Despite recent advances, Handwritten Text Recognition (HTR) for Arabic-script languages still lags behind Latin-script HTR. Part of the problem is dataset quality. To help closing this gap, we propose a two-stage framework (CER-HV) for detecting label errors. Stage 1 (CER) is a Character-Error-Rate-based noise detector built on a Convolutional Recurrent Neural Network (CRNN) architecture. Stage 2 (HV) is the Human-In-The-Loop (HITL) Verification of noisy samples detected by the first stage. Applying the CER-HV framework on multiple Arabic-script datasets can identify samples with label errors including transcription, segmentation, orientation, and non-text content errors that can markedly affect HTR performance. These errors were identified by the first stage of the framework with up to 90% (top-50) precision.We also show that our CRNN achieves state-of-the-art performance across five of the six evaluated datasets, reaching 8.46% Character Error Rate (CER) on KHATT (Arabic), 8.22% on PHTI (Pashto), 10.59% on Ajami, and 10.11% on Muharaf (Arabic), all without any data cleaning. We establish a new baseline of 11.3% CER on the PHTD (Persian) dataset. Applying CER-HV improves evaluation CER by up to 1.8 percentage points after dataset cleaning and retraining. Although our experiments focus on documents written in an Arabic-script language, the framework is general and can be applied to other text recognition datasets.
Biocarbon can replace fossil carbon in hard-to-abate sectors, but its large-scale deployment depends on feedstock availability, production cost, and the effective utilisation of energy-rich pyrolysis by-products. Integration with sawmills may address these constraints by providing access to woody by-products while enabling recovery and utilisation of excess pyrolysis heat. This study evaluates the techno-economic performance of sawmill-integrated pyrolysis-based biocarbon production. Stand-alone and integrated configurations are compared using bark, sawdust, and combined feedstock strategies, focusing on interactions between pyrolysis heat recovery, sawmill heat demand, and biomass allocation. Mass and energy balances were calculated and used as input to a techno-economic model. Sensitivity analyses were performed for char yield and sawmill dryer efficiency, while a scaling analysis assessed the influence of sawmill size on biocarbon production capacity and cost. The results show that integration enables recovered pyrolysis heat to be used in lumber drying, thereby improving energy utilisation and reducing production cost compared with stand-alone production. However, the benefit depends strongly on the balance between pyrolysis heat generation and sawmill heat demand. Higher char yield generally reduces production cost, while lower char yield increases recoverable heat and improves integration performance by reducing on-site bark combustion. Higher dryer efficiency decreases heat demand and external biomass requirements. The scaling analysis shows that integration improves economic performance across plant sizes, but production scale is often constrained by internal sawmill by-product availability. Overall, sawmill-integrated biocarbon production can improve energy, economic and resource performance when plant scale, feedstock strategy, and heat integration are adapted to site-specific conditions.
Biochar performance in applications like soil amendment depends on properties including porosity, surface area, and pore structure, which are influenced by both process conditions and feedstock characteristics such as wood anatomy. However, their effects across micro-and nanometre length scales remain poorly understood. This study therefore investigates how feedstock and process conditions affect properties of pine wood biochar produced in industrially relevant pilot-scale continuous reactor for torrefaction (291-315 degrees C, 6-12 min) and pyrolysis (350-400 degrees C, 25 min). High-resolution X-ray microtomography combined with AI-assisted image analysis was used to quantify earlywood-latewood distribution, porosity, pore size, cell wall thickness, and micrometre-scale surface area, while N(2)and CO(2)sorption probed nanometre-scale structure. The wood microstructure was largely preserved across treatments, resulting in highly anisotropic pore networks. Porosity remained high (33-77%), primarily influenced by earlywood-latewood variability rather than temperature. From torrefaction to pyrolysis, cell wall thickness and pore diameter decreased, while microporosity (
Research interest in reliability and sensitivity analyses of concrete dams has grown in recent decades, yet existing studies remain largely limited to gravity and arch dams and to case studies focused on earthquake events. To address this gap, this study aims to evaluate the reliability and sensitivity of concrete buttress dam monoliths using a simulated population based on existing Swedish designs with varying geometric properties such as height and width. Three failure modes, sliding, overturning, and combined sliding and overturning, were assessed under ice load and overtopping conditions. Reliability indices and local sensitivity measures were first estimated using first-order reliability analysis, after which polynomial chaos expansion metamodels were constructed to facilitate global sensitivity analyses. The results show that sliding has the lowest reliability, whereas overturning exhibits very high reliability and is generally not governing. Local sensitivity analyses indicate that rock-concrete interface friction dominates sliding and the combined failure mode, while concrete density is most influential for overturning. Global sensitivity analyses show that sliding reliability is primarily affected by front plate inclination, while overturning is mainly relevant for low monoliths or those with near-vertical front plates, and the combined mode is mainly influenced by asperity and front plate inclinations.
Abstract In saline-alkali or coastal regions, sulfate attack is a primary factor causing performance deterioration of concrete structures. Hence, concrete cold joints, being weak interfaces of structures, warrant attention regarding their shear deterioration performance in sulfate attack environments. This study employed molecular dynamics simulations to establish models characterizing concrete cold joints, specifically a CSH-a-to-CSH-b (CC) interface and a CSH - a - to - SiO 2 (CS) interface. The influence of different Na 2 SO 4 solution concentrations (0%, 5%, and 10%) on the shear deterioration behavior of these interfaces was investigated. The results indicated that as the Na 2 SO 4 concentration increased, the degree of interfacial erosion damage became more pronounced. The penetration of Na 2 SO 4 solution altered the interfacial ionic interactions, making the substrate ions more likely to bond with the solution ions, thereby further deteriorating the shear performance of the interface. When the Na 2 SO 4 concentration increased from 0% to 5% and then from 5% to 10%, the maximum shear stress of the CC interface decreased by 68.8% and 31.0%, respectively, whereas that of the CS interface decreased by 62.5% and 39.4%, respectively. Additionally, Na 2 SO 4 solution significantly influenced the shear failure mode of the model. In the absence of Na 2 SO 4 erosion, shear failure occurred within the CSH-a substrate with lower strength and unstable bond energy. However, as the Na 2 SO 4 concentration increased, the failure mode shifted to distinct interfacial shear failure. This study reveals the influence of sulfate concentration at the nanoscale on the shear behavior of concrete cold joints, providing a new perspective for understanding their deterioration mechanisms in sulfate attack environments.
