摘要
To obtain the form error of micro-structured surfaces robustly and accurately, a form er- ror evaluation method was developed based on the real coded genetic algorithm （RCGA）. The meth- od employed the average squared distance as the matching criterion. The point to surface distance was achieved by use of iterative method and the modeling of RCGA for the surface matching was also presented in detail. Parameter selection for RCGA including the crossover rate and population size was discussed. Evaluation results of series simulated surfaces without form error show that this method can achieve the accuracy of root mean square deviation （ Sq ） less than 1 nm and surface pro- file error （ St ） less than 4 nm. Evaluation of the surfaces with different simulated errors illustrates that the proposed method can also robustly obtain the form error with nano-meter precision. The e- valuation of actual measured surfaces further indicates that the proposed method is capable of pre- cisely evaluating micro-structured surfaces.
To obtain the form error of micro-structured surfaces robustly and accurately, a form er- ror evaluation method was developed based on the real coded genetic algorithm （RCGA）. The meth- od employed the average squared distance as the matching criterion. The point to surface distance was achieved by use of iterative method and the modeling of RCGA for the surface matching was also presented in detail. Parameter selection for RCGA including the crossover rate and population size was discussed. Evaluation results of series simulated surfaces without form error show that this method can achieve the accuracy of root mean square deviation （ Sq ） less than 1 nm and surface pro- file error （ St ） less than 4 nm. Evaluation of the surfaces with different simulated errors illustrates that the proposed method can also robustly obtain the form error with nano-meter precision. The e- valuation of actual measured surfaces further indicates that the proposed method is capable of pre- cisely evaluating micro-structured surfaces.
基金
Supported by the Programme of Introducing Talents of Discipline to Universities (B07018)
