1 - 10 of 60
-
PublicationInformed Machine Learning - A Taxonomy and Survey of Integrating Prior Knowledge into Learning Systems( 2023)Despite its great success, machine learning can have its limits when dealing with insufficient training data. A potential solution is the additional integration of prior knowledge into the training process which leads to the notion of informed machine learning. In this paper, we present a structured overview of various approaches in this field. We provide a definition and propose a concept for informed machine learning which illustrates its building blocks and distinguishes it from conventional machine learning. We introduce a taxonomy that serves as a classification framework for informed machine learning approaches. It considers the source of knowledge, its representation, and its integration into the machine learning pipeline. Based on this taxonomy, we survey related research and describe how different knowledge representations such as algebraic equations, logic rules, or simulation results can be used in learning systems. This evaluation of numerous papers on the basis of our taxonomy uncovers key methods in the field of informed machine learning.
-
PublicationExperimental Evaluation of a Novel Sensor-Based Sorting Approach Featuring Predictive Real-Time Multiobject Tracking( 2021)Sensor-based sorting is a machine vision application that has found industrial application in various fields. An accept-or-reject task is executed by separating a material stream into two fractions. Current systems use line-scanning sensors, which is convenient as the material is perceived during transportation. However, line-scanning sensors yield a single observation of each object and no information about their movement. Due to a delay between localization and separation, assumptions regarding the location and point in time for separation need to be made based on the prior localization. Hence, it is necessary to ensure that all objects are transported at uniform velocities. This is often a complex and costly solution. In this paper, we propose a new method for reliably separating particles at non-uniform velocities. The problem is transferred from a mechanical to an algorithmic level. Our novel advanced image processing approach includes equipping the sorter with an area-scan camera in combination with a real-time multiobject tracking system, which enables predictions of the location of individual objects for separation. For the experimental validation of our approach, we present a modular sorting system, which allows comparing sorting results using a line-scan and area-scan camera. Results show that our approach performs reliable separation and hence increases sorting efficiency.
-
PublicationCleaning strategies and cost modelling of experimental membrane-based desalination plants( 2021)In Project WASTEC, an experimental Reverse Osmosis (RO) desalination system was developed. It serves as a platform for testing new technologies. For this system, we solved two problems, which are described in this paper. Firstly, we developed and investigated strategies for scheduling chemical enhanced backwashing and chemical cleaning and secondly, due to the experimental nature of the project, several new technological developments with respect to materials and methods were integrated into the system and requires tools for evaluating the economic viability of the new technologies. In this task, the economics of membrane-based desalination will be investigated. Baseline systems of reverse osmosis and pretreatment systems (microfiltration and ultrafiltration) will be economically examined and compared for their investments and operational costs. Sensitivity of the different plant and membrane parameters to the cost will be studied. Results show that with respect to costs, for a 200m3/hr design capacity plant, a volume of water is produced by a MF process at a cost of $0.494 and at a cost of $0.486 by an ultrafiltration process microfiltration. The reverse osmosis process cannot be compared directly, but it required $ 0.49 / m3 for a plant with 56 m3/hour design capacity. The values are in line with the costs reported in literature for membrane-based filtration.
-
PublicationTesting Automated Operation and Control Algorithms for Distribution Grids Using a Co-Simulation Environment( 2021)This publication presents a co-simulation framework that enables the coupling of various simulation components of energy systems that are not only modeled at three geographically distributed Fraunhofer Institutes, but are also vastly different in terms of their functionality, control algorithms, time resolution, speeds and used tools. This framework facilitates their joint usage despite these differences especially with regard to their time resolutions (real-time combined with non-real-time systems) and is applied in the DistributedGridLab of the Fraunhofer Cluster of Excellence Integrated Energy Systems (CINES) [1]. The DistributedGridLab thus allows users - e.g. grid operators, manufacturers and research institutes - to test their solutions before field deployment with special consideration of their interaction with other solutions at different remote testing facilities, and without the need to use the same software and hardware setups. A demonstrator of the DistributedGridLab is introduced where an electric energy system is modeled, which contains a medium voltage grid with three connected low voltage grids. The different control approaches interact to stabilize the entire system without a centrally controlled simulation of the systems.
-
PublicationImpact of industrial environments on visible light communication( 2021)Visible-light communication is a promising technology for industrial environments. However, a variety of physical effects may influence the communication quality in this potentially harsh environment: Dust and other particles lead to increased attenuation. Artificial light sources and industrial processes, such as grinding and welding, cause optical cross-talk. A multitude of reflective surfaces can lead to fading due to multi-path propagation. These three aspects are experimentally investigated in exemplary manufacturing processes at five different production sites in order to estimate the relative importance and their specific impact on VLC transmission in industrial areas. Spectral measurements demonstrate the presence of interfering light sources, which occupy broad parts of the visible spectrum. They give rise to flickering noise, which comprises a set of frequencies in the electrical domain. The impact of these effects on the communication is analysed with reference to the maximum achievable channel capacity and data rate approximation based on on-off keying is deduced. It is found that cross-talk by environmental and artificial light sources is one of the strongest effects, which influences the optical, but also the electrical spectrum. It is also observed that industrial areas differ strongly and must be categorised according to the manufacturing processes, which can induce quite a variation of dust and attenuation accordingly.
-
PublicationSmartSpectrometer - Embedded Optical Spectroscopy for Applications in Agriculture and Industry( 2021)The ongoing digitization of industry and agriculture can benefit significantly from optical spectroscopy. In many cases, optical spectroscopy enables the estimation of properties such as substance concentrations and compositions. Spectral data can be acquired and evaluated in real time, and the results can be integrated directly into process and automation units, saving resources and costs. Multivariate data analysis is needed to integrate optical spectrometers as sensors. Therefore, a spectrometer with integrated artificial intelligence (AI) called SmartSpectrometer and its interface is presented. The advantages of the SmartSpectrometer are exemplified by its integration into a harvesting vehicle, where quality is determined by predicting sugar and acid in grapes in the field.
-
PublicationFormation of spiraling infrared emission patterns by controlled interaction of optical filaments( 2021)We analyzed the formation of mid-infrared conical emission patterns possessing spiral and half-ring shaped wavelength contours from a beam of a few optical filaments. The complex patterns were generated and modified experimentally by adaptive wavefront shaping of the femtosecond laser pulse. Mutual interactions between co-propagating filaments can induce curvature in their paths, and the spiral and half-ring emissions were shown to be a direct consequence of this angular deflection. Based on our experimental and computational results, the spirals form in the far-field due to self-interference of conical emission from a helically moving filament. The presented findings will advance the tailoring of spatial conical emission patterns potentially beneficial for spectroscopic applications and terahertz generation.
-
PublicationMicrogrid Systems: Towards a Technical Performance Assessment Frame( 2021)
-
PublicationAlignment of safety and security risk assessments for modular production systems( 2021)In order to ensure the safety and security of industrial systems with regard to all life cycle phases from development through operation to disposal, specific regulatory and normative requirements are imposed. Due to the digitalization, interconnection, and constantly increasing complexity of manufacturing systems in the context of Industrie 4.0, the manual effort necessary to achieve the required safety and security is becoming ever greater and almost impossible to manage, especially for small and medium-sized enterprises. Therefore, this paper examines the existing challenges in this area in more detail and gives an outlook on the possible solutions to ensure safety and security much quicker and with less manual effort. The overall vision is a (partially) automated risk assessment of modular systems with respect to safety and security, including the alignment of the corresponding processes from both domains and the formalization of the information models needed.
-
PublicationGenerative Machine Learning for Resource-Aware 5G and IoT SystemsExtrapolations predict that the sheer number of Internet-of-Things (IoT) devices will exceed 40 billion in the next five years. Hand-crafting specialized energy models and monitoring sub-systems for each type of device is error prone, costly, and sometimes infeasible. In order to detect abnormal or faulty behavior as well as inefficient resource usage autonomously, it is of tremendous importance to endow upcoming IoT and 5G devices with sufficient intelligence to deduce an energy model from their own resource usage data. Such models can in-turn be applied to predict upcoming resource consumption and to detect system behavior that deviates from normal states. To this end, we investigate a special class of undirected probabilistic graphical model, the so-called integer Markov random fields (IntMRF). On the one hand, this model learns a full generative probability distribution over all possible states of the system-allowing us to predict system states and to measure the probability of observed states. On the other hand, IntMRFs are themselves designed to consume as less resources as possible-e.g., faithful modelling of systems with an exponentially large number of states, by using only 8-bit unsigned integer arithmetic and less than 16KB memory. We explain how IntMRFs can be applied to model the resource consumption and the system behavior of an IoT device and a 5G core network component, both under various workloads. Our results suggest, that the machine learning model can represent important characteristics of our two test systems and deliver reasonable predictions of the power consumption.
