Florian Felten, Gabriel Apaza, Gerhard Bräunlich
et al.
Engineering design optimization seeks to automatically determine the shapes, topologies, or parameters of components that maximize performance under given conditions. This process often depends on physics-based simulations, which are difficult to install, computationally expensive, and require domain-specific expertise. To mitigate these challenges, we introduce EngiBench, the first open-source library and datasets spanning diverse domains for data-driven engineering design. EngiBench provides a unified API and a curated set of benchmarks -- covering aeronautics, heat conduction, photonics, and more -- that enable fair, reproducible comparisons of optimization and machine learning algorithms, such as generative or surrogate models. We also release EngiOpt, a companion library offering a collection of such algorithms compatible with the EngiBench interface. Both libraries are modular, letting users plug in novel algorithms or problems, automate end-to-end experiment workflows, and leverage built-in utilities for visualization, dataset generation, feasibility checks, and performance analysis. We demonstrate their versatility through experiments comparing state-of-the-art techniques across multiple engineering design problems, an undertaking that was previously prohibitively time-consuming to perform. Finally, we show that these problems pose significant challenges for standard machine learning methods due to highly sensitive and constrained design manifolds.
Modern engineering, spanning electrical, mechanical, aerospace, civil, and computer disciplines, stands as a cornerstone of human civilization and the foundation of our society. However, engineering design poses a fundamentally different challenge for large language models (LLMs) compared with traditional textbook-style problem solving or factual question answering. Although existing benchmarks have driven progress in areas such as language understanding, code synthesis, and scientific problem solving, real-world engineering design demands the synthesis of domain knowledge, navigation of complex trade-offs, and management of the tedious processes that consume much of practicing engineers' time. Despite these shared challenges across engineering disciplines, no benchmark currently captures the unique demands of engineering design work. In this work, we introduce EngDesign, an Engineering Design benchmark that evaluates LLMs' abilities to perform practical design tasks across nine engineering domains. Unlike existing benchmarks that focus on factual recall or question answering, EngDesign uniquely emphasizes LLMs' ability to synthesize domain knowledge, reason under constraints, and generate functional, objective-oriented engineering designs. Each task in EngDesign represents a real-world engineering design problem, accompanied by a detailed task description specifying design goals, constraints, and performance requirements. EngDesign pioneers a simulation-based evaluation paradigm that moves beyond textbook knowledge to assess genuine engineering design capabilities and shifts evaluation from static answer checking to dynamic, simulation-driven functional verification, marking a crucial step toward realizing the vision of engineering Artificial General Intelligence (AGI).
The future of software engineering--SE 3.0--is unfolding with the rise of AI teammates: autonomous, goal-driven systems collaborating with human developers. Among these, autonomous coding agents are especially transformative, now actively initiating, reviewing, and evolving code at scale. This paper introduces AIDev, the first large-scale dataset capturing how such agents operate in the wild. Spanning over 456,000 pull requests by five leading agents--OpenAI Codex, Devin, GitHub Copilot, Cursor, and Claude Code--across 61,000 repositories and 47,000 developers, AIDev provides an unprecedented empirical foundation for studying autonomous teammates in software development. Unlike prior work that has largely theorized the rise of AI-native software engineering, AIDev offers structured, open data to support research in benchmarking, agent readiness, optimization, collaboration modeling, and AI governance. The dataset includes rich metadata on PRs, authorship, review timelines, code changes, and integration outcomes--enabling exploration beyond synthetic benchmarks like SWE-bench. For instance, although agents often outperform humans in speed, their PRs are accepted less frequently, revealing a trust and utility gap. Furthermore, while agents accelerate code submission--one developer submitted as many PRs in three days as they had in three years--these are structurally simpler (via code complexity metrics). We envision AIDev as a living resource: extensible, analyzable, and ready for the SE and AI communities. Grounding SE 3.0 in real-world evidence, AIDev enables a new generation of research into AI-native workflows and supports building the next wave of symbiotic human-AI collaboration. The dataset is publicly available at https://github.com/SAILResearch/AI_Teammates_in_SE3. > AI Agent, Agentic AI, Coding Agent, Agentic Coding, Software Engineering Agent
Gottmyers Melwyn John, Bhagyanathan Chandragandhi, Sathiyaseelan Govindaraj
et al.
Recycled aluminium has received greater attraction in the past few decades but the presence of certain impurities often impacts the properties of these alloys. The presence of chromium is known to enhance the properties of aluminium alloys. The influence of nano powders into the recycled aluminium matrix has not been explored to a greater extent. The present study aims at analysing the behaviour of the recycled aluminium under the influence of micro and nano powders of chromium. The research primarily utilized Troma scraps as the foundational material for experimentation. Varied concentrations of chromium-ranging from 0 to 0.7% of the total melt-were introduced in the form of micro and nano powders. Characterization was done by using optical microscope, image analysis, Field Emission scanning electron microscopy (FESEM) and energy dispersive spectroscopy (EDS). The addition of Cr in the form of nano powder to AlFeSi melt resulted in the formation of intermetallic Al-Si-Cr-Mn-Fe-Cu phases with polyhedral and irregular shape that enhanced the hardness and wear properties of aluminium alloy.
In this study, a porous FeAl intermetallic compound with high porosity was synthesized via vacuum sintering using Mg powder as a pore-forming agent, leveraging its high saturated vapor pressure and almost non-reactivity with Fe. The influence of the addition of Mg powder on pore characteristics and microstructure evolution was systematically investigated. The results indicate that variations in Mg content within sintered compacts exhibit a negligible impact on primary phase composition, with the FeAl phase remaining predominant. However, excessive initial Mg content induces the encapsulation of the FeAl phase by minor Fe<sub>2</sub>Al<sub>5</sub> and Al<sub>3</sub>Mg<sub>2</sub> phases, compromising the phase’s purity. The porosity positively correlates with Mg content, and porous material with a porosity of 72.8% is obtained (40 at.% of Mg as an additive). Moreover, the pore structure manifests as an interconnected hole morphology. These findings provide valuable insights for further exploration of the design of porous FeAl material and its performance enhancement in emerging applications.
The grade control drillhole spacing and mining selectivity decisions are typically made using the resource model estimated from exploration and infill drilling data. Once production starts, large quantities of grade control data are collected to delineate ore and waste boundaries for ore control. In this work, a grade control drillhole spacing and mining selectivity optimization workflow is presented which allows the practitioner to use site specific knowledge to determine the most profitable ore control mining scenario. The densely gridded production data is used to simulate a ground truth block model from which scenarios of variable drillhole spacing and selectivity are evaluated against their corresponding costs to determine the maximum profit scenario. Eleven mining scenarios are evaluated using year production data from three distinctly heterogeneous mine with drillhole spacing and mining selectivity varying from 3 × 3 × 3 m (27 m3) to 30 × 30 × 15 m (13,500 m3). The profit differences from the optimum scenario varied by millions of dollars (1–8%) against the next best case depending on the heterogeneity of the deposit. Practitioners could apply this workflow to inform grade control drillhole spacing and mining selectivity decisions for different domains within the mine or multiple pits especially if distinctly heterogeneous volumes exist.
The financial sector, a pivotal force in economic development, increasingly uses the intelligent technologies such as natural language processing to enhance data processing and insight extraction. This research paper through a review process of the time span of 2018-2023 explores the use of text mining as natural language processing techniques in various components of the financial system including asset pricing, corporate finance, derivatives, risk management, and public finance and highlights the need to address the specific problems in the discussion section. We notice that most of the research materials combined probabilistic with vector-space models, and text-data with numerical ones. The most used technique regarding information processing is the information classification technique and the most used algorithms include the long-short term memory and bidirectional encoder models. The research noticed that new specific algorithms are developed and the focus of the financial system is mainly on asset pricing component. The research also proposes a path from engineering perspective for researchers who need to analyze financial text. The challenges regarding text mining perspective such as data quality, context-adaption and model interpretability need to be solved so to integrate advanced natural language processing models and techniques in enhancing financial analysis and prediction. Keywords: Financial System (FS), Natural Language Processing (NLP), Software and Text Engineering, Probabilistic, Vector-Space, Models, Techniques, TextData, Financial Analysis.
There is abundant observational data in the software engineering domain, whereas running large-scale controlled experiments is often practically impossible. Thus, most empirical studies can only report statistical correlations -- instead of potentially more insightful and robust causal relations. To support analyzing purely observational data for causal relations, and to assess any differences between purely predictive and causal models of the same data, this paper discusses some novel techniques based on structural causal models (such as directed acyclic graphs of causal Bayesian networks). Using these techniques, one can rigorously express, and partially validate, causal hypotheses; and then use the causal information to guide the construction of a statistical model that captures genuine causal relations -- such that correlation does imply causation. We apply these ideas to analyzing public data about programmer performance in Code Jam, a large world-wide coding contest organized by Google every year. Specifically, we look at the impact of different programming languages on a participant's performance in the contest. While the overall effect associated with programming languages is weak compared to other variables -- regardless of whether we consider correlational or causal links -- we found considerable differences between a purely associational and a causal analysis of the very same data. The takeaway message is that even an imperfect causal analysis of observational data can help answer the salient research questions more precisely and more robustly than with just purely predictive techniques -- where genuine causal effects may be confounded.
The early warning of coal spontaneous combustion is difficult to solve the perception of fire information. The acoustic emission (AE) signal generated by thermal damage and destruction on coal rock can perceive the temperature, and then issue a warning on the evolutionary state of coal spontaneous combustion. In order to study the change of AE signal in the process of coal heating and combustion, we explore the difference of and the time and frequency changes of the AE signals released by coal in different temperatures in the process of coal spontaneous combustion, and study the generation mechanism of the AE signal, so as to provide a theoretical basis for the inversion of coal temperature and for monitoring coal spontaneous combustion. The AE signal test system is constructed. By choosing the bituminous coal from Dongtan coal mine as the research subject, we test and analyze the change of the AE signal during coal heating process, analyze the evolutionary process of coal deformation and cracking under the influence of temperature and the change of coal composition, and examine the generation mechanism of AE signal in the process of coal heating. The results show that the ringing counts of coal increase sporadically from heating to combustion, and the correlation coefficient exceeds 0.87 by linear fitting, indicating that the frequency of coal deformation and cracking increases during the heating process. With the increase of temperature, the growth rate of ringing counts and the average frequency (AF value) of AE signal generated by coal are lower in the initial heating stage, but are higher when the temperature rises rapidly. The cumulative damage parameters show an inflection point when the coal temperature exceeds 80°C, and the main damage mode changes from rapid expansion damage to progressive damage. By analyzing the macro- and microscopic changes of coal structure and composition at different temperatures, we find that the mineral conversion rate in coal affects the main damage mode inside the coal body with the increase of temperature, and that the high mineral conversion rate corresponds to the rapid expansion damage inside the coal body, which promotes the derivative expansion of coal pores and cracks, and produces gradually increasing AE signals.
Fatemeh Kazemi, Ataallah Bahrami, Yousef Ghorbani
et al.
The interaction and synergic effect of particle size on flotation efficiency were investigated by a comparison study between laboratories (size-by-size flotation modes) and industrial scale operational data (whole mixed size fraction). For this purpose, sampling was done from the feed, concentrate, and tailing of the flotation rougher cells of the Sungun copper processing complex (located in the northwest of Iran). In the size-by-size flotation mode (lab scale), the sample was first subjected to different size fractions, and then flotation tests were performed for each fraction. On an industrial scale, the particle size distribution of feed, concentrate, and tailing of flotation of the rougher stage have been analyzed. According to the results, in the case of industrial flotation mode (whole mixed size fraction), the particles with d80=84 μm were more likely to reach the tailing of flotation, and the particles within the size range of +63-180 μm constituted the highest amount of concentrate particles. In lab flotation mode (size-by-size), the maximum recovery was in the size fraction of +40-60 μm. By comparing the two flotation modes of industrial (whole mixed size fraction) and lab (size-by-size), for fractions 125 μm, the recovery trend was reversed and the lab flotation recovery was greater than the industrial flotation recovery. Coarse particle flotation has significant economic and technological benefits. By improving the recovery of coarse particles during the flotation process, the amount of grinding requirements will be reduced and consequently, it will considerably decrease the amount of energy consumption.
The optimal design of microstructure is the key to further improve the properties of particle reinforced metal matrix composites. In this study, in-situ TiB2 and TiC + TiB2 reinforced Cu matrix composites are used as examples to develop the microstructure-based representative volume element models to describe their damage evolution and strengthening mechanisms. The most critical features including size distribution and morphology of the particles as well as the fracture of metal matrix, brittle damage of ceramic particles and traction-separation debonding of the interface were integrated into the modeling. Simulation results show that the model could well predict the performance of the composite and reveal its damage mechanism. When the particle size is normally distributed, the composite exhibits higher strength due to the size effect of different particles, which is more consistent with the experimental result. Both matrix and ceramic particles in mixed-phase particle reinforced metal matrix composites can be subjected to higher stresses, which results in higher strength of (TiC + TiB2)/Cu composites. Additionally, the spherical TiC particles facilitate the relaxation of stress concentration within the composite, resulting in a synergistic enhancement of strength and ductility. Under compression, localized interface damage first appears around the sharp corners of the hexagonal particles and induces crack initiation toward the matrix. These microcracks are interconnected and propagate along the direction of high stress, eventually leading to the failure of the composite. This work provides the basis and new ideas to reveal the failure mechanism of composites and further optimize their configuration design.
LiBH4 has been considered as one of the most promising energy storage materials with its ultrahigh hydrogen capacity, which can supply hydrogen through hydrolysis process or realize hydrogen-to-electricity conversion via anodic oxidation reaction of direct borohydride fuel cells (DBFCs). However, the realization of practical hydrogen applications heavily depends on the effective synthesis of high-purity LiBH4 and recycling of the spent fuels (LiBO2·xH2O). The present work demonstrates a convenient and high-efficiency solvent-free strategy for regenerating LiBH4 with a maximum yield close to 80%, by retrieving its by-products with MgH2 as a reducing agent under ambient conditions. Besides, the hydrogen released from the regeneration course can completely compensate the demand for consumed MgH2. The isotopic tracer method reveals that the hydrogen stored in LiBH4 comes from both MgH2 and coordinated water bound to LiBO2. Here, the expensive MgH2 can be substituted with the readily available and cost-effective MgH2−Mg mixtures to simplify the regeneration route. Notably, LiBH4 catalyzed by CoCl2 can stably supply hydrogen to proton exchange membrane fuel cell (PEMFC), thus powering a portable prototype vehicle. By combining hydrogen storage, production and utilization in a closed cycle, this work offers new insights into deploying boron-based hydrides for energy applications.
Abstract. The marketing activity of «MasloTrade» LLC, which operates in the field of promotion and distribution of lubricants in the markets of Ukraine, was analyzed. It has been established that the innovative high margin Flagship product is not mainstream, due to its high price in simple comparison with a similar product, and therefore it is quite logical that the Premium segment is the largest. It was determined that the possibility of engineering support from the distributor company should be included in the additional terms of supply. These aspects are taken into account in industrial branding. The sales structure of «MasloTrade» LLC according to the FPS approach, the structure of the range of lubricants of the Mobil brand in 2021 according to PVL, CVL and IND lines were evaluated. The average sales price of MasloTrade LLC for 2021 was determined. It is noted that lubricants from the Flagship segment have a significantly higher price compared to other segments. This confirms the earlier data that these are innovative, ecological and high-margin lubricants. Market segmentation was carried out using modern information technologies in two stages: macro- and micro-segmentation. It was established that the following three segments were distinguished by the type of activity, frequency and volume of purchases: agricultural farms and holdings; enterprises of the metallurgical industry; mining and mining processing enterprises. Keywords: information technologies, international marketing, corporate branding, planning, pricing, product policy.
In the search for versatile and effective weld cladding processes to deposit ultra-wear-resistant Ni-WC MMC (Ni-based tungsten carbide metal matrix composite) overlays for mining applications, there is an increasing interest in exploring advanced low-heat-input cold metal transfer (CMT) method. Depositions of single weld bead tracks of Ni-WC MMCs on steel plates were performed by employing the CMT process; Taguchi’s design of experiments was used to plan the experimental investigation. All weld tracks exhibit continuous and uniform bead profile and sound metallurgical bonding to the substrate. Retained WCs are present in the overlay tracks relatively uniformly. The formation of primary WC and secondary carbides is observed depending on the level of dilution. In contrast to standard gas metal arc welding processes, the volume fraction of retained WC, which is negatively correlated with dilution level, is not directly interrelated with heat input for the CMT process and can reach a high level together with improved weld bead appearance at high deposition rate. Deposition rate has a positive correlation with average instantaneous power, which is, in turn, positively correlated with wire feed speed. The addition of oxygen into shielding gas mixtures promotes carbide transfer from cored feed wire to the weld track and increases the volume fraction of retained WC. Analysis of signal-to-noise ratios shows that it is difficult to find a single set of optimized processing parameters, and trade-offs are needed in engineering practice. The present investigation demonstrates that the Taguchi method is a powerful tool in process improvement for weld cladding of Ni-WC MMC overlays.
B2-ordered FeAl based materials with 0, 10, 20, 30, 40, 50 and 60 wt.% B2-ordered NiAl reinforced phase were prepared by mechanical alloying with subsequent hot-pressing sintering. The microstructure, mechanical properties and oxidation resistance were investigated. The results showed that after adding NiAl, the main phase was composed of B2-ordered NiAl, B2-ordered FeAl phase, a small amount of AlFe3C0.5 and α-Al2O3. With the increase of NiAl content, the hardness and compressive strength of NiAl/FeAl composites first increased and then decreased. When the content of NiAl increased to 30 wt.%, the hardness and compressive strength of NiAl/FeAl composite reached the maximum value of 50.30 HRC and 2053.43 MPa, which was 35.2% and 40.8% higher than FeAl alloy. Besides, with the increase of NiAl content, the formation of dense oxide film on the surface was easier and faster, and the oxidation resistance was gradually improved. When the content of NiAl increased to 60 wt.%, the oxidation weight gain of NiAl/FeAl composite was 0.19 mg/cm2. Compared with FeAl alloy, the oxidation resistance of 60 wt.% NiAl/FeAl was increased by 81%, and the thickness of oxide layer was reduced from 1.527 μm to 0.253 μm. At this time, the oxide on the surface of the oxidized sample was mainly composed of α-Al2O3, α-Fe2O3 and a small amount of NiO.
1 China Metallurgical Construction Research Institute Co., Ltd., 33#, Xitucheng Road, Haidian District, Beijing 100088, PR China; wang.haidong@139.com (H.W); gaokangle@qq.com (K.G); lupotter@163.com (B.L) 2 School of Chemical and Environmental Engineering, China University of Mining and Technology (Beijing), Ding 11#, Xueyuan Road, Haidian District, Beijing 100083, PR China; E-mail: wcrzgz@126.com