1 ACCEPTED SOLUTION
Accepted Solutions
I will try to put the basic steps.
Dynamic time analysis is a linear vibration analysis. It uses the stiffness and mass matrices from a linear modal analysis and then adds damping assumptions. This technique is a much simpler and faster computation than non-linear explicit dynamic analysis. It is up to the analyst to determine whether this with its linear assumption is an appropriate technique.
1. Define a modal analysis
a) constrain the model (unless doing free-free study)
b) set number of modes or frequency range. For dynamic time I suggest at least the first 10 modes.
2. If doing "load functions" add a unit load to your model in the location you need the dynamic load. Base excitation is loaded through the constraints like a fixture placed on a dynamic shaker machine.
3. Define a Dynamic Time analysis
a) reference the previously defined modal analysis (can choose to run on demand or a pre-run study)
b) Choose the load type: Load Function / Base excitation, the acceleration direction or the previously defined load-set.
c) Choose or make the time dependence function. (default is impulse, which returns the transfer function of the model)
d) Choose output steps. Full results can be generated at any or all master steps. Measures can be charted at a higher resolution according to the output per step. For example if your time function ends at 1 second and you leave the default single step, putting 100 as the output per step would have measures every 0.01s. If your time function is 10 seconds long and you have 10 master steps and 10 outputs per step, the measures would be generated every 0.1 seconds.
e) Make your desired measures, ensuring they are valid for Dynamic Time Analysis.
f) Add damping assumptions. A common assumption is 3% which is considered minimal damping.
4. Run the analysis.
5. View full results and/or chart result measures.
I will try to put the basic steps.
Dynamic time analysis is a linear vibration analysis. It uses the stiffness and mass matrices from a linear modal analysis and then adds damping assumptions. This technique is a much simpler and faster computation than non-linear explicit dynamic analysis. It is up to the analyst to determine whether this with its linear assumption is an appropriate technique.
1. Define a modal analysis
a) constrain the model (unless doing free-free study)
b) set number of modes or frequency range. For dynamic time I suggest at least the first 10 modes.
2. If doing "load functions" add a unit load to your model in the location you need the dynamic load. Base excitation is loaded through the constraints like a fixture placed on a dynamic shaker machine.
3. Define a Dynamic Time analysis
a) reference the previously defined modal analysis (can choose to run on demand or a pre-run study)
b) Choose the load type: Load Function / Base excitation, the acceleration direction or the previously defined load-set.
c) Choose or make the time dependence function. (default is impulse, which returns the transfer function of the model)
d) Choose output steps. Full results can be generated at any or all master steps. Measures can be charted at a higher resolution according to the output per step. For example if your time function ends at 1 second and you leave the default single step, putting 100 as the output per step would have measures every 0.01s. If your time function is 10 seconds long and you have 10 master steps and 10 outputs per step, the measures would be generated every 0.1 seconds.
e) Make your desired measures, ensuring they are valid for Dynamic Time Analysis.
f) Add damping assumptions. A common assumption is 3% which is considered minimal damping.
4. Run the analysis.
5. View full results and/or chart result measures.
The time unit is by your unit system for which all of the built-in units are in seconds.
But, a custom unit system could be made, for example the common millisecond time base for dynamic analysis.
As far as acceleration vs time. Can you screenshot your measure definition and also verify that the measurement point will be on an element of the mesh? Did you get some full result steps and do those look correct?
