Examples paper

[artofscience]

Which examples will we feature in the paper?

To show the principle we can use 1D and/or 2D problems.

To make a strong point the VdP beam would be perfect.

IMO to make a good impression we should at least include the following topology optimization benchmark problems, and show improvement on those:

1. MBB beam (compliance with volume constraint) -> implemented already
2. L-beam (volume with stress constraints) -> @artofscience
3. Force inverter (Uout with volume constraint) -> @artofscience
4. Eigenfrequency benchmark problem (max fundamental eigenfreq, volume constraint) -> @aatmdelissen
5. Thermo-mech benchmark?
6. ... your thoughts.

What problems are generally characterized by slow/problematic convergence (apart from stress)?

[MaxvdKolk]

Maybe a problem that is subjected to gravity loading? In some sense that is similar to thermo-mech problem and might also have difficulties to converge? For instance one of those bridge problems?

[aatmdelissen]

Next to an eigenfrequency optimization, which optimizes quite well with MMA, we could do a dynamic compliance example:

This is a very hard problem to converge, because of the very sharp minimum. Maybe an alternative intervening variable might help here?

[artofscience]

1. MBB g0 = compliance, g1 = volume
2. L-beam g0 = volume, g1..N = stress
3. Force inverter g0 = Uout, g1 = volume
4. Eigenfrequency g0 = max (min w0), g1 = volume?
5. Bridge g0 = compliance g1?
6. Dynamic compliance g0 = compliance at omega_x ? g1 = volume?

[artofscience]

Small update:

1. compliance -> check
2. stress -> check
3. mechanism -> check
4. eigenvalues -> check
5. dynamic compliance -> TODO
6. bridge -> TODO

[artofscience]

Small update:

compliance -> check
stress -> check
mechanism -> check
eigenvalues -> check
dynamic compliance -> TODO
bridge -> check

thermomech -> check

@aatmdelissen can you look at dynamic compliance? (or tell me how this works/ how this can be implemented?)