Method for identifying defect ultrasonic signals in steel ingot based on support vector machine

Abstract

The invention discloses a method for identifying defect ultrasonic signals in a steel ingot based on a support vector machine, which belongs to the field of ultrasonic nondestructive testing, and is used for performing data processing on a characteristic parameter vector matrix of the defect ultrasonic signals in the steel ingot through a Gaussian radial basis kernel function based on a support vector machine algorithm. Linear inseparable data points in an original low-dimensional space are mapped to a high-dimensional space to enable the data points to be linearly separable, then an optimal classification hyperplane is constructed, the distance between different types of data on the two sides of the plane and closest to the plane is maximized, and therefore correct classification of different signals is achieved. In order to improve the signal identification rate in the identification process, the characteristic parameters of the signals and the system parameters of the support vector machine are optimized.

Description

technical field [0001] The invention belongs to the field of ultrasonic non-destructive testing, and in particular relates to a support vector machine-based defect ultrasonic signal identification method in steel ingots, which is applicable to the identification and classification of defect signals in the ultrasonic testing of steel ingots. Background technique [0002] With the development of metallurgy, petroleum, petrochemical, aerospace, shipbuilding and other industries, the demand for steel products is increasing, so the quantity and quality requirements for steel ingots are also increasing accordingly. Once a quality defect occurs in the steel ingot, it will directly affect the subsequent rolling and forging process and even the quality of the finished steel product. Therefore, it has become an indispensable link in the production process of steel ingots to detect defects in steel ingots in time and repair the defects that can be remedied to minimize economic losses. ...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problem that the ultrasonic detection signal waveform of typical defects in steel ingots is complex and the defect type identification is difficult, the present invention proposes an identification method of defect ultrasonic signals in steel ingots based on support vector machines

Images