The reuse of foundry silica sand is essential for sustainability in sand casting, yet its quality progressively deteriorates due to the accumulation of oolitic deposits on sand grains during repeated thermal cycling. This work presents a comprehensive numerical framework for predicting the oolitic content distribution in green sand molds, by coupling a mechanistic oolitization model with transient temperature field simulations. The model integrates second-order reaction kinetics, thermodynamic equilibrium described by logistic functions, and linear corrections for seacoal content. Coupled with temperature field simulations using InteCAST CAE, a dedicated computational program was developed to solve the coupled model for each node in a finite‑difference mesh (3512,124 elements). The framework was validated on a production ductile iron crankcase casting. Simulation results show that only 7.2% of the mold volume experiences oolitization, with an average oolitic content increase of 0.168% per pour. Sampled sand exhibited excellent agreement with predictions (the mean absolute error was 0.42%, the root mean square error was 0.65%, and the coefficient of determination R² was 0.96). The proposed approach enables foundries to quantitatively assess sand reusability, optimize reclamation strategies, and reduce waste sand discharge, offering significant environmental and economic benefits for the casting industry.

Simultaneously enhancing the mechanical, electrical, and magnetic properties of functional alloys is critical for realizing next-generation technological applications in fields such as aerospace and biomedical implants. However, this goal is often hindered by inherent performance trade-offs in existing alloy systems, posing a significant challenge in materials science. To overcome these limitations, we propose a precipitation engineering approach using AuPt alloys, utilizing controlled aging heat treatments to precisely tailor the target key properties. Concurrent enhancement of multiple properties was obtained through synchronous control: Ultra-low magnetic susceptibility (from similar to 10(-6) to -1.1 & times; 10(-5)); the ultimate tensile strength and Vickers hardness were significantly improved by approximately 78% and 133%, respectively, while the material maintained a stable electrical conductivity (2.8-4 S/m). Crucially, this synergistic combination, particularly the precisely controlled magnetic susceptibility, enables the material to meet stringent requirements for applications including test masses in space-borne gravitational wave detectors and MRI-compatible implants. Microstructural analysis pinpoints the underlying mechanisms: the magnetic properties is directly linked to the volume fraction and composition of Pt-rich precipitates; electrical conductivity is governed by the dynamic competition between interfacial scattering and solute re-dissolution, achieving its maximum value at the 600 degrees C aging condition; and mechanical properties are dictated by the progression of strengthening mechanisms, with maximum strength achieved in the critical regime approaching the shearing-bypassing transition. This study constructs a unified framework connecting microstructure to property, providing a critical guide for the rational and coordinated design of complex multifunctional alloy systems.

Addressing defect prediction and optimization challenges caused by the strong nonlinear coupling of process parameters in complex sand casting processes, this research proposes a novel paradigm deeply integrating physical mechanisms with explainable machine learning. The primary contribution to artificial intelligence is the development of a physics-constrained data augmentation and interpretation framework. Specifically, an improved Borderline-Physics-Constrained Synthetic Minority Over-sampling Technique (Borderline-PC SMOTE) algorithm is constructed to resolve extreme class imbalance while ensuring synthetic samples strictly conform to metallurgical boundary conditions. Driven by this augmented dataset, the established Borderline-PC SMOTE Backpropagation (BP) neural network model achieves an accuracy rate of 99.48%. Furthermore, by integrating SHapley Additive exPlanations (SHAP) value analysis, the "black box" nature of the model is decoupled to quantify defect sensitivity and causal mechanisms under multiparameter coupling. To navigate high-dimensional spaces, a novel inflation factor algorithm is proposed to systematically search the 18-dimensional parameter space and precisely identify robust fluctuation windows. Regarding the engineering application, taking the QT450-10 ductile iron casting as the prototype, this study validates the proposed artificial intelligence framework in a real-world manufacturing environment. The offered multidimensional optimization method demonstrates significant improvements in casting stability and ensures a comprehensive defect-free yield rate exceeding 98%. This integration of data-driven analysis with underlying physical causality provides both rigorous theoretical support and a practical engineering paradigm for intelligent process control.

Wear and corrosion of heater treater tubes at high temperature similar to 500 degrees C in oil industries is a great problem of chromium/molybdenum low alloy steel (SA 387 Grade 22). These problems can be mitigated by alloying addition of specific rare earth metals in SA 387 Grade 22. Herein, we made alloying of 0.3 wt% and 0.5 wt% of Nb in SA 387 Grade 22 and investigated the effect on wear and corrosion properties. The sample A1 was unalloyed SA 387 Grade 22, while A2 and A3 contained increasing concentrations of Nb as alloying elements, specifically 0.3 % and 0.5 wt%, respectively. All samples were characterized according to the required characterization techniques, which can confirm the improved properties of A1. Scanning electron (SEM) microscopy revealed that the addition of 0.5 wt% Nb modified the microstructure of A3, which exhibited acicular ferrite and precipitates of niobium carbide (NbC). The A3 exhibited superior tensile strength of 879 +/- 7 MPa (up from 580 +/- 7 MPa in A1), Vicker hardness of 384.46 HV (up from 317 HV), and Charpy V-notch toughness of 344 J (up from 305 J), but ductility decreased from 24 % of A1 to 13 % of A3. Pin-on disc test demonstrated that A3 exhibited stable behaviour during the entire sliding time, resulting in a high value of coefficient of friction COF similar to 0.656 (up from 0.591 for A1), directly indicating the improved mechanical strength, hardness due to 0.5 % Nb, with enhanced wear resistance. Barrier properties of A1 were improved by 0.5 % Nb addition in sample (A3), which shifted the corrosion potential towards a noble direction from (-0.507 +/- 0.024 to -0.292 +/- 0.014 V but also decreased the corrosion current density from -5.397 +/- 0.025 to -6.105 +/- 0.041 A/cm(2). A3, in which 0.5 % Nb addition was made, successfully improved the mechanical properties, wear and corrosion resistance of A1. These improved properties of A1 after 0.5 % Nb addition make A3 a suitable candidate for heater treater applications.

Automation and intelligence have become the primary trends in the design of investment casting processes. However, the design of gating and riser systems still lacks precise quantitative evaluation criteria. Numerical simulation plays a significant role in quantitatively evaluating current processes and making targeted improvements, but its limitations lie in the inability to dynamically reflect the formation outcomes of castings under varying process conditions, making real-time adjustments to gating and riser designs challenging. In this study, an automated design model for gating and riser systems based on integrated parametric 3D modeling-simulation framework is proposed, which enhances the flexibility and usability of evaluating the casting process by simulation. Firstly, geometric feature extraction technology is employed to obtain the geometric information of the target casting. Based on this information, an automated design framework for gating and riser systems is established, incorporating multiple structural parameters for real-time process control. Subsequently, the simulation results for various structural parameters are analyzed, and the influence of these parameters on casting formation is thoroughly investigated. Finally, the optimal design scheme is generated and validated through experimental verification. Simulation analysis and experimental results show that using a larger gate neck (24 mm in side length) and external risers promotes a more uniform temperature distribution and a more stable flow state, effectively eliminating shrinkage cavities and enhancing process yield by 15

During the casting process, no-bake resin-bonded sand molds and cores rapidly heat up upon contact with high-temperature molten metal, causing dramatic changes in the resin binder system and a significant deterioration in mechanical properties, which subsequently leads to casting defects. To reveal the mechanism behind the evolution of high-temperature performance, the effects of resin content, base sand type, and particle size on the compressive strength of alkaline phenolic no-bake resin-bonded sand at temperatures ranging from 600 °C to 1,000 °C were investigated. The results show that the temperature range of 600–800 °C represents the primary stage of strength loss, corresponding to intense resin decomposition. Meanwhile, structural reorganization of the carbon skeleton above 900 °C can lead to a partial recovery of strength. This study provides key data and theoretical support for understanding the high-temperature mechanical behavior of resin-bonded sand and its relationship with casting defects.

To address the challenges of high costs in predicting the grain microstructure of TC4 titanium alloy during centrifugal investment casting, this study investigates grain density prediction based on multi-source heterogeneous data. First, a macro-mesoscopic coupled numerical simulation model for solidification grain structure was established and validated. This model can accurately reflect the mapping relationship between thermal history and grain nucleation and growth in stepped castings of varying thicknesses. Subsequently, by augmenting simulated and experimental data, a structured dataset of “thermal history-grain density” comprising 85 samples was constructed. By extracting key features such as high-temperature residence time, mean temperature, peak temperature, and cooling rate, a small-sample prediction model based on Random Forest Regression was developed. The results demonstrate that the model exhibits extremely high accuracy on the test set (coefficient of determination R2 = 0.9961, mean absolute error MAE = 3.02). Feature importance analysis reveals that high-temperature residence time is the most significant process parameter affecting grain density (accounting for 48.1% of the contribution). Finally, the prediction model was successfully applied to the engineering practice of complex titanium alloy strut castings. The predicted results are highly consistent with experimental observations (R2 = 0.9838), providing a reliable data-driven approach for optimizing precision casting processes and precisely controlling the microstructures of titanium alloys.

Lightweight lattice structures are widely desired in various engineering fields for their high stiffness-to-weight ratio and superior energy absorption capabilities. Compared with conventional body-centered cubic (BCC) single-parameter lattice structures (SPLS), the BCC multiple-parameter lattice structures (MPLS) enable better mechanical performance, since they provide more choices for geometric design. This study presents a novel BCC lattice structure with multiple-parameter intelligently designed by the surrogate-assisted particle swarm optimization (ESPSO) algorithm. Using Ti-6Al-4V alloy powder as the material, the as-designed MPLS are successfully fabricated by selective laser melting (SLM) in different build orientations, and the SPLS with the same material usage are also fabricated for a fair comparison. Their surface forming quality of both BCC SPLS and MPLS is excellent observed by ultra-depth of field microscopy. Subsequently, their mechanical properties are studied via quasi-static uniaxial compression testing. The results show that compared with the SPLS, the elastic modulus, yield strength, compressive strength, and energy absorption capacity of MPLS all exhibit a clear improving trend. Among three different building orientations of the MPLS, the 0 degrees orientation exhibits the optimal mechanical properties, while the 45 degrees orientation presents the best energy absorption capacity. The numerical simulation results accurately predict the fracture-deformation mechanism and mechanical properties of the lattice structure, with a prediction error of less than 15%. The mechanical properties of the optimized MPLS are significantly improved, endowing it with promising application prospects for the lightweight design of equipment structures in key fields such as aerospace,- vehicles, and biomechanical engineering.

With the growing demand for lightweight and high-performance components in automotive and aerospace industries,aluminum alloy die-castings are evolving toward larger dimensions and thinner walls,posing significant challenges to thermal management during solidification.Traditional cooling channel designs often fail to ensure uniform temperature distribution,leading to defects such as shrinkage porosity and deformation.This study proposes an automated design framework integrating the moving morphable components(MMC)topology optimization method with particle swarm optimization(PSO)to generate efficient and manufacturable cooling channel layouts for A380 aluminum alloys.Firstly,a systematic initialization strategy was developed with component dimensions of 4-10 mm in width and 15-40 mm in length,along with discrete orientation angles.The optimization process effectively guided components toward high-temperature regions identified through numerical simulation,followed by post-processing operations including temperature-based sorting,overlap removal,and component interconnection.The final design with 20 retained components was selected.Then,castings with a conventional cooling system and without any cooling system were employed as benchmark cases for comparison with the current optimized design.Compared with the conventional and no-cooling cases,the current cooling system exhibits a consistently lower temperature standard deviation after 30 s,maintains superior thermal uniformity throughout solidification,and achieves this improvement without comprising the average temperature.

Vacuum Arc Remelting (VAR) is a critical process for producing titanium alloys; however, the complex melt flow and its interaction with the mushy zone in the solidifying melt pool often result in macrosegregation. In this study, a volume-average based multiphase solidification model is adapted and extended for this issue. Firstly, a novel method, the Adjusted Material Properties (AMP), is proposed to handle the rising melt pool surface during VAR process. Secondly, the complex multicomponent alloy is reduced to a binary-equivalent system for modeling purposes, with the macrosegregation results subsequently back-calculated to reconstruct element-specific distributions. Thirdly, the developed method for handling the rising melt pool surface requires a modified and more refined coupling approach. Given the dominant columnar structure of most titanium VAR ingots, potential equiaxed crystals from fragmentation are neglected. As an example, the VAR process of an engineering-scale ingot (phi 300 mm & times; 1000 mm) with a multicomponent composition of titanium alloy is simulated. Various flow mechanisms-driven by thermo-solutal buoyancy, the self-induced Lorentz force, and an externally applied electromagnetic force - are considered. A typical segregation profile, consistent with results from engineeringscale ingot, is obtained. The formation mechanisms of macrosegregation in the ingot are well explained. Macrosegregation is caused by the transport of solute element in the liquid phase. Melt flow in the direction of the liquid concentration gradient leads to negative segregation, whereas melt flow against the concentration gradient results in the positive segregation. Numerical parameters studies, e.g. mesh sensitivity, are also made to ensure the computational accuracy.

Mold manufacturing is of vital importance in industrial production. The production planning and scheduling for molds have a significant impact on production efficiency and the competitiveness of enterprises. This study delves into fFJSPa, the fuzzy flexible job shop scheduling problem in mold manufacturing, where automated guided vehicle (AGV) constraints are considered. A rigorous mathematical model is formulated for fFJSPa, employing triangular fuzzy numbers to characterize uncertainties in both processing and transportation times, with the objective of minimizing the fuzzy makespan. To solve this complex problem, an ameliorated discrete teaching-learning-based optimization (ADTLBO) algorithm is proposed. ADTLBO enhances the balance between exploration and exploitation through stochastic integration of discrete crossover operators, historical population archives, self-feedback learning, and backtracking learning mechanisms. Comprehensive experiments on fFJSPa test case demonstrate that ADTLBO significantly outperforms other17 algorithms in optimal fuzzy completion time, worst fuzzy completion time, average fuzzy completion time, and mean certain completion time, particularly in large-scale scenarios. Although computational time is marginally higher in specific cases, ADTLBO delivers robust efficacy in resolving intricate fFJSPa instances, thereby providing a reliable and practical decision-support framework for intelligent production scheduling in mold manufacturing environments.

To improve mechanical properties of high-Nb TiAl alloys, boron nitride (BN) reinforcement was introduced, and an optimum Nb content was determined. Ti46Al2Cr0.5BN-xNb (at.

Despite the extensive development of cast iron, the mechanism behind the transition of 3D graphite morphology influenced by Mg has not yet been fully explained. In this study, X-ray tomography (XRT) was employed to systematically investigate the 3D morphology and size distribution of graphite in cast irons with varying Mg contents. The nucleation and growth behavior of graphite with different morphology were discussed. The results showed that Mg reacts with anti-spheroidal elements such as O and S to form Mg-O-S compounds, which act as a scavenger to a certain extent. The addition of Mg increases the undercooling of the melt, and promotes heterogeneous nucleation of graphite. The 3D analysis demonstrated that Mg causes the transition of graphite from lamellar (LG) to vermicular (VG), and further to spheroidal (SG). This transition is accompanied by an increase in both the sphericity and number of graphite particles, as well as an enhancement in the nodularity of the cast iron. LG precipitates at the austenite/liquid interface with high carbon concentration, while a small fraction nucleates heterogeneously on Mn-S compounds. The LG grows between austenite dendrites and interconnects to form large clusters. VG evolves from tadpole-like graphite, which originates from the distortion of SG. Lower Mg content favors the formation of complex coral-like morphology in cast iron. SG primarily nucleates heterogeneously on complex Mg-O-S compounds and grows via divorced growth. Once encapsulated by austenite, further growth of SG depends on the diffusion of carbon atoms.

Parameter optimization in laser powder bed fusion (LPBF) remains challenging due to the high dimensionality of process variables and the cost of extensive experimentation. Here, a hybrid strategy synergizing physical insight with data-driven modelling is proposed to address the data scarcity. A central composite design (31 runs) with rotatable and symmetric characteristics was constructed and constrained within an intermediate laser volumetric energy density regime to fabricate 316 L stainless steel via LPBF. An iterative stepwise regression model with adaptive alpha-level tuning was developed to model the complex and nonlinear relationships between the key parameters (laser power, scanning speed, hatch spacing, layer thickness) and the resulting properties (relative density, tensile strength, yield strength, elongation), achieving R2 values from 82.19% to 91.72%. The findings, derived from model analyses based on the specific experimental design range, unveil a transition in parameter dominance: from geometric factors alone governing density, to the set expanding to include energy inputs for mechanical properties. Experimental validation demonstrates optimized properties, including 99.96% relative density, 834 MPa tensile strength, 685 MPa yield strength, and 44% elongation. This work provides a resourceefficient and physically informed machine-learning framework for LPBF process optimization, offering practical guidance for high-performance additive manufacturing.

The flow behaviour of Ultra-High Performance Concrete (UHPC) plays a critical role in determining its workability at the fresh stage, as well as its mechanical performance and durability after hardening. A three-dimensional computational model is established to simulate UHPC flow, in which the Herschel-Bulkley constitutive law is adopted to represent the nonlinear fresh state rheology of UHPC. The model is validated against L-box experiments, yielding an average relative error not exceeding 7.2%. Building upon this validated framework, this study investigates mould filling evolution in a large precast capping beam shell, with particular attention to the effects of different boundary conditions and related mechanisms. The obtained results indicate a transition in the flow regime from an initial shear-dominated phase to a confined mode governed by hydrostatic pressure accumulation and yield stress resistance. Moreover, both the number of filling locations and the total filling area exhibit strong positive correlations with filling efficiency. From a rheological perspective, the findings elucidate the dynamic mechanisms that govern UHPC flow in confined environments and provide a theoretical foundation for optimising filling processes in large precast UHPC components.

Laser powder bed fusion (LPBF) involves rapid melting, solidification, and repeated layer deposition, making melt-pool evolution and interlayer bonding highly sensitive to powder-bed morphology and process parameters. In this study, a mesoscopic CFD–DEM model coupled with the VOF method was developed to simulate single-track single-layer and double-layer LPBF of Inconel 718 with a non-uniform powder size distribution. The model was validated by comparing simulated and experimental melt-track morphologies, with the maximum relative error in melt-track width remaining within 6%. The results show that laser power and scanning speed exert opposite effects on melt-pool morphology. Increasing laser power enlarges the melt-pool width and depth and improves melt continuity, whereas increasing scanning speed reduces the effective energy input and promotes discontinuous or fractured tracks. The morphology of the first-layer track strongly affects second-layer powder re-spreading and remelting. Smooth and continuous underlying tracks promote uniform powder coverage, sufficient remelting, and dense metallurgical bonding, whereas fractured tracks cause local powder accumulation, insufficient heat penetration, and interlayer lack-of-fusion voids. The parameter set of 200 W and 1000 mm/s provides a balanced energy input and produces the best interlayer bonding quality among the tested cases. In contrast, 250 W and 1500 mm/s leads to fractured tracks and severe interlayer defects. This work clarifies the coupling mechanism among underlying track morphology, secondary powder re-spreading, and interlayer defect evolution in LPBF.

The hot deformation behavior of a novel duplex Ni-42W-10Co-10Mo (wt.%) medium-heavy alloy (MHA) was studied using Gleeble-3800 within temperature ranging from 1150 degrees C to 1300 degrees C under strain rates of 0.001-1 s- 1. The as-cast alloy exhibited a hypoeutectic microstructure, characterized by a face-centered cubic matrix and mu phase distributed in the inter-dendritic regions. The hot deformation activation energy was determined to be 855.6 kJ/mol, primarily due to the precipitation of secondary mu phases during deformation. The mu phase alloys showed a positive effect on dynamic recrystallization (DRX), Particle-stimulated nucleation (PSN) provided numerous nucleation sites for DRX, while the pinning effect of the mu phase inhibited grain growth, leading to formation of a fine-grained microstructure. Furthermore, a strain-induced boundary migration (SIBM) nucleation mechanism was observed in the alloy. The combined effects of PSN and SIBM facilitated formation of refined dynamic recrystallized grains. The hot deformation window of Ni-42W-10Co-10Mo MHA was determined as deformation temperature of 1200 degrees C under strain rates of 0.01-0.1 s- 1 based on hot processing map and microstructural evolution.

This investigation comprehensively examined the influence of various annealing conditions on the microstructural evolution and mechanical properties of rotary-swaged Mg-6Gd-5Y-1Zn-0.3Zr alloys. The annealing treatments were conducted at temperatures of 350 degrees C, 400 degrees C, and 450 degrees C for durations of 1 h, 5 h, and 10 h, respectively. The findings revealed that annealing at 350 degrees C for 1 h resulted in the formation of numerous stacking faults within the grains, coupled with the precipitation of the (3 phase at the grain boundaries. Upon extending the annealing time to 5 h and 10 h, and increasing the temperature to 400 degrees C, lamellar LPSO phases and continuous (3 phases were observed, both of which exhibited progressive coarsening. Annealing at 450 degrees C induced the gradual disappearance of dislocations and twins, accompanied by static recrystallization and progressive grain growth. Most lamellar LPSO phases were eliminated during extended annealing, leaving behind a few coarser phases. The granular (3 phase formed at the grain boundaries and exhibited coarsening with prolonged annealing. Annealing at 350 degrees C and 400 degrees C led to a reduction in strength, which was primarily associated with the consumption of solute atoms in the matrix and the annihilation of dislocations. Moreover, the coarsening of the (3 phase at the grain boundaries impaired elongation. In contrast, annealing at 450 degrees C enhanced elongation, which involved the elimination of dislocations, twins, and intragranular lamellar LPSO phases.

Laser cutting, as an efficient, high-quality, non-contact metal cutting technology, has the potential to replace traditional manufacturing processes for Zircaloy-4 (Zr-4 alloy) cladding materials of nuclear reactors. However, the stability of the laser cutting process has a significant impact on the quality and service safety of Zr-4 alloy and its key components in nuclear engineering. Therefore, this work first proposes a novel approach for recognizing abnormal fluctuations in the laser cutting process using cutting edge morphology characteristics, thereby ensuring the stability of the process. Firstly, the cutting edge images are captured using an ultra-depth of field microscopy, and the edge morphology feature parameters (the length (L) of vertical striations, the inclination angle (theta) of inclined striations, and surface roughness (Ra)) are measured. Secondly, a density-based spatial clustering of applications with noise (DBSCAN) model for process parameter inversion is established with only 2 cutting edge feature parameters (L and theta , Comprising 386 pairs of data) as model input. Then, a three-standard fusion method is proposed to optimize the model and the model can identify process parameters (laser power, defocus amount, cutting speed, and auxiliary gas pressure, etc.) abnormal fluctuations at 80% accuracy. Finally, by incorporating Ra as an additional input feature parameter along with L and theta , the model can identify process parameters abnormal fluctuations at 100% accuracy. This work effectively recognizes abnormal fluctuations of process parameters during laser cutting of Zr-4 cladding materials, thus benefiting the control of these fluctuations and quality management in metal sheet laser cutting.

Residential and public building infrastructure, including bridges and residences, rely heavily on concrete. However, as fractures emerge in concrete structures, their capacity to seal and carry weight decreases, which can cause structural failures and accidents. Finding cracks early saves money because they may be repaired instead of rebuilding the entire building. The most common way to find cracks in concrete is to study the surface visually. On the other hand, human eyesight and perception can lead to subjective mistakes in today's complicated building situations. This research aims to design and evaluate a system that can autonomously detect concrete fractures using convolutional neural networks. The system processes crack images using advanced image processing techniques (e.g., Sobel and ROI methods) and deep learning algorithms to identify crack areas effectively. The experiment results demonstrate that the trained model achieves a classification accuracy of 99.9 per cent regarding concrete cracks. Using the clustering technique to categorize the number of cracks in crack photos can help increase the accuracy of detecting concrete damage levels over time.