This paper introduces an intelligent system for monitoring and recognition of process disturbances during short-short-Cirutiong gas-metal arc welding. It is based on the measured and statisticlly processed dataof welding electrical parameters. A 12-dimensional array of process features is designed to describevarious welding conditions and is employed as input vector of the intelligent system. Three methods, Such as fuzzy c-means, neural netowrk and fuzzy Kohonen clustering network areused to conduct process monitoring and automatic recognition. The correct recoginition rates of these three methods are compared.

Double-sided arc welding (DSAW) is a novel process for joining metals that is capable of achieving deep narrow penetration in a single pass with minimized distortion. Aclear understanding of the process fundamentals is critical in order to fully examine the potential of DSAW, to guide the further developments of the process, to direct the practicability study, and to design the process parameters. This paper aims at developing a numerical model for examining and simulating the dynamic keyhole establishment process, which will be a key in developing an effective control technology for DSAW. The model is used to determine the geometrical shape of the keyhole and the weld pool, and the temperature distribution in the workpiece. Quantitative information on the establishment of the keyhole in DSAW, such as the transient development of the keyhole and the weld pool, the increase rate of the depth of the surface depression, the time interval from full penetration to the keyhole establishment, the minimum span of the weld pool for describing the conditions required to complete the keyhole establishment, and variation of the overflow height of the weld pool surface on the plasma arc welding side, has been obtained through numerical analysis. The DSAWexperiments show that the predicted weld cross-section is in agreement with the measured one. The results lay a foundation for guiding the further development of the DSAW process and its effective control.

The correlation of the heat flux on the anode surface with the plasma properties at the free-fall edge was developed. By using the models of the tungsten inert gas (TIG) welding arc and the anode boundary layer, the heat transfer contributions to the anode heat flux in TIG welding were analyzed. The percentages of the electron flow contribution and the convection contribution were determined. The results show that the energy carried by the electrons constitutes most of the heat on the anode surface. 2001 Elsevier Science B.V. All rights reserved.

Establishing the correlation between transient behaviors of surface deformation of weld pools and workpiece penetration information, and quantitatively analyzing the surface deformation at top and bottom surfaces at the moment when the pool is penetrated and their dynamic responses to welding process parameters will provide many basic data for the realization of topside vision-based penetration control in gas tungsten arc welding (GTAW). A transient numerical model is developed to investigate the dynamic behaviors of full-penetrated GTAW weld pools in this paper. A whole and comprehensive scheme is used in which many factors such as moving arc, 3-D fluid and heat flow fields, transient state, full-penetrated weld pool and surface deformation at both top and bottom surfaces are considered. The transient developments of 3-D surface deformation and shape of weld pool during the period from partial penetration to full penetration are predicted. The simulated results show that the curves of ratios of the maximum depression to the length and width at top surface of weld pool at different times clearly indicate the basic information of penetration. So the relation of the ratios versus time can be used as an indicator to judge whether the weld pool is penetrated.

A mathematical model has been developed to describe the transient heat and fluid flow fields in gas tungsten arc welding (GTAW). The transient development and diminution of the weld pool at two periods after the arc ignites and extinguishes are analysed quantitatively. The data for the weld pool configurations under different welding conditions from the transient state to the quasi-steady state are obtained. The time for the weld pool shape to reach the quasi-steady state and the time for the weld pool to solidify completely are predicted. GTAWexperiments show that the predictions of the weld pool shape based on the model are in agreement with the measured values. The numerical results of the dynamic development and diminution of weld pool configurations could be used to correlate the transient characteristics of weld pool behaviour with the occurrence of weld formation defects.