pSeven User Conference 2022 | VIRTUAL 

This presentation attempts to demonstrate the contribution of reproducible research in MDO/AI with several recent open source attempts of Prof Morlier’s group. During the last few decades, surrogate modeling has gained in popularity, especially in engineering fields, where they are often used in design analysis and optimization to replace expensive numerical simulations. Coupled to Multidisciplinary Design Optimization (MDO) this process lead to an engineering design acceleration through AI.

The first part of the talk will present some novelties in AI for Engineers such as the Surrogate Modeling Toolbox (SMT developed conjointly between ONERA, ISAE-SUPAERO/ICA, Nasa, and University of Michigan). SMT is a Python package that contains a collection of surrogate modeling methods (Mainly Kriging also known as Gaussian Processes in Machine Learning community), sampling techniques and benchmarks. SMT is different from existing surrogate modeling libraries because of its emphasis on derivatives that can be used directly in MDO processes. It also includes surrogate models or options that are not available elsewhere: KPLS (Kriging with Partial Least Square) for automatic inputs dimension reduction, mixture of experts (with automatic clustering) for simulation codes that include singularities (such as buckling in structural mechanics) and mixed variable design space (continuous, discrete, and categorical). The 2022 version currently available propose new applications such as multi-fidelity approach, Bayesian unconstrained optimization (EGO), fully inter-operable and automated.

The second part will present some MDO applications in Aerospace sciences will be highlighted, including launcher and UAV, HALE optimal design especially in an environmental perspective.

The engineering in offshore industry is facing more and more challenges due to several factors: more design constraints, increase of the workload post-Covid, global cost reduction, etc… To overcome those issues, TechnipFMC developed Smart Pipeline Design ARENA, a digital web platform, powered by pSeven, which is developed for rigid pipeline engineering (flowlines and risers). The objective of such a platform is to improve engineering cost, engineering time, product robustness and cost. Within that context, ARENA is mainly developed playing on the following four main levers: (i) standardization, (ii) automation, (iii) optimization and (iv) uncertainty management.

ARENA remains a design platform which can be used by a large spectrum of users, i.e. from junior or less experienced engineers to technical experts. In the first usage scenario, the platform proposes standardized and well-framed micro-services gathered in two toolboxes, i.e. respectively for flowline and riser design. Such first usage mainly considers standardized and validated micro-services and does not require any key experience. For the second usage scenario, the platform offers all the freedom to the user to consider standardized tool but also to chain them for any automated process or to develop their own fit-for-purpose tools (including AI capabilities). The latter can be shared in a second step with the team for the benefit of the project.

This study aimed to optimally design a parametrically generated bumper geometry under operating conditions. By changing the bumper geometry at the specified boundary conditions, the mass of the structure, strength and critical lifetimewere examined. As the optimization target, it was determined to meet in minimum mass, maximum critical lifetimeand minimize the stress results where is below theyield stress. This study was carried out with the pSeven workflow; CATIA for parametric CAD model, Patran & Nastran for finite element analysis and nCode for calculating critical lifetimewere used.

A new steel ingot casting format had to be designed. Here is the target to propose a feasible and optimal design with reduced segregation and centreline porosity in the as-cast ingot and on the other hand increase the metallic yield. By means THERCAST® the evolution of the filling and the solidification can be simulated and the quality of the ingot can be predicted – given a specific combination of geometry, process parameters and thermal boundary conditions. To investigate the impact of the geometry attributes (the mould, the ingot, the riser, riser insulation), the thermal and physical properties (riser insulation, hot topping), process parameters (steel grade, mass flowrate during filling, superheat) and analyse the responses of the simulation such a study would consume more than six months. In the scope of this investigation pSeven® was used to perform an automated closed-loop workflow in which Inventor® (CAD) and THERCAST® are integrated. PSeven® performs the parametrisation (geometrical, process and thermal data), triggers the automatic generation of CAD as well as the setup (meshing and parameters update) and launching of the THERCAST® computation. The responses from the virtual sensors are automatically acquired and are the basis for the next iteration. Using the design space exploration it was possible to gain a holistic understanding of the parameters interactions and relevance. In the further step the performed surrogate based optimisation enabled to create an optimal design of the ingot. This is performed in a record time of less than three weeks.