Search results

The purpose of this paper is to explore a methodology that allows to represent turbomachinery rotating parts by replacing the blades with a body force field. The objective is to capture interactions between a fan and an air intake at reduced cost, as compared to full annulus unsteady computations.

The blade effects on the flow are taken into account by adding source terms to the Navier-Stokes equations. These source terms give the proper amount of flow turning, entropy, and blockage to the flow. Two different approaches are compared: the source terms can be computed using an analytic model, or they can directly be extracted from RANS computations with the blade’s geometry.

The methodology is first applied to an isolated rotor test case, which allows to show that blockage effects have a strong impact on the performance of the rotor. It is also found that the analytic body force model underestimates the mass flow in the blade row for choked conditions. Finally, the body force approach is used to capture the coupling between a fan and an air intake at high angle of attacks. A comparison with full annulus unsteady computations shows that the model adequately captures the potential effects of the fan on the air intake.

To the authors’ knowledge, it is the first time that the analytic model used in this paper is combined with the blockage source terms. Furthermore, the capability of the model to deal with flows in choked conditions was never assessed.

Geometric errors are common in metallic bent tubular parts. Thus, tubes should be inspected and fixed before welding with the joints first. After welding, the relative position of the joints is also necessary to be inspected to judge whether the tube can be assembled reliably. Therefore, the inspection plays an important role in the tube’s assembly. The purpose of this paper is to propose a multi-vision-based system designed to inspect the tube and the relative position of the joints.

For the tube inspection, the small cylinders are taken as the primitives to reconstruct the tube using the multi- vision-based system. Then, any geometric error in the tube can be inspected by comparing the reconstructed models and designed ones. For joints’ inspection, authors designed an adapter with marked points, by which the system can calculate the relative position of the joints.

The reconstruction idea can recognise the line and arc segments of a tube automatically and resolve the textureless deficiency of the tube’s surface. The joints’ inspection method is simple in operation, and any kinds of joints can be inspected by designing the structure of the adapters accordingly.

By experimental verification, the inspection precision of the proposed system was 0.17 mm; the inspection time was within 2 min. Thus, the system developed can inspect a tube effectively and automatically. Moreover, authors can determine how the springback of the arcs behaves, allowing in-process springback prediction and compensation, which can reduce geometric errors in the tubes given the present bending machine accuracy.