Incorrect variable name in mean

[trevilo]

In the time average output, when there are three velocity components (i.e., 3d or axisymmetric), the variable called meanSolution/mean-E is actually the pressure. This was originally fixed on #48 but was never merged to avoid breaking ongoing runs (since the variable names would change upon restart). As this is still a problem but #48 is very stale, I am closing that PR and opening this issue to remind us to fix it once the ongoing flow-only runs are done and before we start production plasma simulations.