Skip to main content

Multi-ObjecTive design Optimization of fluid eneRgy machines

Periodic Reporting for period 2 - MOTOR (Multi-ObjecTive design Optimization of fluid eneRgy machines)

Reporting period: 2017-03-01 to 2018-08-31

The performance of aircraft engines, ship propellers, water turbines and screw compressors, in the following summarized as fluid energy machines (FEM), crucially depends on the shape of propellers, rotors and fans, which are described by functional free-form surfaces, and the housing design. Slight shape variations can lead to significant changes in the overall machine performance. Likewise, approximation errors created by inaccurate computer simulations can lead to drastic mispredictions of the performance and, in turn, to sub-optimal FEM designs. However, optimised FEM designs with reduced emission of CO2/NOx and noise are urgently needed to enable greener aviation and shipping. Likewise, hydropower plants and compressed air systems with improved efficiency are imperative to compensate the ongoing growth of the world’s energy consumption.

The vision of MOTOR is to link all computational tools to the same master geometry to enable geometry-aware simulation-driven multi-objective design optimization of FEMs. To reach the set-out vision, novel concepts for geometry modelling and multi-physics mesh generation, stronger integration of computer-aided design (CAD) and engineering (CAE) tools and new hybrid methods for performing optimization ‘in the design space’ have been developed as modular components and integrated into the partners’ workflows.

The work performed in MOTOR was organized around the following three objectives:
1) Create automatic design optimization tools for FEMs.
2) Improve the effectiveness of design workflows for FEMs.
3) Improve the efficiency of FEMs.

All three objectives have been reached within MOTOR bringing a technology boost to the participating industry partners and enabling the academic partners to advance the state of the art in different disciplines.
MOTOR created novel CAE technologies to accurately model the complex geometries of the four industrial use cases and perform geometry-aware simulation and optimization. This includes the development of adaptive THB-splines, a template mapping approach and free-from deformation techniques for creating and manipulating volumetric parameterizations as well as hybrid volume meshes without intensive manual work. Moreover, structural mechanics and fluid flow solvers have been implemented in the open-source G+Smo ( library in the framework of Isogeometric Analysis forming the basis for geometry-aware multi-physics simulations. These tools are complemented by two new optimization strategies: The CAD-based adjoint method complements the idea of geometry-aware simulation and enables gradient-based optimization with respect to the design parameters; the hybrid approach combines gradient-based and gradient-free methods and enables efficient exploration of large design spaces without ending up in local optima.

The new CAE technologies have been integrated into the partners’ workflows. Demonstrators have been developed with the aid of the new tools and the improvement of the product qualities has been analyzed. Moreover, the effectiveness of the modernized workflows in terms of reduced amount of manual work, shorter time-to-production, and reduction in computational resources needed has been assessed.

Computing times and the amount of manual work have been significantly reduced in all four use cases. The improved geometry modelling and optimization techniques have enabled the creation of an improved design of a particular part of a commercial aircraft engine. Likewise, the free-form deformation approach was successfully employed to reduce the preparation and calculation time making high-fidelity automated optimization of ship propellers attractive for industrial usage. An optimization loop based on low-fidelity methods has been developed and demonstrated to produce competitive propeller designs at significantly reduced computational costs thereby reducing the time-to-production by orders of magnitude. A brand-new design, analysis and multi-stage optimization workflow for Kaplan and Francis water turbines has been realized in the framework of Isogeometric Analysis that reduces the amount of manual work and generates improved turbine designs. Finally, new and innovative computational tools for modeling and optimizing screw rotor geometries – with constant and variable lead – have been developed and used to create the world’s first air compressor with dual-lead rotor technology delivering a considerable efficiency improvement as compared to constant lead machines.

The outcome of the MOTOR project has been disseminated to the different target groups in academia and industry by more than 110 papers, presentations, posters, training materials and press-releases. Several special dissemination events with high impact in the particular communities have been organized by the MOTOR partners. These include the VKI lecture series on Optimization and Multidisciplinary Design 2016 and 2018, the VKI workshop on Turbomachinery Aerodynamics and Multidisciplinary Optimization 2016, a booth at Daimler EDM CAE Forum 2017, the 3rd International Conference on Isogeometric Analysis and Applications 2018, and the 10th International Conference on Screw Machines 2018. To disseminate the results among the MOTOR consortium partners, three G+Smo developer days have been organized in 2015, 2017 and 2018.

The modernized design workflows and the resulting products will be exploited commercially by the relevant project partners. Company ESS integrated the improved mesh generation capabilities into their MERGE software. Geometry modelling and IgA-based simulation and visualization techniques are available in the open-source G+Smo library, whose professional development will continue after the end of the project. The MOTOR results co
The vision of MOTOR – to enable geometry-aware simulation-driven multi-objective design optimization for four industrial use cases – has been achieved in the project time and pushes CAE technology beyond the state of the art.

A major achievement is the development of a fully “in-paradigm” isogeometric workflow ranging from geometry modelling based on adaptive spline technologies over multi-disciplinary CFD/CSM simulation tools to interactive real-time visualization of results. This enables new research directions and strengthens acceptance of Isogeometric Analysis in industry via future IgA-enabled commercial CAE packages. Further achievements are a CAD-based adjoint optimization approach, which is the first of its kind, and the hybrid multi-level optimization strategy, which allows to significantly accelerate convergence to globally optimal solutions.

Socio-economic impact is achieved via the partners’ modernized workflows, which will change the way of engineers working and thinking. Industry partners indicate that they expect an increase in the number of different designs effectively keeping the number of employees at the same or even a slightly increased level.

Product improvements strengthen the competitiveness of Europe’s FEM industry and improve the environment due to reductions in energy consumption and the emission of CO2/NOx and noise.
WP7: Multi-patch parameterisation of water turbine geometry
WP6: Illustration of mesh deformation of a propeller blade using Free-Form Deformation
MP2: Ship Propellers
WP7: Multi-patch parameterisation of water turbine geometry
WP3: Visualisation of the deformation of a screw machine by the new visualisation environment
MP4: Screw Machines
WP2: Illustration of adaptive THB-spline technology
WP7: Multi-patch parameterisation of water turbine geometry
WP3: Simulation results for a water turbine (velocity streamlines)
WP3: Simulation results for a water turbine (velocity streamlines)
MP3: Water Turbines
MP1: Aircraft Engines
WP3: Illustration of combined visualisation by the new visualisation environment
WP2: Geometry-aware meshing of a blade geometry from CAT by the MERGE software
WP7: Water turbine geometry