【摘要】 本文运用模式识别中的统计决策方法对焊接缺陷超声回波信号进行分类识別,以达到区分平面型焊接缺陷(裂纹)和体积型焊接缺陷(夹渣、气孔)的超声回波信号之目的。试验研究表明,本文采用的方法具有一定的可靠性和可行性。经过对32个实际焊接缺陷超声回波信号的分类识别,平面型缺陷的正确分类率为86%,体积型缺陷的正确分类率为78%;总正确分类率为81%。更多还原

【Abstract】 This paper deals with the flaw classification of actual weld defects in wlded specimens, with emphasis on the distinction between the flat flaws (cracks) and the volume flaws (slags and porosities). The results obtained by the ultrasonic frequency spectrum analysis and pattern recognition are proved to be reliable and applicable. The investigation of 32 natural weld defects show that 86% of the plane defects and 78% of the volume defects can be correctly classified. The overall correct index of performance is 81%.更多还原