No. 3

Abstracts. No. 3, 2021 year
 
ABSTRACTS
Shestakov A.L., Qodirov Sh.Sh. (e-mail: admin@susu.ac.ru)
The predictive model of technological tool sticking in the well drilling process on the base of deep neural network

Technological tool sticking has a major effect on drilling efficiency and well costs. This problem is affected by many parameters, such as geological parameters, technical and technological drilling parameters, technological and rheological parameters and properties of the drilling fluid. To account for all of the aspects of technological tool sticking and anticipated nonlinear behavior of the variables involved, neural network modeling can be used as primary predictive tools. This paper presents a study of the application of the concepts of neural network with sliding window to the problem of stuck drill tool. These methods make it possible to estimate the risk of stuck tool occurrence in the well during drilling in real-time. The test of the quality of the developed model is carried out on the real industrial data set of stuck (on the example of oil wells in the Tyumen region, Russian Federation).

Keywords: data analysis, data preparation, sliding window, neural networks, prediction, technological tool sticking.

Rudskoy A.I. (e-mail: rector@spbstu.ru)
Digital industry based on digital twins

The world is changing rapidly, and the COVID-19 pandemic is accelerating many processes, especially that of digital transformation. According to the Accenture Technology Vision 2021 report published in February 2021, leading enterprises have been forced to compress their decade-long transformation agendas into two to three-year plans.

Transition to a new, digital era, to a new technological stage leads to the emergence of new technologies, methods, solutions, while many conventional technologies are becoming obsolete and cannot ensure global competitiveness.

Of the greatest interest are the changes that are unfolding in the high-tech industry. At the same time, one has to be aware of the fact that the contribution (weighting factor) of some technologies to ensuring the global competitiveness of companies and even countries is much higher than that of other technologies. One of these most significant technologies is the digital twin technology.

Keywords: Digital Twins, digital industry, digital platform.

Ragozin A.N. (e-mail: ragozinan@susu.ru)
Processing of measuring signals using the coherent method of spectrum with the reference signal in diagnostic systems of the state of technical equipment

In systems for diagnosing the state of technical equipment, an important place is occupied by a measuring subsystem for identifying parameters and characteristics of signals observed from the outputs of various diagnostic and parametric sensors. The frequency components of measuring signals may contain important diagnostic features discrete spectral components of the signal spectrum. To detect stable in frequency (with insignificant random angular modulation) discrete spectral components of the spectrum of the measuring signal, it is proposed to use the method of spectrum coherent with the reference signal, which is informative in the problems of diagnostics of the tested technical systems. Classical spectral estimation methods do not allow detecting such diagnostic components in a complex diagnostic signal. The study considers the use of the method of a coherent spectrum with a reference signal (CR-spectrum method) for detecting discrete spectral components of the spectrum of the observed diagnostic signal that are stable in the center frequency with insignificant random angular modulation in the signal.

Keywords: spectrum coherent with the reference signal, frequency components, coherent spectrum, coherence function, time-frequency transformation.

Rednikov S.N., Ahmedyanova E.N., Ahmedyanova K.T., Taimurzin D.A. (e-mail: srednikov@mail.ru)
Efficiency of diagnostics of metallurgical equipment

The article presents the experience of diagnosing metallurgical equipment. The most common failures of metallurgical units are considered. The analysis of modern methods of diagnosis and non-destructive control is given. The method of predicting the change in the state of units during operation is provided. The issues of primary diagnostics are considered in detail by analyzing the volume distribution of temperature received using thermal imaging method to control the condition of the equipment. The results of modern methods of pre-diagnosis are considered. The author's experience in diagnosing the equipment of metallurgical enterprises is described. A comparative assessment of diagnostic methods with a calculated definition of effectiveness is given.

Keywords: vibrodiagnosis, thermal equipment, working fluids, diagnostic measures.

Grigorova E.A., Sartasov E.M., Sukhovilov B.M., Tushev S.A. (e-mail: grigorovaea@susu.ru)
Contactless measurements of minimum clearance outline (structure gauge) of railway platforms with PHS-1 photogrammetry system

We describe the solution to the problem of contactless measurement of minimum clearance outline (structure gauge) of railway platforms by the means of the FS-1 photogrammetric system of our own design. We modified FS-1 components to make them problem-specialized, thus making it possible to opt-out of the use of a precision scale bar. This significantly reduces the overall system cost and increases the convenience of measurements while ensuring the required accuracy of the system. A robust computational algorithm for estimating platform dimensions has been developed. The quality of the assessment is confirmed by the results of simulation modeling.

Keywords: photogrammetric system, coded targets, artificial targets, railway platforms, dimension measurement, simulation modeling.

Shnayder D.A., Kalinina E.A. (e-mail: shnaiderda@susu.ru)
Model-predictive control of iron ore concentrate production in the advice-giving mode

At processing units of beneficiation plants, management and control over ore concentrate production process is performed by an operator working from the automated work station. The operator sets the parameters of the equipment operation in many cases intuitively, based on the experience and the provisions of the manufacturing specification for ore concentrate production. The complexity of setting the equipment parameters is conditioned by the necessity to take into account simultaneously a great number of factor (such as: base ore quality, equipment integrity, mode parameters and others).

The paper suggests increasing the efficiency of managerial decisions by means of implementing model-predictive control of iron ore concentrate production in the advice-giving mode. The system is based on an optimization mathematical model for controlling the technological process of concentrate production. The suggested approach based on the information about the technological process enables conducting assessment of the controlled factors effects on iron-ore concentrate qualitative characteristics and energy consumption. It allows the user of the system-adviser to choose options for adjustment of the controlled parameters of a technological process to optimize it according to one or several techno-economic criteria, including productivity, concentrate quality, useful losses in tailings.

Practical implementation of the suggested approach within the framework of the existing automated control system will allow the operator to receive recommendations about the correction of a control parameter of the technological equipment operation regimes in the real time and on the basis of the current values of the section parameters in crushing-and-preparation plants on the basis of the chosen by the user technical or techno-economic optimization criteria.

Keywords: beneficiation plant, model-predictive control, automated system-advisor, Data Mining, decision trees.

Kolesnikov I.., Korzhov A.V., Gorshkov K.E. (e-mail: ivankolesnikov95@yandex.ru)
Unified Model of Digital Diagnostics of the Power Transformer Condition

The article is devoted to the development of digital technical diagnostics of power transformers. In the course of the study, the main methods of technical control and assessment of the state of electrical equipment were considered. Based on the analysis results, a model has been developed in the LabView software package, which is an imitation of the digital twin version of a power transformer. This model simulates the collection and processing of the necessary data from continuously operating sensors in real time, and also calculates various technical parameters based on the information collected. The program allows you to simulate various situations, as well as gain experience in the development and formation of effective diagnostic recommendations based on an assessment of the possibility of an emergency state.

Keywords: power transformer, digital twin, electrical equipment condition diagnostics, digital program, LabView.

Ibryayeva O.L., Shepelev V.D., Slobodin I.S., Zhulev A.E., Chizhova M.K. (e-mail: ibriaevaol@susu.ru)
Adaptive regulation of traffic and pedestrian flows using machine vision

In large cities, there is a continuous increase in the intensity of traffic flows, which leads to an increase in congestion and emissions of harmful substances, a decrease in the mobility of residents, and an increased risk of road accidents. The paper is devoted to the problem of optimal compliance with the interests of drivers and pedestrians in the pedestrian crossing area when turning right on the basis of data collection using neural networks. An algorithm for adaptive vehicle traffic management, taking into account pedestrian traffic, has been developed to improve safety and throughput at regulated intersections when turning right. The adaptive control method is based on the principle of reducing the exit time of pedestrians, depending on the queue of vehicles, to the minimum, which ensures a comfortable transition of the roadway. It is established that this approach allows to increase the bandwidth of the lane with allowed traffic only to the right up to 20 %. The software is written in Python using the OpenCV computer vision library. All calculations are performed in real time using the YOLOv3 neural network and the SORT tracker with some modifications (4-scale prediction in YOLO). The basic characteristics of YOLO have been improved by using an additional mask branch and optimizing the shape of the anchors.

Keywords: neural networks, YOLOv3, SORT tracker, adaptive regulation, throughput, traffic flow regulation.

Kaftannikov I.L., Kozlova A.V., Khlyzov A.D. (e-mail: kaftannikovil@susu.ru)
Prototype of a Li-Fi Communication System for Data Exchange Between Mobile Devices

This paper deals with the design and development of a Li-Fi (light fidelity) simplex communication system for data exchange between Android mobile devices. Li-Fi is an up-to-date technology in the modern world, since it uses visible light for data exchange, allowing for high-speed communication. The paper includes a brief review of Li-Fi technology, a review of the literature used, and a study of technological methods for implementing such systems, based on scientific sources. We propose the algorithms for data exchange, packet formation, and encryption-decryption. The paper presents the developed mobile application and the transceiver device, the development results, as well as experiments with the developed prototype. The results show that Li-Fi technology is workable and is a good alternative to existing communication methods.

Keywords: VLC, Li-Fi, Android, simplex communication, light fidelity.

ack to content