Blade trailing edge contour based on probability density least-square fitting

In order to solve the problems for measurement and modeling of gas turbine blade trailing edge contour, a least-square circle fitting method based on probability density was proposed. After several times of fitting data repeatedly, each fitting center coordinate and radius was obtained. The maximum value of probability density distribution was looked as the optimal value. The errors of fitting data were analyzed by data simulation after theoretical analysis and experimental verification. The feasibility and robustness of the method were verified. The results show that the fitting of blade trailing edge contour and the estimation of parameter data are realized. The fitting error is 0.01mm. The optimal center coordinates and radius of circle fitting is obtained. For the blade trailing edge composed of straight line segments and short arcs, the fitting result is effective and accurate in the case of unknowing the point of tangent and arc equation. The method has an important guiding significance to precision measurement, machining accuracy and parameter design of blade trailing edge contour.