#10 2025

CONTENTS №10

New book

Kuznetsov A.Yu., Reshetnikov I.S.>  ISA-95: Digital production architecture

Discussing a Topic…
Indirect measurements and system identification in process automation

Kiprov R.R., Zotov N.A., Shtakin D.V., Sneghirev O.Y., Torgashov A.Yu.  Developing Kriging-based quality estimator with a case study of industrial fractionator

The development of a quality estimator (QE) for methanol content in fractionator bottoms stream is discussed. A detailed comparative study of quality estimator development techniques was undertaken, which comprised multiple linear regression, decision trees, Gaussian process regression (aka Kriging), support vector machine, ensemble method. The paper shows that the Gaussian process regression method with an exponential covariance matrix ensures QE with the highest precision. An optimization algorithm for regression hyperparameters using lattice search is proposed. Its effectiveness as against the experimental parameter tuning is proved in QE modeling case.
Keywords: quality estimator, Gaussian process regression, optimization of hyperparameters, lattice search, industrial fractionation column.

Mozharovsky I.S., Shevlyaghina S.A.  Neural network-based product quality estimator using genetic algorithm

Neural network-based soft sensors are extensively used in industry for real-time estimation of product qualities. Their development requires much time for selecting the most effective structure of the artificial neural network (ANN). Tuning, learning, and testing procedures are applied to each experimental ANN, whereas its structural parameters, such as the number of hidden layers, the number of neurons in each layer, and the activation function have several possible values. The development of approaches to the selection of these hyperparameters would save the development time of ANN-based soft sensors and facilitate the research. The paper suggests using a genetic algorithm for the search of optimal ANN architecture in quality estimator development. The approach presents a crossover operator modification that makes it possible to obtain a variety of ANN hyperparameters. This results in higher accuracy of the developed ANN-based estimator. The algorithm terminates automatically once the estimator’s accuracy criterion attains the specified value. The effectiveness of the approach was demonstrated by a numerical example and by the case of a naphtha rerun column for estimating the quality of 35…70 С cut.
Keywords: quality estimator, neural network, genetic algorithm, optimization, naphtha rerun process, product quality estimation.

Vlasov A.K., Maksimenkov V.N.  Creation of inductive knowledge bases for digital identifiers using high-fidelity process simulation models

Digital identifiers develop mathematical models of nonlinear dynamic processes. The paper presents an innovative method for creating the knowledgebase of such identifiers. It proposes an approach to the primary knowledge base filling by data generation and acquisition using high-fidelity simulation models of processes under study. The paper shows that simulation models can be used for comparative analysis and validation of various identification algorithms. The effectiveness of the approach is demonstrated by the example of simulating ore grinding process in a ball mill of an ore mining and processing enterprise. The results confirm the method’s applicability for complex industrial nonlinear dynamic systems.
Keywords: associative search, identification, high-fidelity model, knowledge base, regression analysis, process modeling, grinding.

Bochkarev A.V., Panus S.A.  Developing an automated system for express analysis of liquid content in binary oil/water mixtures

The paper examines the possibility of building a system for automated online measurement of volumetric liquid content in binary oil/water mixtures by the sound speed of an ultrasonic wave transmitted through the sample. It presents key mathematical models describing the relationships between the sound speed and the volumetric percentage. The development phases of a machine learning method for volumetric liquid content evaluation in a mixture are outlined. A pilot automated measurement system is proposed as well as a data processing procedure.
Keywords: ultrasound, volume percentage, binary mixtures, sound speed, analytical methods, moisture in petroleum products, machine learning.

Chesalov A.Yu.  Ethical aspects of the use of artificial intelligence in industry

Ethical challenges posed by the integration of artificial intelligence (AI) systems into industry are analyzed. The paper examines the primary risks and concepts underlying the design, implementation and operation of AI systems at all their lifecycle stages based on key national and international documents: UNESCO recommendation on the ethics of AI, Chinese specification of new generation AI ethics, EU Artificial Intelligence Act, and Russian code of ethics in the field of AI. The focus is given on safety, transparency, explainability, and controllability of industrial AI systems in digitalization conditions.
Keywords: artificial intelligence, artificial intelligence ethics, industry, predictive maintenance, industrial safety.

Gerasimov N.G., Muratov A.G., Protalinsky O.M.  Parametric diagnostics system for RU BREST-OD-300 steam generator furnaces using artificial intelligence techniques

The paper describes the design concepts of a parametric diagnostics system for steam generators of RU BREST-OD-300 fourth-generation reactor unit with lead heat-carrier using artificial intelligence technique for leak detection in heat exchange tubes. The problem is reduced the recognizing fuzzy parametric patterns of flaws expressed by specific changes of reactor unit’s control variables available in the system’s knowledgebase. The authors propose a technique for determining parametric images of heat exchange tubes based on thermohydraulic calculations on the plant model using new-generation calculation software. The application of fuzzy inference systems (FIS) as a framework for parametric diagnostics system is substantiated. System’s key development phases are presented as well as its implementation in the SimInTech software suite.
Keywords: parametric diagnostics, artificial intelligence, power industry, reactor unit, pattern recognition, fuzzy production systems, operator support system.

Martinova L.I., Sokolov S.V., Martemianova N.S.  Developing a toolkit for auto-configuration of remote I/O modules in numerical control systems

The paper presents an approach to the development of auto-configuration toolkit for EtherCAT I/O modules. The toolkit is integrated in the AksiOMA Control NC system for developing automatic configuration of slave EtherCAT devices. The proposed solution makes it possible to resolve a number of problems arising during the configuring and reconfiguring of heterogeneous I/O modules in an NC system. The toolkit formalized the process of IO modules configuration and their mapping on SoftPLC shared memory, that makes possible the elimination of human factor in configuration tasks.
Keywords: NC, auto-configuration, buscoupler, remote I/O, EtherCAT, library.

Posevin A.O.  SCADA system in (open, LIBRE) office calc

The paper describes a mini-SCADA system compatible with Linux-based freeware. The SCADA software is available for free download (a link is provided in the paper); it can be used for specialist training and scientific research as well as in small industrial supervisory applications compatible with Modbus RTU.
Keywords: industrial protocol, SCADA system, operating system, freeware, process control system.

Koshelev E.V.  Practical implementation of cluster-oriented security model by the example of smart energy metering infrastructure

The paper continues the earlier one dedicated to the cluster-oriented model for security assessment of distributed information systems, where key concepts and calculations were presented. The current work describes the model’s practical implementation by the example of the smart energy metering infrastructure, wherefrom the data are used for components clustering and the calculation of the integrated security index Z, subject to quality metrics. Based on cluster analysis results and the visualization, the paper demonstrates the effectiveness of the proposed approach to security assessment, that confirms the applicability of the model in real conditions and opens up new prospects for further experimental research.
Keywords: distributed information systems, smart energy metering infrastructure, clustering, resilience metrics, security.

Industrial Automation Companies

I strongly believe in PC-based automation.

In summer 2025, Sinsegye (translated as “new world”) has entered the Russian industrial automation market. Sinsegye is a four years old company, but its industrial automation products are well known both in China and abroad. Dr. Ma Jun, the Sinsegye founder and CEO informs about the company’s goals and objectives.
Keywords: industrial automation, software, I/O modules, drive technology, sensors, mastermind.