The Industrial Science Report: Automotive manufacturing is reshaping material and battery development with data science

Computational model from University of Michigan and GM predicts best process for strong alloys

Aluminum alloys are widely used in automotive and aerospace components because they combine low weight with high strength, which are important features as manufacturers work to improve fuel efficiency and reduce emissions. Among the most widely used are 7000-series aluminum alloys, which contain magnesium and zinc and gain their strength through a carefully controlled heat-treatment process.

To produce the alloy, manufacturers first heat the metal to around 500 °C so the magnesium and zinc dissolve into the aluminum matrix. The alloy is then rapidly cooled in a process known as quenching, followed by a period of natural aging at room temperature that can last from one to three days. During this time, magnesium and zinc atoms cluster together inside the metal, forming nanoscale structures that ultimately determine the alloy’s final strength.

However, the process is notoriously difficult to predict. Rapid cooling during quenching can leave behind microscopic defects known as vacancies—missing atoms in the metal’s crystal structure. These vacancies influence how magnesium and zinc atoms move and cluster during aging, making it challenging to control the final microstructure and mechanical properties. Manufacturers often rely on repeated testing and trial heat-treatment cycles to determine the best processing conditions.

Aerospace manufacturers sometimes avoid this uncertainty by tightly controlling the aging process through specialized methods such as high-temperature deformation or cryogenic storage. While effective, those techniques are expensive and difficult to scale for the high-volume production typical of the automotive industry.

To address this challenge, university and industry researchers have developed a computational model to predict the strength of 7000-series aluminum alloys. By simulating months of aging behavior in minutes, the model could help manufacturers optimize heat-treatment processes, reduce costly experimentation, and expand the use of lightweight aluminum components in vehicles.

A University of Michigan-led research team, in collaboration with General Motors Research & Development, developed a computational modeling framework that predicts how cooling and natural aging affect the long-term strength of high-strength aluminum alloys. Researchers focused on 7000-series aluminum-magnesium-zinc alloys because the strengthening process is unpredictable and costly. By simulating long periods of aging in minutes, the model offers a data-driven alternative to trial-and-error experiments. The model accurately predicts natural-aging behavior and shows how alloy chemistry and cooling can be fine-tuned to control the aluminum alloy properties. The project received funding from the National Science Foundation and support from the Michigan Center for Materials Characterization. 

Chinese researchers use artificial intelligence-driven predictive modeling to optimize materials from additive manufacturing

Powder bed fusion additive manufacturing has helped develop many advanced materials with complex geometries, especially aluminum alloys. The process can manufacture made-to-order components using a high-energy source, such as a laser or electron beam, to melt and fuse atomized powder particles layer-by-layer. Powder bed fusion can produce complex, lightweight consolidated parts, reducing waste compared to machining, but there is a high cost for powder bed fusion machines and material and long post-processing times.

Again, consistency is the core issue. Producing predictable alloys with powder bed fusion is still challenging, and predictive models help guide process optimization. Traditional research in this area uses experimental testing, which is resource-intensive and limited in scalability. Data-driven methods show more promise for scalable applications.

Researchers are developing better aluminum alloys via powder bed fusion by optimizing the process with AI. The model was trained on 3,083 experimental samples and validated across different combinations of process parameters (laser power, scan speed, and layer thickness) and the resulting microstructural properties (the grain size, phase distribution, and porosity) for commonly used metals for additive manufacturing—Titanium-6Aluminum-4Vanadium (Ti-6Al-4 V) and Inconel 718, a nickel-chromium-based superalloy, often used in high-stress components like turbine blades and rocket engines.

The AI model predicts alloy yield strength, tensile strength, and porosity with more than 93% accuracy. For manufacturers, this could shift additive from trial-and-error projects to statistically controlled production. For maintenance and reliability teams, it offers something even more valuable: earlier detection of defect risk and better traceability between build parameters and in-service performance.

Researchers from Shenyang Aerospace University and Jilin University published a study in Nature Communications that presents the AI model for alloy optimization. It uses a Multistage Transfer Learning Model (MTLM), a machine learning framework that trains a model in multiple phases, transferring knowledge from one dataset to another, combined with a Crystal Graph Convolutional Neural Network (CGCNN), which is a deep learning model designed to analyze crystal structures of materials, and Bayesian Neural Network (BNN), which is a neural network that incorporates probability and uncertainty into its predictions, to predict material properties in powder bed fusion additive manufacturing. The approach enables real-time, data-driven optimization of additive manufacturing processes, and the research demonstrates that integrating big data analytics with AI can dynamically refine process parameters, improving consistency, part quality, and design confidence.

Researchers say this work is a significant advancement in predicting material properties for additive manufacturing processes. The model will allow manufacturers to input planned process parameters and powder specifications to obtain predicted properties, supporting proactive quality control, parameter validation, and risk assessment before expensive build operations. During manufacturing, real-time processing for quality monitoring could predict in-process properties and flag potential defects. Future research includes broadening materials an additive manufacturing processes, validating process control with real-time parameter adjustment during manufacturing, and developing inverse design functionality, where manufacturers identify target properties and the model recommends the optimal process parameter combinations.

Nagoya University researchers use 3D printing to create stronger vehicle parts with new aluminum alloys

Researchers at Nagoya University are also using powder bed fusion to engineer new aluminum alloys. This new mixture combines aluminum, iron, manganese, and titanium and can retain strength and flexibility up to 300 °C, directly targeting the thermal limits of many vehicle and aerospace components.

We’ve already discussed how an alloy’s refined microstructure is central to its performance. “It was demonstrated that the Ti [titanium] element contributed to improved tensile ductility (i.e., suppressing the failure of materials) though enhanced grain refinement in the laser power bed fusion process,” says Naoki Takata, Ph.D., professor, department of materials design innovation engineering at Nagoya University.

There are many important process parameters for laser powder bed fusion, including laser power, laser speed, laser beam spot size, powder thickness, laser scanning hatch distance, and atmosphere temperature. The most important parameters for controlling the cooling rate and developing the microstructure, Takata says, are laser power, scan speed, and spot size.

Translating these alloys into automotive parts will depend on a deeper understanding of how they behave under real operating stresses. “We have not yet investigated the high-temperature fatigue or creep strength of the newly proposed alloys (Al-Fe alloy series). It is well known that the high yield strength of the alloys would enhance fatigue and creep strength, but there is no data to support this. They await our future works for the application to safety-critical automotive or aerospace components,” Takata says.

Other future studies will also investigate environmental resistance, such as corrosion or hydrogen embrittlement. “Now we collaborate with electrochemistry scientists to reveal the environmental resistance mechanism,” Takata says. “Of course, further optimization of alloy chemical compositions could be required.

For manufacturers, alloys engineered specifically for powder bed fusion could improve the reliability and repeatability of metal additive manufacturing. Materials that tolerate wider processing parameters while maintaining consistent microstructures may make additive manufacturing more viable for industrial-scale production. More broadly, the research is a shift toward designing metals for additive manufacturing rather than adapting conventional alloys developed for casting or forging.

Researchers at Nagoya University have developed a new series of aluminum alloys optimized for high strength and heat resistance using metal additive manufacturing, specifically laser powder bed fusion. Unlike traditional aluminum alloys that weaken at elevated temperatures, the new materials leverage rapid cooling during 3D printing to trap iron and other elements in metastable microstructures, producing alloys that remain strong and flexible up to 300°C. The team systematically used different combinations of elements and tested the results with electron microscopy, identifying an aluminum, iron, manganese, and titanium composition as outperforming existing 3D-printed aluminum materials in both heat tolerance and mechanical performance. These new alloys use low-cost, abundant elements and are friendly to recycling, addressing industrial needs in automotive and aerospace sectors where lightweight, heat-resistant components, such as compressor rotors or turbine components, can reduce emissions and improve fuel efficiency. The work also provides a general framework for designing next-generation metals tailored for additive manufacturing applications.

Eindhoven University of Technology and Motion Imager secure funding to industrialize volumetric additive manufacturing

More traditional additive manufacturing can be wasteful. Layer-by-layer printing often requires support structures to hold parts in place during printing. In powder bed fusion, unused powder can degrade after repeated heating cycles. Volumetric additive manufacturing has the potential for less waste and a much faster process than traditional additive manufacturing. Researchers say it can improve the “buy-to-fly ratio” or the ratio between raw material purchased and the material that ends up in the final part. 

The layer-less 3D printing technology solidifies entire 3D objects simultaneously, rather than building them layer by layer, including intricate micro-scale geometries. Researchers are working to industrialize volumetric additive manufacturing. 

For automotive manufacturers under pressure to reduce weight, cost, and emissions simultaneously, scalable volumetric additive manufacturing could enable lighter structural and thermal management components with less waste and fewer downstream machining steps.

The researchers use a satellite micro-thruster as an example of the kind of component that could benefit from volumetric additive manufacturing. Micro-thrusters control a satellite’s position and speed in orbit by precisely mixing propellants such as liquid oxygen and methane to generate small bursts of thrust. Because these systems operate under extreme thermal, chemical, and mechanical conditions, their internal channels and chambers must manage heat transfer, fluid flow, and corrosion from reactive propellants.

Manufacturing these devices is particularly challenging. The thruster requires complex internal geometries, including multi-thickness walls only tens of micrometers thick, non-planar structures, micro-scale surface textures, and chambers made from different material compositions. These intricate features make the component difficult to produce with conventional manufacturing or even traditional layer-by-layer 3D printing. The project uses this example to demonstrate how volumetric additive manufacturing could enable the production of highly complex, multi-material structures with the precision needed for advanced aerospace systems.

Researchers at Eindhoven University of Technology (TU/e), including the Mechanics of Materials and Processing and Performance sections within the Mechanical Engineering department, have secured significant funding from the Materials Innovation Institute (M2i) and Holland HighTech to transition volumetric additive manufacturing (VAM) from a scientific concept to an industrially scalable manufacturing technology. The interdisciplinary project, in collaboration with Motion Imager, aims to bridge material science innovations with manufacturable engineering workflows to deliver reproducible volumetric additive manufacturing methods suitable for series production rather than research prototypes. Volumetric AM’s potential lies in producing complex micro- and sub-micron scale features with high material efficiency and reducing waste, compared with traditional layer-by-layer processes. Target industries identified for this technology include automotive, aerospace, space systems, biomedical, and soft robotics, where intricate geometries, multi-material compositions, and tight tolerances are critical for performance and manufacturability. This funding supports efforts to create standardized techniques, computational tools, and workflows that align material design with industrial fabrication requirements.

Next-generation battery materials development enabled by data science

While much of the automotive industry’s materials research focuses on lightweight structural metals like aluminum alloys, the transition to electric vehicles is pushing innovation in another critical area: batteries. Global commitments to carbon neutrality and electrification and the rapid adoption of EVs are accelerating the search for next-generation battery materials that can outperform conventional lithium-ion systems in energy density, durability, and cost.

Developing new battery chemistries is an extraordinarily complex process. Performance depends on a wide range of variables—from temperature and humidity during manufacturing to material mixing ratios, pressure, and processing conditions. Traditionally, engineers refine these parameters through a long series of experiments, gradually improving performance through trial and error. While effective, this approach can take years of laboratory testing and pilot production before a new battery material is ready for large-scale manufacturing.

Automotive supplier DENSO is exploring a different strategy: applying data science and digital analytics to accelerate battery materials development. The company has introduced a platform called AP+DN7 (Analysis Platform + Digital Native Quality Control 7 Tools) that collects and analyzes data from multiple sources across research and production workflows. By integrating artificial intelligence and advanced analytics with traditional quality-control methods, the system helps engineers identify patterns in experimental data and optimize battery materials more quickly.

Using data-driven approaches, DENSO aims to significantly shorten development cycles for new battery chemistries and manufacturing processes. Faster discovery could help manufacturers bring improved batteries to market more quickly. The company also sees potential for applying the same data-science framework to other emerging energy technologies, including solid oxide electrolysis cells (SOECs) used for hydrogen production.

DENSO is using data science methods to drive next-generation battery materials development to support the global shift toward carbon neutrality and electric vehicle (EV) adoption. The company’s strategic focus is on integrating expertise with digital analytics and computational tools to accelerate research into battery performance improvements, cost reductions, and manufacturing readiness essential for EV scale-up. By applying data-driven approaches, DENSO aims to shorten development cycles for new battery chemistries and material combinations that can outperform conventional lithium-ion systems in energy density, durability, and manufacturability. Denso introduced AP+DN7 (Analysis Platform + Digital Native Quality Control 7 Tools), a data analysis platform that enables data ingestion from multiple sources for preprocessing, integration and analysis. DN7 is an evolution of an evolution of the seven foundation tools for quality control, incorporating AI and IT support for data utilization, visualization, and problem solving. DENSO’s work reflects the broader industry drive to harness data science for competitive advantage in battery innovation and EV manufacturing.