Ingenieurwissenschaften und zugeordnete Tätigkeiten
Filtern
Erscheinungsjahr
- 2023 (3) (entfernen)
Dokumenttyp
- Zeitschriftenartikel (3) (entfernen)
Sprache
- Englisch (3)
Referierte Publikation
- ja (3)
Schlagworte
Organisationseinheit der BAM
AbstractLaser beam welding has become widely applied in many industrial fields in recent years. Solidification cracks remain one of the most common welding faults that can prevent a safe welded joint. In civil engineering, convolutional neural networks (CNNs) have been successfully used to detect cracks in roads and buildings by analysing images of the constructed objects. These cracks are found in static objects, whereas the generation of a welding crack is a dynamic process. Detecting the formation of cracks as early as possible is greatly important to ensure high welding quality. In this study, two end-to-end models based on long short-term memory and three-dimensional convolutional networks (3D-CNN) are proposed for automatic crack formation detection. To achieve maximum accuracy with minimal computational complexity, we progressively modify the model to find the optimal structure. The controlled tensile weldability test is conducted to generate long videos used for training and testing. The performance of the proposed models is compared with the classical neural network ResNet-18, which has been proven to be a good transfer learning model for crack detection. The results show that our models can detect the start time of crack formation earlier, while ResNet-18 only detects cracks during the propagation stage.
AISI 2205 duplex stainless steel is used in a variety of industries, including the chemical and petrochemical industries. This is due to its high tensile strength combined with good ductility and corrosion resistance. However, in laser beam welding, these properties are negatively afected by the high cooling rates typical of the welding process. The resulting higher ferrrite content in the weld metal than in the base material leads to a reduction in the ductility and corrosion resistance of the welded joint. To overcome this problem, in this study, thick plates were coated by direct energy deposition (DED) prior to laser beam welding, whereas a duplex powder mixture containing a higher nickel concentration was used as a coating material. To improve the weld quality for the proposed two-step process, a method of additional material deposition instead of conventional tack weld was investigated. The resulting welded joints showed a well-balanced austenite to ferrite ratio and their properties and microstructure were verifed by metallographic analysis, electron backscatter difraction and Charpy impact testing. Using the standard ASTM G48 test method, it was found that the corrosion resistance of the welds was improved by a factor of four in average compared to the conventionally welded joints. The resulting properties, such as good ductility and corrosion resistance, of the welds with pre-coated edges showed good agreement with those of the base metal and confrmed the proposed two-step process as a promising alternative to the conventional approaches for welding thick duplex stainless steel plates.
The strain field can reflect the initiation time of solidification cracks during the welding process. The traditional strain measurement is to first obtain the displacement field through digital image correlation (DIC) or optical flow and then calculate the strain field. The main disadvantage is that the calculation takes a long time, limiting its suitability to real-time applications. Recently, convolutional neural networks (CNNs) have made impressive achievements in computer vision.
To build a good prediction model, the network structure and dataset are two key factors. In this paper, we first create the training and test sets containing welding cracks using the controlled tensile weldability (CTW) test and obtain the real strain fields through the Lucas–Kanade algorithm. Then, two new networks using ResNet and DenseNet as encoders are developed for strain prediction, called StrainNetR and StrainNetD. The results show that the average endpoint error (AEE) of the two networks on our test set is about 0.04, close to the real strain value. The computation time could be reduced to the millisecond level, which would greatly improve efficiency.