'boxTocell' didn't work when I'm using 'DAObjFuncVariance', is there anything wrong?

[wgtfly]

Hello professor, when I use 'DAObjFuncVariance' to do my own optimization, the optimization can proceed smoothly after changing the objective function, but I found that the program does not seem to really calculate in the box I set, because when I change the value of my box objective function and the optimization process are exactly the same, even if I set it outside the grid.
I wonder if there is something wrong with this function, or if it always defaults to calculating according to the value of the entire field?
If I want to use the box of a certain area to caculate through part of the volumemesh, can I use boxtocell?

Here is my objfunc:

        "UFieldVar": {
           "type": "variance",
           "source": "boxToCell",
           "min": [5.0, -1.0, -10.0],
           "max": [8.0, 1.0, 10.0],
           "scale": 0.1,
           "mode": "field",
           "varName": "U",
           "varType": "vector",
           "components": [0, 1],
           "timeDependentRefData": False,
           "addToAdjoint": True,
        },

and the log files:
opt_IPOPT.txt
logOpt.txt

I just want to do the variance caculation through part of the mesh. Would you help me see what the problem is?

[friedenhe]

That is right, the variance function does not use boxToCell at all.. If you set up your function like the above, it will just use all cells from the U field.. If you really need boxToCell, you have to implement into DAObjFunVariance.C. Or, you can use the probePoint option to prescribe discrete point as your refData. On a separate note, we are about to release the v4 version of DAFoam, we will add the boxToCell feature to variance. Stay tuned.