Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Number of plies as optimization response

Hi hi

I am trying to set up a free size optimization model in Hypermesh where the number of composite laminate plies is one of the parameters affecting the objective function. I just wanted to know if it is possible to set up a free size optimization where the number of plies created by the solver can be defined as an output response or as a design variable. I also wanted to know how to set up a composite optimization in Hyperstudy where the design variable is the ply thickness and the objective is an external DRESP3 response defined in Compose.

Thanks!

Find more posts tagged with

Academic

English

Student Edition License

Accepted answers

All comments

Adriano_Koga

1) you can't define explicitely number of plies in Free-Size, BUT by controlling the maximum allowable thickness for each angle, and total thickness, you should be able to indirectly control number of plies.

2) You don't need to use HyperStudy to couple DRESP3 with Compose. This is directly done by OptiStruct. Ply thickness can be easilly defined in a composite size optimization in OptiStruct.

Kaushik Iyer

1) you can't define explicitely number of plies in Free-Size, BUT by controlling the maximum allowable thickness for each angle, and total thickness, you should be able to indirectly control number of plies.

2) You don't need to use HyperStudy to couple DRESP3 with Compose. This is directly done by OptiStruct. Ply thickness can be easilly defined in a composite size optimization in OptiStruct.

thanks for the answer

1. Yup i have placed a constraint on the manufacturable ply thickness and the maximum laminate thickness in free size but can i get an output for the number of plies generated by the free size algorithm so that i can connect it to another response?

2. Yes my compose function is already connected to optistruct currently in a size optimization model but i would like to use the advanced optimization algorithms as well as the pareto front generation capabilities in hyperstudy

Adriano_Koga

thanks for the answer

1. Yup i have placed a constraint on the manufacturable ply thickness and the maximum laminate thickness in free size but can i get an output for the number of plies generated by the free size algorithm so that i can connect it to another response?

2. Yes my compose function is already connected to optistruct currently in a size optimization model but i would like to use the advanced optimization algorithms as well as the pareto front generation capabilities in hyperstudy

1) i'm not aware of any way of getting this directly, unfortunately.

2) you could try to build a "zero-iteration" optimization run in OS, with MAXITER=0, and then import it as your optimization model in HyperStudy. BUT notice that using an external optimizer might lead to higher run times, overall. But you could do that, for sure. Then your design variables would be the initial values into the DESVARs. You can use GRSM as your optimization method for multi-objective.

Kaushik Iyer

1) i'm not aware of any way of getting this directly, unfortunately.

2) you could try to build a "zero-iteration" optimization run in OS, with MAXITER=0, and then import it as your optimization model in HyperStudy. BUT notice that using an external optimizer might lead to higher run times, overall. But you could do that, for sure. Then your design variables would be the initial values into the DESVARs. You can use GRSM as your optimization method for multi-objective.

1. Okay no worries thanks

2. Yes i will try and do that. Can the external response which i imported into optistruct be used in hyperstudy as well?

Adriano_Koga

1. Okay no worries thanks

2. Yes i will try and do that. Can the external response which i imported into optistruct be used in hyperstudy as well?

2. For HyperStudy, the solver itself is a black box. As long as you can run the solver with Compose (keeping track of the necessary files), it will work with no problems. I suggest you to parametrize your OS model, but keeping the Compose file reference as an absolute path, keeping it always in a single place, so that you don't need to copy it everytime to new folders.