Gathering of Process Data through Numerical Simulation for the Application of Machine Learning Prognosis Algorithms
TL;DR: A method is shown by which application it is possible, that only on the basis the general mechanical properties and the use of data-based prognosis models of supervised machine learning to predict directly a result regarding suitable process parameters as well as expected forming result properties are shown.
About: This article is published in Procedia Manufacturing.The article was published on 2020-01-01 and is currently open access. It has received 10 citations till now. The article focuses on the topics: New product development.
Citations
More filters
TL;DR: In this paper , a comprehensive overview of the most utilized joining by forming processes is given, highlighting strengths and weaknesses for industrial applications, as well as an analysis of the current research trends and hot topics.
Abstract: Abstract The progressively more demanding needs of emissions and costs reduction in the transportation industry are pushing engineers towards the use of increasingly lightweight structures. This goal can be achieved only if dissimilar and/or new materials, including polymers and composites, are joined together to create complex structures. Conventional fusion welding processes have often been proven inadequate to this task because of the high heat input reducing the joint mechanical properties or even making the joining process impossible. Joining by forming technologies take advantage on the plastic deformation to create sound joints out of even very dissimilar materials. Over the last 25 years, several new processes, with increasing potential in effectively joining virtually every structural material, have been invented and developed. In the paper, a comprehensive overview of the most utilized joining by forming processes is given. For each process, an analysis of the current research trends and hot topics is provided, highlighting strengths and weaknesses for industrial applications.
9 citations
TL;DR: In this paper , a comprehensive overview of the most utilized joining by forming processes is given, highlighting strengths and weaknesses for industrial applications, as well as an analysis of the current research trends and hot topics.
Abstract: Abstract The progressively more demanding needs of emissions and costs reduction in the transportation industry are pushing engineers towards the use of increasingly lightweight structures. This goal can be achieved only if dissimilar and/or new materials, including polymers and composites, are joined together to create complex structures. Conventional fusion welding processes have often been proven inadequate to this task because of the high heat input reducing the joint mechanical properties or even making the joining process impossible. Joining by forming technologies take advantage on the plastic deformation to create sound joints out of even very dissimilar materials. Over the last 25 years, several new processes, with increasing potential in effectively joining virtually every structural material, have been invented and developed. In the paper, a comprehensive overview of the most utilized joining by forming processes is given. For each process, an analysis of the current research trends and hot topics is provided, highlighting strengths and weaknesses for industrial applications.
8 citations
TL;DR: In this article, a hybrid approach to generate sample data for future machine learning applications for the prediction of mechanical properties in directed energy deposition-arc (DED-Arc) using the GMAW process is presented.
Abstract: This research presents a hybrid approach to generate sample data for future machine learning applications for the prediction of mechanical properties in directed energy deposition-arc (DED-Arc) using the GMAW process. DED-Arc is an additive manufacturing process which offers a cost-effective way to generate 3D metal parts, due to its high deposition rate of up to 8 kg/h. The mechanical properties additively manufactured wall structures made of the filler material G4Si1 (ER70 S-6) are shown in dependency of the t8/5 cooling time. The numerical simulation is used to link the process parameters and geometrical features to a specific t8/5 cooling time. With an input of average welding power, welding speed and geometrical features such as wall thickness, layer height and heat source size a specific temperature field can be calculated for each iteration in the simulated welding process. This novel approach allows to generate large, artificial data sets as training data for machine learning methods by combining experimental results to generate a regression equation based on the experimentally measured t8/5 cooling time. Therefore, using the regression equations in combination with numerically calculated t8/5 cooling times an accurate prediction of the mechanical properties was possible in this research with an error of only 2.6%. Thus, a small set of experimentally generated data set allows to achieve regression equations which enable a precise prediction of mechanical properties. Moreover, the validated numerical welding simulation model was suitable to achieve an accurate calculation of the t8/5 cooling time, with an error of only 0.3%.
4 citations
TL;DR: In this article , the development of algorithm-based process models for the mechanical joining process self-pierce riveting with semi-tubular rivet (SPR-ST) is described.
Abstract: Abstract In this paper, the development of algorithm-based process models for the mechanical joining process self-pierce riveting with semi-tubular rivet (SPR-ST) is described. Therefore, an extensive experimental and numerical database regarding the SPR-ST process and strength of steel and aluminum joints with tensile strengths of the sheets between 200 and 1000 MPa was generated for the building of the models. This process data could then be used for the training and evaluation of different prediction algorithms. Furthermore, the simulation data is applied to predict the entire contour (mesh) of non-simulated joints. This includes the visualization of output values such as strains, stresses and damage for each element and node of the mesh. That approach enables to obtain more information about the joint than just discrete values such as interlock or strength.
2 citations
TL;DR: In this article , the numerical data-acquisition and the development of algorithm-based and analytical models for strength prediction for SPR-ST and clinching for production is described, which enables an immediate prediction of the quasi-static joint strength based on the input parameters such as properties of the parts to be joined and process parameters.
Abstract: Currently the design of mechanical joining processes like self-pierce riveting with semi-tubular rivet (SPR-ST) and clinching for production is subject to complex and experimental test series in which process parameters such as the rivet or die geometry are varied iteratively and based on experience until a suitable joint contour and strength is achieved. To simplify the use of mechanical joining technologies these development cycles and thereby the effort for implementation into production must be reduced. In this paper, the numerical data-acquisition and the development of algorithm-based and analytical models for strength prediction for SPR-ST and Clinching is described. Therefore, an extensive experimental and numerical database regarding the SPR-ST process and strength of steel and aluminum joints with tensile strengths of the sheets between 200 - 1000 MPa was generated for the building of the models. This process data could then be used for the training and evaluation of different prediction models. The goal of the research presented here is to enable an immediate prediction of the quasi-static joint strength, based on the input parameters such as properties of the parts to be joined and process parameters.
1 citations
References
More filters
TL;DR: In this paper, two sampling plans are examined as alternatives to simple random sampling in Monte Carlo studies and they are shown to be improvements over simple sampling with respect to variance for a class of estimators which includes the sample mean and the empirical distribution function.
Abstract: Two types of sampling plans are examined as alternatives to simple random sampling in Monte Carlo studies. These plans are shown to be improvements over simple random sampling with respect to variance for a class of estimators which includes the sample mean and the empirical distribution function.
8,328 citations
TL;DR: It is shown that both the approximation accuracy and execution speed of gradient boosting can be substantially improved by incorporating randomization into the procedure.
5,355 citations
Proceedings Article•
05 Dec 2005TL;DR: By applying feature selection, a non-linear, simple, yet effective, feature subset selection method for regression and using it in analyzing cortical neural activity, this work is able to improve prediction quality and suggest a novel way of exploring neural data.
Abstract: We present a non-linear, simple, yet effective, feature subset selection method for regression and use it in analyzing cortical neural activity. Our algorithm involves a feature-weighted version of the k-nearest-neighbor algorithm. It is able to capture complex dependency of the target function on its input and makes use of the leave-one-out error as a natural regularization. We explain the characteristics of our algorithm on synthetic problems and use it in the context of predicting hand velocity from spikes recorded in motor cortex of a behaving monkey. By applying feature selection we are able to improve prediction quality and suggest a novel way of exploring neural data.
99 citations
TL;DR: An artificial Neural Network model is utilized to predict the behavior of clinched joints produced with a given clinching tools configuration and an optimization tool based on a Genetic Algorithm tool was developed to demonstrate the effectiveness of this approach.
65 citations
TL;DR: In this paper, the authors present untersuchungen mit Hilfe eines Lernverfahrens, das auf der k-nachsten-Nachbarn-regression (kNN-Regression) basiert, vorgestellt.
Abstract: Kurzfassung Zur zerstorungsfreien Ermittlung der Tragfahigkeit von geclinchten Verbindungselementen unter quasistatischer Belastung werden die Untersuchungen mit Hilfe eines Lernverfahrens, das auf der k-nachsten-Nachbarn-Regression (kNN-Regression) basiert, vorgestellt. Als Eingangsgrosen des Lernverfahrens dienen die Prozessdaten, die wahrend eines Fugens aus den aktuellen Signalen wie z.B. der Fugekraft und dem Stempelweg aufgenommen und abstrahiert werden. Die zu ermittelnden Ausgangsgrosen sind die Scheroder Schalzugkraft einer Clinchverbindung. In diesen Untersuchungen wurde der Stahlwerkstoff DC04 ZE mit zwei verschiedenen Dicken 0,8 mm und 1,0 mm im Anlieferzustand eingesetzt. Als Clinchverfahren kam das einstufige nicht schneidende Fugeverfahren mit geschlossener Matrize zum Einsatz. Die Ergebnisse zeigten, dass unter Beachtung verschiedener Verwendungsbedingungen die ermittelten Scher- und Schalzugkrafte mit den gemessenen Werten gut ubereinstimmten.
4 citations