Publikacje
Filtry
wszystkich: 334
Katalog Publikacji
-
Data augmentation for improving deep learning in image classification problem
PublikacjaThese days deep learning is the fastest-growing field in the field of Machine Learning (ML) and Deep Neural Networks (DNN). Among many of DNN structures, the Convolutional Neural Networks (CNN) are currently the main tool used for the image analysis and classification purposes. Although great achievements and perspectives, deep neural networks and accompanying learning algorithms have some relevant challenges to tackle. In this...
-
Hierarchical predictive control of integrated wastewater treatment systems
PublikacjaThe paper proposes an approach to designing the control structure and algorithms for optimising control of integrated wastewater treatment plant-sewer systems (IWWTS) under a full range of disturbance inputs. The optimised control of IWWTS allows for significant cost savings, fulfilling the effluent discharge limits over a long period and maintaining the system in sustainable operation. Due to the specific features of a wastewater...
-
Multi-objective optimization of water quality, pumps operation, and storage sizing of water distribution systems
PublikacjaA multi-objective methodology utilizing the Strength Pareto Evolutionary Algorithm (SPEA2) linked to EPANET for trading-off pumping costs, water quality, and tanks sizing of water distribution systems is developed and demonstrated. The model integrates variable speed pumps for modeling the pumps operation, two water quality objectives (one based on chlorine disinfectant concentrations and one on water age), and tanks sizing cost...
-
Hierarchical dissolved oxygen control for activated sludge processes
PublikacjaA hierarchical controller for tracking the dissolved oxygen reference trajectory in activated sludge processes is proposed and investigated. The removal of nitrogen and phosphorous from wastewater is considered. Typically, an aeration system itself is a complicated hybrid nonlinear dynamical system with faster dynamics compared to the internal dynamics of the dissolved oxygen in a biological reactor. It is a common approach to...
-
Deep CNN based decision support system for detection and assessing the stage of diabetic retinopathy
PublikacjaThe diabetic retinopathy is a disease caused by long-standing diabetes. Lack of effective treatment can lead to vision impairment and even irreversible blindness. The disease can be diagnosed by examining digital color fundus photographs of retina. In this paper we propose deep learning approach to automated diabetic retinopathy screening. Deep convolutional neural networks (CNN) - the most popular kind of deep learning algorithms...
-
Cooperative control in production and logistics
PublikacjaClassical applications of control engineering and information and communication technology (ICT) in production and logistics are often done in a rigid, centralized and hierarchical way. These inflexible approaches are typically not able to cope with the complexities of the manufacturing environment, such as the instabilities, uncertainties and abrupt changes caused by internal and external disturbances, or a large number and variety...
-
Deep neural networks approach to skin lesions classification — A comparative analysis
PublikacjaThe paper presents the results of research on the use of Deep Neural Networks (DNN) for automatic classification of the skin lesions. The authors have focused on the most effective kind of DNNs for image processing, namely Convolutional Neural Networks (CNN). In particular, three kinds of CNN were analyzed: VGG19, Residual Networks (ResNet) and the hybrid of VGG19 CNN with the Support Vector Machine (SVM). The research was carried...
-
Grid Implementation of a Parallel Multiobjective Genetic Algorithm for Optimized Allocation of Chlorination Stations in Drinking Water Distribution Systems: Chojnice Case Study
PublikacjaSolving multiobjective optimization problems requires suitable algorithms to find a satisfactory approximation of a globally optimal Pareto front. Furthermore, it is a computationally demanding task. In this paper, the grid implementation of a distributed multiobjective genetic algorithm is presented. The distributed version of the algorithm is based on the island algorithm with forgetting island elitism used instead of a genetic...
-
Neural Architecture Search for Skin Lesion Classification
PublikacjaDeep neural networks have achieved great success in many domains. However, successful deployment of such systems is determined by proper manual selection of the neural architecture. This is a tedious and time-consuming process that requires expert knowledge. Different tasks need very different architectures to obtain satisfactory results. The group of methods called the neural architecture search (NAS) helps to find effective architecture...
-
Multiobjective Water Distribution Systems Control of Pumping Cost, Water Quality, and Storage-Reliability Constraints
PublikacjaThis work describes a multiobjective model for trading-off pumping cost and water quality for water distribution systems operation. Constraints are imposed on flows and pressures, on periodical tanks operation, and on tanks storage. The methodology links the multiobjective SPEA2 algorithm with EPANET, and is applied on two example applications of increasing complexity, under extended period simulation conditions and variable energy...
-
Genetic Hybrid Predictive Controller for Optimized Dissolved-Oxygen Tracking at Lower Control Level
PublikacjaA hierarchical two-level controller for dissolvedoxygenreference trajectory tracking in activated sludge processeshas been recently developed and successfully validated on a realwastewater treatment plant. The upper level control unit generatestrajectories of the desired airflows to be delivered by theaeration system to the aerobic zones of the biological reactor. Anonlinear model predictive control algorithm is applied to designthis...
-
Fractional neutron point kinetics equations for nuclear reactor dynamics – Numerical solution investigations
PublikacjaThis paper presents results concerning numerical solutions to a fractional neutron point kinetics model for a nuclear reactor. The paper discusses and expands on results presented in (Espinosa-Paredes et al., 2011). The fractional neutron point kinetics model with six groups of delayed neutron precursors was developed and a numerical solution using the Edwards’ method was proposed (Edwards et al., 2002). The mathematical model...
-
Numerical solution analysis of fractional point kinetics and heat exchange in nuclear reactor
PublikacjaThe paper presents the neutron point kinetics and heat exchange models for the nuclear reactor. The models consist of a nonlinear system of fractional ordinary differential and algebraic equations. Two numerical algorithms are used to solve them. The first algorithm is application of discrete Grünwald-Letnikov definition of the fractional derivative in the model. The second involves building an analog scheme in the FOMCON Toolbox...
-
An Automatic Self-Tuning Control System Design for an Inverted Pendulum
PublikacjaA control problem of an inverted pendulum in the presence of parametric uncertainty has been investigated in this paper. In particular, synthesis and implementation of an automatic self-tuning regulator for a real inverted pendulum have been given. The main cores of the control system are a swing-up control method and a stabilisation regulator. The first one is based on the energy of an inverted pendulum, whereas the second one...
-
Nodal models of Pressurized Water Reactor core for control purposes – A comparison study
PublikacjaThe paper focuses on the presentation and comparison of basic nodal and expanded multi-nodal models of the Pressurized Water Reactor (PWR) core, which includes neutron kinetics, heat transfer between fuel and coolant, and internal and external reactivity feedback processes. In the expanded multi-nodal model, the authors introduce a novel approach to the implementation of thermal power distribution phenomena into the multi-node...
-
Fuzzy Multi-Regional Fractional PID controller for Pressurized Water nuclear Reactor
PublikacjaThe paper presents the methodology for the synthesis of a Fuzzy Multi-Regional Fractional Order PID controller (FMR-FOPID) used to control the average thermal power of a PWR nuclear reactor in the load following mode. The controller utilizes a set of FOPID controllers and the fuzzy logic Takagi-Sugeno reasoning system. The proposed methodology is based on two optimization parts. The first part is devoted to finding the optimal...
-
Accurate Computation of IGBT Junction Temperature in PLECS
PublikacjaIn the article, a new method to improve the accuracy of the insulated-gate bipolar transistor (IGBT) junction temperature computations in the piecewise linear electrical circuit simulation (PLECS) software is proposed and described in detail. This method allows computing the IGBT junction temperature using a nonlinear compact thermal model of this device in PLECS. In the method, a nonlinear compact thermal model of the IGBT is...
-
Review and comparison of smoothing algorithms for one-dimensional data noise reduction
PublikacjaThe paper considers the choice of parameters of smoothing algorithms for data denoising. The impact of the window size on smoothing accuracy was analyzed. The parameters of denoising filters were selected with respect to the meansquare error between the computed linear regression and the noisy signal. Finally, we have compared mean, median, SavitzkyGolay, Kalman and Gaussian filter algorithms for the data from the digital sensor....
-
Two-phase optimizing approach to design assessments of long distance heat transportation for CHP systems
PublikacjaCogeneration or Combined Heat and Power (CHP) for power plants is a method of putting to use waste heat which would be otherwise released to the environment. This allows the increase in thermodynamic efficiency of the plant and can be a source of environmental friendly heat for District Heating (DH). In the paper CHP for Nuclear Power Plant (NPP) is analyzed with the focus on heat transportation. A method for effectivity and feasibility...
-
Self-Supervised Learning to Increase the Performance of Skin Lesion Classification
PublikacjaTo successfully train a deep neural network, a large amount of human-labeled data is required. Unfortunately, in many areas, collecting and labeling data is a difficult and tedious task. Several ways have been developed to mitigate the problem associated with the shortage of data, the most common of which is transfer learning. However, in many cases, the use of transfer learning as the only remedy is insufficient. In this study,...