Re: [Phys-l] black start

[John Denker <jsd@av8n.com>]

On 05/03/2011 12:03 AM, Michael Edmiston wrote:

> The Wiki-article states that hydroelectric power plants are the ones that 
> are easiest to black-start from a diesel generator because you only need 
> excitation power plus sufficient power to open and moderate the water intake 
> gates.  This makes a lot of sense, but not all areas of the continent are 
> flush with hydroelectric power plants. 

A gas turbine is also very easy to start.  (In the electricity
industry it's called a CT == combustion turbine.)

You can buy a 1 megawatt turbine that can be easily started by
a car battery.  It would fit under your desk (although your 
office-mates might object to the noise).  I've seen plenty of
them (in aircraft).

That in turn would provide plenty of power for starting a much 
larger CT.

>  A diesel-powered black-start-able generator would need to be huge 
> (and expensive) in order to start a typical-sized coal-fired generator.

So don't do it that way.  

Suggestion:  Suppose you are in Oklahoma, where there is no hydro
power but plenty of natural gas.
  Unit 1 = 100 MW coal-fired
  Unit 2 = 100 MW coal-fired
  Unit 3 = 100 MW coal-fired
  Unit 4 = 100 MW coal-fired
  Unit 5 = 10 MW natural gas CT
  Unit 6 = 1 MW natural gas CT

Wire unit 5 to the grid so it can provide "peak power" to customers.
Thereby it pays for itself.  Also use it to black-start any of the 
coal-fired units.

Unit 6 is super-easy to start.  Maybe hand-cranking would not
suffice, but a small battery does.  Use it to provide station 
power (lights, computers, controls, et cetera) ... and to spin 
up unit 5.

Such an arrangement provides black start capability at essentially
zero cost.  Remember that unit 5 paid for itself /before/ we started
talking about black start capability.

I'm not an expert, but AFAICT the advantages of such an arrangement 
are overwhelming.

> The question then becomes, do we 
> collectively spend some big bucks to gain black-start capability at some key 
> coal-fired plants, or do we spend money to improve the grid to reduce the 
> likelihood of widespread blackouts like the one that happened in 2003, or 
> perhaps some of both?

The obvious answer is both.  Obviously we want the grid to be
reliable, but we also must realize that there are things like
tsunamis, floods, tornadoes, et cetera than can take out even
a well-designed grid.  The idea of a plant without black start
capability is beyond my comprehension.

I can see how upgrading the grid would require big bucks.  I
cannot imagine how providing black start capability would require
big bucks.  I cannot imagine why this is not longstanding standard
practice.

Maybe it is a matter of perverse incentives.  As it stands, the
utility is not liable for the lives and property that are lost
due to a prolonged blackout.  Maybe we need to change the tariffs
so as to give the utility more of an incentive to get the power 
back on promptly.