-
-
May 13, 2026 at 9:35 am
julmeshreeya2210
SubscriberDear Sir/Madam,
I hope you are doing well.
I am currently working on a Notched Semicircle Bend test simulation project in ANSYS LS-DYNA and would greatly appreciate your guidance regarding crack propagation modeling.
In my model, I am using MAT_ELASTIC along with MAT_COHESIVE_MIXED_MODE and MAT_ADD_EROSION to simulate crack initiation and material failure under three point loading. However, despite implementing these material models, the crack is not propagating as expected during the analysis. (the specimen material is Sandstone)
.png)
I suspect the issue may be related to the erosion parameters, cohesive zone calibration, mesh dependency, or contact/timestep settings. Since I am relatively new to this type of coupled damage modeling, I would be very grateful for any suggestions or best practices you could share.
Specifically, I would like to seek guidance on:.
Appropriate failure criteria for realistic crack initiation and propagation.
Proper setup of MAT_COHESIVE_MIXED_MODE parameters.
Recommended element types and mesh strategies for crack tracking.
Possible reasons why cohesive cracks fail to propagate .
Common mistakes or numerical stability issues to avoid in LS-DYNA.Thank you very much for your time and support. I look forward to learning from your experience and insights.
Best regards,
Shreeya Julme
-
May 13, 2026 at 6:50 pm
Ram Gopisetti
SubscriberHi,Â
I would first recommend reviewing the existing literature on established and validated approaches for modeling sandstone, rather than combining multiple material models arbitrarily.
Since sandstone is inherently brittle, it is important to first understand which material models in LS-DYNA are designed to capture brittle behavior and appropriate failure criteria. For example, MAT_017 is commonly used for brittle materials and includes relevant failure mechanisms.
For rock and sandstone-type materials specifically, MAT_193, MAT_173, and very often MAT_198 tend to offer more suitable and robust capabilities. Based on your problem description, these models are likely to provide better representation of the material response for your application.
Cheers,
Ram
-
- You must be logged in to reply to this topic.
-
6400
-
1906
-
1457
-
1308
-
1022
© 2026 Copyright ANSYS, Inc. All rights reserved.