No abstract

<p>The characterization of pores in materials such as Opalinus clay is crucial for understanding the physical properties, including permeability and strength, which are important for the safe disposal of radioactive waste. Scanning electron microscopy (SEM) is a technique that allows high-resolution imaging of these pores at the nanoscale. However, the analysis of SEM images can be challenging due to the resolution limits of nanoscale pores and their manual segmentation. In the development of automatic segmentation methods, approaches of supervised or unsupervised machine learning (ML) and deep learning (DL) methods are increasingly applied. The main advantage of these methods is to achieve fast and more consistent results that do not rely on user input.</p> <p>An essential component in DL is the so-called backbones, which can learn object features that are necessary for object recognition. In image processing, objects are recognized through groups of specific features that allow an unambiguous identification. Pre-trained backbones, which have been trained on large datasets such as ImageNet containing millions of everyday images, possess a wide range of features that are useful during image processing tasks. However, specialized applications, such as the automatic analysis of microscope images using DL may require features that differ from those of pre-trained backbones. The limited availability of SEM images makes it difficult to effectively train DL models, as these models typically require a large amount of data to learn new features. In these cases, ML methods may perform better due to their ability to use carefully selected, expert-defined features [Maitre et al., 2019].</p> <p>In this study, the training behavior of eight different DL backbones was examined using a dataset of 2000 SEM images showing both the background and pores of an Opalinus clay sample. The backbones studied included VGG16, VGG19, ResNet50, Desenet, Xception, and Mobilenet. To train these models with the relatively small amount of training data available, a transfer learning technique was applied. We analyzed gradient-weighted class activation mappings (grad-CAM) [Selvaraju et al.,2019] during the learning process to obtain a general sense of the behavior of the different backbones. Through analysis of the model's adaptation efforts, the present study demonstrates which pre-trained backbones show good training behavior on SEM images and provides an estimation of the amount of data needed for effective training.</p> <p> </p> <p>References</p> <p>[Maitre et al., 2019] Maitre, J., Bouchard, K., and Bédard, L. P. (2019). Mineral grains recognition using computer vision and machine learning. Computers & Geosciences, 130:84–93.</p> <p>[Selvaraju et al., 2019] Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., and Batra, D. (2019). Grad-cam: Visual explanations from deep networks via gradient-based localization.</p>

Controlling complex systems by traditional control systems can sometimes lead to sub-optimal results since mathematical models do often not completely describe physical processes. An alternative approach is the use of a neural network based control algorithm. Neural Networks can approximate any function and as such are able to control even the most complex system. One challenge of this approach is the necessity of a high speed training loop to facilitate enough training rounds in a reasonable time frame to generate a viable control network. This paper overcomes this problem by employing a second neural network to approximate the output of a relatively slow 3D-FE-Pultrusion-Model. This approximation is by orders of magnitude faster than the original model with only minor deviations from the original models behaviour. This new model is then employed in a training loop to successfully train a NEAT based genetic control algorithm.

Automation technologies such as Automated Fiber Placement (AFP) or Automated Tape Laying (ATL) are widely used in the aerospace industry today. However, these processes can still be further improved for higher productivity. Fiber-reinforced plastics allow the production of components with extremely high specific strength and stiffness. Regarding the automated manufacturing processes, the thermoplastic tape placement offers efficiency improvements compared to the nowadays more commonly used thermoset tape placement, especially through the substitution of the expensive and time-consuming autoclave process. The consolidation of thermoplastic Prepregs is achieved with an elastic or rigid roller according to the current state of the art. The Prepregs must be consolidated precisely on the substrate or on top of each other. The most important process parameters for high-quality laminate structure with low porosity are the control of heat source, consolidation force, consolidation roll speed, and tape tension. The efficiency of the AFP process can generally be improved by increasing the speed of the consolidation roller. By increasing the speed of the consolidation roller, porosity is increased and mechanical properties of the laminate are reduced significantly due to the short contact time between consolidation roller and Prepregs. This study investigates a process that can reduce these challenges by increasing the contact time and force duration of the consolidation roller on the Prepregs. The consolidation roller in this study is additionally to be driven by the harmonic oscillations. The new method allows the consolidation roller to oscillate forward and backward during the fiber placement process. This creates another force vector in addition to the compressive force of the consolidation roller and increases the bonding strength between the layers.

The reduction of material defects in the automated fiber placement process is one of the significant factors for manufacturing large and complex components more efficiently in the future. However, the monitoring of complex manufacturing processes usually requires complex sensor and computer systems that are often quite sensitive to disturbances and errors. New techniques such as image segmentation with neural networks provide a new approach to this problem and have the potential to solve complex processes faster and more robustly. In this study, a system is presented that performs monitoring, inspection and measurement tasks simultaneously in automated fiber placement processes. The system is based on the SiamMask network which is used for the automatic image processing. The artificial neural network is trained to recognize individual carbon fiber tapes and segment them for additional analysis. For the creation of the testing- and training data, an analytical approach is presented. The post-processing of the object segmentation, which is the primary output of the SiamMask network and the identification of individual tapes, provides accurate measurements which are demonstrated by an example. We show that image segmentation with modern approaches like SiamMask offers great potential to handle highly complex engineering tasks in a faster and more intelligent manner in comparison to conventional methods.