I am using DSPI (Q-100 Dantec Dynamics) to validate an FE model of bone. I get similar strain patterns with both methods, but the strain values differ by several orders of magnitude. Values from DSPI are within 1 microstrain of range. Could these values be within DSPI error and not be reliable to perform the validation?
The differences might be due to difference in units between FE model and DSPI. Checking for the units used for different material properties in FE model would be a first starting point.
Yes, I agree with Mr. Kadapa. This is the first point you shoud check. Difference between MPa and GPa is usually the most common error source.
Please check the elastic modulus values. Check the bone density values also. It depends on the quality of bone classification also.
Make sure all boundary conditions are propositional and similar. The result should not be difference significantly.
Dear all, thanks for your replies. So far:
- I varied the elasticity modulus of the FE model, in ranges from 15 to 90 GPA. Resulting strains didn't varied much.
- I varied the number of nodes of constraints and the axes of constraints. not much differences were found.
- I cannot run anisotropic models of generate more complex models because the software is not prepared to do it, but colleagues have validated other models of bone and they didn't have problems related to this.
- I added an additional way of validation, which consisted on a force transducer that measured one of the reaction forces. This was the parameter that changed the most when varying boundary conditions. Icould obtain values very close to the experimental.
- The FE software asks for the following inputs: element size in meters (my elements are cubic in average of 0.4 mm per side=0.0004m). Young modulus in Pascal (mine are ca. 17 GPa=17x10E9 Pa). Poison's, in my case 0.3. Applied force in Newtons (in my case, two hanging loads of 4 kilos each=40 N). According to software instructions and the work of other colleagues, using these inputs should give strain values in strains (instead of microtrains). The results are in the order of 2x10E-5, and my Dantec Results, in the order of 0.2 microstrains (if I am correct, 2x10E-7).
I thank any help you can give me, particularly if you can spot something I am missing.
Viviana
... I would like to add that the units from the DSPI device, are (μm/mm). Any thoughts?
Hi Viviana,
I guess the problem is with the way you use powers of 10. For example, Young's modulus is 17GPa which is = 17 * 10^9 N/m^2. But if you input it as 17*10E9 then it becomes 17*10^10. So, actually the stiffness is higher by a factor of 10 which must be reflecting in lower strain values.
Strain from DSPI is 0.2 μm/mm that is = 0.2 * 10^(-3) m/m = 2* 10^(-4) m/m and from FE you are getting 2*10^(-5), which is lower from that of DSPI by a factor of 10. So, I think, if you input the correct value of Young's modulus, stains should match.
Apart from this, I don't see any other thing that could go wrong, based on the information provided. Adjust the Young's modulus and perform the analysis once again and check for the strain. Let me know about the outcome.
Regards,
Chenna.
Dear all, thanks a lot for your help. Indeed, as suggested by many of you, and revisiting units of DSPI as Mr. Kadapa suggested, I found comparable values in DSPI and FEA.
Best wishes from UK,
Viviana
- Badges
- Science method