AI Applications in Technical Risk Management
The application of AI in technical risk management has made significant progress recently. Here, we provide a brief overview of the current applications, organized by their chronological occurrence.
Field Data Analysis
Field data can be analyzed to detect “patterns.” Typical pattern categories include:
- Spatial patterns
- Temporal patterns
- Behavioral patterns
- …
The results of such evaluations are naturally subject to a high category of information protection. No company will share such insights. Therefore, this category does not involve “world knowledge.”
Development of Specifications (Design Spec.)
AI applications for developing the contents of specifications are possible. Analyzing extensive texts and generating content in this context can be significantly supported, qualified, and accelerated by AI applications.
Design/Configuration of Systems
A core part of FMEA development is the functional analysis. This can decompose customer features (functional networks). If necessary, this decomposition can extend to the design characteristics of components. The FMEA identifies knowledge gaps (equal to risks) in the system design through the associated error and risk analysis. These identified and specified gaps can be closed by applying AI systems. The AI systems apply known physical relationships (world knowledge) in combination with defined requirements and noise factors to provide specific guidelines for the design and tolerance of product features. Thus, AI systems can shorten development times and minimize the number of trials.
Process Design
The statements regarding design configuration can essentially be transferred to the development of robust processes. Here, too, it is about generating the necessary knowledge about the target values for product and corresponding process characteristics. The statements regarding design configuration can essentially be transferred to the development of robust processes. Here, too, it is about generating the necessary knowledge about the target values for product and corresponding process characteristics. The interaction of process characteristics with product characteristics is based on physical models, i.e., the possible influencing factors must be known, to which the FMEA contributes. An engineering AI can then create prediction models based on data from small training samples. With these project-specific target values, solutions can be found more quickly and cost-effectively to achieve process robustness.
The current time will later be considered the initial phase for AI applications. This means that no one can accurately predict at this point what roles and functions AI will take on in the various application areas of quality and risk management. It is the beginning of a “journey.” Therefore, it is sensible to observe developments and draw the necessary conclusions.
In the context of this topic, we cordially invite you to our free DC Talk Online Event: “FMEA with Engineering AI Functions – The Dream Team of the Future!” on September 19 from 10:00 to 11:30. Mr. Dipl.-Ing. Winfried Dietz, together with Mr. Dipl.-Ing. (FH) Frank Thurner from mts Consulting & Engineering GmbH and Contech Software & Engineering GmbH, will give an informative technical lecture. After the lecture, they will be available for questions and look forward to an engaging discussion with you. We hope to welcome you to this event.