Filters

2019 - 20204

More filters

20214
Reset filters

Settings

Type: Article × Subject: Artificial neural networks ×

Search Results

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Application of Artificial Neural Networks in Crystal Growth of Electronic and Opto-Electronic Materials
    (Basel : MDPI, 2020) Dropka, Natasha; Holena, Martin
    In this review, we summarize the results concerning the application of artificial neural networks (ANNs) in the crystal growth of electronic and opto-electronic materials. The main reason for using ANNs is to detect the patterns and relationships in non-linear static and dynamic data sets which are common in crystal growth processes, all in a real time. The fast forecasting is particularly important for the process control, since common numerical simulations are slow and in situ measurements of key process parameters are not feasible. This important machine learning approach thus makes it possible to determine optimized parameters for high-quality up-scaled crystals in real time. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.
  • Thumbnail Image
    Item
    Deep neural networks for classifying complex features in diffraction images
    (Woodbury, NY : Inst., 2019) Zimmermann, Julian; Langbehn, Bruno; Cucini, Riccardo; Di Fraia, Michele; Finetti, Paola; LaForge, Aaron C.; Nishiyama, Toshiyuki; Ovcharenko, Yevheniy; Piseri, Paolo; Plekan, Oksana; Prince, Kevin C.; Stienkemeier, Frank; Ueda, Kiyoshi; Callegari, Carlo; Möller, Thomas; Rupp, Daniela
    Intense short-wavelength pulses from free-electron lasers and high-harmonic-generation sources enable diffractive imaging of individual nanosized objects with a single x-ray laser shot. The enormous data sets with up to several million diffraction patterns present a severe problem for data analysis because of the high dimensionality of imaging data. Feature recognition and selection is a crucial step to reduce the dimensionality. Usually, custom-made algorithms are developed at a considerable effort to approximate the particular features connected to an individual specimen, but because they face different experimental conditions, these approaches do not generalize well. On the other hand, deep neural networks are the principal instrument for today's revolution in automated image recognition, a development that has not been adapted to its full potential for data analysis in science. We recently published [Langbehn et al., Phys. Rev. Lett. 121, 255301 (2018)] the application of a deep neural network as a feature extractor for wide-angle diffraction images of helium nanodroplets. Here we present the setup, our modifications, and the training process of the deep neural network for diffraction image classification and its systematic bench marking. We find that deep neural networks significantly outperform previous attempts for sorting and classifying complex diffraction patterns and are a significant improvement for the much-needed assistance during postprocessing of large amounts of experimental coherent diffraction imaging data.
  • Thumbnail Image
    Item
    High-resolution net load forecasting for micro-neighbourhoods with high penetration of renewable energy sources
    (Amsterdam [u.a.] : Elsevier Science, 2020) Kobylinski, P.; Wierzbowski, M.; Piotrowski, K.
    Though extensive, the literature on electrical load forecasting lacks reports on studies focused on existing residential micro-neighbourhoods comprising small numbers of single-family houses equipped with solar panels. This paper provides a full description of an ANN-based model designed to predict short-term high-resolution (15-min intervals) micro-scale residential net load profiles. Since it seems especially relevant due to the specificity of local autocorrelations in load signal, in this paper we put stress on the systematic approach to feature selection in the context of lagged signal. We performed a case study of a real micro-neighbourhood comprising only 75 single-family houses. The obtained average prediction error was equivalent to 5.4 per cent of the maximal measured net load. The issues, i.e.: (1) the feasibility of micro-scale residential load forecasting taking into account renewable energy penetration, (2) the feasibility to predict net load with dense temporal resolution of 15 min, (3) the feature selection problem, (4) the proposed prosumption- and comparison-oriented prediction model key performance measure, could be of interest to engineers designing energy balancing systems for local smart grids. © 2019 The Authors
  • Thumbnail Image
    Item
    Prediction of the biogas production using GA and ACO input features selection method for ANN model
    (Amsterdam [u.a.] : Elsevier, 2019) Beltramo, Tanja; Klocke, Michael; Hitzmann, Bernd
    This paper presents a fast and reliable approach to analyze the biogas production process with respect to the biogas production rate. The experimental data used for the developed models included 15 process variables measured at an agricultural biogas plant in Germany. In this context, the concentration of volatile fatty acids, total solids, volatile solids acid detergent fibre, acid detergent lignin, neutral detergent fibre, ammonium nitrogen, hydraulic retention time, and organic loading rate were used. Artificial neural networks (ANN) were established to predict the biogas production rate. An ant colony optimization and genetic algorithms were implemented to perform the variable selection. They identified the significant process variables, reduced the model dimension and improved the prediction capacity of the ANN models. The best prediction of the biogas production rate was obtained with an error of prediction of 6.24% and a coefficient of determination of R2 = 0.9.