Showing 3 results for Naderi
R. Naderi, A. Rahmati,
Volume 4, Issue 4 (October 2008)
Abstract
Multilevel PWM waveforms can be decomposed into several multilevel PWM
components. Phase-shifted carrier (PSC) is an efficient decomposition technique. In this
paper, we have first demonstrated the equality of PSC and alternative phase opposition
disposition techniques. Second, we have modified PSC to accommodate other disposition
techniques. Third, we have investigated the effects of using asymmetrical carriers on the
spectrum of the resulting PWM waveform. Fourth, we have proposed a logical algorithm
for decomposing all types of multilevel PWM waveforms.
B. Nasersharif, N. Naderi,
Volume 17, Issue 2 (June 2021)
Abstract
Convolutional Neural Networks (CNNs) have been shown their performance in speech recognition systems for extracting features, and also acoustic modeling. In addition, CNNs have been used for robust speech recognition and competitive results have been reported. Convolutive Bottleneck Network (CBN) is a kind of CNNs which has a bottleneck layer among its fully connected layers. The bottleneck features extracted by CBNs contain discriminative and rich context information. In this paper, we discuss these bottleneck features from an information theory viewpoint and use them as robust features for noisy speech recognition. In the proposed method, CBN inputs are the noisy logarithm of Mel filter bank energies (LMFBs) in a number of neighbor frames and its outputs are corresponding phone labels. In such a system, we showed that the mutual information between the bottleneck layer and labels are higher than the mutual information between noisy input features and labels. Thus, the bottleneck features are a denoised compressed form of input features which are more representative than input features for discriminating phone classes. Experimental results on the Aurora2 database show that bottleneck features extracted by CBN outperform some conventional speech features and also robust features extracted by CNN.
A. Hesami Naghshbandy, K. Naderi, U. D. Annakkage,
Volume 18, Issue 3 (September 2022)
Abstract
The most challenging circumstance of forced oscillations (FOs) is when the power system is forced to oscillate at its natural frequencies. This paper uses a novel PMU data-driven mechanism to pinpoint the source of such phenomena under resonance. Following the detection of FOs, the instantaneous changes in the output power and angular velocity of the rotors are calculated. Accordingly, an energy-driven multilateral interaction pattern is obtained for all synchronous generators. Next, an appropriate positive weighted undirected graph is constructed through these functional patterns based on the spectral graph theory. These quantitative indicators are then analyzed through the eigenvalue spectrum of the normalized Laplacian matrix of the system graph reduced to the internal generator buses. Finally, the smallest value in eigenvectors corresponding to the two largest eigenvalues reveals the location of the source. The proposed methodology’s validation and verification studies have been performed on the WECC 3-machine 9-bus and New England 10-machine 39-bus benchmark power systems modeled in the Real-Time Digital Simulator (RTDS) and then analyzed in the MATLAB environment. The proposed methodology revealed to be fast and accurate in locating the source of FOs under challenging resonance situations with promising results while addressing the generator side origins.
