Spinning Reserve And Load Dumping

I've been beating my head against Google trying to find sources for reliable data on two things: a) how much energy is consumed maintaining the spinning reserve,
and b) how much energy is consumed in dumping excess grid capacity.
I suspect (b) is really small, and not significant in the discussion of overall grid efficiency. Efficient scheduling of energy generation is what those guys do for a living, and you'd think they would have figured it out by now. :-)
With regard to (a), ISO requires at least 50% of the operating reserve to be spinning reserve, that is to say, synchronized to the grid and ready for throttle-up on demand from the ISO. Almost all of that comes from hydroelectric and combustion turbine sources.
The operating reserve requirement is variable, depending on the sources used for grid power. The ISO requires an OR for hydroelectric power of 5%, and for all other sources it's 7%. For interruptible imports, it's 100%. If wind or solar were significant sources, the ISO probably would have a requirement for a very high operating reserve for that power, like 100%, because mankind does not yet have control over the clouds and wind.
What I want to know is how much power is being consumed maintaining the spinning reserve. I'd guess it's a lot, because gas turbine engines do not have good efficiency when running at no-load. I would not be surprised if as much as 10% of the total grid power was being spent on maintaining the spinning reserve. But I just cannot find numbers on this question. I would be very grateful for any pointers to reliable sources of information which discuss this question.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Xpost to aee

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

It appears that you think that when you have, say 100MW of capacity on line, and the load is only 20MW, the remaining 80MW must be dumped. This does not occur. (load dumping is an altogether different thing- as a possible response when the system is going unstable). If the gross load requires 20MW, the grid supplies 20MW and the remaining capacity is not used. The capacity is simply the power available if needed. It is not power generated with excess dumped somehow. If you don't need it, it isn't being produced. Spinning reserve is simply unused capacity which is quickly available on demand.
Once a unit is on line, how it is loaded is a well developed technique called "economic dispatch" Except for hydro plants, this is pretty straightforward. Spinning reserve doesn't enter into this as the basis is an assumed group of generators at a given time. The loading on the individual machines will be optimised for the specific load at any given time. The unused capability or spinning reserve will be shifted between machines as load changes in rder to minimise costs. The decisions as to what units should be on line is what is called "unit committment" This is messier. Spinning reserve considerations enter into this. Spinning reserve based on a given % was (may not be so now) giving way to a statistical approach- the general numbers that you give are partially based on this and are minimal.
--

Don Kelly snipped-for-privacy@shawcross.ca
remove the X to answer
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sat, 22 Dec 2007 04:42:24 +0000, Don Kelly wrote:

Then if you have, say, four 100MW machines on line driving a 300MW load, that counts as essentially 100MW of spinning reserve?
Thanks, Don, for your refreshingly informational and civil posts. The more I read about the electrical grid, the more impressed I am with the remarkable technology we take for granted.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
----------------------------

------------ Yep, you got it, and there is no charge in terms of fuel or whatever for the unused 100MW of capacity.
Merry Christmas
--

Don Kelly snipped-for-privacy@shawcross.ca
remove the X to answer
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
The only slight fuel cost is maybe having to operate one or some of the machines outside the most efficient loading for that particular plant. This would be included in the 'economic dispatch' that you mentioned earlier.
So from the original post the 'cost' of maintaining spinning reserve is additional fuel consumed by having to operate a given plant outside best fuel efficiency (bad for the planet), or worse having to operate an 'uneconomic' plant (bad for share holders).
Newsey

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
----------------------------

Actually not. The process of "economic dispatch" is intended to minimize total costs of generation at all loads. The cost of spinning reserve is simply the cost of having extra capability on line-but once a given unit is on line, the variation of its fuel cost with load is taken into account in determining its share of the load. Which units should be on line is a matter of "unit committment" which is another (related) problem.
One loading scheme that was proposed many years ago was loading the most efficient unit first, up to its most efficient point, then the next most efficient unit etc. When all are loaded to the most efficient point, go back to the first unit and load it to capacity, etc. However this doesn't work.
Roughly speaking, optimal dispatch operates every unit at the same incremental cost unless that point is above or below limits on the unit. In other words, minimize the cost of the next small increment of delivered energy. The operation, at any total load is such that any deviation from this equal incremental cost results in higher net costs.
This is modified by location using loss coefficients to account for transmission losses depending on plant location so the minimization is then based on minimal cost of incremental delivered energy. It can also be modified to meet specific other criteria than $cost such as a built in bias toward less polluting plant.
This process was developed about 60-70 years ago -using a rule of thumb. However, a strict, calculus based analysis confirms this as the theoretically best approach so the original users deserve a lot of credit. The loss model has changed because it can be changed interactively in near real time.
Nowadays, economic dispatch may include "unit committment"- what units should be on at a particular time and when should they be shut down or brought on line considering the costs of bringing a plant on line and this cost will be partially dependent on how long it was down before restarting (e.g. a steam turbine needs a period of time to get up to a stable operating temperature before it is loaded). Computer speed is now such that such a process can be carried out relatively quickly and re-assessed repeatedly during the day as load changes.
Hydro is more complex as there is a time factor involved with storage , inflow and outflow from reservoirs as well as the time delays for water out from one plant into another downstream plant. There is a tool for this -based on "Calculus of Variations" which is somewhat messy. http://en.wikipedia.org/wiki/Calculus_of_variations
--

Don Kelly snipped-for-privacy@shawcross.ca
remove the X to answer
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Don Kelly wrote:

Thanks for the informative replies. I was trying to figure out what the impact will be of time-of-use pricing, demand response appliances, and vehicle-to-grid technology. It's my impression that the first two of these trends are inevitable, and they'll improve overall energy efficiency and peak demand.
Right now, the utilities buy power at variable rates and sell it to consumers at fixed rates. This is a crazy pricing system, and the only reason we have it is we didn't have smart meter technology. But now we do, so we will have TOU pricing, and this is the stick which will push consumers toward DR-enabled white goods and V2G cars (though the latter will only be practical when hydrogen fuel-cell vehicles are introduced, which might be never).
Because efficiency at idle is one of the factors that killed the turbine locomotive (they spend a lot of time at or near idle, for example when connecting rail cars), I was thinking that a gas turbine power generator would also have poor efficiency spinning at or near idle. But now I see that you won't have turbine generators running near idle -- the load will be spread around several generators, all operating at or near their peak efficiency.
DR-enabled white goods will allow the spinning reserve to be replaced by a virtual spinning reserve made up of all the smart appliances that will avoid consuming power during peak load. It seems to me this won't actually save much energy, except that it will allow the utilities to run their turbines closer to peak efficiency and avoid losses from long-distance power transmission. The main effect will be the impact on price, by avoiding buying power at peak rates.
Do I now have a correct interpretation of the trends?
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
----------------------------

--------------- They will flatten peaks and reduce the need for high fuel cost peaking plant. Unfortunately de-regulation favours these (i.e. gas turbines) due to their lower capital costs.

----------------- The fixed rate process has worked well in the past but if the intention is to reduce peak loads and even out loads, then the time of day pricing is a better choice.

--------- Most may not be at or near peak efficiency. This will only be the case for a fairly flat efficiency-load curve. Economic dispatch doesn't imply that but rather the best use of the resources on line. That may mean some machines at minimum load, others at maximum load and most at some in between situation -so that the overall system fuel costs will be minimized. This could mean that any gas turbines on line would be at low loads. Peak efficiency concepts are leading you to conclusions which are invalid. Consider two plants a) Ca+0.1Pa +0.01Pa^2 b) Cb =5 +0.05Pb +0.02Pb^2 Incremental costs are 0.1+0.02Pa and 0.05+0.04Pb The total load is P=Pa+Pb so Pb=P-Pa For equal incremental costs 0.1+0.02Pa=0.05+0.04Pb or Pb =5/4+0.5*Pa so for P =1.25 , Pa =0 and Pb = 1.25 For Pb =2.5 then Pa =2.5 and P=5 and while this is the maximum fuel efficiency for a it is far from maximum for b. BUT the overall fuel cost will be minimum. At Pa =5 then Pb= 3.75so P=8.75 In this case b will be far from its peak efficiency but the overall fuel cost, for that load will also be minimized. In fact for P<5 , unit a will take the larger part of the load but this reverses for P>5 so that at P, Pa =9.2 and Pb=5.8 (In the limit Pa would approach 2/3P)
Peak efficiency for a is where dCa/dPa =0 and for b it is where dCb/dPb=0 The optimal loading is where dCa/dPab/dPb . In the above case the unit efficiency points are on the optimal curve at 2.5 and 5 units but that doesn't mean that the overall system efficiency is maximized. and, at any other load near or far from the peaks, the use of "peak efficiency" is moot. Essentially the incremental cost method is trying to maximize the total system fuel efficiency at all loads.

------------------ This is a partial truth. Remember that spinning reserve allows for contingencies in that enough reserve must be available to pick up the slack if there is loss of generation. You could have the peak shaved by smart appliances but this is not going to be of help if a generator has to be disconnected in an emergency situation as smart appliances, while adjusting load, do not produce energy. Spinning reserves are a reliability of supply issue. If a generator fails- capacity to take over is present- the alternative is the ultimate load control- deliberately impose blackouts in order to shed load to below the level of available supply.

closer.
--

Don Kelly snipped-for-privacy@shawcross.ca
remove the X to answer
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Don Kelly wrote:

Don, just to clarify, I think that what you mean here when you say efficiency is generation cost (or its inverse) rather than energy efficiency. The goal is to minimize overall cost, not to maximize overall energy efficiency, if I interpret your correctly. Units with the lowest incremental cost do not necessarily have the highest energy efficiency. Nuclear plants, for example, have the lowest energy efficiency of all thermal plants, but also the lowest incremental cost, since the fuel costs are low compared to fossil fuels. That is one of the reasons that nuclear plants are almost always run at full load (and therefore, do not contribute to spinning reserve).

What demand based metering would contribute, however, is that reducing the peak load potentially allows utilities to (a) avoid building additional generating capacity and/or (b) avoid bringing their oldest, least cost effective units on line at all. The spinning reserve may need to be just as high, but it will be on top of a lower actual load, meaning generally that the incremental cost of the power being generated may be lower (b) or that the utilities fixed costs may be lower (a).

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

---------- That can be. Please note that I specified cost efficiency rather than energy conversion efficiency. However the relative energy efficiency with respect to fossil fuel is not that they are nuclear but more to do with the thermodynamics involved at, generally lower steam temperatures).The basis of economic dispatch is primarily economics -hence the name. The basis of generation planning (long range) is also economic. This also means that when such sources as wind or solar are available- use them as fully as possible as the "fuel cost" component is negligable even though they may be far less energy efficient than nuclear plant. Hydro in some cases will be peaking plant and in other cases may be base load plant because of other factors even though the "efficiency" is far greater than any other source available to us. When you have a variety of sources available and costing $ even when shut down (capital costs) do you let the high capital cost,low fuel cost units idle (even if less efficient as in the case of solar units)? If you want to bias toward the most energy efficient sources then you have to build this bias into the dispatch just as it can be done with respect to pollution- put a price on it. However, such bias must be based on more than wishful thinking.

------------ Certainly the flattening of peaks will be beneficial in terms of a above. If you can control the peaks by whatever means, you can get by with less total capacity on line and also delay new construction to some extent. With respect to (b) it is less clear. Such plant nearing retirement, goes through a period where it is normally off line acting as standby rather than spinning reserve. Dispatch techniques determine the load sharing between plants on line- unit committment considers what plant should be on line at a given time,based on load predictions, taking into account the needed spinning reserve, and, while the two are related, once plant is on line, the sharing considers the best way to share load between those plants independently of the spinning reserve.
--

Don Kelly snipped-for-privacy@shawcross.ca
remove the X to answer
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

<snip>
I can see that for the case of regulated utility that owns/pays all the fuel costs so naturally wants to minimize their total fuel costs. But what about dispatch when different operators own different plant types? For example, if A belongs to one company and B to another, then each is interested in operating *their* unit at best cost. It would seem each owner would consider this when they 'bid' to an ISO for supplying next day's load. That is, A's bidding would have to consider what operating point they are willing to operate at and for what price (and a little guesswork as to what 'price' is likely to win the bidding). If they are asked to operate away from their 'best point', they would naturally want a higher price.
Do ISO's in unregulated areas have to consider this? When we were starting down this road at DTE, we just accepted bids for a fixed price and a fixed amount of generation, with no complex price structures when under/over that load point. Any insight?
daestrom
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
----------------------------
remove the X to answer

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.