Run-up analysis

[verg1lio]

Hi all, good afternoon!

I have been trying to run a Run-up analysis, simulating the transient behavios of the rotor speeding up from 0 to the operating speed. Currently, I believe to have reached a reasonably good approximation, althought the nominal rotor speed is locked at a given value.

Would it be possible to dinamically change the rotor speed to consider gyroscopic effects?

See code below for example:

`# TRANSIENT ANALYSIS - RUN-UP

Variable definitions

rotor_speed_min = 0 # RPM (Run-up final speed)
rotor_speed_max = 10000 # RPM (Run-up final speed)
t_min = 0 # seconds (Run-up initial time)
t_max = 120 # seconds (Run-up time)
time_samples = 100001 # number of time samples
t = np.linspace(t_min, t_max, time_samples)
dt = t[1] - t[0]

Create time and speed ramp

t = np.linspace(t_min, t_max, time_samples)
dt = t[1] - t[0]
speed_rpm = np.linspace(rotor_speed_min, rotor_speed_max, time_samples)
speed_rad_s = speed_rpm * (2 * np.pi / 60)
max_orbits = {}

Analysis loop - TRANSIENT ANALYSIS - RUN-UP

for error_um in errors:
print(f"Run-up analysis from 0 to {rotor_speed_max} RPM for {error_um} µm of concentricity manufacturing error...")

# Unbalance forces
m1 = base_mass_moment_1 * error_um / 2
m2 = base_mass_moment_2 * error_um / 2
F = np.zeros((time_samples, rotor.ndof))

# Time-varying excitation based on instantaneous speed
for i in range(time_samples):
    omega = speed_rad_s[i]
    F[i, 4 * n1 + 0] = m1 * omega**2 * np.cos(omega * t[i])
    F[i, 4 * n1 + 1] = m1 * omega**2 * np.sin(omega * t[i])
    F[i, 4 * n2 + 0] = m2 * omega**2 * np.cos(omega * t[i])
    F[i, 4 * n2 + 1] = m2 * omega**2 * np.sin(omega * t[i])

# Run time response with force excitation (nominal speed is 0)
response_time = rotor.run_time_response(0, F, t)

# Plot results
trs1d = response_time.plot_1d(probe=[probe1, probe2, probe4, probe5, probe3])
trs1d.update_layout(
    title=(f"Run-up transient analysis from 0 to {rotor_speed_max} RPM, "
            f"with {error_um} µm concentricity manufacturing error")
)
trs1d.show()`

[jguarato]

Hi @verg1lio!

It’s possible to pass the speed argument as an array with the same size as the time array to run_time_response(). This will trigger time integration based on the Newmark method. [Examples]