Search

The increasing expansion of photovoltaic power generation leads to unpredictable fluctuations in electricity supply, which can potentially jeopardize the stability of the power grid and escalate the costs associated with grid imbalances. As a result, precise forecasts of photovoltaic power generation play a vital role in optimizing capacity deployment, enhancing consumption levels, improving planning strategies, and maintaining grid balance within systems characterized by significant penetration of solar energy. This paper proposes a structural adaptive grey seasonal model based on data reorganization. Solar photovoltaic power generation data typically exhibit seasonal fluctuations, which pose a challenge to existing prediction techniques. Therefore, this paper adopts the idea of data reorganization to eliminate the seasonal fluctuations of observations, and the adaptive accumulation operator can accurately simulate the change trend of the original data in different periods, overcoming the defect of insufficient adaptability of the traditional accumulation operator. Subsequently, the time trend items are incorporated into the model structure to identify the trend characteristics of system development, which can effectively explain the power generation trend of photovoltaic systems at different time periods and improve the prediction accuracy of the model. In addition, the compatibility and unbiased nature of the proposed model have been demonstrated to help us better perceive the model. The Grey Wolf Optimizer (GWO) is used to optimize the adaptive parameters of the model, endowing it with higher flexibility and stronger adaptability. In order to verify the effectiveness of the model, three practical cases (namely quarterly solar power generation in the United States, Japan, and Germany) were compared with existing econometric techniques, artificial neural networks, and grey prediction methods. The experimental results show that the new model outperforms other benchmark models in both simulation and prediction performance, and enjoys high robustness. • A novel grey seasonal model with data reorganization is established. • The proposed model can grasp nonlinear and fluctuant in various datasets. • The robustness analysis of the novel model is established. • Comparative study shows that the novel model is superior to other models. • Relevant suggestions are put forward according to the prediction results. [ABSTRACT FROM AUTHOR]

Electricity load exhibits daily and weekly cyclical patterns as well as random characteristics. At present, prevailing deep learning models cannot learn electricity load cyclical and stochastic features adequately. This results in insufficient prediction accuracy and the scalability of current methods. To tackle these difficulties, this paper proposes a framework for electrical load prediction based on an Attention Mechanism Time Series Depthwise Separable Convolutional Neural Network (ELPF-ATDSCN). The framework starts by using the Maximum Information Coefficient for exogenous variable selection. It then incorporates a seasonal decomposition algorithm with manual feature engineering to extract the cyclical and stochastic features of the electrical load. Subsequently, the framework employs the ATDSCN to learn the cyclical and stochastic features of the electrical load. In addition, the Bayesian algorithm optimizes model hyperparameters for optimal model performance. Experimental results of point and interval load prediction on datasets from the US and Nordic power markets reveal that the ATDSCN model proposed in this paper enhances load prediction accuracy compared with other models. It can provide more reliable predictions for power system operation and dispatch. • A new attention mechanism is designed to improve the model memory capability. • The ATDSCN model is proposed for learning periodic and stochastic features of electricity loads. • A framework for electricity load prediction is proposed to improve the accuracy of electricity load forecasting. • The effectiveness of the proposed method is validated on the real world dataset. [ABSTRACT FROM AUTHOR]

Wind power is a critical source of renewable energy, but the variability of wind speed can impact the efficiency and reliability of wind power generation. To improve the accuracy of wind speed prediction, this paper proposes a novel system that combines deep learning and decomposition algorithms based on ensemble optimization methods. Firstly, the paper proposes an innovative hybrid modal decomposition (HMD) method that extracts accurate features from wind speed data using Singular Value Decomposition (SVD), modal number selection method, and Variational Mode Decomposition (VMD). Secondly, the variable screening method identifies relevant factors affecting wind speed as fixed inputs. Then, the stationary test method is used to identify the sequences of different features in the mode and establish different prediction models: a bidirectional long-term short-time neural network model (Bi-LSTM) model optimized by artificial hummingbird algorithm (AHA) is used for non-stationary sequences; an ARIMA model is used for stationary sequences. Data from Sotavento and Changma wind farms are used to test this forecasting system. Simulation experiments are carried out from multiple dimensions. The results show that all error indicators are significantly improved, and the proposed prediction system is better than other comparative forecasting schemes. • The research can be practically applied in the field of wind power generation. • The research can help improve the energy utilisation of wind power generation. • A new data decomposition algorithm called HMD is proposed. • A new prediction system based on ensemble optimization idea is proposed. • Combining AHA-optimized neural networks and ARIMA to predict wind speed. [ABSTRACT FROM AUTHOR]

This paper introduces an innovative decision support system for improving ship energy efficiency. It combines an engine optimization model and an artificial neural network (ANN). Real-time data regarding the ship and engine performance is obtained along a specific ship navigation, and an engine optimization model is built using Ricardo Wave software. This model is validated with high accuracy. Ship-specific parameters are derived, and fuel consumption is estimated using ANN models with different structures. The study identifies the most suitable ANN model with 1 hidden layer and 5 neurons based on error metric evaluation where the 0.99697 R2, 0.00035 RMSE, and 0.06470 MAPE scores are calculated. This approach offers a cost-effective solution for shipping companies to monitor critical engine parameters in real-time without investing in sensors or data collection systems. Thus, it contributes to dealing with the problem of the scarcity of analyzable data in the maritime literature, which is one of the significant issues in the papers related to energy efficiency, machine learning, and deep learning. It is an innovative approach since the parameters that are related to real-time operations rather than considering only instruction book information are aggregated, and the produced data is analyzed by an intelligent tool. • A system is proposed to control ship energy efficiency level by combining an optimization model and a neural network • A simulation model is created in the Ricardo Wave tool and the engine model is validated by actual data. • The ship fuel consumption is estimated by an artificial neural network based on the derived input parameters. • The prediction process is evaluated by error metrics to determine its success. [ABSTRACT FROM AUTHOR]

Solar radiation is an essential meteorological parameter for building energy efficiency analysis, and the quality of the data directly affects the analysis results. This paper investigates the estimation of hourly solar radiation based on the generation of the typical meteorological year(TMY) using various real meteorological parameters and limited solar radiation data. The focus of this paper is to use two types of neural network algorithms to improve the estimation accuracy and applicability, and to solve the problem of hourly solar radiation data acquisition in non-radiation areas. First, select two city station data and use three methods to generate TMY. Then, two neural network models, BP Neural Network (BP)，Convolutional Neural Network (CNN) are used to estimate the hourly solar radiation data and verify the results. Finally, by constructing a photovoltaic-integrated office building model, the accuracy of the hourly solar radiation estimation model is verified using energy consumption simulation and photovoltaic (PV) power generation simulation. The results show that this paper can well solve the problem of limited radiation data, which provides a new idea for the study of building energy efficiency in areas where radiation data is missing. • Multiple neural network is used to estimate the generation of solar radiation data using other meteorological parameters. • An advanced neural network is used to provide an effective TMY generation method in missing hourly solar radiation data. • The accuracy of the experimental results is verified by energy consumption simulation and photovoltaic power generation. [ABSTRACT FROM AUTHOR]

Accurately and efficiently predicting the LHV of MSW is vital for designing and operating a waste-to-energy plant. However, previous prediction models possess limited geographical applicability. In this paper, we employ multiple linear regression and artificial neural network (ANN) techniques to predict LHV. These data-driven models utilize 151 globally distributed datasets identified during a systematic literature review, describing the wet physical composition of MSW and measured LHV. The results show that models built via both methods exhibited acceptable and compatible levels of performance in predicting LHV, based on the multiple statistical indicators. However, the ANN model proved to be more robust in handling of datasets of diverse quality. Models developed from both methods demonstrate clearly that the wet proportion of food waste has a negative impact on LHV. Supported by the strong and significant correlation between food waste and moisture content, we concluded that the negative impact of high moisture content in food waste on LHV outweighed its calorific value. Separating food waste or any other waste with high moisture content from the MSW for incineration can significantly improve energy recovery efficiency. Contrary to expectation, the models also reveal a higher contribution of paper waste to the LHV of MSW than plastic waste. • Models estimating LHV based on worldwide MSW composition were generated. • Models using multiple linear regression analysis and ANN techniques were compared.. • Models were found to yield acceptable accuracy and were broadly compatible. • ANN models prove to be more robust than regression models.. • Results indicate that high moisture content waste should be separated at source. [ABSTRACT FROM AUTHOR]

Sustainable energy planning for power generation with renewables requires a multidimensional approach since it directly impacts several sectors such as the national grid, the environment, and the economy. This paper suggests an intelligent decision based on a combination of the Analytical Hierarchy Process and Artificial Neural Networks to assist policymakers in evaluating future renewable energy investments for power generation. A case study of Bahrain's power system is used to illustrate the usefulness of the proposed planning approach and also to discuss its efficiency. The Analytical Hierarchy Process model outcomes revealed that wind turbines are the most appropriate technology for Bahrain with 32.5 % priority, followed by PV and CSP with 32.2 % and 16.3 %, respectively. Then, the Artificial Neural Networks model was structured based on the generated scenarios from the Analytical Hierarchy Process model, and it reached its best performance after 384 cycles. The integrated approach overcomes some of the Analytical Hierarchy Process's limitations and provides an intelligent tool for advanced assessment of a sustainable power system. • Planning addresses intermittent nature of renewable energies. • Sustainability study integrates diverse, context-sensitive perspectives. • Tool merges Analytical Hierarchy Process with Neural Networks. • Developed tool applied to Bahrain's renewable energy context. [ABSTRACT FROM AUTHOR]

The paper proposes an energy management system, which considers the efficiency of the reversible pump/turbine that varies nonlinearly depending on the water flow rate during the pump/turbine modes of operation. A vibration avoidance strategy of the reversible pump hydro storage is developed. A probabilistic approach based on artificial neural networks and Naive-based is used. Through minimizing the levelized Cost of Energy (COE), this study shows the optimal size and reconfiguration of the HMG system as well as purchased/sold energy. Two novel modified optimizers based on the Particle Swarm Optimization (PSO) and the Aquila Optimizer (AO), namely: AO initialized PSO and AO updated PSO are developed. The results via the PSO and AO optimizers are compared in terms of reducing the COE and attaining a low execution time. Based on the results, a COE of 0.22 $/kWh through the developed strategy could be obtained with CO 2 emissions of 1974 ton/year against 0.24 $/kWh and 2460 ton/year using the PSO, which saves 24.6% of the yearly CO 2 emissions. Furthermore, the vibration avoidance strategy avoids the dead zones and enables the reversible pump/turbine machine to operate at higher efficiencies — both of which are impossible to achieve in the occurrence of vibrations. • Study of vibration zones and weather on the reversible pump hydro storage sizing. • Developing hybrid optimization to boost the renewables and the double storage. • Modeling the technical and economic viability of the battery reversible machine. • Sensitivity study of the reversible pump hydro-storage with PV, and wind turbine. [ABSTRACT FROM AUTHOR]

Predictive models serving as virtual sensors for online optimization and feedback control of diesel engines is gaining increasing attention. However, existing prediction models fall short in simultaneously achieving high prediction accuracy and fast computational speed. In this paper, a novel machine learning algorithm called deep autoencoding support vector regression (DASVR) was proposed, which combines the powerful non-linear feature extraction capability of artificial neural network (ANN) with the good adaptability of support vector regression (SVR) to low-dimensional input spaces. ANN-based autoencoder is firstly employed to extract features from the original high-dimensional input space, forming a low-dimensional latent variable space. SVR is then employed to perform non-linear mapping between latent variables and target output variables. Experiments were conducted on a 16-cylinder marine diesel engine under different load, rail pressure, and injection timing conditions. Prediction models for fuel consumption, NOx, PM, HC, and PM emissions were established based on experimental data and DASVR algorithm. The maximum error of DASVR-based models for the five desired output variables is less than 3.8 % under different operating conditions. DAVSR-based predictive models outperform conventional ANN-based and SVR-based models in terms of both prediction accuracy and real-time capability, and can theoretically meet the real-time requirement below 3000 r/min. • Proposal for a novel machine learning algorithm - DASVR. • Reduce the dimensionality of input space by encoder-decoder neural network. • Map the latent variable space to the output space by SVR. • DASVR outperforms ANN and SVR in both accuracy and speed. • Detailed evaluation of prediction accuracy and speed for both BSFC and emissions. [ABSTRACT FROM AUTHOR]

Effectively utilizing low-grade thermal energy is a promising approach to mitigating greenhouse gas emissions while reducing the burden on centralized power grids. Current thermoacoustic heat pumps and power generators face challenges such as high onset temperature differentials and low performance. This paper addresses these challenges by introducing a gas-liquid resonator into a thermoacoustic combined heat and power systems to recover low-grade thermal energy in residential buildings. Through Sage modeling and calculations, the internal characteristics of the proposed system and its output performance under different operating conditions are explored. At a heating temperature of 350 °C, the system can generate 6.4 kW of output thermal power, 0.9 kW of electricity, and overall exergy efficiency is 79.3 %. Combining neural network models with case studies conducted in Spain and Finland, the system can annually save 5.6 MWh and 20.7 MWh in fuel energy, reduce emissions of 1374 kg and 5180 kg of carbon dioxide, and save a total cost of €611 and €2324, respectively. Furthermore, comparisons with other emerging micro-CHP systems highlight the efficiency of the proposed system. These results indicate the high potential of thermoacoustic combined heat and power systems in recovering low-grade thermal energy and achieving energy savings and emission reductions. • Introducing an innovative thermoacoustic micro-CHP system. • Achieving an exergy efficiency of 79.3 % under standard operating conditions. • One system can annually reduce household CO 2 emissions by 5180 kg. • Ideal for households, balancing lower output levels for both heat and electricity. [ABSTRACT FROM AUTHOR]

The purpose of this paper is to model electricity consumption using Artificial Neural Networks (ANN). Total electricity consumption and consumption generated by households (HH) were modeled. The input variables of the ANN were based on nighttime light images from VIIRS DNB. Studies conducted thus far have covered mainly linear models. Most of case studies focused on single countries or groups of countries with only few focusing on the sub-national scale. This paper is pioneering in covering an area of Poland (Central Europe) at NUTS-2 level. The use of ANN enabled the modeling of the non-linear relations associated with the complex structure of electricity demand. Satellite data were collected for the period 2013–2016, and included images with improved quality (inter alia higher resolution), compared to the DMSP/OLS program. As images are available from April 2012 onwards, it is only recently that their number has become sufficient for ANN learning. The images were used to create models of multilayer perceptrons. The results achieved by ANN were compared with the results obtained using linear regressions. Studies have confirmed that electricity consumption can be determined with higher precision by the ANN method. • Innovative modeling of electricity consumption by artificial neural networks based on nighttime light images from VIIRS DNB. • A new way of preparing the input data of the model. • Total electricity consumption and electricity consumed by households were modeled. • The first electricity consumption analysis covered regions of Poland (Central Europe) at NUTS-2 level. [ABSTRACT FROM AUTHOR]

Abstract Without an intervening wall, the direct contact evaporator (DCE) has been already technically proven to improve the overall thermal efficiency of organic Rankine cycle (ORC) used to recover low-grade heat sources and transform them into power. In the estimation of volumetric heat transfer coefficient (VHTC) which is assumed to vary with flow rate, noises signals caused by various unstable factors (e.g., measurement errors) often corrupt the time series of VHTC. For forecasting the heat transfer performance of DCE in ORC more accurately, this paper proposes a novel approach (refers as EMD-RBF-NN), which combines multi-input radial basis function (RBF) neural network (NN) and empirical mode decomposition (EMD) method. Specifically, the original VHTC time series is firstly decomposed by EMD method that is fully data-driven. Then, the proposed method models the resultant decomposition series with flow rates of two fluids (dispersed and continuous phases) and VHTC by using RBF neural network. This simple technique was illustrated by using the ORC direct contact evaporator (ORC-DCE) and data processing system. Via using the experimental datasets of ORC-DCE, this paper demonstrates that the proposed EMD-RBF-NN model that associates flow rates of two phases with VHTC improves the forecasting accuracy of VHTC noticeably comparing with existing models. Highlights • This paper proposes a novel approach for forecasting the heat transfer performance. • This technique is illustrated using direct-contact evaporator and data processing system. • The proposed EMD-RBF-NN model that associates flow rates of two phases and VHTC. [ABSTRACT FROM AUTHOR]

Abstract In this paper, a novel T-S model with nonlinear consequent parts is introduced for the variable speed, variable pitch wind turbine. Because there is an inherent uncertainty in wind speed measurement, a fuzzy observer is proposed to estimate the effective wind speed, acting on the turbine's blades. Then, a robust H ∞ observer based fuzzy controller is designed to control the turbine using the estimated wind speed. Also, two artificial neural networks are used to accurately model the aerodynamic curves. In contrast to traditional controllers, which have different control schemes for different working regions, in this paper, only one controller is used for all operating regions of the wind turbine. As the main goal of a wind turbine is to maximize energy production and minimize mechanical loads concurrently, in addition to rotor dynamics, blade and tower dynamics are taken into account. To show the effectiveness of the proposed controller, simulations are performed on a 5 MW wind turbine simulator, in different wind profiles. Results show that compared with standard baseline controller, whilst power generation is improved, mechanical loads are reduced considerably. Highlights • T-S model with nonlinear consequent part is employed for both controller and observer design. • A robust fuzzy observer for wind speed estimation is introduced. • The separation principle of the proposed observer and controller is proved, despite that the premise variable of the T-S model is unmeasured. • In contrast to traditional controllers, which have different control schemes for different working regions, in this paper, only one controller is used for all operational regions of wind turbine. • Power generation maximization and mechanical load suppression are taking into account concurrently. [ABSTRACT FROM AUTHOR]

Understanding carbon emissions across income levels is vital to limiting global warming to below 1.5 °C and 2 °C, as the Paris Agreement outlines. Unequal emission patterns can affect global remaining carbon budgets, necessitating differentiated strategies. This paper introduces a unique perspective to forecasting carbon emissions and determining the remaining carbon budget relative to income levels and associated socioeconomic factors. Leveraging Pearson correlation analysis and Bayesian-optimised Artificial Neural Network methods, the study examines emissions and carbon budgets of thirty countries, representing 82 % of global carbon emissions in 2022, under a business-as-usual scenario across various income levels. Findings reveal distinct patterns, with high-income countries experiencing declining emissions while middle- and low-income countries anticipate increases. High-income countries are projected to deplete their 2.0 °C budgets later than middle- and low-income countries, with an estimated depletion around 2052 and 2046–2048, respectively. Proposed mitigation and adaptation strategies tailored to income levels include investment in advanced technologies for high-income countries, promoting energy efficiency and renewable energy transition for middle-income countries, and expanding clean energy access for low-income countries. This study advocates the importance of integrating income levels and socioeconomic factors into climate policymaking, offering insights for policymakers, environmentalists, and businesses to tailor interventions and promote equitable climate action. [Display omitted] • Bayesian-optimised ANN methods are used for forecasting across income level. • Forecasted carbon emissions are evaluated against Paris Agreement targets. • High-income nations' future emissions decline; lower-income nations' emissions rise. • Lower-income nations need immediate action due to shorter carbon budget periods. • Mitigation and adaptation strategies are proposed for various income levels. [ABSTRACT FROM AUTHOR]

This paper proposes an innovative algorithm for predicting short-term electrical maximum demand by using historical demand data. The ability to recognize in peak demand pattern for commercial or industrial customers would propose numerous direct and indirect benefits to the customers and utility providers in terms of demand reduction, cost control, and system stability. Prior works in electrical maximum demand forecasting have been mainly focused on seasonal effects, which is not a feasible approach for industrial manufacturing facilities in short-term load forecasting. The proposed algorithm, denoted as the Adaptive Rate of Change (ARC), determines the logarithmic rate-of-change in load profile prior to a peak by postulating the demand curve as a stochastic, mean-reverting process. The rationale behind this analysis, is that the energy efficient program requires not only demand estimation but also to warn the user of imminent maximum peak occurrence. This paper analyzes demand trend data and incorporates stochastic model and mean reverting half-life to develop an electrical maximum demand forecasting algorithm, which is statistically evaluated by cross-table and F-score for three different manufacturing facilities. The aggregate results show an overall accuracy of 0.91 and a F-score of 0.43, which indicates that the algorithm is effective predicting peak demand in predicting peak demand. [ABSTRACT FROM AUTHOR]

This paper analyzes the energy consumption of buildings considering the uncertainty of structural and environmental parameters. To this end, two-point estimate method (2PEM) is used to model the uncertainties. EnergyPlus software is used in this paper to evaluate the energy consumption, the thermal comfort, and the energy cost of a building. We examine the proposed method in a retail building to show the effectiveness of the method in comparison with the Monte-Carlo Simulation and deterministic methods. The results show that the 2PEM method although may cause a bit accuracy loss, it can considerably reduce the simulation time by more than 97%. In addition, a sensitivity analysis is effectuated in this paper to investigate the impacts of different climate zones on the results of energy consumption analysis. [ABSTRACT FROM AUTHOR]

In recent years, wind power has played a significant role in energy generation of micro-grids (MGs). However, randomness nature of wind speed leading to uncertainty in wind power forecast, imposes some problems such as overestimating wind power on optimized scheduling of MG. In this paper, we propose an adaptive probabilistic concept of confidence interval (APCCI) to address these problems. The main purpose of the proposed APCCI is to modify the risk we endure to schedule wind power with other distributed energy resources (DERs) in order to degrade the unnecessary rigors and upgrade the other ones. The forecasting method which is used in this paper is artificial neural network (ANN). In order to increase the accuracy of forecasting, wavelet decomposition (WD) is applied to the wind power time series then the results are sent to ANN. After that, dependable levels for the predicted wind power based on APCCI are obtained. An energy storage system (ESS) is utilized not only to decrease the impact of forecasting errors on the MG but also to increase the flexibility of the planning. A comprehensive formulation with operational constraints is employed to model the optimization problem. An economic dispatch based non-dominated sorting genetic algorithm II (EDNSGA-II) is proposed and applied to solve the multi-objective optimization problem. The optimization algorithm produces some alternatives which consist of different combination of objectives (cost and emission). Techniques for order preference by similarity to an ideal solution (TOPSIS) method is utilized to make a compromised decision between the alternatives. Eventually, the proposed algorithm is applied to a typical MG which consists of micro turbine (MT), fuel cell (FC), photo voltaic (PV), wind turbine (WT) and energy storage system (ESS). Evaluation of the results show that the proposed APCCI works well and can adapt the level of confidence interval in various situations. Moreover, the results confirm the superiority of WNN over ANN. The results also show that the proposed EDNSGA-II is more efficient in comparison with the well-known NSGA-II. [ABSTRACT FROM AUTHOR]

The configuration and development of energy management strategies (EMSs) are generating considerable interest regarding vehicles due to the rapid blossoming momentum of electric vehicles. Battery state of charge is one of the main characterization parameters for evaluating EMSs, so the practical advancement of the range is critical to the evolution of electric vehicles. A novel electric-hydraulic hybrid electric vehicle (EHHEV) is investigated in this paper, which has the characteristics of various working modes and multi-energy sources. According to the hybrid system's energy flow, a rule-based control strategy is established, and the superiority of EHHEV in energy management is verified by steady-state simulation. Further, this paper combines Q-learning with deep neural networks to construct a double deep Q-network (DDQN)-guided EMS to solve traditional control strategy and reinforcement learning issues. After appropriate hyperparameters setting and batch training, the EMS can make EHHEV realize the optimal switching among working modes. Experimental results showcase that the EMS can significantly enhance vehicle economy. This is the first of its kind to apply the DDQN to developing EMS for EHHEV. • Novel electric-hydraulic hybrid electric vehicles can improve energy management performance. • Double deep Q-network algorithm has a positive effect on vehicle working mode switching. • DDQN-guided energy management strategy is proposed to optimize the energy consumption rate. • Off-line training and adaptability verification are applied to prove the applicability of the EMS. • EHHEV realizes the optimal working mode switching under the interference of DDQN-guided EMS. [ABSTRACT FROM AUTHOR]

To solve the problems of the current optimization methods for solar segmented thermoelectric generator performance based on numerical methods, this paper applied deep neural networks to optimize the device geometry for improved thermo-mechanical performance. The motivation for using the deep neural network is to overcome the lengthy computational time and very high computational energy required by the traditional numerical method in optimizing the segmented thermoelectric generator performance. The numerical model is built using ANSYS software and the effects of temperature dependency in the 4 thermoelectric materials are considered to ensure result accuracy. Furthermore, 16 possible geometry parameters which were previously not considered, encompassing the individual and combined segment's heights and cross-sectional areas are optimized to find which set of parameters are the best in maximizing the device performance. The deep neural network is a regressive multilayer perceptron with network hyperparameters comprising 2 hidden layers with 5 neurons per layer. The training process is governed by the Levenberg-Marquardt standard backpropagation algorithm to minimize the mean squared error and maximize the regression correlation between the neural network forecasted outputs and the numerical-generated dataset. The most significant contribution of the proposed deep neural network is that it was able to quickly and accurately forecast the device performance in just 10 s, which was 2880 times faster than the conventional numerical-based optimization approach. Additionally, the optimized device had a maximum efficiency of 18%, which was 78% higher than that of the unoptimized device. Also, the thermal stress of the optimized device was 73% less than that of the unoptimized device design, indicating an extension in the device mechanical reliability and service lifetime. The results reported in this paper will accelerate the ease at which efficient, long-lasting segmented thermoelectric generators are manufactured by harnessing the power of artificial intelligence. [Display omitted] • AI driven geometry optimization of STEG thermo-mechanical performance using DNN. • 3D FEM considers 16 STEG geometry parameters that were previously neglected. • AI fixes shortcomings of conventional FEM method in optimizing STEG performance. • DNN forecasted device performance in just 10s, 2880 times faster than FEM. • Optimized STEG is 78% more efficient than conventional STEG, reduces failure by 73%. [ABSTRACT FROM AUTHOR]

Swarm-based algorithms are widely accepted in different fields of engineering as they have proven themselves effective in solving real-world optimization problems. According to the "No-Free-Lunch" theorem, there is no such algorithm that can perform well in all optimization problems. This paper proposes a novel swarm-based optimization algorithm, influencer buddy optimization (IBO), that mimics human behavior in social environments. This algorithm exploits the fact that humans are influenced by a group of individuals rather than a single person. However, every individual can select a buddy from this group based on their preferences. The mathematical representation of this social human behavior is shown in this paper. The algorithm is tested on twenty-one standard benchmark functions, and twenty-four noiseless black-box optimization benchmarking (BBOB) functions to validate its effectiveness over fourteen optimization algorithms. Further, two hybrid models are proposed by combining IBO with an artificial neural network (ANN) and recurrent neural network (RNN) to solve electricity load and price forecasting problems. The results of hybrid models are compared with nine recent and popular optimization algorithms. The simulation result of hybrid models shows that ANN-IBO and RNN-IBO generate the least forecasting error in their respective classes of network architectures. • Proposed a novel swarm-based optimization algorithm named IBO. • IBO is tested with multiple benchmark functions in different dimensions. • IBO is combined with ANN and RNN for solving electricity prediction problems. • Two electricity datasets are used to test the efficacy of the proposed model. [ABSTRACT FROM AUTHOR]

Thermostatically controlled loads (TCLs) are viewed as a promising demand-side resource for smoothing out the intermittency of renewable energy. However, it is difficult to dispatch these dispersed flexibility resources to serve the microgrid without establishing the aggregation model of TCLs. Motivated by this need, this paper presents a distributed and real-time economic dispatch strategy based on the alternating direction multiplier method, that allows TCLs and distributed generations to work in coordination to maintain the power balance of islanded microgrids under fluctuating renewable energy sources, while considering the occupant comfort and the fair utilization. A semi-online parameter identification method that does not involve occupant privacy is proposed for obtaining parameters of inverter-based TCLs. This method is based on recursive least squares and designed with a changing forgetting factor to avoid data saturation. Also, a virtual battery model considering parameter heterogeneity is established for TCLs, and the scheduling capability of the TCLs cluster under different ambient temperatures is determined using deep neural networks. Besides, this paper introduces a fast distributed consensus algorithm without changing the communication structure to improve the convergence speed of the distributed algorithm. The effectiveness of the proposed strategy is validated on an islanded microgrid with 100 TCLs. • Estimating parameters of TCLs by a non-invasive method to protect privacy. • Evaluating the dispatch capability by a deep neural network. • Distributed ADMM solves the real-time EDP of microgrids in parallel. • Guiding TCLs to participate in power scheduling with a consistent utilization. [ABSTRACT FROM AUTHOR]

This paper aims to provide a thorough guide on how to analyze complex energy demand forecasting models with Shapley Additive exPlanations (SHAP) in order to build trust in their predictions and understand the model and SHAP limitations based on selected real-world use case. There are only few examples described in the literature in energy industry and they present very basic usage. This study fills the gap for the class of energy (heat, electric, gas) demand predicting models by showing step by step, top-down analysis of state-of-the-art, deep neural network model predicting total heat demand (hot water and room heating) in Warsaw District Heating Network – the largest district heating network in EU. The paper shows how SHAP can be successfully used on demand forecasting models to provide practical, easily interpretable insights on inner workings of these models which can be used to assess their reliability and plan further development. • Predictions of deep neural network models forecasting energy demand can be explained. • SHAP can be successfully used to analyze deep ANN with multiple inputs (>100). • Step by step guide on how to analyze energy demand model with SHAP is provided. • Relationships identified by ANN forecasting heat demand agree with expert knowledge. • Trust in predictions of black-box models can be build by analyzing with SHAP. [ABSTRACT FROM AUTHOR]

Recently security issues like investment and financing in China's power industry have become increasingly prominent, bringing serious challenges to the financial security of the domestic power industry. Thus, it deserves to develop financial safety evaluation towards the Chinese power industry and is of practical significance. In this paper, the GA (genetic algorithm) is used to optimize the connection weights and thresholds of the traditional BPNN (back propagation neural network) so the new model of BPNN based on GA is established, hereinafter referred to as GA-BPNN (back propagation neural network based on genetic algorithm). Then, an empirical example of the electric power industry in China during the period 2003–2010 was selected to verify the proposed algorithm. By comparison with three other algorithms, the results indicate the model can be applied to evaluate the financial security of China's power industry effectively. Then index values of the financial security of China's power industry in 2011 were obtained according to the tested prediction model and the comprehensive safety scores and grades are calculated by the weighted algorithm. Finally, we analyzed the reasons and throw out suggestions based on the results. The work of this paper will provide a reference for the financial security evaluation of the energy industry in the future. [ABSTRACT FROM AUTHOR]

The knowledge of the available power provided by the driving wheel of agricultural tractors is required to gain a correct insight into the energy management of agricultural tractors. The design of the tractors is pivotal on the maximization of the traction efficiency and simultaneous minimization of energy dissipation. This paper spearheads the synthesis of the power provided by the agricultural tractors as affected by wheel load, slip and speed by use of the potential of a soil bin facility and a single-wheel test rig. The hybridized artificial neural network-genetic algorithm method was adopted to model the provided power of the driving wheel under the effect of the aforementioned tire parameters. The common drawback of the back-propagation algorithm known as the low speed of convergence and the possibility of being trapped in a local minimum was solved by the use of genetic algorithm. The mean square error equal to 0.02242 was obtained as the most optimal artificial neural network-genetic algorithm configuration using Levenberg–Marquardt training algorithm. Therefore, a 3-9-1 feed-forward with back propagation learning algorithm was selected as the modeling structure. The computed coefficient of determination for the training and test phases of the best artificial neural network-genetic algorithm model was obtained at 0.9696 and 0.9672, respectively. The present study spearheads the required power estimation for the driving wheels of off-road vehicles while the experimental test conduction in a controlled soil bin facility using single-wheel tester and adoption of soft computing tools are of the highlights and added values of the paper. [ABSTRACT FROM AUTHOR]

Leaks in fluidized-bed boilers are typically characterized by slow escalation. Early detection and prediction of such faults is an important task that has not been solved in practice yet. The paper reports a series of research and development works related to achieving early detection and prediction of leaks in fluidized-bed boilers using ANN (artificial neural networks). The obtained results were used in pilot implementation of a diagnostics and prediction system covering six blocks of a professional power plant. The diagnostics and prediction task is divided into two stages: early fault detection by virtual sensors and leak isolation using classifiers of fault state. Models of process variables were created by employing a novel two-stage structure of ANN. The resulting efficiency of leak detection is presented. Also provided is an example of 12 faults of a fluidized-bed boiler, achieving detection of 11 faults with at least two days advance prediction of a boiler shutdown. These results are compared with detections obtained by the authors previously with the use of neuro-fuzzy models. Then, the paper reports the ability to distinguish between three classes of leaks by the developed classifier of the fault state. Further possible improvements of this fault classification system are discussed. [ABSTRACT FROM AUTHOR]

Energy dissipation control has long been synthesized addressing the trafficking of wheeled vehicles. Wheel-obstacle collision has attracted the studies more on ride comfort, stability, maneuvering, and suspension purposes. This paper communicates, for the first time, the energy dissipation analysis through tire-obstacle collision that frequently occurs for the wheeled vehicles particularly those of off-road vehicles. To this aim, a soil bin facility equipped with a single wheel-tester is employed considering input parameters of wheel load, speed, slippage, and obstacle height each at three different levels. In the next step, the potential of classic artificial neural networks was appraised against support vector regression with the two kernels of radial basis function and polynomial function. On account of performance metrics, it was revealed that radial basis function based support vector regression is outperforming the other tested methods for the prediction of dissipated energy through tire-obstacle collision dynamics. The details are documented in the paper. [ABSTRACT FROM AUTHOR]

Lithium-ion batteries have emerged as an energy storage solution for electrified vehicles. A Battery Management System (BMS) is critical for efficient and reliable system operation, in which State of Charge (SoC) estimation and State of Health (SoH) monitoring are of major importance to ensure optimal energy management in battery vehicles for increased autonomy and battery life. This paper presents a neural network with Elman architecture trained for a lithium-ion cell, aiming at SoC and SoH estimation. The multi-objective optimization approach based on the particle swarm algorithm is used for the training in order to lower the root mean square error in calculating the SoC and SoH. For such purposes, the neural network characteristics are optimized, such as the number of hidden layers, the number of neurons in each hidden layer, the activation functions, the bias value, and the weights of the inputs and outputs. The best trade-off solution has an error of 2.56% in the average SoC estimate and 0.003% in the average SoH estimate. [Display omitted] • A neural network to estimate the SoC and SoH of a battery cell is developed. • Multi-objective optimization of the Elman Neural Network using PSO. • Development of an algorithm for assembling battery packs. • Mean SoC estimate for battery cell with 2.56% error and 0.003% error in SoH estimate. • 1.26% difference in SoC estimate when compared with a commercial vehicle. [ABSTRACT FROM AUTHOR]

Wind, as a fluid, has continuity in both space and time. Coupling spatial and temporal information to build prediction models can improve wind speed prediction accuracy. This paper proposes a method that predicts wind speed at multiple locations using both spatial and temporal data. Three deep learning models are introduced: Convolutional Residual Spatial-Temporal Long Short-Term Memory neural network (CoReSTL), Convolutional Spatial-Temporal-3D neural network (CoST-3), and Convolutional Spatial-Temporal Long Short-Term Memory neural network (CoST-L). These models combine Convolutional Long Short-Term Memory (ConvLSTM), Residual Network (ResNet), and 1 × 1 3D convolution to extract spatial and temporal correlations between multi-site wind speeds. The spatio-temporal prediction of wind fields under two terrains was carried out to screen out neural network models with higher accuracy. The results show that CoReSTL, CoST-3, and CoST-L all achieved better prediction results. However, the performance of the CoReSTL model was better than that of CoST-3 and CoST-L, with stronger applicability in complex real terrain. • A multi-sites wind speed prediction methodology that leverages both spatial and temporal correlations. • ConvLSTM, ResNet, and 3D convolution are combined as the prediction model. • The proposed prediction method works well with ideal 3D hills and real complex mountainous terrain. [ABSTRACT FROM AUTHOR]

Abstract: This paper focuses on NO x emission prediction and operating parameters optimization for coal-fired boilers. Support Vector Regression (SVR) model based on CGA (Conventional Genetic Algorithm) was proposed to model the relationship between the operating parameters and the concentration of NO x emission. Then CGA and two modified algorithms, the Quantum Genetic Algorithm (QGA) and SAGA (Simulated Annealing Genetic Algorithm), were employed to optimize the operating parameters of the coal-fired boiler to reduce NO x emission. The results showed that the proposed SVR model was more accurate than the widely used Artificial Neural Network (ANN) model when employed to predict the concentration of NO x emission. The mean relative error and correlation coefficient calculated by the proposed SVR model were 2.08% and 0.95, respectively. Among the three optimization algorithms implemented in this paper, the SAGA showed superiority to the other two algorithms considering the quality of solution within a given computing time. The SVR plus SAGA method was preferable to predict the concentration of NO x emission and further to optimize the operating parameters to achieve low NO x emission for coal-fired boilers. [Copyright &y& Elsevier]

Abstract: For accurate electricity demand forecasting, this paper proposes a novel approach, MFES, that combines a multi-output FFNN (feedforward neural network) with EMD (empirical mode decomposition)-based signal filtering and seasonal adjustment. In electricity demand forecasting, noise signals, caused by various unstable factors, often corrupt demand series. To reduce these noise signals, MFES first uses an EMD-based signal filtering method which is fully data-driven. Secondly, MFES removes the seasonal component from the denoised demand series and models the resultant series using FFNN model with a multi-output strategy. This multi-output strategy can overcome the limitations of common multi-step-ahead forecasting approaches, including error amplification and the neglect of dependency between inputs and outputs. At last, MFES obtains the final prediction by restoring the season indexes back to the FFNN forecasts. Using the half-hour electricity demand series of New South Wales in Australia, this paper demonstrates that the proposed MFES model improves the forecasting accuracy noticeably comparing with existing models. [Copyright &y& Elsevier]

Model predictive control (MPC), although considered a high-potential control approach, usually requires considerable effort for model-creation and parametrization. Moreover, many models can be too computationally intensive for control applications. Distributed model predictive control (DMPC) is a promising approach that avoids the construction of a complex model of the total system and thus facilitates modeling and supports the use of exact simulation models. DMPC divides the optimization problem into sub-problems, with the advantage that pre-fabricated simulation models from standard libraries, models provided by component manufacturers or purely data-driven models can be used. Moreover, each optimization problem, considered for its own, becomes smaller and, in whole, can be solved faster compared to an integrated system model. The distributed optimizations must be coordinated to achieve near-global-optimum performance. This coordination requires a suitable scheme, which, in many publications, is based on iterative data exchange between the subsystems. In previous works, we developed a non-iterative algorithm based on the exchange of lookup tables. In this paper, we compare and benchmark the previously developed approach against a second iterative approach to show advantages and limitations of both algorithms. We apply both approaches to the Modelica simulation model air-handling unit while using artificial neural network models and a non-linear solver for the DMPC algorithms. We make simplifying assumptions to provide mathematical justification regarding the optimality of the approaches. Judging from the indicators for control quality and the monetary operation costs in the case study, we conclude that the algorithms hold high potential for the application in generic building energy systems. • Distributed model predictive control (DMPC) algorithms facilitate the construction or training of models. • Study of a suitable schemes for data exchange between subsystems. • Simulation study with an air handling unit modelled in Modelica. • Usage of artificial neural network models for model-predictive control. • Conclusion on the high potential of DMPC in generic building energy systems. [ABSTRACT FROM AUTHOR]

Predictive control has gained increasing attention for its ability to reduce energy consumption and improve occupant comfort in buildings. The plug loads prediction is a key component for the predictive building controls, as plug loads is a major source of internal heat gains in buildings. This study proposed a novel method to apply the Long-Short-Term-Memory (LSTM) Network, a special form of Recurrent Neural Network, to predict plug loads. The occupant count and the time have been confirmed to drive the plug load profile and thus selected as the features for the plug load prediction. The LSTM network was trained and tested with ground truth occupant count data collected from a real office building in Berkeley, California. Results from the LSTM network markedly improve the prediction accuracy compared with traditional linear regression methods and the classical Artificial Neural Network. 95% of 1-h predictions from LSTM network are within ±1 kW of the actual plug loads, given the average plug loads during the office hour is 8.6 kW. The CV(RMSE) of the predicted plug load is 11% for the next hour, and 20% for the next 8 h. Lastly, we compared four prediction approaches with the office building we monitored: LSTM vs. ARIMA, with occupant counts vs. without occupant counts. It was found, the prediction error of the LSTM approach is around 4% less than the ARIMA approach. Using occupant counts as an exogenous input could further reduce the prediction error by 5%–6%. The findings of this paper could shed light on the plug load prediction for building control optimizations such as model-predictive control. • Accurate plug loads prediction is key to Model Predictive Control in buildings. • LSTM (Long Short Tem Memory) network is used to predict plug loads using occupant count data. • The prediction accuracy is 89% for the next hour, and 80% for the next 8 h. • LSTM is computational demanding but 4% more accurate than ARIMA. • Using occupant counts could reduce plug load prediction error by 5%–6%. [ABSTRACT FROM AUTHOR]

Abstract In this paper, a novel free piston expander-linear generator (FPE-LG) prototype has been developed for small scale organic Rankine cycle system. The effects of three key operating parameters, including intake pressure, operation frequency and external load resistance on piston dynamics, output characteristics of the linear generator, and system energy conversion efficiency are investigated. An artificial neural network (ANN) based prediction model is established after evaluating different learning rates, hidden layer neural numbers and train functions. The ANN model is also validated and tested using the experimental data with consideration of mean squared error and correlation coefficient. Finally, combined the genetic algorithm with the ANN model, a parametric optimization and performance prediction for maximum power output of the linear generator is conducted. The results show that the free piston assembly operates stably with good consistency. Higher intake pressure and external load resistance are beneficial for improving the piston dynamics and output characteristics of the linear generator while the optimal operation frequency corresponding to the maximum peak power output is more dependent on the coordinated variation of the operating parameters. The maximum system energy conversion efficiency can reach up to 28.81% with the intake pressure of 0.2 MPa, operation frequency of 1.5 Hz and external load resistance of 5 Ω. The proposed ANN model shows a strong learning ability and generalization performance. The correlation coefficients between the ANN predictions and experimental data obtained from the validation and test processes are all close to 1. The optimized peak power output can reach up to 100.47 W based on the proposed ANN model. The ANN based method can provide a useful guidance for the performance prediction and coordinated optimization with the least deviation and high accuracy. Highlights • Experimental study of a novel free piston expander-linear generator is presented. • Higher intake pressure and external load resistance are preferred. • The maximum system energy conversion efficiency can reach up to 28.81%. • An artificial neural network (ANN) based prediction model is established and validated. • The proposed ANN model is able to predict the motion and output characteristics accurately. [ABSTRACT FROM AUTHOR]

The paper analyses a model of the EU energy system by means of artificial neural networks. This model is based on the prediction of CO 2 emissions until 2050 taking into account the current Energy Policy of the EU. The results show that artificial neural networks model this system very well and that this model has the ability to predict the behaviour of CO 2 emissions. This will also enable timely response and correction of energy and economic strategy by changing the value of the relevant indicators in order to achieve the ambitious planned reductions of CO 2 emissions by 2050. These plans are specified in the Energy Roadmap 2050 document of the European Commission from 2012 and promote economically cost-effective scenarios that will adapt the European Union's economy to the needs of environmental protection and the reduction of energy consumption. Several structures of Artificial Neural Networks were analysed in order to select the best one for modelling large energy systems. It was determined that the model with the Cascade Forward Back Propagation structure with numerous specific indicators can model such energy systems and predict of CO 2 emissions with acceptable accuracy. • Analysis of European Union energy system using Artificial Neural Network is presented. • Based on key indicators, emissions of CO 2 in the year 2050 are estimated. • Influence of energy policy on energy trends is discussed. [ABSTRACT FROM AUTHOR]

Load forecasting is a challenging task in power markets that require attention in generating accurate and stable load to deal with planning and management strategies. In past few years, several intelligence-based models have been introduced for precise load forecast. Among them, artificial neural network (ANN) seems more effective and capable to handle the non-linear behavior of load and generates an accurate forecast. However, it suffers from overfitting problem thus reducing the accuracy of load forecasts. To overcome this problem, a hybrid methodology namely ANN-IEAMCGM-R, for short-term load forecast is proposed in this paper. ANN is integrated with an enhanced evolutionary algorithm (IEAMCGM-R) to find optimal network weights. This evolutionary algorithm is composed of improved environmental adaptation method with real parameters (IEAM-R) and our proposed Controlled Gaussian Mutation (CGM) method to bring greater diversity within the population resulting in a higher convergence of solutions. The electric load data from the New England Power Pool (NEPOOL, ISO New England) and Australian Energy Market Operator (New South Wales (NSW), Australia) have been used to illustrate the efficacy of the proposed hybrid methodology. Results show that the proposed hybrid methodology generates higher accuracy than other state-of-the-art algorithms. • Proposal of Controlled Gaussian Mutation for enhancing the diversity of IEAM-R. • Proposal of enhanced optimization algorithm, namely IEAMCGM-R. • A hybrid methodology is presented for load forecasting. • Experimental results show that ANN-IEAMCGM-R generates least forecasting error. • Statistical test results ensures the statistical signi_cance of forecasted load. [ABSTRACT FROM AUTHOR]

Abstract This paper presents a new integrated model predictive control (IMPC) method that combines power management and adaptive velocity control during vehicle-following scenarios in reality, for a plug-in hybrid electric vehicle (PHEV). Innovatively, the IMPC is able to plan the battery state of charge (SOC) and vehicular velocity trajectories, in order to improve fue economy and driving safety. To assess the performance of the IMPC, a comparison is performed with common charge-depleting and charge-sustaining (CDCS) and DP-based energy management strategies, where an improved full velocity difference model (IFVDM) is incorporated to simulate vehicle-following behavior. These solutions are examined using a real-world driving cycle. The results reveal an enormous potential of flexibly tuning the inter-vehicle distance to increase fuel economy. This is distinct from the rigid vehicle-following behavior of the IFVDM just for driving safety. Moreover, the proposed IMPC can ensure the battery charge depletion at the end of the trip. The quantitative results witness that the total cost of the IMPC with a preview horizon of 3s can be reduced by 17.9% and 36.1% for a 70 km city-bus route, compared to IFVDM-based DP and CDCS counterparts, respectively. In addition, the effect of the preview-horizon length on both fuel economy and computational time is examined. Highlights • An integrated model predictive power management is proposed. • SOC planning is built into the power management controller. • Driving safety and fuel economy are compromised. • An artificial neural network is established to forecast the leading vehicle speed. • The effect of preview-horizon length is evaluated. [ABSTRACT FROM AUTHOR]

Abstract Dynamic modelling plays a crucial role in the analysis of Organic Rankine Cycle (ORC) systems for waste heat recovery, which deal with a highly unsteady heat source. The efficiency of small scale ORCs (i.e. below 100 kW power output) is low (<10%). Therefore, it is essential to keep the performance of the system as stable as possible. To do so, it is helpful to be able to predict the dynamic behavior of the system, in order to perform a maximization of its performance over the time. In this paper, Feedforward, Recurrent and Long Short Term Memory networks have been compared in the prediction of the dynamics of a 20 kW ORC system. Feedforward Neural Network is the simplest among the architectures developed for machine learning. Recurrent and Long Short Term Memory networks have been proved accurate in the prediction of the performance of dynamic systems. This study demonstrates that the three architectures are capable of predicting the dynamic behavior of the ORC system with a good degree of accuracy. The Long Short Term Memory architecture resulted as the highest performing, in that it correctly predicts the dynamics of the system, showing an error prediction lower than 5% and 10% respectively for what concern the prediction 10 and 60 seconds ahead. Highlights • A small scale ORC system has been developed and tested under transient conditions. • A Neural Network approach has been used to predict the performance of the system. • Feed-forward, Recurrent and Long Short Term Memory networks have been compared. • The LSTM network shows the best performance, with an accuracy higher than 5%. [ABSTRACT FROM AUTHOR]

Abstract: This paper presents an application of Artificial Neural Networks (ANNs) to predict daily solar radiation. We look at the Multi-Layer Perceptron (MLP) network which is the most used of ANNs architectures. In previous studies, we have developed an ad-hoc time series preprocessing and optimized a MLP with endogenous inputs in order to forecast the solar radiation on a horizontal surface. We propose in this paper to study the contribution of exogenous meteorological data (multivariate method) as time series to our optimized MLP and compare with different forecasting methods: a naïve forecaster (persistence), ARIMA reference predictor, an ANN with preprocessing using only endogenous inputs (univariate method) and an ANN with preprocessing using endogenous and exogenous inputs. The use of exogenous data generates an nRMSE decrease between 0.5% and 1% for two stations during 2006 and 2007 (Corsica Island, France). The prediction results are also relevant for the concrete case of a tilted PV wall (1.175 kWp). The addition of endogenous and exogenous data allows a 1% decrease of the nRMSE over a 6 months-cloudy period for the power production. While the use of exogenous data shows an interest in winter, endogenous data as inputs on a preprocessed ANN seem sufficient in summer. [ABSTRACT FROM AUTHOR]

Abstract In order to achieve near-optimal fuel economy for plug-in hybrid electric vehicles (PHEVs) using the equivalent consumption minimum strategy (ECMS), it is necessary to dynamically tune the equivalent factor (EF). Unlike widely used model-based approaches, this paper proposes a data-driven ECMS that determines the EF using an artificial neural network (ANN). First, by comparing Pontryagin's Minimum Principle (PMP) with the ECMS, one can find that the EF is related to the co-state value of the PMP method. Then, an ANN is constructed with three accessible input variables, including the current demanded power, the ratio of the distance travelled to the total distance, and the battery State of Charge (SOC). The neural network is subsequently trained using real-world speed profiles. Simulations are performed considering different initial SOC values. The results reveal that the proposed data-driven ECMS demonstrates satisfactory fuel economy compared to global optimization methods like dynamic programming and PMP methods. The computational time of the proposed method relative to the duration of the entire trip indicates a great potential for the development of a time-conscious energy management strategy. Moreover, the impact of training sample size on the ANN performance is discussed. Highlights • A data-driven ECMS for PHEVs energy management is proposed. • The equivalent factor is determined based on a concise artificial neural network. • Real-world speed profiles are applied to train the neural network. • Impact of sample size on neural network performance is discussed. • Optimality and computational efficiency are compared with other typical methods. [ABSTRACT FROM AUTHOR]

Abstract: This paper presents a fully automatic system intended to detect leaks of dielectric fluid in underground high-pressure, fluid-filled (HPFF) cables. The system combines a number of artificial intelligence (AI) and data processing techniques to achieve high detection capabilities for various rates of leaks, including leaks as small as 15 l per hour. The system achieves this level of precision mainly thanks to a novel auto-tuning procedure, enabling learning of the Bayesian network – the decision-making component of the system – using simulated leaks of various rates. Significant new developments extending the capabilities of the original leak detection system described in and form the basis of this paper. Tests conducted on the real-life HPFF cable system in New York City are also discussed. [ABSTRACT FROM AUTHOR]

The coordinated control systems (CCS) in ultra-supercritical thermal power unit, like many other industrial systems, is a complex multivariable system with severe nonlinearity, strong multivariable coupling and uncertainties. In order to meet the requirements of operational stability, economy. etc in ultra-supercritical unit, it is necessary to establish its accurate mathematical model and further design the advanced controller. Against this background, a new fuzzy neural network modeling method is proposed in this paper. First of all, the incremental model is considered separately to improve the rationality of the local linear model structure. Then, the parameters in antecedent part is initialized by a kernel k -means++ algorithm, in which Xie-Beni index is used to optimize the number of fuzzy rules. Finally, supervised adaptive gradient descent algorithm and artificial immune particle swarm optimization algorithm work in stages to complete the training of the consequent part parameters. The proposed modeling method in this paper is applied to a 1000 MW unit in China and shows satisfactory accuracy. In the established model, the MSE of power output, main steam pressure and separator outlet steam temperature are 0.0099, 1.21E-4, 0.0023, respectively. Both numerical and graphical simulation results confirm the effectiveness of the presented fuzzy neural network in modeling. [ABSTRACT FROM AUTHOR]

Abstract: This paper uses linear and nonlinear statistical models, including artificial neural network (ANN) methods, to investigate the influence of the four economic factors, which are the national income (NI), population (POP), gross of domestic production (GDP), and consumer price index (CPI) on the electricity consumption in Taiwan and then to develop an economic forecasting model. Both methods agree that POP and NI influence electricity consumption the most, whereas GDP the least. The results of comparing the out-of-sample forecasting capabilities of the two methods indicate the following. (1) If given a large amount of historical data, the forecasts of ARMAX are better than the other linear models. (2) The linear model is weaker on foretelling peaks and bottoms regardless the amount of historical data. (3) The forecasting performance of ANN is higher than the other linear models based on two sets of historical data considered in the paper. This is probably due to the fact that the ANN model is capable of catching sophisticated nonlinear integrating effects through a learning process. To sum up, the ANN method is more appropriate than the linear method for developing a forecasting model of electricity consumption. Moreover, researchers can employ either ANN or linear model to extract the important economic factors of the electricity consumption in Taiwan. [Copyright &y& Elsevier]

This paper investigates the influence of incorporating spatiotemporal wind data on the performance of wind forecasting neural networks. While previous studies have shown that including spatial data enhances the accuracy of such models, limited research has explored the impact of different spatial and temporal scales of input wind data on the learnability of neural network models. In this study, convolutional neural networks (CNNs) are employed and trained using various scales of spatiotemporal wind data. The research demonstrates that using spatiotemporally correlated data from the surrounding area and past time steps for training a CNN favorably affects the predictive performance of the model. The study proposes correlation analyses, including autocorrelation and Pearson correlation analyses, to unveil the influence of spatiotemporal wind characteristics on the predictive performance of different CNN models. The spatiotemporal correlations and performances of CNN models are investigated in three regions: Korea, the USA, and the UK. The findings reveal that regions with smaller deviations of autocorrelation coefficients (ACC) are more favorable for CNNs to learn the regional and seasonal wind characteristics. Specifically, the regions of Korea, the USA, and the UK exhibit maximum standard deviations of ACCs of 0.100, 0.043, and 0.023, respectively. The CNNs wind prediction performances follow the reverse order of the regions: UK, USA, and Korea. This highlights the significant impact of regional and seasonal wind conditions on the performance of the prediction models. [Display omitted] • 3D CNN-based wind velocity prediction. • Estimation of CNN learnability through spatiotemporal wind correlation analysis. • Discovery of influence of local geometric and seasonal wind on CNN prediction. • The proposed correlation analysis can aid in selecting yaw-control wind farm sites. [ABSTRACT FROM AUTHOR]

The integrated energy system (IES) can improve energy efficiency and promote the use of renewable energy. It is of great significance to find the optimal alternative from many IES alternatives. However, most studies focused on the IES evaluation with complete information. Given this background, this paper proposes an evaluation method based on generalized regression neural network (GRNN) and probabilistic neural network (PNN) to address the issue of IES evaluation with incomplete information, which takes the running speed of the algorithm and the dynamic updating of the indexes into account. Firstly, the missing data is predicted via GRNN, and a complete evaluation matrix is obtained. Then, the evaluation indexes are clustered by affinity propagation algorithm, and the evaluation matrix of the cluster centers is determined, which can reduce the dimension of the initial matrix and improve the speed of the IES evaluation method. For the newly added evaluation indexes, the PNN is used to classify them, and the new cluster center is calculated, which can address the issue of the indexes updating. Based on the data of a hospital in Henan, case studies were carried out to evaluate the performance of the proposed method. The results show that the proposed method performs better than the existing IES evaluation method. • GRNN is used to address the IES evaluation problem with incomplete information. • The accuracy of GRNN is verified by comparing with common prediction methods. • It is the first work to deal with the issue of indexes updating in IES evaluation. • Compared with the existing evaluation method, the proposed method performs better. [ABSTRACT FROM AUTHOR]

This paper proposes a collection of novel deliverability prediction models for underground natural gas storage (UNGS) in salt caverns based on machine learning algorithms. Considering that the natural gas supply chain is characterized by imbalances between demand and supply on a timely basis, effective and fast models for predicting the deliverability of UNGS would not only be a valuable tool to various stakeholders but also, of great benefit in competitive natural gas marketplace. In this paper, a first step in applying machine learning algorithms to predict the deliverability of UNGS in salt caverns is proposed. To achieve this, the capability of three machine learning algorithms namely, artificial neural network (ANN), support vector machine (SVM), and Random Forest (RF) are examined. The predictive capabilities of these methods were investigated using different monthly field storage data samples for different years with varied data samples of 36 active UNGS in salt caverns in the United States. Experimental results reveal that the machine learning models developed in this study can serve as suitable tools for predicting the deliverability of UNGS in salt caverns with different performances. Overall result shows that RF model exhibits better prediction performance with varied data partitions over ANN and SVM models. • Deliverability prediction for underground natural gas storage in salt caverns. • Artificial Neural Network, Random Forest, and Support Vector Machines are examined. • Random Forest exhibits better prediction performance on different data samples. • Results can facilitate decisions for narrowing the gap between demand and supply. • Results can guide decisions for future design specification of gas storage. [ABSTRACT FROM AUTHOR]

State of charge (SOC) is one of the most critical parameters for indication of the remaining energy which is vital important for the safety and reliability of power system. In this paper, Levenberg-Marquardt (L-M) algorithm optimized multi-hidden-layer wavelet neural network (WNN) model and a series of novel intelligent SOC estimation methods using L-M based WNN are proposed. Particle swarm optimization (PSO) algorithm is used to optimize L-M based three-layer WNN (LMWNN) for SOC estimation problem. Furthermore, it is validated that L-M based multi-hidden-layer WNN (LMMWNN) has better performance than LMWNN. Basing the specific characteristic of SOC estimation, the LMMWNN method is optimized by combining piecewise-network method (PLMMWNN) and seven-point linear smoothing method (smoothed PLMMWNN). Under single driving cycle, such as the New European Driving Cycle(NEDC), the mean absolute error of PLMMWNN can be decreased to 0.6% and the maximum absolute error 5%. A comparison study of the series of WNN-based methods with BP neural network (BPNN) and extend Kalman filter (EKF) is conducted. The robustness evaluation, which is based on untrained driving cycles test, measurement noise test and piecewise training and batteries test, indicates the good performance on estimation accuracy, applicability and robustness of the proposed methods. [ABSTRACT FROM AUTHOR]

To solve the collaboration problem of multi-time scale economic dispatch and generation control in power system, i.e., long term time scale optimization, short term time scale optimization, and real-time control, a real-time economic generation dispatch and control (REG) framework, which as a unified time scale framework, is designed in this paper. With a relaxed operator employed into deep neural network (DNN), relaxed deep learning (RDL) is proposed for the REG framework, which towards an alternative to conventional generation control framework, which combines unit commitment (UC), economic dispatch (ED), automatic generation control (AGC), and generation command dispatch (GCD). Compared with 1200 combined conventional generation control algorithms in two simulations, i.e., IEEE 10-generator 39-bus New-England power system and Hainan 8-generator power grid (China), RDL obtains the optimal control performance with smaller frequency deviation, smaller area control error, smaller total cost, and smaller number of reverse regulation. Although the RDL needs a relative long computation time in the pre-training, the simulation results verify the effectiveness and feasibility of the proposed RDL for REG framework. [ABSTRACT FROM AUTHOR]

This paper proposes a novel hybrid technique for prediction of power generation in wind farm. Using this technique, the effective number of input parameters is determined by data correlation analysis. Then, the type of the effective number of parameters are selected from current parameters via neural network. The WT (wavelet transform) is used for filtering input data related to wind power; while, neural network ‘RBF (Radial basis function)’ is utilized as a preliminary predictor. Main predictor motor is combined three neural networks of MLP (Multilayer perceptron) with learning algorithms: BR (Bayesian regularization), RP(Resilient back propagation), and LM(Levenberg Marquardt). The heuristic technique ‘WIPSO (Weight improved particle swarm optimization)’ is used to improve the accuracy of predictions and escape from local minimum to optimize weights of neural networks. Input data is realistic data of wind farms in Southern Alberta, Canada for recent years. The data is a complete set of wind power and five meteorological characteristics including wind speed, wind direction, temperature, air pressure and humidity. Simulation results verify appropriate selection of the number and type of input parameters and heuristic algorithm application. Additionally, the absolute superiority of this technique is validated compared with the other methods. Prediction error is also reduced perceptibly. [ABSTRACT FROM AUTHOR]

In this paper, an intelligent control strategy for a grid connected hybrid energy generation system consisting of Photovoltaic (PV) panels, Fuel Cell (FC) stack and Battery Energy Storage System (BESS) is proposed. Firstly, the dynamic modeling of the electrical energy resources is carried out. Then, the local controllers of DC-DC power converters are designed to regulate the operating points of the energy generation units. An online-trained Elman Neural Network (ENN) based controller is developed to perform the tracking of the optimal operating point of the PV source. Moreover, a Takagi-Sugeno-Kang based fuzzy gain tuner is adopted for the adjustment of the PI parameters of the FC and BESS controllers. Besides, the Virtual Flux Oriented Control (VFOC) scheme based on neuro-fuzzy gain tuners is adopted to control the power flow between the hybrid system, the local loads and the utility grid. Furthermore, a centralized power supervisor is established to determine the levels of the power references for the local controllers according to the supply and demand conditions. The simulation results obtained in the Matlab/Simulink environment demonstrate the efficiency of the proposed control scheme. [ABSTRACT FROM AUTHOR]

This paper attempts to explore the evolvement rule of energy-saving and emission-reduction with changing economic growth rate based on a novel energy-saving and emission-reduction system with multi-variable constraints. Energy structure adjustment, economic growth rate, new energy technology progress and other constrained variables are introduced into the novel system with the aid of nonlinear dynamics. The dynamic behavior of the new system is discussed by means of nonlinear dynamics theory. The quantitative coefficients of the actual system are identified based on the genetic algorithm-back propagation neural network. A scenario study of the actual system is carried out, and some valuable conclusions are presented. Energy structure adjustment is driven more by economic considerations according to China’s national conditions. Energy structure adjustment could perfectly control energy intensity, without obvious potential negative impact on economic growth. New energy technology progress has a better effect on controlling energy intensity, and the more mature the technology is, the better the effect is. Energy intensity is sensitive to economic growth rate. There is an optimal interval of economic growth rate, and energy intensity could be effectively controlled by adjusting corresponding constraint condition in this interval. [ABSTRACT FROM AUTHOR]