Most Ph.D. programs look for assurance that a Ph.D. candidate can do successful research. Normally, candidates cannot possibly display the type of creativity and grid needed for doctoral work in their previous career. Then they look for substitute measures, making sure that the candidate has at least the prerequisites for doctoral work. From what you describe, your past would raise suspicions that you cannot work successfully on a Ph.D. First, you would be expected to take classes at a high level, and maybe out of your field (e.g. Rare Event Simulation taught by the Math Department when you are studying Engineering.) Second, you need to be able to get along - and more than that - collaborate positively with others, even if your personalities do not mash. So, I would not be sanguine about your chances for admission right now. This especially, since most Engineering programs pay their Ph.D. students and people are more conservative if money is involved.
This does not mean that you have to give up on a Ph.D. For example, as a member of an admission committee, I would discount somewhat previous educational outcomes after a successful few years in industry. I would also value any type of peer-reviewed publication. Finally, if the personality mis-match is not so much you but your advisor and the advisor is well-known to be difficult, and all of this would be known to me, then of course I would discount your not getting along with your advisor. The only problem for you is that to get me to know about your advisor is pretty much impossible. However, even then I would be wondering why you stuck with a bad advisor and would be tempted to assume that the problem lay more with you than with your advisor.
In short, consider spending some years in industry. Publications would solve your problem as well, but might be out of your reach.
