Aim. Choice of method and its validation for the purpose of confirming the feasibility of safe operation of a switch mechanism housing submitted to rolling stock axle loads of 30 tf.
Method. As of today, there are neither regulations regarding the strength parameters of load-carrying elements of electric switch mechanisms, nor verification procedures. Given the above, when evaluating the strength of electric switch mechanism housing, the authors used the method used for the purpose of evaluating the strength of load-carrying elements of locomotives per GOST R 55513-2013. That method has a long history of application as part of calculation and testing of motive power and has shown good results. According to the method, strength is evaluated by comparing the safety factor of fatigue strength n with the allowed value for steel structures [n] = 2.0.
Result. The activities resulted in the validation of the proposed method of strength estimation by the criterion of fatigue strength. The value of fatigue endurance of electric switch mechanism housing was estimated based on the results of fatigue benchmark tests of three items. Each was stressed by stepping up the amplitude of the applied force after the base number of cycles had been reached. At the first stage, the loading was equivalent to the operational value. After the base number of functions had been reached for each item, the following step was initiated. The tests continued until cracks have been found. In the process of testing, the amplitude of stress was recorded at each level of loading. For the purpose of identifying the fatigue strength characteristics, the reduced fatigue strength of the switch mechanism housing was calculated under the hypothesis of linear addition of fatigue damage subject to the condition of deterministic loading and subsequent processing of the findings using statistical methods. The value of fatigue limit has been obtained for the housing of a tie-type electric switch mechanism equal to 48.4 MPa. The safety factor of fatigue strength was found to be equal to 2.86, which is above the minimal allowed value.
Conclusion. It has been shown that the housing of a tie-type electric switch mechanism has a sufficient strength as regards the operational static and dynamic loads. Its safe operation when exposed to axle loads of 30 tf has also been confirmed.

The Aim of the paper is to compare the UDP-2001SM installation for vibration diagnostics of bearings of freight, refrigerator and passenger car wheels and the OMSD-02 mechanism diagnostics system that since recently has been used in the Belarusian Railway.
Methods. By way of analysis of the UDP-2001SM vibration diagnostics installation and the OMSD-02 mechanism diagnostics system, the authors considered the list of identifiable defects, method for defining threshold values (reference levels) of vibroacoustic pulses and noises for the criteria of “faulty” and “good”, statistical data for rejected axleboxes of freight and passenger car wheelsets in respect to the overall number of conducted measurements, the experience of operation in various car shops of the Belarusian Railway.
Results. The paper describes the process of vibration diagnostics of wheelset bearings of freight and passenger cars using the case of stationary acquisition of vibroacoustic signals from acoustic plates that interact with the diagnosed axlebox. The authors present the diagrams of the UDP-2001SM installation and OMSD-02 system, as well as the defects of bearings that are to be identified using the above diagnostics equipment.
Conclusion. The paper showed the advantages and disadvantages of the above diagnostics systems based on the experience of their application in a number of car shops of the Belarusian Railway.

Aim. Improving the reliability of servo drives through partial redundancy of structural components.
Methods. The paper uses methods of harmonic analysis, technical system dependability theory.
Results. The mean times to failure of servo drives that differ in terms of the level of redundancy of structural elements (external connectors, feedback sensors, control boards, electric motors and load gears) were compared. It was shown that structural element redundancy along with the capability to switch to still-operable elements in the course of servo drive operation enable a triple increase of the failure tolerance. The authors suggested criteria for identifying the moment of time for decision-making regarding an electric servo’s structure modification before the moment of failure of the redundant element of such servo.
Conclusion. The method suggested in the paper allows formalizing the process of electric servo structure modification for the purpose of ensuring the required level of failure tolerance.

Aim. The paper aims to develop a process for treating the back side of silicon slabs and giving them a surface texture that enables good adhesion with printed metal with an even thickness and no mechanical stress, cracks or chipping. That is achieved by sandblasting slabs before contact material application. Silicon carbide (SiC) with particle size not exceeding 6 um is used as the abrasive material.
Methods. The proposed method of sandblasting improves chip bonding by eliminating surface stress, formation of microcracks and crystal grain within the chip. The research was conducted using a sandblaster and electronic microscope.
Results. It was shown that, after sandblasting, metal coating in the process of transistor collector area contact delineation improves the adhesion of the substrate semiconductor. Additionally, no mechanical stress was observes as compared to grinding, polishing and chemical treatment that are normally done before transistor contact encapsulation.
Conclusions. The process of thin metal film evaporation as part of semiconductor-based device manufacture affects their dependability. The treatment and preparation of the back side of the silicon slab with finished transistor configurations is one of the key operations as part of transistor manufacturing process that affects the adhesive properties of the surface before evaporation, output characteristics of the semiconductor-based devices, yield and cost.

Aim. Based on an analysis of Russian and foreign literary sources, the paper deduces an estimation of the effect of trust in aviation on the safety and reliability of flights. Trust is a multifunctional phenomenon that appeared in the Russian scientific psychology relatively recently (in the mid-1990s) and soon became an object of research in many branches of science. The phenomenon of trust in scientific psychology is a high-complexity category due to its highly dynamic nature and multifactorial dependence. Objects of trust might include individuals, social groups and organisations, as well as various material and non-material phenomena. The phenomenon of trust is conventionally subdivided into self-trust, trust in people (social) and trust in the world.
Methods. 49 specialised literary sources were analysed. Those include papers in science journals (25), papers in proceedings (8), monographs (15), abstracts of theses (1), out of which 37 are by Russian and 12 are by foreign authors.
Results. The development of trust was traced through the ontogeny, its effect on personality development throughout life was identified. Of great importance is the basic trust that burgeons in the infancy within the mother – child diad ant later defines the view of life that significantly affects the socialization and personality development, behaviour in personal life and professional activities. The phenomenon of trust (cognitive, emotional, behavioural aspects), its phenomenology and psychosocial functions were analysed. It was established, that the phenomenon of trust is of significant importance in many spheres of life and activities of individuals and organisations. It is manifested in the ethical, psychosocial, economic, legal and political functions. The authors examine the applicability of the identified properties of trust in aviation. It was identified that the level of trust in aviation systems (civilian and military) affects flight safety. Of special importance is the trust in pilot operations, as it affects his/her adaptive capability in cases of professional emergencies.
Conclusion. The paper substantiates the evaluation of the phenomenon of trust as a factor of reliability of an aviation system’s operation. At the same time, the effect of trust in the pilot on the adaptation and stress tolerance in emergency situations is identified. It is logical to consider personal trust of a pilot as a personal antistress resource. Consequently, the research should continue in order to identify the value of a pilot’ professional trust for the purpose of improving flight safety.

Aim. The paper aims to examine the matters related to increasing the objectivity of evaluation of the quality of train control by train drivers.
Methods. The study presented in the paper uses statistical analysis and linear algebra.
Results. An algorithm was developed for defining preventive measures and their application efficiency was evaluated for drivers of rapid transit trains. The algorithm for defining preventive measures for drivers of rapid transit trains includes the following: violation prediction; definition of the factors that affect the onset of each type of violations; definition of the characteristics of the drivers that most deviate from the target values. The efficiency estimation is based on the assumption of correlation between the cost of a driving instructor’s work with a driver and the cost of losses that the company might incur in case of violations. The paper shows that the level of an error of the first kind in the train driver violation prediction model is justified, provided that the cost incurred as the result of gross train control violations is significantly greater than that associated with the training of such driver. The paper presents an analysis of the application of the AI-based system in four depots.
Conclusion. The paper presents an algorithm for defining preventive measures for train drivers. An economic criterion was defined for evaluating the efficiency of application of the developed mathematical model for predicting gross violations of train control. The required and sufficient conditions of economic efficiency of the AI-based systems application were analysed. A comparative analysis was presented of the mean number of gross train driving violations in depots with and without the AI-based system.

Aim. The paper aims to analyse the specifics of the use of commercial fault tree (FT)-based software suites as part of engineering practice for the purpose of dependability calculation of emergency shutdown systems (ESS). Standards of the IEC 61508 Functional safety series stress that, in such cases, there is a possibility of incorrect and non-conservative estimates of the mean probability of failure on demand of an ESS safety feature. Incorrect results are primarily caused by the use of approximate and simplified formulas for identifying the dependability indicators of ESS circuit components and calculating the ESS mean unavailability for safety function performance based on the mean unavailability values of its components. In order to correct the FT simulation results, correction factors can be used that take into account the ESS circuit structure along with exact formulas per IEC 61508-6 for calculating the mean probability of failure on demand of the ESS circuit components. Additionally, the type of common cause failure (ССF) model can be chosen.
Methods. A comparative analysis was performed as regards the effects of components of hazardous failures that may be detected or not detected by internal diagnostics on the assessment of the mean probability of failure on demand of an ESS circuit components. It was shown that in less dependable components this dependence significantly affects the unavailability value. The efficiency of correction coefficients that take into account the ESS circuit architecture also depends on the dependability of components, and their introduction is justified for those components whose safety integrity level is between 1 and 2. Engineering estimation of the functional safety indicators can be done using a beta-factor model of common cause failures that is employed as part of design analysis of ESS functional safety.
Results. An analysis of simplified and approximate formulas for calculating the mean unavailability of the non-redundant elements of an ESS circuit has shown that in the case of an over 90-percent diagnostic coverage the use of simplified formulas causes an underestimation of the unavailability indicator caused by the increased effect of detected hazardous failures on the probability of ESS misoperation. If the FT analysis is used for the purpose of deducing a conservative estimate of an ESS circuit unavailability indicator, correction factors should be used, whose values depend on the ESS channels redundancy parameters. Two models of accounting for CCF were examined that are used when calculating ESS functional safety. It was shown that under any ESS model the system’s dependability indicators decrease. This decrease is defined by the value of the beta factor and the dependability of the ESS system elements.
Conclusion. The information presented in the paper indicates the limited applicability of the simplified formula for calculating the mean unavailability of non-redundant ESS elements as the input data for FT construction. When identifying the safety integrity level of an ESS circuit that includes elements with a low dependability, it should be taken into consideration that, if a FT is used, commercial software suites may overestimate the dependability, which is undesirable in respect to functional safety analysis.

Aim. Feature transformation is one of the stages of machine learning application that has a significant effect on the quality of regression models. The paper aims to develop criteria for evaluating the quality of data dimensionality reduction at the stage of feature transformation and adaptation of the UMAP method to the problem of prediction of the number of days to failure in the locomotives of JSC RZD.
Methods. The data transformation methods are divided into two groups, those that attempt to preserve the global data structure, and those that attempt to preserve the distances between points. The paper examines in detail the UMAP no-linear method of dimensionality reduction, whose low-dimensional data presentation is based on a transformation of a nearest neighbour graph retaining the data structure. The structure of the initial data manifold is examined using topological data analysis and simplified fuzzy set construction methods.
Results. The analysis of UMAP theory conducted in the Russian language for the first time enabled a substantiated identification of the three primary parameters of the method, whose variation significantly affects the type of data obtained as the result of a transformation. In particular, that pertains to the quality of class separation over a two-dimensional space. Additionally, the characteristics of the input set of parameters were identified that affect the UMAP results. Practical results of UMAP application were
demonstrated. Intermediate results included a list of nearest neighbours, a weighted graph of nearest neighbours. The fundamental result is a low-dimensional data representation (out of 44 initial measurements) over a two-dimensional space with class separation, which is confirmed both by calculations, and visually.
Conclusions. It was identified that UMAP is an efficient and substantiated method of dimensionality reduction that allows – through parameter variation – transforming data in such a way as to improve the quality of data submitted to machine learning models by the criterion of “evident class separation”. The transformation is an intermediate stage of data preparation for regression model application, and class separation was performed for the purpose of eliminating the probability of gross regression errors.