Periodic Reporting for period 1 - InSoMa (Feasibility study for lancing an innovative software tool based on mathematical methods in the B2B market)
Berichtszeitraum: 2019-07-01 bis 2019-12-31
NAVASTO is a technology company focusing on engineering services for Computational and Experimental Fluid Dynamics (CFD/EFD), including workflow and software development. Founded in 2014 in Berlin with only two employees, we are currently looking for number ten to join our quickly growing team of experts.
Our SME instrument phase 1 project focused on determining the feasibility of establishing an innovative engineering tool based on state-of-the-art mathematical methods in the business to business market. The software product envisioned addresses a problem common to many companies in high-tech industries worldwide: Large amounts of simulation and experiment data are produced during the design process - constantly increasing due to the wide availability of high-performance computing capacities. A lot of this data perishes into the abyss of data graveyards. Our software uses machine learning, reduced order modelling and interpolation methods to enable engineers to exploit the entirety of the available data sets and data sources. With those methods, novel optimization strategies, such as Surrogate Based Optimization, can be used to significantly speed-up the design process and validation strategies can be improved by applying data fusion for numerical and experimental data to enhance the validity of the result.
The software will enable customers to reduce time to market, reduce development costs and increase the quality of their products by contributing to the tools necessary for the digitalization of the design process.
Within this project we addressed three aspects relevant to the further development of our software, namely technical, economic and legal aspects.
Our SME instrument phase 1 project focused on determining the feasibility of establishing an innovative engineering tool based on state-of-the-art mathematical methods in the business to business market. The software product envisioned addresses a problem common to many companies in high-tech industries worldwide: Large amounts of simulation and experiment data are produced during the design process - constantly increasing due to the wide availability of high-performance computing capacities. A lot of this data perishes into the abyss of data graveyards. Our software uses machine learning, reduced order modelling and interpolation methods to enable engineers to exploit the entirety of the available data sets and data sources. With those methods, novel optimization strategies, such as Surrogate Based Optimization, can be used to significantly speed-up the design process and validation strategies can be improved by applying data fusion for numerical and experimental data to enhance the validity of the result.
The software will enable customers to reduce time to market, reduce development costs and increase the quality of their products by contributing to the tools necessary for the digitalization of the design process.
Within this project we addressed three aspects relevant to the further development of our software, namely technical, economic and legal aspects.
With respect to the technical dimension of the project, we implemented a new interface to a primary data source (ESI’s PAM-CRASH software) and produced a surrogate based optimization application that was successfully tested on a real-live use case of an automotive OEM. The legal dimension, which focused on licensing aspects of the software, addressed the selection of a suitable licensing model including the draft of a preliminary license agreement that can now be used for the collaboration of our development partners in different industries. Finally, the work performed with respect to economic aspects helped us to clearly identify our main potential customers, develop our pricing strategy and to determine the ideal distribution channels.
The main achievement beyond the state of the art is the proof of concept of the usability of the mathematical methods on an industry case. In contrast to most academic demonstrations of the methods, the data we worked with was noisy and sparse. Nonetheless our methods proved useful for the optimization of a real live problem and therefore the project contributed not only to increasing the confidence in the methods themselves, but also to the maturing of the technology towards industry relevant cases