Research On Sliding Model Variable Structure Control Of Real Time Seam Tracking Based On Magnetron Arc Sensor

At present, the level of welding automation is low in our country. A majority of welding operation are still staying in traditional manual welding and semi-automatic welding. It is a long way for each welding researcher to achieve welding automation and intellectualization. In order to improve the level of welding automation and intellectualization, the first important task is to realize automatic welding seam tracking. There are a lot of deficiencies when the seam tracking common methods are applied into the practical production, the main problems include: the complex control system and low performance-price ratio, the unbalance between parameter adjustment flexible and filtering effect, the delayed tracking caused by phase delay and high requirements to the working environment. It is imperative to develop real-time seam tracking method of high performance-price ratio.The intelligent seam tracking method based on magnetron arc sensor has the potential to solve the above problems, so the further research on it is of great significance. Basing on magnetron arc sensor, This paper combines the real-time signal processing, the welding deviation prediction, the tangent fitting ideas and sliding mode variable structure to realize automatic seam tracking. It contributes to perfect the theoretical system of automatic seam tracking based on magnetron arc sensor.1、The magnetron arc sensor has been optimized. Arc scanning v-groove model has been established and verified by simulation and experiment.Aiming at the influence on swinging arc made by CO2 gas and the distance between two poles and on magnetic field made by the temperature, the function of magnetron arc sensor has been perfected. Meanwhile, the principle of automatic seam tracking based on magnetron arc sensor has been elaborated and arc scanning V groove model has been established. The theoretical feasibility of the model has been verified by simulation. In the experiments,when the frequency of swinging arc is set to 3Hz,4Hz,5Hz or 6Hz, The feature signal of groove seam tracking based on magnetron arc sensor has been extracted. The experimental results have verified the correctness of the arc scanning v-groove model and proved that it has greater potential that magnetron arc sensor applied in automatic groove seam tracking.2、Two real-time signal processing methods are used for filtering magnetron arc signal. The advantages and disadvantages of two methods are elaborated.Aiming at nonlinear and non-stationary characteristics of seam tracking signals based on magnetron arc sensor, the extraction method of seam tracking characteristic signal based on combining matching pursuit and the nonparametric basis function is presented. Meanwhile, the voltage signal extracted by HR sensor is filtered with real time wavelet de-noising method. The signal filtered with two methods can be used to achieve automatic seam tracking. The wavelet de-noising is widely used for signal processing and is development-maturity theory, but the theoretical basis based on Fourier analysis lead to its poor flexibility.3. The dynamic weld tangent method has been presented by combining the tangent fitting idea and Sliding Mode Variable Structure Control and the theoretical feasibility of the method has been verified by the simulation results.On the basis of real time signal processing, the dynamic weld tangent method has been presented by combining seam deviation prediction, the tangent fitting idea and variable structure control method. According to the determined seam deviation, the tangent fitting model is used to work out the deflection angle. Subsequently, the variable structure control method is used to adjust the slider or welding tractor or the slider and welding tractor. Meanwhile, Seam tracking is realized rapidly and the shaking is prevented.4. The feasibility of automatic fillet seam tracking based on combining magnetron arc sensor and the dynamic weld tangent method has been verified by the experimental.Finally, on the seam tracking platform built by ourselves, the fillet seam tracking experimental has been realized by combining real time signal processing, the seam prediction, the tangent fitting and variable structure control method.