droop setting of governors 2

[RAgrawal]

I would appreciate, if any one can explain me exact meaning of droop in governor and how it is set along with setting range.

 

[xnuke]

Here's some basics:

Droop allows generators operating in parallel to share load in proportion to their capacity. Voltage droop is used to provide reactive load sharing (kVAR), and frequency (speed) droop is used to provide real load (kW) sharing.

The concept of voltage droop: from a no-load voltage condition, as reactive load is added the terminal voltage of the generator lowers due to the change in armature reaction. A voltage regulator causes the terminal voltage to lower linearly by controlling the rotor excitation. Voltage droop is usually expressed as the percent change in voltage from no-load terminal voltage as the reactive load varies over the full reactive load range.

The concept of frequency or speed droop: from a no-load frequency condition, as real load is added the the prime mover slows and the frequency lowers. A speed regulator causes the frequency to lower linearly by controlling the prime mover speed (governing). Frequency droop is usually expressed as the percent change in frequency from no-load frequency as the real load varies over the full load range.

The generators I've worked with usually had less than 5% droop, and were fairly linear over the full load ranges.

Regulation allows load sharing by: (assume regulators for G1 and G2 are set equally as an initial condition)

As G1's voltage regulator is raised and G2's is lowered, G1 picks up more of the system's reactive load and G2 sheds reactive load while keeping the system voltage constant.

As G1's speed regulator is raised and G2's is lowered, G1 picks up more of the system's real load and G2 sheds real load while keeping the system frequency constant.

The load should be shared in proportion to the capacity of the generators. If G1's power rating is twice G2's, G1 should carry two-thirds of the system's real and reactive load. This minimizes heating effects in each machine.

 

[TheBlacksmith]

xnuke,

Good explanation of droop, however, a good electrical operator or automatic system will manipulate the speed and voltage settings of both machines in parallel to maintain 60 hertz and rated output voltage on the buss. While the methodology of droop is to affect the machine's speed and voltage, the big picture is that the prime mover's (diesel, turbine, etc) power output and the field excitation are being adjusted proportionally to share the loads betwwen two connected machines.

Been there, done that.

Blacksmith

 

[hitu]

DROOP is the % of variation (increase) in speed is allowed, if the full load (base load) of the turbine is thrown off immediately(by opening its main breaker)

 

[sailer12000]

I like to take this question a step further and ask how you would actually program the logic for an electro-hydraulic governor which is controlled by a PLC. Basically the PLC "knows" generator frequency & output, turbine speed, final control element position (in this case a hydro wicket gate), and frequency / droop setpts.

I assume I'll need a PID block with actual speed as PV and a speed setpoint SP. One of these terms is somehow modified by the droop variable, but I am not sure which or how.

The output of the PID block would be treated as a gate position setpoint. An inner control loop would position the gate to this setpoint using a hydraulic proportional valve.

Any thoughts?

Thank you

 

[bogdal]

Hi, Sailer 12000
It's right, you will use a PID algorithm, but not a compact one. For a electro-hydraulic governor, we have a cascade loop, so the functions of the PID are distributed on the feedback reactions, depending on the regulating principle. The governor's droop can be defined either as a "power droop", when it operates on the generator's power feedback, either as an "opening droop" when the power signal from the generator is disconected and the droop is defined on the gate position. Assuming that the droop formula is (df/f0)/(dP/P0) where f0 and P0 are the rated values for the frequency and for the power, and df and dP are the variations at an arbitrary moment, that means you will take from the power signal only a part to compensate the frequency variation. The same applies for the gate position feedback. So when you write the PLC program you will multiply the power feedback with the droop value and sum the result with the frequncy error. The entire algorithm form depends on the way you design the speed governor.
Sorry for my english.

 

[sailer12000]

I am picturing a simple PID block with inputs of Speed Setpoint (SP) and actual Speed Present Value (PV). The output from this block represents a gate position setpoint which is cascaded to an inner loop which positions the gate.

I would calculate the Droop Adjustment as (%Droop) x (%Gate) so that as the Gate Position increases, I'll get a larger droop adjustment value. It is what I do next that I am confused about.

I assume that the PID block's setpoint (PV) input value should be actual speed (60HZ) and that I calculate the SP value by either adding or subtracting a value for Droop Adjustment from the desired speed setpoint (typically 60HZ).

Therefore the inputs to my PID block are as follows:

PV=actual measured speed (fixed by utility bus @60HZ)

SP=Desired Speed (60HZ) + (Max Droop Adjust - Calculated Droop Adjust)

I think I want the SP calculation as noted so that as load increases (gate %) the calc'd SP value gets smaller and approaches 60HZ, tending to balance out the PID equation. This way I am always trying to drive the unit towards full load (100% gate). But if all the connected generators produce to much power and the bus frequency speeds up, the unit will start to shut the gate as PV increases.

Comments and suggestions are greatly appreciated!

 

[bogdal]

The speed governor actually makes the frequency error (frequency setpoint - actual frequency) to become zero at no load functioning. After the generator is coupled to the grid, you can assume that the grid power is infinite, compared to the generator's power. So in load or full load functioning, the gate position variation modifies the active power of the generator. The governor droop will say how much the grid frequency makes the generator's power to increase or decrease. Let's take an example: assuming the droop is set to 10%, a gate position variation of 100% and the rated frequency of 50Hz, results a frequency variation of 5 Hz (using the droop formula (df/f0)x(y0/dy)x100). The droop value is applied to the power feedback of the govrnor, or the gate position. To be more clear, I can send if you want a block diagram so you can understand better what I'm trying to say.

 

[sailer12000]

I appreciate your feedback, perhaps you could explain how the governor logic should work in terms of what my PLC can detect. The PLC "knows" the following information:

SPrpm= Desired Generator Speed in RPM entered by operator
PVrpm= Actual Generator Speed in RPM read from speed sensor
W= Actual Generator Output in Watts
F= Actual Generator Frequency in HZ(1800RPM=60HZ)
%Gate= Actual Gate Position from 0-100%
%Droop= Droop setpoint from 0-5% entered by operator

The PLC program has a PID block with a setpoint (SP), present value (PV), and output (CV). The output (CV) will be fed as a setpoint to an inner loop (gate positioner) that will position the gate between 0-100%.

I assume that I want the speed loop's SP=SPrpm as entered by the operator (typically 60HZ or 1800RPM). Therefore I believe that the "droop adjustment" should be made to the PV value. A formula of PV= PVrpm + (MaxDroopVal - DroopVal) should provide a PV of between 1800 and 1890RPM with 5% droop setpoint. As the actual %gate increased towards 100% (the unit is assuming more load), the PV value would decrease from 1890RPM (at 0% gate) to 1800RPM at 100% gate. Therefore the PID's blocks error signal would get smaller as the unit took on more load, therefore making it less likely to take on additional load. Am I on the right track?

Thanks again

 

[bogdal]

The droop is applied to the generator power value. The frequency error determines the gate movement to increase or decrease the active power, so the error at the inner loop (gate positioner) becomes zero. You can use the following formula:

CV = [(df+dP*bp)(Kp+Ki/s)+df*Kd*s+setpoint]
where:
CV = the inner loop command
df = frequency error (desired frequency - actual frequency)
dP = power error (desired Power - actual power)
Kp = proportional gain
Ki = integral rate
Kd = derivative amplitude
setpoint = desired power

You can see that if the frequency error is zero, the power error is zero (so the frequency derivative is also zero) the gate positioner setpoint is zero for a steady-state opperation. You can also see that the PID is distributed on different terms so in your PLC you must use multiple calls to the PID routine (or function block) according to the mathematical operation you need in the loop.

 

[sailer12000]

Thank you, this is a very clear and useful post. Can you specify what the bp and s terms are in the equation?

 

[bogdal]

Ples escuse me, I've forgot to mention them:

bp - permanent droop
s - Laplacian operator

 

[sailer12000]

I'm afraid the math is beyond me. How would you calculate the bp and s terms based on the values available to the PLC program?

 

[Brimy]

What benefit is there to performing speed control via PLC when an off-the-shelf speed-governor can be purchased with desired PID characteristics? Wouldn't it be practical (& economical) to simply bias an existing speed-governor?

 

[tomatge]

I agree with Brimy - the only reason for developing these algorithms in a PLC is to use that PLC to replace the machine governor. Any off the shelf governor, that has been developed, tested QA'd etc. for that machine will include speed droop. If you want to use a PLC you will need to provide a mechanism for that PLC to directly control generator speed. I think this is a rather dangerous path to follow, bearing in mind the consequence to the generator if you get it wrong !!