The straight wing, all things being equal, should produce less drag since it should have a smaller wetted surface area when compared to the curved wing.
As far as how you altered the curved wing to match the lift, that sounds very iffy. I think it would be better in general to just change the sweep profile of the exact same airfoil/wing section and then compare L/D.
I don't see why to curve the wing for any reason but down wash/messy air caused by the air box. I think the better experiment would lie along those lines. Obviously there will be a lot of variability so you could only hope to find general trends.
One interesting aspect is that in aircraft, the trailing vortices create down wash which effectively reduces the wings effective angle of attack. You could intentionally induce a down wash upstream of the central section of the wing and mess around with twist or washout/wash-in.
Also, if you try standard airfoils with smaller angles of attack and less extreme cambers your simulations will likely be faster and more reliable since you will be dealing with less flow sep. In doing this you can actually learn about the lift slope and changes in the sweep profile, or more or less dihedral. Seeing actual trends in data is much more useful than nice colorful post processing, and honestly more gratifying once you learn what you're doing.
Apologies to all Engineers that may be facepalming over all the over simplification.