We engineers are great at estimating energy and carbon emissions and dealing with concrete systems: pipes, wires, flues – that’s our bag. One of the things we do poorly (but for some reason are too willing to do) is financial modeling relating to low and zero carbon generation.
For the last couple of years I’ve been working alongside financial and commercial bods who actually do know what they’re on about and it’s been a real eye opener. They might not know how to size a duct but they can tell you where your business is making money – and where it isn’t.
On a recent project I was looking at small CHP engines (5 – 30kWe) on a sheltered housing scheme. As part of that work, I put together a simplified financial model (with guidance from the bods) to quickly test whether a given option was worth looking at in detail. It was hugely useful and threw up some surprising results – for example, none of the small engines I looked at could pay back its capital cost in its lifetime. Ouch.
So based on that work, here’s the model. I’m using micro-CHP as an example but it’s just as easy to use for renewables.
The main principle is that we want to simplify everything down to a single year. Here are the steps:
1. Estimate your revenues (or value of output)
Assume you’re selling heat and electricity. The value of the heat is the price of the offset fuel with an allowance for plant efficiency. So assuming you’re offsetting heat from an 86% efficient gas boiler, the value of heat is:
heat price = gas price / 86%
The value of electricity is equal to the price of gas multiplied by a spark spread (the multiple of electricity price to gas price). We’re doing it this way so that it will be easy to do some sensitivity analysis and try out the effects of changing the spark spread.
There’s also a value associated with the fact that we’re generating electricity on site – this comes from the fact that we’re avoiding distribution and some other charges. Let’s say it’s 1.5p/kWh. So the value of electricity is:
electricity price = gas price * spark spread + 1.5p
So you can now take a range of gas prices and spark spreads to use in your sensitivity analysis:
min | avg | max | |
Gas price (p/kWh) | 1.7 | 2.9 | 4.0 |
Spark spread | 2.5 | 3.0 | 3.5 |
Table 1 – Sensitivity variables
Now for each kWh of heat and electricity delivered by your CHP, you know exactly what the value is. In this example, the electricity prices (p/kWh) break down as shown in the following table. We’re going to ignore heat for now and take it off our costs rather than add it to our revenues (hopefully this will be clear in section 2 below).
Spark Spread |
||||
Min |
Average |
Max |
||
Gas Price |
Min |
5.71 |
6.58 |
7.44 |
Average |
8.59 |
10.03 |
11.46 |
|
Max |
11.46 |
13.48 |
15.49 |
Table 2 – Electricity price (p/kWh)
2. Estimate your operating costs
Costs are based on the characteristics of the engine. So let’s take a well known small CHP engine:
5.5kWe Micro CHP | |
Electrical efficiency | 24% |
Thermal efficiency | 68% |
Electricity output (kWh/yr) | 33,000 (6k hrs x 5.5kWe) |
O&M contract (£/yr) | £1,250 |
Operating cost (p/kWh) | 3.79p (O&M cost / elec output) |
Table 3 – engine characteristics
We know that for each kWh of electricity produced, it costs 3.79p for O&M. The next thing to do is work out what it costs us to generate that kWh of electricity. To do this, you take the cost of the fuel input and subtract the value of the heat that’s created as a byproduct (this assumes all energy is being used on site):
cost of electricity = (gas price / electrical efficiency) – (heat efficiency / electrical efficiency * heat value)
So taking the sensitivity variables and engine characteristics laid out above, for our engine the cost of electricity looks like this:
min | avg | max | |
Gas Price (p/kWh) | 1.7 | 2.9 | 4.0 |
Spark Spread | 2.5 | 3 | 3.5 |
CHP elec cost (p/kWh) | 1.50 | 2.51 | 3.51 |
Table 4 – Cost of electricity generation (p/kWh)
3. Calculate your operating profit based on 1 and 2
If you take your electricity value per kWh and subtract your O&M costs and cost of electricity per kWh, you get your gross operating profit. At the simplest level we’re interested in whether the engine generates any profit. Here’s how it looks:
Spark Spread |
||||
Min |
Average |
Max |
||
Gas Price |
Min |
0.42 |
1.28 |
2.15 |
Average |
2.29 |
3.73 |
5.17 |
|
Max |
4.16 |
6.18 |
8.19 |
Table 5 – Operating profit (p/kWhe)
From this table, you can see that at average gas price and spark spread, the engine will generate 3.73p/kWh of profit. It’s helpful to put this in a graph.
So the engine is making money, but is that profit enough to pay back the installed cost of the engine?
4. Factor in the installed cost
In this next step, we want to calculate a cost per kWh based on the installed cost and total kWh run. But we also need to allow for the fact that the capital has a cost. At the simplest level this is equal to the cost of borrowing the money from the bank. Our financial bods use weighted average cost of capital (WACC) based on cost of borrowing and the minimum return on investors’ equity. But whatever, the principle is roughly the same. Let’s say the cost of capital is 7%.
Our engine has an expected life of 80k hours, which at 6k hours a year equates to 13 years. The installed cost is £18.5k. Putting this all together:
installed cost = pmt(7%,13,18500) / kWh per year
PMT is a function in excel to calculate the periodic repayment of a loan over so many years at such and such an interest rate. For our example, that gives us an installed cost per kWh of 6.7p.
So factoring in our cost of capital, the bar chart above now looks like this:
It’s immediately clear that the unit doesn’t support its own costs. In other words, it will never pay itself back. You can install it for other reasons (as a planning requirement, because you think it will save carbon, etc) but just understand that the engine will cost you money over its life, not make you money.
The basic principles of the method above are useful for CHP but you can also use them for biomass boilers or other LZC generation.
While I appreciate the argument I think that the financial model is a little pejorative.
You assume that the capital cost is covered by a loan charging 7% interest but take no account for how inflation drives up the price of energy over the life of the product.
In addition £1250 PA for O+M for what is basically a domestic product seems to be more than a little extortionate.
Plus, £18.5k for what is basically a boiler providing some electricity on the side seems to be unrealistic for a domestic market.
It seems to me that the Micro CHP manufacturer is living in cloud cuckoo land if they think they can charge that much. it seems to make much more sense to install a condensing gas boiler and spending the change on wind turbines or PV
Michael, thanks for the comment.
First, about the CHP provider living in cloud cuckoo land, I don’t entirely disagree. But the capital costs and O&M costs come directly from a quotation – they’re not my figures. This is a very well known small CHP unit aimed at sheltered housing, primary schools, etc. and it seems to be selling quite well (maybe because of good marketing). It’s not a dwelling scale unit.
But that’s just for the engine I happened to use as an example. The aim of the post is really to present a simple model to see whether it’s worth looking at a technology in more detail. The model isn’t meant to give an exhaustive assessment of financial performance. Having said that, using the model it’s easy to see what energy prices would be necessary for the unit to support its own CAPEX: just play with the gas prices and see when the EBT figures become greater than zero. If things look good enough, then you know it’s worth doing full financial statements for the life of the engine. If they don’t look good enough, you’ve just saved yourself a lot of time.
Re 7%, I think I was being generous. In reality, a developer’s WACC is likely to be significantly higher, making it even tougher for the engine to measure up.
I’m a little confused – please excuse any daft questions:
1. CHP is eligible for a number of support measures, like CCL exemption and enhanced capital allowances. I believe small-scale CHP will also qualify for the forthcoming RE heat and electricity tariffs.
What impact would that have on project economics?
2. Why introduce a spark spread? I know the economics of gas-fired generation depends on this but it’s not a measure most people are familiar with. You have to add distribution and other costs to make it meaningful.
If the aim is to produce a simple model, why not just consider retail electricity and gas prices directly?
3. As I understand it, the economics of CHP are crucially dependent on local heat and electricity demand profiles, and whether (and at what price) you can sell any excess heat or electricity. I appreciate that you want to keep it simple, but it’s not clear to me that you can ignore these factors and still have a useful model.
4. Could you confirm for me under what conditions your modelled Net Profit/Costs would apply? I assume it’s under the condition that either there’s no excess heat or electricity generated, or that any excess can be sold at retail prices.
Hi Gavin,
The idea with this model is to keep things simple and flexible to get an overall idea of profitability before moving on to more detailed modelling. Having said that, it’s pretty straightforward to build in other functionality, for example to allow for a proportion of the electricity to be exported or heat to be dumped. Going through your questions:
1. CCL charges are around 0.43p/kWh (small!). You can build this in or just take a look at the EBT figures and see whether you think LEC sales will tip the scales. Re the FiT and RHI, yes that’s my understanding too. Once the terms of these become clearer, it’ll be worth building in.
ECA’s become an important factor once you start looking at tax. The 100% write down in the first year means the date of your first tax payment is pushed back and there’s a small boost to your IRR. This model looks at earnings before tax so ECA’s don’t come into it.
2. The idea behind using spark spread is, again, to keep things simple. Electricity price almost always tracks gas. In the case of gas CHP, your economics are totally dependent on this figure so in my view it’s worth making it explicity in the model so you see the results in your sensitivity. Building in retail electricity prices would do the same job except you’d have to do the figures in your head to make sure you were looking at a range of spreads.
3. Good point. The ideal scenario is one where you’re selling all of your heat and electricity on site. As soon as you start to export electricity or dump heat, you lose revenues. If a technology doesn’t stack up in the best possible scenario (100% on site use) then you know it’s not a goer.
Having said that, again it’s easy to build in a factor for on site use.
4. Yes, that’s right. For the reasons described in 3 above.
I hope that helps.
Thanks Casey, though I’m still a bit confused about point 2.
You’re probably right to focus on the spark spread given that it’s so crucial to CHP project economics – a developer considering CHP needs to become familiar with the term if they aren’t already.
However, I’m not quite sure where the spark spread you refer to comes from. For example, if I want to base my model on historical values for this figure, where can I look it up?
Gavin, the CHPA might be able to help for historical spark spreads. The further back in time you go, the more useful it is to have someone on hand who knows why it was the way it was. As a rough sanity check for the last 5 or so years, try the price indices(pdf) published by Buy Energy Online.
Let me try a little thought experiment (apologies for the length. and the German).
I want to use your model to assess the suitability of a micro CHP unit for my house. First need to come up with reasonable figures for the gas price and the spark spread. The latest quarterly energy prices from BERR (tables 2.2.3 and 2.3.3) give unit prices for households going back to 1998.
Figures for DD payments in Edinburgh show gas prices in the range 1.5 – 2.8 p/kWh and electricity prices from 7.7 to 11.5 p/kWh. By inverting the formula you give above I can use these figures to calculate the historic spark spread:
spark spread = (electricity price – 1.5 p/kWh) / gas price
This yields a spark spread of between 3.5 and 4.4. That’s above the top end of your range, and the gas prices are consistently at the low end. However, I presume that’s because you consider mini rather than micro CHP, and so different prices apply.
I now want to use your model to assess a mini CHP unit installed in a small business. Repeating the exercise above but looking at industrial prices for the last 3 years (BERR table 3.1.2, small manufacturing) yields gas prices in the range 2.2 – 3.6 p/kWh and a spark spread from 2.0 – 2.7, which are close to the ranges you consider.
So here are my questions:
1. I want to use your model for different types of project, and I need to feed in reasonable values for the gas price and spark spread. Is the approach I’ve sketched out above reasonable?
2. Why do the distribution costs for electricity enter into the model? By generating my own electricity I avoid paying the retail price of electricity for every unit I generate. Why do I care what proportion of the retail price consists of distribution costs?
3. Why have you expressed the spark spread as a ratio? Wherever I can find the term used elsewhere it’s defined as the gross margin per unit of electricity generated by a gas generator. E.g. from Wikipedia:
Spark Spread = Price of Electricity – (Cost of Gas * Heat Rate)
Where Heat Rate is the electrical conversion efficiency of the generator. i.e. the spark spread is the difference between the marginal cost of producing electricity and the price at which it can be sold, measured in £/MWh (or p/kWh).
I like your idea of building a simple model. However I think you’ve used a peculiar definition of the spark spread.
This is really interesting and informative. There are, at current prices, many renewable projects that are unviable. Your WACC is generous, but lets assume that 7% is fair. I don’t understand how you have calculated the net profit, but I’ll assume you have done a simple NPV approach.
Energy prices are very low at the moment and are they likely to stay low? I would factor some increase over time. Secondly, there are Renewable Obligation Certificates (ROCs) involved in CHP (if surplus electricity is sold) and this is extra revenue. In fact biomass CHP ROCs are double. You also gain revenue from recycled values too. Do you have to use gas?
From what I understand, CHP is an alternative energy source, not renewable. And, although it is much more efficient than most energy sources, it is not as low carbon as, say, a biomass CHP.
I’m still learning a lot about this field – I’m studying for a Strategic Carbon Management MBA at the UEA – and enjoy reading your blog so please correct any errors made and I look forward to your comments.
Hi Ben, thanks for the comment. Just to point out that gas CHP isn’t eligible for ROCs – though very small gas CHP is likely to be eligible (maybe perversely) for FiT’s and renewable heat incentive.
5.5 kWe is pretty small for CHP, so not surprising it doesn’t stack up – that is why it has been decided to support CHP up to 50 kWe through the FiT mechanism (not RHI by the way unless it is renewably fuelled).
One thing I think is missed here is that the value of heat is not just the price of gas / efficiency. The avoided capex and opex for a conventional boiler should also be included unless it is retained as a backup / topup unit.
Hi Rufus. A conventional boiler will still be needed as backup, so I don’t think there should be offset costs here. You could argue that there will be some savings in consumables on the boiler opex but on this scale I don’t think it will be significant.
Re the FiT, I’m ambivalent about small CHP. The FiT should stimulate the LZC technology market so that costs come down (and subsidies become unecessary). For PV, I can see how this would work. For micro gas CHP, I’m not sure where the economies will come from.
I was hoping to use your model to test the economic viability of a household micro CHP system. Will I need to install a conventional backup here too?
Hi Gavin,
A simple model for a household system is going to be slightly less simple because effectively the backup boiler and CHP unit are bundled into one (so you can’t test the CHP element in isolation). As a result your heat to electricity ratio is going to change from moment to moment. I suspect that the only way to model this system is to do it on an hourly basis. In other words, it’s whole hog or nothing.
Thanks for your response. I just wanted to share that CHP does recieve ROCs IF energy comes from waste or biomass. See http://www.lowcarboneconomy.com/community_content/_tips_did_you_know/5473 for details. So, you’re right that gas doesn’t. But, why would you use gas?