Yes, even if they can prove violations were rampant, does it get any votes thrown out?
Doubtful.
The goal in a state like Michigan is to convince Republican state legislators to vote for the Trump electors.
The Republicans control the state house and the state senate, and under Article 2 Section 1 of the Constitution and the Twelfth Amendment _they_ have total discretion and are the _sole_ final authority on the Michigan electors.
In 2016, because of all the mess, the city of Detroit could not do a recount and the state was forced to stay with the results because there was no change in the winner... Trump.
In my opinion, the count cannot be certified unless the machines that count the votes are open to all scrutiny to the public, and are watched closely. That we let private companies sell machines to count the ballots using closed-source proprietary software that nobody can inspect is a major civil rights violation.
In the video I posted above, these professors explain that voting machines have built-in algorithms that will automatically switch votes from one to the other candidate based on thresholds. And if you watch 10 or so minutes you will see that the greater the GOP registration in a certain precinct, the LESS they were going for Trump. It is not impossible, but it is not logical and it should not be something you see repeat over and over again across 100s of different precincts. Some by as many as 25% or so - the greater the GOP registration the less likely they were going for Trump? Makes no sense at all.