Image-based electrode tip displacement in resistance spot welding

Abstract The paper presents a new image processing-based algorithm for electrode tip displacementmeasurement. Its accuracy is evaluated by a comparison with a conventional optical sensor.The results show a clear advantage of the proposed image processing algorithm. A specialprogram was written in order to analyze the algorithm. The electrode tips are detected by theprogram. It also determines the size of the reference areas on the electrode tips and tracks theseareas during the welding process. Due to a multi-thread implementation of the algorithm theexecution time was reduced significantly as compared to previously presented work.Consequently this approach is now applicable for control-related resistance spot weldingpurposes. Keywords: resistance spot welding, electrode displacement, digital image correlation(Some figures may appear in colour only in the online journal) 1. Introduction Resistance welding (RW) can be defined as a group of weldingprocesses that produce coalescence of the faying surfacesusing the heat obtained from the resistance of the workpiecesto the flow of electrical current in the circuit which theworkpieces are a part of and by the application of pressure [1].There are several different versions of RW. The most widelyused is resistance spot welding (RSW). RSW is one of themajor welding technologies used in the appliance, electricand aviation industries. The automotive industry, however, isthe major user due to its high efficiency, low cost and highdegree of automation. There are thousands of spot welds onan automobile body [2, 3] and in a steel car body; in general,90% of assembly work is performed with RSW [4]. As theproperties of the spot welds significantly affect the durabilityand crashworthiness of the vehicle [5], improving their qualityis an ongoing process in RSW research [6–10].At least in theory an optimal spot weld can be obtained ifan appropriate set of welding parameters is chosen (weldingtime, welding current, welding force, electrode tip geometry,etc). There are, however, two problems associated with thisapproach. The first one is the vast array of different materialsof different thicknesses, without or with different coatings.These materials are sometimes combined. So an enormousdatabase is needed in order to select the correct parameters.The second problem, which is even more pronounced, is thefactthattheweldingprocessisalwaysassociatedwithdifferentdisturbances (variable contact area between the electrodesand the workpieces, variable surface conditions, variablesheet and coating thicknesses, poor fit-up of the workpieces,shunting, small edge distances, disturbances in the powersupply, etc) [1].The most elegant approach to solve these kinds ofproblems is a feedback control system. The feedback signalmust be a signal whose values are closely correlated with thewelding nugget development. The most commonly appliedfeedback signals are electrical signals (welding voltage,welding current and dynamic resistance) [11], electrodedisplacement [12–14] and ultrasound transmission [15]. Othersignals like welding force, acoustic and sonic emission,infrared light emission and thermoelectric voltage are usedfor various control algorithms as well [1].The electrode displacement signal seems to be one ofthe most researched feedback signals. Its close correlationwith the welding nugget development has been known forquite some time [16, 17]. The most common approach forthe acquisition of the displacement signal utilizes some kindof optical sensor (digital optical linear quadrature encoder[18], laser triangulation sensor [19], etc). The application of