Scenario Generation Research Articles

Hybrid stochastic and robust optimization of a hybrid system with fuel cell for building electrification using an improved arithmetic optimization algorithm

This paper proposes a hybrid stochastic-robust optimization framework for sizing a photovoltaic/tidal/fuel cell (PV/TDL/FC) system to meet an annual educational building demand based on hydrogen storage via unscented transformation (UT), and information gap decision theory-based risk-averse strategy (IGDT-RA). The hybrid framework integrates the strengths of UT for scenario generation and IGDT-RA (hybrid UT-IGDT-RA) for optimizing the system robustness and maximum uncertainty radius (MRU) of building energy demand and renewable resource generation. The deterministic model focuses on minimizing the cost of energy production over the project’s lifespan (CEPLS) and considers a reliability constraint defined as the demand shortage probability (DSHP). The study utilizes an improved arithmetic optimization algorithm (IAOA) to optimize component sizes and MRUs, incorporating a neighborhood search operator to enhance performance and prevent premature convergence. The deterministic findings revealed that the PV/TDL/FC system configuration offers the lowest CEPLS and the highest reliability level (lowest DSHP) compared to the hybrid PV/FC and TDL/FC configurations. Additionally, these results indicated that enhancing the reliability of the energy supply for the educational building entails higher CEPLS, particularly due to increased costs associated with hydrogen storage. The robust framework findings for the PV/TDL/FC system using IGDT-RA show that for an uncertainty budget of 21%, the MRUs for educational building demand and renewable generation are obtained at 10.34% and 2.65%, respectively, which are higher compared to other configurations. This indicates that the hybrid PV/TDL/FC system is more robust in handling worst-case scenario uncertainties. Furthermore, the hybrid UT-IGDT-RA outcomes found that the stochastic scenarios incorporated to simulate a range of uncertainties beyond the conventional IGDT-RA based-nominal scenario, and it provides a broader range of robust solutions, enabling operators to align strategies with their risk tolerance and improves system flexibility, and decision-making precision in the face of uncertainties.

Security-constrained stochastic optimal power flow analysis using optimally reduced scenarios for wind generation

Security-constrained stochastic optimal power flow analysis using optimally reduced scenarios for wind generation

Novel demand response-based stochastic optimisation in a grid-connected microgrid considering outlier detection and imputation

To address the challenges in renewable generation, the proposed work presents a risk-based stochastic optimisation approach. This method considers uncertainties by employing Monte Carlo simulations for scenario generation and scenario reduction employing K-means clustering. The unsupervised K-means clustering technique is particularly effective for outlier detection and imputation, ensuring more reliable data for decision-making. Additionally, the authors proposed a novel demand response programme called Dynamic Sustainability Pricing (DSP), designed to reduce both the environmental impact and the total cost of operating a microgrid. To evaluate the performance of the DSP, several indices such as the personalised sustainability index, the carbon intensity index, the renewable energy participation index, and the environmental impact index have been proposed. It is evident that the DSP has caused the PAR to decrease from 5.73 to 4.98 and that the load factor's percentage increase from before to after the DSP application has improved to 3.603%. The improvement in LF from 0.48 to 0.51 is the result of DSP employment. In addition, the TCPI was raised from 1 to 1.148, which helps consumers by lowering the overall consumer tariff and raising the profit index. The PSI increased from 3.3 to 3.069, indicating that RES's participation was maximized.

Capacity Planning Method for Wind–Solar–Thermal-Storage Bundled HVDC Sending System Considering Transient Overvoltage Constraints

High-voltage direct current (HVDC) sending systems have been the main means of renewable power cross-regional sharing and consumption. However, the transient overvoltage problems restrict the transmission capacity and renewable energy accommodation. The allocation of wind–solar–thermal storage capacity has become an important factor affecting the safety and stability of renewable energy sending. A capacity planning method is proposed for a wind–solar–thermal-storage bundled HVDC sending system considering transient overvoltage constraints. Firstly, based on quantile regression analysis and Gaussian mixture modeling, the typical scenario generation method is proposed to depict the uncertainty of renewable energy. Then, the transient overvoltage characteristics of the integrated HVDC transmission system are analyzed. The relationship between the power output of power sources and the system short-circuit capacity is derived. Meanwhile, the calculation method of the minimum short-circuit capacity of the HVDC system is proposed. Based on the calculation method, the transient overvoltage constraint corresponding to the voltage support strength is constructed. Finally, considering the transient overvoltage constraints, the capacity planning model of the wind–solar–thermal storage is established. The upper-layer model optimizes the configuration scheme of the wind–solar–thermal storage to minimize the total system cost. The lower-layer model optimizes the operation scheduling under the typical operation scenarios of renewable energy and delivery load. The optimal capacity planning scheme for the wind–solar–thermal storage is determined through the coordinated optimization of the two-layer model. The feasibility and effectiveness of the proposed method are verified through a case analysis. The results show that the proposed planning method can effectively maintain a higher short-circuit ratio and improve the voltage support strength under the premise of completing the sending plan.

A Survey of Scenario Generation for Automated Vehicle Testing and Validation

This survey explores the evolution of test scenario generation for autonomous vehicles (AVs), distinguishing between non-adaptive and adaptive scenario approaches. Non-adaptive scenarios, where dynamic objects follow predetermined scripts, provide repeatable and reliable tests but fail to capture the complexity and unpredictability of real-world traffic interactions. In contrast, adaptive scenarios, which adapt in real time to environmental changes, offer a more realistic simulation of traffic conditions, enabling the assessment of an AV system’s adaptability, safety, and robustness. The shift from non-adaptive to adaptive scenarios is increasingly emphasized in AV research, to better evaluate system performance in complex environments. However, generating adaptive scenario is more complex and faces challenges. These include the limited diversity in behaviors, low model interpretability, and high resource requirements. Future research should focus on enhancing the efficiency of adaptive scenario generation and developing comprehensive evaluation metrics to improve the realism and effectiveness of AV testing.

Systematic Analysis Based on the Optimal Design of Photovoltaic Inverter Circuit

As a clean and renewable energy source, solar energy is of great significance to reduce greenhouse gas and harmful gas emissions, stabilize energy supply, promote industrial innovation, and help achieve sustainable development and carbon neutrality goals. In this paper, the structure and working principle of photovoltaic cells are introduced, and the characteristics of solar photovoltaic cells and thin film photovoltaic cells are compared and analyzed. Then, the application scenarios of solar photovoltaic power generation and thin film photovoltaic power generation are compared, and the future development direction of photovoltaic cells is prospected. Finally, the structure and working mechanism of the inverter circuit are introduced, some problems in the current photovoltaic inverter circuit are summarized, and then some optimization design of photovoltaic inverter circuit based on the maximum power point tracking technology and leakage current related problems are analyzed, and suggestions on how to optimize the design of photovoltaic inverter circuit are given.

Hybrid Proximal Policy Optimization—Wasserstein Generative Adversarial Network Framework for Hosting Capacity Optimization in Renewable-Integrated Power Systems

The rapid integration of distributed energy resources (DERs) such as photovoltaics (PV), wind turbines, and energy storage systems has transformed modern power systems, with hosting capacity optimization emerging as a critical challenge. This paper presents a novel Hybrid Proximal Policy Optimization-Wasserstein Generative Adversarial Network (PPO-WGAN) framework designed to address the temporal-spatial complexities and uncertainties inherent in renewable-integrated distribution networks. The proposed method combines Proximal Policy Optimization (PPO) for sequential decision-making with Wasserstein Generative Adversarial Networks (WGAN) for high-quality scenario generation, enabling robust hosting capacity enhancement and operational efficiency. Simulation results demonstrate a hosting capacity improvement of up to 128.6% in high-penetration scenarios (90% renewable), with average operational cost reductions of 22%. Voltage deviations are minimized to within ±5% of nominal levels, while energy losses are reduced by 18%. Scenario quality, evaluated using the Wasserstein metric, achieved convergence with an average score of 0.95 after 80 iterations, highlighting the WGAN’s ability to generate realistic and diverse scenarios. This study advances the state of the art in distribution network optimization by integrating machine learning techniques with robust mathematical modeling. The PPO-WGAN framework enhances scalability, ensures grid stability, and promotes efficient renewable integration, providing a robust foundation for future applications in modern power systems.

Personalized Shared Control for Automated Vehicles Considering Driving Capability and Styles.

The shared control system has been a key technology framework and trend, with its advantages in overcoming the performance shortage of safety and comfort in automated vehicles. Understanding human drivers' driving capabilities and styles is the key to improving system performance, in particular, the acceptance by and adaption of shared control vehicles to human drivers. In this research, personalized shared control considering drivers' main human factors is proposed. A simulated scenario generation method for human factors was established. Drivers' driving capabilities were defined and evaluated to improve the rationality of the driving authority allocation. Drivers' driving styles were analyzed, characterized, and evaluated in a field test for the intention-aware personalized automated subsystem. A personalized shared control framework is proposed based on the driving capabilities and styles, and its evaluation criteria were established, including driving safety, comfort, and workload. The personalized shared control system was evaluated in a human-in-the-loop simulation platform and a field test based on an automated vehicle. The results show that the proposed system could achieve better performances in terms of different driving capabilities, styles, and complex scenarios than those only driven by human drivers or automated systems.

Robust planning of production networks at an automotive supplier

As global trends like individualisation continue to drive significant changes in industrial production within international networks, it has become increasingly crucial for companies to maintain competitiveness through the efficient utilisation of resources. The rising complexity in production networks originates from an increasing number of constraints due to company-specific requirements, coupled with expanding networks that broaden the solution space, ultimately leading to prolonged planning processes. Furthermore, current planning tasks are predominantly performed manually, as the extensive efforts required for data acquisition often render the use of solution algorithms infeasible due to incomplete or inaccurate data. Therefore, this study explores robust planning of production networks, employing Monte–Carlo simulation and clustering for scenario generation, stochastic modelling concepts to tackle the mathematical problem, and utilising DT concepts for data integration at the network level.

Secondary Reconnection Between Interlinked Flux Tubes Driven by Magnetic Reconnection With a Short X‐Line

AbstractA three‐dimensional particle‐in‐cell simulation is performed to study secondary reconnection between two interlinked flux tubes produced by neighboring guide field reconnection x‐lines. The reconnecting magnetic fields of this secondary reconnection is enhanced toward the diffusion region, agree well with that in observations. The magnetic field pileup is attributed to the upstream magnetic tension force, that smashes the flux tubes into each other. We propose that the primary reconnection x‐line length is a key parameter to determine the formation of interlinked flux tubes and secondary reconnection therein. Interlinked flux tubes will form only if the x‐line is short; when the x‐line is long enough, the regular flux ropes are formed instead. The critical x‐line length to form interlinked flux tubes is determined by the distance between two neighbor x‐lines and the magnetic shear angle of the primary reconnection. The results provide a novel scenario of secondary reconnection generation during three‐dimensional reconnection.

Sensitivity analysis of thermal phenomena in APB (Annular Pressure Buildup) in oil wells

This paper presents a sensitivity analysis aimed at understanding how certain parameters of the thermal problem affect the increase in Annular Pressure Buildup (APB) in oil wells. The structure of an oil well experiences high-temperature gradients throughout its lifespan, directly affecting its components such as casings, cement, and rock formation. Among the undesirable effects caused by temperature variation, APB stands out, related to the expansion of fluids confined between casings, being a severe phenomenon that can lead to well collapse. This justifies studies aimed at better understanding temperature distributions and, consequently, the increase in pressure in the annular spaces of oil wells. On the other hand, changes in fluid and component temperatures during well operation depend on many factors, such as operating flow rate, inlet pressure and temperature, operating time, formation type, and properties of the produced/injected fluid. Some of these parameters are chosen for inclusion in the parametric analysis proposed in this work. To achieve the proposed objective, the methodology adopted in this work is based on four steps: a) selection of a reference well used for the analyses; b) definition of the variables to be studied and their assumed values; c) generation and simulation of scenarios formed by the combination of chosen variables; and d) selection and description of the sensitivity analysis method to be employed. The main contribution of this work is to present a sensitivity analysis of certain well parameters in annular pressure buildup, allowing inference about the importance of some isolated properties in the thermomechanical response. Additionally, the methodology presented can be applied to other parameters and thus improve understanding of how each property influences well APB.

Insights into Building VR Solutions in the Police Domain

Advanced technologies such as Virtual Reality (VR) hold immense significance for society, particularly in the law enforcement domain. VR offers a unique platform for immersive, experiential learning, allowing Police Officers (POs) to engage with dynamic scenarios that mirror real-world situations, incidents, and challenges. In the Police domain, the integration of VR solutions provides valuable insights into enhancing POs’ training, situational awareness, and community relations. Learning objectives in VR-based Police training encompass a spectrum of skills, including decision-making under pressure, effective communication and collaboration, and de-escalation techniques. Furthermore, evaluation perspectives span both quantitative and qualitative measures, assessing training efficacy, officer performance, and societal impact. This research adopts a systematic literature review approach where it becomes evident that the future of VR-enabled policing lies in interdisciplinary collaboration, technological innovation, and ethical considerations. Future directions entail refining VR solutions to reflect evolving societal dynamics, integrating Artificial Intelligence (AI) for adaptive scenario generation and analytics, and addressing concerns related to algorithmic transparency, fairness, security, and privacy. By leveraging VR technologies, law enforcement agencies can enhance operational effectiveness and foster trust, accountability, and responsibility within the communities they serve.

Generative Artificial Intellegence (AI) in Pathology and Medicine: A Deeper Dive.

This review article builds upon the introductory piece in our seven-part series, delving deeper into the transformative potential of generative artificial intelligence (Gen AI) in pathology and medicine. The article explores the applications of Gen AI models in pathology and medicine, including the use of custom chatbots for diagnostic report generation, synthetic image synthesis for training new models, dataset augmentation, hypothetical scenario generation for educational purposes, and the use of multimodal along with multi-agent models. This article also provides an overview of the common categories within generative AI models, discussing open-source and closed-source models, as well as specific examples of popular models such as GPT-4, Llama, Mistral, DALL-E, Stable Diffusion and their associated frameworks (e.g. transformers, GANs, diffusion-based neural networks), along with their limitations and challenges, especially within the medical domain. We also review common libraries, and tools that are currently deemed necessary to build and integrate such models. Finally, we look to the future, discussing the potential impact of generative AI on healthcare, including benefits, challenges, and concerns related to privacy, bias, ethics, API costs, and security measures.

On the dynamics of treasury bond yields: From term structure modelling to economic scenario generation

On the dynamics of treasury bond yields: From term structure modelling to economic scenario generation

Quantity bidding strategy of inter-provincial and intra-provincial two-level peak regulation market of coal-fired units considering prediction error

Abstract Under the background of the ‘carbon peaking and carbon neutrality’ target, the proportion of new energy installed capacity has been greatly increased. The problem of wind and photovoltaic power abandonment has become more and more serious. In the construction of inter-provincial and intra-provincial multi-level electricity markets, it is urgent to establish a source-network coordination mechanism and tap new consumption capacity to adapt to the rapid development of new energy. This paper first sorts out the joint operation process of the inter-provincial and intra-provincial two-level electric energy market and the peak regulation market. Then, with the goal of minimizing the total cost of the inter-provincial peak regulation market and provincial joint market, a collaborative optimization model of inter-provincial and intra-provincial peak regulation quantity bidding strategy is established. The expected scenario generation method considering prediction error is proposed. The collaborative optimization solution process is explained. Finally, the proposed model is simulated to verify the effectiveness of the collaborative optimization of inter-provincial and intra-provincial peak load regulation.

Corrigendum to ‘Low-dimensional scenario generation method of solar and wind availability for representative days in energy modeling’[Applied Energy Volume 306 PB (2021) 118075

Corrigendum to ‘Low-dimensional scenario generation method of solar and wind availability for representative days in energy modeling’[Applied Energy Volume 306 PB (2021) 118075

Food and land system transformations under different societal perspectives on sustainable development

Abstract The future of food and land systems is crucial for achieving multiple UN Sustainable Development Goals, given their essential role in providing adequate nutrition and their significant impact on Earth system processes. Despite widespread consensus on the need for transformation, discussed strategies vary widely, from technology-driven to sufficiency-focused approaches, emphasizing different agents of change and policy mixes. This study assesses the implications of a new generation of target-seeking scenarios incorporating such diverse sustainability perspectives. We apply two integrated assessment models to explore food and land futures under three whole-economy Sustainable Development Pathways (SDPs): Economy-driven Innovation, Resilient Communities, and Managing the Global Commons. Our assessment shows that all SDPs align sufficient food supply with progress towards planetary integrity, halting biodiversity loss, mitigating adverse impacts from irrigation, and significantly reducing nitrogen pollution. While all pathways comply with the Paris climate target, they diverge in the timing of climate mitigation efforts and focus on different greenhouse gases and emission sources. The Economy-driven Innovation pathway rapidly achieves net-negative CO2 emissions from the land system, whereas Resilient Communities and Managing the Global Commons significantly decrease agricultural non-CO2 emissions. Moreover, sustainability interventions attenuate trade-offs associated with narrowly focused mitigation pathways and reduce reliance on carbon dioxide removal strategies like re/afforestation and bioenergy with carbon capture and storage.

Conditional neural field latent diffusion model for generating spatiotemporal turbulence

Eddy-resolving turbulence simulations are essential for understanding and controlling complex unsteady fluid dynamics, with significant implications for engineering and scientific applications. Traditional numerical methods, such as direct numerical simulations (DNS) and large eddy simulations (LES), provide high accuracy but face severe computational limitations, restricting their use in high-Reynolds number or real-time scenarios. Recent advances in deep learning-based surrogate models offer a promising alternative by providing efficient, data-driven approximations. However, these models often rely on deterministic frameworks, which struggle to capture the chaotic and stochastic nature of turbulence, especially under varying physical conditions and complex, irregular geometries. Here, we introduce the Conditional Neural Field Latent Diffusion (CoNFiLD) model, a generative learning framework for efficient high-fidelity stochastic generation of spatiotemporal turbulent flows in complex, three-dimensional domains. CoNFiLD synergistically integrates conditional neural field encoding with latent diffusion processes, enabling memory-efficient and robust generation of turbulence under diverse conditions. Leveraging Bayesian conditional sampling, CoNFiLD flexibly adapts to various turbulence generation scenarios without retraining. This capability supports applications such as zero-shot full-field flow reconstruction from sparse sensor data, super-resolution generation, and spatiotemporal data restoration. Extensive numerical experiments demonstrate CoNFiLD’s capability to accurately generate inhomogeneous, anisotropic turbulent flows within complex domains. These findings underscore CoNFiLD’s potential as a versatile, computationally efficient tool for real-time unsteady turbulence simulation, paving the way for advancements in digital twin technology for fluid dynamics. By enabling rapid, adaptive high-fidelity simulations, CoNFiLD can bridge the gap between physical and virtual systems, allowing real-time monitoring, predictive analysis, and optimization of complex fluid processes.

On-demand reverse design of polymers with PolyTAO

The forward screening and reverse design of drug molecules, inorganic molecules, and polymers with enhanced properties are vital for accelerating the transition from laboratory research to market application. Specifically, due to the scarcity of large-scale datasets, the discovery of polymers via materials informatics is particularly challenging. Nonetheless, scientists have developed various machine learning models for polymer structure-property relationships using only small polymer datasets, thereby advancing the forward screening process of polymers. However, the success of this approach ultimately depends on the diversity of the candidate pool, and exhaustively enumerating all possible polymer structures through human imagination is impractical. Consequently, achieving on-demand reverse design of polymers is essential. In this work, we curate an immense polymer dataset containing nearly one million polymeric structure-property pairs based on expert knowledge. Leveraging this dataset, we propose a Transformer-Assisted Oriented pretrained model for on-demand polymer generation (PolyTAO). This model generates polymers with 99.27% chemical validity in top-1 generation mode (approximately 200k generated polymers), representing the highest reported success rate among polymer generative models, and this was achieved on the largest test set. Importantly, the average R2 between the properties of the generated polymers and their expected values across 15 predefined properties is 0.96, which underscores PolyTAO’s powerful on-demand polymer generation capabilities. To further evaluate the pretrained model’s performance in generating polymers with additional user-defined properties for downstream tasks, we conduct fine-tuning experiments on three publicly available small polymer datasets using both semi-template and template-free generation paradigms. Through these extensive experiments, we demonstrate that our pretrained model and its fine-tuned versions are capable of achieving the on-demand reverse design of polymers with specified properties, whether in a semi-template generation or the more challenging template-free generation scenarios, showcasing its potential as a unified pretrained foundation model for polymer generation.

(Invited) Marine Plastic Waste-Based Supercapacitors for Saving Remote Island

Marine plastic waste is a global environmental issue that needs urgent attention as it adversely affects the environment and living organisms, including human beings. More than 32–280 million metric tons of plastic are produced every year; however plastic production decreased in 2020 owing to a drop in demand caused by COVID-19[1-4]. Plastic production is expected to reach 53 million metric tons by 2030[5] and 155–265 million metric tons by 2060[3]. As the amount of plastic waste from sources other than packaging is decreasing, most of the waste is not treated as waste and is discharged into the ocean. Approximately 4.7–12.7 million metric tons of plastic waste is dumped into the sea annually[4,6], covering an area of 1.6 million km2[2]. About 150 million tons of plastic waste is reportedly present in the sea, which will likely remain in the sea until at least 2050. Properties of plastic further exacerbate this problem. Plastic shopping bags take more than a thousand years to completely decompose naturally, and plastics that have once entered the sea continue to harm the environment for an extremely long time[7]. Approximately 700 species, including endangered marine species, have been injured and killed[8]. Such deaths were attributed to glass and trees in the past; however, in recent years, 92% of these deaths have occurred owing to marine plastic waste/microplastic particles[8-10]. The effects of plastic waste and microplastic particles also affect coral reefs[11] and the Arctic[12]. This research will increase the applicability of activated carbon (AC) in marine plastic recycling. However, marine plastics are a mixture of various plastics, and it is difficult to recycle all of them from deposits, additives, and paints. Therefore, a recycling destination with high added value is required. In contrast, applications of AC have evolved from those in improving water quality to those in electrode materials, and especially as a storage battery. Its application prospects differ from other storage batteries in terms of charge and discharge characteristics; therefore, it is regarded as a power storage device whose demand will continue to rise in the future. Moreover, impurities in the electrode materials of ordinary secondary batteries can have a large adverse effect on battery capacity, internal resistance, and cycle life. Supercapacitors with carbon electrodes can be charged and discharged even if some impurities are present. Therefore, this study presented the production of AC as one of the treatment methods for marine plastic waste and evaluated the performance of AC and its performance as a supercapacitor electrode. The aim of this study was to develop methods to manage marine plastic waste (Fig.1(A)) through the production of activated carbon (AC). The specific surface area, micropore volume, and mesopore volume of marine plastic AC prepared under arbitrary temperature and activator weight ratio were measured (Fig1(B),(C)), and the specific capacitance and supercapacitor electrode performance were evaluated. At an activation temperature of 800 °C and a weight ratio of 1:7 between the raw material and 8 M KOH solution, a specific capacitance of 201 F/g and a high surface area of 2389 m2/g were obtained. Accordingly, marine plastic waste-based AC can be used as the electrode material in supercapacitors.[1] C. M. Rochman, et al., Policy: Classify plastic waste as hazardous, Nature 494 (2013), 169–171.[2] L. Lebreton, et al., Evidence that the Great Pacific Garbage Patch is rapidly accumulating plastic, Sci. Rep. 8 (2018), 4666.[3] L. Lebreton, A. Andrady, Future scenarios of global plastic waste generation and disposal, Palgrave Commun. 5 (2019), 6.[4] J. R. Jambeck, et al., Marine pollution. Plastic waste inputs from land into the ocean, Science 347 (2015), 768–771.[5] T. M. Adyel, Accumulation of plastic waste during COVID-19, Science 369 (2020), 1314–1315.[6] E. MacArthur, Beyond plastic waste, Science 358 (2017), 843.[7] C. Giacovelli, Single-use plastics: A roadmap for sustainability. International Environmental Technology Centre (2018).[8] S. C. Gall, R. C. Thompson. The impact of debris on marine life. Mar. Pollut. Bull. 92 (2015), 170–179.[9] O. M. Lonnstedt, P. Eklov, Environmentally relevant concentrations of microplastic particles influence larval fish ecology, Science 352 (2016), 1213–1216.[10] Animal behaviour: Plastic smells good to marine birds, Nature 539 (2016), 332.[11] J. B. Lamb, et al., Plastic waste associated with disease on coral reefs, Science 359 (2018), 460–462.[12] M. Bergmann, et al., Plastic pollution in the Arctic, Nat. Rev. Earth Environ. 3 (2022), 323–337. Figure 1