A few observations:
The EIG calls anything 4 Hz and lower as "Flexible," and thus prone to the highest rate of failure due to the EIG.
The EIG's estimates for natural frequencies is higher than that of CAESAR's, I find. This may be due to lack of mass (fluid, insulation) in their estimations.
You should be careful when using 4 Hz as your final arbiter. It's important to understand the kind of problem you have.
By specifying 4 Hz as a minimum natural frequency, you're assuming that small slugs in your system will never exceed 4 events per second and cause a runaway condition.
However, you could have large slugs at 0.1 Hz that can still cause your system to fail or fatigue, despite not having a runaway condition.