#4 2023

CONTENTS №4

Discussing a Topic…

Automation as the key to efficient production and use of electricity

Golubev A.V., Muravyov I.K., Nikonorov A.N., Naumov Yu.V., Obuvaev A.S., Tselishchev E.S.  Instrumentation training system and its integration with power plant simulator

The paper presents simulation models of field instrumentation of power plants with visualization aimed at training operations and maintenance staff. The models are integrated with training simulator of a heat power plant.
The system enables the training of instrumentation staff in control maintenance and troubleshooting of field instruments and actuators. The skills acquired include equipment inspection, sensing line purging, naught check, sensor contact broaching, limit switch tuning on actuators, scrolling, equipment diagnosis, transmitter replacement, removal, and tuning, etc. The functionality for the certification of operations and maintenance staff is also available.

Keywords: instrumentation training system, computerized training simulators, operator interface, plant model, simulation modeling.

Golubev P.A.  Automated validation of video frames of applied software of upper-level process control systems at A-plants

The paper discusses the automated validation of video frames from applied software of upper-level process control systems at A-plants. It outlines the problems and algorithms of video frame validation to be addressed for improving the quality both at the control subsystem design phase and during further modernizations of A-plants power units.

Keywords: upper-level process control system, A-plant, validation, applied software, databases, video frames.

Kuznetsov R.S.  Automated system for commercial accounting and regulation of thermal energy of an industrial enterprise

Modern approaches to creating a digital enterprise in relation to the management of the heat supply system are described. The project of the system for monitoring, analysis and regulation of thermal energy at the instrument-making enterprise Dalpribor of the industrial complex of Primorsky Krai is shown. The emphasis is on the analytical modules of the automated system, which allow you to quickly solve the problems of equipment diagnostics, monitoring the efficiency of heat supply processes and energy saving by setting the correct operating modes of the heating controllers.

Keywords: digital enterprise, heating controller, heat meter, information and analytical system, energy saving.

Samovolik A.N., Piskazhova T.V., Osipova V.A.  Troubleshooting procedure for industrial electrical automation boards

A stepwise troubleshooting procedure for industrial electrical automation boards with discrete control is developed. The procedure addresses the needs of end-user maintenance staff and third-party system integrators. It reduces qualification requirements, systematizes and accelerates the troubleshooting process and thus improves operations efficiency. The algorithms developed can be used in automatic troubleshooting systems.

Keywords: troubleshooting, electrical boards, fault diagnosis, troubleshooting algorithm, discrete systems.

Kalashnikov A.A.  Automatic calibration of measuring channels based on digital technologies

The paper summarizes the research in the field of developing methods and technologies for on-line calibration of measuring channels of industrial process control systems. With the example of pressure gages, the “digital calibration” mode was undertaken with successful full-scale rig test on automatic calibration. The results obtain aroused the interest of instrumentation services in nuclear, oil and gas, and other industries, and laid the scientific foundation for developing GOST R 70518-2022 National standard “Instrumentation and control systems of objects using atomic energy. Metrological assurance. Basic provisions” with reference to the recommended methods for automatic calibration of measuring channels at A-plants. Industrial application of the methods and technologies developed contributes to further development of automated measurement assurance of process control systems in nuclear industry as well as other Russian strategic industrial segments.

Keywords: automatic calibration, measuring channel, digital calibrator mode.

Tyurin O.G., Kornelyuk O.A., Ikonnikov S.V.  A method for controlling and maintaining article’s weight and dimensions in high-energy material pressing

Quality control is an importantt task in articles production from high-energy materials by through pressing in continuous extrusion machines. The paper proposes a method for controlling and maintaining article’s weight and dimensions within specified limits during high-energy material pressing.

Keywords: length measurement, weight measurement, laser ranger, quality management, process control system.

Malyshev G.S., Kamnev M.A., Drumov I.V., Terekhin G.V.  Selecting the dimensionality of fully connected classifier of a neural network dependent on the training set size in the binary flaw recognition problem

With the example of siliconized graphite articles, the paper compares the operation results of fully connected classifiers trained on various-size image sets. The graphs demonstrating the changes of percentage of correctly recognized images in process of classifier training are included, their features are analyzed. In the case of binary classification of flaws on siliconized graphite articles the number of neurons in the classifier may vary within a range where network retraining problems can be avoided. The architecture of a fully connected network classifier is presented along with the settings used for its training. The paper shows that further training of the deepest convolutional block of the neural network does not result in any significant changes in the network operation. An attempt is made to explain this effect.

Keywords: binary classification, flaws, dropout, machine vision, network retraining, fully connected network, pretrained network, convolutional neural networks, siliconized graphite.

Poleshchenko D.A., Tsygankov Yu.A., Sergeychik M.S.  Application of hybrid neural networks for predicting iron content in iron-ore concentrate

Iron ore dressing process is a discrete-continuous sequence of production operations performed under varying feed properties (due to unstable mineralogical makeup) and equipment parameters (due to work wear) and other factors impeding automatic control. The key quality indicator of the ore dressing process is iron percentage in the concentrate, which is measured only in the laboratory. Samples are taken at time intervals exceeding the process duration. To overcome this challenge, the ore content prediction subsystem was developed. It comprises of hybrid neural networks deployed on each processing unit.

Keywords: mining industry, predicting system, machine learning, deep neural network.

Krasnov A.E., Golovkin M.E.  Recognition of texture images with the help of incoherent image blurring

The paper offers a simple approach to the recognition of texture images invariant to their rotation on the plane. It provides a technology for describing texture images based on their incoherent image blur by finite-size windows with subsequent histogram analysis. Numerical experiments with various texture images show the importance of selecting right sizes of blur window apertures.

Keywords: image, incoherent blur, blur window, aperture, histogram, shift and rotation invariance, aperture adaptation.

Savkin M.A., Trifonov A.V.  Improving the reliability of foundation state prediction with the help acoustic emission methods

Industrial and civil construction objects may come to emergency state or suddenly break down because of the defects not detected in time or wrong assessment of their risk level, poor quality of building materials, or design faults. The paper discusses the ways to improve the reliability of foundation state prediction with the help acoustic emission methods in predictive data analysis systems.

Keywords: predictive analysis, acoustic emission, foundation, defects.

Grishin D.A.  K15 modular programmable controllers: effective domestic solution based on the best practices

The paper reviews design features and functionality of domestic K15 modular programmable controllers. It discusses the principles of their programming in C++ language and presents an application case.

Keywords: modularity, programmable controller, CPU, I/O modules.

Shcheglov D.K., Fyodorov D.A., Saibel A.G.  Analysis of possible application of artificial intelligence for predicting industrial production development trends

With the example of ChatGPT system, paper analyzes the possibility of applying conversational artificial intelligence systems in exploratory research and discusses the study methodology. The results of the analysis of possible ChatGPT application for predicting the development trends of the theories of engineering works and industrial production are expounded.

Keywords: artificial intelligence, large language models, exploratory research, ChatGPT, industrial production, development trends.