Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

validation of output result of LOAD/PBLAST card

Dear Forum Members,

I hope this message finds you all well. I am currently working on utilizing the LOAD/PBLAST card from Radioss for Air Blast-related applications. During my exploration, I have encountered some doubts and questions regarding this card. I would greatly appreciate it if you could provide some insights or clarification on the following points:

In the experimental data of /PBLAST the unit system {cm, g, µs} is used. Should we define the problem using this specific cm, g, µs system, or are there other unit systems that can be used?
It appears that extracting the pressure output on the applied surface is not possible using this card. How can we validate the external force evaluated on the area?
In the context of validating the Air pressure developed on the surface, which formulation or methodology is commonly employed?

If any of you have validation data, examples, or any information related to the above queries, I would be grateful if you could share it. Your contributions would greatly assist me in my project.

Thank you in advance for your time and support.

Best regards,

Sachin

Find more posts tagged with

English

Radioss

v2021.1

Accepted answers

All comments

Mathis

Hi Sachin,

Great questions about /LOAD/PBLAST, thanks for asking.

"In the experimental data of /PBLAST the unit system {cm, g, µs} is used. Should we define the problem using this specific cm, g, µs system, or are there other unit systems that can be used?"

Yes, you can use any Unit System you want. It is however mandatory to specify your unit system in the /BEGIN cards in your Radioss model. Otherwise, Radioss won't do any conversion and you will get bad results.

"It appears that extracting the pressure output on the applied surface is not possible using this card. How can we validate the external force evaluated on the area?"

It is true that it is currently not possible to output directly applied pressure. You can use the request /H3D/NODA/FEXT to output external force. Then, by dividing by the area of the element, you can retrieve the applied pressure.
In Radioss 2023 (available next August), you should be able to output applied pressure throuth /TH/SURF card.

"In the context of validating the Air pressure developed on the surface, which formulation or methodology is commonly employed?"

For free air burst, I would use Exp_data=1.
You can set I_tshift=2, so the blast is directly applied at t=0 to the surface without having to wait the time for the wave to travel from the det point to the surface.

I want to add 2 very important warnings:

The normal of your elements should be oriented towards the blast

/LOAD/PBLAST is a simplified loading method because the arrival time and incident pressure are not adjusted for obstacles.
It also does not consider confinement or ground effects.

I am going to share with you an example attached to next post.

I hope this was helpful to you.

Regards,

Mathis.

Mathis

Here is an example model

PBLAST_EXAMPLE.zip

sachin awasare

Dear Mathis Loverini,

I greatly appreciate your prompt response, which has cleared up most of my doubts.

I have thoroughly examined the example model provided, and it is quite self-explanatory. However, I have a couple more questions regarding the topic.

Could you please provide a brief explanation of how the "scaled distance" Iz option works? I would like to gain a better understanding of its functionality.
In my attempt to validate the results obtained from the given model, I have encountered a discrepancy with the calculated pressure values. I have attached my calculations for your reference (refer to Result calculations.docx). I kindly request you review them and provide your valuable feedback on the matter.

I am eagerly awaiting your response and appreciate any insights you can provide.

Thank you in advance.

Result calculations.docx

Mathis

Hi Sachin,

Thanks for the positive feedback, I appreciate it.

"1. Could you please provide a brief explanation of how the "scaled distance" Iz option works? I would like to gain a better understanding of its functionality."

As you may know, the pressure applied to the structure is depending on the scaled distance (ratio between the distance between the explosive and the mass^(1/3)).

This Iz option is set by default to 2, which means that the R of the scaled distance is updated during the simulation which is interesting when your structure is moving.

If you set it to 1, R is not updated during the simulation.

I would recommand to keep the value to 2.

"2. In my attempt to validate the results obtained from the given model, I have encountered a discrepancy with the calculated pressure values. I have attached my calculations for your reference (refer to Result calculations.docx). I kindly request you review them and provide your valuable feedback on the matter."

I think the thing is you are probably missing the real peak pressure here due to the h3d output frequency. You would need to have a much bigger frequency to be sure not to miss the peak.

Could you please check what happens if you increase the number of h3d outputs wrt time (with /H3D/DT)?

Thanks,

Mathis

sachin awasare

Dear Mathis,

I have a couple of follow-up questions based on your previous recommendations:

In cases where the structure is fixed, as you mentioned, using Iz option 2, will the results still provide considerable verification? I am curious to know how effective this approach would be under such conditions.
Following your suggestion, I increased the output h3D frequency from 0.1 ms to 0.001 ms and re-ran the simulation. However, the peak forces did not increase significantly. The maximum recorded external force was 416 N before, and now it is 508 N, resulting in a peak pressure of 1.27 N/mm². However, the empirically calculated pressure indicates 20 MPa. Please check the below image

Could you kindly advise if the empirical pressure calculation method I used is correct? I would greatly appreciate your guidance on this matter.

Regards,

Sachin

Mathis

Hi Sachin,

"In cases where the structure is fixed, as you mentioned, using Iz option 2, will the results still provide considerable verification? I am curious to know how effective this approach would be under such conditions."

In such cases Iz=1 and Iz=2 should deliver the same results. Iz=1 might be a bit quicker to compute, but nothing significant. THat's why Iz=2 is enable as default.

for your question 2, I need more time to investigate. I am going to be in vacation (1 week). I hope I can answer you before, but I am not sure.

Regards,

Mathis.

sachin awasare

Dear Mathis,

thank you for your reply.

Whenever you have the information ready, please let me know meanwhile I will try from my side

Regards,

Sachin

Mathis

Hi Sachin,

I am very happy to say that I found out that Radioss seems to be giving the expected results.

There were 2 issues in the verification document:

- We need to consider Preflected and not Pincident in the model as the elements are oriented towards the blast. Hence we are in the case where theta=0. So we have to look at the Preflected red curve in the graph.

- The scaled distance Z is not unitless. So when calculating Z, we have to be with the same unit system as the graph, which is g, cm, bar.

I took the liberty to update your document accordingly. You will see that now, everything is matching well.

Please have a look at the attched document and tell me if it is ok for you.

I hope it was helpful.

Regards,

Mathis

Result calculations (1).docx

sachin awasare

Dear Mathis,

I wanted to express my gratitude for your assistance. I have thoroughly reviewed the updated document and I am pleased to inform you that the results align perfectly. Apart from the given example, I am interested in validating some more examples.

I am also keen on implementing /PBLAST for a problem where I need to measure the pressure or external forces on points outside the Ship container while the blast is being carried out inside the container. Previously, I attempted to tackle this using the CEL approach and TYPE 18 contact, but it turned out to be computationally expensive due to the requirement of large elements. Please find below a picture of the problem definition.

Therefore, I would like to explore the possibility of utilizing /PBLAST for the above purpose. Could you please confirm whether it is possible to achieve this by using the /PBLAST command? Your expertise in this matter would be invaluable in guiding me towards an effective solution.

I am eagerly awaiting your response. Thank you for your time and consideration.

Best regards,

Regards,

Sachin Awasare

Community-Pasted-Attachment-1688964384263.png

Mathis

Hi Sachin,

Sorry for my late reply, I was off last week.

Happy to hear that the results are now matching and that you intend to validate PBLAST with more example. Maybe if these examples are public, you could create some Knowledge Base article to share with the community. I plan to do this in a few months myself

About the simulation of your Ship container, I understand that you may want to gain computation time, as CEL method require to mesh the fluid and can therefore be rather expensive.

As I mentionned, /LOAD/PBLAST does not take into account confinement effects. With the TNT inside the container, there will be confinement effects. So /LOAD/PBLAST is not an obvious candidate for such case. But it doesn't mean that we can't find a solution

Depending on what you want to analyze in the end, we can think about some different methods:

- If you are interested in what happens really close to the container, you can model TNT with SPH, or solid elements (with sol2sph enabled).

- If you have a target structure "a bit far" from the container and want only to model the pressure applied on the target, you could record on your CEL model, the pressure seen by some elements close to the structure and apply it on a second run with just your target and a PLOAD applying the same pressure.

- you could also replace the PLOAD of the previous method by a smartly calibrated PBLAST if the shape of the pressure wave is close to friedlander pattern.

As I mentionned, it depends on what you want to do and analyze in the end, so if you can tell me a little bit more about it, I could point you the best method for you.

Regards,

Mathis.

sachin awasare

Dear Mathis,

I wanted to discuss the second method you suggested, which involves recording pressure by CEL on the target surface and using PLOAD. However, I have found that this method requires a significant amount of computation time.

I have tried to modify the given box problem with two surfaces. The two surfaces are one after another to check how /PBLAST will work if the same command I apply on container blast.

After analysing, I have found that the measured pressure on the second plate is less than on first plate

Upon analysis, I observed that the measured pressure on the second plate is lower than that on the first plate. While the phenomenon appears to be functioning well, there is some uncertainty regarding the accuracy of the results. If there are published results available for validation, it would greatly enhance the usefulness of using this method on containers.

For your reference, please find below image where I applied /PBLAST to a container and measured the pressure on two small areas outside the container.

I would greatly appreciate your assistance in validating these results.

Regards,

SACHIN

Community-Pasted-Attachment-1689585221635.png

Mathis

Hi Sachin,

"Upon analysis, I observed that the measured pressure on the second plate is lower than that on the first plate. While the phenomenon appears to be functioning well, there is some uncertainty regarding the accuracy of the results."

The results you have seems to make sence. You just need to do the same thing as before to validate the values by looking at the graph.

The following case:

highlights a limitation of load PBLAST: you shouldn't have any pressure applied on the second plate as you have the first plate as an obstacle. As I mentionned "/LOAD/PBLAST is a simplified loading method because the arrival time and incident pressure are not adjusted for obstacles."

This case:

is really not good for Pblast. It is really too close to the explosive. Also, PBLAST is neglecting confinement effects.

In such case, I would recommand an ALE model. I think you can have a smaller Air domain than this one if you are just interested in the container:

Regards,
Mathis

sachin awasare

Dear Mathis,

For ALE method the main constraint is the modelling of the Air domain. As I need pressure on points that are far away from the container the size of the problem getting increased and taking more than 2 weeks to solve.

Regards,

Sachin Awasare.

Mathis

Hi Sachin,

2 weeks seem pretty huge for such simulations. What hardware do you have (number of cpus?). I am sure there are ways to speed-up your computation. Is it possible for you to share your ALE model ?

Regards,
Mathis

sachin awasare

Hi Sachin,

Great questions about /LOAD/PBLAST, thanks for asking.

"In the experimental data of /PBLAST the unit system {cm, g, µs} is used. Should we define the problem using this specific cm, g, µs system, or are there other unit systems that can be used?"

Yes, you can use any Unit System you want. It is however mandatory to specify your unit system in the /BEGIN cards in your Radioss model. Otherwise, Radioss won't do any conversion and you will get bad results.

"It appears that extracting the pressure output on the applied surface is not possible using this card. How can we validate the external force evaluated on the area?"

It is true that it is currently not possible to output directly applied pressure. You can use the request /H3D/NODA/FEXT to output external force. Then, by dividing by the area of the element, you can retrieve the applied pressure.

In Radioss 2023 (available next August), you should be able to output applied pressure throuth /TH/SURF card.

"In the context of validating the Air pressure developed on the surface, which formulation or methodology is commonly employed?"

For free air burst, I would use Exp_data=1.

You can set I_tshift=2, so the blast is directly applied at t=0 to the surface without having to wait the time for the wave to travel from the det point to the surface.

I want to add 2 very important warnings:

The normal of your elements should be oriented towards the blast

/LOAD/PBLAST is a simplified loading method because the arrival time and incident pressure are not adjusted for obstacles.
It also does not consider confinement or ground effects.

I am going to share with you an example attached to next post.

I hope this was helpful to you.

Regards,

Mathis.

Dear Mathis,

Please find below the reply from you.

In Radioss 2023 (available next August), you should be able to output applied pressure through/TH/SURF card.

As per this reply, The applied pressure output is available in Latest Radioss version?

kindly revert.

Regards,

Sachin Awasare

Mathis

Dear Mathis,

Please find below the reply from you.

In Radioss 2023 (available next August), you should be able to output applied pressure through/TH/SURF card.

As per this reply, The applied pressure output is available in Latest Radioss version?

kindly revert.

Regards,

Sachin Awasare

Hi Sachin,

Yes, it should be available in 2023 version of Radioss.

On my sde, I noticed a specific case, where it could be improve, so it will be in later version.

You should give it a try anyway.

Regards,

Mathis

sachin awasare

Hi Sachin,

Yes, it should be available in 2023 version of Radioss.

On my sde, I noticed a specific case, where it could be improve, so it will be in later version.

You should give it a try anyway.

Regards,

Mathis

I have tried, it is giving output in terms of pressure contours.

But some are the bugs I have encountered while using.

1) In Exp_data flag 3 is absent

2) When I enter the Pmin value the pressure is not dropping to zero. It is showing a Negative phase also.

3) The peak measure from the explosion matches with the Reflected pressure value calculated by the empirical formula of Kinney Graham, not incident pressure.

Regards,

Sachin Awasare