Performance improvement for optimization of the non-linear geometric fitting problem in manufacturing metrology

Product quality is a main concern today in manufacturing; it drives competition between companies. To ensure high quality, a dimensional inspection to verify the geometric properties of a product must be carried out. High-speed non-contact scanners help with this task, by both speeding up acquisition speed and increasing accuracy through a more complete description of the surface. The algorithms for the management of the measurement data play a critical role in ensuring both the measurement accuracy and speed of the device. One of the most fundamental parts of the algorithm is the procedure for fitting the substitute geometry to a cloud of points. This article addresses this challenge. Three relevant geometries are selected as case studies: a non-linear least-squares fitting of a circle, sphere and cylinder. These geometries are chosen in consideration of their common use in practice; for example the sphere is often adopted as a reference artifact for performance verification of a coordinate measuring machine (CMM) and a cylinder is the most relevant geometry for a pin-hole relation as an assembly feature to construct a complete functioning product.In this article, an improvement of the initial point guess for the Levenberg–Marquardt (LM) algorithm by employing a chaos optimization (CO) method is proposed. This causes a performance improvement in the optimization of a non-linear function fitting the three geometries. The results show that, with this combination, a higher quality of fitting results a smaller norm of the residuals can be obtained while preserving the computational cost. Fitting an 'incomplete-point-cloud', which is a situation where the point cloud does not cover a complete feature e.g. from half of the total part surface, is also investigated. Finally, a case study of fitting a hemisphere is presented.

[1]  D. Marquardt.  An Algorithm for Least-Squares Estimation of Nonlinear Parameters , 1963 .

[2]  G. Cheng,et al.  On the efficiency of chaos optimization algorithms for global optimization , 2007 .

[3]  Robert G. Wilhelm,et al.  Task Specific Uncertainty in Coordinate Measurement , 2001 .

[4]  Jean-Pierre Kruth,et al.  Uncertainty determination for CMMs by Monte Carlo simulation integrating feature form deviations , 2009 .

[5]  Craig M. Shakarji,et al.  Least-Squares Fitting Algorithms of the NIST Algorithm Testing System , 1998, Journal of research of the National Institute of Standards and Technology.

[6]  P. Neittaanmäki,et al.  Fixed domain approaches in shape optimization problems , 2012 .

[7]  Blaine Lilly,et al.  Mechanical Assemblies: their Design, Manufacture, and Role in Product Development , 2013 .

[8]  Guojin Tang,et al.  Hybrid approach for solving systems of nonlinear equations using chaos optimization and quasi-Newton method , 2008, Appl. Soft Comput..

[9]  Theodore B. Trafalis,et al.  A novel metaheuristics approach for continuous global optimization , 2002, J. Glob. Optim..

[10]  P. E. Castro.  Compact Numerical Methods for Computers: Linear Algebra and Function Minimization , 1978 .

[11]  Stefano Petrò,et al.  Early cost estimation for tolerance verification , 2011 .

[12]  A. Weckenmann,et al.  Productive Metrology - Adding Value to Manufacture , 2005 .

[13]  Mohammad Saleh Tavazoei,et al.  An optimization algorithm based on chaotic behavior and fractal nature , 2007 .

[14]  Stephen Jay,et al.  Case Studies (3) , 2014 .

[15]  Bing Li,et al.  Optimizing Complex Functions by Chaos Search , 1998, Cybern. Syst..

[16]  Wahyudin P. Syam,et al.  Comparison of Three Meta Heuristics to Optimize Hybrid Flow Shop Scheduling Problem with Parallel Machines , 2010 .

[17]  Tony DeRose,et al.  Surface reconstruction from unorganized points , 1992, SIGGRAPH.

[18]  Marek Magdziak,et al.  Usability assessment of selected methods of optimization for some measurement task in coordinate measurement technique , 2012 .

[19]  Robert Hocken,et al.  Operating a Coordinate Measuring Machine , 2016 .