Research On Shape Error Analysis Methodology In Multi-point Forming

Multi-point forming (MPF) is a flexible forming technology for 3D sheet metal. Its basic idea is to make traditional mold discrete, using a matrix of adjustable discrete elements which arrayed regularly to approach"flexibility Multi-point die"to combine forming surface. The MPF technology can be applied extensively in a lot of fields, such as aerospace, shipbuilding, vehicle plate and so on.There is a shape error between workpiece form and objective model due to springback and positioning error in the process of MPF, how to get the shape error between workpiece form and objective model is one of key problems in improving forming precision of MPF.The data in this paper is got from the measurement of laser scanning, with preprocessed technique the noise elimination of data is through random filtering. The measured data is reconstructed with B-spline method. ICP algorithm and geometric hashing algorithm are particularly introduced, the geometric hashing algorithm is firstly used to the registration of data, the result of registration is the initial value of ICP algorithm. The optimal registration of workpiece form and objective model is got through the combination of these two algorithm, the error analysis is done through the comparison of workpiece form and objective model. The main contents and results are as follows:In this paper, the measured data is measured with Laser scanning measurement machine and reconstructed with B-spline method. Random filter is applied to improve the influence of noise data. Then the surface denoised is output in IGES format to transfer data to other CAD software or MPF CAD software. The shape error between workpiece form and objective model is calculated with the method combination of ICP algorithm and geometric hashing algorithm. Geometric hashing algorithm is not fit for high precision registration because of the inevitable range error in the field in discretization of it. ICP algorithm, which can obtain high precision results, is simple but requires good initial positions. Therefore, the result calculated by geometric hashing algorithm is used for ICP algorithm. The registration algorithm is checked by ideal curved surface. The performance is more excellent than the ICP algorithm, the algorithm is implemented by visual C++ and OpenGL, feasibility is proved by the error analysis of curved surface.In this paper, the author analyzed the error of Multi-point forming through the data preprocessing, the surface reconstruction and the surface registration, solved the problem appeared in the production practice. It is proved by some examples using data preprocessing and surface registration technique to implement accuracy control to be feasible, and possesses definite use value.