基于旋转电弧传感器的水下焊缝偏差识别算法

提出了基于旋转电弧传感器的Nu-SVR水下焊缝偏差识别算法.Nu-SVR通过对基于旋转电弧传感器采集到的不同偏差的水下焊接信号进行学习,然后对水下电流信号进行焊缝偏差识别得出偏差.相对于传统的回归算法--区间积分法和神经网络法,算法具有更好的识别能力.最后通过水下焊接实验,其最大的识别误差仅为0.554 mm,证明了该方法十分有效.

Seam Offset Identification of Underwater Arc Welding Based on Rotating Arc Sensor

A learning algorithm for underwater arc welding seam offset identification using Nu-SVR is proposed. Nu-SVR is used to identify the welding seam offset using the learning result from the periodical the welding current signals. Compared to traditional regression method, advantage of the method is that it has better generalization ability. It outperforms existed methods, such as interval integral and artificial neural network method. The max error of the underwater arc welding seam offset identification error is 0.554 mm, which proves that Nu-SVR is effective in the underwater arc welding seam offset identification.