Project Details
Figure 1: Position of the QVT supply line in the machine.
In order to achieve the most uniform fiber distribution possible after a cross-flow distributor, the suspension should already have a uniform flow profile at the inlet (see Figure 1). This is made more difficult by the bend located directly in front of it. As part of this project, we compared both options with which geometry optimizations can be implemented in our simulations.
In all projects, we use a simulation of the geometry currently in use as a starting point in order to identify the critical sections in the flow pattern (see Figure 2). In this simulation, the RANS equations are solved and, in particular, the shear rate-dependent viscosity behavior of the pulp suspension is mapped in order to obtain representative results. The comparability of the various geometries is ensured by the fact that the boundary conditions in Figure 4 are defined identically for each of the simulations.
Figure 2: Flow velocity distribution in the initial geometry.
Figure 3: Flow velocity distribution in the optimized geometry.
There are then two possible approaches: an analytical parameter study and an optimization algorithm. In this specific case, the changes were concentrated on the manifold. In an analytical approach, parameters relevant to the design, such as diameter changes, radii of curvature or inlet and outlet lengths, are systematically varied until the desired flow pattern is achieved. With clearly defined parameter ranges and only a few independent parameters, this quickly leads to the desired result. However, if your problem is more complex, as was the case here, an analytical solution would require many variants and a lot of time, without necessarily arriving at the optimum result.
This is where our optimization algorithms come into play, which can incorporate a wide range of dependencies. In the case of the pipe bend, a solution using a free surface provided the best result. This forms an arbitrary geometry that provides the ideal flow pattern. However, as this cannot be produced at all or only at great expense, our engineers simplify these solutions into a design suitable for production, as shown in the sequence of images in Figure 5. This geometry then provides the best possible flow pattern, which can be seen in Figure 3.
Figure 4: Boundary conditions of the calculation domain.
The disadvantage of the optimization algorithm is the development time required to apply it to the respective project. Therefore, the decision on the optimization method for your project is always a trade-off between the desired result, the available time frame and the design scope of our employees.
Figure 5: Optimization via the algorithm.
In our projects, it is particularly important to us not only to realize the ideal flow behaviour, but also to take into account the structural limitations of the system and the simplest possible production, so that the optimization remains advantageous across the board. In this case, the improved flow pattern can reduce the required outlet length and thus open up new optimization possibilities even in existing systems.
