In this paper, ultrasonic echo signals of four kinds of stainless steel resistance spot welds, namely failed weld, stick weld, good weld, and defective weld with gas pore, are analyzed in the time domain, frequency domain, and time-frequency domain based on wavelet packet transform. Fourteen ultrasonic characteristic signals which can reflect the different kinds of spot welds are extracted and can be automatically identified and classified by back-propagation (BP) neural network. The method of this paper can realize the intelligent identification of resistance spot welding defects, and the feasibility of this method has been verified in the experiment.

Defect intelligent identification in resistance spot welding ultrasonic detection based on wavelet packet and neural network

The aim of this paper is to provide an integrated real-time control system for resistance spot welding (RSW), which is capable of producing welds with predetermined nugget diameters. Nugget diameter is a commonly used criterion for estimating the weld quality; thus, the system can be used for online quality control of RSW. The proposed system consists of two parts: a constant current controller and an online nugget diameter estimator. The constant current controller is used to guarantee the consistency of the nugget formation and growth during the welding process, while the online nugget diameter estimator can be used as a sensor for online estimation of the nugget diameter. The proposed online nugget diameter estimator is a mathematical function of the heat energy absorbed by the weld, which is taken into account only after the time when the first melting point of a weld appears. This time can be obtained by means of the dynamic resistance curve with a high resolution. A mathematical model of the online nugget diameter estimator is proposed and realized based on the experiment results and mathematical analysis. The welding process would be terminated when the difference between the actual nugget diameter and the predetermined desired value is within a given tolerance. According to a series of experiments, the welds used in the experiments had nugget diameters, which were well within the limits and the errors were within the given tolerance.

Online nugget diameter control system for resistance spot welding

Resistance spot welding is one of the most important welding procedures. Therefore, a lot of research is done in order to increase its cost effectiveness. One of the problems is short electrode life, especially when coated materials are welded. This paper presents a fuzzy logic-based controller which is capable of detecting expulsion and stopping the welding process when it occurs. Consequently, electrodes are spared of unnecessary high stresses which occur when the energy is being poured into the weld region after expulsion. Their life is therefore substantially increased.

Resistance spot welding control based on fuzzy logic

Deep learning assisted vision inspection of resistance spot welds

To develop an effective nondestructive evaluation method for the welding quality of the resistance spot welding, the electrode displacement signal during the resistance spot welding process is monitored, and the acquisition data of the signal are innovatively presented as the radar chart format. Some geometric features of the radar charts are extracted to reflect the welding quality. The decision tree classification technique is adopted to build a classifier and to provide a visible and intuitive diagnostic procedure for welding quality assessment. Test results of the decision tree classifier show that it is feasible and reliable to evaluate weld quality based on the graphics features of the radar chart. The features and the weld quality are closely related, and their extraction avoids complex algorithm. Meanwhile, when there are small samples, the decision tree classifier can identify good or bad weld accurately and rapidly, even though the weld is from an abnormal welding process, such as expulsion, current shunting, greasy surface, and small edge distance condition.

A new method for nondestructive quality evaluation of the resistance spot welding based on the radar chart method and the decision tree classifier

This paper introduces a novel method for the quality evaluation of resistance spot welds. The evaluation is based on computer vision methods, which allow non-destructive on-line real-time processing. The input of the system is the image of a weld imprint on a metal band which covers the electrodes against wear and soiling. The shape and size of the structures within the imprint correlate with the nugget area and, therefore, allow an accurate estimation of the quality of the spot weld. The system segments the electrode imprint and computes the nugget area from the minimum and maximum axis of a fitted ellipse. This method does not need training samples to perform reliable quality estimation. Additionally, the used algorithms are easy to implement and efficient, which guarantees real-time ability. Since there is only a single low-cost camera needed, the hardware can be placed directly on the gun arm, which makes a fast evaluation possible.

Jürgen Biber

Papers

Quality evaluation in resistance spot welding by analysing the weld fingerprint on metal bands by computer vision

RGB-D Structural Classification of Guardrails via Learning from Synthetic Data