Where does the modulation to F happen?

If you don't know already, I am writing a fantasia in the key of F major. I already posted about it here:

Well, I have made some progress with the Fantasia. I have another section in place. Now, I might change some of those triplet arpeggios into sixteenth note arpeggios or eighth note arpeggios, but I have an arpeggiated bass line. I'm wanting to gradually harmonize the melody and make it richer to make it sound like the river is getting wider. But, in order to do that, I need to know where the modulation back to F major occurs(because, like I said in the previous post, I am reserving dissonance for those moments where I go into a minor key). I know I am in C major for a while before I modulate back to F major(Settling the drama of the C minor with a Picardy Third followed by a passage in C major). But, I have no idea where I actually modulate back to F major. That is a bit of a problem. Because what if I by accident put a Bb over a B natural or vice versa(those G chords do change in sonority from major to minor)? That is going to lead to a major seventh interval in there and an urge to resolve. As an example of what I mean by the major seventh having an urge to resolve, whenever I hear a CM7 chord, my brain is always like "This chord so wants to go to C major, it is so close, just raise that B and it will feel right" or in other words, it hears a dominant function from a "tonic" chord.

Anyway, back to my question relating to modulation. There are a few places where I can hear a tendency towards F major, with the first one being more subtle than the others. Here they are:

Quote

Bar 44: Change in the sonority of the G chord from a major chord to a minor chord by introducing back Bb.

Bar 46: Leap from C to F, in both the bass and the melody, suggests that it is modulating back to F

Bar 47: Bb major chord in the bass really suggests the key of F major

Bar 51: Cadence in F major

But then again, between bars 38 and 51, C major feels pretty settled, like it is the tonic, and it is only after bar 51 that I am sure it is acting as a dominant chord in the key of F major. However, I'm pretty sure that bar 51 is not the point of modulation back to F. I'm pretty sure that the modulation happens before that cadence in F and is only confirmed by the cadence. But then, if the modulation happens before bar 51, why would C major feel so settled before bar 51?

You see what I mean? On the one hand, there are things that push the music towards the key of F major, like the Bb major chord at bar 47. If I were to still describe C major as being the tonic at bar 47, the Bb major chord would be a secondary subdominant chord. But on the other hand, until that cadence in F, C major feels as settled as can be, like it is the tonic chord. When I am wanting to make the melody richer, this could potentially become disastrous, this uncertainty of where the modulation happens because C major feels settled until the cadence in F. Here is what I have now of the Fantasia in F(C major section that modulates back to F starts at 1:26 in the audio and bar 38 in the PDF)

To my ears, beat 4 of m. 44 has a G minor chord (ii in F), m. 45 is C (V in F), and m. 46 is F's arrival. 

From m. 38 to 43 it seems in Cmaj: C - Dm - C - Dm - C... C - G (not Gm in m. 44) - C (m. 43).... This is all in Cmaj, the two following measures rest again in Cmaj. Notice in the melody that C is the longest note in this part.

The Bb appears in m. 47 in a Bb chord, and in a Gm in m. 50 and 51. Perhaps here there is the true modulation with Gm - Bb - C7 - C7 / C7 - F in m. 52. From here it seems clear we are in F maj but notice the cadences or resolutions in F are in weak moments of the measurese. The G7 in m. 58 is a V7/V7.... perefect cadence in m. 62.

It sounds nice. Only if you want to have a strongers F tonality from the beginning of the part, establish it with C7-F or Gm-C7-F.....