I really enjoyed Brunel - it was actually my first choice anyway, made on the basis of its good rep for science and engineering and the bonus was that it was within reasonable commuting distance from the then Casa Reynard. I ended up there for five years all told, and, for the most part, loved every minute. There were a few modules on my course I didn't enjoy, but they were mandatory for IMechE accreditation, so I just had to stick through it.
I didn't do the wrong A-levels (Pure & Applied Maths, Chemistry, Physics), but I was going through a tough time mentally and I didn't get the support I needed - but that's another story. Ergo I flubbed them badly because my heart wasn't in it, but got offered the place on the foundation year on the strength of the interview and personal statement. Also, I didn't get the right careers / further education advice at school - traditional girls public school, and girls don't go off to do mechanical engineering. Hey ho! I ended up with a 2:1 for my BEng, missed a first by a gnat's todger due to my fannying around on a management module and scraping a pass on the exam...
Mech Eng & Automotive Design here too.
Went in with designs of being a race engineer, but found that dynamics and fluid mechanics weren't nearly as much my cup of tea as the stress analysis and materials tech side of things. The electrical modules we did were ok, I hated CAD and 3D CAD as I much preferred doing that the traditional way.
Robotics and integrated manufacture weren't offered on the Mech Eng courses, nor were the more in-depth materials modules, which is why I stayed on to do the Masters. For my sins (AKA masters project), I ended up developing a full-chassis computer model for Reynard Cars to do stress analysis with - at that time, it was still very much a theoretical thing for them, so I proved it was possible, and with commercially available software to boot (ANSYS).
My PhD was in F1 crash testing, concentrating on materials selection for side impacts. I designed and developed a small-scale testing rig for one of the FIA crash test procedures that could be used on an Instron in a standard testing lab, and then proved it could provide valuable information to reduce the number of iterations needed during full scale testing. The full-scale testing needs to be done in specialist places like MIRA as opposed to being done in-house, so it's expensive and time consuming and not conducive to real experimentation. I should add, that having my favourite racing driver die in a preventable accident, driver safety is something that really matters to me.
I did end up in motorsport after all that, but as a photographer and journalist.