Position 1: Black is one move away from castling, however with the queen and bishop battery the h6 pawn will be lost if black castles. Besides that, the exchange on e5 (1. dxe5 Nxe5 2. Nxe5 dxe5) also leaves black's bishop on g7 miserably staring at it's own pawn (if 2. ... Bxe5 then white may proceed with the gain of tempo f4 followed by Rae1 and eventually f5 crashing through when the time is right).
Position 2: It's not unreasonable to want to keep queens on since white's pieces are much better than blacks here. The reason trading queens is preferred is mainly due to the fact that white has a much better pawn structure and that white's queen really doesn't have a way to be impactful if it's tucked away on the queenside. If you have the better pawn structure, an endgame will favour you more since your pieces will have more activity and your pawns are less weak. Vice versa applies due to the fact that keeping queens on can allow for dynamic play and attacks that lessen the importance of pawn structure in some cases. When you enter an endgame, the structural weaknesses are amplified because there's less means of defending the weakness or seeking compensation for them. After trading queens, both of black's bishops are awful, both of his rooks have no prospects and the knight will take a while to maneuver to an even semi-useful square. White simply dominates on structure and activity. In this position retreating the queen puts your queen temporarily out of the game and gives black a free move to potentially create some unnecessary counterplay with Ng6 Nf4 Qg4. Shouldn't be dangerous with best play from white, but it's suboptimal since black is completely paralyzed and without play in the immediate endgame.
Position 3: It's difficult to say without knowing where it is moving from, assuming it moved from a square like c3, it would simply be undeveloping the queen and tucking it away near the corner of the board for the sole reason of attacking a pawn that can be easily defended with no way to further mount pressure considering you still need to undevelop your great knight on b5 and then maneuver the f3 knight (what you should start with) to d2 to c4 which is the maneuver you should play regardless to put pressure on b6 and get out of the way of black's possible kingside pawn storm. Only then should you consider a comital move with the queen like Qb2. Focus first on developing the pieces that are on the worst squares, only after that should you consider delegating your queen to such a task.
I just finished a game and went through the analysis, but I still have a couple of questions I hope someone can help me with.
Position 1: How am I supposed to know that dxe5 is the best move? Sure, his king is in the center, but he’s only one move away from castling.
Position 2: How can I know if trading queens here should be good or better to save it?Position 3: Why is this Qb2 not a good move? Preparing an attack on weak b6 pawn?