I set the excitation frequency to be 140hz and observe the modal shape. I found the excited modal shape also same as that before thickness changed.
The animation of the harmonic results is showing you the system response to the load/frequency event. (This is not the same as the animation of the natural modes of vibration.) While changing the wall thickness will change the natural frequencies, the response of the piping system to a given load may or may not be all that different. The overall shape (what gets animated) can be similary visually, but the loads and stresses can be different.
For my undersatnding, this time, the excited modal shape sould be also almost same as that shape(back and forth along X-axis) under first natural frequency (142hz).
Maybe. The overall response of the system is a combination of the responses off all of the individual modes. By changing the wall thickness, if one of the (major) contributing modes changes, then the harmonic animation will change. If the major contributing modes don't change, then the overall harmonic response won't change.