Prediction method of surface roughness of elastic workpiece robot grinding and polishing based on virtual samples

Abstract

The invention belongs to the field of robot grinding and polishing, and specifically discloses a method for predicting the surface roughness of elastic workpiece robot grinding and polishing based on virtual samples. Including: building a sample set based on the robot grinding and polishing experiment data, denoising the sample points in the sample set, and then dividing the remaining sample points according to the set ratio to obtain training samples and verification samples; The distribution space and distribution dispersion generate the confidence space of the training samples, and generate relevant virtual samples in the confidence space, construct the loss function of the mapping relationship model, and then use the virtual samples and the training samples together as the input of the mapping relationship model, and the mapping Gradient iterative training is performed on the relational model to solve the poles of the loss function, and verification samples are used to verify the mapping relational models at the poles to obtain the optimal mapping relational model. The method of the invention can still maintain high prediction accuracy and good robustness under small sample size.

Description

technical field [0001] The invention belongs to the technical field of industrial robot automatic processing, and more specifically relates to a method for predicting surface roughness of elastic workpiece grinding and polishing based on virtual samples. Background technique [0002] Robot grinding and polishing is one of the key processes to improve the surface quality and dimensional accuracy of workpieces. As an important evaluation index to measure the quality of robot grinding and polishing, surface roughness not only represents the manufacturing accuracy level of the part surface, but also has a huge impact on the performance of parts and the reliability of the overall equipment. However, due to the influence of processing materials, grinding environment and robot body stiffness, the mapping relationship between process parameters and surface roughness is difficult to unify. In grinding and polishing, it is usually necessary for technicians to repeatedly experiment an...

AI Technical Summary

Problems solved by technology

[0004] For elastic materials (such as rubber, plastics, etc.), not only the processing thickness will change during the grinding and polishing process, but also a certain deformation will occur in itself. The correction scheme proposed by the patent document CN109794813A exists in the application of elastic workpieces. Certain limitations, did not take into account the impact of the angle between the grinding disc and the surface of the workpiece on the surface roughness; at the same time, the linear regression used in the patent document CN110118543A has a high requirement for the accuracy of the model itself, when the physical model is not complete , there are certain limitations in its prediction effect and universal ability. In the process of surface roughness prediction, problems such as weak generalization ability and poor robustness still exist.

Images