中文摘要：

通过采集TIG，MIG焊接过程的光谱辐射信息，基于等离子体辐射的基础理论，对其焊接电弧辐射进行了分析，TIG焊与MIG焊的光谱分布由于气氛中金属元素浓度的差别，辐射强度和分布都存在较大差别：MIG焊不仅金属线谱数量多，辐射强度比TIG焊大，且随熔滴过渡波动明显。针对其光谱分布和变化特点，选择了特征谱段，用于焊接质量的检测；对于TIG焊选取线谱聚集的紫外区辐射（230-300nm），对于MIG焊选取以连续辐射为主的可见光区辐射（570～590nm），建立了不同焊接方法下，焊接电弧光谱信息在焊接质量检测上的应用理论基础。还进一步通过在焊接过程中预设干扰因素，采集焊接过程在特征谱段的信号，对选择谱段的试验验证表明：基于建立的理论基础，可以有效地利用焊接电弧光谱信息，对焊接质量及焊接过程干扰因素实现判识，特征谱段的信号具有很好的信噪比。

英文摘要：

Through collecting the spectral signals of TIG and MIG welding arc with spectrometer, the arc light radiations were analyzed based on the basic theory of plasma physics. The radiation of welding arc distributes over a broad range of frequency, from infrared to ultraviolet. The arc spectrum is composed of line spectra and continuous spectra. Due to the variation of metal density in the welding arc, there is great difference between the welding arc spectra of TIG and MIG in both their intensity and distribution. The MIG welding arc provides more line spectra of metal and the intensity of radiation is greater than TIG. The arc spectrum of TIC, welding is stable during the welding process, disturbance factors that cause the spectral variations can be reflected by the spectral line related to the corresponding element entering the welding are. The are spectrum of MIG welding will fluctuate severely due to droplet transfer, which produces ＂noise＂ in the line spectrum aggregation zone. So for MIG welding, the spectral zone lacking spectral line is suitable for welding quality diagnosis. According to the characteristic of TIG and MIG, special spectral zones were selected for welding quality diagnosis. For TIG welding, the selected zone is in ultraviolet zone （230-300 nm）. For MIG welding, the selected zone is in visible zone （570-590 nm）. With the basic theory provided for welding quality diagnosis, the integral intensity of spectral signal in the selected zone of welding process with disturbing factor was studied to prove the theory. The results show that the welding quality and disturbance factors can be diagnosed with good signal to noise ratio in the selected spectral zone compared with signal in other spectral zone. The spectral signal can be used for real-time diagnosis of the welding quality.