Browse

You are looking at 1 - 10 of 604 items for :

  • Artificial Intelligence x
Clear All
Open access

A. Godlewska

Abstract

Nowadays, the increasing number of non-linear loads influences the grid, causing grid voltage disturbances. These disturbances may be very dangerous for the equipment and can create faults in converter behaviour. However, the right control algorithm can improve the reliability of the work. For a current source rectifier, the finite control set model predictive control has been proposed. This method is very flexible because of the variety of the possible cost function forms. It has been examined under grid voltage disturbed by the higher harmonics and the voltage drop. Simulation results prove the ability to damp the distortions and to ensure the unity power factor. Summing up, the algorithm is a very good solution for use in applications such as battery charging, active power filtering and low-voltage direct current load feeding.

Open access

Maciej Chojowski

Abstract

The purpose of this paper was to present a method for the estimation of the rotor speed and position of brushless DC (BLDC) motor. The BLDC motor state equations were developed, and the model was discretised. Extended Kalman filter has been designed to observe specific states from the state vector, needed for the sensorless control (rotor position) and to determine the speed, which may be useful to use as a feedback for the controller. A test was carried out to determine the noise covariance matrices in a simulation manner.

Open access

Jacek Rąbkowski and Rafał Kopacz

Abstract

This paper presents a new concept for a power electronic converter - the extended T-type (eT) inverter, which is a combination of a three-phase inverter and a three-level direct current (dc)/dc converter. The novel converter shows better performance than a comparable system composed of two converters: a T-type inverter and a boost converter. At first, the three-level dc/dc converter is able to boost the input voltage but also affects the neutral point potential. The operation principles of the eT inverter are explained and a simulation study of the SiC-based 6 kVA system is presented in this paper. Presented results show a serious reduction of the DC-link capacitors and the input inductor. Furthermore, suitable SiC power semiconductor devices are selected and power losses are estimated using Saber software in reference to a comparative T-type inverter. According to the simulations, the 50 kHz/6 kVA inverter feed from the low voltage (250 V) shows <2.5% of power losses in the suggested SiC metal oxide-semiconductor field-effect transistors (MOSFETs) and Schottky diodes. Finally, a 6 kVA laboratory model was designed, built and tested. Conducted measurements show that despite low capacitance (2 × 30 μF/450 V), the neutral point potential is balanced, and the observed efficiency of the inverter is around 96%.

Open access

Maciej Komosinski

Abstract

The rapid development of technology has allowed computer simulations to become routinely used in an increasing number of fields of science. These simulations become more and more realistic, and their energetic efficiency grows due to progress in computer hardware and software. As humans merge with machines via implants, brain-computer interfaces and increased activity involving information instead of material objects, philosophical concepts and theoretical considerations on the nature of reality are beginning to concern practical, working models and testable virtual environments. This article discusses how simulation is understood and employed in computer science today, how software, hardware and the physical universe unify, how simulated realities are embedded one in another, how complicated it can get in application, practical scenarios, and the possible consequences of these situations. A number of basic properties of universes and simulations in such multiply nested structures are reviewed, and the relationship of these properties with a level of civilizational development is explored.

Open access

Gang Mu, Teodor Godina, Antonio Maffia and Yong Chao Sun

Abstract

In this paper, we make use of a Bayesian (supervised learning) approach in pricing American options via Monte Carlo simulations. We first present Gaussian process regression (Kriging) approach for American options pricing and compare its performance in estimating the continuation value with the Longstaff and Schwartz algorithm. Secondly, we explore the control variates technique in combination with Kriging to further improve the estimation of the continuation value. This method allows to reduce dramatically the standard errors and to improve the stability of the Kriging approach. For illustrative purposes, we use American put options on a stock whose dynamics is given by Heston model, and use European options on the same stock as control variates.

Open access

Tamás Majoros and Balázs Ujvári

Abstract

Neural networks are used as triggers at highenergy physics detectors. These triggers can separate the event that must be saved for later analysis from the other events or noises. Using the raw data of the detector, the signal and the background can be separated offline. After separation, sets of signals and backgrounds can be used to train the neural network. A gas-filled detector (multiwire proportional chamber) was used to study the trigger at different noise levels to find the most stable neural network that tolerates the random hits. The ratio of the recognized and the unrecognized signal and background events is used for the measurement. Its stability is part of the systematical uncertainty.

Open access

A. P. Vancea and I. Orha

Abstract

Our project describes a system for the automation and monitoring of a smart house. The system consists of several sensors such as: temperature sensor, humidity sensor, light sensor and presence sensor. The data from sensors is processed and transmitted to the central module via Xbee-ZigBee or to a smartphone through the Bluetooth module. The central module connects at the Internet via WiFi and through an application we can remotely monitor the state of the house or control various devices within the enclosure.

Open access

Tomáš Huszaník, Ján Turán and Ľuboš Ovseník

Abstract

Optical fiber has the great advantages of capacity and reliability. That is why network providers started to deploy FTTx (Fiber-To-The-x) optical access using various PON (Passive Optical Network) architectures. The leading technology right now is Gigabit PON (GPON). However, with increasing amount of multimedia we need to further develop existing technologies to go on with these high demands. Fiber-To-The-Home (FTTH) using 10G-PON technology for broadband access application is effective solution for high speed networks with high capacity. In this paper, we look at the passive optical network in the city of Košice and based on the real network we created simulation model of downlink of 10G-PON based FTTH with triple-play service.

Open access

Daniel Vamos, Stefan Oniga and Anca Alexan

Abstract

Personal activity tracker are nowadays part of our lives. They silently monitor our movements and can provide valuable information and even important alerts. But usually the user’s data is stored only on the activity tracker device and the processing done is limited by this modest processing power device. Thus it is very important that the user’s data can be stored and processed in the cloud, making the activity tracker an IOT node. This paper proposes a simple IOT gateway solution for a custom user monitoring device.