Toward fixtureless inspection of automotive fenders
Analytics
56 views ◎29 downloads ⇓
Abstract
With nonrigid parts, it is most convenient for designers to specify the desired shape in the design condition; that is, the shape is specified with all loads present that the part will experience in service (gravity, assembly constraints, etc.). The flexibility of the part begins to pose challenges for the dimensional inspection of the part profile when the deformations due to design loads exceed 10% of the dimensional tolerances. A common approach to negotiate the inspection of nonrigid parts is to construct an inspection fixture that mimics the design condition (identical mounting points and orientation to gravity). Although effective, inspection fixtures have the limitation of cost, calibration maintenance, procurement time, and the inspection is limited to a small subset of parts. This thesis builds on the fixtureless inspection literature which has emerged in response to these limitations.The proposed fixtureless inspection method uses finite element simulations to adjust the nominal design shape into the fixtureless measurement condition. Finite element simulations are used to remove deformations from the nominal shape due to design condition loads and add deformations due to measurement condition loads. In this way, the part can be inspected under different constraints than the design condition and without specialized fixtures. The details of the method are outlined including the finite element simulations (using Abaqus), conditioning of the finite element mesh (using MATLAB®), optical point cloud acquisition (using a ROMER Absolute Arm with integrated scanner), and the processing of the point cloud data (using MATLAB®).This method is first demonstrated on a cantilevered flat plate where the design condition is defined with a mass hanging on the end and the measurement condition is defined under the influence of gravity only. The proposed method is used to measure the profile deviation of the plate in the measurement condition (gravity only). For validation, the profile deviation is also measured in the design condition (mass hanging on the end). The profile deviation from the measurement condition is shown to match the directly measured profile deviation from the design condition to within 25 μm. This is two orders of magnitude lower than the 3 mm design condition deflections of the cantilevered plate.The method is extended to an automotive fender where an original modal decomposition technique is used to deform the nominal model to the measurement data. The modal decomposition compensates for deformation of the part during assembly and provides a means to predict the required assembly forces. The profile deviation is measured using the proposed method with the fender in a fixtureless state resting on a flat table. For validation, the profile deviation of the fender is also measured in a specialized fixture to hold the fender in the design condition. The two profile deviations match within 0.6 mm, more than an order of magnitude lower than the 10 mm measurement condition deflections of the fender. This thesis provides one of the first automotive examples of fixtureless inspection and offers improved computational efficiency as finite element simulations for every measured part are not required.