To my ear, "a manufacturing process for high-precision lenses" sounds every bit as acceptable as "a process for manufacturing high-precision lenses." So now, having given my subjective answer to your first question, I turn to the question of change in the phrases' frequency of usage.
Here is an approximation of the Ngram chart that you refer to in your question, matching "manufacturing process for" (blue line) against "process for manufacturing" (red line) and "manufacturing process of" (green line) for the period 1800–2005:
Even with a smoothing factor of three years, it's a pretty jumpy-looking chart. Now let's look at the same results but with a fourth line added for "process of manufacturing" (yellow line):
Compared to the mountain range of "process of manufacturing," the other three plotted lines seem to have undergone rather nondramatic changes. There is no obvious correlation of the changes either, except between "manufacturing process for" and "manufacturing process of," which seem to have been followed a generally similar trajectory since 1900 and especially since about 1975.
The period of sharpest increase for "process of manufacturing" occurred roughly between 1900 and 1920—and yet all three other phrases showed at least modest gains over the same period; then, from the late 19540s to 1980, when "process of manufacturing" went through a long and continuous decline in frequency of use, "process for manufacturing" fell only slightly and the "manufacturing process of/for" lines grew steadily but (relatively) modestly. I don't see what correlation one can draw between the fortunes of the yellow elephant in the room and its red, blue, and green juniors.
Between 1980 and 1990, "manufacturing process for" continued its ascent at essentially the same rate as during the previous dozen years, and "manufacturing process of" joined it after a period of relatively flat performance; but beginning in the early 1990s, both of those lines leveled off again. Since 1920, "process for manufacturing" has been the steadiest line in terms of end to end frequency of use.
My conclusion is that the Ngram data doesn't support the existence of any clear connection between the rise in frequency of "manufacturing process for" and "manufacturing process of" (on the one hand) and the either the decline of "process for manufacturing" or the mostly level performance of "process of manufacturing" (on the other). That's not to say that no connection exists, but I wouldn't argue that the line graphs tend to show one.
A stronger (though not ironclad) case might be made that the frequency of use of "process of manufacturing" was, in the long run, adversely affected by the rise in popularity of the form "manufacturing process." Look at what happens when we add "manufacturing process" (lavender line) to the previous Ngram chart:
The frequency line for "manufacturing process" takes off around 1900 and grows so large that even the changes in the line for "process of manufacturing" seem rather modest. The lesson here may be that the importance of scale in Ngram charts should not be underestimated.
As to why the phrases "manufacturing process for" and "manufacturing process of" have become so much more common on recent decades than they were a century ago, I have no hypothesis to offer beyond the fairly obvious one that they benefited from emergence of the "manufacturing process" behemoth over the course of the twentieth century.
In general, phrases that were not popular in the past sometimes become popular later, as they become familiar to more people and cease to strike the ears of hearers as odd-sounding. The first and second Ngram charts above suggests that "manufacturing process for" and "manufacturing process of" made a gradual transition from unusual to mainstream together—but the third Ngram chart suggests that they did so, as it were, as barnacles on the back of the larger phenomenon of the rise of "manufacturing process."
