Generative Artificial Intelligence (AI) is generating significant interest in engineering communities as they explore how it can help in product engineering. Their interest is in creating next-generation products that bring the best of knowledge and experience to an enterprise and the ecosystem. Generative AI can influence the entire product engineering value chain across engineering, manufacturing, operations, maintenance and disposal.
Generative AI can significantly improve productivity in engineering product development. It can change how product engineering is performed by augmenting individual engineers’ capabilities by automating some day-to-day activities that were previously considered impossible to automate. As organizations adopt this technology across the value chain, the engineering product development process will change steeply.
Engineering product development is highly knowledge-intensive and experience-driven. A significant portion of the current process involves retrieving relevant information, analyzing and interpreting in the context of a problem and making engineering decisions. Generative AI is promising in terms of natural language processing. It enables engineers to retrieve the documented and stored knowledge by querying in the same way they do humans. This empowers designers and engineers to access the most relevant information and enables them to make informed engineering decisions rapidly. It increases the effectiveness and efficiency of the engineering process encompassing the entire value chain with various stages, as illustrated in Figure 1.
Figure 1: Typical Stages in Product Development Value Chain
Generative AI can help in various stages of the product value chain:
Product definition & conceptualization
During the product inception and definition phase of product development, Generative AI can analyze market trends, user feedback, customer complaints, prior product related safety issues, warranty issues and regulatory compliance requirements. Researchers can use Generative AI to enhance market reporting, ideation and product specification drafting. Possible use cases include:
- Generate new ideas/features for products by analyzing customer feedback and identifying areas where new products or features could be developed and enhanced.
- Explore research papers, patents, and related technical publications to get insights into new product design, design improvements and innovative ideas.
- Utilize historical data of similar products across engineering, manufacturing, operations and maintenance functions to conceptualize new products.
- Preliminary validation of product specification against various regulations and standards knowledge across engineering, manufacturing, operations and maintenance.
Based on the nature of the product and industry, Generative AI can be adopted in the core product designs considering domain aspects. It can help rapidly generate new, innovative designs that traditional design methods cannot deliver. Generative AI can develop multiple design solutions based on specific design criteria and constraints. This enables engineers to explore many design options and variations and quickly evaluate and refine them based on specifications. Some of the scenarios of how it can help are –
- Product design related to physical products often involves physics-based simulations such as computer-aided engineering (CAE), finite element analysis (FEA) and computational fluid dynamics (CFD) to generate computer-aided design (CAD) models. Generative AI with deep learning techniques can help evaluate many designs and enable faster decision-making.
When applied to physical product design, it is known as Generative Design and can be a next-generation trend in CAD design for all manufacturing industries. It helps solve complex multi-disciplinary design constraints, reduce component weights and manufacturing costs, scale customization and optimize performance.
- It can recommend personalized or custom designs or configurations for custom products based on customer preferences.
Generative AI technology can also help automate and enhance various facets of the product realization process. This could include production planning, supply chain optimization and inventory planning. Some possibilities include:
- Utilize historical data to predict demand, enabling more accurate production schedules and optimal inventory levels.
- Generate and recommend product manufacturing plans based on product features, production setup and historical production plans of prior products.
- Foresee production issues based on historical data of manufacturing deviations bringing multiple aspects of product features and manufacturing process parameters.
Product Testing & Quality Assurance
Generative AI can improve quality control processes in product manufacturing. Images of good quality and defective products can be used to predict whether a newly manufactured product is likely to be faulty. It can also help identify and implement corrective actions in defect detection and root cause analysis. Some use cases include:
- Investigation of production issues based on incident reports, customer complaints, and manufacturing process plans.
- Analysis and querying of various documents related to manufacturing specifications, inspection data, and defect descriptions to identify defect patterns and control quality.
- Generation of synthetic defect images to improve model performance for image based inspection and detection.
Product Maintenance & Support
Generative AI can deliver significant value in product maintenance and support functions. Potential use cases are:
- Plan maintenance and troubleshooting activity by searching through historical maintenance logs, product inspection manuals and maintenance manuals.
- Search product support knowledge repositories such as product specifications.
- Generate responses to customer inquiries, complaints, reviews, chats and emails.
Generative AI can influence the creation of new or next-generation products, improve product features, and troubleshoot existing products. It can help engineers and designers bring multiple aspects of product manufacturing, operations, and maintenance during the design phase. It can help realize circular and sustainable products.
However, there are certain things to watch out for. Generative AI can bring scale to the entire design process, but accuracy must improve. Further, the computation cost is still high, and there is a need to assure enterprises of the confidentiality of data and knowledge. The technology cannot now interpret multiple data sources like text, images, time series data and acoustics.
Despite these issues, Generative AI promises plenty and can transform engineering enterprises in the future.