What are the effects of a leading power factor?

------------ I may not have been clear- at any pf other than unity -losses per unit of useful energy are increased. Customer pays for the losses on his side of the metering.

transformers,

Demand metering doesn't replace energy metering- it is additional to the energy metering. Demand metering doesn't measure efficiency or losses. All that it indicates is the peak sustained kVA during a billing period. Customer's inefficiencies will affect this peak and the loading of transformers etc so, in that respect, the demand is affected correctly. The customer pays directly for his losses in the energy charges (actual energy use). In addition, losses in the utility system are not generally attributable to a given customer so the overall energy rates include these losses along with other operating costs, executive lunches, and investor dividends.

|> | That may be true- it was routinely done in the area that I used to live | in. |> | Note that KVA demand metering was based on monthly peak where the meter | had |> | a thermal(?) lag so that it rode through starting transients and short | peaks |> | and represented what was important to the utility - temperature rises in |> | equipment. |>

|> Sounds like exactly the right way to measure for the component of costs |> that relates to purchasing, installing, and maintaining transmission and |> distribution facilities. |>

|> So, given 2 customers that periodically have a 1000 kVAR demand, but run |> regularly around 100 kW, where one of them has pf 0.97 and the other has |> pf 0.66, how are they going to measure the added charges for the second? |> A kVAR usage meter? | --- | The meters measured kVA (not kVAR) demand and the needle pushed another | needle so that the peak kVA during the period is registered. Your two | customers would have the same kVA demand charge as both require the utility | to provide the same equipment capability but the first would have a higher | kWH bill as he uses more energy. The effective cost per kWH would be lower | for him than for the second person. Both are being charged for a high kVA | (not necessarily kVAR) peak demand because of a fluctuating load and the | second gets a whammy in this respect from poor pf. The KVA demand | penalises high peak (over 15 to 30 minutes duration) to average loads as | well as poor power factor.

The example I am giving is that both customers use the _same_ energy level (an average of 100 kWH over the month) and both will run a peak kVA and kVAR at the same high level (and hence the same demand reading). But what the difference is, when at times other than the peak, one customer has a lower power factor despite the same energy. So that customer takes more power, but then gives much of it back each half cycle, netting the same usage and loss internally as the first customer, but imposing unmeasured (by that meter) losses on the utility distribution. What I am saying is it can be measured in terms of measuring both kW(H) and kVAR(H). Each of those figures would have a rate. The kWH rate would be the usual energy rate. the kVARH rate would be a secondary energy rate prportioned to the loss that would be experienced sloshing reactive power around. The first customer would have a lower kVARH and the second customer would have a higher kVARH, while they both have the same kWH and kVA peak demand.

| Take another case - steady 100KVA loads - first one at 0.97 pf lag and the | second at 0.66 pf lag. Each pay $0.08 per KWH so that over the billing | period (say 720 hours) he will pay $5587 for energy and (at $2/KVA) $2000 | for demand - total $7787 for $0.11 "average cost per KWH" | The second pays $2000 for the demand charge and $3802 for energy - "average | cost per KWH" becomes $0.13 per KWH. In this case it would be beneficial for | the second user to install capacitors.

In this case energy usage is definitely different.

| Another case- first one 100KW steady at 1.0 pf and second 100KW steady at | 0.66pf lag | First pays $200+5760=$5960 Second pays $303+7760 =$8063 for the same actual | energy used. | | Third case: both at 100KW average- first as above and the second with a peak | of 1000kVA- both at unity pf. First pays $5960, second pays $7760 | | A demand charge based on KVA peaks which exist long enough for equipment to | reach a steady state operating temperature reflects both the effect of high | load peaks as well as poor power factor.

Now go with BOTH having a peak of 1000kVA. Both demand meters will read the same thing (1000kVA peak since reset). It could be a one day long peak. So they would both have the same demand charge. The rest of the time both use power such that over the whole month (720 hours) the energy used, including that used during the one day peak, totals 72000 kWH (the equivalent of 100kW over 720 hours). So their energy readings will be the same, too. But the difference is that during the times other than peak, they have different pf.

What I am saying is that kVA demand peak and kWH total energy can still have exactly the same readings even though the kVAH would differ. The customer with the lower pf is, of course, putting more load on the distribution due to more kVAR, but this load isn't contributing to the peak demand since that one bad day set the peak level. Still, it is overall heating of equipment, and more importantly, is energy wasted in distribution that does not get measured by the kWH reading (I'm not talking about what little loss there is in the customer's own wires, which would be measured by the kWH meter); I'm talking about the extra current drawn in by the kVAR, not see by the kWH meter at the customer, but has some partial loss in the distribution coming in. And that is an energy loss (in addition to anything else). Maybe we don't need both kW and kVA peak demand readings, but I do think we need to have both kWH and kVA(R)H readings to measure the cost of the customer (not just what they benefit from).

We had 'demand metering' at one facility, so we coordinated the starting of our largest machinery (2000 hp emergency pump started monthly for testing) and testing our emergency diesel generators (4000 kW EMD's) so the pump start would *not* be on-grid and affect our monthly demand charge. This is the type of demand metering that Don seems to be talking about. The motor start was less than 10 second surge, but it would 'kick the needle' up for the month.

But you seem to be talking about kVAR metering/charging. I've seen that in the Detroit area and it *is* a different animal. A couple of industries that run really poor pf caused a lot of trouble for Detroit Edison. Billing based on the energy used was not enough. Billing based on 15-minute demand was not enough. The voltage control DE needed to implement because of the very poor pf justified VARH metering, in addition to kWH and peak demand. This was very rare though, I think there was only one or two customers in the service area that had that. ISTR, in the early 90's, the company got tired of paying DE and went to some sort of solid-state, electronically controlled 'condenser' system that could react to the changing var loading and save them the extra charges. Don't know why they didn't use 'synchronous condensers', *thats* how DE was serving them. Maybe they didn't have the space for them, but did for electronic version.

daestrom

| We had 'demand metering' at one facility, so we coordinated the starting of | our largest machinery (2000 hp emergency pump started monthly for testing) | and testing our emergency diesel generators (4000 kW EMD's) so the pump | start would *not* be on-grid and affect our monthly demand charge. This is | the type of demand metering that Don seems to be talking about. The motor | start was less than 10 second surge, but it would 'kick the needle' up for | the month.

That is definitely the way to do it to keep down costs.

| But you seem to be talking about kVAR metering/charging. I've seen that in | the Detroit area and it *is* a different animal. A couple of industries | that run really poor pf caused a lot of trouble for Detroit Edison. Billing | based on the energy used was not enough. Billing based on 15-minute demand | was not enough. The voltage control DE needed to implement because of the | very poor pf justified VARH metering, in addition to kWH and peak demand. | This was very rare though, I think there was only one or two customers in | the service area that had that. ISTR, in the early 90's, the company got | tired of paying DE and went to some sort of solid-state, electronically | controlled 'condenser' system that could react to the changing var loading | and save them the extra charges. Don't know why they didn't use | 'synchronous condensers', *thats* how DE was serving them. Maybe they | didn't have the space for them, but did for electronic version.

Yeah, I am talking about that, in addition to other measures. My point is that a customer can have the same energy usage, and the same demand peak, and still vary in the reactive power effect on distribution. I think, to be fair, all factors need to be measured, and charged an appropriate rate for each. But kVARH (yes, the H is in there) is not a measure of demand, or how big the equipment to serve the customer has to be, but rather, it is a measure of how much energy is being LOANED to the customer. It's rate would be like charging interest on the loan, figured on the basis of how much real energy is lost in transporting this loaned energy. It is an energy charge (the wasted energy has to be generated by the energy provider) and a distribution charge (the wasted energy is a little more heat load that could add up if other customers are at greater peak demands at that time). But for most customers, these charges would be small, and probably do not warrant the cost of measuring them. Still, it is a valid charge (and should be charged if measuring and billing didn't add to the cost).

My next argument is that demand metering itself should be measured differently. Suppose you started that emergency pump on the grid (perhaps because you didn't have emergency diesel generators). If you do that during the day, you are putting more load on the grid than if you do that during the night (generally). You should be charged less for doing that at night. But of course this would require more sophisticated metering. Imagine a gang of 96 demand meters, each of which operates only during a unique designated 15 minute slice of the day. Billing would be done by multiplying each of the 96 demand levels by the respective rate for that

15 minute slice. Then the monthly demand charge is which of these 96 charges is the highest. Thus if you have a peak demand at night, when the rates would normally be lowest, and none in the day, your demand charge would be lower.

And I am not mixing up demand and energy. Of course energy rates are generally lower at night because of greater availability of generating capacity then. But demand rates would be as well, because you demand peak, while still stressing the transformers feeding your loads, would not be stressing the distribution and transmission as much because it is at a time when the overall demand from all customers is down.

Of course a gang of 96 demand meters is totally impractical. But there is a way to do this reasonably well with electronic metering. One meter that measures kWH, kVARH, and peak kVA, separately in each 15 minute period, and sends the up to 2976 x 3 data points to the billing computer (it would be no more than about 128K bytes of data), can accomplish all these measures, and do it all on a time of day, and day of week (do your pump test on Sunday when many industrial users and commercial office air conditioners are at reduced load), basis. Computer software can readily sort out all this data, as well as provide detailed time of day load trends for engineering to compare with their distribution metrics. This is what I think the meter of the future (today, really) needs to be. For the home, a less sophisticated version (cheaper to make because it only measures kWH in each 15 minute period) would be used.

Your not the utility. They pick how the rate structures are set due to the customer that that have. If you were in an area of steel production you would see higher PF billing rate that just for homes. I know of places on the eastern seaboard that if a large user fails to keep the PF above 85 they the utilitiy has the option to disconnect them. My experence of demand billing is done by using an either fixed or sliding demand window. With an sub demand window below that of 3 or 5 minutes. They record the maxium demand in each 24 hours and reset it once a month. This all has to do with generator capisity. Try northern Mexico or the LA basin for some utility demand rates that get complicated. Most utilities use Summer and Winter rates. They also introduce Peak demand, Mid Peak and low. They also use up to 4 time periods a day. ( at least that is the most I have ever seen).

The problem with a computer data base is that the information becomes huge very fast. Using the CH PowerNet system I have seen an single meter create an 40-50 meg file in 30 days. ( 15 minute window and 5 minute sub window) MS Access crapps out when the data base climbs much over 700 meg. Now what happens when you miss a time point? You interpolate around it. Bad news if it was an Peak billing period. The customer billing systems that I have installed are used for demand side management. They need to know where the load is going and when. Square D has an system and so does PML. I like the PML meters because you can install more storage in the meter. CH meters, the

6000/6600 had only a meg of storage when I worked with them. One meg was only about a weeks worth of data depending on your rate structure.

Here in AZ there are residential demand meters. One used has up to 4 times of day in it and the other is stricly demand based on a 15 minute window. The utlilites keep the monthly high for a year. Make a mistake and once and you are penalized for a year. When I worked for ASU, the electric bill was around a million a month. Load varied from 24-28 meg demand in the day time to 30-34 meg demand at night. Chillers ran only off peak. We were able to choose our demand for a month. Breaking the bubble as it was called cost $40,000 dollars a meg until the end of the month. You do not see many night games during the week at Sun Devil Stadium for this very reason. I called central plant one Monday night game and told them they had 20 minutes to shed 3 meg. 3 Meg is what the feild lights draw. They though I was crazy, until the load hit them. I had worked on the feild light upgrade for the SuperBowl XXX and knew exactly what the lights drew. We once measured the feeders to the stadium, just about 130 amps at 15kv. But when the kick off happen the load dropped to about 90. They were done cooking and everything was in warmers after kick off. By half time the load dropped to about 75 amps.

The metering you describe had been available in various forms for more than

10 years maybe longer than that.

------------------- There seems to be some confusion here: 100kWh is, I assume the monthly energy usage- if so then a peak kVA reading would last for less than 0.1 hours and a demand meter would register a negligable kVA for both loads. If you are talking about 100kW average for the month (about 720 hours) then the kWH would be 72000 kWH for both loads. If both loads indeed had a peak of

1000kVA with the pf's that you cited, the first load would have a peak power of 970kW and the second a peak power of 660 kW but the same kVA peak.

-------------- >But what the

----------- I have to disagree with you. The energy that is being shuffled around will not appear on the kWH meter. That is true. The meter (and kW meters) measure average power. The difference is that 100kW at 0.97 pf will have a kVA of

103 kVA while 100kW at 0.66 pf will have 151.+ kVA. Assuming the same voltage and line resistances , the internal losses of the second customer will be (151./103)^2 = 2.16 times that of the first customer. This loss WILL be registered by the kWH meter and the customer pays for it. Measuring kVARH will not give any useful information as the customer's losses are already measured and this information will not be of use in sizing of equipment or, as a matter of fact, determining the utility losses. KVA demand will not determine utility losses either but will affect equipment sizing and costs (a point that you agree with). Utility energy losses will be determined by utilities measuring the difference between their generated energy and the customers' usage.

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

The first customer would have

------------- Interesting- the demand meters that I have seen (and tested one about 50 years ago) had a thermal lag such that the reading would take about 20 to 30 minutes to reach 90% of the actual kVA in a steady situation. In 10 seconds the "kick" would be very small if any. The recognition that motor starting occurs and does not normally have any real effect on equipment heating was recognised. Repeated starts- another animal but I bet there were limits on how often the pump could be started.

As for the VARH metering - in fact this appears to be a case of a specific solution to a specific situation. Again, interesting- thanks.

------------- The problem is that the customer returns this energy within the period (1/60 second) that it he used itIt is not an "energy charge"- loaned or otherwise. What interest would be charged on a loan for 1/2 of a cycle? It is being charged through the energy metering. In addition, as I have said earlier this evening- KVARH metering does not reflect the actual losses seen by the utility. In the case above it is not an energy related issue but a voltage control issue which so far is quite a specific situation.

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

---------- My initial comment was "Does the size and cost of a utility transformer and lines feeding a specific plant depend on the time of day? If it were so, then your suggestion would be great." However, I think that you are going beyond this - doing "time of day" to bias the demand rate as is now done in many cases for energy rates. Interesting but considering that the demand charges are mainly related to the cost of equipment to supply a particular load ( further back in the system the effect becomes more diffuse so the costs of the immediate feed to the customer is the only one that can fairly be attributed to that particular customer- the more diffuse costs are built into the energy rate) and the equipment isn't going to change on a 15 minute cycle- there is a problem. What sort of "demand" metering most accurately represents costs?

You are absolutely right in suggesting that such testing be done at night or off-peak times- The KVA demand is reduced and if time of day metering of energy is in place, the energy cost would also be reduced. Note that many large industries negotiate energy contracts which have a kVA limit- if the peak kVA is below a certain value then a "good" energy rate is available:-BUT if the kva demand exceeds this limit- the energy charge increases drastically. Many industries with their own generation, use such generation to shave the peaks and keep the charges from the utility down. If one has such capability it works well to the advantqage of both the utility and the customer. Many moe utilities use some form of pf correction- reduction of kVA demand as well as improved voltages.

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

-------- The idea is great but the problem is that back in the system- your load is absorbed into a melting pot. What demand that is measured at a customer level can be isolated as chargeable to that customer. How much effect that it has further up the chain depends on the size of that customer with respect to the sum total of all customers. Load diversity rears its head as well. Demand charges are meant to address costs associated with supply to a specific customer rather than overall system costs. . Can this be extended to consider overall system costs - in theory - Yes. However, in practice, when the cost of getting the data exceeds the saving gained from use of the data- don't bother.

| Your not the utility. They pick how the rate structures are set due to the

Rate structures are regulated by appropriate state agencies that do take consumer feedback. I've done this in telephone situations before. I may do so with electrical some day.

| customer that that have. If you were in an area of steel production you | would see higher PF billing rate that just for homes. I know of places on | the eastern seaboard that if a large user fails to keep the PF above 85 they | the utilitiy has the option to disconnect them. My experence of demand | billing is done by using an either fixed or sliding demand window. With an | sub demand window below that of 3 or 5 minutes. They record the maxium | demand in each 24 hours and reset it once a month. This all has to do with | generator capisity. Try northern Mexico or the LA basin for some utility | demand rates that get complicated. Most utilities use Summer and Winter | rates. They also introduce Peak demand, Mid Peak and low. They also use up | to 4 time periods a day. ( at least that is the most I have ever seen).

Obviously the utility has to plan ahead and make investments to support custoemrs with low PF.

By concern is that the utilities do charge them appropriately, so as to keep the rates lower for the rest of us.

| The problem with a computer data base is that the information becomes huge | very fast. Using the CH PowerNet system I have seen an single meter create | an 40-50 meg file in 30 days. ( 15 minute window and 5 minute sub window) MS | Access crapps out when the data base climbs much over 700 meg. Now what | happens when you miss a time point? You interpolate around it. Bad news if | it was an Peak billing period. The customer billing systems that I have | installed are used for demand side management. They need to know where the | load is going and when. Square D has an system and so does PML. I like the | PML meters because you can install more storage in the meter. CH meters, the | 6000/6600 had only a meg of storage when I worked with them. One meg was | only about a weeks worth of data depending on your rate structure.

You don't have to tell me what size of data you use for me to know that MA Access will crap out. It does that on just about anything.

There are plenty of poorly programmed systems out there. But that does not stop me from promoting the idea of using them; I'll just insist that they hire more competent programmers.

The size of data depends on how much reduction is done in advance to make summaries of metering periods, such as 15 minutes. Of course you could go to the extreme of recording their current and voltage waveforms at high resolution, and get a good picture of all the quality issues, including harmonics. But what I suggested was 3 variables 96 times a day, which is much smaller than the numbers you mention. So either they are gathering much more data, or doing it very poorly.

| Here in AZ there are residential demand meters. One used has up to 4 times | of day in it and the other is stricly demand based on a 15 minute window. | The utlilites keep the monthly high for a year. Make a mistake and once and | you are penalized for a year.

Sounds like a tariff issue to me.

| The metering you describe had been available in various forms for more than | 10 years maybe longer than that.

Then it's time to deploy them more widely and fix the software issues. Back to a regulatory agency.

| The problem is that the customer returns this energy within the period (1/60 | second) that it he used itIt is not an "energy charge"- loaned or | otherwise. What interest would be charged on a loan for 1/2 of a cycle? It | is being charged through the energy metering. In addition, as I have said | earlier this evening- KVARH metering does not reflect the actual losses seen | by the utility. | In the case above it is not an energy related issue but a voltage control | issue which so far is quite a specific situation.

It is hard to tell from your writing whether your are saying that it is, or is not, an energy charge. kVAR is certainly not energy used. But a fraction of it is energy wasted, such as in voltage drop over the lines. In these days of separate energy providers, accounting for that energy waste is now important because a different company has to generate it.

| My initial comment was "Does the size and cost of a utility transformer and | lines feeding a specific plant depend on the time of day? If it were so, | then your suggestion would be great." However, I think that you are going | beyond this - doing "time of day" to bias the demand rate as is now done in | many cases for energy rates. Interesting but considering that the demand | charges are mainly related to the cost of equipment to supply a particular | load ( further back in the system the effect becomes more diffuse so the | costs of the immediate feed to the customer is the only one that can fairly | be attributed to that particular customer- the more diffuse costs are built | into the energy rate) and the equipment isn't going to change on a 15 minute | cycle- there is a problem. What sort of "demand" metering most accurately | represents costs?

Of course further back is more diffuse. With one supplier and distributor as the same company, it wasn't much to worry about. Today it is more so. And there is some fraction where the distrbution costs are increased due to industrial users who make extra demands on the system and don't get metered for it (if that is the case, which I think it is in many cases).

| You are absolutely right in suggesting that such testing be done at night or | off-peak times- The KVA demand is reduced and if time of day metering of | energy is in place, the energy cost would also be reduced. | Note that many large industries negotiate energy contracts which have a kVA | limit- if the peak kVA is below a certain value then a "good" energy rate is | available:-BUT if the kva demand exceeds this limit- the energy charge | increases drastically. Many industries with their own generation, use such | generation to shave the peaks and keep the charges from the utility down. If | one has such capability it works well to the advantqage of both the utility | and the customer. Many moe utilities use some form of pf correction- | reduction of kVA demand as well as improved voltages.

And of course there should be a drastic increase, since the utility has to invest in a specific level of capacity to provide the service. If you now expect them to just supply you with more energy or more VARs or whatever, they have to gear up for it.

| The idea is great but the problem is that back in the system- your load is | absorbed into a melting pot. What demand that is measured at a customer | level can be isolated as chargeable to that customer. How much effect that | it has further up the chain depends on the size of that customer with | respect to the sum total of all customers. Load diversity rears its head as | well. Demand charges are meant to address costs associated with supply to a | specific customer rather than overall system costs. . | Can this be extended to consider overall system costs - in theory - Yes. | However, in practice, when the cost of getting the data exceeds the saving | gained from use of the data- don't bother.

Things like load diversity can actually be figured out with all that data. Overall system costs need to be considered, and those costs apportioned to customers properly based on their usage, demand, waste, etc. But yes, if the cost of getting and working with the data exceeds the cost, then don't do it. But those costs _are_ getting lower while the costs of maintaining an electric grid go up (especially with the reliability and security issues now days being raised). Electric utilities will be expected to invest in more capacity, more reliability, etc. Those who are served more by that are the ones that should pay more for it.

| Interesting- the demand meters that I have seen (and tested one about 50 | years ago) had a thermal lag such that the reading would take about 20 to 30 | minutes to reach 90% of the actual kVA in a steady situation. In 10 seconds | the "kick" would be very small if any. The recognition that motor starting | occurs and does not normally have any real effect on equipment heating was | recognised. Repeated starts- another animal but I bet there were limits on | how often the pump could be started.

That's a good way to model the demand since it represents actual impact. When going to software based metering, it should be simulated if the data sensed is just raw samples.

Sometimes technology gets in the way.

A common demand recorder in the 1970's -1980's was a wide paper chart recorder. The one we had actually recorded dots on a calibrated strip, one dot per minute. The utility changed the chart once a month and interpreted the data manually. In the mid 1980's the paper gave way to magnetic tape.

FWIW, that plant took service at 115kV from two feeders, and had two metering systems. Demand ran around 30 MW. The utility had one customer who was larger and took about 50 MW, but in any event there were not more than a handful of customers who were metered like that.

I really don't know why you would need more than one data point per minute, and more probably 4 data points per hour for each parameter measured. The argument about vast amounts of data seems like a red herring to me.

I was hoping somebody would say that!!

I have never, ever seen a demand meter or tariff that measures demand with an integration period of less than five minutes, and the vast majority that I have seen are 15 minutes. Yes, yes, there are peak reading meters that respond more quickly but I am referring to meters used for utility revenue metering.

Does any one have a specific reference to such a tariff? Not anecdotal evidence, but an actual utility tariff that determines demand over less than

5 minutes?

Thanks in advance.

From a plant perspective, a leading (ie-capacitive) power factor generally gives you a little higher voltage. From a utility standpoint, it could be good or bad. Leading or lagging power factors impact losses the same way for a utility. So in the old days you would get penalized for either extreme. With the impact of de-regulation (actually re-regulation) and the developement of digital metering, more factors can be recorded and more billing variants can be negotiated.

In the Texas de-regulated market (a joke second only to California)there is a bitter fight going on as to who has to provide corrective vars...the generators or the transmission companies. The transmission companies want to dump that responsiblity on the distribution companies. And on and on!

Most loads are inductive (lagging). In plants, the Kvar penalty either compenstes the utility for additional losses/investment, or forces the customer financially to correct his own power factor. For residential loads, the utility has to correct the power factor at it's own cost.

So.....if you have a leading power factor, in most cases you are actually helping the utility by compensating for the other predominately inductive loads. Start a little negotiating with your utility and see if you can wrangle a deal.

Well, *most* of our motors and loads were under 500 hp. But the two emergency pumps were so much larger than any others that it warranted the effort to coordinate with DG testing. Your thermal-equivalent makes sense to me too, but what we had was a near instantaneous 'peak demand' recorder. Highest peak during the month cost us an extra fee. Not sure if it was just thermal KVA rating of utility equipment, or a 'penalty' for the voltage dip we caused on the utility lines. But it was there, so we dealt with it.

Yeah, very specific case. As I said, only saw this on one or two customers.

daestrom

electronically

I can attest to that. While working with DTE during 'deregulation', they had to supply a new meter to any industrial customer that wanted to buy power from an independent. These meters required that the customer have a dedicated phone line next to the meter. The meters were setup to 'phone home' once a week and download their 15-minute readings to DTE. If the line was busy or what-have-you, the meter would keep trying and store the information for as long as it could. How to deal with 'missing' data was a tangled issue with the service commission.

Then all the meter readings had to be allotted to each IPP and compared with their scheduled power. If the IPP generated more or less than their customers used, $$$ for DTE. Reconciling the monthly scedules of IPP's along with non-metered, and DTE customers was a b----.

daestrom P.S. You should have seen the number of ad-nauseum meetings about even such straightforward things as setting the clocks ahead/back each year for Daylight Savings Time ;-)

The near instantaneous measurement affecting costs appears to be a specific penalty for a specific situation. I think you were being screwed.

The demand charges that I have seen are not energy charges but simply a measure of the highest 15 minute or 30 minute) peak in the billing period. As I said, it doesn't account for system losses. Nor is kVAR a measure of energy wasted. If the pf is poor, the kVA/kW ratio will be higher and the losses higher but kVARH or kVAH does not provide a measure of this. However, a very good point that you have is that de-regulation (in addition to wipig out proper load forecasting and optimisation of the generation mix) does present problems in accounting for where the losses occur and whose responsibility they are.

-----------

If all industrial users do have demand charges then that is one way to get a handle on the problem-it doesn't cure all but the cure-all depends on a hell of a lot of information to be transferred and interpreted. There will be situations where more complex specific metering is needed to provide a fair assessment.

----- True. I think we agree. However, one could measure each and every load kW, kVAR,kVA, and all of these with H attached as well, but still not have a proper assessment of system losses - where, when and how much, that is sufficient for proper charging of losses to the customers. I would respectfully submit that KISS (keep it simple, stupid) still has a place. Utility x has %y of the total load and gets %y of the cost of losses - I bet the utility would then look at its customers and bill accordingly.

However, I am from the era where each utility had its own area and customer base. They could buy bulk from other suppliers and did but were able to plan their generation and transmission needs on a long term basis. Nowadays the planning horizon has shrunk, and optimisation of generation, along with economic dispatch and emphasis on reliability has been compromised.

Don,

The two utilities that I worked for in the "old days" that used customer electromechanical meters with either a demand register or that sent impulses to a printing demand meter did not record leading kVArh demand. The meters both for real and reactive energy were equipped with detents that prevented reverse rotation. The reactive meter was a Watt-hour meter with an external phase shifting transformer that shifted the voltage 90 degrees. In order to measure outgoing or leading reactive, a second VAr meter would have been required and I never saw such an installation. A fourth meter would have been similarly required for outgoing real energy. This has all changed with the new metering devices.

Regards,

John Phillips