Driven by advances in machine learning and sensors, data-driven state estimation has become a key technology for supporting operational decisions in safety-critical systems. However, a single point estimate fails to reveal the uncertainty and confidence associated with a prediction, which can lead to overconfident decisions; as reliance on such estimates grows, the quantification and management of uncertainty becomes a prerequisite for reliable operation. This comprehensive review analyzes uncertainty management in data-driven state estimation, taking battery state estimation, in which current state estimation and future state prediction coexist, as a representative case study. This paper organizes uncertainty along two axes, its source (Aleatoric/Epistemic) and the mechanism that handles it (memory-based before modeling, model-based within the model), and applies this taxonomy to analyze how the dominant source and the management strategy suited to it shift along the prediction horizon. In doing so, it goes beyond a simple classification of techniques to present the correspondence between uncertainty sources and management strategies, and it summarizes the main open challenges and future research directions. The insights of this paper provide a foundation for reliable and risk-aware decision-making in predictive maintenance and safety-critical systems beyond batteries.
Purpose: Authorship can be defined as a combination of content and style. Modern open-source transformer foundational authorship models apply contrastive learning techniques. When naively contrasting texts to an authorship task some amount of semantic leakage is present, as authors frequently repeat their topic preferences. Our aim is to reduce spurious correlations due to topic leakage born from contrastive objectives.Methodology: We present a technique to modify a well-established contrastive learning objective (InfoNCE) using synthetic hard negative examples for in-domain topic leakage improvements, while preserving competitive out of domain performance. This topic leakage mitigation technique aims to distance the content embedding space from the style embedding space. Our experiments aim to demonstrate this technique in detail, using exclusively affordable encoder-only models instead of costly hard negative mining.Results: We showcase the performance with ablations on two different datasets and compare them on out-of-domain challenges. We improve on challenging evaluations with prolific authors, with up to 10% increase in accuracy on highly diverse bodies of work. Trials with standard challenges also demonstrate the preservation of zero-shot capabilities of this method as fine-tuning.
Concentrated photovoltaic–thermal (CPV/T) technologies offer a promising approach for simultaneous electricity generation and thermal energy recovery. However, the integration of dual-concentration receivers combining multi-junction (MJ) and polycrystalline (Poly-PV) solar cells with solar desalination systems, together with their coupled thermo-electrical and desalination performance, has not been comprehensively investigated. To address this gap, this study presents a coupled numerical investigation of an integrated concentrated photovoltaic multi-junction thermal system (CPV–MJ/T) coupled with a stepped solar still (SSS) for simultaneous power generation and freshwater production under concentrated solar operating conditions. The proposed system integrates MJ and Poly-PV cells operating at concentration ratios of 400× and 20×, respectively, within a V-shaped CPV/T receiver. A coupled CFD–MATLAB framework was developed to predict the thermo-fluid, thermal, electrical, and desalination performance of the integrated system. The CFD analysis shows that increasing coolant flow rate significantly improves thermal regulation and temperature uniformity. Under peak operating conditions, the Poly-PV and MJ cell temperatures decrease from 86 to 73 °C and from 92 to 71 °C, corresponding to thermal reductions of 15.9% and 23%, respectively. In addition, electrical power loss decreases from 26% to 20% for the Poly-PV cells and from nearly 7% to 5% for the MJ cells. The recovered thermal energy exceeds 2.2 kWh/day for the Poly-PV section and 1.0–1.2 kWh/day for the MJ section, increasing freshwater productivity by up to 39.5% and improving the thermal efficiency of the stepped solar still by 44.6%. These findings demonstrate the potential of the proposed CPV–MJ/T–SSS system for efficient solar-driven electricity generation, thermal recovery, and freshwater production.
Remote sensing is a valuable tool for creating site-specific maps to reduce the environmental impact of excessive nitrogen (N) fertilization and for predicting crop yield and quality in harvest planning and food security. The aim of this study was the application of remote sensing to improve N management and harvest assessment in bread wheat (Triticum aestivum L.). A wheat field experiment with four N levels and two water regimes was conducted in Central Spain over 2 years. Ground-truth measurements of biomass, plant N concentration, and nitrogen nutrition index (NNI) were collected at three growth stages, with grain yield and N concentration recorded at harvest. Close to the dates of the ground measurements, hyperspectral imagery was acquired covering the visible and near-infrared regions (400–850 nm) and part of the short-wave infrared (950–1750 nm) from an aircraft flying 300 m above the experiment. Sentinel-1 and Sentinel-2 imagery of the site was downloaded and processed. Vegetation indices extracted from the airborne imagery were tested to assess NNI and combined with satellite data by ensemble models (multiple linear regression, artificial neural network, random forest) to predict wheat traits at harvest. The canopy chlorophyll content index (CCCI) was the best proxy for crop N status, assessing NNI with root mean square error of 0.21. The NNI maps from aerial imagery reflected the spatial distribution of wheat N requirements and enabled identification of N-responsive and nonresponsive sites for yield at stem elongation and for grain N concentration at flowering. Visible and near-infrared regions provided reliable yield estimates, and bands from the red-edge and the short-wave infrared regions improved prediction of N-related crop traits. Models using hyperspectral imagery and Sentinel-2 data performed comparably. The findings highlight the effectiveness of hyperspectral and multispectral imagery for crop monitoring, N-fertilizer management, and harvest planning. Canopy chlorophyll content index (CCCI) correlates with nitrogen nutrition index (NNI). Crop N status maps based on NNI can guide N fertilizer recommendations. Yield response to N fertilization was assessed more accurately than grain quality response. Airborne hyperspectral and multispectral satellite sensors estimate wheat traits similarly at flowering. Assessing wheat N-related traits improves when SWIR and narrow red-edge bands are included. The data provided in this manuscript enable the creation of in-season N fertilizer application maps and the assessment of wheat yield and grain quality at harvest using sensors on aerial or satellite platforms. These maps reflect the spatial variability of the crop N status and are needed for site-specific N management at the field scale.
Abstract According to Global Cement and Concrete Association (GCCA), the cement industry is responsible for about 7–8% of all anthropogenic CO2 emissions released into the atmosphere as reported by GCCA et al. (CCUS in the Indian Cement Industry: A Review of CO2 Hubs and Storage Facilities, 2024). Clinker production results in high energy costs and the release of CO2 due to the calcination or decarbonation of its main raw material, limestone. In addition, energy is consumed to heat and maintain the temperature inside the rotary kiln, involving another significant CO2 release. This work presents a review of mitigation processes in seven Ibero-American countries (Argentina, Brazil, Chile, Colombia, Peru, Portugal and Spain) that guide the cement industry toward carbon-neutral production by 2030 and 2050. The mitigation parameters are analysed based on the global roadmap published by GCCA (Global Cement and Concrete Association), FICEM (Inter-American Cement Federation) and CEMBUREAU (European Cement Association), as well as the national roadmaps. The six main parameters identified across the countries’ roadmaps were the reduction of total direct CO2 emissions in cement production, thermal efficiency, electrical efficiency, the clinker factor, the use of alternative fuels and (re)carbonation. The adoption of biomass as an alternative fuel and the reduction of the clinker factor through the use of SCM were identified as the main strategies for lowering CO2 emissions from Portland cement production in all seven countries analysed. Despite current efforts, CO2 mitigation in the cement industry has limits. Achieving carbon neutrality by 2050 will require substantial investment in CO2 capture, storage and utilization technologies.
