Publications Repository - Gdańsk University of Technology

Page settings

polski
Publications Repository
Gdańsk University of Technology

Show publications from the year

Publications from the year 2019

Show all
  • Damage Detection in the Wind Turbine Blade Using Root Mean Square and Experimental Modal Parameters
    • Łukasz Doliński
    • Marek Krawczuk
    • Arkadiusz Żak
    2019

    The paper presents results of an experimental study related to a non-destructive diagnostic technique used for preliminary determination the location and size of delamination in composite coatings of wind turbine blades. The proposed method of damage detection is based on the analysis of the ten first mode shapes of bending vibrations, which correspond to displacements of rotor blades perpendicular to the rotor plane. Modal parameters depend on the physical properties of the structure. On the other hand, failures can affect these properties (e.g. locally reduce the stiffness of the structure). Monitoring of selected modal parameter can allow determination the technical condition of the structure. The main assumption of the presented method is a comprehensive analysis of the measured data by determination the root mean square value (RMS) for each measurement point from all forms of free vibration obtained from the experiment. As a result, information contained in all modes of vibrations that may indicate damage of the blade will be included in a single characteristic. The investigations were carried out on a scaled-down model of a wind turbine blade of a rotor diameter of 36 m. The modal parameters have been determined only experimentally using a Laser Doppler Scanning Vibrometer. Damage was simulated for three localizations by additional high stiffness elements fixed to the surface of the blade. The results of the research presented in this paper confirm the effectiveness of RMS calculation in detection damage using modes of vibrations.

  • DANE I DOKUMENTACJA MEDYCZNA
    • Jacek Rumiński
    2019

    Rozdział analizuje aspekty danych i dokumentacji medycznych z odniesieniu do międzynarodowych norm technicznych takich jak DICOM, HL7. Wskazuje również definicje dokumentacji medycznej związane z polskim prawem. Porusza również aspekty związane z kolekcjami danych (np. biobanki) jak i bezpieczeństwem danych.

  • Data librarian and data steward – new tasks and responsibilities of academic libraries in the context of Open Research Data implementation in Poland
    • Anna Wałek
    2019 Full text Przegląd Biblioteczny

    Thesis/Objective – The policy of Open Access (OA) for researching resources in Europe has been implemented for more than 10 years. The first recommendations concerning providing OA to scientific materials were defined during the implementation of the 7th Framework Programme. Introducing another set of recommendations concerning OA to research data was the next stage. The recommendations were transformed into obligations under the Horizon 2020 Programme. In 2018, research-funding institutions were associated in the Plan S document issued by CoalitionS ,which aims to ac celerate the transition to full and immediate OA to publications from publicly funded research until January 2021. Academic libraries have always been pioneers in implementing OA to research, creating the necessary tools (platforms and repositories), and preparing training workshops for researchers. OA policy implementation, including both access to research resources and data, is accelerating. That is why the role of academic libraries and academic librarians has become crucial. The article presents how library services and the scope of tasks of their employees change in connection with the introduction of open access policies for research data in Poland. Research methods – A critical review of the literature was used to analyse the content of foreign and Polish LIS literature published in the years 2009-2019. In addition, official documents issued by the European Commission were analysed, as well as websites devoted to Open Research Data (ORD). Results and conclusions – Some new specialisations in librarianship have been introduced – e.g. a data librarian who is responsible not only for academic staff training sessions on Open Research Data, but also for assistance for research teams in the field of data management and data curation. In the future, academic libraries will be responsible for coordinating the work of data stewards responsible for supporting the process of research data creating and managing at university departments and in research teams.

  • Database of speech and facial expressions recorded with optimized face motion capture settings
    • Andrzej Czyżewski
    • Miłosz Kawaler
    2019 Full text JOURNAL OF INTELLIGENT INFORMATION SYSTEMS

    The broad objective of the present research is the analysis of spoken English employing a multiplicity of modalities. An important stage of this process, discussed in the paper, is creating a database of speech accompanied with facial expressions. Recordings of speakers were made using an advanced system for capturing facial muscle motion. A brief historical outline, current applications, limitations and the ways of capturing face muscle motion as well as the problems with recording facial expressions are discussed. In particular, the scope of the present analysis concerns the registration of facial expressions related to emotions of speakers which accompany articulation. The camera system, instrumentation and software used for registration and for post-production are outlined. An analysis of the registration procedure and the results of the registration process was performed. The obtained results demonstrate how muscle movements can be registered employing reflective markers and point at the advantages and limitations of applying FMC (Face Motion Capture) technology in compiling a multimodal speech database. A short discussion pertaining to the usage of FMC as ground truth data source in facial expression databases concludes the paper.

  • Daylight evaluation for multi-family housing in Poland
    • Natalia Sokół
    2019 Full text

    This PhD dissertation focuses on methods of daylight appraisal useful in the design of the contemporary multifamily housing. The theoretical part of the thesis offers a review of daylight indicators, evaluations methods and tools within the built environment. It covers a review of daylight recommendations found in building standards and other normative documents affecting the design of the residential spaces. A pilot work survey carried out among 140 architecture students aimed to verify students', perception, preferences on daylight as well as their knowledge about contemporary daylight metrics and assessment methods and regulations. The empirical part of the thesis presents the experiment focused on the appraisal of daylight conditions within 20 rooms spaces via on-site repeated luminance measurements, the questionnaire focusing on inhabitants' perception, preferences and satisfaction with the daylight conditions within their dwellings, and luminance, illuminance simulation. The obtained results are presented in numerical form and graphical notation, in the form of multi-coloured graphs - images. In the summary, the potential and level of difficulty of the presented daylight appraisal methods were assessed. The influence of normative documents on the design solutions affecting propagation of light in the multi-dwelling structures was described. The results confirm that the analysis and assessment of daylight is crucial at all stages of planning residential architecture. A well-thought-out design of daylight affects the perception of residential interiors by their users and impacts their quality of life.

  • DAYLIGHT WITHIN A ROOM IN THE EYES OF ARCHITECTURE STUDENTS
    • Natalia Sokół
    • Giuliani Federica
    • Justyna Martyniuk-Pęczek
    2019

    A questionnaire was conducted to investigate how 140 architecture students apprise daylight conditions within the classrooms. The participants were requested to evaluate the luminous environment and their luminous comfort. They were also asked about light preferences and knowledge on daylight metrics and regulations. The students’ subjective appraisals results were compared with the experts’ assessment and the on-site illuminance measurements. Later on, the students had to carry out a series of daylight indices simulations summarizing daylight conditions in their private rooms (a user’s experience), and in a room, within a building, they had been designing (a designer’s experience). The perception of the luminous environment was analysed against participants’ comprehension of daylight simulations data. The issues students encountered during the daylight indicators analysis suggested that more coherent daylight education should be offered for future architects.

  • Debonding Detection in Reinforced Concrete Beams with the Use of Guided Wave Propagation
    • Beata Zima
    2019

    One of the most frequent damage of the reinforced concrete structures is debonding between steel bar and concrete cover. In the case of debonding occurrence not only the strength of the structure decreases, but also it is more vulnerable to corrosion damages. For this reason fast and effective methods of debonding detection in an early stage of its development need a significant boost. The paper presents analytical and experimental investigation of debonding detection in reinforced concrete beams using non-destructive method based on guided waves propagation. Concrete beams of rectangular cross-section consisting of four steel rebars with pre-existing debonding are investigated. Lack of adhesive connection between one rod and concrete cover is provided by introducing cellophane film of a small thickness (90 μm). The research is focused on detection of debonding with variable length on the basis of time-domain signals captured by piezoelectric sensors. Presented method of damage detection takes advantage of the time of flight of the reflections registered at the ends of the specimens.

  • Decision making process using deep learning
    • Olgun Aydin
    2019

    Endüstri 4.0, dördüncü endüstri devrimi veya Endüstriyel Nesnelerin İnterneti (IIoT) olarak adlandırılan sanayi akımı, işletmelere, daha verimli, daha büyük bir esneklikle, daha güvenli ve daha çevre dostu bir şekilde üretim yapma imkanı sunmaktadır. Nesnelerin İnterneti ile bağlantılı yeni teknoloji ve hizmetler birçok endüstriyel uygulamada devrim niteliği taşımaktadır. Fabrikalardaki otomasyon, tahminleyici bakım (PdM – Predictive Maintenance) modellerigibi gelişmeler işletmelere, iş modellerini değiştiren yenilikçi çözümler sunmaktadır. Günümüz endüstriyel uygulamalarında ihtiyaç duyulan yüksek otomasyon seviyesini sağlamak için, daha verimli ekipman kullanılmalı, daha akıllı sistemler oluşturulmalıdır. Bu sayede, hem üretimdeki verim artacak hem de daha güvenli bir çalışma ortamı da sağlanmış olacaktır. İşletmeler, üretimlerine minimum aksama süresiyle, optimum hızda devam etmek istemektedir. Bu sebeple, üretimde kullanılan, hareketli parçaları olan makinelerin daha verimli kullanılmasını sağlamak için yapılacak bakım planlamaları kritik önem taşımaktadır. Bu konu ile ilgili yaklaşımlardan bir tanesi, ekipmanların durumuna bakılmaksızın bakım süreçlerini sabit aralıklarla gerçekleştirmektir. Bu yöntem planlanması basit bir yöntemdir ancak, zaman zaman ekipmanların arızası gerçekleştikten sonra bakım işleminin gerçekleştirilmesine ya da ekipmanlarda hiç bir problem yokken bakım işleminin gerçekleştirilmesine neden olabilmektedir. Bu da sistemde uzun süreli aksamalar, gereksiz bakım maliyetleri gibi sonuçları doğurmaktadır. Bakım süreçlerine farklı bir yaklaşım olan PdM, makinenin gözlemlenen durumuna bağlı olarak bakım süreçlerinin yönetilmesine olanak kılmaktadır.PdM yaklaşımına yeni bir akış açısı getirmek amacıyla yapılan bu çalışmada yeni bir derin yapay sinir ağı mimarisi önerilmiştir. Bu mimaride bir girdi katmanı bir LSTM katmanı, bırakma (DO) ve ardından yine bir LSTM katmanı, bir gizli katman ve çıktı katmanı bulunmaktadır. Mimaride kullanılan iterasyon sayısı, parti büyüklüğü Genetik Algoritma (GA) kullanılarak, kayıp fonksiyonunu optimize eden optimizasyon algoritması, çıktı katmanında sonra kullanılan aktivasyon fonksiyonu ve DO oranı ızgara araması (GS) kullanılarak belirlenmiştir.

  • Decisional DNA based intelligent knowledge model for flexible manufacturing system
    • Syed Imran Shafiq
    • Edward Szczerbicki
    • Bogdan Trawiński
    • Cesar Sanin
    2019 JOURNAL OF INTELLIGENT & FUZZY SYSTEMS

    Modeling an effective mechanism for design and control strategies for the implementation of a flexible manufacturing system (FMS) has been a challenge. Consequently, to overcome this issue various techniques have applied in the past but most of these models are effective only for some specific situation or an element of FMS. In this study, the knowledge representation technique of Decisional DNA (DDNA) is applied to FMS to develop a generic model to achieve effective scheduling and manufacturing flexibility. Decisional DNA based Virtual Engineering Objects (VEO) are used as communicating media between machines, equipment and works pieces. The concept of Virtual Engineering Process (VEP) is applied for modeling routing flexibility. VEOs combined with VEPs form FMS-DDNA model, which facilitates in enhancing the performance of FMS, by inducing intelligence based on its own previous experience thus making it practical and smart.

  • Decisional-DNA Based Smart Production Performance Analysis Model
    • Syed Imran
    • Edward Szczerbicki
    • Cesar Sanin
    2019 Full text CYBERNETICS AND SYSTEMS

    n order to allocate resources effectively according to the production plan and to reduce disturbances, a framework for smart production performance analysis is proposed in this article. Decisional DNA based knowledge models of engineering objects, processes and factory are developed within the proposed framework. These models are the virtual representation of manufacturing resources, and with help of Internet of Things, are capable of capturing the past experience and formal decisions. A case study for the smart tool performance analysis is presented in which information of key tool parameters like tool life, surface integrity, tool forces and chip formation can be sensed in real-time, and predictions can be made according to specific requirements. This framework is capable of creating a cyber-physical conjoining of the bottom-level manufacturing resources and thus can work as a technological basis for smart factories and Industry 4.0.

  • Decontaminating Arbitrary Graphs by Mobile Agents: a Survey
    • Dorota Osula
    2019 UTILITAS MATHEMATICA

    A team of mobile agents starting from homebases need to visit and clean all nodes of the network. The goal is to find a strategy, which would be optimal in the sense of the number of needed entities, the number of moves performed by them or the completion time of the strategy. Currently, the field of distributed graph searching by a team of mobile agents is rapidly expanding and many new approaches and models are being presented in order to better describe real life problems like decontaminating danger areas by a group of robots or cleaning networks from viruses. A centralized searching, when a topology of a graph is known in advance is well studied. This survey presents comprehensive results focusing mainly on an issue of the distributed monotone contiguous decontamination problem, including recent results for clearing graphs with and without a priori knowledge about its topology. We introduce a bibliography for various models, which differ on e.g., knowledge about a graph, properties of agents, time clock or size of the available memory.

  • Decyzje przedsiębiorstwa na rynku finansowym
    • Gabriela Golawska-Witkowska
    • Ewa Mazurek-Krasodomska
    • Anna Rzeczycka
    2019

    Prowadzenie działalności gospodarczej, w tym realizowanie zamierzeń rozwojowych oraz bieżących, wiąże się z ryzykiem niezrealizowania wytyczonego celu. Oznacza to, że firma działa w warunkach ryzyka, a niekiedy niepewności. Główne ryzyka dotyczą utrudnień w pozyskaniu niezbędnych środków finansowych, wahań kursów walut, zmian stóp procentowych itp. Przedsiębiorstwo musi więc znaleźć miejsce, w którym firma pozyska niezbędne kapitały – własne i obce – oraz ograniczy generowane ryzyko. Może to zrobić na rynku finansowym, na określonych jego segmentach, w zależności od potrzeb. Niniejsza publikacja została poświęcona problematyce decyzji przedsiębiorstwa związanych z realizacją wytyczonych celów w kontekście pozyskiwania źródeł finansowania oraz minimalizowania ryzyka. Główne założenia pracy obejmują wskazanie zasad podejmowania przez przedsiębiorstwo decyzji na rynku finansowym oraz określenie miejsca instrumentów tego rynku w bilansach polskich przedsiębiorstw. Monografia przeznaczona jest dla studentów kierunków ekonomicznych, w szczególności studiujących zarządzanie finansami przedsiębiorstwa i pragnących poszerzyć wiedzę dotyczącą decyzji podejmowanych przez firmy na rynku finansowym.

  • DEDUKCJA ZACHOWAŃ WĘZŁÓW TRANZYTOWYCH W WIELOSKOKOWEJ SIECI BEZPRZEWODOWEJ W OBECNOŚCI ZAKŁÓCEŃ
    • Jerzy Konorski
    • Karol Rydzewski
    2019 Full text Przegląd Telekomunikacyjny + Wiadomości Telekomunikacyjne

    Przedstawiono nowy algorytm dedukcji zachowań (metryki reputacji) węzłów tranzytowych w wieloskokowej sieci bezprzewodowej na podstawie potwierdzeń końcowych. Algorytm stosuje znane metody matematyczne i jest odporny na zakłócenia naturalnie występujące w sieciach bezprzewodowych oraz intencjonalne zmiany zachowania węzłów. Informacja zwracana przez algorytm, poza wydedukowanym zachowaniem węzłów, zawiera dane o możliwym błędzie dedukcji.

  • Deep eutectic solvents based highly efficient extractive desulfurization of fuels – Eco-friendly approach
    • Patrycja Makoś
    • Grzegorz Boczkaj
    2019 Full text JOURNAL OF MOLECULAR LIQUIDS

    The developed process is based on alternative, green and cheap solvents for efficient desulfurization of fuels. Several deep eutectic solvents (DESs) were successfully synthesized and studied as extraction solvents for desulfurization of model fuel containing thiophene (T), benzothiophene (BT) and dibenzothiophene (DBT). The most important extraction parameters (i.e. kind of DES, DES: fuel volume ratio, hydrogen bond acceptor: hydrogen bond donor mole ratio, time of extraction and temperature) were optimized using central composite design model. Furthermore, the mutual solubility of DES and model fuel and influence of multistage extraction, reusability, regeneration of DES and content of aromatic groups in fuel are discussed followed by explanation of desulfurization mechanism, by means of density functional theory (DFT) as well as FT-IR analysis. The studies revealed high desulfurization effectiveness resulting in 91.5%, 95.4% and 99.2% removal of T, BT and DBT respectively in a single stage extraction. A three stage desulfurization provide >99.99% removal of T, BT and DBT. The research on the desulfurization mechanism revealed that π-π interaction is the main driving force for desulfurization process based on DES.

  • Deep learning in the fog
    • Andrzej Sobecki
    • Julian Szymański
    • David Gil
    • Higinio Mora
    2019 Full text International Journal of Distributed Sensor Networks

    In the era of a ubiquitous Internet of Things and fast artificial intelligence advance, especially thanks to deep learning networks and hardware acceleration, we face rapid growth of highly decentralized and intelligent solutions that offer functionality of data processing closer to the end user. Internet of Things usually produces a huge amount of data that to be effectively analyzed, especially with neural networks, demands high computing capabilities. Processing all the data in the cloud may not be sufficient in cases when we need privacy and low latency, and when we have limited Internet bandwidth, or it is simply too expensive. It poses a challenge for creating a new generation of fog computing that supports artificial intelligence and selects the architecture appropriate for an intelligent solution. In this article, we show from four perspectives, namely, hardware, software libraries, platforms, and current applications, the landscape of components used for developing intelligent Internet of Things solutions located near where the data are generated. This way, we pinpoint the odds and risks of artificial intelligence fog computing and help in the process of selecting suitable architecture and components that will satisfy all requirements defined by the complex Internet of Things systems.

  • Deep Learning Optimization for Edge Devices: Analysis of Training Quantization Parameters
    • Alicja Kwaśniewska
    • Maciej Szankin
    • Mateusz Ozga
    • Jason Wolfe
    • Arun Das
    • Adam Zajac
    • Jacek Rumiński
    • Paul Rad
    2019 Full text

    This paper focuses on convolution neural network quantization problem. The quantization has a distinct stage of data conversion from floating-point into integer-point numbers. In general, the process of quantization is associated with the reduction of the matrix dimension via limited precision of the numbers. However, the training and inference stages of deep learning neural network are limited by the space of the memory and a variety of factors including programming complexity and even reliability of the system. On the whole the process of quantization becomes more and more popular due to significant impact on performance and minimal accuracy loss. Various techniques for networks quantization have been already proposed, including quantization aware training and integer arithmetic-only inference. Yet, a detailed comparison of various quantization configurations, combining all proposed methods haven’t been presented yet. This comparison is important to understand selection of quantization hyperparameters during training to optimize networks for inference while preserving their robustness. In this work, we perform in-depth analysis of parameters in the quantization aware training, the process of simulating precision loss in the forward pass by quantizing and dequantizing tensors. Specifically, we modify rounding modes, input preprocessing, output data signedness, bitwidth of the quantization and locations of precision loss simulation to evaluate how they affect accuracy of deep neural network aimed at performing efficient calculations on resource-constrained devices.

  • Deep Learning-Based LOS and NLOS Identification in Wireless Body Area Networks
    • Krzysztof Cwalina
    • Piotr Rajchowski
    • Olga Błaszkiewicz
    • Alicja Olejniczak
    • Jarosław Sadowski
    2019 Full text SENSORS

    In this article, the usage of deep learning (DL) in ultra-wideband (UWB) Wireless Body Area Networks (WBANs) is presented. The developed approach, using channel impulse response, allows higher efficiency in identifying the direct visibility conditions between nodes in off-body communication with comparison to the methods described in the literature. The effectiveness of the proposed deep feedforward neural network was checked on the basis of the measurement data for dynamic scenarios in an indoor environment. The obtained results clearly prove the validity of the proposed DL approach in the UWB WBANs and high (over 98.6% for most cases) efficiency for LOS and NLOS conditions classification.

  • Deep neural network architecture search using network morphism
    • Arkadiusz Kwasigroch
    • Michał Grochowski
    • Mateusz Mikołajczyk
    2019

    The paper presents the results of the research on neural architecture search (NAS) algorithm. We utilized the hill climbing algorithm to search for well-performing structures of deep convolutional neural network. Moreover, we used the function preserving transformations which enabled the effective operation of the algorithm in a short period of time. The network obtained with the advantage of NAS was validated on skin lesion classification problem. We compared the parameters and performance of the automatically generated neural structure with the architectures selected manually, reported by the authors in previous papers. The obtained structure achieved comparable results to hand-designed networks, but with much fewer parameters then manually crafted architectures.

  • Deep neural networks for human pose estimation from a very low resolution depth image
    • Piotr Szczuko
    2019 Full text MULTIMEDIA TOOLS AND APPLICATIONS

    The work presented in the paper is dedicated to determining and evaluating the most efficient neural network architecture applied as a multiple regression network localizing human body joints in 3D space based on a single low resolution depth image. The main challenge was to deal with a noisy and coarse representation of the human body, as observed by a depth sensor from a large distance, and to achieve high localization precision. The regression network was expected to reason about relations of body parts based on depth image, and to extract locations of joints, and provide coordinates defining the body pose. The method involved creation of a dataset with 200,000 realistic depth images of a 3D body model, then training and testing numerous architectures including feedforward multilayer perceptron network and deep convolutional neural networks. The results of training and evaluation are included and discussed. The most accurate DNN network was further trained and evaluated on an augmented depth images dataset. The achieved accuracy was similar to a reference Kinect algorithm results, with a great benefit of fast processing speed and significantly lower requirements on sensor resolution, as it used 100 times less pixels than Kinect depth sensor. The method was robust against sensor noise, allowing imprecision of depth measurements. Finally, our results were compared with VGG, MobileNet, and ResNet architectures.

  • Degradable poly(ester-ether) urethanes of improved surface calcium deposition developed as novel biomaterials
    • Justyna Kucińska-Lipka
    • Alicja Lewandowska
    • Paweł Szarlej
    • Marcin Stanisław Łapiński
    • Iga Gubańska
    2019 JOURNAL OF BIOACTIVE AND COMPATIBLE POLYMERS

    Bones, which are considered as hard tissues, work as scaffold for human body. They provide physical support for muscles and protect intestinal organs. Percentage of hard tissues in human body depends on age, weight, and gender. Human skeleton consists of 206 connected bones. Therefore, it is natural that the hard-tissue damage such as fractures, osteoporosis, and congenital lack of bone may appear. The innovative way of bone healing is an application of so-called tissue scaffolds. There are many synthetic polymers used in this field, but polyurethanes play a great role in this field. It is due to the possibility to control their degradation rate and to tune their surface to improve the calcification process, required for proper bone regeneration. In this article, we described the fabrication of degradable poly(ester-ether)urethane materials, having different hard-segment content (28% or 47%). PEEURs-28HS and PEEURs-47HS materials were obtained by two-step polymerization method and characterized by mechanical properties, ability to undergo oxidative degradation and surface calcification. Performed studies indicated that the PEEURs-28HS material possessed suitable properties to be proposed as a material for possible application in the bone tissue engineering.