Energy Conversion and Economics最新文献

This article proposes a novel online reinforcement learning-based linear quadratic regulator for the three-level neutral-point clamped DC/AC voltage source inverter. The proposed controller employs online updated fixed-weight recurrent neural network (NN) and policy iteration to dynamically adjust the optimal control gains based on real-time measurements without any knowledge of the system model or offline pre-training. Moreover, it produces a constant switching frequency with low current harmonics. Compared to the existing control methods, it provides superior control performance, guaranteed control stability, and simplified NN design. Experimental results are presented to verify the effectiveness of the proposed control method.

Human activities have been driving massive greenhouse gas emissions, causing global warming, and triggering increasingly frequent extreme weather events that severely threaten the environment. Power generation is the leading contributor to anthropogenic emissions, making precise, real-time measurement and monitoring of power plant carbon emissions crucial in reducing climate change. This study uses a new sophisticated pipeline that combines tropospheric monitoring instrument satellite data, power plant attributes, and advanced artificial intelligence algorithms to build a predictive carbon emission model. The approach utilizes multimodal data processing, encoding, and model optimisation. Experimental results confirm that this pipeline can automatically extract and utilize vast amounts of relevant data, thereby enabling the artificial intelligence model to accurately predict power plant carbon emissions and providing a vital tool for reducing global warming.

Load forecasting with distributed energy resources (DERs) behind-the-meter is more challenging owing to transformed data patterns. Traditional forecasting method which is only based on unmasked-load could not suit the present limited masked-load. To bridge the divergence between unmasked-load and masked-load, this article proposes a masked-load forecasting (MLF) method based on transfer learning technique and Bayesian optimization, which is Maximum Mean Discrepancy-Neural Network with Bayesian optimization (MMD-NN^b). At first, common feature vectors between unmasked-load and masked-load are extracted and an outcome predictor could be established based on feature vectors from historical unmasked-load. The feature vectors from masked-load could therefore accommodate to the outcome predictor, and the masked-load could be forecast. Owing to the excessive hyperparameters involved in training, Bayesian optimization is adopted for hyperparameters fine-tuning. MMD-NN^b was tested and compared with four related models. The improvements from MMD-NN^b were observed in all comparison scenarios. Also, MMD-NN^b was proved to have high resilience to the different DERs and not requiring additional DERs-data.

The integration of seasonal loads, such as cereal baking and aquatic-product processing loads, often leads to significant voltage deviations and severe peak loads of the distribution system during specific periods, resulting in increased network losses. Traditional approaches for reducing network losses are becoming less effective and cost-efficient due to the spatiotemporally uneven distribution characteristics of seasonal loads. To address this issue, this study proposes an optimisation model that collaboratively integrates mobile energy storage, switching capacitors, and tie lines to minimise annual network losses in special planning scenarios affected by seasonal loads. The deployment strategies of multiple reinforcement methods are thoroughly analysed, greatly enhancing the explainability and feasibility of the collaborative deployment model. Then, the proposed model is reformulated to a mixed-integer linear programming model using the inscribed regular dodecagon approximation approach, thereby making it trackable for state-of-the-art solvers. To illustrate the effectiveness of the model, case studies are conducted on a unique 55-bus distribution system located in East China, which contains feeders with substantial seasonal variation aquaculture loads and with general loads. The effectiveness of multiple reinforcement methods is thoroughly analysed through detailed numerical results. Furthermore, a sensitivity analysis of the investment budget is conducted.

Lithium-ion battery state-of-health (SOH) monitoring is essential for maintaining the safety and reliability of electric vehicles and efficiency of energy storage systems. When the SOH of lithium-ion batteries reaches the end-of-life threshold, replacement and maintenance are required to avoid fire and explosion hazards. This paper provides a comprehensive literature review of lithium-ion battery SOH estimation methods at the cell, module, and pack levels. Analysis and summary of the SOH definition based on the resistance, capacity, and energy indices are presented at each battery hierarchy level. A Comparison of SOH indices in terms of modelling complexity, required measurement time, and accuracy is provided. To the best of knowledge, a comprehensive classification of SOH estimation methods at different battery hierarchy levels is presented for the first time in this review. In addition, SOH estimation methods are further classified based on the applied methodologies, including direct measurement, model-based methods, data-driven methods, and hybrid model-data methods. Advantages and disadvantages of SOH estimation methods are summarized and compared across different battery hierarchy levels. A detailed summary of typical SOH estimation methods is presented along with the battery topology, operating conditions, and performance. The challenges and research prospects of lithium-ion battery SOH estimation are discussed from the cell to pack levels.

Greenhouses need to supply CO₂ to crops while simultaneously emitting CO₂. To effectively harness the dual functionality of greenhouses as a carbon source and carbon consumer, this work incorporates carbon capture and emissions trading into a multi-energy greenhouse (MEG), which is equipped with various power and heat sources such as photovoltaic (PV) panels and a combined heat and power (CHP) unit and proposes that the captured CO₂ should be used to feed crops on-site. A low-carbon economic operation method is proposed for the coordinated environment-energy-carbon management of the MEG, and it considers various factors, including the power purchase/carbon supply costs, carbon emissions trading income, temperature/humidity/light intensity and CO₂ concentration requirements for crops, and operational constraints of various energy/environmental regulation equipment. The proposed method is validated using a tomato MEG. The results highlight the significant economic and environmental benefits of introducing carbon capture, emissions trading, and utilisation into MEGs.

Accurately and efficiently estimating the carbon market risk is paramount for ensuring financial stability, promoting environmental sustainability, and facilitating informed decision-making. Although classical risk estimation methods are extensively utilized, the implicit pre-assumptions regarding distribution are predominantly contained and challenging to balance accuracy and computational efficiency. A quantum computing-based carbon market risk estimation framework is proposed to address this problem with the quantum conditional generative adversarial network-quantum amplitude estimation (QCGAN-QAE) algorithm. Specifically, quantum conditional generative adversarial network (QCGAN) is employed to simulate the future distribution of the generated return rate, whereas quantum amplitude estimation (QAE) is employed to measure the distribution. Moreover, the quantum circuit of the QCGAN improved by reordering the data interaction layer and data simulation layer is coupled with the introduction of the quantum fully connected layer. The binary search method is incorporated into the QAE to bolster the computational efficiency. The simulation results based on the European Union Emissions Trading System reveals that the proposed framework markedly enhances the efficiency and precision of Value-at-Risk and Conditional Value-at-Risk compared to original methods.

Renewable energy-based microgrids (MGs) strongly depend on the implementation of energy storage technologies to optimize their functionality. Traditionally, electrochemical batteries have been the predominant means of energy storage. However, technological advancements have led to the recognition of hydrogen as a promising solution to address the long-term energy requirements of microgrid systems. This study conducted a comprehensive literature review aimed at analysing and synthesizing the principal optimization and control methodologies employed in hydrogen-based microgrids within the context of building microgrid infrastructures. A comparative assessment was conducted to evaluate the merits and disadvantages of the different approaches. The optimization techniques for energy management are categorized based on their predictability, deployment feasibility, and computational complexity. In addition, the proposed ranking system facilitates an understanding of its suitability for diverse applications. This review encompasses deterministic, stochastic, and cutting-edge methodologies, such as machine learning-based approaches, and compares and discusses their respective merits. The key outcome of this research is the classification of various energy management strategy (EMS) methodologies for hydrogen-based MG, along with a mechanism to identify which methodologies will be suitable under what conditions. Finally, a detailed examination of the advantages and disadvantages of various strategies for controlling and optimizing hybrid microgrid systems with an emphasis on hydrogen utilization is provided.

Packetized energy encapsulates energy into modulated, routable, and trackable energy packets, enhancing the flexibility of managing distributed energy resources and expediting prosumers’ participation in transactive energy markets. In the context of packetized energy trading (PET), energy prosumers are naturally deemed as self-interested agents seeking to obtain their own benefits. To align with prosumers’ demand, supply, quality of service (QoS), and system-level social welfare, it is necessary to explore the design of prosumers’ bidding strategies and the market clearing methods, considering prosumers’ utility and the best demand response to markets. This study addresses challenges arising from prosumers’ selfishness and asymmetric preferences by proposing a PET-oriented iterative double auction (IDA-PET) design, where prosumers are allowed to iteratively change the bids before the auctioneer clears the market. Moreover, IDA-PET accommodates system capacity constraints, energy balance, and economic constraints, providing cooperative strategies for both prosumers and the auctioneer. To validate the effectiveness of IDA-PET, a novel and dedicated co-simulation platform based on the hierarchical engine for large-scale infrastructure co-simulation platform is developed and case studies are conducted within a residential microgrid. The simulation results demonstrate that IDA-PET can efficiently enhance the revenue of the auction market while meeting prosumers’ QoS requirements.

To optimally harness the adjustable capabilities of reactive power sources for voltage control, a dynamic partitioning method that uses reactive power flow tracking for branch cutting through Binary Particle Swarm Optimisation (BPSO) is proposed for Active Distribution Networks (ADNs). Initially, the limitations of existing Voltage/Var Sensitivity (VVS) calculation methods are analysed, leading to the proposition of a novel VVS calculation method capable of capturing variations in source-load timing characteristics. Subsequently, the fuzzification of the VVS matrix between nodes is used to derive the membership degree matrix. Next, based on the membership relationship between reactive power source nodes, these nodes are pre-partitioned, and the number of leading nodes and zones alongside are preliminarily determined. Then, the range of the branch to be cut is established, guided by the reactive power flow direction of the branch. Employing the zonal comprehensive coupling degree as the objective function of the BPSO facilitates the identification of optimal branch cutting points, thereby determining the partitioning outcome. Finally, a reactive power reserve check is executed to rectify any non-compliant zones. In this study, numerical simulations are conducted using the enhanced IEEE 33-node power system to demonstrate the efficacy of the proposed method.