The problem is that if I keep increasing the cutoff frequency the stresses continue to increase. The time history has very high frequency components, but this frequency has a very low amplitude. As if it was an electrical signal with a lot of noise. The pipe is not large, the longest section is 6 meters.