Perpustakaan Digital ITB

High-speed train is one of the most used ground transportation modes in several countries. The use of high-speed trains also has a positive impact on the economy. Due to the advantages they offer, Indonesia has begun developing high-speed trains, a project conducted by PT INKA. Commuter rail, one of the most widely used modes of transportation in Indonesia, are experiencing an increase in passenger numbers, which is beginning to cause overcapacity. To increase passenger capacity and replace conserved trains, Indonesia planned to retrofit aging commuter rail. Safety is a crucial factor to consider in the development process of these trains. One way to ensure passenger safety is by focusing on the structural aspects of the carbody. In this regard, calculations, simulations, and experiments are necessary, employing the same loading conditions experienced in actual operations. In this final thesis, static loading will be simulated numerically using ABAQUS software. The static loading simulation will be done by applying vertical load while the carbody experiencing full passenger load condition. Static simulation shows that the high-speed train carbody meets the requirement with maximum tensile and shear stress result, 198.1 and 95.5 MPa, at 62.49% and 30.13% of the yield strength, and bending mode that exceeds the minimum requirement with the simulation result 15.1 Hz. Static simulation also shows that the commuter rail carbody does not meet the requirement with maximum tensile and shear stress result, 566 and 184.5 MPa, at 288.78% and 104.24% of the yield strength. Static experiment shows a desirable high-speed train positive camber of 10.58 mm under full load, with a camber change of -3.18 mm from empty to full load. Simulation results for camber (-3.47 mm) deviate by 8% from experimental findings. Stress analysis reveals similarities in trends but significant differences in specific areas like the underframe, side sill, windows, and doors, likely due to unaccounted camber effects in finite element method simulations. Meanwhile for the commuter rail, undesirable negative camber of -8.19 mm under full load, with a camber change of -14.19 mm from empty to full load. Simulation results average -11.92 mm camber, showing a 16% error from experimental values. Stress analysis shows significant errors in several load cases, with errors above 10% attributed to experimental and simulation, although both results show similar trends. These errors may be caused by non-uniform load distribution, overly rigid finite element models, and the structure's aging. Experimentally, lateral stress at the middle underframe increases sharply at specific loads, while the longitudinal stress decreases unexpectedly under higher loads. Non-linear behavior observed suggests elastic deformation inconsistencies. Previous studies analyzed track irregularities on the HST carbody, weighing 38165 kg fully loaded, with the updated design reducing full-load weight to 28910 kg (24.25% lighter). Comparing the 2023 and 2024 designs at a critical point (rear roof-side wall weld), the 2024 version shows a stress range 66% lower (25 MPa vs. 75 MPa) under dynamic load. Simulations used tangent track irregularity at 250 km/h over 1 km. The 2024 design improves fatigue performance, significantly enhancing the critical point's durability. Further research is needed to confirm actual fatigue life improvements. Dynamic load experiment was conducted on Depok – Bogor – Depok route counted as 1 trip. From this experiment, strain data measured was then smoothed to remove noise, but it accidentally reduced strain range magnitudes. Only the MA10 channel retained strain ranges close to original data after smoothing. There is no definitive way to confirm the removed data is solely noise. Averaging data introduces uncertainties, making this method incompatible with such studies. FFT analysis of strain data in the frequency domain shows the largest magnitude near 0 Hz, indicating data concentration in low-frequency areas. A peak at 60 Hz is observed because noise frequencies generated by vehicles, motors, power lines, and the like frequently dominates the recordings in 60 Hz area, but low-pass filters with cutoff frequencies between 1-40 Hz show negligible changes. Rainflow counting analysis reveals that moving averages reduce stress ranges, corresponding to lower strain ranges. Larger data sets used for averaging result in decreased high exceedance stress ranges. The method proves incompatible with this study due to its impact on stress range accuracy. Low-pass filter shows no substantial differences for channels 6 and 7, supporting FFT results of no impactful deformation in the 1-40 Hz range. Applying a 1 Hz cutoff filter reduces stress ranges for both channels, suggesting noise frequencies higher than 1 Hz have low stress ranges. This confirms the presence of noise at higher frequencies.