This work aims to improve the low accuracy caused by interference information during real-time weld surface detection. Firstly, a weld surface dataset with 7580 pictures containing 8 types of defects was established. Secondly, an improved detection model named Yolo-MSAPF was designed based on Yolo-v5 model and verified by the self-established database. Finally, a real-time detection system was built to analyze the performance of the detection model in an industrial environment. The design principle of the Yolo-MSAPF is to eliminate interference information but enhance necessary features in each scale as much as possible by multi-scale alignment fusion (MSAF) with parallel features filtering (PFF) modules (i.e., MSAPF strategy). For the MSAF, not only the accuracy but the richness of fused features are guaranteed by aligning features at one level to fuse all other scales. After that, the fused features in each scale were individually filtered out in parallel spatial and channel in the PFF module. The results show that rate for the images with missing detection for 8 types of defects sharply drops from 21.47% to 1.68% when the MSAPF strategy is used. Moreover, it is worth noting that the mAP @0.5 of the Yolo-MSAPF model reaches 95.3%, which is similar to the Yolo-v7 model, while the number of parameters is reduced by 30.1%, compared with the baseline model. Additionally, the great ability to screen out unqualified weld is also verified in the industrial environment. Soon, code and dataset will be available at https://github.com/Luckycat518/Yolo-MSAPF.

The dynamic recrystallization (DRX) features and the evolution of the microstructure of a new hot isostatic pressed (HIPed) powder metallurgy (P/M) superalloy are investigated by hot-compression tests. The sensitivity of grain dimension and DRX behavior to deformation parameters is analyzed. The results reveal that the DRX features and grain-growth behavior are significantly affected by deformation conditions. The DRX process is promoted with a raised temperature/true strain or a reduced strain rate. However, the grains grow up rapidly at relatively high temperatures. At strain rates of o.1 s⁻¹ and 1 s⁻¹, a uniform microstructure and small grains are obtained. Due to the obvious differences in the DRX rate at various temperatures, the piecewise DRX kinetics equations are proposed to predict the DRX behavior. At the same time, a mathematical model for predicting the grain dimension and the grain growth behavior is established. To further analyze the DRX behavior and the changes in grain dimension, the hot deformation process is simulated. The developed grain-growth equation as well as the piecewise DRX kinetics equations are integrated into DEFORM software. The simulated DRX features are consistent with the test results, indicating that the proposed DRX kinetics equations and the established grain-growth model can be well used for describing the microstructure evolution. So, they are very useful for the practical hot forming of P/M superalloy parts.