Paul Scholz

Over the past six decades, many companies have discovered the potential of computer-controlled systems in the manufacturing industry. Overall, digitization can be identified as one of the main drivers of cost reduction in the manufacturing industry. However, recent advances in Artificial Intelligence indicate that there is still untapped potential in the use and analysis of data in industry. Many reports and surveys indicate that machine learning solutions are slowly adapted and that the… Mehr anzeigen

Over the past six decades, many companies have discovered the potential of computer-controlled systems in the manufacturing industry. Overall, digitization can be identified as one of the main drivers of cost reduction in the manufacturing industry. However, recent advances in Artificial Intelligence indicate that there is still untapped potential in the use and analysis of data in industry. Many reports and surveys indicate that machine learning solutions are slowly adapted and that the process of implementation is decelerated by inefficiencies. The goal of this paper is the systematic analysis of successfully implemented machine learning solutions in manufacturing as well as the derivation of a more efficient implementation approach. For this, three use cases have been identified for in-depth analysis and a framework for systematic comparisons between differently implemented solutions is developed. In all three use cases it is possible to derive implementation patterns as well as to identify key variables which determine the success of implementation. The identified patterns show that similar machine learning problems within the same use case can be solved with similar solutions. The results provide a heuristic for future implementation attempts tackling problems of similar nature. Weniger anzeigen

Due to the increasing digitalization and connection of production areas, more and more companies have large amounts of process and machine data at their disposal. Utilization of this data by machine learning systems (MLS) offers productivity potential and thus combines self-control (adaptivity) with increased economic efficiency. Yet, in industry the diffusion of MLS into practical applications has been slow and the potential is only being tapped in isolated cases. This can be traced back to a… Mehr anzeigen

Due to the increasing digitalization and connection of production areas, more and more companies have large amounts of process and machine data at their disposal. Utilization of this data by machine learning systems (MLS) offers productivity potential and thus combines self-control (adaptivity) with increased economic efficiency. Yet, in industry the diffusion of MLS into practical applications has been slow and the potential is only being tapped in isolated cases. This can be traced back to a missing understanding of the technology-inherent operating principles of machine learning (ML). This paper draws upon the
assumption that types of MLS can be formed to structure the domain of ML. This enables a detailed understanding of MLS and their technology-inherent operating principles in a completely new manner. Thus, a framework for forming types of MLS is proposed and five types of MLS are derived. Furthermore, each type of MLS is detailed by an analysis of real world use cases. Weniger anzeigen

The increasing digitalization across all business sectors creates ever larger amounts of data. When analyzing this data to extract information, traditional data analysis methods easily meet their limits of performance. Conversely, methods from the machine learning (ML) spectrum promise to be a versatile tool for solving highly complex data-related tasks. Yet, organizations fail to identify relevant applications for ML due to a lack of systematic understanding of the applicability of the… Mehr anzeigen

The increasing digitalization across all business sectors creates ever larger amounts of data. When analyzing this data to extract information, traditional data analysis methods easily meet their limits of performance. Conversely, methods from the machine learning (ML) spectrum promise to be a versatile tool for solving highly complex data-related tasks. Yet, organizations fail to identify relevant applications for ML due to a lack of systematic understanding of the applicability of the technology. This research paper draws upon the assumption that real-world problems exhibit distinctive characteristics that indicate their suitability for the application of ML. A framework for describing those constituent problem characteristics is proposed. Investigating the functional differences between traditional data analysis methods and ML, constituent characteristics are derived based on the distinctive technological abilities of ML. In order to differentiate simple ML methods from advanced ML methods regarding their technological abilities, a framework is presented. We suggest that advanced ML methods such as Deep Learning or Transfer Learning provide different potential benefits than simple methods such as Decision Trees, therefore necessitating this additional distinction. Weniger anzeigen

In an increasingly dynamic and complex environment, manufacturing systems must respond quickly to changes in order to remain productive. Hence, existing tasks and decisions in manufacturing have to be aligned in an ever more complex system of connected and dependent machines and devices. Various data-enabled assistance systems that help to coordinate tasks and support decisions are already existent. However, due to the volatile, uncertain, complex, and ambiguous (VUCA) environment, further… Mehr anzeigen

In an increasingly dynamic and complex environment, manufacturing systems must respond quickly to changes in order to remain productive. Hence, existing tasks and decisions in manufacturing have to be aligned in an ever more complex system of connected and dependent machines and devices. Various data-enabled assistance systems that help to coordinate tasks and support decisions are already existent. However, due to the volatile, uncertain, complex, and ambiguous (VUCA) environment, further demands for the assistance systems emerge. Increasing availability of data and decreased costs for computing such as storage and computing capacities for the use of machine learning (ML) indicate promising potential to address the ever-changing conditions securing productivity and thus remain competitive. But the promising potential widely remains untapped. The challenges faced by manufacturing companies especially lie in the identification of attractive application areas and the recognition of the associated learning tasks. Therefore, the aim of this paper is to derive and systematize challenges in the future VUCA-submissive manufacturing landscape to effectively design ML applications. The results provide a target system of objectives in which challenges can be positioned, such as an application navigator for archetypical challenges to be addressed by ML. Weniger anzeigen

Although machine learning (ML) methods have already been well described in science, the transfer into manufacturing business practice is only slowly taking place. One of the reasons is that current research is lacking a comprehensive analysis of working ML methods and their characteristics. Therefore, this paper systematically analyzes successfully implemented ML solutions to facilitate the design process for future machine learning systems (MLS) implementations in manufacturing companies… Mehr anzeigen

Although machine learning (ML) methods have already been well described in science, the transfer into manufacturing business practice is only slowly taking place. One of the reasons is that current research is lacking a comprehensive analysis of working ML methods and their characteristics. Therefore, this paper systematically analyzes successfully implemented ML solutions to facilitate the design process for future machine learning systems (MLS) implementations in manufacturing companies. First, a systematic literature review based on 18 scientific publications is conducted to confirm the lacks assumed. Second, 15 MLS approaches are analyzed based on a technology framework to solve the shortcomings identified and extract further findings. In total, we identified two general MLS design patterns. Furthermore, we extracted seven suitable data and data model requirements as well as technology potentials. The results show that theory-based ML approaches are often based on linear methods requiring low-dimensional data, e.g. in image recognition. This points towards the hypothesis, that the application of non-linear ML methods processing high-dimensional data could increase the number of possible use cases. Thus, further high technology potentials regarding the application of MLS in the manufacturing industry would arise. Weniger anzeigen

With an increasing digitalization of more and more domains, the need to use the available data accordingly increases. To do so in complex application domains (e.g. mobility or production) the data must often be processed with high accuracy and in real time. As machine learning systems (MLS) have the potential to meet those requirements, they are increasingly relevant. However, it is difficult for potential solution providers (e.g. innovation managers) to fully assess the technology potential of… Mehr anzeigen

With an increasing digitalization of more and more domains, the need to use the available data accordingly increases. To do so in complex application domains (e.g. mobility or production) the data must often be processed with high accuracy and in real time. As machine learning systems (MLS) have the potential to meet those requirements, they are increasingly relevant. However, it is difficult for potential solution providers (e.g. innovation managers) to fully assess the technology potential of MLS, to identify relevant cases in complex application domains and thus to develop functional solutions. This research work is therefore concerned with the development of a methodology for the systematic design of MLS for complex application domains. In a first step, problem and solution domains are analyzed and deficits of existing approaches to the systematic design of MLS are identified. From this, requirements for a design methodology are derived and a specific solution concept is developed. Finally, the individual steps are linked in a basic concept for a MLS design methodology. Weniger anzeigen

A significant amount of data is generatedand could be utilized in order to improve quality, time, and cost related performance characteristics of the production process. Machine Learning (ML) is considered as a particularly effective method of data processing with the aim of generating usable knowledge from data and therefore becomes increasingly relevant in manufacturing. In this research paper, a technology framework is created that supports solution providers in the development and… Mehr anzeigen

A significant amount of data is generatedand could be utilized in order to improve quality, time, and cost related performance characteristics of the production process. Machine Learning (ML) is considered as a particularly effective method of data processing with the aim of generating usable knowledge from data and therefore becomes increasingly relevant in manufacturing. In this research paper, a technology framework is created that supports solution providers in the development and deployment process of ML applications. This framework is subsequently successfully employed in the development of an ML application for quality prediction in a machining process of Bosch Rexroth AG.For this purpose the 50 mostrelevant features were extracted out of time series data and used to determine the best ML operation. Extra Tree Regressor (XT) is found to achieve precise predictions with a coefficient of determination (R2) of constantly over 91% for the considered quality characteristics of a boreof hydraulic valves. Weniger anzeigen

The technological change in the form of the electrification of vehicle concepts is making its way into the mobility sector - increasing legal regulation is promoting this trend. For example, for locally emission-free driving in inner cities, electric mobility is absolutely essential. For electrified driving, there are various technology options of varying degrees of maturity. Battery storage technologies (especially lithium-ion batteries) are currently receiving the most attention. In addition… Mehr anzeigen

The technological change in the form of the electrification of vehicle concepts is making its way into the mobility sector - increasing legal regulation is promoting this trend. For example, for locally emission-free driving in inner cities, electric mobility is absolutely essential. For electrified driving, there are various technology options of varying degrees of maturity. Battery storage technologies (especially lithium-ion batteries) are currently receiving the most attention. In addition to technological advantages, however, batteries also have a number of systematic disadvantages, such as a high constant power-to-weight ratio, very limited cold start capability or time-consuming recharging times, which represent an obstacle to the use of the technology in numerous application scenarios. Especially in the commercial vehicle sector these application scenarios can be identified. A possible end to the technological openness in favour of battery technologies, as is being promoted by some OEMs, among other things due to the higher level of technological maturity, provides them with planning security and greater design freedom, but also prevents the further development of potential alternative energy storage technologies such as the fuel cell. The main obstacle to a more intensive use of fuel cell technology in the commercial vehicle sector is the relatively low level of technological maturity and a lack of competitiveness in terms of TCO in this cost-sensitive industry. Against the background of largely production technology challenges in the further technological development of the fuel cell and the extreme scalability of certain key components, a more in-depth examination of this energy storage technology in particular is appropriate. Weniger anzeigen

An increasing number of electrified vehicles can be expected across all vehicle classes. The growing importance of electrified vehicle concepts is causing major changes in the automotive industry in general and in the German automotive industry in particular. The technologies of electric vehicle concepts differ fundamentally from those powered by combustion engines - many parts and components are no longer needed in the electrified powertrain. Although conventional vehicle components will be… Mehr anzeigen

An increasing number of electrified vehicles can be expected across all vehicle classes. The growing importance of electrified vehicle concepts is causing major changes in the automotive industry in general and in the German automotive industry in particular. The technologies of electric vehicle concepts differ fundamentally from those powered by combustion engines - many parts and components are no longer needed in the electrified powertrain. Although conventional vehicle components will be needed in the long term in automotive engineering, their technical and economic importance is gradually declining. In Germany, the production of a motor vehicle is based on a division of labor along various stages of the value chain, with automotive suppliers contributing around three-quarters of the total value added. These are focused on the manufacture of a technology-specific product range and are also often medium-sized. Current know-how in the field of conventional drives is partially devalued. A high degree of division of labor across different supplier levels and the associated specialization, particularly at lower levels, could mean that such technologically narrowly focused suppliers will have to revise their technology and product portfolios in the medium term if they want to continue to participate in market growth in the future. For this reason, increasing consolidation is also expected in the automotive industry. The paper provides a conceptual approach (already successfully tested in industrial practice) to quantify the shifts in the value-added structure between different market participants, along the various value-added steps and between existing vehicle systems. Weniger anzeigen

In any industrial sector an increasing number of interconnected objects along with more sensors relying on shortened query rates cause large data volumes that can be utilized for product and process improvement. Methods from the Artificial Intelligence (AI) technology spectrum have the potential
to uncover complex interdependencies in data sets instantly, improve analysis results steadily and adjust to changing external factors dynamically. AI is a heterogeneous technology bundle mainly… Mehr anzeigen

In any industrial sector an increasing number of interconnected objects along with more sensors relying on shortened query rates cause large data volumes that can be utilized for product and process improvement. Methods from the Artificial Intelligence (AI) technology spectrum have the potential
to uncover complex interdependencies in data sets instantly, improve analysis results steadily and adjust to changing external factors dynamically. AI is a heterogeneous technology bundle mainly originating from statistics, advanced analytics and machine learning (ML), which is built up in different layers.
Current research is lacking a comprehensive analysis of these different AI technology layers and their corresponding characteristics that can serve as an orientation guideline especially for manufacturing companies. This research paper derives a nomenclature for the AI technology ecosystem in order to facilitate the discussion of this topic. Moreover, a systematic framework (morphology) is derived in order to classify current AI applications and to identify crucial AI technology composition patterns that might be helpful for future AI application development. Potentially promising scopes for the derivation of AI
technology composition patterns are discussed and exemplary settings for employment of the proposed method are evaluated. Weniger anzeigen