Intelligence Computing Method For Evaluating Flatness Errors Of Big Plane

The most basic geometrical element of large-scale equipment is huge size flatness which is also frequently used in equipment machining process and at the same time it is the standard of other geometry. The failure of flatness assess measurement will lead to unsuccessfulness of the entire product geometric quantity thus seriously affect the function of equipment and life expectancy, even cause equipment scrap. Therefore, huge size flatness of assessment provide with important value for study. Take"big size flatness of intelligent assess"as the title and systematically study the genetic algorithms in the flatness calculation of mathematical model and actual measurement applications. That can play an important role of large size measurement of development and academic value as well as practical significance.The thesis analyzes the flatness of approximation, the minimum area search and the convex hull theory of research domestic and overseas and confirmers that the research will use a kind of intelligent algorithm----genetic algorithm to figure out the error of large size flatness. It also excavates the intelligent algorithm which works as a mathematical model to assess large size flatness and hopes to improve the calculation accuracy and reduces the computation time or even further improves the applicability of behavior-error evaluation system. Major work are listed below:1. According to ISO international standard definition of flatness, we can deduce the mathematical model of flatness. And also according to the principle of genetic algorithm and flat degrees minimum zone method, we can structure individual variables and constraints and design the fitness function as well as crossover and mutation methods.2. According to the new generation of GPS uncertainty theory, we design the flatness intelligence evaluation of uncertain analysis method. Analyses of intelligent assessment of the measurement uncertainty of the source points, measuring distribution method, genetic algorithm crossover and mutation and probability and the number of iterations. In addition, we use residual error checking method to judge system errors, Grove Booth Law to judge bulky errors and t distribution method for limit errors.3. In MATLAB environment, we compiled calculation procedures of large size and flatness, at the time designed visual flatness assess platform which can be used in manual and automatic loading points three coordinates data and calculated the flatness. Meanwhile, we also analyzed the iterations calculation accuracy and the computational time influence. According to the contrast experiment, we determine the algorithm of iterations.In order to get the data, the paper's experiments used tri-ordinate measuring machine and digital level for large flat surface, then respectively by using least square method, the diagonal method and genetic algorithm. The results showed that the flatness calculation based on genetic algorithm of intelligent evaluation more accurate. At the same time, it also shows that the combination of intelligence evaluation method, three coordinates measuring and on-line measurement is feasible. The feasibility is strongly supported by the paper's standpoints which are used in industry measurement.