Gustavo Silva
Civil/Environmental
- Oct 29, 2020
- 2
Hi everyone,
I'm doing some work about topology optimization of fiber reinforced concrete and i have one question regarding the consideration of the self weight through a gravity load.
So, in order to explain, the goal here is to minimize the strain energy (maximize stifness of the structure) for 40% of the initial volume, and here is the geometry that will be optimized, with the support conditions defined:
For the load case to be used in the optimization processes the initial goal was to take into consideration the weight of the beam plus an overload 5 kN/m in order to simulate a real-life example. In Abaqus the self-weight and the overload are applied through a gravitational load and a pressure load respectively.
When considering that only the self-weight (gravitational load) is applied the result of the optimization is the following:
When considering that only the overload (pressure load) is applied that result is the following:
As you can see, for the self-weight the algorithm focusses on removing material located on the center of the beam, on the other hand, when it’s considered the overload the optimization outcome becomes more similar to a trussed beam, which tottaly makes sense. But now when i ran an optimization process considering the application of the two load it was expectable that the outcome would be something between the ones presented before, but more similar to the "trussed" one, giving that the overload value is bigger than the self-weight, which did not happen as you can see in the picture below:
I even tried change the values of the loads but even when i put the gravity very very low the result continues to be the same :/
I know that i could "work arround" thist problem by just applying the self-weight as many forces uniformely distributed by the geometry, but would be nicer the keep with the gravity load because with that the algorithm would always take into consideration the weight of the beam throughout the optimization cycles.
Do any of you know a solution for this problem? Or maybe i'm doing something wrong? Would really help me alot!
Thanks anyway
I'm doing some work about topology optimization of fiber reinforced concrete and i have one question regarding the consideration of the self weight through a gravity load.
So, in order to explain, the goal here is to minimize the strain energy (maximize stifness of the structure) for 40% of the initial volume, and here is the geometry that will be optimized, with the support conditions defined:
For the load case to be used in the optimization processes the initial goal was to take into consideration the weight of the beam plus an overload 5 kN/m in order to simulate a real-life example. In Abaqus the self-weight and the overload are applied through a gravitational load and a pressure load respectively.
When considering that only the self-weight (gravitational load) is applied the result of the optimization is the following:
When considering that only the overload (pressure load) is applied that result is the following:
As you can see, for the self-weight the algorithm focusses on removing material located on the center of the beam, on the other hand, when it’s considered the overload the optimization outcome becomes more similar to a trussed beam, which tottaly makes sense. But now when i ran an optimization process considering the application of the two load it was expectable that the outcome would be something between the ones presented before, but more similar to the "trussed" one, giving that the overload value is bigger than the self-weight, which did not happen as you can see in the picture below:
I even tried change the values of the loads but even when i put the gravity very very low the result continues to be the same :/
I know that i could "work arround" thist problem by just applying the self-weight as many forces uniformely distributed by the geometry, but would be nicer the keep with the gravity load because with that the algorithm would always take into consideration the weight of the beam throughout the optimization cycles.
Do any of you know a solution for this problem? Or maybe i'm doing something wrong? Would really help me alot!
Thanks anyway