While the CDS crew had systems down yesterday, it appears the HEPI Pump Servo Pressure sensors were affected and changed their reading.  This happened before in May 2015.  SImilar to then, while the loads on the building changed, the voltage offsets changed on the pressure transducers and they stepped their value.

The attached trends shows the three building's HEPI Differential Pressures and their servo output drive to the VFD.  You can tell that it is the pressure that is affected as it goes high the controller drive drops to lower the pressure back to the setpoint.  When all the system loads are brought back on line, the pressures drop lower and the drive recovers to put the pressures back.  The pressures come back to where they were as evidenced on the CONTROL VOUT.  At EndY, I had put the servo into manual and the pressure only recovered to 65psi.  This morning, I put the servo back in auto and now the drive is back to pre event level.

When the recovery happens and the pressures drop, the servo will drive hard to get the pressure back up and this can be enough to trigger the fluid level shutdown.  This happened yesterday in the corner as evidenced by the pressure drop to zero for a while.  I expect the increase drive on the pump literally pulls fluid out of the reservoir faster than it  can be replenished from the return flow (until the system reaches the new equilibrium) and trips the level switch.  I set these pretty close to nominal running state so as to not make as big a mess if there is a bad system leak.  Other than resetting the corner level switch and reenergizing the controllers, the servos kept right on going through everything.

Barker and I will try to design an experiment as to what or how many 'power offs' cause this.  The solution is pretty simple though.  When this sort of activity is going to occur, just put the servos into manual mode.  We'll get alarms when the pressure changes but we'll know they aren't real.