charging scenarios. Specifically, a convolutional neural network (CNN) model is proposed to explain the relationship between the charging data and the SOH. Following this endeavor, the transfer learning is exploited for model mitigation and SOH estimation on different battery types, relying on much reduced amount of data for efficient CNN model re-training. The validation experiments are conducted based on the aging data obtained on LiNiCoAlO 2 (NCA) and LiCoO 2 (LCO) cells. Results suggest that the proposed method realizes accurate SOH estimation requiring only a short segment from the CC-CV transient phase, so that can meet a broad range of real-world charging scenarios. Moreover, the efficient model transfer promises expected performance with different battery types. The short version of the paper was presented at ICAE2021, Nov 29 - Dec 5, 2021. This paper is a substantial extension of the short version of the conference paper. [ABSTRACT FROM AUTHOR]

Rooftop photovoltaic (PV) segmentation based on remote sensing images is highly applied in solar potential assessment and prediction. Still, such methods often feature dataset limitations of PV data, poor robustness, and are non-generalizable. General Generative AI eliminates the need for pre-training emerging to improve the sample diversity and algorithm robustness and generalizability of the segmentation. This paper designs a PV sample generation method based on the generative model, which leverages the text-guided stable diffusion inpainting model to augment the PV dataset and generate massive multi-background rooftop PV panel samples. The real and generated samples are mixed in different proportions to form a new training set for ablation experiments. Results show that a small number of real datasets mixed with generated data could reach a high relative IoU and Precision value. In small sample learning, the generated data achieves similar effects as real data during the segmenting process even better than without generated data. It demonstrates that the generated datasets outperform traditionally augmented data and that the manual text prompts are tested more accurately than ChatGPT-generated ones. This study highlights the efficiency and robustness of generated datasets in PV segmentation tasks and moves beyond the constraints of remote sensing data acquisition and limited data diversity. Further, it would facilitate large-scale assessments of the urban PV potential for urban planners and policymakers using an efficient and low-cost method. • This study leverages a text-guided stable diffusion inpainting model to augment the PV dataset and improve data diversity. • It proves that a small number of real datasets mixed with generated data could reach high relative precision and robustness. • The General Generative AI-based data augmentation reaches a higher accuracy than conventional data augmentation. • The manual text prompts perform better than the ChatGPT-generated versions. [ABSTRACT FROM AUTHOR]

Establishing building energy models (BEMs) for building design and analysis poses significant challenges due to demanding modeling efforts, expertise to use simulation software, and building science knowledge in practice. These make building modeling labor-intensive, hindering its widespread adoptions in building development. Therefore, to overcome these challenges in building modeling with enhanced automation in modeling practice, this paper proposes Eplus-LLM (EnergyPlus-Large Language Model) as the auto-building modeling platform, building on a fine-tuned large language model (LLM) to directly translate natural language description of buildings to established building models of various geometries, occupancy scenarios, and equipment loads. Through fine-tuning, the LLM (i.e., T5) is customized to digest natural language and simulation demands from users and convert human descriptions into EnergyPlus modeling files. Then, the Eplus-LLM platform realizes the automated building modeling through invoking the API of simulation software (i.e., the EnergyPlus engine) to simulate the auto-generated model files and output simulation results of interest. The validation process, involving four different types of prompts, demonstrates that Eplus-LLM reduces over 95% modeling efforts and achieves 100% accuracy in establishing BEMs while being robust to interference in usage, including but not limited to different tones, misspells, omissions, and redundancies. Overall, this research serves as the pioneering effort to customize LLM for auto-modeling purpose (directly build-up building models from natural language), aiming to provide a user-friendly human-AI interface that significantly reduces building modeling efforts. This work also further facilitates large-scale building model efforts, e.g., urban building energy modeling (UBEM), in modeling practice. • This study represents the pioneering effort in customizing LLM for auto-modeling. • A large language model is used to digest natural language descriptions for modeling. • Integration of four prompts enhances the flexibility and versatility of our platform. • Our platform provides a human-AI interface and reduces building modeling efforts. [ABSTRACT FROM AUTHOR]

Carbon emissions have become particularly urgent with the acceleration of industrialization. Currently, AI-based carbon emission modeling provides essential support for solving the problem of predicting, managing, and optimizing carbon emissions. However, there are still challenges in accurately predicting carbon dioxide emissions and establishing an appropriate carbon emission regulatory model to manage carbon dioxide emissions effectively. Firstly, this paper introduces a framework for learning seasonal and trend representations called DeTF (Decoupling representation for Time and Frequency domain of time series). The framework is designed for time series data of CO 2 emissions used for power generation from 1973 to 2023. Secondly, we use various data enhancement techniques to improve the robustness of the learned representations. Finally, the prediction results establish a tripartite evolutionary game relationship between the state, regulatory agencies, and enterprises. We use MSE and MAE as evaluation metrics, and our model overall outperforms current state-of-the-art end-to-end algorithms. It outperforms the best performing end-to-end prediction method by 27.08% (MSE) in the prediction results. Besides, the time series prediction results are innovatively combined with the tripartite game model to analyze the main factors determining the optimal strategy. • Time domain augmentation boosts robustness with strong and weak methods. • The New DeTF framework splits time and frequency in time series analysis. • The overall loss consists of two main components: the time domain and the frequency domain. • Time series prediction merges with a tripartite game and explores strategies and equilibria. [ABSTRACT FROM AUTHOR]

• This paper focuses on daily natural gas consumption forecasting. • A novel hybrid model is developed for daily natural gas consumption forecasting. • The most recent datasets of four cities in three climate zones are collected. • An improved singular spectrum analysis is first proposed. In daily natural gas consumption forecasting, the accuracy of forecasting models is vulnerably affected by the noise data in historical time series. Singular spectrum analysis (SSA) is often introduced into hybrid models for denoising. However, as a deterministic-based algorithm, SSA does not give good results when a time series is contaminated with a high noise level. Considering this fact, this paper proposes an improved SSA (ISSA) that modifies the determination method of subseries selection in the reconstruction stage of SSA. Combining ISSA with long short-term memory (LSTM), a novel hybrid model, ISSA-LSTM, is thus developed. Additionally, for validating the robustness and superiority of ISSA-LSTM, the historical datasets of four representative cities located in three climate zones are collected as the training and testing datasets, and a comparison of ISSA-LSTM with five advanced models is performed. The results reveal that SSA would generate negative values when time series close to zero and the contribution of SSA in improving the forecasting accuracy of LSTM is insignificant. In contrast, ISSA avoids generating negative values and reduces the mean absolute range normalized error (MARNE) of LSTM by a range of 0.86–11.86%. Among the models, ISSA-LSTM achieves the best performance and its MARNEs for London (temperate zone), Melbourne (subtropical zone), Karditsa (subtropical zone), and Hong Kong (tropical zone) are 4.68%, 5.72%, 5.76%, and 14.10%, respectively. The MARNE of the tropical city is higher than that of others, which is caused by the complex natural gas consumption pattern of itself. [ABSTRACT FROM AUTHOR]

Highlights • Realize IT load demand response controllability with data center resource scheduling. • Design server cluster workload scheduling method with time shift mechanism. • Design data center UPS energy storage dynamic response mechanism. • Holistic dynamic response method smoothens data center tie-line power fluctuation. Abstract With the rapid development of cloud computing, artificial intelligence technologies and big data applications, data centers have become widely deployed. High density IT equipment in data centers consumes a lot of electrical power, and makes data center a hungry monster of energy consumption. To solve this problem, renewable energy is increasingly integrated into data center power provisioning systems. Compared to the traditional power supply methods, renewable energy has its unique characteristics, such as intermittency and randomness. When renewable energy supplies power to the data center industrial park, this kind of power supply not only has negative effects on the normal operation of precision equipment, such as CPU/GPU chips and hard disk, in data center, but it would also impact the stability of the utility power grids operation. To solve this problem, this paper presents a novel tie-line power fluctuation smoothing algorithm with consideration of data center's holistic demand response. The contributions of this paper are: (1) overcoming the limitations of treating IT load as uncontrollable workload in the traditional demand response research, we design a data center resource scheduling model to realize IT load demand response controllability; (2) two novel mechanisms are proposed: (i) the server cluster workload scheduling method with time shift mechanism, and (ii) the data center UPS (Uninterruptible Power Supply) energy storage dynamic response mechanism. (3) Combining these two mechanisms as holistic demand response of data center, we present a tie-line power fluctuation smoothing algorithm to improve power supply reliability, which is beneficial to both the high density and precision IT equipment in the data center and the utility power grid. In the experiments, the results show that the new algorithm can effectively regulate the tie-line power fluctuations under different server cluster utilization ranges and scenarios of large-scale penetration of distributed renewable energy scenarios. The new algorithm is hence able to contribute beneficially to the reliability and stability of intelligent industrial park micro-grid and utility power grids. [ABSTRACT FROM AUTHOR]

Energy storage systems enable balancing supply and demand and facilitate the integration of intermittent renewable energy sources. In particular, latent heat thermal energy storage units are attractive for deployment in thermal systems due to their high energy density. However, knowledge of the state-of-charge of a thermal store is crucial to effectively regulate its charging and discharging cycles. To achieve this, continuous-time non-linear observers may be employed to estimate the state-of-charge at the expense of a high computational cost. The emergence of artificial intelligence solutions may be helpful to reduce computational burden, but their adoption for thermal stores has been limited. This paper bridges this research gap by presenting a novel approach to predict the state-of-charge of a latent heat unit. It employs a discrete-time estimator based on a recurrent neural network, which is based on a long short-term memory structure and the regression method for estimation. The estimator offers a reduced computation time for state-of-charge estimation and allows flexible sampling adjustments without sacrificing accuracy. Additionally, the presented approach simplifies data collection by independently handling charging and discharging processes through internal state resets. The estimator, trained using MATLAB's deep learning toolbox, uses a dataset comprising various operating conditions obtained from simulations of a physics-based model. When compared against a more traditional state-of-charge estimation method using a discrete-time non-linear observer, the advantages of a recurrent neural network-based estimator are evidenced, highlighting its potential for practical applications. The presented method exhibited high accuracy with a maximum root mean square error under 0.73% and a mean absolute error below 0.41% with respect to direct state-of-charge calculation. Although the discrete-time non-linear observer exhibited a marginal higher accuracy, the recurrent neural network-based estimator achieved a significant improvement in computational efficiency. These findings make the proposed approach a robust tool facilitating enhanced control strategies, optimised energy management, and increased overall thermal system performance. • Highly accurate AI-based SoC estimator for thermal stores. • RNN training framed within a regression problem to predict SoC. • Implementation of the SoC estimator through simple matrix operations. • Significantly reduced computational resource compared to other methods. [ABSTRACT FROM AUTHOR]

Energy plays a vital part in stimulating economic progress, and the shift towards a cleaner energy system is highly significant for ensuring the sustainable development of the economy. China's energy structure urgently needs to be transitioned. The fast advancement and implementation of artificial intelligence (AI) has provided a new and important tool for promoting the transition of energy structure. So, what is the relationship between the application of artificial intelligence and the transition of the energy structure? This research introduces artificial intelligence into the energy sector, focusing on the relationship between artificial intelligence and energy transition. Since nonlinear models are better able to study the complex effects and phase differences of artificial intelligence. Using China's provincial panel data spanning from 2006 to 2019, this study employs nonlinear modeling to explore the stage differences in the process of AI in facilitating energy structure transformation. This paper derives the following findings based on empirical research. First, there is a U-shaped relationship between artificial intelligence and the transition of energy structure. Specifically, before the inflection point, the initial application of artificial intelligence, artificial intelligence may adversely impact energy transition. When the inflection point is passed, AI will help facilitate the energy transition. Second, the U-shaped relationship between AI and energy transition is more pronounced in coastal and non-resource-based regions. Third, energy intensity, government investment in science and technology, and informatization will moderate the U-shaped relationship between artificial intelligence and energy transition, changing the steepness of the original U-shaped relationship and even reversing it. Hence, it is imperative to effectively utilize the technological benefits of artificial intelligence through the development patterns and distinctive features of different regions, thereby facilitating the smooth transition of the energy structure. [Display omitted] • The role of artificial intelligence in the energy transition process is studied. • Stage differences in the role of AI are analyzed. • The nonlinear relationship between AI and energy transition is investigated. • Provide insights into the factors affecting the relationship between AI and energy transition. • This research helps the government to advance the energy structure transition. [ABSTRACT FROM AUTHOR]

At present, operational power forecasts are primarily based on the predicted wind speed of a single-valued deterministic Numerical Weather Prediction (NWP) simulation. However, due to the unavoidable uncertainties from model initialization and/or model imperfections, recent numerical techniques cannot directly meet the actual needs of grid dispatch in many cases, which means that achieving accurate forecasts of wind speed and power is still a critical issue. On this topic, our paper contributes to the development of a new multi-step forecasting method termed CSFC-Apriori-WRF, providing a one-day ahead wind speed and power forecast consisting of 96 steps. This method is based on a Weather Research and Forecasting (WRF) simulation, a Cuckoo search (CS) optimized fuzzy clustering, and an Apriori association process. First, a wind speed forecast is generated by running a configured WRF model. Next, the wind speed forecasting series is divided into segments that meet certain conditions and are defined as “waves” in this paper. Next, combining the CS-optimized fuzzy clustering and Apriori algorithm, the proposed method extracts the association rules between the shape characteristics and the forecasting error of the divided waves. Applying the association rules in the final optimization process, the proposed method significantly reduces the uncertainties of the WRF simulation and performs best among other models to which it is compared. [ABSTRACT FROM AUTHOR]

This paper presents a solution to the energy planning problem in buildings by implementing the net Zero Energy cost Building (nZEcB), a new concept that refers to buildings with zero or almost zero annual energy costs, using Artificial Intelligence (AI) techniques such as bidirectional long short-term memory, ordinary least squares linear regression, K-means, Pearson's correlation, decision tree, and binary gravitational search algorithm. AI techniques are used to design the optimal structure of a distributed generation system. The system includes wind and photovoltaic renewable energy sources, a battery bank, and an automated capacitor bank for power factor compensation. A case study was conducted in a real public building with an annual consumption of 1.748 GWh, which resulted in the specification of a distributed generation system with a generation of 2.805 GWh per year. This system meets approximately 160.5% of the building's electrical demand and 99.999% of annual energy costs, i.e., attending the nZEcB concept. The excess production of approximately 60% of the energy is necessary because the exported energy (feed-in tariff) has a lower value than that imported from the network, having to deduct costs with demand, in addition to having to compensate for losses in the battery banks. The project has a payback period of 6.79 years. This novel study demonstrates the feasibility and effectiveness of using AI techniques to achieve nZEcBs more efficiently, economically and sustainably. • Efficient integration of AI techniques in nZEcB design. • BiLSTM employed to forecast weather variables of subsequent years. • Load demand prediction using OLS Regression for forthcoming years. • K-means pre-selection of nZEcB equipment sizes based on generation and consumption. • AI outcomes integration in a BGS Algorithm to optimize annual net energy and payback. [ABSTRACT FROM AUTHOR]

The soaring potential of edge computing leads to the emergence of cloud–edge collaboration. This hierarchy enables the deployment of artificial intelligence models in the cyber–physical venue. This paper presents Dual Digital Twin, the next level of digital twin, in the presence of two levels of communication availability, for battery system real-time monitoring and control in electric vehicles. To implement the dual digital twin concept, an online adaptive model reduction problem is formulated with time scale differences induced by the time sensitivity property of industrial applications and limitations of infrastructure. To minimize the model reduction error and battery system control penalty, the online adaptive battery reduced order model framework is proposed, consisting of the gated recurrent unit neural network to construct battery internal states given Internet of things sensor measurements, and incremental learning techniques to facilitate the update of the reduced-order model given data stream. Moreover, we design the physics-informed update of the neural network using the Lyapunov stability theorem to enhance the synchronization with the physical battery behavior. A Li-ion battery and single particle digital twin model with electrolyte and thermal dynamics are utilized in the simulation to justify the effectiveness of the proposed framework. Numerical results demonstrate 1.70% average reduced-order model prediction error and 43.3% accuracy improvement with the novel physics-informed online adaptive framework. The method is also robust concerning varying environmental factors and noise. • Introduce the dual digital twin concept and its realization in EV batteries. • Online adaptive model reduction problem with intrinsic time scale difference in AIoT system. • Cloud–edge collaborated AIoT online adaptive battery ROM framework. • Physics-informed ROM incremental learning algorithm to track the real system dynamics. [ABSTRACT FROM AUTHOR]

Modeling of the chiller performance is essential for the implementation of optimal energy-efficient control strategies in a heating, ventilation, and air conditioning (HVAC) system. Though classical data-driven chiller performance models are widely adopted in the industry, they generally suffer from poor out-of-sample generalization abilities, which refers to the model's capability to extrapolate for new data outside the range of the training dataset. In practice, however, the available chiller operation data for model development are often insufficient or collected from a few limited operating conditions, such that extrapolation is unavoidable after the model is applied for control purposes. To deal with this issue, this paper proposed a physics-guided neural network (PGNN) to model the energy performance of chillers. By adopting a new neural network architecture, modifying the loss function, and adding limited out-of-sample data, the PGNN incorporates domain knowledge into the data-driven model to achieve better out-of-sample generalization performance. Meanwhile, the convexity and monotonicity between the dependent and independent variables in the PGNN are properly addressed. The proposed PGNN is applied to model the chiller serving a high-rise building, and results show that PGNN performs much better in extrapolation than classical models and the multi-layer perceptron model. The research demonstrated the usefulness and effectiveness of the PGNN in modeling HVAC equipment. • A physics-guided neural network is developed to model chiller energy performance. • Physical constraints are added through new architecture and modified loss function. • Limited out-of-sample data are used as additional constraints for model training. • The out-of-sample generalizability of the proposed model is significantly enhanced. [ABSTRACT FROM AUTHOR]

Accurate and timely carbon emission accounting (CEA) is vital to industrial corporations, especially those who participate in the carbon market. With the rapid development of artificial intelligence and power systems, the power data-based method provides a new way for real-time CEA. However, the extensive installation of distributed photovoltaics (PV) significantly increases the accounting difficulty of corporate carbon emissions. This paper proposes a non-intrusive method of real-time CEA for industrial corporations from the perspective of modern power systems. First, a device operation state (DOS) estimation model based on a modified Informer algorithm is proposed to calculate corporate direct carbon emissions. Wherein, an equivalent distributed PV output estimation model is used to decrease the impact of invisible PVs on direct emission accounting. Second, an improved carbon emission flow model is proposed to calculate corporate indirect carbon emissions, which considers "prosumers" arising from the installation of distributed PVs. Finally, the total corporate carbon emissions, including direct and indirect parts, are obtained by using the CEA model. Case studies based on four typical high‑carbon-emission factories in Zhejiang province, China demonstrate that the proposed method can make accurate CEA for industrial corporations by effectively lessening the impact of distributed PVs. • A non-intrusive carbon emission accounting method for industrial corporations. • A novel perspective of modern power systems for carbon emission accounting. • Calculating carbon emissions considering distributed photovoltaics' impacts. • The method becomes superior as the photovoltaic output proportion increases. [ABSTRACT FROM AUTHOR]

The Atikokan Generating Station, operated by Ontario Power Generation, has a 200 MW, biomass-fired tower boiler that operates on a dispatch schedule with a five-minute cycle. The boiler is generally operated in the range of 40–100 MW using two of five burner levels. In order to optimize boiler performance, we propose the implementation of a unique digital twin. Our digital twin abstraction couples Bayesian inference from science-based models and from observations (machine learning) with decision theory to predict operating-variable set points that optimize the physical asset (the boiler) in the presence of uncertainty (artificial intelligence). We focus this paper on the continuous Bayesian machine learning part of the Atikokan Digital Twin; we discuss decision theory in a companion paper. We identify and learn about 12 operational, model, and measured-output parameters and their uncertainties from high-fidelity, science-based simulations of the Atikokan boiler and from the observed measurements at the power plant. Since the goal of the Atikokan Digital Twin is to implement it online in real time, we require fast function evaluations for the quantities of interest extracted from the simulations in the Bayesian analysis. We use Gaussian process regression/interpolation to create accurate, robust surrogate models. We define the Bayesian priors and likelihood function and solve for the posterior distributions of the 12 parameters. We then propagate these distributions (i.e., parameters with uncertainty) into the predicted distributions of 790 quantities of interest to learn about the relative importance of various sources of error including experimental, model, and operating-parameter errors. • Everything is a model and all models have uncertainty. • Uncertainty is represented with distributions. • Bayesian machine learning tools predict key process behavior with uncertainty. • Digital twin learns continuously from science-based model and experimental data. • Digital twin of biomass boiler produces dynamic predictions of 790 outputs. [ABSTRACT FROM AUTHOR]

"Framework" and "case-study" are the two most prominent features in the optimization of architectural building design. The first can improve the speed of the process, the accuracy and integration of the outcome. The second can evolve the design knowledge of the various types of buildings in practice. This paper presents the "PCA-ANN integrated NSGA-III" framework that is a fast, accurate, and integrated solution for practical, erosive, and complicated building performance optimization problems, especially when the case study is a multi-type building such as a dormitory. The novelty and contribution of this paper exist in all three phases of its research framework. First, the dormitory building was parametrically modeled as a case study in two conventional dormitory building types. Second, an artificial model called PCA-ANN is presented in which problems that have a more complicated and practical case study with a multi-type building can be solved simply, integratively, and accurately. Third, the capabilities of NSGA-III are introduced and showed that it could be more accurate in comparison with NSGA-II, and then NSGA-III was integrated with the PCA-ANN model to solve the many-objective problem posed in this study. R2 score for PCA-ANN in the best-presented structure reached 0.99. In the overall best building alternative, energy efficiency and daylight performance for the whole year and thermal comfort performance in the free running period improved by 41.27%, 42.24%, and 15.57%, respectively, compared to its reference model. The presented framework can be applied and generalized to building energy optimization problems in practical and complex cases. • PCA-ANN and NSGA-III are integrated to optimize vast and complicated solutions. • Generative design of two dormitory types is set up to evaluate the main potentials. • The design space is discovered and datasets are generated for both types of building. • PCA-ANN is selected for the practical problem raised in this study from a comparison. • Final selection is made based on NSGA-III and the architect's conditional decisions. [ABSTRACT FROM AUTHOR]

• A new hybrid prediction system is proposed to predict wind speed. • Simplex and chaotic mapping optimization cuckoo algorithm are proposed. • An uncertainty prediction model based on Monte Carlo method is proposed. • The energy theory is proposed to optimize the decomposition mode number. • Theories proofs and experiments verify the effectiveness of the model. Wind energy has strong volatility and intermittent. Accurate wind speed prediction can not only improve the safety of the system, but also optimize dispatch and reduce economic losses. However, previous studies tend to ignore the influence of virtual components and lack effective identification of wind speed characteristics and a robust interval prediction scheme, resulting in poor results. To bridge these gaps, this paper proposes an energy theory method to solve the problem of modal over-decomposition. The study also combines effective modal recognition, uses different prediction methods according to modal characteristics and proposes a set of new optimization algorithms to improve nonlinear prediction capabilities. Finally, based on Monte Carlo theory, a set of interval prediction schemes that can adapt to different error characteristics are proposed. Under the verification of wind speed data in Changma, China and Sotavento, Spain. The mean absolute percentage error of wind speed deterministic prediction reaches 4.22% and 5.82%, respectively. The coverage rate of wind speed uncertainty prediction meets different confidence requirements, and the average interval width is still less than 2.5 m/s at 90% confidence. The results show that the forecasting system proposed in this paper is significantly better than all the comparative forecasting schemes, which can reduce the risk of fluctuations and improve the stability and safety of the wind power system. [ABSTRACT FROM AUTHOR]

Highlights • Discrete Choice Models model occupant actions through machine learning algorithms. • Deep Learning shows outstanding potential for sequential decision making. • Artificial Intelligence can generate high dimensional dynamic occupancy usage data. • Gamification in human-centric cyber-physical systems promotes energy efficiency. • Participation in social game communicates usage data and facilitates forecasting. Abstract In this paper, we propose a gamification approach as a novel framework for smart building infrastructure with the goal of motivating human occupants to consider personal energy usage and to have positive effects on their environment. Human interaction in the context of cyber-physical systems is a core component and consideration in the implementation of any smart building technology. Research has shown that the adoption of human-centric building services and amenities leads to improvements in the operational efficiency of these cyber-physical systems directed toward controlling building energy usage. We introduce a strategy that incorporates humans-in-the-loop modeling by creating an interface to allow building managers to interact with occupants and potentially incentivize energy efficient behavior. Game theoretic analysis typically relies on the assumption that the utility function of each individual agent is known a priori. Instead, we propose a novel benchmark utility learning framework that employs robust estimations of occupant actions toward energy efficiency. To improve forecasting performance, we extend the benchmark utility learning scheme by leveraging Deep Learning end-to-end training with deep bi-directional Recurrent Neural Networks. We apply the proposed methods to high-dimensional data from a social game experiment designed to encourage energy efficient behavior among smart building occupants. Using data gathered from occupant actions for resources such as room lighting, we forecast patterns of resource usage to demonstrate the performance of the proposed methods on ground truth data. The results of our study show that we can achieve a highly accurate representation of the ground truth for occupant resource usage. For demonstrations of our infrastructure and for downloading de-identified, high-dimensional data sets, please visit our website (smartNTU demo web portal: https://smartntu.eecs.berkeley.edu) [ABSTRACT FROM AUTHOR]

Highlights • Propose an incentive-based demand response algorithm with artificial intelligence. • Reinforcement learning is employed to obtain the optimal incentive rates. • Achieve real-time performance with the aid of deep neural network. • Customer diversity is taken into account by provision of different incentive rates. • Service provider payment under cases without and with demand response is compared. Abstract Balancing electricity generation and consumption is essential for smoothing the power grids. Any mismatch between energy supply and demand would increase costs to both the service provider and customers and may even cripple the entire grid. This paper proposes a novel real-time incentive-based demand response algorithm for smart grid systems with reinforcement learning and deep neural network, aiming to help the service provider to purchase energy resources from its subscribed customers to balance energy fluctuations and enhance grid reliability. In particular, to overcome the future uncertainties, deep neural network is used to predict the unknown prices and energy demands. After that, reinforcement learning is adopted to obtain the optimal incentive rates for different customers considering the profits of both service provider and customers. Simulation results show that this proposed incentive-based demand response algorithm induces demand side participation, promotes service provider and customers profitabilities, and improves system reliability by balancing energy resources, which can be regarded as a win-win strategy for both service provider and customers. [ABSTRACT FROM AUTHOR]

Highlights • A combined model is proposed for multi-step ahead wind speed forecasting. • Data preprocessing technology is introduced to improve the forecasting performance. • Quasi-Newton algorithm is used to increase the particle diversity of water cycle algorithm. • The accuracy and stability of wind speed forecasting are improved simultaneously. • The simulation results are validated well in China. Abstract Owing to the complexity and uncertainty of wind speed, accurate wind speed prediction has become a highly anticipated and challenging problem in recent years. Researchers have conducted numerous studies on wind speed prediction theory and practice; however, research on multi-step wind speed prediction remains scarce, which hinders further development in this area. To improve upon the accuracy and stability of multi-step wind speed prediction, this paper proposes a combination model based on a data preprocessing strategy, an improved optimization model, a no negative constraint theory, and several single prediction models. To improve upon forecasting performance, an improved water cycle algorithm based on a quasi-Newton algorithm is proposed to optimize the weight coefficients of the single models. In the empirical research, 10-min and 30-min wind speed data from Shandong Province in China, collected for case studies, were used to assess the comprehensive performance of the proposed combination model. Finally, we used 10-fold cross-validation and multiple error criteria to evaluate the comprehensive performance of the proposed combination model. The simulation results indicate that (a) the quasi-Newton algorithm can effectively increase the diversity of the water cycle algorithm particles, resulting in improved water cycle algorithm optimization performance; (b) the combination model exhibits superior predictive performance to a single model by taking advantage of each single model; and (c) the proposed combination model can effectively improve multi-step wind speed prediction results. [ABSTRACT FROM AUTHOR]

Highlights • An adaptive Fuzzy Logic PID approach is proposed to control steam turbines. • A CSPP has been modeled focusing on power loop with variable steam conditions. • Design of FPID controller through the knowledge acquired during simulation phase. • FL allows to control steam turbines in off-design points with optimal performance. • Control algorithm with reduced design and implementation time on PLC platforms. Abstract In Concentrated Solar Power Plants, steam turbines controlled with standard Proportional Integrative Derivative (PID) methods may suffer from performance downgrading in power generation when the steam conditions deviate from nominal ones. An enhancement of standard steam turbine controller can be the key to achieve optimal performance also in non-nominal steam conditions. This paper presents the improvement of the PID control concept by exploiting Fuzzy Logic, an artificial intelligence technique that allows taking into account the human experience and knowledge on the system behavior. A real Concentrated Solar Power Plant has been modeled focusing on generated power control loop, its stability and performance analysis, knowledge useful to design a Fuzzy Inference System. A fuzzy logic controller is proposed to continuously adapt the PID parameters, to improve the steam turbine governor action. Its performance is compared to the classical PID tuned according to three different approaches. The fuzzy logic PID controller extends the simplicity of PID and adapts the control action to actual operating condition by providing the system with a sort of “decision-making skill”. The possibility to design implementable algorithms on a Programmable Logic Controller, which have stringent computational speed and memory requirements, has been explicitly taken into account in the developed work, through the minimization of the controller complexity with a reduced number of fuzzy sets and fuzzy rules within the fuzzy inference system. [ABSTRACT FROM AUTHOR]

Reinforcement learning is a new research hotspot in the artificial intelligence community. Q learning as a famous reinforcement learning algorithm can achieve satisfactory control performance without need to clarify the complex internal factors in controlled objects. However, discretization state is necessary which limits the application of Q learning in energy management for hybrid electric bus (HEB). In this paper the deep Q learning (DQL) is adopted for energy management issue and the strategy is proposed and verified. Firstly, the system modeling of bus configuration are described. Then, the energy management strategy based on deep Q learning is put forward. Deep neural network is employed and well trained to approximate the action value function (Q function). Furthermore, the Q learning strategy based on the same model is mentioned and applied to compare with deep Q learning. Finally, a part of trained decision network is analyzed separately to verify the effectiveness and rationality of the DQL-based strategy. The training results indicate that DQL-based strategy makes a better performance than that of Q learning in training time consuming and convergence rate. Results also demonstrate the fuel economy of proposed strategy under the unknown driving condition achieves 89% of dynamic programming-based method. In addition, the technique can finally learn to the target state of charge under different initial conditions. The main contribution of this study is to explore a novel reinforcement learning methodology into energy management for HEB which solve the curse of state variable dimensionality, and the techniques can be adopted to solve similar problems. [ABSTRACT FROM AUTHOR]

Forecasting the price of crude oil is a challenging task. To improve this forecasting, this paper proposes a novel hybrid method that uses an integrated data fluctuation network (DFN) and several artificial intelligence (AI) algorithms, named DFN-AI model. In the proposed DFN-AI model, a complex network time series analysis technique is performed as a preprocessor for the original data to extract the fluctuation features and reconstruct the original data, and then an artificial intelligence tool, e.g., BPNN, RBFNN or ELM, is employed to model the reconstructed data and predict the future data. To verify these results we examine the daily, weekly, and monthly price data from the crude oil trading hub in Cushing, Oklahoma. Empirical results demonstrate that the proposed DFN-AI models (i.e., DFN-BP, DFN-RBF, and DFN-ELM) perform significantly better than their corresponding single AI models in both the direction and level of prediction. This confirms the effectiveness of our proposed modeling of the nonlinear patterns hidden in crude oil prices. In addition, our proposed DFN-AI methods are robust and reliable and are unaffected by random sample selection, sample frequency, or breaks in sample structure. [ABSTRACT FROM AUTHOR]

With the modern advanced information and communication technologies in smart grid systems, demand response (DR) has become an effective method for improving grid reliability and reducing energy costs due to the ability to react quickly to supply-demand mismatches by adjusting flexible loads on the demand side. This paper proposes a dynamic pricing DR algorithm for energy management in a hierarchical electricity market that considers both service provider’s (SP) profit and customers’ (CUs) costs. Reinforcement learning (RL) is used to illustrate the hierarchical decision-making framework, in which the dynamic pricing problem is formulated as a discrete finite Markov decision process (MDP), and Q-learning is adopted to solve this decision-making problem. Using RL, the SP can adaptively decide the retail electricity price during the on-line learning process where the uncertainty of CUs’ load demand profiles and the flexibility of wholesale electricity prices are addressed. Simulation results show that this proposed DR algorithm, can promote SP profitability, reduce energy costs for CUs, balance energy supply and demand in the electricity market, and improve the reliability of electric power systems, which can be regarded as a win-win strategy for both SP and CUs. [ABSTRACT FROM AUTHOR]

• Methodology and model for creating AI expert learning via AR and humans. • Positive climate impact when some services can be carried out remotely over AR calls. • Energy savings and CO₂ emissions reduction have been estimated. • Supporting the development of smart industries and autonomous operations. • Economic benefits as for health improvement and increased business productivity. Initially, Artificial Intelligence (AI) focused on diagnostics during the 70s and 80s. Unfortunately, it did not gain trust and few industries embraced it, mostly due to the extensive manual programming effort that AI required for interpreting data and act. In addition, the computer capacity, for handling the amounts of data necessary to train AI, was lacking the disc dimensions we are used to today, which made it go slowly. Not until the 2000 s confidence in AI was established in parallel with the introduction of new tools that was paving the way for PLS, PCA, ANN and soft sensors. Year 2011, IBM Watson (an AI application) was developed and won over the jeopardy champion. Today's machine learning (ML) such as "deep learning" and artificial neural networks (ANN) have created interesting use cases. AI has therefore regained confidence and industries are beginning to embrace where they see appropriate uses. Simultaneously, Internet of Things (IoT) tools have been introduced and made it possible to develop new capabilities such as virtual reality (VR), augmented reality (AR), mixed reality (MR) and extended reality (XR). These technologies are maturing and could be used in several application areas for the industries and form part of their digitalization journey. Furthermore, it is not only the industries that could benefit from introducing these technologies. Studies also show several areas and use cases where augmented reality has a positive impact, such as on students' learning ability. Yet few teachers know or use this technology. This paper evaluates and analyze AR, remote assistance tool for industrial purposes. The potential of the tool is discussed for frequent maintenance cases in the mining industry. Further on, if we look into the future, it is not surprising if we will be able to see that today's concepts of reality tools have evolved to become smarter by being trained by multimedia recognition and from people who have thus created an AI expert. Where the AI expert will support customers and be able to solve simple errors but also those that occur rarely and thus be a natural part of the solution for future completely autonomous processes for the industry. The article demonstrates a framework for creating smarter tools by combining AR, ML and AI and forms part of the basis creating the smarter industry of the future. Natural Language Processing (NLP) toolbox has been utilized to train and test an AI expert to give suitable resolutions to a specific maintenance request. The motivation for AR is the possible energy savings and reduction of CO₂ emissions in the maintenance field for all business trips that can be avoided. At the same time saving money for the industries and expert manhours that are spent on traveling and finally enhancing the productivity for the industries. Tests cases have verified that with AR, the resolution time could be significantly reduced, minimizing production stoppages by more than 50% of the time, which ultimately has a positive effect on a country's GDP. How much energy can be saved is predicted by the fact that 50% of all the world's business flights are replaced by one of the reality concepts and are estimated to amount to at least 50 Mton CO₂ per year. This figure is probably slightly higher as business trips also take place by other means of transport such as trains, buses, and cars. With today's volatile employees changing jobs more frequently, industry experts are becoming fewer and fewer. Since new employee stays for a maximum of 3–5 years per workplace, they will not stay long enough to become experts. Introducing an AI expert trained by today's experts, there is a chance that this knowledge can be maintained. [ABSTRACT FROM AUTHOR]

A number of magnetic refrigeration prototypes have been developed in recent years. Nevertheless, their optimization remains a challenging process. This paper presents an efficient approach based on artificial intelligence for optimizing the cooling performance of multi-layer active magnetic regenerators. To this end, a validated numerical model was used to predict the temperature difference between the hot and cold sources of a four-layer active magnetic regenerator. By using a nanofluid as a heat transfer fluid, the maximum temperature span of the device can be increased by approximately 20%. More importantly, by simultaneously optimizing a set of 10 key parameters, including the geometric parameters and the working conditions, the thermodynamic performance of the four-layer active magnetic regenerator prototype can be markedly enhanced by almost 40%. The newly established approach will be of considerable practical importance to both scientists and engineers since it will enable them to avoid costly experimental trials by optimizing a wide number of parameters. • A validated numerical model of a multilayer active magnetic regenerator is developed. • Thermodynamic performance optimization using artificial intelligence is performed. • A set of 10 geometric parameters and working conditions are optimized. • An enhancement of roughly 20% in cooling performance is achieved using a nanofluid. • An optimal set of parameters yields a nearly 40% improvement in cooling performance. [ABSTRACT FROM AUTHOR]

• Data driven method is used for Fuel Cell prognostics under a stationary load profile. • The parameters of the Echo State Network are optimized via an optimization algorithm. • Cycle repetetion and markov chains are used to generate the stationary load profile. • The remaining useful lifetime is estimated using the developed approach. • The prognostics performance is validated using experimental data. Improving Proton Exchange Membrane Fuel Cell durability is a key that paves the way to its large scale industrial deployment. During the last five years, the prognostics discipline emerged as an interesting field for Proton Exchange Membrane Fuel Cell state of health prediction and lifetime estimation. The information provided by the prognostic module is crucial for optimizing the control strategy to extend the fuel cell lifetime. In this paper, an approach based on Echo State Network for fuel cell prognostics under a variable load is developed. The novelty of this paper is to perform prognostics under a variable load profile without prior knowledge of this latter. Two solutions are developed in this work. The first one consists of evaluating the remaining useful lifetime under a repeated load cycle. The second one is based on using Markov chains to generate estimations of the future load profile, allowing thus to overcome the need of real future load profile prior knowledge. Both proposed solutions give accurate prediction results of proton exchange membrane fuel cell remaining useful lifetime, with low uncertainties. [ABSTRACT FROM AUTHOR]

• Taking real-time price as a feature affecting power load. • Selecting features by minimal redundancy maximal relevance. • Distinguishing holiday and non-holiday when selecting the historical load sequence. • Using an improved particle swarm optimization for parameter optimization. Accurate power load forecasting contributes to guaranteeing safe dispatch and stable operation of a power system. As a great forecasting tool, support vector machine is widely used in power load forecasting. However, due to the rapid development of information technology, the prediction result of simple support vector machine is no longer accurate enough to forecast in the smart grid. To enhance the prediction accuracy, this paper makes some improvements on support vector machine, and proposes a hybrid model integrated with intelligent methods for feature selection and parameter optimization. Firstly, real-time price becomes an important influencing factor of power load as people increasingly rely on demand and real-time price to adjust their electricity consumption patterns. Thus, real-time price, together with other factors that affect power load, is taken as a candidate feature, and minimal redundancy maximal relevance is applied to derive informative features from candidate features. Secondly, as for another feature, the historical load sequence, to make its selection more general, this paper employs the weighted gray relation projection algorithm for holidays to be predicted. Finally, second-order oscillation and repulsion particle swarm optimization is used for optimizing parameters of support vector machine. Moreover, the proposed model is tested via simulations on datasets of Singapore. By comparing prediction results of the proposed model, the support vector machine before improvement and other three forecasting models, this paper confirms that the improvements on support vector machine are effective, and the proposed model outperforms the other forecasting models in aspect of accuracy, stability and effectiveness. [ABSTRACT FROM AUTHOR]

Smart grid is the new trend for clean, sustainable, efficient and reliable energy generation, delivery and use. To ensure stable and secure operation is essential for the smart grid, which needs effective stability analysis and control. As the smart grid has evolved through a growing scale of interconnection, increasing integration of renewable energy, widespread operation of direct current power transmission systems, and liberalization of electricity markets, the stability characteristics of it are much more complex than the past. Due to these changes, conventional stability analysis and control approaches have a series of drawbacks in terms of speed, effectiveness and economy. On the contrary, the emerging artificial intelligence (AI) techniques provide powerful and promising tools for stability analysis and control in smart grids and have attracted growing attention. This paper aims to give a comprehensive and clear picture of recent advances in this research area. First, we present a general overview of AI, including its definitions, history and state-of-the-art methodologies. And then, this paper gives a comprehensive review of its applications to security assessment, stability assessment, fault diagnosis, and stability control in smart grids. These applications have achieved impressive results. Nevertheless, we also identify some major challenges these applications face in practice: high requirements on data, imbalanced learning, interpretability of AI, difficulties in transfer learning, the robustness of AI to communication quality, and the robustness against attack or adversarial examples. Furthermore, we provide suggestions for potential important future investigation directions to overcome these challenges and bridge the gap between research and practice. • An overview of state-of-the-art artificial intelligence (AI) is provided. • A review of AI's applications to smart grid stability and control is undertaken. • This paper identifies some major challenges these applications face in practice. • Suggestions and potentially important future research directions are put forward. [ABSTRACT FROM AUTHOR]

• More accurate and explanatory model is proposed for natural gas storage deliverability. • A hybrid Sparse Autoencoder and Random Forest model shows on average of 5.7% increase in prediction accuracy. • The first use of machine learning and explanatory methods (XAI) in deliverability prediction for natural gas storage. Numerous subsurface factors, including geology and fluid properties, can affect the connectivity of the storage spaces in depleted reservoirs; hence, fluid flow simulations become more complicated, and predicting their deliverability remains challenging. This paper applies Machine Learning (ML) techniques to predict the deliverability of underground natural gas storage (UNGS) in depleted reservoirs. First, three baseline models were developed based on Support Vector Regression (SVR), Artificial Neural Network (ANN), and Random Forest (RF) algorithms. To improve the accuracy of the RF model as the best-performing baseline model, a unified framework, referred to as SARF, was developed. SARF combines the capabilities of Sparse Autoencoder (SA) and that of Random Forest (RF). To achieve this, the internal representations of the SA, which constitute extracted features of the input variables, are used in RF to develop the proposed SARF framework. The predictive capabilities of the baseline models and the proposed SARF model were validated using 3744 real-world storage data samples of 52 active storage reservoirs in the United States. The experimental result of this study shows that the proposed SARF model achieved an average 5.7% increase in accuracy on four separate data partitions over the baseline RF model. Furthermore, a set of eXplainable Artificial Intelligence (XAI) methods were developed to provide an intuitive explanation of which factors influence the deliverability of reservoir storage. The visualizations developed using the XAI method provide an easy-to-understand interpretation of how the SARF model predicted the deliverability values for separate reservoirs. [ABSTRACT FROM AUTHOR]

The pressure difference between the inlet and outlet of the reactant in fuel cells is called the pressure drop, which is related to the water amount inside the fuel cells. In recent years there have been many studies that used the pressure drop to detect the water content and diagnose water fault of proton exchange membrane fuel cells (PEMFCs). To our knowledge, there has not been a systematic review of these studies. In this paper, the effect variables of pressure drop are reviewed firstly. Then estimations of the theoretical pressure drop are reviewed mainly based on the following four aspects: Bernoulli’s equation, two-phase flow multiplier, Darcy’s law and artificial intelligence. Afterward, the water fault diagnosis based on the pressure drop using the following six indicators are reviewed: indicator of direct pressure drop, its deviation, frequency, multiplier, the ratio of pressure drop to flow rate and the flooding degree. In addition, the strategies of water fault recovery are also summarized. Finally the merits, demerits and application prospects of pressure drop-based water fault diagnosis are presented. [ABSTRACT FROM AUTHOR]

Different oil price shock incentives under different domestic and international environment will cause different oil price shocks and bring different impacts to China’s macroeconomy. However, there are few empirical studies on early warning prediction of the co-movements between oil price shocks and macroeconomy. This paper presents an incentive-oriented artificial intelligent (AI) early warning system (EWS) with ontology supported case based reasoning (CBR) method, called “relationship between oil price shocks and economy-an early warning system (ROSE 2 )”, to forecast the co-movements between macroeconomy and oil price shocks in China. Simultaneously, multi-galois lattice (MGL), which is more suitable for matching multiple attributes, is used to improve the recall and precision capability of ROSE 2 . Finally, several practical queries called Q1–Q4 are presented for verifying the validation and efficiency of the ROSE 2 system. [ABSTRACT FROM AUTHOR]

• A novel system aiming at gas demand response from domestic gas boilers is presented. • A novel methodology to calculate the gas demand baseline is provided. • Neural networks are used to forecast the indoor conditions and the boiler modulation. • Optimization of consumer's welfare from gas supply and market participation. • The presented gas DR system handles efficiently the so-called "rebound effect" This paper presents a novel, fully automated system for the implementation of explicit gas Demand Response in the Gas Balancing Market, with real-time control instructions provided to domestic gas boilers. The Demand Response is applied in both the upward and downward direction, enabling the respective gas consumer to provide both upward and downward Balancing Services to the Gas Transmission System Operator. The system targets at the welfare maximization of domestic gas consumers, i.e., the revenues attained by providing Balancing Services to the Gas Transmission System Operator minus the cost of gas supply, while maintaining the house's indoor temperature within the residents' comfort limits. Real-time data are derived from interconnected commercial thermostats and used by two Recurrent Neural Networks for each house, in order to attain forecasts for the indoor temperature change and the expected boiler's modulation levels for the next future time intervals. Such forecasted condition changes within each house are then considered in an optimization model that results in the optimal instruction signal that must be provided to the boiler. The signal concerns the boiler operation level for the following 5 min duration. A Genetic Algorithm is employed for the optimization problem solution. The whole system is deployed using a containerized software architecture to ensure scalability and full-service availability. The implemented real-world tests exhibit that domestic consumers can increase their profits from both gas consumption minimization and from the provision of gas DR services in real-time. [ABSTRACT FROM AUTHOR]

[Display omitted] • The first ever AI-enabled thermo-mechanical optimization of a TEG using DNN is presented. • Nine strategic parameters covering overall performance of TEGs are studied. • DNN accurately learns and predicts FEM data that took 8 h to generate in just 8 s. • Optimization guidelines for manufacturing efficient TEGs with longer lifespans are given. • Possible avenues for applying DNN in optimizing cutting-edge TEG designs are disclosed. The rising levels of global warming in the environment owing to emissions from fossil-fuel-based engines has increased the search for efficient clean energy systems. Thermoelectric generators (TEGs) standout as a promising energy conversion device which can directly convert heat to electricity. Several optimization studies have been carried out on these devices to improve their power generation rate and efficiencies while guaranteeing long lifespan. However, the limitations of finite element methods (FEMs) in easily providing optimization guidelines at a fast rate has hindered the manufacture of TEGs with high thermo-mechanical performance. This is why this paper presents the first ever artificial intelligence enabled optimization of a TEG conducted via deep neural networks (DNNs). Previous research on this topic completely neglected the mechanical performance and consequently, the service lifetime of the TEG when exposed to thermal operating conditions. To fill this gap, the effects of strategic parameters on the power output, efficiency and thermal stress performances of the TEG are investigated. These parameters are the hot and cold junction temperatures/heat transfer coefficients, incident heat flux, external load resistance, TE leg height, area, and area ratio. The DNN is fed with verified three-dimensional FEM simulations carried out on the ANSYS Workbench platform. The FEM results exhibit almost perfect correlation with experimental data which establishes the precision of the model. Results are that the DNN model is able to provide the necessary optimization guidelines for maximum thermo-mechanical performance in just 8 seconds compared to the 8 hours required by the FEM. The TEGs that were modelled using the DNN optimization guidelines improved the power, efficiency and mechanical performance of the unoptimized TEG by 11.94%, 14.17%, and 91%, respectively. These results are sufficient to provide useful guidelines at a fast rate for the fabrication of high-efficiency TEGs that will operate for a long time. [ABSTRACT FROM AUTHOR]

• An adaptive boosting charging strategy for optimizing charging time and capacity attenuation for batteries in EVs is proposed. • A coupled thermoelectric-aging battery model is established. • Multi-objective optimization for optimising capacity attenuation and charging time simultaneously is performed. • Results shows the proposed strategy reduce the capacity loss by 3.6% when compared with CCCV strategy. This paper presents an adaptive boosting charging strategy incorporating the capacity estimation method based on the operation parameters, which is used for the state estimation and the adaptive adjustment of the charging strategy during the charging/discharging cycling process. Firstly, a coupling thermoelectric-aging battery model involving second-order resistor-capacity equivalent circuit model, two-state thermal model, semi-empirical aging models is set up for the simulation purpose. Secondly, numerical surrogate models describing the charging time and capacity loss in different aging states and charging strategies and a capacity estimation model are generated based on the coupling battery model, design of experiment method and artificial intelligence method. Subsequently, a multi-objective optimization framework based on surrogate models and the second non-dominated sorting genetic algorithm is used to solve the optimization problem of the boosting charging strategy to balance the charging time and capacity attenuation under different aging states. Finally, the simulation of the processes of capacity estimation and charging parameter optimal selection is carried on, and the simulation and comparison of the charging/discharging cycles applying the proposed and standard constant current constant voltage charging strategy are conducted. The results show that the proposed strategy can restrict the capacity loss at 4.44% which is 4.2% higher than the capacity loss by applying standard CCCV, the average charging time decreases from 3792.3 s to 2881.4 s. The average temperature increment is increased significantly, while it is still in the ideal operating temperature, and the average charging time can be significantly decreased. Based on this work, further researches can be conducted to develop the actual boosting charging management systems including aging state estimation and adaptive charging parameter selection. [ABSTRACT FROM AUTHOR]

• A risk-averse deep learning method for real-time emergency load shedding. • Deep learning algorithm to enhance prediction accuracy on load shedding strategy. • A shedding amount inference model to speculate the stability after load shedding. • A risk-averse loss function to minimize the risk of load under-cutting events. • Experiments on multiple power systems to verify the practical values of the method. Emergency load shedding is an effective and frequently used emergency control action for power system transient stability. Solving the full optimization models for load shedding is computational burdensome and thus slow react to the intense system variations from the increasing renewable energy sources and the more active demand-side behavior. Other sensitivity-based methods impair the control accuracy and may not guarantee global optimality. Artificial intelligence methods, as the data-driven approaches, have recently been well-recognized for its real-time decision-making capability to tackle the system variations. The existing artificial intelligence methods for emergency load shedding are based on shallow learning algorithms and can lead to both load under-cutting and over-cutting events. However, when the loads are under-cut, the power system will be exposed to a high risk of post-control instability that can propagate into cascading events, which incurs significantly higher cost than an over-cutting event. Being aware of such unbalanced control costs, this paper proposes a risk-averse deep learning method for real-time emergency load shedding, which trains deep neural network towards the reluctance to load under-cutting events, so as to avoid the huge control cost incurred by control failure. The case studies on two renewable power systems demonstrate that, compared to the state-of-the-art methods, the proposed risk-averse method can significantly improve the control success rate with negligible increase in prediction error, ending up with lower overall control cost. The results verify the enhanced control performance and the practical values of the proposed method. [ABSTRACT FROM AUTHOR]

[Display omitted] • A transfer learning routine is proposed to reduce workload in control development. • Fuzzy learning and Gaussian process regression are incorporated in a control model. • Fuzzy learning achieves 27% higher distinction rate than a neural network. • The 'deeper' architecture outpefroms the 'broader' one in transfer learning. • Real-time control functionality is verified by hardware-in-the-loop testing. Electric vehicles, including plug-in hybrids, are important for achieving net-zero emission and will dominate road transportation in the future. Energy management, which optimizes the onboard energy usage, is a critical functionality of electric vehicles. It is usually developed following the model-based routine, which is conventionally costly and time-consuming and is hard to meet the increasing market competition in the digital era. To reduce the development workload for the energy management controller, this paper studies an innovative transfer learning routine. A new transferable representation control model is proposed by incorporating two promising artificial intelligence technologies, adaptive neural fuzzy inference system and Gaussian process regression, where the former applies k-fold cross valudation to build a neural fuzzy system for real-time implementation of offline optimization result, and the later connects the neural fuzzy system with a 'deeper' architecture to transfer the offline optimization knowledge learnt at source domain to new target domains. By introducing a concept of control utility that evaluates vehicle energy efficiency with a penalty on usage of battery energy, experimental evaluations based on the hardware-in-the-loop testing platform are conducted. Competitive real-time control ultility values (as much as 90% of offline benchmarking results) can be achieved by the proposed control method. They are over 27% higher than that achieved by the neural-network-based model. [ABSTRACT FROM AUTHOR]

Abstract: The electric grid is changing. With the smart grid the demand response (DR) programs will hopefully make the grid more resilient and cost efficient. However, a scheme where consumers can directly participate in demand management requires new efforts for forecasting the electric loads of individual consumers. In this paper we try to find answers to two main questions for forecasting loads for individual consumers: First, can current short term load forecasting (STLF) models work efficiently for forecasting individual households? Second, do the anthropologic and structural variables enhance the forecasting accuracy of individual consumer loads? Our analysis show that a single multi-dimensional model forecasting for all houses using anthropologic and structural data variables is more efficient than a forecast based on traditional global measures. We have provided an extensive empirical evidence to support our claims. [Copyright &y& Elsevier]

The use of wind power, a pollution-free and renewable form of energy, to generate electricity has attracted increasing attention. However, intermittent electricity generation resulting from the random nature of wind speed poses challenges to the safety and stability of electric power grids when wind power is integrated into grids on large scales. Therefore, accurate forecasting of wind speed and wind power (WS/WP) has gradually taken on a key role in reducing wind power fluctuations in system dispatch planning. With the development of artificial intelligence technologies, especially deep learning, increasing numbers of deep learning-based models are being considered for WS/WP forecasting due to their superior ability to deal with complex nonlinear problems. This paper comprehensively reviews the various deep learning technologies being used in WS/WP forecasting, including the stages of data processing, feature extraction, and relationship learning. The forecasting performance of some popular models is tested and compared using two real-world wind datasets. In this review, three challenges to accurate WS/WP forecasting under complex conditions are identified, namely, data uncertainties, incomplete features, and intricate nonlinear relationships. Moreover, future research directions are summarized as a guide to improve the accuracy of WS/WP forecasts. • Review and compare deep feature extraction models in terms of computation cost. • Review and discuss various single and hybrid deep models for relationship learning. • Present an overview of intelligent optimizers for model configuration determination. • Summarize challenges and future research directions in wind energy forecasting. [ABSTRACT FROM AUTHOR]

[Display omitted] • A novel structure-adaptative grey model with a time power is proposed. • The Cultural Algorithm is employed for determining the optimal time power item. • Monte-Carlo Simulation and Probability Density Analysis are originally introduced to enhance the model's robustness. • Empirical results comprehensively verify the efficacy of our novel method. • The future demands for nuclear energy are predicted by using the promising model from 2019 to 2023. Accurate estimations of nuclear energy consumption are an essential process for formulating appropriate policies and plans in the energy sector and associated companies. This paper presents a novel structure-adaptive grey model with an adjustable time power based on the nonlinear and complicated characteristics of nuclear energy consumption, in which three core innovations are summarized below. Initially, the generalized time response function for projections is theoretically deduced, which overcomes the fundamental flaws in the conventional grey model. Subsequently, the Cultural Algorithm is employed to determine the optimum values of the time power item to improve the adaptability and flexibility to confront diverse forecasting issues. Further, Monte-Carlo Simulation and Probability Density Analysis (PDA) are originally introduced to enhance the robustness of the proposed model. For illustration and verification purposes, experiments on predicting nuclear energy consumption in China and America are conducted in comparison with a range of benchmark models, including other prevalent grey models, conventional econometric technology, and artificial intelligences. The performance of the novel technique is evaluated from two different perspectives of PDA and level accuracy, confirming that this model is a very promising and powerful tool for predicting nuclear energy demands in China and America from 2019 to 2023. [ABSTRACT FROM AUTHOR]

In this paper, we present a newly developed eXplainable artificial intelligence (XAI) model to analyze the impacts of climate change on the cooling energy consumption (E c) in buildings, predict long-term E c under the new shared socioeconomic pathway (SSP) climate change scenarios, and explain the underlying reasons behind the predictions. Such analyses and future predictions are imperative to allow decision-makers and stakeholders to accomplish climate-resilient and sustainable development goals by leveraging the power of meaningful and trustworthy projections and insights. We demonstrated that the XAI is capable of predicting the E c under future climate scenarios with high accuracy ( R 2 > 0. 9) and reveals the critical inflection points of the daily average outdoor air temperature (T a) beyond which the E c increase exponentially. We applied the XAI model for residential and commercial buildings in hot–humid and mixed–humid climate regions to quantify the incremental impacts of climate change on E c under the different SSPs. The XAI-based analysis concluded positive and persistent incremental changes in the E c from 2020 to 2100 under all future SSP scenarios, with the maximum incremental impact of 24.5%, 33.3%, 57.8%, and 87.2% in hot–humid and 37.1%, 47.5%, 85.3%, and 121% in mixed–humid climate regions under the sustainable green energy (SSP126), business-as-usual (SSP245), challenges to adaptation (SSP370), and increased reliance on fossil fuels (SSP585) scenarios, respectively. Potential increases in the E c in future climates could have significant adverse impacts on the local and regional economy if necessary adaptation and mitigation measures are not implemented a priori. • XAI predicts building cooling energy demand under climate change. • XAI reveals the critical reason for increase in cooling energy consumption. • Space cooling energy demand is predicted to increase exponentially after 2050. • Climate change impacts on cooling energy needs are worse in residential buildings. [ABSTRACT FROM AUTHOR]

Enormous amounts of data are being produced everyday by sub-meters and smart sensors installed in residential buildings. If leveraged properly, that data could assist end-users, energy producers and utility companies in detecting anomalous power consumption and understanding the causes of each anomaly. Therefore, anomaly detection could stop a minor problem becoming overwhelming. Moreover, it will aid in better decision-making to reduce wasted energy and promote sustainable and energy efficient behavior. In this regard, this paper is an in-depth review of existing anomaly detection frameworks for building energy consumption based on artificial intelligence. Specifically, an extensive survey is presented, in which a comprehensive taxonomy is introduced to classify existing algorithms based on different modules and parameters adopted, such as machine learning algorithms, feature extraction approaches, anomaly detection levels, computing platforms and application scenarios. To the best of the authors' knowledge, this is the first review article that discusses anomaly detection in building energy consumption. Moving forward, important findings along with domain-specific problems, difficulties and challenges that remain unresolved are thoroughly discussed, including the absence of: (i) precise definitions of anomalous power consumption, (ii) annotated datasets, (iii) unified metrics to assess the performance of existing solutions, (iv) platforms for reproducibility and (v) privacy-preservation. Following, insights about current research trends are discussed to widen the applications and effectiveness of the anomaly detection technology before deriving future directions attracting significant attention. This article serves as a comprehensive reference to understand the current technological progress in anomaly detection of energy consumption based on artificial intelligence. • Anomaly detection of energy consumption in buildings. • Taxonomy of artificial intelligence based anomaly detection methods. • Current trends and new perspectives of anomaly detection. • Non-intrusive anomaly detection. • Multimodal anomaly detection visualization techniques. [ABSTRACT FROM AUTHOR]

A new methodology, that combine three different Artificial Intelligence techniques, is proposed in this paper to solve the energy demand planning in Smart Homes. Conceived as a multi-objective scheduling problem, the new method is developed to reach the compromise between energy cost and the user comfort. Using an Elitist Non-dominated Sorting Genetic Algorithm II, the concept of demand-side management is applied taking into account electricity price fluctuations over time, priority in the use of equipment, operating cycles and a battery bank. The demand-side management also considers a forecast of a distributed generation for a day ahead, employing the Support Vector Regression technique. Validated by numerical simulations with real data obtained from a smart home, the user comfort levels were determined by the K-means clustering technique. The efficiency of the proposed Artificial Intelligence combination was proved according to a 51.4% cost reduction, when Smart Homes with and without distributed generation and battery bank are compared. • Three Artificial Intelligence techniques were used for energy planning in Smart Home. • The methodology uses Support Vector Regression for power forecast for a day ahead. • The user comfort levels were determined by a K-means clustering technique. • NSGA-II was applied to solve energy planning, taking into account a battery bank. [ABSTRACT FROM AUTHOR]

• Online network reconfiguration is introduced to manage distribution networks. • A Deep Q-learning algorithm is used to find the optimal network topology. • The proposed algorithm is more scalable than other algorithms. • The computation time of the algorithm is low enough for practical applications. • The algorithm can increase DRES utilization without heavy capital investment. With increasing number of distributed renewable energy sources integrated in power distribution networks, network security issues such as line overloading or bus voltage violations are becoming increasingly common. Traditional capital-intensive system reinforcements could lead to overinvestment. Moreover, active network management solutions, which have emerged as important alternatives, may become a financial burden for distribution system operators or reduce profits for owners of distributed renewable energy sources, or both. To address these limitations, this paper proposes an online network reconfiguration scheme based on a deep reinforcement learning approach. In this scheme, the distribution network operator modifies the network topology to change the power flow when the reliability of network is threatened. Because the variability of distributed renewable energy is large in self-sufficient distribution networks, the reconfiguration process needs to be performed online within short time intervals, which involves the use of conventional algorithms. To solve this problem efficiently, a deep q-learning model is utilized to determine the optimal network topology. Performances of proposed and other algorithms were compared in modified CIGRE 14-bus and IEEE 123-bus test network, as well as varying penalties for frequent switching operation in consideration of physical characteristic of the network. Simulation results demonstrated that the proposed algorithm showed almost identical performances with brute-force search algorithm in both test networks, satisfying network constraints over almost all timespans. Further, the proposed method required very small computation times - under a second per each state and its scalability was verified by comparing the computation time between two test networks. [ABSTRACT FROM AUTHOR]

Non-domestic buildings contribute 20% of the UK's annual carbon emissions. A contribution exacerbated by its ageing stock of which only 7% is considered new-build. Consequently, the government has set regulations to decrease the amount of energy take-up by buildings which currently favour deep energy retrofitting analysis for decision-making and demonstrating compliance. Due to the size and complexity of non-domestic buildings, identifying optimal retrofit packages can be very challenging. The need for effective decision-making has led to the wide adoption of artificial intelligence in the retrofit strategy design process. However, the vast retrofit solution space and high time-complexity of energy simulations inhibit artificial intelligence's application. This paper presents an energy performance prediction model for non-domestic buildings supported by machine learning. The aim of the model is to provide a rapid energy performance estimation engine for assisting multi-objective optimisation of non-domestic buildings energy retrofit planning. The study lays out the process of model development from the investigation of requirements and feature extraction to the application on a case study. It employs sensitivity analysis methods to evaluate the effectiveness of the feature set in covering retrofit technologies. The machine learning model which is optimised using advanced evolutionary algorithms provide a robust and reliable tool for building analysts enabling them to meaningfully explore the expanding solution space. The model is evaluated by assessing three thousand retrofit variations of a case study building, achieving a root mean square error of 1.02 k g C O 2 ∕ m 2 × y e a r equal to 1.7% of error. • Addressed impracticality of optimisation methods in supporting deep energy retrofit. • A machine learning model to calculate non-domestic buildings' energy performance. • Identified essential features of non-domestic buildings affecting energy performance. • TOptimised the model by tuning hyperparameters using an evolutionary algorithm. • Evaluated model performance by scoring thousands of variations in a sample building. [ABSTRACT FROM AUTHOR]

• DRL-based control methods are proposed for ORC for engine waste heat recovery. • The DRL control for ORC superheat performs considerably better than PID control. • The DRL-based PID control is more robust than the DRL control. • The switching DRL control performs well and exhibits sufficient robustness. • Useful reference and motivation for application of DRL to thermodynamic systems. The organic Rankine cycle (ORC) is a promising technology for engine waste heat recovery. During real-world operation, the engine working condition varies frequently to satisfy the power demand; thus, the transient nature of engine waste heat presents significant control challenges for the ORC. To control the superheat of the ORC precisely under a transient heat source, several optimal control methods have been used such as model predictive control and dynamic programing. However, most of them depend strongly on the accurate prediction of future disturbances. Deep reinforcement learning (DRL) is an artificial-intelligence algorithm that can overcome the aforementioned disadvantage, but the potential of DRL in control of thermodynamic systems has not yet been investigated. Thus, this paper proposes two DRL-based control methods for controlling the superheat of ORC under a transient heat source. One directly uses the DRL agent to learn the control strategy (DRL control), and the other uses the DRL agent to optimize the parameters of the proportional–integral–derivative (PID) controller (DRL-based PID control). Additionally, a switching mechanism between different DRL controllers is proposed for improving the training efficiency and enlarging the operation range of the controller. The results of this study indicate that the DRL agent can satisfactorily perform the control task and optimize the traditional controller under the trained and untrained transient heat source. Specifically, the DRL control can track the reference superheat with an average error of only 0.19 K, whereas that of the traditional PID control is 2.16 K. Furthermore, the proposed switching DRL control exhibits excellent tracking performance with an average error of only 0.21 K and robustness over a wide range of operation conditions. The successful application of DRL demonstrates its considerable potential for the control of thermodynamic systems, providing a useful reference and motivation for the application to other thermodynamic systems. [ABSTRACT FROM AUTHOR]