Community Research and Development Information Service - CORDIS

H2020

MOTOR Report Summary

Project ID: 678727

Periodic Reporting for period 1 - MOTOR (Multi-ObjecTive design Optimization of fluid eneRgy machines)

Reporting period: 2015-09-01 to 2017-02-28

Summary of the context and overall objectives of the project

The creation of competitive product design workflows for fluid energy machines requires the effective interaction and integration of various computer-aided technologies (CAx). R&I actions have created many smart materials (shape-memory alloys, piezo-structures) and (additive) manufacturing (CAM) technologies (3D printing, laser powder forming), which have the potential to revolutionize manufacturing processes, especially by enabling rapid prototyping, and to create innovative products. The full potential of novel CAM technologies is, however, enchained by sub-optimal design (CAD) and engineering analysis (CAE) workflows, which have become the major bottleneck of the entire process chain.

The functionality of fluid energy machines depends on the shape of propellers, rotors, fans and the housing design. These geometries are described by functional free-form surfaces, which are constructed indirectly with the aid of design parameters like thickness distributions and blade angles. The fact that already a slight variation of the resulting CAD shape can lead to a significant change in the overall performance is the grand challenge for the CAE toolchain. The accurate solution of multi-physics problems requires the exact capturing of the interface between the solid and the surrounding fluids. A poor approximation of one sub-problem propagates through the interface to other sub-problems and pollutes the entire simulation result, causing the optimiser to produce sub-optimal designs. Shape optimisers are most effective if they have access to the exact geometry and its generating design parameters instead of the discrete surface approximation.

The vision of MOTOR is to link all CAE tools to the same accurate representation of the master geometry to enable geometry-aware simulation-driven multi-objective design optimisation of fluid energy machines. To reach the set out vision, novel CAE technologies are developed as modular components and integrated into existing design workflows for:

Aircraft Engines
Ship Propellers
Water Turbines
Screw Machines

The overall objectives of MOTOR are:
- To create automatic multi-objective design optimisation technologies for fluid energy machines to enable integrated design workflows.
- To improve the effectiveness of the integrated design workflows in terms of shorter times to production and increased mass-customisation capacity.
- To improve the efficiency of fluid energy machines.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

MOTOR created CAE technologies to accurately model the geometries of the four use cases and perform geometry-aware simulation and optimisation. To maximise tool reusability and exploitation potential, the open-source Geometry + Simulation Modules library (gs.jku.at/gismo) has been selected for common software development.

Geometry Modules: JKU extended the theory of multivariate adaptive splines (THB-splines) and improved the efficiency of their implementation. Moreover, a template mapping approach has been devised and combined with an existing isogeometric segmentation pipeline for the creation of multi-patch domain parameterisations. The advantages of this approach in terms of higher flexibility in geometry modelling, and higher accuracy at only a fraction of the computational costs of tensor-product B-splines has been evaluated for demonstrator geometries (MTU, TUDO). Worth mentioning is the unplanned development of a THB-spline plugin for the geometry modelling tool Rhino by MARIN. The MERGE meshing tool (ESS) has been extended by mesh repairing and improving algorithms to enable the creation of high-quality surface and volume meshes for fluid energy machines, thereby ensuring that meshes for multi-physics simulations match perfectly at the interfaces.

Simulation Modules: Solvers for elasticity problems and incompressible and compressible fluid flow problems have been implemented (UNIKL, UNI PILSEN, TU Delft) in the framework of isogeometric analysis (IgA), forming the basis for the geometry-aware simulation of the water turbines and screw machines. A highlight is the GPU-accelerated visualisation environment (UNIKL). It enables interactive visualisation of huge IgA-data sets at unparalleled level of detail and accuracy. Moreover, a multi-domain non-uniform rational B-splines (NURBS) coupling approach has been realised (VKI), which makes it possible to simulate multi-physics problems without introducing interpolation errors at domain interfaces.

Optimisation Modules: Two new optimization strategies have been developed by VKI. The CAD-based adjoint method complements the idea of geometry-aware simulation and enables gradient-based optimisation with respect to the design parameters that control the CAD shape. The computational cost of the adjoint-based method is practically independent of the number of design variables enabling fine-grained shape optimisation. It is combined with gradient-free methods in a novel hybrid multi-level algorithm, which enables fast optimisation cycles and, at the same time, extensive exploration of the entire design space.

Integration of the new tools into the design workflows has started.

Aircraft Engines: MTU has integrated THB-splines and the template mapping approach into their workflow. As a first assessment, a turning-mid-turbine-frame has been re-designed and optimised by the in-house toolchain leading to significant reductions in computing times.

Ship Propellers: MARIN has developed a free-form deformation (FFD) approach that will enable efficient high-fidelity propeller optimisation. Initial tests demonstrate that the volumetric deformation method can be applied to generic propeller surfaces together with existing propeller meshes without reducing the quality of the deformed mesh. The FFD approach reduces the initialisation time of a single calculation by days and makes high-fidelity automated optimisation of ship propeller attractive for industrial usage. An assessment of the new tool by the industrial partner CAT is underway.

Water Turbines: UNI PILSEN and MAVEL created tools for the automatic generation of B-spline/NURBS surface and volume models of Kaplan and Francis turbines and devised methods for the generation of initial turbine designs. The hybrid multi-level optimisation approach described above will be adapted and combined with the developed IgA-solvers in the second period.

Screw Machines: TUDO has created a new chamber model generator which can create initial design

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Geometry-aware simulation-driven multi-objective design optimisation for industrial use cases goes beyond the state of the art.

It is expected that the industrial partners will largely benefit from the modernisation of their design workflows in terms of shorted development times (due to the automated optimisation), higher flexibility in product optimisation (due to more powerful CAE technologies) and, finally, the creation of products with increased performance.

Related information

Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top