A common claim suggests that energy storage can lower the cost of natural gas generation by enabling plants to operate within a more efficient operating window, thereby reducing fuel consumption.
While this idea appears plausible at first glance, it oversimplifies the complexities of total system costs. Read on to uncover the full picture…
A Model Based Analysis
To evaluate this claim, we will analyze a simplified model of the Texas Grid (ERCOT) using 2024 data (one year of 5-minute interval data).
We will assess two scenarios:
In this analysis, "Gas Demand" refers to the combined total of gas generation and storage discharge. This assumption is realistic because, currently, the Texas Grid has no surplus wind or solar energy available to charge storage systems. As a result, natural gas is used exclusively to charge storage.
1) Base Case:
Natural Gas generation using CCGT Plus SCGT (Peakers) during peak demand periods.
CCGT is set to cover the first 15,000 MW of Gas Demand.
The balance is covered by SCGT (Peakers).
2) Time-Shift Case
Gas Demand that was covered with SCGT in the Base Case is covered by Energy Storage in this case.
Will cause gas generation to operate in a more efficient envelope.
Storage discharge demand is set to mirror SCGT Generation in the Base Case.
Figure 1 provides an overview of the system.
Discussion
Summary of Results
As anticipated, the Time-Shift Case shows a marginal efficiency improvement (57.13% vs. 57.09% in the Base Case).
However, any savings from this efficiency gain are offset by energy losses (54 GWh) incurred during storage charging.
To compensate for these losses, gas generation in the Time-Shift Case increases from 126,681 GWh in the Base Case to 127,222 GWh, a net increase of 541 GWh over the year.
Total annual costs for the Time-Shift Case are $129,293 Million compared to $117,114 Million in the Base Case, resulting in a net cost increase of $12,179 Million annually.
Conclusion: Utilizing Energy Storage to replace SCGT results in Increased Costs.
Methodology
Source for LCOE & LCOS was Lazard V18 (2025).
Current gas and storage generation was downloaded from the Texas (ERCOT) grid for 2024 so therefore represent real-time data.
Nat Gas generation was adjusted to compensate for:
Average Capacity factor effect on LCOE
Variable Costs (including Fuel) were tied to actual generation during each 5-minute period over the year.
Storage
LCOS was based on 100 Mw / 400 MWh facilities which represent the lowest cost option.
Charge and Discharge rate was constrained by maximum power rating. This was determined by the hours of storage. Although Lazard uses a 4-hour duration, this model modified that to 1.8-hours to increase power rating.
State of Charge limitations were imposed. SOC(High) was set at 90% and SOC(Low) was set at 10%.
Storage round trip efficiency was set at 90%.
Storage opening balance was set to match closing balance to remove any effect from a change in the balance over the year.
Storage Annual Output was compared to Lazard data. The annual output of the model was ~485,000 MWh vs 1,764,000 MWh used in Lazard’s calculation of LCOS.
The impact of this is that storage facilities would have to use Arbitrage to make up the difference to meet their fixed costs.
This would further increase the cost of the Time-Shift case. However for the purpose of this discussion, these are ignored.
Figure 2 shows real-time arbitrage by storage facilities on the Texas grid. This increases storage losses, and therefore gas generation. The cost of which is passed onto the consumer.
Efficiency Calculations
Where on the curve the model range resides is a function of Capacity Factor, which is itself reliant on on-line Capacity.
It is important to note, that while increasing capacity on line will increase the differential efficiency between cases, it will also increase the Fixed LCOE to pay fixed costs.
The following curves were used.
Gas Generation & Storage Discharge
Storage
Figure 6 shows the storage level.
Note that the model does not use the entire available capacity.
This is due to the restraint imposed by the total power rating of the storage facilities.
Efficiency
Although shifting gas generation to later in the day, the net gain in average efficiency is minor.
Love your state of charge/discharge graph. We will be stealing that concept for our work!
Is there really not enough wind and solar to charge the batteries? Is that because they are dispatched to meet load first rather than charge them?
I am travelling and don't have time to read fully but may be a jack hammer to crack a walnut. Consider the cost of storage to get through a few nights of a wind drought with no coal in the system.
Also consider that the storage will never be even partly charged without substantial fossil fuel in the grid, even when there is enough RE capacity to deliver 100% of demand on a good day.
Back of envelope estimates should do it.
See also Schernikau and Smith.