Fatal Error E5004

[Anonymous]

This is the first time I have used Nastran In-CAD, I am using it with Inventor 2018. I have created a 4 part box frame to be used as a lifting and transportation frame, with some brackets added to the side flanges to bolt the mass on to be lifted and 4 lifting eyes at each corner of the frame for the cables to attach onto.

The four bars have bonded surface contacts on each other to simulate being welded together. The brackets have offset bonded contacts between the edge of the bracket and adjacent flange face, and the lifting eyes have bonded contacts with the top flange face to simulate being welded together.

I have created four cable connectors at each corner of the frame. In each case, a 3D point was created at the centre of the lifting eye component as one end point, and a 3D point was created in the centre of the frame and a designated height above the frame as the other end point.

I have assigned a tension in each of the cables, and a vertical load acting down on the faces of the brackets. I have constrained the top centre point where the 4 cables meet to be completely fixed.

When I try to run the solution I receive Error E5004: Stiffness matrix singularity or non-positive definite. I am looking for help regarding 2 points:

1. It's my understanding this error occurs if a model is not fully constrained, and the model cannot be free to translate in X,Y or Z direction or rotate in the X,Y or Z axes. Is this the case? And if so, my model is a design which is to be lifted and moved and so theoretically in real life could move in 6 degrees of freedom, therefore is it possible to constrain this in Nastran and still simulate real life application?
2. I'm unsure if the way I have attached the cable connectors is right, since I only selected as an end point a 3D point on the model, when the model does run will the cables actually 'pull' on the frame in the same they would in real life? As this is my first time using Nastran I'm not totally sure on how the connectors work or if there is another way to do this? Although I think I will have to resolve my E5004 error before addressing this.

[Anonymous]

This is the first time I have used Nastran In-CAD, I am using it with Inventor 2018. I have created a 4 part box frame to be used as a lifting and transportation frame, with some brackets added to the side flanges to bolt the mass on to be lifted and 4 lifting eyes at each corner of the frame for the cables to attach onto.

The four bars have bonded surface contacts on each other to simulate being welded together. The brackets have offset bonded contacts between the edge of the bracket and adjacent flange face, and the lifting eyes have bonded contacts with the top flange face to simulate being welded together.

I have created four cable connectors at each corner of the frame. In each case, a 3D point was created at the centre of the lifting eye component as one end point, and a 3D point was created in the centre of the frame and a designated height above the frame as the other end point.

I have assigned a tension in each of the cables, and a vertical load acting down on the faces of the brackets. I have constrained the top centre point where the 4 cables meet to be completely fixed.

When I try to run the solution I receive Error E5004: Stiffness matrix singularity or non-positive definite. I am looking for help regarding 2 points:

1. It's my understanding this error occurs if a model is not fully constrained, and the model cannot be free to translate in X,Y or Z direction or rotate in the X,Y or Z axes. Is this the case? And if so, my model is a design which is to be lifted and moved and so theoretically in real life could move in 6 degrees of freedom, therefore is it possible to constrain this in Nastran and still simulate real life application?
2. I'm unsure if the way I have attached the cable connectors is right, since I only selected as an end point a 3D point on the model, when the model does run will the cables actually 'pull' on the frame in the same they would in real life? As this is my first time using Nastran I'm not totally sure on how the connectors work or if there is another way to do this? Although I think I will have to resolve my E5004 error before addressing this.

[John_Holtz]

Hi @Anonymous Welcome to the In-CAD forum.

Here are some answers to your questions.

1. Yes to the question "is it possible to constrain this in Nastran and still simulate real life application?". I assume that you are performing a linear static analysis, so you are only simulating one instant in time where the load is being held without moving. Since it is not moving, you can put constraints on the model that do not affect the displacement and stress but provide a statically stable model that can be solved mathematically. See the attached image for a suggestion on how to restrain the model.
2. If I understand your description about the cable connection to the lifting eye, the cable is not connected to any geometry. In other words, it is not connected to any nodes on the mesh. Is that correct? If so, you need to create a rigid connector from the center of the eye to an edge or face on the eye. Then the cable will be connected to the eye through the rigid connectors.

By the way, I think you do not need to apply a tension to the cables if the arrangement is anything like what I show in my figure. In that arrangement, the tension in the cables is developed because of the vertical load (simulating the weight of the frame and the load being lifted). The tension is not created because you pretension the cables and then lift it.

[John_Holtz]

Hi @Anonymous Welcome to the In-CAD forum.

Here are some answers to your questions.

1. Yes to the question "is it possible to constrain this in Nastran and still simulate real life application?". I assume that you are performing a linear static analysis, so you are only simulating one instant in time where the load is being held without moving. Since it is not moving, you can put constraints on the model that do not affect the displacement and stress but provide a statically stable model that can be solved mathematically. See the attached image for a suggestion on how to restrain the model.
2. If I understand your description about the cable connection to the lifting eye, the cable is not connected to any geometry. In other words, it is not connected to any nodes on the mesh. Is that correct? If so, you need to create a rigid connector from the center of the eye to an edge or face on the eye. Then the cable will be connected to the eye through the rigid connectors.

By the way, I think you do not need to apply a tension to the cables if the arrangement is anything like what I show in my figure. In that arrangement, the tension in the cables is developed because of the vertical load (simulating the weight of the frame and the load being lifted). The tension is not created because you pretension the cables and then lift it.

[Anonymous]

Hi @John_Holtz, thanks for your reply.

I have tried to constrain the model as your graphic depicted. The top end point of the cables is fully fixed, with 3 edges fixed with the translational contraints you suggested. I had a feeling about the pretension in the cables, thanks for confirming I've now removed that pretension value.

I have tried to connect the cables to the model via the rigid body connectors, however I'm not sure I've done it correctly based on the results plot?

On another note, is there any way to show my stress results as a fully contoured plot and not just the dotted outline of the model? I was hoping to be able to analyse the whole frame and look at specific stress values through elements at the brackets, at the bonds between beams etc. (I realise this may just be a setup problem but I am new to In-CAD and can't find a way to change it!).

I have attached images of my rigid body connectors, the whole model and my results.

Thanks,

Daniel 

[Anonymous]

Hi @John_Holtz, thanks for your reply.

I have tried to constrain the model as your graphic depicted. The top end point of the cables is fully fixed, with 3 edges fixed with the translational contraints you suggested. I had a feeling about the pretension in the cables, thanks for confirming I've now removed that pretension value.

I have tried to connect the cables to the model via the rigid body connectors, however I'm not sure I've done it correctly based on the results plot?

On another note, is there any way to show my stress results as a fully contoured plot and not just the dotted outline of the model? I was hoping to be able to analyse the whole frame and look at specific stress values through elements at the brackets, at the bonds between beams etc. (I realise this may just be a setup problem but I am new to In-CAD and can't find a way to change it!).

I have attached images of my rigid body connectors, the whole model and my results.

Thanks,

Daniel 

[John_Holtz · Accepted solution]

Hi Daniel,

I think you need to display the mesh so that you can see the results. When the contours do not show, it is usually because the mesh is hidden.

Otherwise, the images look okay. You may want to zip the Inventor files (.iam and .ipt) and attach the model if you have any other questions or problems. That makes it easier to check the setup and results in detail.

[John_Holtz · Accepted solution]

Hi Daniel,

I think you need to display the mesh so that you can see the results. When the contours do not show, it is usually because the mesh is hidden.

Otherwise, the images look okay. You may want to zip the Inventor files (.iam and .ipt) and attach the model if you have any other questions or problems. That makes it easier to check the setup and results in detail.