As a result of lasting operation of the undercarriages of traction rolling stock (frame and body, their pivot assemblies, bogie frames, axles of wheel pairs, etc.) under alternating cyclic loads, the strength properties of the metal parts degrade, their fatigue resistance declines, which may eventually cause the structure’s destruction. Therefore, the strength of the structures of locomotives and wagons is to be validated not only through safety margins as per the current regulations, but also through durability (life) predictions for such structures with predefined levels (dependability). Aim. To improve the conventional approaches to assessing and ensuring safe operation of rolling stock. Method. The life of structural elements is estimated using computations and experiments involving methods developed by the Mechanical Engineering Research Institute of the Russian Academy of Sciences (IMASH RAN) adapted to the structures of railway rolling stock given the existing experience of their operation and using the data collected by VNIKTI. Results. Calculations of the durability and life of critical items, definition of the acceptable level of risk of locomotive operation. Conclusion. A risk-based approach helps improve railway traffic safety. 

Passenger cars are complex technical products. They consist of units, assemblies, and components that are characterized by a certain combination of interacting parts. Additionally, modern passenger cars feature significant numbers of automatic subsystems and automated components: air conditioning, electric heating systems, lighting systems, compartment doors, exterior doors, etc. The process of collecting data on the technical condition of products is to ensure the regularity, reliability, timeliness, and completeness of information. It is known that products most clearly manifest their quality and dependability in operation. A competent organisation of the collection and processing of information on a product’s dependability allows obtaining reliable information on its health and performance. In the course of operation, the connections between individual units and components of passenger cars may become disrupted, the fasteners of individual parts and sensors may become loose, rubber seals may become naturally worn. All of that causes performance decline, as well as malfunctions and failures. Preventing a sharp increase in the number of failures requires performing a number of preventive actions aimed at identifying and eliminating faults, as well as preventing their root causes. First and foremost, such measures include rolling stock maintenance and overhaul. All such activities are strictly regulated in the operating manuals of both a car and its components. The specified life of passenger cars, as well as their components and units, varies roughly from 20 to 40 years. Some components of passenger cars have been in production with no major modifications since the early 2000s. That suggests that a product’s dependability can be evaluated comprehensively throughout the entire life cycle. But that can only be done by collecting and processing a significant amount of information on malfunctions obtained both during the warranty and post-warranty periods. The information is to come from various sources, i.e., operating companies, service depots, car repair plants that carry out overhauls. This most valuable information is to be accumulated and be digitalisable. This paper addresses a number of matters associated with the collection, validation, and recording of faults and failures of passenger car components. Aim. To examine the state-of-the-art systems that collect and process fault data in engineering companies and to suggest algorithmic and methodological solutions to improve the degree of automation of failure information processing. Methods. The paper uses methods of system analysis and software engineering. Conclusions. An algorithm for recording product failures according to incoming documents is proposed. Software solutions have been developed to automate the process of collecting and processing data on malfunctions of passenger car components. The authors examined a method of tracking the warranty fleet required for defining the total operating time as part of calculating the dependability indicators of passenger car components in operation. A failure code list was proposed that takes into account the specificity of the structural relationships between passenger car components.

Aim. This paper examines the construction of an intelligent system for analysing and classifying pseudorandom number generators (PRNGs) that combines the capabilities of machine learning and directed search for determining the type of the source of a random sequence of numbers. The focus is on identifying weaknesses in non-cryptographic PRNGs that may be predictable, which entails risks for their use in information security. Methods. The research used machine learning methods, including neural networks, correlation analysis, and NIST statistical tests. The developed models were trained on large samples of PRNG output strings, which allowed estimating the predictability of the PRNG and internal state restorability. Neural network structures were chosen taking into account the results of optimisation of the neural network hyperparameter values. The paper shows the effect of the sample size on the obtained results. Results. The analysis and classification of a PRNG involves a number of steps: calculating the autocorrelation function of the output strings and their spectrum; execution of statistical tests developed by the NIST laboratory; classification of PRNGs based on the output strings analysis; identifying the specificity of the PRNG’s internal structure or its internal states; prediction of the output values. For the Xorshift128 algorithm, the neural network showed a high accuracy of output value restoration, which confirms its vulnerability. An analysis of the Mersenne Twister algorithm revealed certain patterns, but required more complex architectures to completely reconstruct the strings. Using machine learning algorithms, the authors managed to identify the structure building patterns for the “stop-and-go” algorithm, but failed to highly accurately predict the PRNG output value based only on the prior output string values with no knowledge of the internal states. Directed search algorithms allow classifying and predicting a linear congruentional generator and a Geffe generator. The models combined into a system classify PRNGs according to their characteristics and predict their eventual output values. An analysis of the obtained results confirms the significance of not only the selected PRNG structure, but also the numerical parameters and the bits within numbers involved in the computation. Conclusion. The conducted study confirms the efficiency of the combination of machine learning and directed search as part of the analysis and classification of PRNGs. The findings allow recommending the developed system for use in practical PRNG safety assessment. Further research will focus on expanding the set of analysed PRNGs and examining other types of neural networks for improving the quality and performance of models.

Aim. To substantiate the semantic content of the concept of energy system dependability. To assess the applicability of such terms as short-term, long-term, overall and cross-mode dependability, as well as adequacy as part of collections of recommended terms and reference books. To scientifically substantiate the revised content of Energy system dependability. Collection of recommended terms published in 2007. Methods. The paper employs logical procedures that consist in giving an unchanging meaning to such a concept as energy system dependability. Results. Major accidents that have occurred in energy systems of various countries over the past 30 to 40 years indicate that the problem of safety deserves closer attention. However, the term “safety” is not featured in Energy system dependability. Collection of recommended terms published in 2007. The terminology states that “the exclusion of safety from the properties of energy facility dependability was primarily motivated by the fact that it had been in no demand for 27 years”. There is no theoretical substantiation for excluding the term “safety” from the terminology handbook. How can we talk about no demand for the property of safety given the Chernobyl nuclear power plant disaster on April 26, 1986 and the Sayano-Shushenskaya HPP accident of August 17, 2009 that claimed the lives of 75 people? The 2007 terminology notes that its development was motivated by “... changing conditions surrounding the operation and development of energy systems (liberalisation of the economy, the emergence of energy markets)...” Dependability is a property (a characteristic that constitutes a distinctive feature) of energy systems, therefore the liberalisation of the economy cannot affect the content of the concept of “dependability”. It may only affect the required level of dependability, the decision-making criteria and dependability models. To simplify understanding, experts (involved in the study of energy system dependability) have come up with a number of terms that are a set of simplified (abbreviated or figurative) words, e.g., cross-mode dependability, short-term dependability, long-term dependability, overall dependability, etc., that represent an abstract idea of energy system dependability that is reflected – with certain assumptions – in the respective models of dependability research. There is no need to integrate such terms in collections of recommended terms or reference books (since the list of specified functions of energy systems does not change, therefore the property remains unchanged). Conclusions. Energy system dependability is a complex multidimensional property both in terms of public regulation of the economy, and under market conditions. The content of the concept of energy system dependability does not depend on the form of ownership, does not depend on an entity’s change of ownership, does not depend on advance decisions associated with ensuring dependability. 

Aim. To study the process of operation of the existing level crossing devices. To examine a detailed barrier operation algorithm that takes into account the proper operation of each component. The existing level crossing protection systems use simplified operation algorithms. To focus on the process of crossing warning generation when a train enters the warning section. To examine in detail the work of operators who identify the distance to a vehicle and, consequently, the barrier activation time. That will allow adjusting the barrier activation time and, consequently, reduce the rolling stock downtime. Methods. The paper suggests a method that involves continuously measuring the vibration displacements of the rails at the crossing using an accelerometer. By continuously measuring the values of vibration velocity (V) and vibration acceleration (a) at the boundary of a crossing, a set of measured values Vi and ai is generated. These values are used for making a system of equations for the train location coordinate, whose right part is equated to the values of the train coordinates. By solving the system of train coordinate equations using the values of vibration velocity (V) and vibration acceleration (a) at the boundary of a crossing at the approach of a train, the coordinate of the vehicle can be defined at specific points in time. Consequently, the speed and nature of its movement can be identified throughout the warning section. Out of the obtained values, the time of the train’s head clearing the crossing can be identified. By comparing the estimated barrier activation time with the time derived from the nature of a train’s movement, the moment of crossing barrier activation can be identified. Results. The study showed that the existing level crossing warning systems deployed within stations normally activate barriers in advance. That may cause motor vehicles to spend excessive amounts of time before a crossing. If the length of the approach section is fixed, the actual warning time is inversely proportional to the train’s speed and may significantly exceed the minimum required time. Excessive warning time may have negative consequences, which requires a solution. If a crossing is regularly closed for long periods of time, drivers will be attempting to cross faster once the restrictive signal has turned on, which may lead to accidents, train-to-motor vehicle collisions. Should data on the actual time of a train’s arrival to a crossing be available, the time spent by motor vehicles at closed crossings will be able to be reduced down to the estimated closing time and the time it takes for the train to clear the crossing. The average train speed in the Russian Federation is 35.7 km/h [1], while warnings are calculated for the top speed. Conclusion. The time spent by vehicles before a barrier can be reduced about 3 to 4 times. Consequently, the number of accidents at crossings can be reduced as well.

The paper examines concepts associated with ensuring comprehensive security of the control area of upper-level automated railway control and management systems and the primary regulatory requirements for information protection. It is shown that the presented concepts do not contradict the requirements of the FSTEC of Russia.

Aim. The application of novel technologies in intelligent water transportation systems (IWTS) is associated with additional security risks that are due to the emergence of new types of threats. The automated corporate and process management systems that are part of IWTS are critical information infrastructure (CII) facilities. That imposes increased safety requirements on IWTSs. Hardware and software systems that implement such solutions are undergoing active development. In many cases, physical prototyping of IWTS facilities within reasonable periods of time is difficult and economically unviable. Modern simulation methods efficiently solve the above problems. They allow creating digital prototypes of IWTSs and IWTSs proper within secure virtual environments. That represents the subject matter of this paper. Methods. The paper uses system analysis, operations research, simulation, and IWTS security. Results. The authors examine the evolution of simulation and provide subject-matter terminology. The paper defines standard CII facilities as part of IWTS and facilities to be digitally simulated. It analyses the tools that contribute to the creation of a digital testbench for analysing the IWTS CII security. It provides a description of the digital testbench for analysing IWTS CII security, as well as examples of its operation. Conclusion. The digital testbench presented in the paper allows incorporating both existing Russian secure software and hardware systems, and those under development. It also enables IWTS security risk management. That allows using the testbench at various lifecycle stages of IWTS CII facilities. Further development of the testbench is associated with the development of digital twins of inland waterways, Russian-made software and hardware systems of IWTS CII facilities, IWTS CII computer attack protection tools and methods for hybrid security management.

Aim. To examine the applicability of the breadth-first path searching algorithm for spatial development of linear land transportation infrastructure facilities. Methods. The paper uses Breadth-First Searching, a graph path searching algorithm that is widely used as part of various graph theory applications, including path tracing and path planning. A number of simple experiments were carried out with this algorithm in order to determine the quantitative indicators of its asymptotic complexity, i.e., the number of performed operations and the algorithm execution time. The series of experiments has a different structure that is defined by the search direction (unidirectional and bidirectional) and the method of cell scanning (direct and mixed). Conclusion. Experiments involving various implementations of the algorithm show that bidirectional search can significantly reduce the number of performed operations and the search time. Thus, the number of operations for bidirectional search is 2.75 times less for direct and 2.78 times less for mixed (direct and diagonal) cell scanning. Moreover, it is concluded that the bidirectional implementation of the algorithm has its own scope of efficient use. First, bidirectional search is effective in highly-branched graphs. The number of operations for bidirectional maze search decreases 57.07%, while the time of the same experiment decreases 76.92% as compared to the unidirectional search. In a corridor environment that, by definition, has weak branching, the difference in the number of performed operations between bidirectional and unidirectional search was 1.06%, while the execution time remained the same. Secondly, the efficiency of the algorithm is significantly reduced when the graph structure is complex. Thirdly, using this implementation requires confidence in the fact that a path between the starting and target nodes exists.