More on Cogeneration Operating Costs

The last post showed that running cogeneration saves about 20-25% in operating costs when combining heating and electricity. These numbers are based upon the current retail fuel and retail electric prices and assume the standard single rate price for electricity. What would the economics be in the case of a varying electricity price? When would it make sense to run the Ecopower if the electric price wasn't fixed but varied as the wholesale price does on a real-time basis?

On the high power price side, it is easy to calculate. At $2/gallon propane cost, if power costs exceed $.34 per KWH, then it makes sense to run the Ecopower at full capacity. Under present conditions here in the Northeast, this can occur probably for a few hours during the heart of the cooling season in July or August, but the current regulatory inefficiencies in power pricing restrict this potential capability. We could sure use a smart grid with smart pricing.

On the lower power price side, what is the limitation? The economic-based decision that justifies cogeneration in the first place is that the cogenerator runs when heat and power are required, and when it does run, then the combined heat and power costs are lower than heating locally and buying power externally. We turn the problem around from the last post and ask what would the power cost need to be to break even by using the cogeneration versus the standard historical setup of purchasing all the power from the grid?

To get the same amount of heat from a co-generator, 1.3 gallons must be burned for every 1 gallon in a standard boiler. When we burn that extra 0.3 gallons, we need to make up enough money to pay for the power costs. The answer I get (after some algebra) for $2/gal propane cost is $.079/KWH.

Let's check that. In the traditional arrangement, we burn 1 gallon of propane and buy 5.82 KWH from the grid. At $.079/KWH, the total cost is $2.46 for heat and power. In the cogeneration case, we burn 1.3 gallons of fuel and generate 7.56 KWH for a total fuel cost of $2.6 minus a credit of $.14 for the extra power beyond what we would traditionally buy, giving a total cost of $2.46. Good, we got the right number.

Now let's vary the propane cost. At $1/gallon, the break-even power cost is $.04/KWH, at $3/gallon, the break-even power cost is $.12/KWH, and etc. The numbers are even better with a cheaper fuel like natural gas. Hopefully all these numbers aren't too eye-glazing. There is a more important point here, which we now finally get to.

The break-even power costs calculated are for the total power cost. In CT, the bill is divided into generation cost, delivery cost(s), and monthly service fee ($16/month). In the single rate bill, the generation cost last month was $.12217/KWH and delivery costs (including taxes, etc) totaled $.0574/KWH. So, at $1/gal propane cost, it would be cheaper to make the power in cogeneration mode than it would be to deliver it, let alone generate it. At current propane prices of around $2/gallon, with a delivery cost of $0.0574/KWH, we would need to get to a generation cost of $0.032/KWH (currently about 25% of the current single-rate power cost) in order for cogeneration NOT to make sense. Does this ever happen? Yes.

The wholesale generation power costs can be found at the ISO-NE website (www.iso-ne.com). In 2009 through April, the average generation price has been just under $0.05/KWH and about 14% of the time the cost has been under $0.032/KWH. In fact, on 25-Apr-2009 in the early morning, wholesale prices went to zero! They were literally giving it away (as seen in the following figure).




The irony is that during part of this time, we were generating power (see above figure), and given that it was early Saturday morning, we were almost certainly selling back to the grid--and we were saving money, because our average cost to purchase from the grid is $0.18/KWH and break-even generation cost is around $0.08/KWH. If during that time we had purchased the power from the grid ($0/KWH for the generation(!) plus $.0574/KWH for the delivery services), would could have heated the hot water cheaply via electric heat. On the other hand, if we had just delayed the hot water heat for a few hours, system costs would have been close to break-even with the wholesale price close to the break-even generation price. As a third alternative, if we were able to store more of the heat from 12 hours earlier during the prior day's peak price time, we could have saved some money and helped deliver power when prices were higher during the afternoon of 24-Apr.

The inefficiencies in the power grid system can be pretty glaring at times. What this note has tried to show is that with the single-rate power structure that most people use, cogeneration has a cheaper operation cost. Further, other system inefficiencies could be eliminated if real-time pricing were available and utilized by the consumer.

Co-generation Operating Costs

The generator in the basement provides for a good deal of local flexibility, at least in principle, with respect to our energy choices. In the present setup, the generator runs when heat is needed, either space heat or water or both. The plan or goal is to optimize this with our geothermal units, but that is a big project, and it will take a while. For the rest of this year's heating season, we will be running only the generator when we need heat--unless we get some anomalous cold wave, but given the lateness of the season, that is rather unlikely.

The question I've been thinking about is how to best reduce operating costs, and this is a first look at those issues. First up is the question of what are the co-generation operating costs? One would think this would be an easy question. For now, I will try to make it easy, but it could be complicated.

The State Legislature in Connecticut has been fairly progressive with respect to energy issues. In the recent past, they have opened up the electric grid to competitive pricing where companies can compete for customers, so some competition has come in. The Legislature has also required net metering, so any kilowatt hours sent back from solar or generators etc., will net against those that are pulled from the grid. One of the recent mandates coming from Hartford is that residential customers must have a variable pricing option. This is to encourage consumers to reduce demand during the peak hours. We haven't explored this option too far. An initial inquiry with our supplier, Connecticut Light & Power, was discouraging, for they didn't believe people with generation capabilities were allowed to use variable pricing. I've just looked at the CL&P website, and it appears that we are. This is definitely something to look into.

For now, let's use the KISS principle and keep it simple. Our current rate is the standard single rate residential rate that is a constant charge per kilowatt hour (KWH) used (plus a monthly fee that seems to ever be creeping up in price and is currently $16/month). This power rate has also been going up and is currently $.18/KWH. The propane costs are currently around $2/gallon. This has come way down from last summer, but it is still historically high.

So what does it cost? Per gallon of propane, the Ecopower makes 5.82 KWH/gallon. That number is derived from the data on the Ecopower spec sheet, and I'm suspicious of the accuracy of that last decimal place, but okay. At $2/gal, the price then is $.34/KWH, or almost twice the amount we are paying. Hmm....what's wrong? That's easy! We are also using the heat. If we were using the generator to make electricity only, the cost would be too high. So let's look at it from a heating perspective.

We (currently) use the Ecopower only when heat is needed, and because we are also generating electricity, we have to use more fuel than we would if we were only heating. The question is how much more? A typical new boiler is about 85% efficient. Some of the older ones are a lot worse, but let's compare to new equipment. The Ecopower, from a heating perspective, is 65% efficient (and 27% efficient for power generation). Let's assume then we use one gallon in a standard boiler. In the Ecopower, we need to use 1.3 gallons for the same heating needs. The advantage is that we'd also generate 7.56 KWH of electricity along with that heat. Let's compare costs.

In a standard residential setting here in CT, the heating and power costs would be $2 for the propane plus (7.56 KWH * $.18/KWH = ) $1.36 for the power costs or $3.36. With the Ecopower, we are charged the cost of 1.3 gallons of propane which in this example would be $2.60, and the power generated comes as a free by-product. Per the heating requirements for a traditional boiler using a gallon of propane, we are saving $0.76 or 23% in combined heat and power operation costs.

For this comparison to work, there are a few assumptions, namely that one also uses power, that the power can be sold back at the same rate it is generated (net-metering laws), that the amount of power used is more than what is produced with the co-generation. There are also issues with the relative pricing of power versus heating. For instance, in New England, our power prices are driven mainly by natural gas prices, so co-generation works well here. In the Midwest, power is driven mainly by coal prices, so one would have to compare local power prices with heating fuel prices.

These assumptions are important and drive the applicability of co-generation. For instance, it only makes sense to keep part of the local needs as co-generation. You don't want more power generation capacity than you use, and one must have good heating requirements to even consider the idea (so it may not be a good fit in the Southern USA unless it can be sized primarily for hot water needs, and even that may not make sense, because solar heating would be much better in the South).

For our needs here in CT, the 5 KW Ecopower is a good sizing fit, but we could have settled for something a little smaller if one had been available. For smaller houses with smaller demands, a smaller unit (1-2 KW in size) would be preferable. For larger demands, say small businesses, etc., a larger unit would be better.

In this simple comparison, it is clear the co-generation has lower operation costs. There is the question of whether or not it is worth the machine cost, and I defer that question to another post. It is also important to know that there are other advantages coming from the co-generation. Less energy is lost in the generation and transmission of the power with the typical energy losses as high as 70% in standard power generation. The vast majority of those losses are saved with co-generation. Further, the co-generation requires less transmission grid capacity which will be a big deal going forward. Getting more power lines up in the current environment will neither be easy nor cheap. Needless to say, I'm a big fan of co-generation. It has just come 20 years too late.

First Look at Ecopower Operation

It has almost been a year now since the generator was installed in the basement, and there are more data points to analyze than there is time to do it. The photo shows the engine from the side view. First let's review the motivation for buying this thing and then see how it has performed.

The decision to purchase the co-gen unit goes back to late 2006 shortly after the geothermal was installed. Geothermal is a great heating device, but the power usage is high. In New England, geothermal peak power is in the winter, and if everyone went to geothermal for cooling, the power peak would be in the winter instead of in the summer as it is now. That is pretty much true across the USA, although it gets hard to use geothermal for cooling in the deep south due to the high ground temperature. I'm digressing.

The concern was the large power requirements and dependence upon the grid for heating needs. Given a choice, I'd choose the domestic power grid over heating oil supplied by foreign providers who don't seem to care for us, which is why we went to geothermal in the first place. The geothermal unit sizing still required second-stage heating, and moving from heating oil to propane (natural gas is not available at our house still) was a better choice, and with a co generation unit, I'd get both heat and power and a fuel source that is about 70% domestic instead of about 70% foreign.

So, why not more solar. Definitely, more solar would be better, but the largest problem with solar, aside from the cost which is becoming less and less of an issue, is its intermittent nature. Solar power has a fixed generation schedule with frequent interruptions. These interruptions are difficult to forecast as the behavior of clouds is rather hard to predict. The other difficult problem is that these solar outages can extend for a couple of days, so there is some need for something other than solar (or lots of power and heat storage) to help smooth out the rather jagged production.

The other reason is the gross inefficiency of the electric grid. With close to 70% of the energy lost as waste heat, there is good reason to be generating power close to where heat is required, and residential heat is a large fossil fuel consumer, so generating power at the residential level has the potential of recapturing the heat lost in power production. In terms of cost, it is a trade-off between building a small number of large-scale power plants that are 30-50% efficient versus a larger number of small-scale generators that are 95%+ efficient. Historically, the large-scale system has dominated for a variety of reasons, not all of them rational, but with the semiconductor revolution of the 1980s, it is my opinion that the micro-generation paradigm is more efficient, potentially cost-advantageous, and definitely better in terms of energy security.

The following figure shows the amount of generated daily power from June, 2008 through March, 2009. The big gap in October is due to a computer outage, so daily Ecopower files that are normally available were not saved.





Notice in the figure there are three main periods. When the system was started, it was put in a maximum power generation mode. The amount of power generated was limited by how fast the heat could be dissipated. No system was installed to dissipate excess heat other than losses from the water tank, because it was not really cost-effective to do so (I may be changing my mind on that). Starting in late July, in an effort to reduce operating costs, we switched it to water heating and second-stage space heating mode only, so the Ecopower only ran when hot water (or second-stage house heat) was needed, and the average daily power was reduced as a result. It ran in this mode until the end of 2008.

At the beginning of 2009, the unit was switched to first-stage heat. The problem with this mode is that the generator is not sized to the heating needs of the house. It comes at a maximum 5KW power generation capability, and to meet the standard design criteria for house heating needs, it would have to be 2-3 times larger than it is. When thinking about the size of generator I wanted, 5KW seemed to be about the right size to match up the with geothermal needs and about what I'd want for backup power in case it was necessary. An important point is that with a co-generation unit, one still wants a second stage heating source for the very cold days. Otherwise, the unit required will be way too large for the local power needs.

Two points after 2009 are worth mentioning. The drop shortly after the New Year was a result of a discovered propane leak, so we shut off the Ecopower until it was fixed. The second one at the end of February was over a vacation and the house heating needs were dramatically reduced. The rest of the variability is mainly explained by changing temperatures and the required heating needs.

Currently, I think 5KW is too much given our plans and our current other equipment, but for now, it is still too little to heat the house by itself. We found this out shortly after switching to the cogen unit running by itself with no geothermal heating, for it was not enough to keep up with the heating load, and the house slowly became colder. For the colder days, I've come up with a compromise which is to put part of the house onto geothermal and part on co generation. This happens when the average daily outside temperature is near freezing or colder as seen in the next graph.



In this graph, we plot the average daily power generation versus outside temperature (which is recorded every 15 minutes by a thermometer outside the house and read by the Ecopower computer). Each dot represents one day. Remember that zero degrees Celsius is 32 degrees Fahrenheit. Only the days from August, 2008 are included. The important features are (1) the lower boundary where the Ecopower is used for hot water and 2nd stage heating needs, (2) the upper boundary where the limit is the 4.7KW power capability, and (3) the sloping boundary on the upper right where the Ecopower is used as first-stage heating. Above 15C (60 degrees Fahrenheit), there is no space heating need and the Ecopower is used only to make hot water. For the points in between the 3 boundaries, the heating was a combination of Ecopower and geothermal.

There have been a few issues with operation, but that will have to be discussed in another post. Bottom line is that most of the issues have been resolved, and the Ecopower is now running well and represents yet another step down for us in terms of fossil fuel usage and another step in the direction of energy security.

The next question is how best to operate the generator by itself and/or with the geothermal. This is not an easy question to answer, and it is one of the problems I'm working on now.

Solar Power Update

While the most interesting thing here is still the new generator in the basement (I'll get to it, I promise!), here is an update on the solar panels. We are coming up on the 3rd year in May, 2009, and there are enough data to get a good sense of how the system is performing.

The above graph shows the solar generation results. This energy issue has many moving parts, and the figure attempts to document the main changes associated with the solar panels. The black line is the actual amount of kilowatt-hours generated for each month from May, 2006 to Feb, 2009. The yellow filled area shows the expected amount of power production as calculated by the installers, Sunlight Solar in Milford, CT. The first month's value is low because the system wasn't put in until the first week or so had passed, but otherwise, the first year's worth of data fit the expectation pretty well. The differences between the expected and the actual can be from two reasons: (1) weather, and (2) model errors such as estimating the amount of shading from trees each month.

Continuing chronologically, there is a vertical red line in May, 2007 that marks the point where we cut down a tree in the front yard on the Eastern side and relatively close to the house. Notice how the amount of power generated increased after doing so. This fits with the notion that shading is a really big issue. Also note that the increase in power seems to be dominated by the summer months (May-Sep). The tree that was taken down affected the morning light when the sun was at higher angles. The Oct-Apr time was not affected (probably) because there are other trees that still shade the roof in the morning during those months.

Next up is the replacement of the older panels and the addition of 2 more panels in September, 2008 (as told in a prior blog entry). The light blue filled in area is the expected power using the original advertised expected generation multiplied by 1.24. The 24% increase is a quick estimate that comes from the addition of 2 panels plus the increase in panel efficiency. Also of note is the new power inverter that takes the direct current from the panels and converts it to the alternating current in sync with the power from the power company. I'm guessing that the actual generated power will exceed this expectation, but by how much we will have to wait and see.

There are a couple of other things worth noting. Even with the new panels, it appears that the array under performs in the winter months. This is probably a consequence of the shading model underestimating the amount of shade in the winter months. The winter is when the sun angle is lowest, and its light spends more time in the neighbor's trees then. This segues well to the next obvious point which is that more power is generated in the summer than winter. Actual power demand also peaks in the summer, and solar fits this pattern well. Solar power is peak power!

There are important economic consequences of the peak power fact, because peak power is much more expensive than base-line power. Power use increases during the hot summer months, primarily because of air conditioning. Solar power can meet a significant part of that demand while wind almost certainly won't. Wind is erratic and peaks in the Fall and Spring when the wind is the strongest. Wind will do nothing for you on those hot still summer days. Solar, however, will.

In terms of return on capital, the solar panels have saved over $1600 so far from a cost base of $10,878 (net cost after the CT 50% rebate but not including the federal tax credit, because we used that money to cut down another tree!). On an annualized basis, the panels have returned 5.4% (that is an after-tax equivalent number), and the rate of return is growing because of the newer efficient panels, the increased exposure due to the removed tree, and the increased power cost. Last year's return was an even 6%, and this year should be better, because power prices are still a high $0.18/KWH here in Southern CT. Sure has beaten the stock market over the last 3years!

New Solar Panels

Way behind on updates. Most interesting thing happening here is probably the co generation installation and operation. There are many items related to that which will take some time to write up. Until that can get updated, here is an update on the solar panels.

As mentioned in a previous post, the older Sunpower panels had a defect associated with them as one can see in the following photo. Apparently an outer coating was coming off of the surface.



Here is a picture of the Sunlight Solar employees (Milford, CT) taking down the old panels. They did this in mid September, 2008--shortly after they took delivery of the replacement panels from Sunpower.



Here is a picture of one of the new panels. These are the newer 220W panels. The older panels were rated at 215W peak power. The difference is the technological improvement over the last 2 years (about 1% improvement per year).




Finally, here is a recent shot of the new installed array. Before there were 12 panels, and now there are 14.



There are other upgrades with the new system besides the 2 additional panels. The improved efficiency has already been mentioned. The other improvement is the inverter that takes the direct current out of the panels and converts it to alternating current in sync with the electric grid. The earlier inverter was rated at 2 Kilowatts (KW) while the old panels had 2.5 KW peak output, so there was some loss when the old inverter maxed out in its capacity to convert DC to AC power. In fact, the old peak power I ever saw during operation of the old system was 1.948 KW, and during good sunny days, the system would stay just under 2KW in output for about an hour. The new inverter maxes out at 3 KW which is just under the peak power rate. I haven't measured the new actual peak rate yet....a good thing to put on the list of things to do. Given the inverter and transmission (from the roof to the basement where the inverter is located) losses, there should be some spare inverter capacity, so there should be not maxing out of the actual power output now with the new array versus the old array.

Before, the old array consisted of 12 panels with peak output of 2.5 KW with an inverter limit of 1.95KW. Now, there are 14 panels with peak output of 3 KW. This peak output has yet to be verified, but I have seen output above 2KW. In the spring when the peak output is expected, I'll make sure this is correct.

The new system represents approximately a 30% increase in actual output. While only a small step, it is nonetheless a positive one in the right direction.

Co-Gen Unit Installed--Finally!

After 14 months, finally, the Ecopower combined heat and power unit (CHP) is installed. It has been running now for about 3 weeks, and after some initial debugging, it seems to be working well. Above is a picture of the unit in our basement. The Ecopower is clearly in the foreground. The silvery thing behind and to the right is a new water storage tank that is in addition to my normal hot water tank (to the left and behind the copper pipes), and this new tank acts as a heat storage unit.

According to the installer, this unit is indeed the first residential co-gen unit installed in Connecticut.

The unit is very quiet. We never hear it upstairs while it is running. This surely beats having a noisey back-up generator in the back yard. Our CHP is presently grid-tied which means that it is not set up for backup power. In the case of a power failure, the CHP will also go down. This will change once the "island mode" electronics become available. Current target date for that is December, 2008, but given the amount of time it took for this thing to become installed, I'm not so optimistic that it will be ready then. We'll see.

I've been monitoring the power this generates (5 KW max power) on a daily basis and will post some graphs in due time. It seems to be generating about 40% of our monthly baseload power while satisfying our hot water usage (I can't say "hot water needs" because we definitely use more than we need!). The heating season is over, so we won't know about the winter power generation until it gets cold enough again in the late fall.

This past winter provided some opportunities to determine our (past) heating oil use vs. our geothermal heat pump power use, and I want to post that data first for comparison to our new system.

Funny thing about this co-gen unit. We agreed to purchase it in February, 2007. Since then, the oil situation has changed for the worse. While I'm still glad we got this thing, it seems our next goal will be to make the ecopower less critical to the daily house operation. This CHP technology is a great improvement efficiency-wise, but I can't help but wonder if it isn't coming 20 years too late.

Discoloration of Solar Panels

One problem I've seen with our Sunpower solar panels was that about half of them have air-pockets in them. You can see this in the above picture as the light green areas on the panels.

I've let this problem fester for some time, but while at the Earthday celebration (where our efforts were acknowledged), the company who installed the panels had a booth, and I told them about the problem. They said that this was a known defect, and that Sunpower will replace the panels.

After a few days letting the wheels turn on this, Sunpower has decided to replace all of my panels, because replacing only half of them would not work out well. Apparently, my panels are now obsolete (after 2 years!). So far, they have worked as advertised, despite the defects.

Bottom line: We get a new solar panel array sometime near Labor Day this year. Will update on that when appropriate. Until then, there is plenty to write about.