Excitation, Torque and Volts/Hz 2

[Schmoleskin]

How much effect does changing synchronous generator excitiation have on the mechanical load that the prime mover sees? Is it the same as KVA is to KW and KVAR, in the imaginary axis? I know that PSS uses this concept, but don't quite understand the relationship.

Have been working with a CAT technician who says they always set up their voltage regulators to operate in the 2 Volts/Hz mode to lessen the mechanical load on the prime mover while recovering from big load steps. I would like to hear if anyone disagrees with that philisophy.

Doesn't the V and 2V/Hz functionality make the generator look more like a big VFD while recovering from large load steps, lowering excitation to maintain proper V/Hz out to motor loads while frequency is depressed?

What are your thoughts on selecting V/Hz as opposed to 2 V/Hz? Does the type of load being served figure into this? For example, which would you want for primarily direct-connected motor load; or for load mostly served through VFDs?

Thanks, DJO

 

[LionelHutz]

For motor loads, you would basically want the V/Hz reduction to match the motor rated V/Hz calculated.

If you are powering loads such as a VFD then maintaining the voltage to a lower frequency would likely be better. Basically, the input side of a VFD is a rectifier and a filter capacitor bank. So, maintaining voltage means the capacitor voltage is maintained and the VFD can continue to operate. The same would basically apply to any electronic type load where the supply is rectified to DC (electronic light ballists, UPS's, computers etc).

 

[ScottyUK]

The synchronous machine itself will not be able to maintain output voltage if frequency drops significantly because of stator core saturation so while what Lionel suggests is in principle a good idea for a solid state load there are limits to how far down in frequency this could be sustained.

I wonder if this is a trick used by the field service guys to buy time for a turbocharged engine to recover from a big load change, i.e. to give the turbo time to spin up? It sounds like the Cat technician is using an aggressive voltage rolloff as a crude means of temporarily shedding load in order to minimise the probability of stalling the prime mover, at the likely expense of tripping some or all of the load on undervoltage. A fast-acting load shedding scheme and a less aggressive rolloff would sacrifice the less critical loads while keeping the supply to the critical loads within reasonable boundaries. The techinique being used by the Cat tech has a hint of Russian roulette in terms of which loads drop out first - there's no guarantee that the most sensitive or critical load isn't the one which is dumped. I'd rather have a defined schedule of load shedding which sequentially dumps the least critical loads to protect the most critical.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[catserveng]

The CAT tech is using the factory recommended settings. As indicated above, the reason for the agressive voltage drop during a frequency dip is to allow the engine to recover more quickly. Unfortunately CAT doesn't do a very good job these days of training the field techs to do actual field startups where they look at the installed loads and make adjustments that would give best possible transient performance while still meeting customers minimum voltage and frequency needs.

In their defence I would also say that some customer expectations are sometimes beyond what modern packages can deliver effectively. The trend has been to get more horsepower out of less displacement by turbocharging the heck out of the engine, then still requiring fast starts, ability to accept large block loads, and meet more stringent emmissions levels. Instead of Russian roulette I personally find the standby generation business a pretty good game of liars poker. I recently got involved with a site that was using a package fom another supplier with similar issues. Taking some time, looking at the loads, seeing what the end user really needed and making adjustments based off actual dynamic tests and real time information got things where they needed to be. But you may find as I have that typically units are shipped with "factory" settings, and the field techs are under pressure to get the unit signed off as quickly as possible, and if setting changes are requested, there is usually some resistance on the part of the factory because "the standard settings should work, sounds like a system problem".

Selection of the volts/hz slope is dependent on the site loads and the transient response you are trying to achieve. The older AVR's had a jumper selectable set of options, single or double volts per hertz. These settings gave you a voltage rolloff and about 8 and 16 volts per hertz drop. The newer digital AVR's have this parameter programmable, usually in a range from 0 (basically sets the AVR to constant voltage mode) to about 20. On the newer AVR's there are also knee frequency and cutoff frequency settings that define the start and end of the slope. You should also be aware that some of the dynamic settings of the digital AVR's can affect the actual operation of the v/hz response during a transient.

Best way I found to setup the transient response is to use a strip chart recorder or some other means of seeing the voltage and frequency response, I use a Fluke Scopemeter and the trending in the FlukeView software, not as fast as a real strip chart recorder but stands the abuse in the field a lot better. The make sure you define your maximum allowable voltage and frequency drop and recovery time. I usually start with speed control (governing) dynamics first, but you may find with newer electronically controlled engines this can be difficult. Then make adjustments to the AVR dynamics and volts/hertz settings. Make sure the knee frequency (point at which the volts per hertz rolloff will start) and cutoff frequency are set based on your site requirements. Most factory defaults for 60 Hz systems are 59.5 Hz for knee frequency and 45 hz for cutoff.

If you have voltage and frequency sensitive critical site loads you may not have the transient or block load capability as defined by the manufacturer, usually referenced to an ISO standard.

Hope that helps.

 

[ScottyUK]

Thanks CatServEng.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[rovineye]

Some good points CatServEng.

Our installation testing is complicated by dynamic load rejection testing in addition to the unreasonable expectations you correctly cited.

How ofter do you have reactive load banks available when commissioning?

 

[Schmoleskin]

Thanks all, wasn't expecting so much good info so fast.

DJO

 

[catserveng]

I used to work for a engine/generator dealer for a number of years, and did a lot of contract startup work for the manufacturer. Our dealership did a large enough power business that reactive load bank capability was available and used often. While it helped in a lot of areas, it is still very hard to simulate a site load, especially nowadays with all the power electronics suach as VFD's and UPS systems. It is also really hard to take a fairly general industry specification and apply it across the board to users of generator power. A dairy and a data center have really different loads, and the same generator system usually can be applied to both pretty easy, but how the react dynamically to each site can be very different.

I guess your big challenge is to figure out what you can really live with and find a way to test it so that your installed equipment won't be damaged. Double volts/hz is fine if you can live with it, but it should be your decision, not the vendors. Of course now you'll get involved with what was promised, purchased and installed, and what it really does once it's sitting there.

Hope that helps.

I am a firm believer that a real systems approach needs to be taken in applying standby and prime power. Manufacturer sizing programs can help define the power levels needed, but can't take into account all of the possible contributors that can affect system response and stability. And the amount of time actually spent on startup, commissioning and testing has seemed to decline, at least in the markets I normally deal in, even though the overall power systems and customers needs have become more critical and complex. Problem is no one really wants to trash around a power system with all the cirtical loads on, so how do you adjust and test to meet dynamic needs of the site. Been more than a few times we did the load bank tests, packed up and went away, only to get a frantic call after the first outage and hear that the standby system didn't work as expected.

 

[rovineye]

The digital world really helps me here. On the first ship of a class I struggle TRYING to pass the requirements laid out for block loading and rejection. With the bigger slower machines we don't make all recovery times or maximum excursion. Once we take the ship to sea and fine tune AVR and governors we just import those setting to follow-on ships. Those follow-ons may or may not pass dynamic testing, but I accept the results anyway. I am much more concerened that the final porduct performs and don't care that we don't quite pass dynamic testing using random load criteria. Random in that is pulled off a generic load chart and not related to the load load ramps of our propulsion motor.

 

[Andy512]

How would you size a generator to run a bank of VFD's?

Number of motors * motor FLA * Safety factor?

This would be a continuous load as grid power will not be available.

 

[rovineye]

You need to better define the operational scenario of the VFDs. Are they always running or can you introduce a useage factor?

If all VFDs, you may need to account for harmonic heating of the generator.

 

[Andy512]

The motors will all be running at or near same load. This includes periods where maximum motor power is required.

Between 15 and 20 400HP motors.