Saturday, April 22, 2023

Levelized Cost of Electricity Revisited: Accounting the Costs of Grid Integration that Increase Per Unit of Power as Variable Generation Increases

 

     Most determinations of levelized cost of electricity (LCOE) ignore grid integration costs. As share on the grid of variable generation, or variable renewable energy (VRE), basically wind and solar, grows the cost of accommodating that generation also grows. This means that the cost to accommodate a unit amount of variably generated power grows as percentage of wind and solar on the grid grows. LCOE has been criticized as a metric for investment decisions with good reason. It is a misleading metric. Some models consider accounting for avoided costs of building new generation (say for distributed resources addressing peak shaving) but those should also account for unavoidable costs (say for battery back-up, peaking plant back-up, or start-stop damage to peakers). Clearly, LCOE as a metric is inadequate and needs to be revised into something more functional. Scientist Joseph Fournier notes that those very significant extra hidden costs, including transmission and distribution, are passed on to consumers:

Ultimately, grid operators amalgamate the total costs of the system and pass these onto the retail consumer.”

“Those who use unit Levelized Cost of Energy (LCOE) of a unit facility are ignoring reality and guilty of misrepresentation.”

     In the early years of solar and wind cost accounting the cost of grid integration could be quasi-ignored but as integration costs per unit of power increase those costs become harder to ignore. Grid integration for renewables is more expensive than grid integration for more concentrated fossil, hydro, and nuclear power plants. Intermittency is the costliest variable of wind and solar when comparing to other resources. Part-time resources can’t run a full-time world without humungous spending on back-up generation resources: batteries, gas peakers, distributed energy resources that can be tapped, and the extra transmission lines and components to enable the shifting of loads when needed. Solar and wind need to be moved via transmission to where they are needed when they are being overgenerated at peak generation times or the power is lost in curtailment.

     Various methods of estimating full system costs attempt to quantify cost per MWh produced in order to compare full system costs for different energy sources. The EIA has a method and papers have been written suggesting different costs and ways to classify and attribute grid integration costs. In my 2022 book Natural Gas and Decarbonization I coined the phrase levelized cost of grid integration (LCOGI) The basic formula is: Levelized full system cost of electricity (LFSCOE) = Levelized cost of generation (LCOG) + Levelized cost of grid integration (LCOGI), or LFSCOE = LCOG + LCOGI. Grid integration costs are of several types. All energy sources have some grid integration costs, but variable generation has the highest costs.

 

Robert Idel writes in his 2022 paper in Energy:

 

“ … the function of supply in electricity markets is not to generate electricity but to provide a specified amount of electricity to a specific place at a particular time. The locational aspect adds significant additional costs to renewables that are generally less flexible about where they can be sited than fossil fuel plants. As a result, a larger grid is required to transport the electricity from, e.g., hydropower plants to the demand in urban areas. These transmission costs are partly taken care of in some LCOE estimates when a transmission cost adder is included in the LCOE. But the timing aspect turns out to be even more crucial and the focus of this paper.”

 

His explanation of the crux of the issue is pretty useful:

 

As long as the share of intermittent generation is low, sufficient dispatchable generation capacity will usually be available to step in and replace missing intermittent generation output. Economically, the fact that intermittent generation has no obligation to meet the demand can be seen as a hidden subsidy. One can even go one step further and argue that intermittent generation is of zero value if it cannot be made available to consumers who demand a steady electricity flow. To do that, however, supply and demand on the network must always be in balance. In effect, the ability to schedule other generators to continuously maintain that balance is necessary to give value to renewable output. The dispatchable generators thus raise the value of renewable generation, but the subsidy is “hidden” because the latter does not have to pay for it. Once the share of intermittent generation increases to a certain level (and dispatchable capacity is shut down), efforts have to be taken to maintain system reliability. But who should be responsible for these costs? How can the cost of integrating renewables into the system (which increases significantly with their market share) be included in the evaluation of their cost?

 

     As modern humans living in the modern world, we have expectations of affordable electricity being available to us. Up to $4 billion people around the world do not have adequate electricity access, about half of the people in the world. Intermittent electricity is obviously not an option so all intermittent sources must be backed up. Those needs drastically increase the costs of providing a certain amount of electricity for all 24 hours of a day and all seasons of the year. Thus, the real value of dispatchable electricity is much higher than that of electricity that is subject to the limitations of intermittency. While this is no doubt true, the total net difference in value is difficult to quantify. The value of electricity changes through time. The juice must be apportioned in a balanced way so on a grid electricity supply and demand must be kept in balance. This requires availability of generation to replace other generation when it becomes unavailable. It becomes unavailable on different time and intensity scales: no sun at night, less sun in the winter, less sun on cloudy days. Night and day are quite predictable. The seasons are generally predictable. The clouds are not predictable. Even though night and day are quite predictable electricity demand may or may not be predictable. We know there is a commonly occurring electricity supply-demand imbalance, the solar duck curve, on hot late afternoons in places with high solar penetration on the grid like California during high air conditioner usage after people return home from work and increase demand just as solar resources are going dark. But as we saw in August 2020 in California when this caused forced power outages, it was reserve capacity that was inadequate. There is a cost to having reserve capacity on stand-by ready to provide power. The inadequacy was partially due to the effects of heatwaves that cause outages at gas-fired plants so that dropped off some reserves, but it was mostly due to the sheer need for power. Other contingencies like low hydro output and unavailability of imported power from a nearby grid due to low wind output and their own higher needs due to heatwave, also contributed to the problem. Power system operators must prepare and make available if needed multiple power generation sources to be on stand-by ready to provide power if and when needed. Gas combustion turbine peaking plants that only run 5% of the time are a poor investment, can require significant maintenance due to starting and stopping frequency, put out more carbon emissions per MWh than combined cycle gas plants, and some or much of those emissions are attributable to changes in supply from intermittent sources. However, they are needed as an affordable way to provide reserves. There should be an accounting difference applied between more efficient gas power (combined cycle) and less efficient gas power (simple cycle gas turbine).


 


    Household Electricity Prices: U.S. Avg vs. California and Germany. Source: Alex Epstein. Fossil Future. 2022.



System LCOE or Levelized Full System COE (LFSCOE) – Models and Accounting

 

     Before I look at Robert Idels LCOE accounting I want to mention a few other attempts to quantify full system costs for economic analysis and for comparing the cost and emissions attributes of different energy sources. The following section is excerpted from my 2022 book, Natural Gas and Decarbonization:

 

New Ways to Compare: System LCOE to Account for Grid Integration Costs and Other Comparisons

 

     Through time the lack of fuel costs of solar and wind add up to operational savings. There is no need to pay-as-you-go. That is why levelized cost of electricity often show wind and solar as being the cheapest forms of energy during the full life of the projects. While this may be true, the upfront costs of wind and solar are quite high, as noted, over three times those of natural gas combined cycle. For solar in particular, but also wind, there may be grid integration costs that may not be attributed in the graphs below and those costs will increase as more of that variable generation penetrates grids. Gas is available 24/7. Solar is available in the daytime but much less in the short cloudy days of winter in the most populated parts of the US. Wind is intermittent and variable. The second of the two graphs below, shows that it is lifetime fuel costs that make lifetime LCOE slightly higher for natural gas relative to wind and solar.

 



 

2026 Expected Total Capacity-Weighted Levelized Cost of Electricity (for life of plant). Data Source: Energy Information Administration. (Note: by 2026 wind PTC is expected to be phased out and solar ITC is expected to be phasing lower and adding it in would put solar PV at 29.04)


  


  

2026 Expected Total Capacity-Weighted Levelized Cost of Electricity (for life of plant) and Comparison of Upfront Costs and Lifetime Fuel Costs. Data Source: Energy Information Administration. (Note: by 2026 wind PTC is expected to be phased out and solar ITC is expected to be phasing lower and adding it in would put solar PV at 29.04)

 

 

     The EIA’s method of calculating final or total capacity-weighted levelized cost of electricity is a complex formula. They consider and calculate levelized avoided cost of electricity (LACE) and levelized cost of storage (LCOS) where applicable. LACE seeks to estimate the value of dispatchability which addresses some of the grid integration costs. No fuel costs help wind and solar catch up economically through time, somewhere around year 30. Wind and solar are slow to break even, but they are good investments through time. As we gain better knowledge of how to integrate them effectively and lower those costs, we can reduce future costs. Perhaps we need to factor in a levelized cost of grid integration (LCOGI). This cost will vary by generation sources, supply, and demand. It will also increase through time as grid integration becomes more complex as expected with higher amounts of variable generation. This will also add to the solar and wind costs through time as integration costs are expected to rise through time which offsets both efficiency gains and cost reductions.

     In comparing levelized costs there are different ways to present those costs to suit the narrative that the comparer wishes to advance. LCOE is often not a good way to compare real world costs. There are differences in cash flow, opportunity costs, discount rate, and present value. The simple fact that wind and solar have the lowest “discounted lifetime average generation costs per unit of energy ($/MWh),” or LCOE, does not mean wind and solar are more competitive. Different kinds of energy have different values. Reliable dispatchable energy is worth more to the whole than variable generation. A 2016 paper in the International Association for Energy Economics, Why Wind is Not Coal: On the Economics of Energy Generation, does indeed attempt to quantify value of different sources of energy. In the abstract of the paper they explain this value as follows:

 

Electricity is a paradoxical economic good: it is highly homogeneous and heterogeneous at the same time. Electricity prices vary dramatically between moments in time, between location, and according to lead-time between contract and delivery. This three-dimensional heterogeneity has implication for the economic assessment of power generation technologies: different technologies, such as coal-fired plants and wind turbines, produce electricity that has, on average, a different economic value. Several tools that are used to evaluate generators in practice ignore these value differences, including "levelized electricity costs", "grid parity", and simple macroeconomic models.”

 

     The authors also note that LCOE and other common comparison methods introduce bias by not properly valuating the reliability, dispatchability, and flexibility of resources like natural gas. They introduce the concept of System LCOE as a comparison method. I’m guessing this should include my LCOGI designation. They mention two biases introduced by LCOE: 1) it favors base-load generation over peak-load generation and 2) at high renewables penetration it favors that renewables generation over dispatchable generation. To get a better understanding of the limitations of LCOE and why it needs reconciled I include a quote from the World Resources Institute from an August 2019 article by Laura Malaguzzi Valeri, INSIDER, Not All Electricity is Equal – Uses and Misuses of Levelized Cost of Electricity (LCOE):

 

The LCOE metric is concerned only with costs. By ignoring the revenue or value of the electricity generated, it implicitly assumes that all technologies provide similar services. However, picking electricity sources is not the same as choosing among brands of gasoline, as not all electricity is created equal. Electricity generated now provides a different value than electricity generated several hours from now because demand for electricity varies over time and electricity storage is expensive. Electricity generated farther from consumption centers is more expensive than nearby generation because transmission is costly. Electricity generation that is easy to predict is more valuable than unpredictable generation because it helps electricity system operators maintain the balance between demand and supply. Electricity generation that emits more pollutants is more harmful than cleaner generation. This is why LCOE on its own is insufficient for determining which technology investors or utilities should build.”

    

     A February 2021 research paper in the International Journal of Energy Research seeks to quantify the grid integration costs of solar PV. This study is from Italy where combined cycle natural gas is currently the largest electricity generator. Integration costs are separated into grid costs and balancing costs. Grid costs include transmission costs, adequacy costs, curtailment costs, and the costs of reinforcing the distribution network. Balancing costs include start-up costs and decay of efficiency costs. Italy was separated into regions for the study. One of the conclusions I gathered based on charting and graphing the solar PV integration costs was an avg. of about 31% additional costs for system LCOE than for the solar PV LCOE alone. This is for solar PV without storage. Adding storage in a solar-plus-storage has the additional costs at 24% for system LCOE but the costs of storage would cancel that out and more, especially in the short-term.  Comparing this to the EIA’s total system LCOE suggests that the EIA is leaving out some amount of integration costs, particularly for PV solar. It calculates LACE which addresses some of that and also accounts some distributed energy advantages of renewables, but the challenges outweigh the advantages in many cases at present. The EIA total system LCOE comparison shows that in 2026 it will be similar for CCGT, solar, and wind, the three cheapest widespread resources. CCGT will be a little higher in that scenario but does retain its value as a dispatchable fuel, its lower up-front costs per electricity produced, and its pay-as-you-go cashflow advantage. The Italian study suggests high integration costs for solar PV at 31% of the standard LCOE of solar PV. It is not certain how the EIA analysis calculates integration costs. With LACE they attempt to measure economic competitiveness of different technologies. They use a metric LACE-to-LCOE (or LACE-to LCOS in the case of storage), This is called the value-cost ratio. A value-cost ratio of less than one indicates that cost exceeds value and a value-cost ratio of greater than one indicates that value exceeds cost. The EIA cost-value ratios for different technologies suggest that by 2026 solar, some solar-plus-storage, geothermal, and CCGT will be the most attractive investments in some areas. Solar PV is best at 1.06, CCGT is even at 1, and onshore wind is at 0.98.  


 



  

 

System Levelized Cost of Electricity and its Components, a solar PV case study. Source: Veronese, E, Manzolini, G, Moser, D. Improving the traditional levelized cost of electricity approach by including the integration costs in the techno-economic evaluation of future photovoltaic plants. Int J Energy Res. 2021; 45: 9252– 9269. https://doi.org/10.1002/er.6456

 

     Now we return to include other attempts to quantify System LCOE. We can see that there are many things that can be accounted to get an idea of the true cost of intermittent resources. Unfortunately, many have to be apportioned or attributed to renewables costs or fossil energy costs in varying amounts. Fossil fuel power plants need transmission lines but far less total length in lines than renewables do. Apportionment can be tedious and error prone. The goal of comparing resources is to compare apples to apples, oranges to oranges, without skipping anything major or apportioning incorrectly. Fossil systems must account for the fuel, which costs varying amounts through time. Robert Idel’s analysis is perhaps starkly realistic, showing that it is the hidden subsidy that values renewables at low cost to provide when in fact they are expensive to provide. We know wind and solar cost more to build as they have high upfront costs, but renewables do not have fuel costs, so renewables increase in value through their life cycle compared to resources that need fuel.

     We know that grids with high amounts of wind and solar provide electricity that costs more for consumers than grids with lower amounts of it. One might also say that cheap natural gas has masked some of those costs for grids that increased natural gas as well as wind and solar on their grids. We also know that grid integration costs rise as more variable generation is added to the grid. How much is the question.

 

     A 2018 modeling study applied to the EU-7 plus Norway and Switzerland concluded that “the system LCOE for VRE increases linearly with the penetration level range of 20%–80%, above which it increases sharply.” This suggests that variable renewables can be accommodated at reasonable costs up to 80% penetration. Grids with high VRE share typically use wind as the main generating source and have a significant transmission buildout to accommodate that wind power. Indeed, the study also concluded that adequate transmission capacity is the key to accommodating variable generation. However, it is also often expensive and slow to add transmission capacity. The authors also noted that system LCOE estimates for grids with high VRE share can vary widely. They delineate grid integration costs as “transmission, storage, ensuring the flexibility of other units, balancing costs, and for reducing the full-load hours in thermal power plants or curtailing VRE generation.”

     Ueckerdt et al in 2013 first described the notion of System LCOE to account for grid integration costs. The paper describes LCOE as “the full life-cycle costs (fixed and variable) of a power generating technology per unit of electricity (MWh).” They also noted then that LCOE is it is/was is a flawed method that needs to be refined and that more variability = higher costs. Power systems were modeled after Ueckerdt etal’s marginal system LCOE of VRE2 for up to 60% VRE but not higher. The 2018 EU modeling study mentioned above attempts to evaluate higher than 60% VRE. The authors note that the two biggest strategies for managing variation are transmission and storage. They also note that earlier modeling did not consider trade between nearby grids, a variation management strategy enabled by transmission. Robert Idel notes that Ueckerdt etal’s System LCOE accounts for integration and balancing but is incomplete. He notes that LCOE of renewable sources of electricity depend highly on their market share. Wind generation costs remain constant when more is added but costs increase significantly. He notes a calculation from the Ueckerdt etal method where the System LCOE for wind in Germany increase from 60 EUR/MWh to almost 100 EUR/MWh if the share increases from 0% to 40%.

     The goal of System LCOE is to account for all costs, mostly grid integration costs, per unit of energy produced. Authors of a 2022 paper on redefining System LCOE point out the non-linear nature of power generation. Intermittency creates a demand for balancing that increases non-linearly with higher VRE on the grid. They advocate to make System LCOE more mathematically precise so that it can be standardized to weigh for policy. They consider that System LCOE isolated for each technology should be developed in order to make more detailed analyses. In terms of markets the great variability and ranges in estimated LCOE for power projects makes it hard to compare the economic merits of projects.

 


Avg. System LCOE (not a good metric). Source: 
Re-Defining System LCOE: Costs and Values of Power Sources. Yuhji Matsuo.The Institute of Energy Economics, Tokyo 104-0054, Japan. Energies 2022, 15(18),6845; https://doi.org/10.3390/en15186845. September 2022.




Marginal System LCOE (a better metric). Source: As above.


As can be seen from the graphs above, the Marginal System LCOE for wind doubles that of coal for a VRE share of about 75% and increases much faster after 80%. We can also see that wind was never cheaper than coal under that metric, as it is in the graph above of the inadequate and incomplete accounting of the Avg. System LCOE, where wind is cheaper than coal till about 50% VRE share and only accelerates slightly beyond 75-80% share.

     A table from Idel’s paper seems to give solar and wind very high LFSCOE’s but I am not sure the context here, it could be for higher VRE penetration.


 


     


     The bottom line perhaps should be a reminder that when people say that wind and solar are the cheapest forms of energy, that is not quite true at all, not even close really.   

 

 

References:

Levelized Full System Costs of Electricity. Robert Idel. Energy. Volume 259, 15 November 2022, 124905. Levelized Full System Costs of Electricity - ScienceDirect

Joseph Fournier - LinkedIn post, April 2023.

Natural Gas and Decarbonization. Kent C. Stewart. 2022. 

Re-Defining System LCOE: Costs and Values of Power Sources. Yuhji Matsuo.The Institute of Energy Economics, Tokyo 104-0054, Japan. Energies 2022, 15(18), 6845; https://doi.org/10.3390/en15186845. September 2022.

Lion Hirth, Falko Ueckerdt, and Ottmar Edenhofer. Why Wind Is Not Coal: On the Economics of Electricity Generation. The Energy Journal, 2016, vol. Volume 37, issue Number 3. Abstract. EconPapers: Why Wind Is Not Coal: On the Economics of Electricity Generation (repec.org)

Valeri, Laura Maguzzi. INSIDER, Not All Electricity is Equal – Uses and Misuses of Levelized Cost of Electricity (LCOE). World Resources Institute, Ausut 1, 2019. INSIDER: Not All Electricity Is Equal—Uses and Misuses of Levelized Cost of Electricity (LCOE) | World Resources Institute (wri.org)

Veronese, E, Manzolini, G, Moser, D. Improving the traditional levelized cost of electricity approach by including the integration costs in the techno-economic evaluation of future photovoltaic plants. Int J Energy Res. 2021; 45: 9252– 9269. https://doi.org/10.1002/er.6456

Levelized Costs of New Generation Resources in the Annual Energy Outlook 2021. Energy Information Administration, February 2021. Levelized Costs of New Generation Resources in the Annual Energy Outlook 2021 (eia.gov)

The marginal system LCOE of variable renewables – Evaluating high penetration levels of wind and solar in Europe. Lina Reichenberg, Fredrik Hedenus, Mikael Odenberger, and Filip Johnsson. Energy. Volume 152, 1 June 2018, Pages 914-924. The marginal system LCOE of variable renewables – Evaluating high penetration levels of wind and solar in Europe - ScienceDirect

 

System LCOE: What are the costs of variable renewables? Falko Ueckerdt, Lion Hirth, Gunnar Luderer, Ottmar Edenhofer. Energy. Volume 63, 15 December 2013, Pages 61-75. System LCOE: What are the costs of variable renewables? - ScienceDirect

 

 

No comments:

Post a Comment

     The SCORE Consortium is a group of U.S. businesses involved in the domestic extraction of critical minerals and the development of su...

Index of Posts (Linked)