Search

Besides the sequence itself also additional factors serving as moderator variables affect the value of scheduling objectives. For mixed-model assembly lines, especially number and heterogeneity of different products, their volume mix proportions, average workload of the jobs to process and the degree of grouping of identical jobs within the sequence play a major role. By means of a simulation study based on data from a real unpaced mixed-model assembly line in the automotive industry, this work analyzes the impact of these moderating variables on various scheduling objectives. The analyzed scheduling objectives encompass flow-related objectives like mean flow time, productivity-related objectives like makespan, customer-related objectives like mean earliness, the supplier-related objective part usage rate variation and the human-related objectives mean learning effect and mean deterioration effect per job. Simulation scenarios are defined that differ regarding number and heterogeneity of products from three homogeneous to seven more heterogeneous products. Within every simulation scenario the volume mix proportions of the products, and inherently also the average workload of jobs, are systematically varied. Every simulation scenario is analyzed for five sequence types differing in the degree of grouping of identical jobs. For almost all scheduling objectives, strong dependencies on the volume mix proportions can be perceived, particularly for mean flow time. Homogeneous volume mixes with a dominating product in the mix often lead to other objective values compared to heterogeneous volume mixes that allow using alternation effects between various products in a sequence. Concerning degree of grouping, while some scheduling objectives like part usage rate variation are always strongly affected by the degree of grouping for every volume mix, other objectives like throughput show strong dependence only for some mixes and makespan does not even show any tendency. Average

The ongoing digitization of online learning resources has led to a proliferation of collaboration platforms for specific areas of application and disciplines. Simultaneously, the need for collaboration in the field of data analysis and knowledge management increases, especially for manufacturing companies. In this paper, collaborative and competency-based requirements for applying industrial data analytics are derived into tangible specifications for implementing a collaboration platform. The currently absent requirements of Industrial Data Science (IDS) projects are determined and then translated into platform-specific functions. Using the practical implementation of an ongoing research project, the defined requirements are transformed into features and applied in an online platform. The validation in a system of dynamic value networks serves to confirm the practical suitability of the specified requirements. The innovation of the platform is its clear focus on IDS project practitioners, who are typically comprised of several different domains. They often use a variety of analytics tools and aim for long-term use of deployed solutions. A pilot implementation of the developed platform is available online and additionally serves to validate the platform suitability.

In recent years, there has been an increasing interest in using industrial data science (IDS) in manufacturing companies. Structured IDS projects proceed according to process models such as the cross industry standard process for data mining (CRISP-DM), knowledge discovery in databases (KDD), or the process chain of industrial data science. Because of the process Chain’s transdisciplinary procedure, the participation of different people in the analysis with different tasks and competencies is necessary. Therefore, a concept to define specific roles is required, since it provides unequivocal descriptions of the respective tasks. As no role concept for the process chain of IDS exists, this paper develops and presents a transdisciplinary role concept including the four essential roles: Data engineer, analyst, user, and project manager. These roles are described in terms of their characteristics to enable structured cooperation between them. In addition, this paper presents the AKKORD platform for learning and collaboration, which especially should make an important contribution for small and medium-sized enterprises (SME) to develop knowledge in the usage of the process chain of IDS. The platform provides the opportunity for the users to train the essential roles with their characteristics in the company through targeted competence development.

Companies with Make to Order MTO manufacturing have always faced the conflict of meeting a large volume of individual customer orders on time while remaining as flexible as possible Unlike Make to Stock MTS manufacturing planning the production requirements for MTO is a challenging task since incoming orders may vary in time and quantity while also being subject to a number of variables In some cases the delivery time allocated by the customer can be less than the required Order Lead Time to fulfil the order Manufacturers can respond to this with either with approached from production management or from data science This paper presents a framework to leverage the benefit of both domains We conduct a literature review and present the results of an expert workshop We propose criteria for a suitable data enrichment and the application of Machine Learning ML methods in sales and demand forecasting In conclusion the proposed framework helps to equip manufacturing companies with a structured strategy for data management and to utilize the benefit of ML for sales and demand forecasting

Since bolted joints are ubiquitous in manufacturing their effective and reliable quality assurance is particularly important Most tightening processes rely on statistical methods to detect faulty screw connections already during assembly In this paper we address the detection of faulty tightening processes using a clustering based approach from the field of Unsupervised Machine Learning In particular we deploy the k Means algorithm on a real world dataset from the automotive industry The model uses Dynamic Time Warping to determine the similarity between the normal and abnormal tightening processes treating each one as an independent temporal sequence This approach offers three advantages compared to existing supervised methods 1 time series with different lengths can be utilized without extensive preprocessing steps 2 errors never seen before can be detected using the unsupervised approach and 3 extensive manual efforts to generate labels are no longer necessary To evaluate the approach it is applied in a scenario where actual class labels are available This allows evaluating the clustering results using traditional classification scores The approach manages to achieve an accuracy of up to 88 89 and a macro average F1 score of up to 63 65

Industrial Engineering, through its role as design, planning and organizational body of the industrial production, has been crucial for the success of manufacturing companies for decades. The potential, expected over the course of Industry 4.0 and through the application of Data Analytic tools and methods, requires a coupling to established methods. This creates the necessity to extend the traditional job description of Industrial Engineering by new tools from the field of Data Analytics, namely Industrial Data Science. Originating from the historic pioneers of Industrial Engineering, it is evident that the basic principles will remain valuable. However, further development in view of the data analytic possibilities is already taking place. This paper reviews the origins of Industrial Engineering with reference to four pioneers, draws a connection to current day usage, and considers possibilities for future applications of Industrial Data Science.

Manually defined control limits remain a common strategy for quality control in manufacturing due to their ease of deployment on the shop floor compared to more advanced data analysis approaches. Despite their continued importance, there is no systematic method of defining these control limits. However, sub-optimal control limits can lead to undetected faults or cause unnecessary interruption to production. This manuscript presents an algorithm that systematizes this manual process into an efficient search task. We conceptualized the search task as a sequence of sub-problems that are based on the conventional steps taken by process experts when defining control limits. This algorithm can be integrated into an expert tool for shop floor personnel to automate the definition of control limits in annotated time series data. We demonstrate the efficacy of the control limits found by our algorithm by comparing them to those manually defined by process experts in real-world process data from the automotive industry. Furthermore, we show that our algorithm generalizes to traditional time series classification problems and achieves state-of-the-art performance on selected benchmark datasets. Our work is the first effort in automating the otherwise manual definition of control limits for fault detection.

Many real-world big data applications in domains like finance, telecommunication and manufacturing rely on the detection of exceedingly rare patterns in large time series data sets. In principle, machine learning models can be trained to detect and classify such patterns. However, these models often lack the necessary robustness for practical applications and do not generalize well in production. Additionally, their intransparent decision-making hampers systematic debugging and improvement. Time series shapelets are a popular data mining primitive that can be used to extract a shape-based feature representation from the data. However, existing algorithms do not adequately consider the robustness and redundancy of these features. While these drawbacks can be compensated if sufficient labeled data is available, this is not possible for highly imbalanced data. We propose alterations to the current state-of-the-art shapelet algorithm that consider the margin of separation and the multivariate dependencies between the extracted features. This results in more robust and diverse features, which in turn translates to higher classification accuracy. We compare our algorithm to the current state-of-the-art using a public benchmark data set. Additionally, we showcase its applicability to highly imbalanced data using a suitable data set from the manufacturing domain.

Improving the recall of information retrieval systems for similarity search in time series databases is of great practical importance. In the manufacturing domain, these systems are used to query large databases of manufacturing process data that contain terabytes of time series data from millions of parts. This allows domain experts to identify parts that exhibit specific process faults. In practice, the search often amounts to an iterative query–response cycle in which users define new queries (time series patterns) based on results of previous queries. This is a well-documented phenomenon in information retrieval and not unique to the manufacturing domain. Indexing manufacturing databases to speed up the exploratory search is often not feasible as it may result in an unacceptable reduction in recall. In this paper, we present a novel adaptive search algorithm that refines the query based on relevance feedback provided by the user. Additionally, we propose a mechanism that allows the algorithm to self-adapt to new patterns without requiring any user input. As the search progresses, the algorithm constructs a library of time series patterns that are used to accurately find objects of the target class. Experimental validation of the algorithm on real-world manufacturing data shows, that the recall for the retrieval of fault patterns is considerably higher than that of other state-of-the-art adaptive search algorithms. Additionally, its application to publicly available benchmark data sets shows, that these results are transferable to other domains.

Capacity-limiting bottlenecks in manufacturing systems form the ideal starting point for measures of improvement. However, the inherent variability of modern systems leads to dynamic bottleneck behavior, causing them to shift between stations. Numerous methods for the detection of shifting bottlenecks exist in literature. In this paper, we present and compare three methods: Bottleneck Walk (BNW), Active Period Method (APM), and an adaptation of Interdeparture Time Variances (ITV). The comparative study deploys the methods in a serial production line with seven stations and eight buffers. We vary the individual locations of the bottlenecks by adding more process time. To compare the methods, we determine the overall average ratio of agreement between the three detection methods. APM and ITV have the highest agreement at an average of 80.10%. Pairings with BNW achieve significantly lower rates of agreement, with 56.33% for ITV, and 62.03%% when compared to the APM.

Numerous methods for bottleneck detection, along novel approaches for bottleneck prediction, are available in literature. To facilitate the development and application of such methods, this paper proposes a holistic methodology for Bottleneck Analysis in dynamic value streams. Analogous to established data analytics levels, namely descriptive, diagnostic, predictive, and prescriptive analytics, the methodology specifies objectives for data-driven Bottleneck Analysis. Based on state-of-the-art bottleneck detection methods, the methodology provides measures for the diagnosis of bottleneck severity and frequency. Additionally, it considers prediction methods to anticipate emerging bottlenecks, depending on available databases. Finally, the methodology provides a context for the yet unexplored field of bottleneck prescription, which aims to mitigate bottleneck effects by data-driven control recommendations. Further practical application of the methodology has to confirm its suitability as a holistic framework for analyzing bottlenecks in dynamic value streams.