Applied Research

Applied research remains the core of our mission as we provide our customers with the latest in innovations and technology transfer solutions. Our research portfolio and programs form the core focus of our available competences. By addressing market needs that will shape our future, our pre-competitive research specifically targets projects that have high commercial potential, thereby enhancing our impact on society and across multiple sectors. Contact us today and review our selected projects below, highlighting our ability to bring technology to market, work with our university and government partners, and further transatlantic research collaboration. Our three research centers have decades of research expertise and are ready to solve your toughest challenges. At the Center for Manufacturing Innovation, Center Mid-Atlantic, and Center Midwest, our customers quickly discover that #WeKnowHow.

Competences

 

Research Competences

  • Materials and Surfaces
  • Production Technologies
  • Information and Communication
  • Energy and Climate Technologies
  • Health

Highlights ->

Battery Technology

Sustainability

Battery Cell Production

The Fraunhofer Competence Center Battery Production is a a joint initiative of research facilities with-in Fraunhofer-Gesellschaft for the further development of battery production in an international environment.

Sustainability

Our expert researchers and scientists are ready to help your organization decarbonize, implement circular supply chains, or innovate, all while working toward the global goal of reducing greenhouse gas emissions.

Programs

State Alliance Program

The State Alliance Program offers state governments, economic development agencies and academic institutions the opportunity to develop technical assistance programs based on the Alliance template and tailored to states’ specific needs and interests. The program works to assist local bussinesses with the challenges and opportunities presented by rapid technological change in manufacturing processes, product development and service delivery.

TechBridge

The Fraunhofer TechBridge Program works with corporations and startup companies to identify and de-risk promising technologies to solve industry challenges. By performing targeted technical searches and conducting validation and demonstration work, TechBridge evaluates and prepares innovative early-stage products for investors and industry.

Applied Research Consortia

The ARC Initiative is a Fraunhofer USA initiative, with lead academic partner The Jacobs School of Engineering at UC San Diego. 

Bringing Technology to Market

Fraunhofer USA brings cutting-edge research and development and a highly trained staff to tackle the toughest problems for our customers. We bridge the gap between academic research and industrial needs, and leverage both in doing so. Our industrial clients include large multi-national companies, SMEs, and startups, in addition to government organizations. We also collaborate with renowned research organizations, universities, and other networks to fulfill our mission of improving the world through the application of advanced technologies. Our creative and enthusiastic team of scientists and engineers are solution driven.

Application of AI to Verify Weld Quality

Fraunhofer USA Center Mid-Atlantic CMA

In the automotive industry, vehicle body production requires from about 3,500 to 14,000 individual resistance welds, known as spot welds, per vehicle to join sheet metal components. These welds must be verified for quality and structural integrity. Current inspection processes rely on static inspection methods, where all welds are manually checked over several shifts using ultrasound. Additionally, weld integrity is periodically verified through destructive testing, providing critical data on weld quality but further adding to the delay between the production of a weld and confirmation of its quality. The overall process flow for the quality checking of spot welds is highly time consuming and labor intensive, making it a focus point for improvements in efficiency in the automotive industry. Advances in machine learning  offer an approach to greatly reduce this inspection effort and the duration of the feedback process to approve welds and thus optimize the entire welding and associated quality assurance endeavor. Engineers at Fraunhofer USA CMA have worked with colleagues at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA in Stuttgart, Germany and Clemson University to address this opportunity with a major automative manufacturer by collecting extensive data directly from the welding equipment and manual verification systems during production. The team trained artificial intelligence (AI) models on historical data to be able to predict weld quality with high precision in real-time, allowing for targeted manual inspections specifically focused on high-risk welds. This approach also allowed for continuous improvement by refining weld parameter settings to enhance their quality. The outcome of this project was to significantly reduce the time and resources required for manual quality checks while maintaining overall production quality. This effort targets at least a 15% reduction in labor hours dedicated to manual inspections and an estimated return on investment for the manufacturer in under a year. The approach undertaken here should also be applicable to optimizing other production processes requiring inspection and validation, in particular other joining technologies used in the automotive and other manufacturing industries

University and Government Collaboration

Human Agent Teaming for Intelligence Tasks

Fraunhofer USA Center Mid-Atlantic CMA 

Team communication and coordination is of critical importance for intelligence gathering by the military and government security agencies, with weaknesses in gathering and processing information often associated with shift handovers, resulting in team cognition challenges. These challenges include inaccuracy blindness, group sharing and storing of knowledge, known as transactive memory systems, and shared mental models. Artificial intelligence (AI) has often been proposed as a possible solution to these problems, since, for example, AI can support teaming by augmenting individuals’ production capabilities, summarize machine read documents and convert them to summary output text, and organize intelligence analysis around entities, such as people and places, rather than freeform text. However, it is not clear how to best align rapidly developing AI technologies with intelligence analysis work. Engineers at Fraunhofer USA CMA have worked on a project with colleagues at the University of Maryland and Duquesne University for the United States Army Research Office to assess the application of AI and machine language analysis to mitigate team communication and coordination problems such as information overload, ignoring potentially relevant data and erosion of trust between team members. The goal of the project was to provide much needed insight into how human teams can work together with AI, especially AI that provides sensemaking support, to improve outcomes in intelligence analysis and avoid exacerbating team interactions. Based on insights gathered from interviews with intelligence analysts, the team developed a software platform and an experimental infrastructure testbed to experimentally study the role of different types of AI during intelligence analyst shift handovers. They also conducted controlled immersive behavioral experiments to test the effect of AI manipulations on sensemaking, problem solving, workload, and transactive memory systems. The testbed consisted of task-relevant input materials, such as mission descriptions and source documents, simulated team members, activity recording tools, such as search tools and scratchpads, experimental monitoring capabilities, such as recording and survey systems, and AI support tools for human analysts, such as AI that can summarize large quantities of information by, for example, constructing topic models. The experiments simulated the 5Vs challenges associated with big data: a high volume of material, a wide variety of material sources, a rapid velocity of information accrual, questionable veracity of some sources, and extractable value being dependent on linking information from multiple sources. The testbed was most recently applied to analyze interactive shifthandovers, comparing relatively simple AI tools with an entity-based AI drawing on developments with ChatGPT and theories on information science and intelligence analysis. The approach shows great promise for assessing AI tools being applied with the goal of improving the efficacy of intelligence analysis.

Transatlantic Collaboration

Fraunhofer USA exemplifies the power of transatlantic cooperation in applied research and development. Through our unique partnership with Fraunhofer-Gesellschaft in Germany, we create a vital bridge between two of the world’s leading innovation ecosystems. This collaboration goes far beyond traditional institutional partnerships – it represents a strategic alliance that accelerates technological advancement and creates lasting positive impact for both continents.

AI-assisted Laser Material Processing

Prototype for AI-assisted laser welding.
Improvements in the welding process that can be realized through an AI-assisted approach.

Fraunhofer USA Center Mid-Atlantic CMA

There are several challenges with laser welding of metal materials and laser cutting of thick metal materials. These include handling materials of diverse thicknesses and qualities, achieving the required detailing and precision in the product, meeting efficiency and time constraint targets, and limiting material wastage. New technologies are being developed to overcome these challenges, including new laser sources with increased power and tailored attributes, high-frequency power modulation of the laser beam, high-frequency oscillation of the laser beam and the focal plane of the laser, and plasma keyhole welding, which allows for the welding of high-alloy and unalloyed materials in a single pass of the laser. However, these new technologies are of high complexity, involving interacting non-linear effects, resulting in unstable processes when they are combined. Therefore, to incorporate these new technologies into a reliable operation requires comprehensive monitoring and control, best achieved through real-time monitoring and computer control guided by artificial intelligence (AI) to recognize deviations from the desired quality attributes and issue commands to promptly rectify deviations and maintain process stability. Engineers at Fraunhofer USA CMA have worked with colleagues at the Fraunhofer Institute for Material and Beam Technology IWS in Dresden, Germany and the Fraunhofer Institute for Applied Optics and Precision Engineering IOF in Jena, Germany to develop controlled optimized laser material processing assisted by AI. The team implemented multimodal process monitoring equipment that provided input data for AI-based process evaluation that then allowed for AI-based closed-loop feedback control. This AI-based solution reduced energy consumption, increased processing speed, improved weld strength, and reduced distortion. Through this project, the team achieved the seamless integration of diverse monitoring technologies, including input from high-speed cameras and microphones, and ultra-fast data processing to provide real-time process control, allowing for dynamic adjustment of laser cutting and welding parameters during operation. The integrated system targets up to 30% higher speeds and up to 40% lower energy use for laser cutting and welding, while it also reduces the risk of cutting and welding failures. It is suitable for a range of materials, including the cutting of thick metal sheets and the welding together of different materials. The built systems are adaptable and scalable to various industrial needs and have broad application potential.