Nonlinear Wheel Rail Research Articles

Dynamic modelling and experimental validation of a dynamic track stabiliser vehicle–track spatially coupling dynamics system

A dynamic track stabiliser vehicle is an unconventional vehicle for maintaining the quality of ballasted tracks, and its dynamic performance has a significant impact on its working efficiency. For the first time, this paper proposes a comprehensive dynamic track stabiliser vehicle–track coupled vertical and lateral multi-body dynamics model based on classical vehicle–track coupling dynamics theory. In this model, detailed attention is given to the vertical, lateral, roll, yaw, and pitch motions of the components, including the vehicle body, bogies, wheelsets and stabilisers. The Shen–Hedrick–Elkins theory is employed to describe non-linear wheel–rail contact relationships between the bogie wheelsets and the rails, and three various methodologies are adopted to address complicated wheel–rail interaction problems between the stabiliser wheelsets and the rails. This comprehensive multi-body dynamics model enables a detailed investigation of the dynamic performance of the system under complex excitations, such as the lateral excitation of the stabilisers and vertical downward pressure exerted by hydraulic cylinders, especially the dynamic response of the stabilisers in the vibration environment of the complete vehicle. The dynamics model was fully validated by comparing its results with time- and frequency-domain data obtained from field tests. The results indicate that the model is capable of being used for analysis of the dynamic performance of dynamic track stabiliser vehicles.

Running safety control performance of high-speed trains on bridges under seismic excitation utilizing tuned mass damper

The safety of high-speed trains on bridges during seismic conditions is of paramount concern, especially for high-pier bridges characterized by lower lateral stiffness. The risk of train derailment during earthquakes can increase due to heightened low-frequency excitation. While Tuned Mass Dampers (TMDs) have proven effective in mitigating structural responses, their influence on time-varying vehicle-bridge interactions under earthquakes remains an unexplored area of study. In this research, a 1:10 scale model of a high-speed train-track-bridge coupling system, incorporating a nonlinear wheel-rail interface, was subjected to testing using a four-shaking-table system. This system exhibited good control accuracy and synchronization rates. The experimental results were used to validate the accuracy of the numerical model for the train-track-bridge coupled system under various train speeds. Subsequently, the seismic control performance of TMDs on high-speed trains was assessed through a combination of experimental tests and numerical simulations. Both sets of results, from the experimental and numerical approaches, confirmed the effectiveness of TMDs in controlling acceleration and force-related parameters for both moving and stationary trains. Notably, the control performance is most pronounced when vibrations closely align with the TMD's operating frequency. Additionally, this study delved into the mechanism of train vibration control and conducted a comprehensive parametric investigation, considering factors like pier heights (24 m–50 m), train speeds (0–300 km/h), ground motion intensities (0.04 g–0.21 g), and TMD mass ratios (0.03–0.08). The findings revealed that TMDs have a more significant impact on enhancing the safety index of trains for high-pier bridges characterized by lower frequencies. For instance, in the case of a 50 m pier, the running train's derailment coefficient can be reduced by 34.94 %. However, the control action is narrowband, effectively mitigating train vibrations near the bridge's natural frequency but less effective outside the working frequency range and absolute base acceleration.

Research on the Identification of Nonlinear Wheel–Rail Adhesion Characteristics Model Parameters in Electric Traction System Based on the Improved TLBO Algorithm

The wheel–rail adhesion is one of the key factors limiting the traction performance of railway vehicles. To meet the adhesion optimization needs and rapidly obtain wheel–rail creep characteristics under specific operating conditions, an engineering identification method for wheel–rail adhesion characteristics based on a nonlinear model is proposed. The proposed method, built upon the traditional Teaching-Learning-Based Optimization (TLBO) algorithm, has been adapted to the specific nature of nonlinear wheel–rail adhesion model parameters identification, enhancing both the search speed in the early stages and the search accuracy in the later stages of the algorithm. The proposed identification algorithm is validated using experimental data from the South African 22E dual-flow locomotive. The validation results demonstrate that the proposed identification algorithm can obtain a nonlinear wheel–rail adhesion characteristics model with an average adhesion coefficient error of around 0.01 within 50 iteration steps. These validation results indicate promising prospects for the engineering practice of the proposed algorithm.

Influence of unsupported sleepers on flange climb derailment of two freight wagons

Abstract Two freight platform wagons connected together conventionally or in the form of so-called rigid connection were taken into account to investigate the influence of unsupported sleepers on climb flange derailment on transition curve. Presented results are related to the freight wagons equipped with two-axle Y25 standard bogies, entering at an acceptable speed in the transition curve of the track with the radius of 150 or 1,000 m. Three cases have been analyzed: only the inner rail, only the outer rail, or both rails had unsupported sleepers. Simulation models were prepared using the VI-Rail software and nonlinear wheel–rail contact model. The dynamic calculations showed that wheel flange climb derailment caused by unsupported sleepers will only occur if deflection of the rail or track on transition curve is deep enough. These tests allowed us to determine the relationship between the radius of curvature and the depth of the trough causing derailment of the wagons. Calculation results allowed us to answer also the questions whether the method of connection and the weight of freight wagons have a significant impact on the wheel flange climb derailment.

An efficient method for predicting wheel-rail forces in coupled nonlinear train-track-bridge system using artificial neural networks

Wheel-rail forces are an essential indicator for vehicle safety evaluation. The calculation of wheel-rail forces for the coupled nonlinear train-track-bridge system using the direct numerical integration method is time-consuming, hindering the timely safety and reliability assessment of train operations. In this paper, an efficient method based on the nonlinear autoregressive with exogenous input artificial neural network (NARX-ANN) was proposed to predict the nonlinear wheel-rail force of trains on bridges caused by track irregularity. Firstly, samples of track irregularity time history were randomly generated using the stochastic harmonic function method. The vertical and lateral wheel-rail forces of the coupled nonlinear train-track-bridge system were calculated by numerical integration. Secondly, the NARX-ANN model was established and trained with a small number of randomly selected wheel-rail force samples. Finally, the wheel-rail forces obtained by numerical integration were regarded as the ground truth to verify the prediction accuracy of the NARX-ANN model. The influence of training configuration, initial output, and the number of time delays on prediction accuracy was systematically analyzed. The results showed that the NARX-ANN model could accurately predict the time history of wheel-rail forces. In addition, its computational efficiency is 19.35 times higher than that of the numerical integration method. It is hoped that this study can guide the stochastic analysis of the train–track–bridge system.

NONLINEAR DYNAMIC RESPONSES OF LOCOMOTIVE EXCITED BY RAIL CORRUGATION AND GEAR TIME-VARYING MESH STIFFNESS

Rail corrugation is usually generated in modern railway transportations, such as high-speed railway, urban railway, and heavy-haul railway. It is one of the major excitations to the wheel–rail dynamic interaction, which will cause extra vibration and noise, failures, or even risk of derailment to the vehicle and its components. A dynamics model of a heavy-haul locomotive considering the traction power from the electric motor to the wheelset through gear transmission is employed to investigate the nonlinear dynamic responses of the locomotive. This dynamics model couples the motions of the vehicle, the track, and the gear transmission together. In this dynamics model, excitations from the rail corrugation, the nonlinear wheel–rail contact, the time-varying mesh stiffness, and the nonlinear gear backlash are considered. Then, numerical simulations are performed to reveal the dynamic responses of the locomotive. The calculated results indicate that different nonlinear phenomenon can be observed under the excitation of the rail corrugation with different amplitude and wavelength. The high frequency vibrations excited by the time-varying mesh stiffness are usually modulated by the low frequency vibrations caused by the rail corrugation. However, this is likely to vanish under the chaotic conditions with some corrugation wavelength. The vibration level of the vehicle and the gear transmission increases generally with the corrugation amplitude. However, some corrugation lengths have been found to be more responsible for the vibration of the dynamics system, which should be concerned greatly during the locomotive operation. Meanwhile, involvement of gear transmission systems will cause different dynamic responses between the wheelsets under rail corrugation and gear mesh excitations.

Random dynamic analysis on a high-speed train moving over a long-span cable-stayed bridge

ABSTRACT This paper presents random dynamic analysis on a high-speed train moving over a long-span cable-stayed bridge. In the study, the train, slab track, and bridge are regarded as an integrated system, each vehicle is modelled as a four-wheelset mass-spring-damper system with 10 degrees of freedom (DOFs), and the rail and track slabs are modelled as Euler-Bernoulli beams supported by spring-damper elements. The long-span cable-stayed bridge model is established by the ANSYS software, based on which its frequencies and mode shapes are exported and implemented into the dynamics programme executed in the MATLAB platform. The train and substructures are coupled via nonlinear wheel–rail interaction and the large-scale dynamics model is solved by means of a fast explicit integration algorithm. Several numerical examples are performed involving modal convergence study, the effect of train speeds on system dynamic responses and stochastic analysis based on the probability density evolution method.

Bifurcation analysis of a railway wheelset with nonlinear wheel–rail contact

Stability is a key factor for the operation safety of railway vehicles, while current work employs linearized and simplified wheel/rail contact to study the bifurcation mechanism and assess the stability. To study the stability and bifurcation characters under real nonlinear wheel/rail contact, a fully parameterized nonlinear railway vehicle wheelset model is built. In modeling, the geometry nonlinearities of wheel and rail profiles come from field measurements, including the rolling radius, contact angle, and curvatures, etc. Firstly, four flange force models and their effects on the stability bifurcations are compared. It shows that an exponent fitting is more proper than a quintic polynomial one to simulate the flange and works well without changing the Hopf bifurcation type. Then the effects of each term of the nonlinear geometry of wheel/rail contact on the Hopf bifurcation and limit circle bifurcation are discussed. Both the linear term and nonlinear term of rolling radius have a significant influence on Hopf bifurcation and limit point of circle (LPC) bifurcation. The linear critical speed (Hopf bifurcation point) and the nonlinear critical speed (LPC bifurcation point) changes times while within the calculated range of the linear term of the rolling radius. Its nonlinear term changes the bifurcation type and the nonlinear critical speed almost by half. The linear term of the contact angle and the radius of curvature of the wheel and rail profiles should be taken into consideration since they can change both the bifurcation point and type, while the cubic term can be ignored. Furtherly, the field measured wheel profiles for several running mileages are employed to examine the real geometry nonlinearities and the corresponding Hopf bifurcation behavior. The result shows that a larger suspension stiffness would increase the running stability under wheel wear.

Wheel–rail impact loads and axle bending stress simulated for generic distributions and shapes of discrete wheel tread damage

Wheel–rail impact loads generated by discrete wheel tread irregularities may result in high dynamic bending stresses in the wheelset axle, leading to a decrease in component life and an elevated risk for fatigue failure. In this paper, a versatile and cost-efficient method to simulate the vertical dynamic interaction between a wheelset and railway track, accounting for generic distributions and shapes of wheel tread damage, is presented. The wheelset (comprising two wheels, axle and any attached equipment for braking and power transmission) and track with two discretely supported rails are described by three-dimensional finite element (FE) models. The coupling between the two wheel‒rail contacts (one on each wheel) via the wheelset axle and via the sleepers is considered. The simulation of dynamic vehicle–track interaction is carried out in the time domain using a convolution integral approach, while the non-linear wheel–rail normal contact is solved using Kalker's variational method. Wheelset designs that are non-symmetric with respect to the centre of the axle, track support conditions that are non-symmetric with respect to the centre of the track, as well as non-symmetric distributions of tread damage on the two wheels (or irregularities on the two rails) can be studied. Time-variant stresses are computed for the locations in the wheelset axle which are prone to fatigue. Based on Green's functions for stress established using the wheelset FE model, this is achieved in a post-processing step. An extensive parametric study has been performed where wheel–rail impact loads and axle stresses have been computed for different distributions and sizes of tread damage as well as for different train speeds.

Numerical simulation platform for slab track systems subjected to a moving vehicle

With the rapid development of high-speed and urban rail transit, ballastless slab tracks have been widely used in many areas. The work on construction, design, and mechanics of ballastless slab tracks is abundant and achievements have been made. However, a platform for numerical simulation and dynamic analysis of slab track systems subjected to a moving train still needs to be probed into. In the MATLAB®, effective and easily programable methods are presented to guide the construction of the vehicle-slab track interaction system and the realization of long length computation in this present study. Based on previous studies, the formulation of slab track dynamics in the framework of finite element theory is founded with simplicities, and then the wheel-rail coupling matrices totally derived from the nonlinear wheel-rail normal contact/tangential creep forces using the energy variation method are presented. Following the dynamics modeling construction, an improved method for long or infinite length computation is proposed. With an implementation of this method, the effect of the spatial variability of track properties on vehicle-track system dynamic performance can be evaluated conveniently. Using this numerical simulation platform, extensive studies have been conducted to model and analyze the influence of spatial variability of filling layer support stiffness, the track slab finite element type and support type of steel spring for floating-slab tracks, etc.

Stochastic dynamic analysis of vehicle-bridge coupled systems with nonlinear Hertz contacts by explicit time-domain method

ABSTRACT Vibration analysis of vehicle-bridge coupled systems has received considerable attentions over the past two decades. This work is dedicated to the stochastic dynamic analysis of vehicle-bridge coupled systems with nonlinear Hertz wheel-rail contacts subjected to random track irregularities. An explicit time-domain method (ETDM) is proposed for efficient nonlinear dynamic analysis of vehicle-bridge coupled systems with a dimension-reduced iteration scheme focusing on the nonlinear wheel-rail contact forces. The high-efficient ETDM is further utilised to implement the numerous nonlinear dynamic analyses required in the Monte-Carlo simulation (MCS) for obtaining the response statistics of the vehicle-bridge coupled systems. A numerical example involving a single-wheel vehicle traversing a simply-supported beam with nonlinear Hertz wheel-rail contacts considering random track irregularities is presented to demonstrate the computational accuracy and efficiency of the proposed method, and an engineering application involving a four-wheel vehicle moving across the Egongyan suspension bridge is further analysed to indicate the effectiveness of the proposed method for large-scale engineering problems.

Estimating the wheel lateral position of a mechatronic railway running gear with nonlinear wheel–rail geometry

The “Next Generation Train” (NGT) is a double-deck high speed train concept in light-weight design that has been established as a technical and project framework in which the German Aerospace Center gathers its long-term railway vehicle research. To reduce wheel and rail wear and to enhance the passenger capacity, a mechatronic running gear with independently rotating wheels (IRW) is a key vehicle component under research. This running gear requires an advanced control of the lateral dynamics in order to fully exploit its potential in minimizing wear and noise, but this relies on information on its lateral position relative to the track. In daily railway operation it is difficult to directly measure this displacement of the wheel-pair. However, according to previous work it is possible to design an appropriate observer to estimate the lateral position, but its estimation accuracy highly depends on the chosen sensor configuration. Nevertheless, it was shown that the lateral dynamics of the running were not completely reproduced by the system description especially at suddenly changing directions in the movement of the IRW. In this work the system description is complemented by a nonlinear description of the wheel–rail geometry. Therefore, an observability analysis and an observer synthesis to estimate the lateral position for the nonlinear system of the 1:5 scaled hardware running gear is carried out. For validation purposes an extended Kalman filter with the U-D formalism is implemented at the real-time environment of the testbed and the estimation accuracy of the observer configurations is compared.

Bifurcation of a modified railway wheelset model with nonlinear equivalent conicity and wheel–rail force

This paper presents the bifurcation behaviors of a modified railway wheelset model to explore its instability mechanisms of hunting motion. Equivalent conicity data measured from China high-speed railway vehicle are used to modify the wheelset model. Firstly, the relationships between longitudinal stiffness, lateral stiffness, equivalent conicity and critical speed are taken into account by calculating the real parts of the eigenvalues of the Jacobian matrix and Hurwitz criterion for the corresponding linear model. Secondly, measured equivalent conicity data are fitted by a nonlinear function of the lateral displacement rather than are considered as a constant as usual. Nonlinear wheel–rail force function is used to describe the wheel–rail contact force. Based on these modifications, a modified railway wheelset model with nonlinear equivalent conicity and wheel–rail force is set up, and then, some instability mechanisms of China high-speed train vehicle are investigated based on Hopf bifurcation, fold (limit point) bifurcation of cycles, cusp bifurcation of cycles, Neimark–Sacker bifurcation of cycles and 1:1 resonance. In particular, fold bifurcation of cycles can produce a vast effect on the hunting motion of the modified wheelset model. One of the main reasons leading to hunting motion is due to the fold bifurcation structure of cycles, in which stable limit cycles and unstable limit cycles may coincide, and multiple nested limit cycles appear on a side of fold bifurcation curve of cycles. Unstable hunting motion mainly depends on the coexistence of equilibria and limit cycles and their positions; if the most outward limit cycle is stable, then the motion of high-speed vehicle should be safe in a reasonable range. Otherwise, if the initial values are chosen near the most outward unstable limit cycle or the system is perturbed by noises, the high-speed vehicle will take place unstable hunting motion and even lead to serious train derailment events. Therefore, in order to control hunting motions, it may be the easiest way in theory to guarantee the coexistence of the inner stable equilibrium and the most outward stable limit cycle in a wheelset system.

Dynamic response of vehicle–bridge–soil system using lumped-parameter models for structure–soil interaction

Prediction of vibrations generated by railway traffic experiences an increasing interest, as new lines are being constructed and planned in many countries. The paper proposes a numerical model to analyse a coupled vehicle–bridge–soil system, taking into account the most important phenomena affecting the structure, while at the same time being comparatively computationally efficient. Such model is useful in the early design phases of a project, when analysing a range of possible configurations or conducting parametric analysis is required.A simplified vehicle model, nonlinear wheel–rail interaction, a bridge structure modelled using the finite-element method and a semi-analytical model for layered soil are all introduced in the model. A pure time-domain solution procedure is used, utilizing lumped-parameter models (LPMs) for the soil–foundation system. Representation of the dynamic stiffness matrix using LPMs is investigated by analysing the cross-coupling between footings and a novel procedure necessary for computationally stable LPMs is introduced and utilized. Further, pure time-domain solution is compared with an iterative mixed-domain solution. Finally, to illustrate the capabilities of the model, analyses are carried out to determine the resulting maximum and minimum excitation limits resulting from wheel–rail interaction on a number of randomly generated uneven track profiles.

Influence of out-of-round wheels on the vehicle–flexible track interaction at rail welds

The dynamic analysis of the vehicle–flexible track interaction involves the study of vehicle motion and its dynamic impact transmitted to the track structure. This paper studies the influence of the out-of-round vehicle wheels running over rail welds on a flexible ballast track. The rails are modeled as an Euler-Bernoulli beam discretely supported by a spring-damper force element that represents the flexibility of the track structure. The dynamic behavior of the vehicle–flexible track interaction is studied using the combination of the finite element method and the multi-body system. In this paper, the simulation of the vehicle with the out-of-roundness wheel running over rail welds on a flexible ballast track in the high-frequency range and the vehicle–track interaction is coupled by a non-linear wheel–rail contact model. The effects of the out-of-roundness wheel on the vehicle–flexible track interaction at rail welds are investigated by comparing the effects of the round wheel under different vehicle speeds. Results indicate that the out-of-roundness wheel at rail welds creates a high magnitude dynamic effect on the vehicle and track components. The obtained simulation results were used to set a safety limitation for the wheel and rail irregularity size.

Uncertainty and sensitivity analysis for train-ballasted track–bridge system

ABSTRACTThis paper is devoted to present an integrated uncertainty and sensitivity procedure for use in performance assessments for train–bridge interactions. Firstly, a train-ballasted track–bridge dynamic model with proper consideration of the nonlinear wheel–rail contact geometries is constructed in three-dimensional space; and then the uncertainties related to this system are illustrated, in which the material, geometry, structural damping of bridge superstructure and the temporal–spatial ergodic properties of track irregularities are highlighted; and furthermore, the probability density evolution method accompanied with an advanced number theoretical method is introduced to describe the transmission of probability from excitation input to response output for the uncertainty analysis; finally, the key factors affecting the dynamic behaviours of system components are clarified by the variance-based global sensitivity analysis which has taken the interactions between factors into account. Numerical examples show that the uncertainty analysis algorithm is of high efficiency and accuracy by comparing with the robust Monte-Carlo simulation method; the variability of bridge parameters and track irregularities show remarkable effects on the dynamic indices; and the most sensitive factors are rather different with regard to varied dynamic indices.

Investigation of coupled dynamics of a railway tank car and liquid cargo subject to a switch-passing maneuver

The movement of the liquid cargo within a partly filled tank car is known to impose additional slosh forces and moments that may adversely affect the dynamic responses of the vehicle. This study is aimed at analyzing the liquid cargo slosh in a partly filled tank car and its effects on vehicle responses during a switch-passing maneuver. A two-dimensional analytical liquid slosh model is formulated for the analyses of the liquid load shift in the roll plane, lateral slosh force, and roll moment through summation of first four antisymmetric modes of the liquid. The analytical slosh model is integrated to a 114 degrees-of-freedom multibody dynamic model of the railway tank car comprising nonlinear wheel–rail contact and contact pairs of the suspension system. The validity of the slosh model is illustrated by comparing the responses with those reported in other studies and those obtained from a nonlinear computational fluid dynamic model. The coupled fluid–vehicle model is subsequently used to study the effects of fluid slosh during switch-passing maneuvers on different response measures, namely roll motion of the tank car, lateral and vertical wheel–rail contact forces, and derailment ratio. The significance of the liquid cargo slosh in the partially filled state is demonstrated by comparing the responses with those of the car with equivalent rigid cargo. The results show that liquid sloshing within the partly filled car can lead to higher magnitudes of car body roll angle and thereby the unloading ratio compared to the conventional rigid cargo car. Switch-passing critical speeds are further identified for different fill ratios and switch geometries. For fill ratios below 80%, the switch-passing critical speeds of the partly filled car are substantially lower compared to those of the equivalent rigid cargo car. Neglecting the contributions due to dynamic slosh force and roll moment arising from a partially filled railway tank car may thus lead to underestimation of the critical speed in switch-passing maneuvers.

Observer Synthesis for the Adhesion Estimation of a Railway Running Gear

The presented work illustrates an investigation on the observer synthesis for a highly nonlinear railway running gear system. The specific focus is on the longitudinal dynamics and the associated adhesion effects, i.e. the friction conditions between wheel and rail. This interface strongly influences the brake and the traction performance of trains and, thus, its detailed consideration is of crucial importance in terms of safety and comfort.The observer design is accomplished for the experimental running gear developed at the German Aerospace Center (DLR). The article focuses on the first steps of the observer synthesis: (i) the set-up of the observer model, (ii) the selection of an appropriate observer method, (iii) and the validation of the observer in a simulation environment. The determining aspect of the model implementation is the consideration of the nonlinear wheel-rail contact. To cover this nonlinear behavior an Extended Kalman Filter (EKF) is used in combination with a parameter estimator for the friction characteristics in the wheel-rail interfaces. In the end, the received results prove that the observer accurately estimates the system behavior and provides reliable information on the adhesion characteristics and related longitudinal dynamics.Based on this findings one of the upcoming steps at the DLR is to validate the observer in the real-time environment of the test rig. Regarding the use of the presented approach in a train some significant improvements of traction control systems are enabled. A first application could be an enhanced brake control to narrow down the variance of brake distances in daily operation. Furthermore, an advanced usability of the condition based monitoring of traction and brake systems is facilitated with the additional information on the friction conditions.

Development of a coupled numerical model for the interaction between transient fluid slosh and tank wagon vibration

In this paper, a coupled model is developed to evaluate the effect of transient fluid slosh on the railway tank wagon dynamic vice versa. This model has computational complexity in solving the Navier–Stokes equations and nonlinear differential equations of tank wagon vibration with nonlinear wheel–rail contact. The coupled model can be used as an effective and robust tool compared to simplified models for assessing the stability of tank wagon. The transient fluid slosh model is analysed using the computational fluid dynamic method combined with the volume of fluid technique. The tank wagon dynamic model is solved using the fourth-order Runge–Kutta method based on the 19 degrees of freedom model with longitudinal, vertical, roll and pitch vibrations. The wheel–rail contact is considered according to nonlinear Hertzian and Kalker linear rolling contact theories. The fluid slosh model is validated using experimental data. The dynamic response characteristics of the partially filled railway tank wagon are investigated under straight-line braking manoeuvre using the coupled model. The results obtained from a parametric study, including the cross sectional shape and the filled volume show that the modified-oval cross section improves the dynamic response characteristics, which are attributed to its lower fluid's centre of gravity coordinate in the longitudinal direction and low lateral moment transfer of the fluid.

Vibration feature evolution of locomotive with tooth root crack propagation of gear transmission system

Abstract As the prompt development of modern railway transportation towards high-speed and high load-capacity, the high-power locomotive is urgently required. Under this situation, the wheel–rail dynamic interaction is becoming more and more intensified which will deteriorate the vibration condition of the key elements of locomotive, such as the gear transmissions. Once gear failures are present, such as gear tooth crack or breakage, it is likely to threaten the operation safety of the locomotive. Thus, deep insight into the fault features of the locomotive gear transmission is urgently necessary for prevention of the induced disastrous consequences. This paper is concentrated on the fault vibration feature extraction of a locomotive in presence of gear tooth root crack under the complicated dynamic excitations from both the gear transmissions and the nonlinear wheel–rail interactions. The locomotive–track coupled dynamics model considering the dynamic effect of the mechanical power transmission path is employed, and the time-varying mesh stiffness of the gear pair with tooth root crack fault and the rail geometric irregularities are then incorporated into the dynamics model to obtain the vibration responses. Then, angular synchronous average technique is proposed to enhance the fault vibration features, and the statistical indicators extracted in frequency domain are developed to reveal the evolution law for the crack propagation scenarios along crack depth or tooth width. The analyzed results indicate that the angular synchronous average technique could effectively reveal the fault vibration feature, and the M8A in the selected statistical indicators is most sensitive to the tooth crack propagation in frequency domain.