McKinsey identified rapid experimentation and simulation as one important industry 4.0-lever to shorten time to market in their article Digital in industry: From buzzword to value creation.
The logic is that by introducing simulations earlier in the product design process, bad product design can be eliminated at an earlier stage and the development and design iterations become faster. Thus, can time to market be reduced. Flow simulations are complex but extremely valuable. Predicting flow can give an aerospace engineer valuable knowledge on which design minimises fuel consumption and noise generated from landing gear, or where an architect should place the ventilation in an office building to achieve the best level of comfort.
But today there is a significant competence threshold to performing simulations. With few exceptions, only the experienced CFD engineers of today are able to perform accurate flow simulations, for example. Even if design engineers had the competence to conduct a CFD simulation, the time they would have to invest in this is considerable. It is not unheard of that an experienced CFD engineer can spend one to several weeks preparing a mesh, refining it after iterations and post-processing the result before even starting to analyse the result of the simulation.
In order to introduce simulations earlier on in the design process, the cost and competence threshold need to be reduced significantly. This can only be achieved by introducing fully- automated flow simulation capability- which preferably is integrated into the existing PLM-software (Product Lifecycle Management software), or product design software.
We are currently on the homestretch of developing such capability to the industry. Our solution for Automated Flow Prediction enables engineers, product designers and innovators to simulate and compare different designs (and optimise them) early in the design process, without any prior knowledge or experience from CFD. By eliminating almost all manual steps, CFD engineers can use our tool to liberate time from tedious tasks and produce value-creating analysis of the simulation results.
Our user interface is intuitive and extremely easy to use. The simulations are accurate and powerful with computations performed on a supercomputer. We offer this capability as a flexible SaaS solution without any need for IT or educational investments. Simply put: we make simulations easy!
The groundbreaking algorithms in our model are adaptive and parameter-free, which enables an automated simulation process as well as a reliable estimation of error margin. Our innovation is the result of a spin-off from world-leading research conducted over the last decade by a team of researchers at KTH Royal Institute of Technology in Stockholm.
The method is based on a quantitative a posteriori error estimation, which means that the user prescribes what they want to compute, and then the mesh is automatically optimised to compute that quantity to a prescribed tolerance with minimal computational cost. This is achieved by solving a so called adjoint problem that connects local errors to the error in the quantity of interest.
If you would like to help us improve our service by providing feedback on functionalities and user interface, then please join our upcoming beta phase (free of charge during the beta). We also welcome pilot customers who wish to customise an app in our solution, or who need our help solving a specific problem.
If you have any questions, please feel free to send us a message!