NachhaltigF: Europan Campus Rottal-InnS: TC Freyung
Javier Valdés, Sebastian Wöllmann, Andreas Weber, G. Klaus, Christina Sigl, M. Prem, Robert Bauer, Roland Zink
A framework for regional smart energy planning using volunteered geographic information
Advances in Geosciences, vol. 54, no. 10 December 2020, pp. 179-193
This study presents a framework for regional smart energy planning for the optimal location and sizing of small hybrid systems. By using an optimization model – in combination with weather data – various local energy systems are simulated using the Calliope and PyPSA energy system simulation tools. The optimization and simulation models are fed with GIS data from different volunteered geographic information projects, including OpenStreetMap. These allow automatic allocation of specific demand profiles to diverse OpenStreetMap building categories. Moreover, based on the characteristics of the OpenStreetMap data, a set of possible distributed energy resources, including renewables and fossil-fueled generators, is defined for each building category. The optimization model can be applied for a set of scenarios based on different assumptions on electricity prices and technologies. Moreover, to assess the impact of the scenarios on the current distribution infrastructure, a simulation model of the low- and medium-voltage network is conducted. Finally, to facilitate their dissemination, the results of the simulation are stored in a PostgreSQL database, before they are delivered by a RESTful Laravel Server and displayed in an angular web application.
DigitalF: Maschinenbau und MechatronikI: Fraunhofer AWZ CTMT
F. Heilmeier, R. Koos, M. Singer, C. Bauer, Peter Hornberger, Jochen Hiller, W. Volk
Evaluation of Strain Transition Properties between Cast-In Fibre Bragg Gratings and Cast Aluminium during Uniaxial Straining
Sensors, vol. 20, no. 21
Current testing methods are capable of measuring strain near the surface on structural parts, for example by using strain gauges. However, stress peaks often occur within the material and can only be approximated. An alternative strain measurement incorporates fibre-optical strain sensors (Fiber Bragg Gratings, FBG) which are able to determine strains within the material. The principle has already been verified by using embedded FBGs in tensile specimens. The transition area between fibre and aluminium, however, is not yet properly investigated. Therefore, strains in tensile specimens containing FBGs were measured by neutron diffraction in gauge volumes of two different sizes around the Bragg grating. As a result, it is possible to identify and decouple elastic and plastic strains affecting the FBGs and to transfer the findings into a fully descriptive FE-model of the strain transition area.We thus accomplished closing the gap between the external load and internal straining obtained from cast-in FBG and generating valuable information about the mechanisms within the strain transition area.It was found that the porosity within the casting has a significant impact on the stiffness of the tensile specimen, the generation of excess microscopic tensions and thus the formation of permanent plastic strains, which are well recognized by the FBG. The knowledge that FBG as internal strain sensors function just as well as common external strain sensors will now allow for the application of FBG in actual structural parts and measurements under real load conditions. In the future, applications for long-term monitoring of cast parts will also be enabled and are currently under development.
F: Angewandte WirtschaftswissenschaftenF: Maschinenbau und Mechatronik
M. Bauer, O. Buchtala, T. Horeis, R. Kern, Bernhard Sick, et al.
Technical Data Mining with Evolutionary Radial Basis Function Classiers
Applied Soft Computing, vol. 9, no. 2, pp. 765-774
This article deals with two key problems of data mining, the automation of the data mining process and the integration of human domain experts. We show how an evolutionary algorithm (EA) can be used to optimize radial basis function (RBF) neural networks used for classification tasks. First, input features will be chosen from a set of possible input features (feature selection). Second, the number of hidden neurons is adapted (model selection). It is known that interpretable (fuzzy-type) rule sets may be extracted from RBF networks. We show how appropriate training algorithms for RBF networks and penalty terms in the fitness function of the EA may improve the understandability of the extracted rules. The properties of our approach are set out by means of two industrial application examples (process identification and quality control).