Is there any opening for a win for white in this

Any possible win for white?

No. Black was better for most of the game.

Moves 17-20 lead to a small advantage to black from the advanced pawns. Any possible opportunity would lie in 17-20 since it defines the pawn structure. After these moves, draw becomes quite certain.

18. Bg3/Bg5 is a possibility with the former still reeking of a draw. The latter can potentially open the f-file and so more opportunities.

I actually might take white at the end of move 16. White can try Bg5/Bg3 and then try to swing a rook onto the d file trying to push the d pawn and either lock up Black's fianchettoed bishop or gain space and undouble the c pawns.

4.Bf4?! =

4.Qc2 Lb4 e3 Ne4 Nge2 +=

4.e3 +=

I tried those variations but almost all lead to draws.

It is difficult to comment as early as move 4. 4 Bf4 seems weaker than 4 e3 or 4 Qc2 or 4 Nf3. More promising seems 3 g3 instead of 3 Nc3, so as to avoid the pin ...Bb4 and the resulting loss of control over the vital central square e4.

@tpr, yes, that is also a good move. All these try to solve the positional threat Bb4xc3 bxc3 (known from the Nimzo) in a different way. That is why Bf4 is not optimal, it allows black to do so.

While we are at it, 2.e4 prevents 2...f5 right away.

The French defence 1 e4 e6 2 d4 d5 is usually held in higher esteem for black than the Dutch defence 1 d4 f5 2 c4 e6, so theoretically it does not make much sense to play 1 d4 e6 2 e4. Also from a practical point of view: a player who opens 1 d4 is not usually prepared to meet the French defence.

The issue was discussed at length during the Botvinnik-Bronstein match for the World Championship in 1951. The 1st 9th game went 1 d4 e6 2 c4 f5 3 g3. The 3rd game 1 d4 e6 2 e4 The 10th game 1 d4 e6 2 c4 f5 3 Nc3 Bb4 4 Qc2 Nf6 5 e3 the 12th game 1 d4 e6 2 c4 f5 3 e3 Nf6 4 Nc3 d5

@tpr valid point. I am just suggesting this as the OP said in #6 that he tried all this but all 'lead to draw'.