Hikaru Yamamoto

Image processing and control of weld pool in switch-back welding without backing plate

Abstract To achieve high quality in multilayer welding, it is important to obtain stable weld results for the first layer.1–5 For this reason, welding is usually conducted by using a backing of the same kind as the base material or a ceramic backing plate. In using a backing metal, the weight of the base material increases. Also it sometimes induces notches to occur at the weld joint between the base material and the backing. This can cause cracking. On the other hand, in using the backing plate, its removal is required after welding, and this makes the method unsuitable for automating the process. Therefore, in the authors previous report, a proposal was made of the switch-back welding method in which welding is executed without using a backing plate, and also an investigation was made of the welding conditions for the root gap.6 In this welding method, it is necessary to adjust the wire feeding rate, the torch motion and the power property in accordance with the gap so that the back bead and the deposit can be controlled. Even if the gap is measured before welding, it may change owing to welding distortion. Hence, when welding is conducted by fixing the welding conditions, the change in the gap causes the back bead to become unstable. For this reason, in this study, a method is discussed to detect the gap on-line by processing images of the molten pool observed by a CCD camera. Then, using the results, the back bead was feed-forward-controlled in real time. Moreover, to examine the controlling performance, welding was conducted by varying the gap between 4.6 mm and 2.5 mm without tack welding. Very good weld results were obtained.

Application of switch back welding to V groove MAG welding

The formation of stable back beads in a root pass weld during one side multi-layer welding is important to achieve high quality welded metal joints in MAG welding. The authors employed the switch back welding method for V groove joints without backing plates. In this switch back welding method, the torch moves forward and backward with an oscillation frequency of 2.5 Hz. In order to achieve this welding, personal computers control the conventional welding robot, the power source characteristic and the wire feeder unit. During the forward, the torch is weaving on the V groove gap without the weld pool. If the weaving width becomes wider than the proper width, the tip of the wire becomes high and a good back bead cannot be obtained. The weaving width is adjusted so as to get the proper width in the switch back welding. The suitability of the welding conditions for each root gap was verified by observation of the arc, the weld pool and the external appearance of back beads. A good back bead was obtained under V groove welding in 2–4 mm gap.

Detecting and tracking of welding line in visual plasma robotic welding

The formation of stable back beads in joining of the thick materials is important in order to achieve high-quality welded metal joints. Plasma welding uses the high welding current density, which is suitable for thick materials. The keyhole in the plasma welding depends on the pilot gas and the welding current. The voltage behaviour depends on the keyhole situation. If the torch is moved away from the welding line in conventional GMA welding, the welding voltage and the welding current are changed due to variations of the arc length. But the welding voltage does not change with the arc length in plasma welding, because the welding voltage depends on the situation of the keyhole. The authors tried to observe the weld pool on the top side by using a CCD camera. The timing of the shutter in the CCD camera is investigated to take clear images of the weld pool. As a result, the clear images of the weld pool were taken when the welding current was reduced to 30 A and an interference filter of 950 nm was attached to the CCD camera. The weld pool shape was changed with the torch position in the groove. The image-processing method was developed to detect the top of the weld pool. The torch position was estimated by processing the weld pool images. The digital control was designed to trace the welding line. The performance of the controller was verified by carrying out tracking experiments.

Estimation of arc length and wire extension using neural network

The automatisation of arc welding using robots has progressed in recent years as skilled welding technicians become more and more scarce. For automatic welding, it is essential to implement methods for sensing the welding phenomena; the requirements for weld quality are high so improvement and movement towards highly functional detection accuracy are desirable. It is important to measure the arc length in order to obtain a continuously stable weld quality, regardless of the torch position deviation from the base metal or the molten pool. The objectives for this article are to show how real-time sensing of the arc length and the electrode extension are feasible not only under steady state conditions but also under transient state conditions and to indicate how to construct a highly versatile system by devising a simple model using a neural network. Even for weaving welding within the groove, the detection of arc length and wire extension in the transient state is indispensable for arc length control, torch position control and the weld line tracking control of the weave centre. This article explains the most fundamental examples of the detection of arc length and wire extension when the arc is started on a flat plate.

Estimation of wire extension length using neural network in MIG welding

Measuring arc length is important to obtain good welding quality in spite of variation of torch height. Therefore, it is necessary to detect arc behaviour in the transient state in addition to the steady state. For this purpose, this paper proposes neural network models which output the present wire extension from the data relating to wire melting, such as welding current, current pickup voltage and wire feed rate in every sampling period. Since performance of the neural network model depends on threshold functions, authors investigate the performance of the neural network models based on both sigmoid function (SF) and radial base function (RBF). To confirm the validity of these systems, fundamental experiments were carried out. The arc was directly observed and recorded as image data using a high-speed camera. The output data from the neural network were compared with the measured data, which were obtained from every captured image. It was found that the neural network model based on the RBF is useful than the SF to estimate the wire extension length in MIG welding because of better responses in the transient state and smaller steady state error.

Cooperative Control of Robotic Welding System in One Side Backingless V Groove Welding

The formation of stable back beads in the first layer weld during one side multilayer welding is important to achieve high quality welded metal joints. The authors thus employed the switch back welding method for the welding of V groove joints without backing plates. In this method, the personal computers control a welding robot, a digital welding power source and a wire feeder, simultaneously. Each unit is connected together with Ethernet and UDP (User Datagram Protocol). The personal computer for controlling the welding robot is synchronized with other computers. The suitability of the welding conditions for each root gap was verified by observation of the arc, molten pool and external appearance of back beads.