An in-situ detection method for surface defects of high temperature forgings based on pca and cnn

Abstract

The invention discloses an in-situ detection method for surface defects of high-temperature forgings based on PCA and CNN, comprising: obtaining the three-dimensional point cloud data group of the finished forging to be detected by a three-dimensional laser scanner according to the set sequence rules and calculating its principal curvature to obtain the finished product The five-dimensional point cloud data set of the forging; the PCA method is used to reduce the dimensionality of the five-dimensional point cloud data set of the finished forging to two dimensions, and input it into the successfully trained CNN forging surface defect detector, the CNN The forging surface defect detector outputs judgment results. The CNN forging surface defect detector is trained by using the point cloud data of standard forgings, and then the trained CNN forging surface defect detector is used to detect whether there are defects on the forging surface at a high temperature above 1000 ° C, such as missing corners and protrusions Structural defects such as dents or depressions, to achieve rapid and accurate in-situ detection of high-temperature forgings, to ensure the production quality of forgings.

Description

technical field [0001] The invention relates to the field of detection of high-temperature forgings, in particular to an in-situ detection method for surface defects of high-temperature forgings based on PCA and CNN. Background technique [0002] In industrial production, the free forging process is a relatively common production method, and the production of forgings is also in a very important position. The manufacturing process of free forging is carried out under extreme conditions such as strong earthquakes, high temperature, and high pressure. Forgings need to invest a lot of manpower and material resources before production, and the manufacturing process is continuous and complicated. The material and energy consumption is huge, and the cost is expensive. In the forging process, the size of large forgings is an important indicator that must be detected in time in the production of forgings. [0003] At present, the detection accuracy is low through manual visual insp...

AI Technical Summary

Problems solved by technology

[0003] At present, the detection accuracy is low through manual visual inspection of high-temperature forgings, resulting in the size of forgings generally being larger than the specified value, with an average loss of up to 15%; the characteristic dimensions of key parts of forgings cannot be timely The information is fed back to the host computer system, resulting in failure to improve the internal quality of forgings, improve machining accuracy, and reduce production efficiency

Moreover, during the free forging process, the mold is easily damaged due to repeated extrusion under high temperature and high pressure, and the forging obtained by using the damaged mold for free forging will be a scrap

Images