Search

Mit der Einführung von gesellschaftlichen und politischen Vorgaben wie dem „European Green Deal“ wird die Umsetzung der Kreislaufwirtschaft zur Unternehmensaufgabe. Deren wirtschaftliche Umsetzung erfordert jedoch die Anpassung industrieller Prozessabläufe hinsichtlich des angestrebten Grades an Adaptivität an eine dann zunehmende Varianz von Eingangsmaterialien in die industrielle Wertschöpfung. Strategien wie das Remanufacturing bieten nicht nur ein ökologisches Potenzial, sondern bewirken eine Steigerung der Resilienz und eröffnen dadurch neue wirtschaftliche Geschäftsperspektiven. Innerhalb des Beitrags wird die Bedeutung geeigneter Demontagesysteme für die Kreislaufwirtschaft aufgezeigt. Aufbauend auf den spezifischen Herausforderungen der Kreislaufstrategien wird ein Zielbild für wirtschaftliche Demontagesysteme abgeleitet. Drei industrielle Use Cases zeigen verschiedene Ausprägungen solcher Systeme auf. Es besteht der Bedarf, die Demontage zu automatisieren und gleichzeitig die Adaptivität des Systems zu maximieren, um verschiedenen Produktzuständen gerecht zu werden. Die adaptive automatisierte Demontage stellt hierfür die zentrale Befähigung dar. Zur Umsetzung dienen Befähigertechnologien wie der Einsatz von Sensorik und KI, aber auch das „Design for Disassembly“. Aus organisatorischer Perspektive sind Standardisierungen notwendig, um die Verfügbarkeit von Daten sowohl aus der Produktion, dem User-Zyklus als auch aus automatisierten Zustandsbewertungen als ein Entscheidungsfundament für adaptive Prozesse nutzbar zu machen. Die Bereitstellung und Analyse der notwendigen Daten stellt folglich ein zentrales Anliegen für die Umsetzung der Demontageprozesse dar., With the introduction of social and political requirements such as the "European Green Deal", the implementation of the circular economy is becoming a corporate task. However, its economic implementation as an answer to regulatory demands requires the adaptation of industrial processes regarding the desired degree of adaptivity to a then increasing variance of input materials in industrial value creation. Strategies such as re-manufacturing not only offer ecological potential, but also bring about an increase in resilience and thus open new economic business perspectives. Within the contribution, the importance of suitable disassembly systems including dismantling abilities for the circular economy is shown. Based on the specific challenges of the circular strategies, a target picture for economic disassembly systems is derived. Three industrial use cases show different forms of such systems. There is a need to automate disassembly while maximizing the adaptivity of the system to cope with different product states. Adaptive automated disassembly is the key enabler for this. Enabling technologies such as the use of sensors, AI or "design for disassembly", serve to implement this. From an organizational perspective, standardization is necessary to make data from production, from the user cycle and from automated condition assessments available and usable as a decision-taking foundation for adaptive processes. Consequently, the provision and analysis of the necessary data represents a central concern for the implementation of disassembly processes.

The era of Industry 4.0 opens up the possibility of optimizing production systems in a data-driven way. To turn data into value, machine learning (ML) models are trained on production data aiming at identifying patterns to optimize processes. A crucial prereq-uisite for achieving performant ML models is the availability of high quality data. Since raw data generated in production exhibits multiple quality issues, data preprocessing (DPP) is required to increase the quality of the data. One of the key design decisions in any ML project is the choice of suitable DPP methods. The search space further increases when DPP methods are configured into DPP pipelines. Due to the high num-ber of possible DPP pipelines, data scientists commonly select suitable pipelines man-ually and via trial and error. For these reasons, DPP nowadays accounts for approximately 80 % of the time in ML projects.To guide data scientists, decision support systems (DSS) have been developed that assist in the selection of suitable DPP pipelines but do not cover productionspecific requirements. Therefore, the main research question was: Can a DSS be developed that supports in recommending DPP pipelines for ML applications in production? To be able to answer the main research question, a meta learning-based decision sup-port system, called Meta-DPP, was developed. Meta-DPP relies on three core compo-nents: the meta target selector, meta features database, and meta model. The meta target selector chooses between two preselected sets of overall well performing pipe-lines, called pipeline pools, for both classification and regression tasks. Further, the meta features database stores learning taskspecific information about the data set, e. g., the number of instances, as well as past ML algorithm and DPP pipeline performances. The meta model then recommends a pipeline from the pipeline pool based on the meta features from the database. When applying Meta-DPP, a user interface enables the data scientist, production expert, or IT expert to input their data set, learning task, ML algorithm and information about explainability. Given these four inputs, Meta-DPP provides a ranked recommendation of the DPP pipelines from the pool. Probabilities provided by the meta model further indicate how certain Meta-DPP is about the recommendation. Verifying and validating revealed the correct development and implementation of Meta-DPP. The validation on 324 production use cases further prove that Meta-DPP outperform essential pipelines on average, whereby essential pipelines ensure the function-ing of ML algorithms by performing minimum DPP. As a conclusion, the main research question was positively answered.

Dissertation, RWTH Aachen University, 2022; Aachen : Apprimus Verlag, Ergebnisse aus der Produktionstechnik 3/2023, 1 Online-Ressource : Illustrationen, Diagramme (2023). = Dissertation, RWTH Aachen University, 2022, The era of Industry 4.0 opens up the possibility of optimizing production systems in a data-driven way. To turn data into value, machine learning (ML) models are trained on production data aiming at identifying patterns to optimize processes. A crucial prereq-uisite for achieving performant ML models is the availability of high quality data. Since raw data generated in production exhibits multiple quality issues, data preprocessing (DPP) is required to increase the quality of the data. One of the key design decisions in any ML project is the choice of suitable DPP methods. The search space further increases when DPP methods are configured into DPP pipelines. Due to the high num-ber of possible DPP pipelines, data scientists commonly select suitable pipelines man-ually and via trial and error. For these reasons, DPP nowadays accounts for approximately 80 % of the time in ML projects.To guide data scientists, decision support systems (DSS) have been developed that assist in the selection of suitable DPP pipelines but do not cover productionspecific requirements. Therefore, the main research question was: Can a DSS be developed that supports in recommending DPP pipelines for ML applications in production? To be able to answer the main research question, a meta learning-based decision sup-port system, called Meta-DPP, was developed. Meta-DPP relies on three core compo-nents: the meta target selector, meta features database, and meta model. The meta target selector chooses between two preselected sets of overall well performing pipe-lines, called pipeline pools, for both classification and regression tasks. Further, the meta features database stores learning taskspecific information about the data set, e. g., the number of instances, as well as past ML algorithm and DPP pipeline performances. The meta model then recommends a pipeline from the pipeline pool based on the meta features from the database. When applying Meta-DPP, a user interface enables the data scientist, production expert, or IT expert to input their data set, learning task, ML algorithm and information about explainability. Given these four inputs, Meta-DPP provides a ranked recommendation of the DPP pipelines from the pool. Probabilities provided by the meta model further indicate how certain Meta-DPP is about the recommendation. Verifying and validating revealed the correct development and implementation of Meta-DPP. The validation on 324 production use cases further prove that Meta-DPP outperform essential pipelines on average, whereby essential pipelines ensure the function-ing of ML algorithms by performing minimum DPP. As a conclusion, the main research question was positively answered., Published by Apprimus Verlag, Aachen

Dissertation, RWTH Aachen University, 2022; Aachen : Apprimus Verlag, Ergebnisse aus der Produktionstechnik Band 8/2023, 1 Online-Ressource : Illustrationen, Diagramme (2023). doi:10.18154/RWTH-2023-02364 = Dissertation, RWTH Aachen University, 2022, The era of Industry 4.0 opens up the possibility of optimizing production systems in a data-driven way. To turn data into value, machine learning (ML) models are trained on production data aiming at identifying patterns to optimize processes. A crucial prereq-uisite for achieving performant ML models is the availability of high quality data. Since raw data generated in production exhibits multiple quality issues, data preprocessing (DPP) is required to increase the quality of the data. One of the key design decisions in any ML project is the choice of suitable DPP methods. The search space further increases when DPP methods are configured into DPP pipelines. Due to the high num-ber of possible DPP pipelines, data scientists commonly select suitable pipelines man-ually and via trial and error. For these reasons, DPP nowadays accounts for approximately 80 % of the time in ML projects.To guide data scientists, decision support systems (DSS) have been developed that assist in the selection of suitable DPP pipelines but do not cover productionspecific requirements. Therefore, the main research question was: Can a DSS be developed that supports in recommending DPP pipelines for ML applications in production? To be able to answer the main research question, a meta learning-based decision sup-port system, called Meta-DPP, was developed. Meta-DPP relies on three core compo-nents: the meta target selector, meta features database, and meta model. The meta target selector chooses between two preselected sets of overall well performing pipe-lines, called pipeline pools, for both classification and regression tasks. Further, the meta features database stores learning taskspecific information about the data set, e. g., the number of instances, as well as past ML algorithm and DPP pipeline performances. The meta model then recommends a pipeline from the pipeline pool based on the meta features from the database. When applying Meta-DPP, a user interface enables the data scientist, production expert, or IT expert to input their data set, learning task, ML algorithm and information about explainability. Given these four inputs, Meta-DPP provides a ranked recommendation of the DPP pipelines from the pool. Probabilities provided by the meta model further indicate how certain Meta-DPP is about the recommendation. Verifying and validating revealed the correct development and implementation of Meta-DPP. The validation on 324 production use cases further prove that Meta-DPP outperform essential pipelines on average, whereby essential pipelines ensure the function-ing of ML algorithms by performing minimum DPP. As a conclusion, the main research question was positively answered., Published by Apprimus Verlag, Aachen

Steigende Rechenleistungen und bessere Datengrundlagen bei gleichzeitig sinkenden Kosten für Rechen- und Speicherkapazitäten stellen die Basis für den Einsatz von Machine Learning (ML) in der Produktion dar. Herausforderungen bestehen in der Identifizierung aussichtsreicher Anwendungsgebiete, dem Erkennen der mit diesen verbundenen Learning Tasks sowie dem Aufdecken passender Datensätze. In diesem Beitrag werden daher folgende Fragen beantwortet: Welche Anwendungsgebiete in der Produktion bieten das größte Potenzial für den Einsatz von ML? Welche frei zugänglichen Datensätze eignen sich, um eigene Erfahrungen zu sammeln und welche Learning Tasks sind damit verbunden? Was sind Best Practices für die Anwendungsgebiete?