文字のサイズ
- 小
- 中
- 大
Denso to Maximize Use of Accumulated Information with Specialized AI to Vertically Launch Production Equipment Development
DENSO Corporation continues to make changes to realize a safer and greener future of mobility under the slogan “Delivering a future of smiles to the earth, society, and all people” in its “2030 Long-Term Policy”. DENSO Corporation continues to make changes to realize a future of safer and more environmentally friendly mobility under the slogan “Delivering a future of smiles to the earth, society, and all people.
The pursuit of highly efficient and high-quality manufacturing is the driving force that creates value, and in order to realize this in a wide variety of products on a global scale, the development of production facilities is a particularly important theme. The company’s production facility development is facing increasingly complex product specifications, accelerating business expansion, and growing demand for shorter lead times when starting up mass production lines, and the requirements for facilities are rapidly becoming more diverse and sophisticated. In addition, the aging and retirement of skilled designers is diluting the technical succession and creating a shortage of design resources. To address these issues, the company has embarked on a design transformation aimed at vertically launching new production facilities in its Industrial Machinery Department, with the aim of delivering high-quality products more rapidly by further strengthening product quality and its global production system.
Evolving the advanced technical experience and tacit knowledge of design review into a mechanism through process innovation and AI
The equipment design work of the company’s Industrial Equipment Department is thoroughly standardized and already guarantees a high level of quality and efficiency, but there is an urgent need to address the increasingly sophisticated and complex requirements of production equipment and the speed of change. The roles and requirements of each region, plant, and production line are also becoming more diverse under the expansion of global production systems, the replacement of business portfolios in line with the development of CASE, and efforts to achieve carbon neutrality. In addition to these influences, equipment development tends to be individualized design specific to industrial machinery, with many variations depending on the combination of production applications, specifications, destination requirements, etc., increasing the number of specification adjustments each time and increasing the amount of coordination by the designer. As measures to meet these requirements were implemented on a case-by-case basis, there were many rework issues in terms of quality during equipment start-up. In order to address the increasingly sophisticated and complex requirements for production facilities, it is particularly important to identify all technical issues and ensure the appropriateness of countermeasures during design reviews (DR). To strengthen the company’s ability to respond to these issues, the Engineering Department’s Reform Team initiated drastic reform activities in DR operations, which are the key to the quality gate.
The DR that the company is implementing refers to the following sophisticated series of business processes.
The DR process involves the identification of numerous prerequisites and requirements, including product development plans, production plans, production site specifications, destination requirements, production line concepts, process capacity requirements, and the company’s major policies.
We identify novel technologies and related technical risks, and incorporate them into design specifications and countermeasures, taking into account the degree of impact.
While the company has steadily accumulated a variety of DR-related study concepts, principles, and technical standards, the “core know-how” for considering and reporting on these standards in an integrated manner had become “tacit knowledge of experienced designers and belonged to them,” and “past knowledge that would serve as material for technical studies could not be consolidated and utilized. The workload of searching and creating documents was heavy (there were many variations, and even skilled designers had not fully experienced some patterns and information),” and “the know-how to point out DR depended on the experience of each reviewer,” which made it difficult to fully mature the quality of the review process.
The team held a series of discussions based on the actual business situation at the worksite, and came up with the following ideas as a goal for the design of production facilities,
Thoroughly consolidate and utilize knowledge at the technical review stage, and comprehensively identify concerns and countermeasures from the point of change or modification.
The team discussed the following points based on the actual situation of work in the field, and came up with the following vision for the design of production facilities: – Improvement of the quality and efficiency of the technical review as a mechanism, not by human resources, so that designers can focus on more essential technical review (≒opening up the potential of engineers).
Based on the premise of this vision, we steadily promoted essential human-centered DX that goes beyond mere tool changes and system implementation.
Process restructuring, starting with the conversion of DR tacit knowledge into formal knowledge
While the concepts, principles, and technical standards used in the existing DR process were sufficient as a background and basis for thinking, the skill level and experience of the designers in charge and the number of man-hours required to execute the work using them were a bottleneck, making it impossible to use and complete the work.
Therefore, the transformation team implemented measures aimed at creating a system that anyone can use in their work in line with practical operations, while formalizing the tacit knowledge of skilled designers and quality assurance professionals.
Restructuring the business process from DR to QA
The transformation team started by fundamentally restructuring the DR process as a QA (Quality Assurance, hereafter) process. The team promoted a business process reform that merged the traditional “DR: a design-centered place where concerns and countermeasures are fully addressed” with “QA: a place where the design and quality assurance departments maximize their respective audit functions and technical perspectives to guarantee certainty from the upstream stage. We also established a system that anyone can carry out in practice, which includes the “development of standards for QA processes and gates,” the “development of the order and methods for using the standards,” and the “development of operations for using the standards.
2. restructuring of change point management and concern point extraction operations
Based on the structure reconstructed in 1. above, for the core area of “reviewing highly relevant concerns based on changes/changes in facility specifications,” we made maximum use of the knowledge of experienced staff, developed various formats, meeting bodies, and operational procedures, and incorporated them into a reproducible work process.
AI recommends just-in-time concerns and countermeasures based on changes/changes, enhancing exhaustive risk examination
In the core task of extracting concerns, the transformation team linked “a vast amount of actual information identifying root causes and countermeasures for past quality rework” with “natural language processing AI” to comprehensively provide information on highly relevant concerns and countermeasures from changes in equipment specifications on a just-in-time basis (i.e., the necessary information is provided to the necessary people at the necessary time). We have established a system that recommends information on highly relevant concerns and countermeasures based on changes and changes in facility specifications in a comprehensive and just-in-time manner.
SpectA DKM” was adopted as the core natural language processing AI solution, and by learning the large amount of technical information accumulated by the company’s Engineering Department, the tacit knowledge and experience of skilled workers were converted into digital assets, enabling anyone to perform the task of extracting concerns efficiently.
All the designer has to do is input “information on specification changes” and select “unit/function,” and the AI specialized for equipment quality recommends axes of priority conditions for additional consideration, such as “mechanism structure” and “components,” to support information extraction, and displays “quality trouble examples” and “anticipated risks and countermeasure know-how” in order of relevance. The information recommended by the AI is presented in an easy-to-understand manner, including a summary of problem factors and measures to deal with possible risks, and image information, etc., to encourage comprehensive risk studies and countermeasures while raising the level of technical knowledge of each individual (promotion of practical self-review).
Transcending the structure and changing the mindset of employees through the power of true change by thoroughly confronting people.
The digital transformation that the team is promoting is not limited to simply introducing the latest digital technology or restructuring business processes, but also emphasizes the importance of using this as a catalyst to “change the mindset of employees” and “create a culture of change in the organization. We are steadily building on small successes by gaining an accurate understanding of the actual business situation onsite, carefully communicating in our own words the significance, purpose, and goals of the transformation, and rapidly repeating consensus building and system improvements in close cooperation with the onsite staff so that each and every employee can work on the transformation as if it were their own.
As a result of these activities, not only has the quality of production facility design been enhanced, but also man-hours have been reduced by improving work efficiency and reducing rework, which has contributed to shifting designers to more creative and value-added work and unlocking the potential of engineers.