Search results for "NEURAL NETWORK"

A Neural Network Based Approach for the Design of FSW Processes

2009

Friction Stir Welding (FSW) is an energy efficient and environmentally "friendly" welding process. The parts are welded together in a solid-state joining process at a temperature below the melting point of the workpiece material under a combination of extruding and forging. This technology has been successfully used to join materials that are difficult-to-weld or ‘unweldable’ by fusion welding methods. In the paper a neural network was set up and trained in order to predict the final grain size in the transverse section of a FSW butt joint of aluminum alloys. What is more, due to the relationship between the extension of the “material zones” and the joint resistance, the AI tool was able to…

On-line adaptive neural network in very remote control system

2006

Remote control involves several issues that degrade seriously the performance of the plant to be controlled. This paper presents a strategy improving the characteristics of the remote control system, using an on-line adaptive neural net, in order to learn the variations of the remote system parameters to minimize the errors. This strategy is successfully applied to a client-server remote control system for a two link robot arm. Tests show that an error position in a remote control brushless motor can be highly reduced since its first "reference command" using a prevision of that error to modify the original reference. The neural net, used only by the client, is previously trained using loca…

Sensorless Control of PMSM Fractional Horsepower Drives by Signal Injection and Neural Adaptive-Band Filtering

2012

This paper presents a sensorless technique for permanent-magnet synchronous motors (PMSMs) based on high-frequency pulsating voltage injection. Starting from a speed estimation scheme well known in the literature, this paper proposes the adoption of a neural network (NN) based adaptive variable-band filter instead of a fixed-bandwidth filter, needed for catching the speed information from the sidebands of the stator current. The proposed NN filter is based on a linear NN adaptive linear neuron (ADALINE), trained with a classic least mean squares (LMS) algorithm, and is twice adaptive. From one side, it is adaptive in the sense that its weights are adapted online recursively. From another si…

Application of Neural Networks and Expert Systems in a hierarchical approach to the intelligent greenhouse control

2003

A novelty methodology based on the hierarchical combination of neural networks and expert systems is proposed in a centralized approach for the intelligent greenhouse control. The knowledge-based system is in charge of carrying out the determination of PH land value, composition, carbonic anhydride artificial atmosphere, external and internal temperature, wind and humidity measurements. From the results obtained, and by means of a neural network developed and trained for this application, the land quality is evaluated. On the other hand, the expert system, apart from supervising the system function and implementing fault tolerance mechanisms, performs the opportune actions in function of th…

Factors Affecting Attrition among First Year Computer Science Students: the Case of University of Latvia

2015

<p class="R-AbstractKeywords"><span lang="EN-GB">The purpose of our study was to identify reasons for high dropout of students enrolled in the first year of the computer science study program to make it possible to determine students, who are potentially in risk. Several factors that could affect attrition, as it was originally assumed, were studied: high school grades (admission score), compensative course in high school mathematics, intermediate grades for core courses, prior knowledge of programming. However, the results of our study indicate that none of the studied factors is determinant to identify those students, who are going to abandon their studies, with great precisio…

2014

The problem of robust decentralized adaptive neural stabilization control is investigated for a class of nonaffine nonlinear interconnected large-scale systems with unknown dead zones. In the controller design procedure, radical basis function (RBF) neural networks are applied to approximate packaged unknown nonlinearities and then an adaptive neural decentralized controller is systematically derived without requiring any information on the boundedness of dead zone parameters (slopes and break points). It is proven that the developed control scheme can ensure that all the signals in the closed-loop system are semiglobally uniformly ultimately bounded in the sense of mean square. Simulation …

Vibration analysis for bearing fault detection and classification using an intelligent filter

2014

Abstract This paper proposes an intelligent method based on artificial neural networks (ANNs) to detect bearing defects of induction motors. In this method, the vibration signal passes through removing non-bearing fault component (RNFC) filter, designed by neural networks, in order to remove its non-bearing fault components, and then enters the second neural network that uses pattern recognition techniques for fault classification. Four different categories include; healthy, inner race defect, outer race defect, and double holes in outer race are investigated. Compared to the regular fault detection methods that use frequency-domain features, the proposed method is based on analyzing time-d…

The Jiles Atherton Model for Description Of Hysteresis in Lithium Battery

2013

In this paper Jiles Atherton (JA) Model is used to obtain a mathematical model of the hysteresis in lithium battery. JA Model allows to describe both the hysteresis and the dynamical features of charging and discharging cycles in a lithium battery. The identification of the model is obtained by using a neural network technique developed for magnetic systems. The model is validated on some experimental tests on commercial batteries.

Comparative analysis of vehicle to pole collision models established using analytical methods and neural networks

2010

This paper presents a comparison between two modeling approaches of vehicle to pole collision. Firstly, analytical and curve fitting methods are explained and subsequently they are utilized to create lumped parameter models. Having parameters of such systems and their responses we proceed to brief description of the radial basis function neural network and its application to the linear models' coefficients' identification. Comparative analysis of the models formulated according to those two different manners is done. (6 pages)

Application of learning pallets for real-time scheduling by use of artificial neural network

2011

Author's version of a chapter in the book: 5th International Conference on Software, Knowledge Information, Industrial Management and Applications (SKIMA). Also available from the publisher at: http://dx.doi.org/10.1109/SKIMA.2011.6089986 Generally, this paper deals with the problem of autonomy in logistics. Specifically here, a complex problem in inbound logistics is considered as real-time scheduling in a stochastic shop floor problem. Recently, in order to comply with real-time decisions, autonomous logistic objects have been suggested as an alternative. Since pallets are common used objects in carrying materials (finished or semi-finished), so they have the possibility to undertake the …