| Compared with traditional manufacturing technology,laser manufacturing(LM)can reduce costs while providing better quality,and can process complex parts with special-shaped structures.Therefore,it has been increasingly widely used in automotive,shipping,aerospace,biomedical and other fields.In the LM process,the temperature of the melt pool is closely related to the micro properties and morphology of the product.The accurate measurement of the three-dimensional(3D)temperature field of the melt pool is helpful to improve the LM process,thus improving the quality and dimensional accuracy of the product.Among the current temperature measurement technologies,the contact method can only realize the measurement of a few local characteristic points.While the spectral-radiation method has been able to achieve high-precision measurement of the two-dimensional(2D)distribution of the melt pool temperature field because of its non-contact and 2D imaging characteristics.To further improve the LM process and the quality and dimensional accuracy of the product,it is necessary to accurately measure the 3D temperature field of the molten pool.Thus,the key issues of highprecision measurement of 3D temperature field of melt pool using spectral radiation thermometry are studied in this paper.The main research content and innovation points are as follows.Aiming at the problem that the radiation proportion coefficient matrix error is large due to the uniform-approximation and depth-information-flatterning assumption of the radiation model in the 3D temperature field inversion,a non approximate 3D radiation transmission model with the spatial location difference of the radiation source point as the core is established.According to the radiometric theory,the paper firstly establishes the functional relationship among the radiation amount of the source point,the 3D-space position of the source point,and the optical parameters of the observation system,By this way,a radiation proportion model that is strongly related to the 3D spatial distribution of the radiation field is formed and the model error introduced by the assumption that the radiation proportion of different source points are approximately the same is eliminated.Then,according to the principle of optical imaging,a dispersion imaging model of radiation point sources at different observation depths is established.This model eliminates the error caused by the pseudo-3D model in which the pinhole imaging conditions is adopted and the difference of radiation distribution on the image plane caused by different depth of source points is ignored.The transmission model established in this paper can improve the accuracy of the radiation proportion coefficient matrix.As a result,the reconstruction accuracy of the three-dimensional radiation field can be ensured,laying the foundation for high-precision inversion of the temperature field.The simulation results show that the reconstruction accuracy of this model is better than 7% for symmetrical radiation field,and is 30% higher than that of the traditional model.For the asymmetric radiation field,the reconstruction accuracy of this model is better than 11%,which is 60% higher than that of the traditional model.Aiming at the problem that the accuracy of radiation temperature calculation is difficult to achieve when the emission efficiency of the target material is unknown,a high-precision genetic optimization algorithm for emission efficiency is proposed with the emission efficiency changing trend with wavelength as the constraint condition.On the one hand,the algorithm uses the magnitude relationship between spectral emission efficiencies at different wavelengths and the slope characteristics of the emission efficiency versus wavelength curve to establish constraint conditions,which avoids the problem that the accuracy of emission efficiency calculation decreases significantly due to the deviation between the assumption model and actual situation.As a comparision,the traditional algorithm using emission efficiency assumption model as the constraint condition.On the other hand,the algorithm uses the global optimization genetic algorithm to solve the emission efficiency iteratively,overcoming the limitation that the solution accuracy is restricted by the search range and the selection of the initial iterative value.Our method does not need to reasonably select the initial iteration value,but only needs the initial emission efficiency within the natural physical range(0< ε < 1).At the same time,there is no need to limit the search scope in the iterative optimization.This method not only reduces the experience requirement for the algorithm user,but also achieves high-precision solving for the target material emission efficiency with strong adaptive ability.The simulation results of typical tungsten materials show that the accuracy of the algorithm is better than 5% for emission efficiency calculation and 1% for temperature.Aiming at the problem that the acqusiton synchronization of the time-domain and the acqusiton consistency of spatial characteristics of the temperature field cannot be guaranteed,which affects the measurement accuracy of the 3D spatio-temporal evolution characteristics of the temperature field,a method for 3D temperature measurement based on orthogonal dualchannel single-camera three-wavelength spectral imaging is proposed.In this method,dichroic prism is used for three-channel light splitting,and X prism is used to image the information collected by these channels with equal magnification at three different positions on the same CCD target surface.That is,the 2D spatiotemporal information of the temperature field under three spectral channels can be obtained using a single CCD.Because only a single CCD is used as the detector,the frequency response difference caused by using different detectors in different spectral channels is eliminated,and the time synchronization of collection of spectral intensity data of the temperature field is ensured.In this way,the real-time measurement accuracy of temperature field can be improved.On the basis of the above three key technologies,a set of system based on orthogonal dualchannel single-camera three-wavelength spectral imaging is built for verification experiments.The system was used to measure the 3D temperature field of the melt pool produced by highenergy laser cutting 304 stainless steel.The distribution characteristics of the section temperature fields at different depths and that of the profile temperature field of the melt pool was obtained.Besides,the characteristics of the changes of the maximum temperature,width,depth and depth-width-ratio of the melt pool during laser cutting were obtained.The maximum temperature,width and depth all increased rapidly at the beginning,then gradually slowed down and became stable.The depth-width ratio decreased rapidly at first,then gradually slowed down and stabilized.The error analysis shows that the relative error of radiation temperature measurement of the system is better than 1.1%,which is 84.41%,74.76% and 83.69% higher than that of single-color method,two-color method and infrared thermal imaging method respectively.The results of the experiments prove that the method proposed in this paper can obtain the spatio-temporal evolution characteristics and laws of the three-dimensional temperature field of the melt pool in the laser manufacturing process with high accuracy.The method can provide measurement support for the improvement and enhancement of LM processes such as laser cutting and laser a welding. |
