Games from a 45+5 Tournament

Hi, I'm wondering if anyone could possibly provide comments and criticism on a few games I played in a tournament recently.

The first game I played white: lichess.org/NtyxFyns/white/ It was a standard Slav Defense; I didn't feel like I had any disadvantage in the opening. However, after I played the seemingly-natural move 18. Rac1 and black played ...Nb6, I thought for a long time on this move and just couldn't find a comfortable plan. I ended up moving the rook back to b1 because I wanted to do a minority attack and couldn't get in a central break.
I played the move 21.b4-b5? rather quickly as it looked good at first glance, but the computer ridicules this move. After a few more moves, he got a passed pawn, I didn't like my position at all and ended up falling into the trap of completely reacting to my opponent's moves, being pushed onto the defensive and lost the game. My two main questions for this game are: 1. What kind of plan should white try to achieve in a quiet position like the one at move 17? 2. At move 27, is white forced on the defensive and completely lost, or does white have any offensive juice left?

-

The other two games can be found Chapters 1 and 4 of this study:
They were both Caro Kann: Classical Variations. I have provided my thought process in the comments of the chapters. I felt pretty comfortable in those games, and I'm just curious as to if there are any different ideas I could have done.

Thanks and Happy Checkmating!

@FunnyAnimatorJimTV #1
Starting with a very quick comment. In your first game 10.Bd3 is imho a clear mistake. One : Trading your attacking bishop [1] (and at the same time your good bishop looking at the colors of the pawn structure) for no apparent reason. Two : You have the move 10.Nf3-e5 for free, threatening to win a pawn with Nxg6, fxg6 Bxe6. After e.g. 10...Bg6-h7 you can play Qd1-b3 threatening to take on b7 and maybe also on e6.
Furthermore the plan with b2b4 turned out to be not as promising as it might have looked like. Your pieces were still more in the center than on the queenside, and blacks pieces were surprisingly solid. Instead of 15.b2b4 you could have played 15.Be3 (15...Qxb2?? 16.Rdb1 1-0) followed by Rd1-d2 or Qd3-d2, and then look at options to play d4d5 or e4e5 or even g4g5 (The h7h6 move makes g2g4 and g4g5 more attractive).

Quoting Siegbert Tarrasch about "the attacking bishop" :
"As Rousseau could not compose without his cat beside him, so I cannot play chess without my king's bishop. In its absense the game to me is lifeless and void. The vitalizing factor is missing, and I can devise no plan of attack."

The analysis says that you were going to get mated in 12 moves

Agree with @achja.
Your problems start with the obscure 7 a3. Normal moves are 7 Qb3, 7 g4, 7 Be2 intending to punish the bishop's sortie by attacking b7 or chasing it with gain of space. 7 a3 does nothing useful and thus is a loss of time. Nothing bad per se, but it starts a string of weak moves.
8 Bd3 loses a tempo, nothing bad per se, but loss of time anyway.
10 Bd3 loses another tempo. The trade is undesirable as pointed out in #2, but playing the bishop f1-d3xc4-d3 is also a considerable loss of time. 18 Rac1 and 19 Rb1 is again loss of time. Your intended minority attack is not the right plan here, but even if it were the execution with loss of time is doomed to fail.
You should consider your position more thoroughly and plan ahead where you want your pieces in the long term. By useless moves 7 a3, Bf1-d3xc4-d3, Ra1-c1-b1 you allow your opponent to develop his position.