switching proportional valve power supply causing vibration sensor output spikes?

Hi

We have 12x Bosch 4WRPEH 100l/min proportional valves. These valves have a 4th spool position that closes the actuator if the proportional valve power supply is turned off.

Due to the hydraulic system design we were forced to de-energize the proportional valve power supplies with the emergency stop button.

When the proportional valve power supplies are switched on or off we are observing spikes in the output of the vibration sensors (5 on each unit). This is a hydroelectric power plant with two units. The vibration sensors on both units report vibration over the trip threshold if the e-stop on either unit is pressed. This occurs even if the unit is at a stand-still and there is no oil pressure in the hydraulic system (so no chance of any actual vibration).

You can imagine the confusion when pressing the e-stop on the unit that is not running results in a trip (due to false vibration) on the other unit.

The only thing the 2 units have in common is the 24V DC power supplies are fed from the same battery bank.

I think some EMF kickback from the proportional valve linear motors or inrush current is causing a voltage fluctuation and the vibration sensors (Hansford HS-420) are overly sensitive to changes in supply voltage.

I put the fluke in min max mode on the supply to the vibration sensors and it changed by 0.2V when the e-stop was pressed.

Has anybody ever had to use a freewheeling diode on a proportional valve power supply? We do this for all the solenoids on the HPU.

Has anybody ever had an instrument that was sensitive to transients in power supply?

thanks for any thoughts.

It would seem that a diode on this would be the same sort of application as a diode on a solenoid valve. Spool or solenoid, they are still a coil with the same voltage kickback upon de-energization.
I am curious- why do you have E-Stops on your hydro units? Our utility does not have any E-Stops however most of our stuff is quite old. Is this a new application? Or maybe a code that only applies to Canada?

The problem was not isolated just to the de-energization of the proportional valves, it also occurred due to the voltage dip due to inrush when they were energized in which case the diode wouldn't help. In the end we increased the delay on all of the vibration trips to 3 seconds and I will know to watch out for the 'settling time' spec on instruments going forward.

I am mostly working on greenfield run of river hydro in the 5-50 MW range. The head is usually between 70 and 800 metres, flow anywhere from 2 to 50 m3/s and penstocks 1-8 km long, without surge tanks with one exception that I know of. The plants have 1-4 units and are primarily operated remotely.

After commissioning the E-stop button should never be used as there are stipulations in the water licenses for the plants that say the plant shall not cause the river level to change by more than 2.5 cm/hr. Pressing the e-stop button closes the nozzles or wicket gates in 60-180s and also closes the turbine inlet which takes about 5 minutes, which can cause large changes in river level downstream of the plant depending on the flows. During commissioning they will wait for a high flow event or position biologists in the creek at potential fish stranding sites and press the e-stop button in order to record the transient pressure waves in the penstock and make sure they are as expected.

A non emergency stop either maintains flow through the nozzles and controls the unit speed with the deflector, or moves the flow to other units, to a bypass valve energy dissipation system, or decreases the penstock flow at a slow rate usually between 0.2 and 2.5 m3/s/hr in order to avoid changes in the river level. Some plants can take a day to decrease penstock flow to zero.

These plants have no storage, they make as much power as there is water available in the river. The units start and stop by themselves as required to maintain a constant head pond level. The utility does not schedule this generation it ramps up and down with snow melt, rain, freshets etc.

I don't know why we have e-stops other than because we always do, and every other plant I've been to in BC has them. I can't remember if the plants in Ontario I went to when I lived there had them or not.

Check grounding, the 24 Battery should be grounded on one side,
same for the secondary side of eah power supply.
put a diode in front of every PSU to prevent the current flowing back, same as you start car the radio is dead just because the voltage is too low
estop should go to safetyrelais, to be sure they switch off.

we run floating DC systems so there are no grounds.

I've never considered a diode on the circuit feeding PLC power supplies or other DC-DC converters. can you provide some links to literature which describes when and why this is necessary?

browning out equipment can be a concern, so far I've only run in to one device which was a consumer grade wireless radio that was damaged due to low voltage.

For installations where there is no diesel generator to keep the battery chargers on when there is an outage we use an undervoltage relay that disconnects all the load from the DC system when the voltage gets below 22 and then doesn't re-connect the load until the voltage is above 24.

the devices switched by the e-stop are wired directly to the e-stop contacts not through relays.

thanks for your thoughts shooter

I thought about this and while we don't have traditional e-stops, we do have a big pistol grip knob that is the E-Stop on our plants (labeled 86E if you are following the IEEE relay numbering common in utilities). This stops the unit as quickly as safely possible, but will leave the governor, etc powered up. A couple things can trigger this- burst penstock, low penstock pressure, remote from the control center, and a few other catastrophic triggers like overspeed and certain vibration thresholds on some plants. However we leave everything on and ready for the next command (breakers, governor system, hydraulic HPU's). Interesting to see the difference in plants. By the way our licenses usually let us get away with some changes in water levels due to tripping however things are getting tougher as we relicense, we are faced with very specific and seasonal ramp rates that require new governor controls sometimes. If there is a PRV (Penstock Pressure Relief, for plants with no surge chamber), or bypass valve, it gets even more fun!

Note that we do have a delay on our vibration trips, likely to filter out noise and the like. If it takes 3 minutes for your plant to spin down (as is common with medium and large Francis units), a couple second delay on a vibration trip means squat.

Maybe shooter meant a set of diodes, each fed from different power sources kind of like a "best battery" arrangement that some generators have. You get burned by the 0.6V drop across the diode however whatever source with the most voltage takes most of the load. I would also check the transient performance of your DC_DC convertors, maybe they are not imune to dips in input voltage. One idea would the the ride through capacitors that Phoenix Contact, etc make that supposedly help you out with transients under a few seconds.

We also use two pistol grip lockout relays for each unit. One is for electrical problems (86E) and is usually tripped by the generator protection relay and one is for mechanical problems (86M) and is usually tripped by the PLC or overspeed relay. When the 86E is tripped the PLC will maintain the flow through the unit or slowly ramp the flow to zero over the course of hours but the turbine rotation stops in minutes. When the 86M trips the flow through the unit is stopped as fast as possible (15-120 seconds). An example 86E would be a differential trip. an example 86M would be loss of HPU pressure.

We have provided control systems for half a dozen plants with various arrangements of pressure reducing valves, bypass valves, and energy dissipation systems. It does get interesting. The specific ramp rates are probably what keeps me employed otherwise an off the shelf speed controller would probably do just fine!