I was just wondering whether there is a way to sequentially feed the Nord Modular with the results of some equations...?

Here's an equation I did.

Sample & hold and mixer modules take care of addition very easily. Morph 1 changes the value of x and knobs 3 and 6 change the lower and upper limits.

I am a huge fan of "automatic music" ... music that follows specific guidelines to create itself, ever-evolving, ever-changing, etc. I'm also an utter newbie at creating nord modular patches. Hence my question: what are my options for creating random event-firing engines? These engines would be used to modify, on the fly, various 'n sundry patch parameters: pitch, duration filter cutoff/resonance, volume, density of notes, etc.

If you like algorithmic composing you should check SoftStep: http://geneticmusic.com/software/

As far as using the NM to create random events, modulations, etc. There are literally thousands of ways to do this.

If you need total and exact control of e.g. probability curves and change them over time within precisely define borders, nothing beats a normal computer language like C or Basic. But that means having to program everything upfront and minimal user interaction when the process runs.

The NM doesn't offer that very exact control over the process, but interaction is unsurpassed. There are several modules / module combinations that can generate almost all types of probability curves to be used to generate those parameter curves or clusters of events. Imho a big advantage of the NM is that the modulas that generate random values work independently for every voice. So making a two voice automated patch can be regarded as two processes running at the same time instead of merely a second transposed version of the process for the first voice, if somewhere one of those modules is used. Like always there is some drawbacks, if the process has a feedback of some generated parameter you have to run that through a S&H module clocked at e.g. 1kHz, as the calculation order of the modules is uncertain and there is some latency when running a parameter through everal processing modules before it feeds back. The S&H at the feedback input point lets the modules "settle", and then you get the expected curves, although now at a samplerate of 1kHz.