Filtern
Dokumenttyp
- Vortrag (6) (entfernen)
Referierte Publikation
- nein (6) (entfernen)
Schlagworte
- Digital Twin (6) (entfernen)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (6)
The future competitiveness of the process industry and their providers depends on its ability to deliver high quality and high value products at competitive prices in a sus-tainable fashion, and to adapt quickly to changing customer needs. The transition of process industry due to the mounting digitalization of technical devices and their pro-vided data used in chemical plants proceeds. Though, the detailed characteristics and consequences for the whole chemical and pharmaceutical industry are still unfore-seeable, new potentials arise as well as questions regarding the implementation. As the digitalization gains pace fundamental subjects like the standardization of device interfaces or organization of automation systems must be answered. Still, process in-dustry lack of sufficient system and development concepts with commercial advantage from this trend.
Compared to traditional batch processes, intensified continuous production allows new and difficult to produce compounds with better product uniformity and reduced consumption of raw materials and energy. Flexible (modular) chemical plants can pro-duce various products using the same equipment with short down-times between campaigns, and quick introduction of new products to the market.
Full automation is a prerequisite to realize such benefits of intensified continuous plants. In continuous flow processes, continuous, automated measurements and closed-loop control of the product quality are required. Consequently, the demand for smart sensors, which can monitor key variables like component concentrations in real-time, is increasing. Low-Field NMR spectroscopy presents itself as such an upcoming smart sensor1,2 (as addressed, e.g., in the CONSENS project3).
Systems utilizing such an online NMR analyzer benefits through short development and set-up times when applied to modular production plants starting from a desired chemical reaction3. As an example for such a modular process unit, we present the design and validation of an integrated NMR micro mixer based on computational mod-elling suited for a desired chemical reaction. This method includes a proper design of a continuous reactor, which is optimized through computational fluid dynamics (CFD) for the demands of the NMR sensor as well as for the given reaction conditions. The system was validated with a chemical reaction process.
References:
[1] M. V. Gomez et al., Beilstein J. Org. Chem. 2017, 13, 285-300
[2] K. Meyer et al., Trends Anal. Chem. 2016, 83, 39-52
[3] S. Kern et al., Anal Bioanal Chem. 2018, 410, 3349-3360
Chemical companies must find new paths to successfully survive in a changing environment. The potential of digital technologies belongs to these. Flexible and modular chemical plants can produce various high-quality products using multi-purpose equipment with short down-times between campaigns and reduce time to market for new products. Intensified continuous production plants allow for difficult to produce compounds. Therefore, fully automated “chemical” process control along with real-time quality control are prerequisites to such concepts while being based on “chemical” information.
As an example, a fully automated NMR sensor is introduced, using a given pharmaceutical lithiation reaction as an example process within a modular pilot plant. Therefore, a commercially available benchtop NMR spectrometer was adapted to the full requirements of an automated chemical production environment such as, e.g., explosion safety, field communication, and robust evaluation of sensor data. It was thereof used for direct loop advanced process control and real-time optimization of the process. NMR appeared as preeminent online analytical tool and allowed using a modular data analysis tool, which even served as reliable reference method for further PAT applications.
A full integration and intelligent interconnection of such systems and processes progresses only hesitantly. The talk should encourage to re-think digitization of process industry based on smart sensors, actuators, and communication more comprehensively and informs about current technical perspectives such as the “one-network paradigm”, edge computing, or virtual machines. These give smart sensors, actuators, and communication a new perspective.
At the Bundesanstalt für Materialforschung und -prüfung (BAM) full scale specimens for nuclear transport and storage containers (casks) are tested for their structural integrity in a series of drop tests on the Test Site Technical Safety in Horstwalde, 50 km south of Berlin. These drop tests cause a major stress not only on the casks, but also on the steel tower structure of the test facility, itself. The load pattern makes the structure very interesting for detailed investigation. The focus of the monitoring lies on the bolted joints of the flange connections that are a typical connection for cylindrical elements if welding is technical or economical unfavorable. The definition of the monitoring takes was done by investigating the existing documents and inspection results accompanied by building an initial digital representation of the structure, consisting of two finite element (FE) models and a geometrical 3D point cloud representation. As a first step the structures behavior during static and dynamic loading was analyzed using measurement data and an updated numerical FE Model. The idea behind is to use models for a digital planning and operation/evaluation of the structural health monitoring. A static FE simulation and a dynamic FE simulation are generated, to investigate how the structure behaves under the load conditions.
In recent years, the use of simulation-based digital twins for monitoring and assessment of complex mechanical systems has greatly expanded. Their potential to increase the information obtained from limited data makes them an invaluable tool for a broad range of real-world applications. Nonetheless, there usually exists a discrepancy between the predicted response and the measurements of the system once built. One of the main contributors to this difference in addition to miscalibrated model parameters is the model error. Quantifying this socalled model bias (as well as proper values for the model parameters) is critical for the reliable performance of digital twins. Model bias identification is ultimately an inverse problem where information from measurements is used to update the original model. Bayesian formulations can tackle this task. Including the model bias as a parameter to be inferred enables the use of a Bayesian framework to obtain a probability distribution that represents the uncertainty between the measurements and the model. Simultaneously, this procedure can be combined with a classic parameter updating scheme to account for the trainable parameters in the original model. This study evaluates the effectiveness of different model bias identification approaches based on Bayesian inference methods. This includes more classical approaches such as direct parameter estimation using MCMC in a Bayesian setup, as well as more recent proposals such as stat- FEM or orthogonal Gaussian Processes. Their potential use in digital twins, generalization capabilities, and computational cost is extensively analyzed.
Simulation-based digital twins have emerged as a powerful tool for evaluating the mechanical response of bridges. As virtual representations of physical systems, digital twins can provide a wealth of information that complements traditional inspection and monitoring data. By incorporating virtual sensors and predictive maintenance strategies, they have the potential to improve our understanding of the behavior and performance of bridges over time. However, as bridges age and undergo regular loading and extreme events, their structural characteristics change, often differing from the predictions of their initial design. Digital twins must be continuously adapted to reflect these changes. In this article, we present a Bayesian framework for updating simulation-based digital twins in the context of bridges. Our approach integrates information from measurements to account for inaccuracies in the simulation model and quantify uncertainties. Through its implementation and assessment, this work demonstrates the potential for digital twins to provide a reliable and up-to-date representation of bridge behavior, helping to inform decision-making for maintenance and management.