A method for stabilizing structural anomaly detection under additive noise conditions as well as an algorithm for formal selection of the parameters of the solver rule in the structural anomaly detector based on the Robust Random Cut Forest (RRCF) method are proposed. In the framework of the developed approach, in order to stabilize the process of structural anomaly detection under the influence of additive noise, it is proposed to feed to the input of the RRCF-detector a data stream which is pre-processed by one of the digital filtering methods. In this case, the decision rule for anomaly detection is strictly formalized and transparently interpreted. The selection of parameters of the RRCF-based anomaly detector stabilized by pre-filtering methods of the input data stream is formalized. The RRCF-detector parameters choice within the proposed scheme guarantees a predetermined upper bound for the false alarm probability when deciding to detect a structural anomaly. This property is rigorously proved and formalized as a theorem. The performance of the stabilized RRCF-detector is investigated numerically. The achieved results confirm the performance of the proposed approach provided that the detection threshold is selected in the way proposed in this paper. An example of practical application of the proposed method is presented. The developed approach is promising for the detection of structural anomalies in conditions of observation additive noise, in a situation where it is important to guarantee an upper bound for the probability of false alarm. In particular, the approach can find application in monitoring technological regimes of liquid pumping in pipeline systems or in systems for detecting pre-failure states of technological equipment.

As part of the development of the technological concept of industry 4.0, of considerable interest is the issue of predicting the economic and business effect that a particular company may receive from the digital twin’s implementation in the management circuit of its key business processes. Currently, there is no widely accepted methodology for obtaining a predictive estimate of this type. The way to create such methodology, which is based on predictive estimates of the relative effect of the implementation of digital twins on a set of formal metrics, looks promising. The current configuration of business processes is used as the base. This paper proposes a transparent methodology of the mentioned type, which is designed to obtain a predictive estimate of the quantitative type for the value of the economic and business effect that a company will receive after the digital twin’s concept implementation. This assessment has a transparent interpretation and is quite naturally associated with a group of metrics that are commonly used to evaluate the company performance. For example, the metrics that a company uses within the Balanced Scorecard framework can be used as a group of such metrics. It also proposed an original composition of digital twins, designed for the railway traffic management. On the example of this digital twin composition step by step demonstrated the mechanism of using the proposed predictive methodology to quantify the value of the economic effect that a railway company will receive from the digital twin’s implementation. The methodology proposed in the article to obtain a predictive estimate of the digital twins implementation economic effect is quite transparent and simple. Therefore, it can be used in the practice of financial planning without much difficulty.

This paper proposes a system designed to detect hydrogen leaks from a pipe system. The system is characterized by a “light” architecture and is installed on the pipeline according to the principle: attach the IIot-sensor to the pipe surface and forget about it. This means that the IIoT-sensor itself “integrates” into the monitoring system, communicates with adjacent IIoT-sensors, initializes a self-test procedure and starts to operate. The IIoT-sensors do not re-quire any external power supply and can function autonomously for several years. Communication between IIoT-sensors is organized on the mesh-network principle, the transmission protocol (LoRa based) is such that data are transferred directly from IIoT-sensor to sensor without the use of base stations. The algorithmic part of the system implements a new two-stage leak detection method, which provides high reliability of detection of the target event with low probability of false alarms. No information about the probability of noise and signal distribution is used. The a priori information about signals and noise, - is minimal. The algorithms are very simple to implement and work efficiently in real time with minimal computing power requirements.

Oil refining processes require large quantities of fuel gas. For rational consumption of fuel gas, digitalization of fuel gas consumption is required. The object of the study is the hydrotreating process. The subject of the study is digitalization of fuel gas consumption during hydrotreating. Main research methods: analysis, comparison, econometrics methods. The authors considered the process of hydrotreating diesel fuel. The authors applied reliable statistics from the refinery. Box-Jenkins methods were applied to the obtained time series of specific fuel gas consumption to construct a forecast model of fuel gas consumption during hydrotreatment. The obtained fuel gas flow model will allow predicting the fuel gas flow rate depending on the planned load of the hydrotreating unit. Fuel gas flow control will reduce the cost and price of diesel fuel. Accordingly, reducing the transport component will reduce the prices of goods for which transportation is needed. The further direction of the study is to draw up forecast models for fuel gas consumption at other plants of the enterprise.

A new method of guaranteed solution for multiclass classification problem of stochastic objects is proposed. Within the framework of the proposed approach, the classification result is a finite set of class indices which with a predetermined confidence coefficient contains the index of the class to which the object being classified corresponds. In this case, the classification itself is realized on the basis of using a classifier of the new type which is called a confidence Lipschitz classifier. The definition of the confidence Lipschitz classifier is given and its main properties have been studied. Among them, the property of guaranteed reliability of the classification which is expressed in the construction of a confidence set of limited size containing the index of the true class with a predetermined coefficient of confidence, has been studied. The case of the assembly of Lipschitz classifiers, the properties of which are formalized in the form of a theorem, is considered. We consider a practically important example of using the proposed approach in the problems of compensation of the noise process dynamics in the channels of the fiber-optic monitoring system. The proposed approach is promising for use in those classification tasks in which the number of classes has an order higher than the second, including large-scale biometric identification systems as well as multi-channel systems for monitoring extended objects.

Binary and ternary composites (CM) based on M-type hexaferrite (HF), polymer matrix (PVDF) and carbon nanomaterials (quasi-one-dimensional carbon nanotubes-CNT and quasi-two-dimensional carbon nanoflakes-CNF) were prepared and investigated for establishing the impact of the different nanosized carbon on magnetic and electrodynamic properties. The ratio between HF and PVDF in HF + PVDF composite was fixed (85 wt% HF and 15 wt% PVDF). The concentration of CNT and CNF in CM was fixed (5 wt% from total HF + PVDF weight). The phase composition and microstructural features were investigated using XRD and SEM, respectively. It was observed that CM contains single-phase HF, γ- and β-PVDF and carbon nanomaterials. Thus, we produced composites that consist of mixed different phases (organic insulator matrix-PDVF; functional magnetic fillers-HF and highly electroconductive additives-CNT/CNF) in the required ratio. VSM data demonstrate that the main contribution in main magnetic characteristics belongs to magnetic fillers (HF). The principal difference in magnetic and electrodynamic properties was shown for CNT- and CNF-based composites. That confirms that the shape of nanosized carbon nanomaterials impact on physical properties of the ternary composited-based magnetic fillers in polymer dielectric matrix.

A practically effective solution to the problem of automated processing of ice reconnaissance data in high latitudes is proposed. The intermediate result of ice reconnaissance is huge aerial survey data set consisting of images of low quality; this is a consequence of the difficult conditions of aerial survey in high latitudes. The goal of the study is to create a high-level method that can either efficiently process this pre-collected data set or perform real-time processing of similar images while ensuring high reliability in solving the problem of recognizing ice class distribution on the water surface with minimal computing resources. In particular, the problem of automatic classification of ice-floe size distribution (FSD) type for a three-class model based on aerial survey data is solved. The practically important case of low-quality images is considered, a common situation for the meteorological conditions of the Far North. The proposed approach is based on the use of machine learning methods, in particular on the well-known multi-class SVM (Support Vector Machine), which is extremely undemanding to computing resources and therefore can be implemented even by the onboard computer of an ice reconnaissance UAV. From the input images of low quality some numerical characteristics of the image are calculated which informatively characterize the image. These characteristics (features) are invariant to scaling, rotation and illumination as well as have a much smaller dimensionality than the original image. The main idea underlying the proposed method is to form an original set of features which are implemented in the original feature space. These features characterize large fragments of the analyzed image and are “stable”, in contrast to the features that characterize small details. A new method of FSD type classification based on the processing of aerial survey data by using machine learning methods, which is sufficiently effective for processing low-quality images, has been proposed. Also, the original feature space for classification was proposed which ensured high practical efficiency of this method. The method has shown high efficiency when it is tested on a data set composed of low-quality real images (high blurriness, vagueness, presence of meteorological noises). The developed algorithm can be used for express analysis of ice reconnaissance data, including an ice reconnaissance UAV on-board software component.

The article proposes a new method of monitoring the infiltration processes developing inside the body of hydraulic structures. The method is based on the use of DAS (distributed acoustic sensing) fiber-optic technology which provides high spatial continuity of hydraulic structure seismoacoustic field analysis; digital twin infiltration dynamics and efficient signal processing methods based on machine learning. As a distributed sensor of the object’s seismoacoustic field, a DAS system is used the fiber-optic sensor of which is installed inside the body of the structure according to the principle of maximum coverage. The infiltration activity inside the structure body is estimated based on the analysis of an infiltration flows ensemble which are detected and classified by machine learning (ML) methods. These infiltration flows are sources of seismoacoustic emission and are therefore confidently detected by the DAS system. A digital twin of the infiltration dynamics based on the equations of mathematical physics is used as the normal basis for estimating the current state of fluid activity in the body of the structure. The risk of a structure failure under the influence of the observed infiltration flow is estimated within the framework of the proposed formal method based on the digital twin data. Based on the analysis of Data Set, consisting of real signals of infiltration processes, the high efficiency of detection and classification of this type of signals with the special ML-classifier included in the monitoring system is proved. A digital twin model of the infiltration processes dynamics in the body of a hydraulic structure is proposed. On the basis of the digital twin model, a method for estimating the risk of damage to the body of a hydraulic structure, which may occur as a result of the observed infiltration activity, is proposed. The method of controlling infiltration processes inside hydraulic structures can be used to monitor the operational condition of almost any hydraulic structures, including those in the cryolithic zone.

The paper proposes a novel and practically effective method for detecting, classifying, and estimating the coordinates of the image center of a small-size target object on a noisy scene, which is invariant to linear conformal transformations (rotation, shift, and scale). We consider a binary classifier that decides whether a particular part of the scene contains the desired image or only the background. The proposed approach implies an interactive procedure for finding an extremum of a function that approximates the likelihood function of the binary classifier. A two-step procedure based on the Nelder-Meade method is used to implement the extremum search. In order to ensure the robustness to noise and linear conformal transformations, both special training methods and the approach based on using an ensemble of classifiers, each of which corresponds to a certain scale, are applied in training the classifier. The author created a method for detecting a blurred image of a small-sized object in a scene that is distorted by correlated noise and proposes simultaneous estimation of the coordinates of the center of the target image. The method is robust to linear conformal distortions and has been successfully tested both on the artificial model and real images. The results of numerical study confirmed the robustness of the method to correlated noise of additive type and to linear conformal transformations. Within the framework of the proposed approach, the problem of constructing a confidence set for the coordinates of the target image center has been formally solved, and the efficiency of the obtained solution has been numerically investigated. The properties of the confidence set are formalized in the form of a theorem. The work also makes a comparison with the classical correlation-extreme method. If necessary, the proposed method can be easily generalized to the multiclass case. The method can be applied to machine vision systems, including online analysis of aerial survey data and to systems for video monitoring of the mechanical condition of complex technical equipment under conditions of strong meteorological disturbance.

Coatings made of the materials that effectively absorb radiation, e.g., ferrite materials, are used to reduce the level of electromagnetic radiation in rooms containing household or industrial equipment. It is known that significant dissipation of the radiation energy is provided by the thickness of the shielding coating which should be comparable to the length of the electromagnetic wave in the material which, in turn, significantly decreases at high values of the magnetic permeability and permittivity of the radio-absorbing material. Ferrite radio-absorbing coatings are characterized by the high heat resistance, low flammability and small (10 – 20 mm) thickness. However, at frequencies less than 40 MHz, plates with a thickness of more than 30 mm are to be used to provide the effective absorption, and the weight and cost of the coatings increase significantly. The results of studying the effect of the sintering temperature and micro-additives of titanium, calcium and bismuth oxides on the dielectric constant of Ni- and Mn-Zn radio-absorbing ferrites are presented. Reactively pure starting oxide components with a basic substance content of more than 99.6 % wt. were used to synthesize samples using traditional oxide technology. It is shown that alloying with bismuth and titanium oxides is rather effective for obtaining radio-absorbing ferrites with a combination of high values of the magnetic permeability and dielectric permittivity. The obtained results can be used in production of ferrite radio- absorbing materials operating in the megahertz range.

Barium hexaferrite thin films were prepared by ion beam deposition on sapphire (1 0 2) single crystal substrates. The structural properties of the films were analyzed by means of X-ray diffraction, atomic force microscopy, magnetic force microscopy and scanning electron microscopy. The obtained results were interpreted as a presence of layered structure with different types of texture and its formation mechanism is proposed. The significant influence of BaFe12O19 thickness on its microstructure was observed. Films obtained under certain conditions are potentially fit for further growing of oriented hexaferrite with out-of-plane orientation of c-axis.

The problem of monitoring the state of landfills is described in the article. There are a lot of such objects in the world. At the same time, there are no standard solutions for monitoring the state of these facilities both in Russia and abroad. It is proposed to develop a technical solution based on autonomous sensors for measuring the concentrations of hazardous fumes, radiation background, geotechnical factors and other environmental variables. Such system can be easily installed at any site and can work offline for a long time. The design and implementation of this system are undoubtedly connected with the issue of investment analysis. Positive economic results of the use of such systems can be an important target component in concept of green economy.

The article examines the principles of developing wireless networks of autonomous gamma sensors in order to create systems for spatial environmental radiation monitoring. The main task of such systems is to control the level of gamma radiation in areas where potential sources of ionizing radiation are located. An autonomous gamma-ray spectrometer is used as a measuring sensor. The authors propose to apply measuring sensors based on a silicon photomultiplier to create autonomous wireless networks of the industrial Internet for radiation monitoring. To confirm the possibility of using this class of receivers as part of gamma spectrometers, the main structural elements of the system were modeled, and the experimental model of the gamma spectrometer was prototyped. The linearity and energy resolution of the experimental sample were also investigated. To test the model for constructing a gamma spectrometer, a CsI (Tl) scintillation crystal and a Sensl Array-60035-4P photomultiplier were used. The established range of recorded energies is in the range from 121 keV to 1332 keV, the relative energy resolution for the 137Cs peak is 11.07 %, the linearity of the transfer characteristic is 99.91 %. Based on this sensor, the architecture of an automated wireless system for monitoring the spatial distribution of gamma radiation has been developed. The results of the work allow the use of radiation monitoring systems in accordance with the requirements of Industry 4.0.

We proposed a new approach to solving the problem of operational analysis and medium-term forecasting of the greenhouse gas generation (CO2, CH4) intensity in a certain area of the cryolithozone using data from a geographically distributed network of multimodal measuring stations. A network of measuring stations, capable of functioning autonomously for long periods of time, continuously generated a data flow of the CO2, CH4 concentration, soil moisture, and temperature, as well as a number of other parameters. These data, taking into account the type of soil, were used to build a spatially distributed dynamic model of greenhouse gas emission intensity of the permafrost area depending on the temperature and moisture of the soil. This article presented models for estimating and medium-term predicting ground greenhouse gases emission intensity, which are based on artificial intelligence methods. The results of the numerical simulations were also presented, which showed the adequacy of the proposed approach for predicting the intensity of greenhouse gas emissions.

BaFe12O19 films on Al2O3 (001) substrates were synthesized by ion beam deposition with post-annealing. Three deposition regimes were applied, resulting in unique surface morphology and X-ray diffraction patterns of obtained samples. In the first case, the deposition process was interrupted several times and the sample was subjected to ex situ annealing at 900 degrees C after each interruption. As a result, a film with high texture degree (about 97 %) was obtained. In the second case, the deposition process was interrupted the same number of times, but during each pause the film remained in the vacuum chamber for 10 min at 300 degrees C. This approach resulted in a medium texture degree of the sample (about 72 %). In the third case, there were no breaks in the deposition process and the texture of the resulting film was extremely poor. An explanation for the observed differences is proposed. To improve the texture of the films, a second annealing at 1280 degrees C for 10 h was applied, but as a result strong interdiffusion and misorientation of the grains occurred.

The current state of landfills is an important issue for monitoring. There are a lot of municipal solid waste landfills in the world. But there are no standard solutions for monitoring the current state of these facilities both in Russia and abroad. It makes developing of technical solution, based on Autonomous sensors for measuring the concentrations of hazardous fumes, radiation background, geotechnical factors and other environmental variables, actual. Such system can be easily installed and can work in stand-alone mode for a long time. Decision-making technology in such system can be based on using of modern achievements of Machine Learning.

We present a fully-staring THz video camera prototype intended for security screening. The camera utilizes so-called kinetic inductance bolometers to detect THz radiation in the bandwidth of 0.3-1 THz. The imaging distance is 2.5 m with the field-of-view being 2 m × 1 m. The camera is equipped with a kilo-pixel detector array, large field-of-view optics, intermediate-scale cryogenics operating at 6 K, and low-noise electronics to read out the whole detector array. The imaging capabilities of the system are demonstrated through radiometric performance characterization and actual imaging experiments.

Subject of Research. The paper presents a simple and practically effective solution for hyperparameter tuning in classification problem by machine learning methods. The proposed method is applicable for any hyperparameters of the real type with the values which lie within the known real parametric compact. Method. A random sample (trial network) of small power is generated within the parametric compact, and the efficiency of hyperparameter tuning is calculated for each element according to a special criterion. The efficiency is estimated by the value of a real scalar, which does not depend on the classification threshold. Thus, a regression sample is formed, the regressors of which are the random sets of hyperparameters from the parametric compact, and regression values are classification efficiency indicator values corresponding to these sets. The nonparametric approximation of this regression is constructed on the basis of the formed data set. At the next stage the minimum value of the constructed approximation is determined for the regression function on the parametric compact by the Nelder-Mead optimization method. The arguments of the minimum regression value appear to be an approximate solution to the problem. Main Results. Unlike traditional approaches, the proposed approach is based on non-parametric approximation of the regression function: a set of hyperparameters – classification efficiency index value. Particular attention is paid to the choice of the classification quality criterion. Due to the use of the mentioned type approximation, it is possible to study the performance indicator behavior out of the trial grid values (“between” its nodes). As it follows from the experiments carried out on various databases, the proposed approach provides a significant increase in the efficiency of hyperparameter tuning in comparison with the basic variants and at the same time maintains almost acceptable performance even for small values of the trial grid power. The novelty of the approach lies in the simultaneous use of non-parametric approximation for the regression function, which links the hyperparameter values with the corresponding values of the quality criterion, selection of the classification quality criterion, and search method for the global extremum of this function. Practical Relevance. The proposed algorithm for hyperparameters tuning can be used in any systems built on the principles of machine learning, for example, in process control systems, biometric systems and machine vision systems.

Subject of Research. The paper proposes a novel organization technique for preventive maintenance systems (including condition-based and predictive maintenance systems) based on the use of modern machine learning methods. The systems are operating using an original, non-parametric identification method for the current degradation phase of serviced equipment. Method. The proposed method comprises reducing the task of the current phase identification of the equipment degradation phase to interval estimation of the value of the so-called “health index” parameter of the equipment. This parameter is represented as a step function with the arguments in terms of a set of the measurable equipment objective parameters. The current equipment degradation phase is determined by classification approach. At this, based on the analysis of the observed data, it is decided upon what class (state phase) these data correspond to. Measurements from a group of sensors, in general, of various physical nature, which are located both on the surface and inside the equipment being monitored are used as data for identification of the equipment degradation stage. Mathematically, the proposed approach is reduced to a weighted combination of two classifiers. One of the classifiers of this combination is based on solving a group of binary classification problems. The second classifier is based on “Remaining Useful Life” parameter estimation by the method of nonparametric regression. Main Results. As distinguished from traditional approaches, the proposed approach uses a minimum of a priori information about the principles of operation and the internal structure of the equipment being serviced. The approach is based on the usage of the “health index” equipment parameter presented in the form of a step function. The novelty of the approach lies in the simultaneous use of the “health index” step function and the weighted combination of two classifiers with various structure. The proposed method showed good results when being tested on the C-MAPPS Dataset database, which contains data on failures of turbofan engines modeled using a thermodynamic simulation model. The pre-failure status of the equipment is identified with the probability of 99%. Practical Relevance. The obtained results and algorithms can be used in preventive maintenance systems aimed at reliable identification of the equipment degradation current stage.

This chapter provides some practical aspects and peculiarities of the use of Machine Learning based Predictive Maintenance for the infrastructure facilities in the cryolithozone. Some mathematical models of Machine Learning based Predictive Maintenance are described, which have shown their practical effectiveness. The solutions of several important problems of Predictive Maintenance for pipelines located in cryolithozone are considered, including: problem of leak detection from pipelines taking into account the possible damage to the pipeline foundation due melting of permafrost; problem of automatic classifying of defects that led to leaks; problem of prompt corrosion spot detection in the pipelines as well as problem of identifying the current state of the corrosion process in the pipeline. The problem of optimizing the procedure for incident tickets processing in the Predictive Maintenance system for oil pipelines was also considered.