"Using Generative Engineering ("Bots") applications [....], it is possible to produce the same results [...] with a time saving of over 80%, using at least 3 times less resources."
Jean Baptiste Chancerelle - Digital Catalyst @Renault Group
In order to make the engineering process more efficient and increase the performance of its automotive products, Renault has entered into a collaboration with the French startup Dessia Technologies to test and use its Generative Engineering software solution.
"The advent of advanced techniques in the information sector brings us two major elements: the ability to collect and process massive amounts of information - "Data" - and a range of powerful tools capable of interpreting and deciphering the wealth of information collected - "AI". We expect partners like DESSIA to help us equip ourselves with tools that capitalize on these new technologies, to serve the performance of our engineering." - Renault
Source: Essai et Simulation 151
In order to make the engineering process more efficient and increase the performance of its automotive products, Renault has entered into a collaboration with the French startup Dessia Technologies to test and use its Generative Engineering software solution.
"The range of tools represented by what we call AI, supported by new capacities to collect information, allows us to reinforce our developments with the experience we have acquired. This opens up new opportunities to explore the field of possibilities in record time." - Renault
The first contribution of Generative Engineering is located in the upstream phases of development, at a time when the expenses are still contained and when the definition hypotheses are numerous. The stake here is to ensure a fast and robust convergence between the requirements that the product must satisfy and the quality/cost/time target assigned to the project. The future performance of the product is largely determined during this period. Similarly, a number of experts are mobilized to ensure this convergence. When the work has been well done, this should limit the costly iterations that would be carried out later in the development.
The upstream project phases implement intra- or inter-system functional approaches whenever possible. For complex systems or inter-systems, it is not possible to correctly establish the impacts or incompatibilities of requirements choices between each system. Iterative loops are then used until a coherent whole is obtained, generally with a certain number of points of vigilance to be dealt with later. These iterations are costly in terms of expert knowledge and study time. Moreover, it is not possible to explore the whole field of possibilities in the time usually allocated to these upstream project phases. If a reference definition is put forward, it is the result of a laborious process and there is no guarantee that it is optimal.
The main ambition fixed with Generative Engineering is to associate very early "functional specifications and 3D architecture drawing" and to multiply the number of iterations of technical definition via generative AI algorithms in order to explore as soon as possible the field of possible solutions. In the scoping part of the project, the idea is that each system brings up its requirements and different definition hypotheses that meet them, linked together by a simplified physical model in such a way that it is possible to compile all these requirements and definition hypotheses into a single model.
Thanks to this model, it is possible to go through the design space very widely. In concrete terms, it is possible to carry out studies of the sensitivity of the requirements fields, analyses of the incompatibility of inter-system choices or even analyses of the consequences of local choices on the overall structure of the product. Analyzed and reprocessed, macro-energy balances, macro-quality/cost/deadlines balances, or a first list of potential hard points are presented to the project team and thus produce an optimized and enlightened reference project framework with two or three alternatives to cover opportunities or risks.
At the end of these Generative Engineering iterations, a technical definition deemed optimized with two or three alternatives is consolidated. The systematic exploration of the field of possibilities by engineering bots has made it possible to consolidate both the performance of the technical response while strongly limiting the time spent and the number of expert people involved.
This response is informed, chosen and robust so that it can limit the number of iterations in later stages of development, when any change is costly and involves a large number of actors. Finally, the use of Generative Engineering applications allows us to be agile in the evaluation of different architectures and different technologies.
