You make a good point if technology is constant. In such cases the gradual reduction in availability, due to the increasing amount of down time for repairs would be the primary driver of any reduction in a plant's load factor. However, over the last 30 years the thermal efficiency of new CCGTs has increased from roughly 48% to 64% now. So the marginal cost of running a new CCGT has fallen by about 25%.
The effect is that new CCGTs push older one down the merit order, which means that new plants run all the time, while older plants are run less frequently and their load factors decline. The rate of decline depends on the rate of addition of new plants relative to the growth or decline in the requirement for gas generation, so the changes are not smooth or certain but over time the trend is clear.
Nuclear plants have low fuel costs and high capital costs. They are only built on the assumption that they will run for at least 40 years. They are expensive to decommission, so you don't want to do that until you have to. CCGTs (and OCGTs) are, in effect, disposable assets. They are (relatively) cheap to build and are run as long as they make money, but because their fuel costs are high it is not worth keeping them in service if the ratio of power to gas prices becomes unfavourable.
Can I offer a clarification. The decline in load factor as CCGTs is not due to a degradation in performance per se. It is a matter of choice or economic efficiency as more modern and higher efficiency plants enter the merit order above them. Older plants either can't cover their operating costs when competing with new, more efficient, plants or their start-up costs exceed the amount of money that they can make by running for 2 or 3 hours. Not only are newer plants usually more thermally efficient than older plants but a lot of effort has been devoted to reducing the cost and time required to go from zero output to 50% or full output.
There is a larger point. Advocates of renewable energy point to improvements in the size and efficiency of solar and wind plants as a result of experience and better technology. What is frequently forgotten is that gas CCGTs and OCGTs and coal plants have been transformed by technological developments over the last 30 years. One of the reasons why China and India stick with coal is that supercritical and now ultracritical coal plants are far more efficient than their predecessors. Further these improvements have been standardised. Two decades ago it was necessary to have highly skilled staff to run modern CCGTs or supercritical coal plants at close to their best. Now those skills have become widespread via standardisation.
Excellent article and should be essential reading at DENZ but wont be but surely NESO gets this stuff though and will at least exert influence. How they manage keeping gas generation on the system is now mission critical although at least we've got past the first stage in that they acknowledge that gas will be needed even in Milibrains NZ fantasy world.
In respect of your data analysis can you see how load factor in gas vs unreliables is changing as installed generation has been commissioned. Im thinking next year when Doggerbank, and others, comes on line that will further eat into gas generation (when its windy of course) further lowering the utilisation of the stations and thus their financial viability to owners.
Re the question in your second paragraph, I have been thinking about how best to capture this effect. It was too much for this article but I will return to the question in another article. There will be a tipping point at which no CCGT can expect to operate with a load factor of 80% or above. Keadby 2, which is new and has a high thermal efficiency, is the canary in the mine.
There is more going on than just thermal efficiency. I was very surprised to discover that Great Yarmouth & Kings Lynn CCGTs, which are 22 & 26 years old, had load factors of 82% & 85% in 2023-24. They were followed by Enfield Energy (24 years) at 79%. These three plants account for about 1.2 GW of capacity.
This is partly location - East of England and London - and I suspect that they have very favourable gas purchase contracts. Newer plants such as Carrington have significantly lower load factors - Carrington is 64% for the whole plant. West Burton B (10 years) is just over 60% for the whole plant. In that case there is a systematic pattern for several years of using Unit 2 much less than the other two units.
Thank you for this Gordon. Do you happen to know why the load factor of an individual gas generation plant would decline with time? I imagine the running efficiency of the machine stays fairly constant, but with age and wear and tear, there is more down time for repairs and maintenance.
The relatively short lifetime surprised me as I understand there are quite a few nuclear plants around the world running for several decades more. Then again, because of the high capital costs and energy output, maintenance is planned at the outset. So much to learn, so little time.
You make a good point if technology is constant. In such cases the gradual reduction in availability, due to the increasing amount of down time for repairs would be the primary driver of any reduction in a plant's load factor. However, over the last 30 years the thermal efficiency of new CCGTs has increased from roughly 48% to 64% now. So the marginal cost of running a new CCGT has fallen by about 25%.
The effect is that new CCGTs push older one down the merit order, which means that new plants run all the time, while older plants are run less frequently and their load factors decline. The rate of decline depends on the rate of addition of new plants relative to the growth or decline in the requirement for gas generation, so the changes are not smooth or certain but over time the trend is clear.
Nuclear plants have low fuel costs and high capital costs. They are only built on the assumption that they will run for at least 40 years. They are expensive to decommission, so you don't want to do that until you have to. CCGTs (and OCGTs) are, in effect, disposable assets. They are (relatively) cheap to build and are run as long as they make money, but because their fuel costs are high it is not worth keeping them in service if the ratio of power to gas prices becomes unfavourable.
Again, thank you.
Can I offer a clarification. The decline in load factor as CCGTs is not due to a degradation in performance per se. It is a matter of choice or economic efficiency as more modern and higher efficiency plants enter the merit order above them. Older plants either can't cover their operating costs when competing with new, more efficient, plants or their start-up costs exceed the amount of money that they can make by running for 2 or 3 hours. Not only are newer plants usually more thermally efficient than older plants but a lot of effort has been devoted to reducing the cost and time required to go from zero output to 50% or full output.
There is a larger point. Advocates of renewable energy point to improvements in the size and efficiency of solar and wind plants as a result of experience and better technology. What is frequently forgotten is that gas CCGTs and OCGTs and coal plants have been transformed by technological developments over the last 30 years. One of the reasons why China and India stick with coal is that supercritical and now ultracritical coal plants are far more efficient than their predecessors. Further these improvements have been standardised. Two decades ago it was necessary to have highly skilled staff to run modern CCGTs or supercritical coal plants at close to their best. Now those skills have become widespread via standardisation.
Excellent article and should be essential reading at DENZ but wont be but surely NESO gets this stuff though and will at least exert influence. How they manage keeping gas generation on the system is now mission critical although at least we've got past the first stage in that they acknowledge that gas will be needed even in Milibrains NZ fantasy world.
In respect of your data analysis can you see how load factor in gas vs unreliables is changing as installed generation has been commissioned. Im thinking next year when Doggerbank, and others, comes on line that will further eat into gas generation (when its windy of course) further lowering the utilisation of the stations and thus their financial viability to owners.
Re the question in your second paragraph, I have been thinking about how best to capture this effect. It was too much for this article but I will return to the question in another article. There will be a tipping point at which no CCGT can expect to operate with a load factor of 80% or above. Keadby 2, which is new and has a high thermal efficiency, is the canary in the mine.
There is more going on than just thermal efficiency. I was very surprised to discover that Great Yarmouth & Kings Lynn CCGTs, which are 22 & 26 years old, had load factors of 82% & 85% in 2023-24. They were followed by Enfield Energy (24 years) at 79%. These three plants account for about 1.2 GW of capacity.
This is partly location - East of England and London - and I suspect that they have very favourable gas purchase contracts. Newer plants such as Carrington have significantly lower load factors - Carrington is 64% for the whole plant. West Burton B (10 years) is just over 60% for the whole plant. In that case there is a systematic pattern for several years of using Unit 2 much less than the other two units.
A fascinating article, and have to say I’m shocked at the degradation in performance
Thank you for this Gordon. Do you happen to know why the load factor of an individual gas generation plant would decline with time? I imagine the running efficiency of the machine stays fairly constant, but with age and wear and tear, there is more down time for repairs and maintenance.
The relatively short lifetime surprised me as I understand there are quite a few nuclear plants around the world running for several decades more. Then again, because of the high capital costs and energy output, maintenance is planned at the outset. So much to learn, so little time.