This work presents a methodology used to filter the roughness profile of soft metals and natural materials. This methodology is based on polynomial regression, Robust Gaussian Regression filter (RGRF) and Abbot filter. The advantages of this method come from the possibility of elimination the deep valleys introduced by scratches after manufacturing or associated with anatomy of materials like wood. The method involves three steps: i) fitting roughness raw data with polynomial regression to remove profile form errors; ii) using the RGRF to filter the profile waviness; iii) applying the Abbot curve method to remove the remained deep valleys. The resulting profile was used to determine roughness parameters Ra and Rmax. Samples of wood Eucalyptus Camaldulensis and aluminium were prepared to carry out the measurements and the calculations where performed with algorithms developed using MatLab software. The results proved that the proposed approach is robust against modifications introduced after processing of soft materials surface and suitable to apply to materials having particular anatomic components.
Keywords: roughness, filtering, polynomial regression, RGRF, Abbot curve.
RESUMO
Este trabalho apresenta uma metodologia para filtrar o perfil efetivo na medicao da rugosidade superficial de metais moles e materiais naturais, baseado na aplicacao de regressao polinomial, regressao gaussiana robusta (RGRF) e curva de filtragem de Abbot. A vantagem deste metodo reside na possibilidade de remover os vales profundos introduzidos apos a fabricacao ou associados a anatomia de materiais como madeiras. O metodo deve ser aplicado em tres etapas sequenciais: ajustar os dados do perfil efetivo a uma regressao polinomial para remover o erro de forma; ajustar este novo perfil a uma regressao RGRF para filtrar a ondulacao; aplicar a curva de filtragem Abbot para remover os vales profundos remanescentes. O perfil resultante foi usado para determinar os parâmetros de rugosidade Ra e Rmax. Amostras de madeira Eucalyptus Camaldulensis e aluminio foram preparadas para executar as medicoes e os calculos foram feitos atraves de algoritmos desenvolvidos no programa MatLab. Os resultados mostraram que a abordagem proposta foi robusta em relacao as modificacoes introduzidas apos o processamente de materiais moles e adequada para aplicacao em materiais com componentes anatomicos destacados.
Palavras-chave: rugosidade, filtros, regressao polinomial, RGRF, curva de Abbot.
[1]
B. Muralikrishnan,et al.
Recent advances in separation of roughness, waviness and form
,
2002
.
[2]
T. Nomura,et al.
High-order spline filter and ideal low-pass filter at the limit of its order
,
2007
.
[3]
Pawel Pawlus,et al.
Digital filtering of surface topography: Part I. Separation of one-process surface roughness and waviness by Gaussian convolution, Gaussian regression and spline filters
,
2010
.
[4]
M. Irle,et al.
Filtering the roughness of a sanded wood surface
,
2006,
Holz als Roh- und Werkstoff.
[5]
Y. Fujii,et al.
Assessment of wood surface roughness: comparison of tactile roughness and three-dimensional parameters derived using a robust Gaussian regression filter
,
2004,
Journal of Wood Science.
[6]
P. Dobrzanski,et al.
Digital filtering of surface topography: Part II. Applications of robust and valley suppression filters
,
2010
.
[7]
Bala Muralikrishnan,et al.
Computational Surface and Roundness Metrology
,
2008
.
[8]
M. Irle,et al.
Processing roughness of sanded wood surfaces
,
2005,
Holz als Roh- und Werkstoff.
[9]
Ebrahim Shayan,et al.
Analysis of roughness of a sanded wood surface
,
2006
.