Remove rigid body modes from modal analysis

[jguarato]

For the 6 dof model, running modal analysis has typically revealed modes with zero natural frequency, known as rigid body modes. This may result from the lack of constraints in the model and can be observed by comparing the Campbell diagrams for the 4 and 6 dof models of the same rotor system, as shown below:

However, a problem related to the convergence of Scipy's eigen solver was also noticed when changing the frequencies argument:

samples = 31
speed_range = np.linspace(315, 1150, samples)
campbell = rotor.run_campbell(speed_range, frequencies=6)
fig = campbell.plot(harmonics=[0.5, 1])

To address this, my suggestion is to remove these zero-frequency modes after calculating the eigenvalues and increase the number of eigenvalues to be obtained from the solver:

As a result, the new Campbell diagrams are: