Devices with tunable characteristics and parameters are used in many technical fields. Such devices can be based on memristors, which serve as programmable potentiometers. The quality of the tuning is higher by means of memristors than with mechanical and digital potentiometers. We investigate a bandpass filter in the form of an active Wien bridge with a memristor. The filter is analyzed with the help of the nodal voltage method. The dependence of the resonance frequency on the parameters of the Wien circuit, the dependence of the quality factor, and the filter gain at resonant frequency on the parameters of the voltage divider are obtained. The dependences of the resonant frequency, quality factor, and gain at the resonant frequency on the parameters of the Wien filter were formed. The tuning of the main frequency features (the filter gain, quality factor, and resonance frequency) is shown to be independent. Under different values of memristance, the frequency features result from a simulation in LTspice. These features are less than 1 percent different from the corresponding features obtained analytically. Thus, the high precision of modeling and tuning of the frequency characteristics of the memristive Wien filter is demonstrated.

Memristors are devices built on the basis of fourth passive electrical elements in nanosystems. Because of the multitude of technologies used for memristor implementation, it is not always possible to obtain analytical models of memristors. This difficulty can be overcome using behavioral modeling, which is when mathematical models are constructed according to the input–output relationships on the input and output signals. For memristor modeling, piecewise neural and polynomial models with split signals are proposed. At harmonic input signals of memristors, this study suggests that split signals should be formed using a delay line. This method produces the minimum number of split signals and, as a result, simplifies behavioral models. Simplicity helps reduce the dimension of the nonlinear approximation problem solved in behavioral modeling. Based on the proposed method, the piecewise neural and polynomial models with harmonic input signals were constructed to approximate the transfer characteristic of the memristor, in which the current dynamics are described using the Bernoulli differential equation. It is shown that the piecewise neural model based on the feedforward network ensures higher modeling accuracy at almost the same complexity as the piecewise polynomial model.

Problems in measuring electromagnetic parameters for the sensors of different physical quantities, such as temperature, humidity, and radiation intensity, are well known. In some cases, for example, when measuring cryogenic temperatures, it is fundamentally important to minimize the number of cables connecting a sensor to the measuring device. In this paper, a method is developed for determining the electromagnetic parameters of a group of four sensors using only four communication cables between a sensor and a measuring device. For different configurations connecting the sensors to a group, the interrelations between the resistances of each sensor with the measured quadripole A-parameters, as well as the identities connecting between the elements of the A-parameter matrices of the sensor group with each other in the cases of different methods for connecting to a measuring device have been obtained. The proposed method makes it possible to provide an increase in the determination accuracy for sought-for electromagnetic parameters via performing multiple independent measurements.

When studying the features of electrical circuits with nonlinear elements in the course “Theoretical Fundamentals of Electrical Engineering”, students face with the problem of analytical analysis complexity. The use of software allows to optimize the learning process and consider more different options for tasks. Dynamic processes in three-phase alternating current circuits with a nonlinear load in the form of a coil with a ferromagnetic core and distinctive features of the analysis of such circuits are described. In connection with the difficulties of analytical calculations, the use of numerical analysis methods by means of the MatLab software is proposed.

Tunable band-stop filters are important for selective conversion of electrical signals in radio technical devices and systems, in equalizers for the noise suppression, in measuring complexes, in acoustic and hydroacoustic equipment, etc. These filters are usually complex due to a large number of amplifiers and passive elements. The implementation of tunable active RC filters with several operational amplifiers ensures the improved filters characteristics and low sensitivity of the basic parameters of filters to scattering of electrical elements values. The active RC filter, which is characterized by independent tuning within the wide limits of a resonance frequency, quality factor and transmission coefficient, is presented. The resonance frequency of the tunable band-stop filter is adjusted with the aid of a potentiometer with the maintenance of the transmission coefficient and small changes in the pole quality factor of a transfer function. The pole quality factor and transmission coefficient are independently adjusted by means of two variable resistors. The properties of the represented tunable band-stop filter provide high technological efficiency and convenient manner of the operating modes tuning.

The numerical calculation method of dynamic models with nonlinearities described by elementary non-power functions is represented. The computational difficulties in applying power series are analyzed. To overcome computational difficulties, a constructive procedure based on the analytical-numerical method of analysis developed by the authors is proposed. The numerical procedure is applicable to the models described by a system of nonlinear differential equations in the normal form of Cauchy. The proposed procedure, while preserving all the advantages of the power-series apparatus, eliminates the necessity for constructing the power-series compositions and special estimates, as well as raises the formalization of dynamic model calculations with non-power nonlinearities. The proposed numerical procedure uses the evaluations of the desired solutions and manages the boundaries of one-dimensional domains of exact solutions in order to bring them to unknown exact solutions. The advantages of the procedure are highlighted during the calculation of nonlinear autonomous dynamic model of the type «damped pendulum».

Atomic spectroscopy experiments have reached a level where the transition line profile has become measurable with high precision. One of the most important task of such experiments is the determination of the transition frequency and an accurate comparison with the corresponding theoretical value. However, a detailed theoretical description of the line profile requires consideration of increasingly tiny effects which can play a crucial role in determining fundamental physical constants. In particular, it was shown recently that the nonresonant corrections arising in the description of the scattering processes of photon by atoms can lead to significant changes in the determination of the transition frequency. In present work, an adaptive method of moments for extracting the transition frequency from the experimental spectral data is discussed. (C) 2022 Elsevier B.V. All rights reserved.

A dual autoencoder employing separable convolutional layers for image denoising and deblurring is represented. Combining two autoencoders is presented to gain higher accuracy and simultaneously reduce the complexity of neural network parameters by using separable convolutional layers. In the proposed structure of the dual autoencoder, the first autoencoder aims to denoise the image, while the second one aims to enhance the quality of the denoised image. The research includes Gaussian noise (Gaussian blur), Poisson noise, speckle noise, and random impulse noise. The advantages of the proposed neural network are the number reduction in the trainable parameters and the increase in the similarity between the denoised or deblurred image and the original one. The similarity is increased by decreasing the main square error and increasing the structural similarity index. The advantages of a dual autoencoder network with separable convolutional layers are demonstrated by a comparison of the proposed network with a convolutional autoencoder and dual convolutional autoencoder.

The analyses of current-voltage characteristics of metal oxide memristors resulted from the investigation of a real memristor is represented. The characteristics of the real memristor are measured with the help of the NI ELVIS workstation. The electrical properties of the memristor with bipolar resistive switching and the transformation of hysteresis curve into a line under increasing the frequency of the harmonic input voltage are observed on oscillograms. The estimation of resistances in low and high states on the bases of measured current-voltage characteristics is discussed.

Naive Bayes, logistic regression, linear support vector machine, and deep neural networks are estimated and compared from the point of fulfilling text classification. The naive Bayes approach can be used in real-time with big datasets, but it assumes that each feature makes an independent and equal contribution to the outcome. Logistic regression is a linear model that is easy to implement in classification problems but has poor performance on nonlinear data. The support vector machine algorithm is a linear algorithm that performs well in higher dimensions, but it requires significant time to process. The main advantage of the deep neural network model over the other techniques is that it reduces the need for extensive feature extraction and selection and the distributed representation of words as features input into the network.

In electrical engineering, radio engineering, robotics, computing, control systems, etc., a lot of nonlinear devices are synthesized on the basis of a nanoelement named memristor that possesses a number of useful properties, such as passivity, nonlinearity, high variability of parameters, nonvolatility, compactness. The efficiency of this electric element has led to the emergence of many memristor technologies based on different physical principles and, as a result, to the occurrence of different mathematical models describing these principles. A general approach to the modeling of memristive devices is represented. The essence is to construct a behavioral model that approximates nonlinear mapping of the input signal set into the output signal set. The polynomials of split signals, which are adaptive to the class of input signals, are used. This adaptation leads to the model's simplification important in practice. Multi-dimensional polynomials of split signals are built for the rectifier bridge at harmonic input signals. The modeling error is estimated in the mean-square norm. It is shown that the accuracy of the modeling is increased in the case of using the piecewise polynomial with split signals.

A full-wave rectifier is considered as a nonlinear device, widely used in domestic electric appliances, in distribution networks, electric driven complexes and therefore useful for studying in the course of electrical engineering. The rectifier implementation in the form of a diode bridge and the results of its investigation with the aid of the control and measuring complex NI ELVIS are shown. The properties and characteristics of the rectifier are evaluated using the obtained signal spectra. The rectified signal was detected by means of a low-pass filter and the error of this filtering was estimated.

A lot of electrotechnical systems are nonlinear, that is why the study of dynamic processes in nonlinear electrical circuits is very important and advisable. Dynamic analysis is performed in electrical circuits with nonlinear passive elements. Such elements include nonlinear resistors, which are described by the piecewise-linear volt-ampere characteristics, inductances and capacities. The analytic analysis of some nonlinear circuits, as well as the results of their numerical analysis in the MATLAB system are represented. The importance of the numerical analysis study is emphasized, since the complexity of electrotechnical systems makes it difficult to build their analytical models in practice.

In recent decades, memristive systems, synthesized on the basis of the fourth passive electrical element, the memristor, are developed. A memristor is a controlled resistor with memory, its resistance is nonlinearly dependent on the history of the current change in it. The application of memristor systems is wide, for example, memory devices, neuromorphic systems, neural networks, nanoelectronics devices used for various purposes. This diversity is due to the advantages of memristive systems (compactness, energy efficiency, high performance). Memristors are created using new materials and technologies. As a result, there are difficulties in their analytical and numerical modeling. The main mathematical models of memristors, their advantages and disadvantages, the possibility of implementation in LTspice are represented.

In this paper, the structure of a separable convolutional neural network that consists of an embedding layer, separable convolutional layers, convolutional layer and global average pooling is represented for binary and multiclass text classifications. The advantage of the proposed structure is the absence of multiple fully connected layers, which is used to increase the classification accuracy but raises the computational cost. The combination of low-cost separable convolutional layers and a convolutional layer is proposed to gain high accuracy and, simultaneously, to reduce the complexity of neural classifiers. Advantages are demonstrated at binary and multiclass classifications of written texts by means of the proposed networks under the sigmoid and Softmax activation functions in convolutional layer. At binary and multiclass classifications, the accuracy obtained by separable convolutional neural networks is higher in comparison with some investigated types of recurrent neural networks and fully connected networks.

Cellular neural network is a well-known mathematical model for investigation of deterministic, stochastic and chaotic processes in nonlinear dynamic systems. The turn to a reaction-diffusion cellular neural network (RDCNN) enables to model different types of autowave processes (concentric, circular and spiral waves, Turing structures, etc.) RDCNN is a convenient architecture for the mathematical description of the systems of nonlinear differential equations with diffusion and their solution based on numerical methods. Running circular and spiral waves in RDCNN are synthesized and their properties are demonstrated (preservation of the shape and amplitude of the running waves, as well as the destruction of waves which results from their collision.) Autowaves are built in the MATLAB system on using the fourth-order Runge-Kutta numerical method.

The tasks of identifying points of unauthorized power takeoff in a distributive network by measurements in some selected nodes have a lot of applications. The practical interest in solving them increases with a reduction in the number of network nodes necessary for monitoring. At a relatively constant voltage in the distributive network nodes, the determination of consumer capacities is confined to determining the parameters of the equivalent circuits of the latter. The determination of parameters by measurement results is a classic inverse problem from the circuit theory. This article proposes the solution of this problem for a 0.4-kV triphase distributive network with 9–12 loads to be controlled at measurements only in two nodes of the network. To solve this problem, it is proposed to set variable conductivity in one of the measurement nodes. The general approach to solving the problems of the considered type is provided. The derived solutions can be used to control the state of insulators in high-voltage lines and temperatures or humidity in hard-to-reach zones. The solutions derived for the 9- to 12-load network are relatively simple and allow promptly obtaining information on the network loads from the measured data. In addition, a solution for determining the EMF of several parallel-connected thermocouples and used as temperature sensors is provided.

The cascade structure of behavioral models in nonlinear dynamic systems is considered. This structure includes a splitter, implemented in the form of a delay line, and a nonlinear inertialess converter in the form of a feed forward neural network. The splitter provides the unique mapping of the input signals subset into the output signals subset. It adapts the form of a mathematical model to the class of the input signals; as a result, the behavioral model becomes simpler compared with universal models for the given class of inputs. The influence of split operation on the approximation of the Bernoulli memristor operator is demonstrated on the example of the harmonic input signals. The splitter is constructed as a delay line with a unit memory, the nonlinear inertialess converter – as a two-layer neural network.