Active control of high speed train vibrations under wind and flexible foundation using artificial intelligence algorithms

Due to the interaction problem between the train and the rail in high-speed trains, the railway vehicle's driving safety and passenger comfort are adversely affected. At the beginning of these problems are rail irregularity, flexible foundation effect, and external effects such as wind load and seismic loads. In this thesis, simulation software has been developed to determine the dynamic responses of the train-track-bridge interaction (TTBI) system. While the full train model was modeled with 31 degrees of freedom to examine the dynamic responses in the lateral and vertical directions, the bridge and the rail were modeled according to the Euler-Bernoulli beam theory. In addition, in this thesis, the use of artificial intelligence algorithms and the active suspension system also have been examined. The primary purpose of the active suspension control is to eliminate the disturbing effects of the train with the linear actuator placed on the secondary suspension system. It is also crucial to reduce the effect of the wind load that the high-speed train may be exposed to. Because in the past trains have derailed and bridges have been damaged by the wind. With the use of state space forms, the Lagrangian approach is used to extract the second-order equations of motion for the train and bridge beam and then convert them to first-order differential equations. Simply supported bridge and rail beams, the fourth-order differential equations, are transformed into second-order equations of motion using the Galerkin method. These differential equations are solved quickly and with high precision in the time domain using the fourth-order Runge-Kutta algorithm. To verify the numerical results of the presented method, this study has also been compared with several accepted studies in the literature. The first four natural frequencies of the beam, which have been calculated using the bridge parameters, have been determined, and the resonance velocities, which are the critical speeds of the train corresponding to this determined frequency, have been calculated. The train moving at resonance speeds causes the maximum acceleration amplitudes to increase, especially in low-damped beams. In this thesis, maximum dynamic responses have been determined at variable speeds of the train. It has been understood that critical speeds are an important concept in terms of train-rail interaction. It has also been seen that well-damped beams reduce the maximum dynamic responses. In this thesis, the effect of the track system consisting of sleeper and ballast between the high-speed train and the bridge has also been examined. Since the track system has been used in high-speed railways for almost two centuries, it is essential to include it in the content of this study. In this context, the effect of the track system and the impact of the track parameters on the dynamic behavior of high-speed trains has been studied in detail. To reduce the disruptive effects caused by crosswind and bridge vibrations during bridge crossings of high-speed trains in windy areas, three different artificial intelligence-supported active controllers, namely PID, self-tuning fuzzy PID, and sliding mode control, have been used. The performances of these controllers on the dynamic responses of the high-speed train's wagon in the vertical and lateral directions have been studied in detail. In this context, variables such as the effect of the wind angle on the train and the wind speed have been analyzed separately. It has been observed that it exhibits superior performance compared to passive control by using artificial intelligence-supported active controllers. As a result, thanks to the TTBI system software modeled in this thesis, the effect of all parameters of the train, track, and bridge, the impact of train velocity, and the effect of variables such as wind angle and velocity can be determined without the need for time-consuming and expensive experimental studies. The thesis presented in this study has been examined in seven sections. The first section emphasizes the features and importance of high-speed trains in the world and our country. In addition, engineering problems caused by increasing train speeds have been mentioned, and solutions have been proposed. In the second part of the thesis, high-speed trains and related subjects have been researched, and a detailed literature review has been carried out. In the third section, a mathematical model has been created to conduct detailed analyzes of high-speed trains. Since a completely realistic analysis will be made in this thesis study, a full model with 31 degrees of freedom has been created by considering all the parameters that can be examined in a train. In addition, the detailed modeling of the bridges where high-speed trains pass and the modeling of the track structure on the railway line are also included in the analysis. In addition to the bridges, which are considered flexible ground, wind load is also taken into account in this thesis. By examining the wind load analysis in the literature, the wind load exposed to the high-speed train, especially during bridge crossings, is also modeled. After the mathematical modeling of the high-speed train was created, the equations of motion were obtained for analysis, and then the equations of motion were solved using Matlab software. This study's proposed solution method is verified in the fourth section of this thesis. Since the examined high-speed train model is very complex, two different models were created and solved with the proposed method used in this study. The accuracy of the study was made by comparing the results in the literature. The analysis results of the high-speed train and bridge interaction model presented in this thesis are given in the fifth section. First of all, a train-bridge interaction analysis was carried out, and then a train-track-bridge interaction analysis was carried out considering the track structure, and the effect of the track structure has been examined by comparing these two analyses. Before these analyses, the number of vibration modes of the bridge and the effect of the step size were determined. In addition, in this section, parameters such as the mass of the wagon, the length of the bridge, the effect of the track parameters, and the damping of the bridge are investigated in detail. The sixth section presents artificial intelligence-supported active controllers used to control vibrations in this thesis study. It is aimed to reduce the vibrations that occur under the influence of the wind load and bridge crossings of high-speed trains by using PID, self-tuning fuzzy PID, and sliding mode control. In this context, a controller simulation software has been developed that aims to minimize the vertical displacement of the wagon by means of four actuators placed between the wagon and the bogie. The performances of the controllers used have been compared by considering many parameters, such as the constant and variable speed of the train, the different mass of the wagon, the variable bridge length, and the speed and angle of the wind load. As a result, sliding mode control, self-adaptive fuzzy PID, and PID control were more successful in controlling the vibrations of the high-speed train, respectively. In addition, it is understood from the analysis results that the sliding mode control gives very successful results compared to other controllers. In the seventh and last part of this thesis, the controversial results of all the analyzes made in the study are given in detail.