So first off my opinion
Combined Heat and Power is one of those ideas that makes me immediately ask:
Why don’t we do this everywhere?
Which normally means I’ve missed something.
The basic idea seems incredibly sensible.
Generating electricity creates heat.
We need heat.
So rather than throwing that heat away, let’s use it.
That’s Combined Heat and Power.
CHP.
Electricity generation produces a lot of heat
A conventional thermal power station doesn’t turn all of the energy in its fuel into electricity.
Far from it.
A substantial amount becomes heat.
That’s just thermodynamics doing its thing.
Normally that heat has to go somewhere.
Cooling towers.
Cooling water.
Eventually the atmosphere.
So we’ve taken fuel containing useful energy, converted some of it into electricity and deliberately disposed of another part of that energy because we can’t use it.
CHP asks a fairly obvious question.
What if somebody nearby actually wants the heat?
So what does CHP actually do?
Instead of producing electricity alone, a CHP system produces useful electricity and useful heat.
That heat might be used by:
- factories;
- hospitals;
- universities;
- large buildings;
- district heating systems;
- industrial processes.
This can make the overall use of the original fuel much more efficient.
We’re getting two useful products from the same energy input.
I like that.
Location suddenly becomes really important
Electricity is relatively easy to move long distances.
Heat isn’t.
And that’s probably the first big limitation.
You can build a power station a long way from a city and transmit its electricity through the grid.
Moving useful heat tens or hundreds of miles isn’t nearly as convenient.
So CHP makes most sense where electricity generation and heat demand are reasonably close together.
A factory is an obvious example.
If an industrial process needs heat and electricity anyway, producing both on site starts looking interesting.
But what happens when you don’t need the heat?
This is where my initial:
Why don’t we do this everywhere?
starts running into reality.
Electricity demand and heat demand aren’t necessarily the same.
A district heating network might need huge amounts of heat on a freezing January evening.
It probably needs considerably less during a warm afternoon in July.
So you’ve now got two outputs from one system with two different demand patterns.
That’s an optimisation problem.
And obviously optimisation problems eventually lead me back to data.
This could produce some brilliant datasets
If I was working with a CHP plant I’d want to understand both sides of the system.
Fuel input.
Electrical output.
Heat output.
Temperatures.
Flow rates.
Electricity demand.
Heat demand.
Weather.
Equipment efficiency.
Electricity prices.
Potentially gas prices.
Then I’d want to understand how all of those interact.
If electrical efficiency falls, did something change in the plant?
Or did we deliberately change how it was operating because heat demand increased?
That’s an important distinction.
A number that looks like deteriorating performance might actually represent a perfectly sensible operational decision.
Again, context matters.
District heating makes this even more interesting
This is another area I’ve started finding quite interesting.
Instead of every building having its own boiler, heat can be produced centrally and distributed through insulated pipes.
Now you’ve effectively created another energy network.
Not an electricity grid.
A heat grid.
You’ve got producers.
Consumers.
Pipes.
Losses.
Temperatures.
Demand patterns.
Storage might even enter the picture.
From a data point of view that starts looking surprisingly familiar.
Metering.
Forecasting.
Asset monitoring.
Demand modelling.
Maintenance.
The commodity has changed, but a lot of the data problems haven’t.
Efficiency can be misleading without understanding the system
This is probably my biggest takeaway.
If I’m looking at a normal generator, I might calculate:
Energy in → Electricity out
and use that to understand efficiency.
For CHP that isn’t enough.
Some of the energy has deliberately become useful heat.
Ignore the heat and the plant might look terrible.
Include useful heat and suddenly the picture changes.
That’s a really nice example of why data people need domain knowledge.
The SQL query could be absolutely perfect.
The calculation could be mathematically perfect.
And the answer could still be completely misleading because we didn’t understand what the machine was actually designed to do.
So where is my head at?
CHP seems very sensible where you’ve got a reliable demand for both electricity and heat.
Particularly industrial sites and places where heat demand is concentrated.
But geography and demand matter.
You can’t just bolt “useful heat” onto every power station if nobody nearby actually wants it.
The bit I really like, though, is how clearly CHP demonstrates the relationship between engineering and data.
Before deciding whether a machine is efficient, you need to understand what outputs the machine is supposed to produce.
Sounds obvious.
But I suspect we forget that more often than we’d like to admit.
That’s where my head is at the moment.
Gareth Winterman