Question about selection of objects for lift monitor

I have studied the Symscape tutorials about generating drag and lift monitors but still have questions about creating a lift monitor.

The first times I created and used a lift monitor, I got lift measurements about a magnitude greater than I would have expected. I was suspect of the lift reported from my simulation, which was a “pie-slice” fraction of a full circular simulation; I made it that way for more manageable computing requirements as the object of study is axisymmetric with both radial and axial flow relative to its axis of rotation. I have since created a full circular simulation like your simulation of the Tower Bridge where the model of the bridge is subtracted from a giant cylinder, leaving a single "solid" flow volume. I likewise created a cylinder, then Boolean subtracted my object of study from it.

The cylinder’s bounding faces are all inlet-outlets with no other flow specifics assigned (though I should probably make them full slip-surfaces as well). Each inlet-outlet has a coincidental seed face. The subtracted object of my study has 2 inlets and 13 outlets with fixed flow velocities. Both inlets have coincidental seed faces.

The volume meshes and the solver runs fine. Around 1,200 iterations it settles down with all residuals around 1e-5 and showing some ongoing minimal oscillatory behavior thereafter for hundreds more iterations without any apparent divergence. I understand that this does not constitute ideal convergence, though the streamlines do not move to a degree that would be at all relevant in a real-world example of my model. Their movement is really not even perceptible.

My question is: What should I select for determining lift of the object in the center of the flow cylinder?
1) Do I select the entire volume (cylinder plus subtracted object)?
2) Do I select all the faces of only the object of my study that "floats" in the cylinder (but none that bound the cylinder itself?)
3) Do I select faces that that do not represent physical surfaces in the real world, such as inlets/outlets, even though they may impact lift? If so which ones? All of them, or only ones to which an inlet/outlet velocity has been assigned?
4) Do I include particle seed faces, or do they even matter since they are not walls and have no physics assignments besides their accuracy settings to specify the number of streamlines?
5) Do I simply select faces representing the intended lifting surfaces of that object? (This seems too limited.)

To me, option #2) seemed to make the most sense; it seems then I would have the most accurate results in case certain faces are creating negative lift. I created a group by selecting all the faces of the whole volume and de-selecting the top, bottom, and curved wall of the cylinder that bound the entire volume.

I have tried all three of these options, however, and though I can get other kinds of results (color-mapped pressure maps and streamlines, etc.) I now either get lift as zero or only a very few pounds; other calculations and the pressure mapping indicates that I should be getting several thousand pounds of lift.

The lift should be in the + Z-axis direction and I have the force vector variable properties set to that direction. I have the lift monitor itself set and indicating "Force" along its Y-axis, and cannot discern what I am doing wrong.

Any thoughts? Thanks much in advance for your time and comments.