On the Pump Station output line below the reservoir is found a pressure relief valve.  This valve is factory set to 125psi but this looks very coarse.  The output from the valve is plumbed to a small drum.  I found this clear line full of fluid.  This suggests to me there have either been several pressure spikes on pump station restart--not good; or, a slowly leaking valve--less bad but still not ideal.  I've drained the line as best I can and marked the hose for monitoring.

The pressure spikes upon restart are caused by poor operation (not following the restart guidance: https://cdswiki.ligo-wa.caltech.edu/wiki/SEI) and should be avoided to limit fluid loss and disposal pain down the road.  Plus, the system will very likely have not successfully restarted if the best practice has not been followed.

The attached plot shows the pump drive and the output pressure closest to the relief valve during the July 29 OU3 fault.  The bottom plot is zoomed into the pressure during restart attempts.  The bottom line to remember when operating this system is to manually zero the PID output before pressing the red button or hitting the fault reset.  The PID loop knows nothing about the VFD (maybe this will change with potential Beckoff upgrade.)  It knows only that the process variable differential pressure is not at the process setpoint.  So the PID increases the output to max.  In this state, as shown in this trends, the reset button was pushed with the VOUT at max and as a result the pressure spikes to more than 100psi very quickly.  Whether this spike opens the relief valve is unknown--these EPICS trends could easily not show the highest pressure, the relief valve and this pressure gauge are likely not tightly calibrated.  Couple things could happen with this spike occurs: the relief valve opens and spills fluid; and/or, the fluid level in the reservoir is pulled down too quickly and the pump station trips.  Obviously something like this happened twice here.  Once the VOUT was zero'd, restart proceeded nicely.

Again, bottom line, Operators--If the HEPI Fluid  Pressure is not okay, manually reduce the output of the PID before pushing any hardware buttons.