# Shell buckling

I’ve been evaluating the size of flotation tanks needed to install a large jacket substructure. The circular cylindrical tank sections in my case have a D/t ratio upwards of 300 — too slender and non-compact to withstand severe buckling hazard under stress. Thin walled large tanks such as these are typically reinforced with uniformly-spaced internal ring frame stiffeners. And so I’ve been testing to see what ring frame stiffener spacing would be reasonable to maintain sufficient buckling strength for this tank to perform its job as intended, which is to safely assist in floating, upending and in setting-down the jacket structure on sea-bed.

For each of the eight hundred ring frame spacing variations, this little script generates elastic buckling strengths for (a) axial stress, (b) bending, (c) torsion and shear force, (d) lateral pressure, and (e) hydrostatic pressure respectively — all in one go. And it powers through this in 1.0s. The bottleneck is not the speed at which it does this but in validating the results I do by hand for one or two unit values.

```
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
ebs.py -- Elastic buckling strength of un-stiffened circular
cylindrical shell for a range of distances between ring frame
stiffeners, based on DNVGL-RP-C202 (2019), 2020 ckunte
"""
import numpy as np
import matplotlib.pyplot as plt
def bcoeff_ebs(D, t, l):
# Mechanical properties of structural steel
# Poisson's ratio
nu = 0.3
# Young's modulus (MPa)
E = 2.1E5
# Mid-plane radius of cylindrical tank
r = 0.5 * (D - t)
# The curvature parameter
Z_l = (l**2 / (r * t)) * np.sqrt(1 - nu**2)
"""
Buckling coefficients (psi, xi, rho) for un-stiffened
cylindrical shells, mode (a) shell buckling from Table 3-2,
Buckling strength of shells, DNVGL-RP-C202 (2019) for:
(a) axial stress,
(b) bending,
(c) torsion and shear force,
(d) lateral pressure, and
(e) hydrostatic pressure respectively
"""
psi = [
1.0,
1.0,
5.34,
4.0,
2.0
]
xi = [
(0.702 * Z_l),
(0.702 * Z_l),
(0.856 * np.power(Z_l, 0.75)),
1.04 * np.sqrt(Z_l),
1.04 * np.sqrt(Z_l)
]
rho = [
(0.5 * np.power((1 + r / (150.0 * t)), -0.5)),
(0.5 * np.power((1 + r / (300.0 * t)), -0.5)),
0.6,
0.6,
0.6
]
# For each array of [psi, xi, rho] coefficients, compute reduced co-effcient (C)
# elastic buckling strength of un-stiffened circular cylindrical shell (f_E)
C = []
f_E = []
for i, j, k in zip(psi, xi, rho):
# Reduced buckling coefficient, see equation 3.4.2, DnV-RP-C202
c = i * np.sqrt(1 + (k * j / i)**2)
C.append(c)
"""
Elastic buckling strength, see equation 3.4.1, DnV-RP-C202
Also applicable to:
Torsion and shear force when
(l / r) < 3.85 * sqrt(r / t)
Lateral / hydrostatic pressure when
(l / r) < 2.25 * sqrt(r / t)
"""
f_e = c * (np.pi**2 * E / (12.0 * (1 - nu**2))) * (t / l)**2
f_E.append(f_e)
"""
Replace values of elastic buckling strength from f_E list
for long cylinders for torsion and shear force:
"""
if any(l / r) > (3.85 * np.sqrt(r / t)):
f_E[2] = 0.25 * E * np.power((t / r), 1.5)
pass
"""
Replace values of elastic buckling strength from f_E list
for long cylinders for lateral / hydrostatic pressure:
"""
if any(l / r) > (2.25 * np.sqrt(r / t)):
f_E[3] = 0.25 * E * np.power((t / r), 2)
f_E[4] = f_E[3]
pass
# any() is used (in above conditionals) as "l" is an iterable list
return C, f_E
def ebs_plot(D, t, l):
# Call only the C list (not used)
C = bcoeff_ebs(D, t, l)[0]
# Call only the f_E list
f_E = bcoeff_ebs(D, t, l)[1]
# Plot labels
lbl = [
'$f_{Ea}$ for axial stress',
'$f_{Em}$ for bending',
'$f_{Et}$ for torsion and shear',
'$f_{El}$ for lateral pressure',
'$f_{Eh}$ for hydrostatic pressure'
]
# Generate plots
for i, j in zip(f_E, lbl):
plt.plot(l, i, label=j, linewidth=2)
pass
plt.xlabel('Distance between ring frames, l (mm)')
plt.ylabel('Elastic buckling strength, $f_E$ (MPa)')
plt.title('Un-stiffened circular cylindrical shell (%d$\\times$%d)'%(D,t))
plt.grid(True)
plt.legend(loc=0)
plt.savefig('lfE_%dx%d.svg'%(D, t))
plt.close()
pass
def main():
# -- USER INPUTS --
# Size of circular cylindrical tank (mm)
D = [6000.0]
t = [20.0]
# Distance between ring frames (mm)
l = [np.arange(400.0, 1200.0, 1.0)]
# -- END of USER INPUTS --
for i, j, k in zip(D, t, l):
ebs_plot(i, j, k)
pass
if __name__ == '__main__':
main()
```