Research On Detection And Control Of Melt Pool Geometrical Morphology In Wide-beam Laser Cladding

Parts such as automobile cover mould are complex in structure,and the qualities of the surfaces are required to be high.As they are operated in complex heavy load conditions for a long time,they are easily to be ineffective because of the surface abrasion.The dimension and performance of the part surface can be effectively recovered by remanufacturing technology,and then the part can be put back into use.The high power semiconductor diode has its own characteristics of high output power,short laser wavelength and large spot size.The forming precision and quality of the cladding layer can be effectively improved by the accurate detection and reasonable control of the melt pool morphology during the laser cladding process.Therefore,the high quality and high efficiency remanufacturing and repairing of the complicated parts can be realized.In the paper,in order to solve the problems of detection and control of the wide-beam laser cladding melt pool caused by the high temperature and high brightness,a vision detection scheme for wide-beam laser cladding melt pool is designed,the geometrical and dynamic characteristics of the melt pool are analyzed,and a corresponding melt pool control strategy is developed.The main research contents are as follows:Firstly,the selection and design of the melt pool detection system are carried out.The inverse perspective conversion model between the pixel coordinate system and the world coordinate system is established by camera calibration.And a melt pool contour extraction algorithm based on the local segmentation framework and the active contour model is written.Therefore,the contour extraction and geometrical characteristics measurement of the melt pool are realized.A software system for laser cladding melt pool detection is developed,which combines with many functions including camera calibration,image preprocessing and melt pool contour extraction.Secondly,the effects of laser power,scanning speed and powder thickness on the width,length and area of the melt pool are studied by the full factorial experiments.The process parameters and the geometrical characteristics of the melt pool are analyzed by regression,and a geometrical model is established to further characterize the melt pool contour in different process parameters.Then,the step input experiments of process parameters are carried out,a single input and single output model by using transfer function is established between the laser power,the scanning speed and the melt pool width,respectively.The scanning speed is selected as the control object,and the melt pool width is selected as the controlled object,the random input signal of the scanning speed is designed for laser cladding experiment.The obtained data of the melt pool width is dynamically expanded,and the initial weights and thresholds of the multi-layer neural network model are optimized by genetic algorithm.The prediction model of the melt pool width at the current moment,which lays a foundation for the design of the control strategy of melt pool,is established.Finally,In order to realize the reasonable control of the melt pool,the fuzzy inference system is used to adjust the PID controller parameters adaptively and the fuzzy PID control strategy is designed.Combined with the neural network prediction model of the melt pool width,the fuzzy PID control strategy,PID control strategy and fuzzy control strategy are simulated and compared.The research shows that the adjustment time and the maximum overshoot of the fuzzy PID control strategy is 1.8s and 0.44%,respectively,indicating that the fuzzy PID control strategy has a good control effect on the melt pool width.The above research realizes the accurate detection and reasonable control of the melt pool morphology in the wide-beam laser cladding,which has certain academic and engineering significance.