Radiator mechanical limits

I'm not an engineer, but I would like to learn more about the mechanical limits of radiators. In particular, how far out they can extend from the spacecraft before deforming plastically at maximum thrust. It seems to me that there are two case study: the first is when the spacecraft is thrusting in a straight line, the second is when it is rotating/spinning. I think I found how to model the first case by looking at this video and this website. For semplicity's sake, I'll assume the radiator is a full parallelepiped of just one material, with no pipes, no fluids, and no tapering.

In order to understand each other, for the first case let's use:

• σ_y: Yield strength of the material (found in the appropriate material file)
• P_w: Panel width (found in the radiator module editor)
• P_h: Height (found in the radiator module editor)
• P_t: Thickness (found in the radiator module editor)
• P_a: Armor thickness (found in the radiator module editor)
• P_n: Panels (found in the radiator module editor)
• m: Mass of the radiator (found in the radiator module editor)
• a: Dry acceleration of the spacecraft (found in the ship editor)
• L = P_w·P_n: Total length of the panels perpendicular to the surface of the spacecraft's armor
• h = P_h: Length of the panel parallel to the main engine thrust
• b = P_t+P_a·2: Length of the panel parallel to the surface of the spacecraft's armor and perpendicular to the main engine thrust
• g₀ = 9,80665 m/s²: Standard gravity
• F = m·a·g₀: Weight of the radiator at dry acceleration
• M = F·L/2: Moment of the radiator at dry acceleration
• I = b·h³/12: Second moment of area, second area moment or area moment of inertia
• σ = M·(h/2)/I: Maximum stress the radiator has to withstand

σ ≤ σ_y: The radiator can withstand the dry acceleration and will deform elastically
σ > σ_y: The radiator can not withstand the dry acceleration and will deform plastically or break

For the second case, probably the values of b and h have to be reversed. I don't know how to calculate the new σ though.