The factory design phase is highly complex: In addition to constraints, such as siting and permitting, its physical and logistical organization is particularly challenging. Design often comprises several thousand individual components, from machines and systems to logistics and storage components to IT and energy infrastructures.

Many interconnected processes have to be designed and repeatedly examined for interactions and efficiency. All these activities are extremely demanding and, painstaking design notwithstanding, prone to many errors. Moreover, all their data are available in widely varying formats. They are stored in different systems, forms of representation and software applications or can even only be gleaned from conversations. Sorting and assessing these data is very time-consuming. Artificial intelligence approaches are intended to reduce the labor required substantially.

The objective of the GenAI4FFD research project is the development of an assistance system that uses generative intelligence (GenAI for short) to upgrade and optimize existing design process in companies systematically by delivering knowledge supported by data and boosting the efficiency of decision-making in factory design.

The research scientists will be developing a prototype assistance system for three key activities of factory design:

Requirements specification and analysis:

The assistance system is intended to manage knowledge and facilitate communication with clients. The goal is to make actionable, high-quality decisions rapidly, under the given constraints. A dialogue function will make it possible to query the collected base data while ascertaining which data are still needed.

Factory design:

Manufacturing and logistics processes are derived from the requirements specification and analysis. The assistance system will help with concept development and the design of factory systems with specially generated layouts. Constraints, laws and standards, such as for workplaces, the environment or structural design, will be included automatically.

Modeling and evaluation:

The design results generated will have to be evaluated objectively and quantitatively. The assistance system is intended to help implement simulation models that reflect the structure and process design of the particular design in-progress. Simulation can be used to size equipment and the electrical system, plan the materials supply and mathematically model the structural system.

The GenAI assistance system will autonomously assemble knowledge about and experience with factory design. This generic approach will make it possible to expand the application to more than just other factory design activities. It will in fact be generally usable in building design. Its transferability to housing development will consequently also be examined in the project. Small and midsize architectural and engineering firms will particularly be able to profit from this. The GenAI assistance system will be able to expedite new factory and factory redesign and expansion projects. This will make it possible to meet demand for new technologies and products, such as battery cells to transform transportation or electrolyzers to establish the German hydrogen economy, faster and more efficiently.

The following partners are involved in the project together with Fraunhofer IFF:

• Ingenics AG, Ulm
• IPK - Ingenieurplanungs- und Komplexbaugesellschaft mbH, Magdeburg
• METOP GmbH, Magdeburg
• Otto von Guericke University Magdeburg
• Schaeffler Technologies AG & Co. KG, Herzogenaurach
• Tragwerk Ingenieure Döking + Purtak GmbH, Dresden

The project is being funded by the Federal Ministry for Economic Affairs and Climate Action and will run for three years.