Catálogo de publicaciones - libros

The proliferation of cyber-physical systems and the advancement of Internet of Things technologies have led to an explosive digitization of the industrial sector. Driven by the high-tech strategy of the federal government in Germany, many manufacturers across all industry segments are accelerating the adoption of cyber-physical system and Internet of Things technologies to manage and ultimately improve their industrial production processes. In this work, we are focusing on the EU funded project MONSOON, which is a concrete example where production processes from different industrial sectors are to be optimized via data-driven methodology. We show how the particular problem of waste quantity reduction can be enhanced by means of machine learning. The results presented in this paper are useful for researchers and practitioners in the field of machine learning for cyber-physical systems in data-intensive Industry 4.0 domains.

With the onset of ICT and big data capabilities, the physical asset and data computation is integrated in manufacturing through Cyber Physical Systems (CPS). This strategy also denoted as Industry 4.0 will improve any kind of monitoring for maintenance and production planning purposes. So-called bigdata approaches try to use the extensive amounts of diffuse and distributed data in production systems for monitoring based on artificial neural networks (ANN). These machine learning approaches are robust and accurate if the data base for a given process is sufficient and the scope of the target functions is curtailed. However, a considerable proportion of high-performance manufacturing is characterized by permanently changing process, workpiece and machine configuration conditions, e.g. machining of large workpieces is often performed in batch sizes of one or of a few parts. Therefore, it is not possible to implement a robust condition monitoring based on ANN without structured data-analyses considering different machine states – e.g. a certain machining operation for a certain machine configuration. Fuzzy-clustering of machine states over time creates a stable pool representing different typical machine configuration clusters. The time-depending adjustment and automatized creation of clusters enables monitoring and interpretation of machine tool characteristics independently of single machine states and pre-defined processes.

With an ongoing digital transformation towards industry 4.0 and the corresponding growth of collected sensor data based on cyberphysical systems, the need for automatic data analysis in industrial production lines has increased drastically. One relevant application scenario is the usage of intelligent approaches to anticipate upcoming failures for maintenance. In this paper, we present a novel approach for anomaly detection regarding predictive maintenance in an industrial data-intensive environment. In particular, we are focusing on historical sensor data from a real reflow oven that is used for soldering surface mount electronic components to printed circuit boards. The sensor data, which is provided within the scope of the EU-Project COMPOSITION (under grant no. 723145), comprises information about the heat and the power consumption of individual fans inside a reflow oven. The data set contains timeannotated sensor measurements in combination with additional process information over a period of more than seven years.

This paper presents an application of a random forest based classifier that aims at recognizing flawed products in a highly automated production environment. Within the course of this paper, some data set and application features are highlighted that make the underlying classification problem rather complex and hinders the usage of machine learning algorithms straight out-of-the-box. The findings regarding these features and how to treat the concluded challenges are highlighted in a abstracted and generalized manner.

Modern water system infrastructures are equipped with a large amount of sensors. In recent years machine-learning (ML) algorithms became a promising option for data analysis. However, currently ML algorithms are not frequently used in real-world applications. One reason is the costly and time-consuming integration and maintenance of ML algorithms by data scientists. To overcome this challenge, this paper proposes a generic, adaptable platform for real-time data analysis in water distribution networks. The architecture of the platform allows to connect to different types of data sources, to process its measurements in realtime with and without ML algorithms and finally pushing the results to different sinks, like a database or a web-interface. This is achieved by a modular, plugin based software architecture of the platform. As a use-case, a data-driven anomaly detection algorithm is used to monitor the water quality of several water treatment plants of the city of Berlin.

Due to the increase in digitalization Machine Learning (ML)- algorithms bare high potentials for process optimization in the production quality- domain. Nowadays, ML-algorithms are hardly implemented in the production environment. In this paper, we present a tangible use case in which MLalgorithms are applied for predicting the quality of products in a process chain and present the lessons learned we extracted from the application. In the described project, the process of choosing ML-algorithms was a bottleneck. Therefore we describe a promising approach how a decision making tool can help selecting ML-algorithms problem-specifically.

Small humps on the floor go beyond the detectable scope of laser scanners and are therefore not integrated into SLAM based maps of mobile robots. However, even such small irregularities can have a tremendous effect on the robot’s stability and the path quality. As a basis to develop anomaly detection algorithms, kinematics data is collected exemplarily for an overrun of a cable channel and a bulb plate. A recurrent neuronal network (RNN), based on the autoencoder principle, could be trained successfully with this data. The described RNN architecture looks promising to be used for realtime anomaly detection and also to quantify path quality.

Deep learning (DL) is one of the key technologies in the artificial intelligence (AI) domain Deep learning neural networks (DLNN) profit a lot from the overall exponential data growth while on the other hand the computational effort for training and inference strongly increase. Most of the computational time in DLNN is consumed by the convolution step, which is based on a general matrix multiplication (GEMM). In order to accelerate the computational time for DLNN different highly optimized GEMM implementations for Graphic Processing Units (GPUs) have been presented in the last years [1] most of these approaches are GPU hardware specific implementations of the GEMM software kernel and do not incorporate the performance dependency of the training data layout. In order to achieve a maximum performance the parameters of the GEMM algorithm have to be tuned for the different GPU hardware and specific data layout of the training task. In this paper we present a two step autotuning approach for GPU based GEMM algorithms. In the first step the kernel parameter search space is pruned by several performance criteria and afterwards further processed by a modified Simulated Annealing in order to find the best kernel parameter combinations with respect to the GPU hardware and the task specific data layout. Our results were carried out on 160 different input problems with the proposed approach an average speedup against the state of the art implementation from NVIDIA (cuBLAS) from around 12 on a NVIDIA GTX 1080 Ti accelerator card can be achieved.

Today, the optimization of the press hardening process is still a complex and challenging task. This report describes the combination of linear regression with least squares optimization to adjust the process parameters of this process for quality improvement. The FE simulation program AutoForm was used to model the production line concerned and various process and quality parameters were measured. The proposed system is capable of automatically adjusting the process parameters of following process steps based on the quality estimate at each step of the production line. An additional benefit is the identification of likely defective parts early in the production process. Based on the results derived from 1000 observations a better understanding of the process was obtained and in the future the combined regression and optimization approach can be extended to more complex production lines.

Digital transformation and evolution of integrated computational and visualisation technologies lead to new opportunities for reinforcing knowledge-based maintenance through collection, processing and provision of actionable information and recommendations for maintenance operators. Providing actionable information regarding both corrective and preventive maintenance activities at the right time may lead to reduce human failure and improve overall efficiency within maintenance processes. Selecting appropriate digital assistance systems (DAS), however, highly depends on hardware and IT infrastructure, software and interfaces as well as information provision methods such as visualization. The selection procedures can be challenging due to the wide range of services and products available on the market. In particular, underlying machine learning algorithms deployed by each product could provide certain level of intelligence and ultimately could transform diagnostic maintenance capabilities into predictive and prescriptive maintenance. This paper proposes a process-based model to facilitate the selection of suitable DAS for supporting maintenance operations in manufacturing industries. This solution is employed for a structured requirement elicitation from various application domains and ultimately mapping the requirements to existing digital assistance solutions. Using the proposed approach, a (combination of) digital assistance system is selected and linked to maintenance activities. For this purpose, we gain benefit from an in-house process modeling tool utilized for identifying and relating sequence of maintenance activities. Finally, we collect feedback through employing the selected digital assistance system to improve the quality of recommendations and to identify the strengths and weaknesses of each system in association to practical usecases from TU Wien Pilot-Factory Industry 4.0.