Three-Temperature Storage Design Widens Compressed Air Energy Storage Power Range

Engineers studying grid-scale energy storage have proposed a new compressed air energy storage configuration that uses three distinct temperature levels of stored heat to widen the range of power a single plant can absorb and deliver. The system, called a multi-temperature graded compressed air energy storage system, or MTG-CAES, was modeled by Gen Li, Huan Guo, Yujie Xu, Haisheng Chen and colleagues in a study published in the journal Carbon Neutrality. Under the baseline design parameters, the modeled output power range reaches from 279.17 megawatts to 682.36 megawatts, while the input power range spans 210.44 megawatts to 511.32 megawatts. The authors report an exergy efficiency of 70.47 percent and a round-trip efficiency of 72.34 percent under idealized assumptions, figures they explicitly describe as theoretical upper bounds rather than measured performance of a built facility.

The motivation comes from a well-known weakness of conventional compressed air energy storage. The technology, which compresses air into storage and later expands it through turbines to generate electricity, is valued for its large capacity, long storage duration and long operational lifetime, and it is one of the candidate technologies for coping with the intermittency and variability of solar and wind power. But the relatively narrow power range of conventional systems makes it hard to integrate them smoothly with fluctuating renewable generation and variable grid demand. When renewable input falls outside the range a plant can absorb, operators are forced into off-design operation, unstable power absorption or, ultimately, curtailment of wind and solar generation. The research team argues that most prior work on coupling renewables with compressed air storage has concentrated on efficiency, economics and dispatch strategies, while efforts to broaden the operating range have focused on internal components rather than new system configurations.

The new design attacks that problem by combining several energy conversion routes in one plant. During charging, electricity from wind turbines and photovoltaic modules drives a two-stage adiabatic compressor followed by a single-stage isothermal compressor, which pressurize air into a storage tank at a final pressure of 10 megapascals. Heat released during adiabatic compression is captured by thermal oil and stored in a second-gradient heat tank. Meanwhile, high-frequency wind power fluctuations that the compressors cannot absorb are not discarded: they are converted into heat by electrical resistive heaters and stored alongside solar heat collected by solar collectors. The isothermal compressor, which the authors note could be a spray-assisted reciprocating piston compressor, suppresses unnecessary temperature rise through precise thermal control, avoiding additional heat generation that would otherwise go to waste.

Discharging works through three temperature gradients. High-pressure air released from storage is first throttled down to 7 megapascals and preheated in a recuperator that captures waste heat from the last expander’s exhaust. In the first gradient, the air is then heated by high-temperature oil from a solar-charged first-gradient heat tank before expanding through the first expander, with interstage heaters also drawing on that same tank to maximize work output. The oil leaves that stage at 244 degrees Celsius and flows into the second-gradient tank, where it mixes with compression heat of similar temperature. The second gradient runs the same flow path on that blended oil, discharging oil at 168 degrees Celsius into the third-gradient tank, and the third gradient uses that recovered heat before the oil returns to the solar cold tank and compression cold tank in its original proportion. Valves routing the oil allow flexible switching among gradients, so the expander inlet air temperature, and with it the power output, can be adjusted to match grid demand rather than following a fixed sequence.

The study is a steady-state thermodynamic analysis, not an experiment or a demonstration plant. The system was modeled in MATLAB with fluid properties retrieved through REFPROP, and the model deliberately neglects pipeline pressure losses, thermal dissipation from air and fluids, and air leakage, on the grounds that previous studies have found these effects small. The authors state plainly that the results represent an upper-bound performance under ideal operating conditions and serve as a theoretical benchmark rather than direct practical performance. Charging time was fixed at 8 hours to represent a typical large-scale daily storage scenario, and the resulting baseline discharging time is 4 hours, determined by the stored air mass and the discharge flow rate.

Under those baseline parameters, the three gradients deliver designed output powers of 568.63 megawatts, 484.50 megawatts and 398.82 megawatts respectively, with gradient exergy efficiencies of 88.70 percent, 75.58 percent and 62.21 percent and discharge capacities of 140.87, 959.27 and 706.84 megawatt-hours. The designed input power range runs from 300 to 450.63 megawatts; when combined with a flow regulation range of 0.7 to 1.2 and off-design operation, this stretches to 210.44 to 511.32 megawatts, while the output range stretches to 279.17 to 682.36 megawatts. The sequential decline in gradient efficiency and output power follows directly from the falling heat source temperature at each stage: hotter oil heats the pressurized air more, producing more expansion work. Exergy destruction analysis identified the compression-expansion unit group and the heat exchange subsystem as the largest loss centers, at 16.21 percent and 8.03 percent of total input exergy, with the air tank throttle valve also contributing considerably.

The sensitivity analysis reveals trade-offs that any real design would have to navigate. Raising the adiabatic compressor pressure ratio increases the temperature of the recovered compression heat, lifting the temperatures of the second and third gradient tanks and improving overall exergy, energy and round-trip efficiency. But it also reduces the total stored air mass and shrinks the share of air allocated to the first, solar-heated gradient, which narrows the system’s power operating range. Very high pressure ratios also pose compressor design and manufacturing challenges and raise costs. Increasing the solar temperature works in the opposite direction on flexibility: it substantially raises first-gradient power and widens the operational power range, although energy and round-trip efficiency decline because the extra work gained cannot fully compensate for the increased solar input, and thermal oil is not suited to prolonged high-temperature operation.

The number of expander stages and the discharge-to-charge time ratio matter as well. More expander stages lower the expansion ratio per stage, raise exhaust oil temperatures and increase the solar thermal input required, and all system and gradient efficiencies decline as stages are added, while the duration of high-power operation lengthens. A smaller discharge-to-charge time ratio widens the power output range, because longer relative discharging times reduce air flow rates and compress the gap between the three gradients’ discharge powers. The wind-to-solar heat ratio offers another lever: shifting more heating duty to surplus wind power raises the energy storage power, which includes both compressor consumption and wind-powered heating, and broadens the charging power range, but it lowers efficiency because wind energy follows an electric-thermal-electric pathway with additional irreversible losses, unlike the thermal-to-electric route of solar heat.

The authors also sketch the economics. For systems of similar scale, a single thermal storage tank typically accounts for roughly 1 percent of total equipment cost, so the additional tanks required for the three-gradient configuration are expected to increase initial equipment investment by about 3 percent. Against that, the widened power regulation range could support more effective peak shaving and load-following, potentially increasing revenue through ancillary service markets and reducing renewable curtailment, though the authors caution that operational revenue depends strongly on market mechanisms and policy frameworks. They note that the added components and the flexible switching between gradients impose higher demands on control systems and regulation strategies, and they identify economic performance, transient behavior and coordinated control as targets for future work. Within those limits, the study presents a modeled route to compressed air storage plants that can flex across a much wider band of input and output power while, on paper, keeping efficiency high.

Subject of Research: Thermodynamic modeling of a multi-temperature graded compressed air energy storage system coupled with wind and solar power

Article Title: Thermodynamic characteristics of multi-temperature graded compressed air energy storage system

Article References: Li, G., Guo, H., Xu, Y., & Chen, H. (2026). Thermodynamic characteristics of multi-temperature graded compressed air energy storage system. Carbon Neutrality, 5(1), Article 21. https://doi.org/10.1007/s43979-026-00173-y

Image Credits: AI Generated

DOI: 10.1007/s43979-026-00173-y

Keywords: compressed air energy storage, MTG-CAES, energy storage, exergy efficiency, round-trip efficiency, wind power, solar thermal, thermal energy storage, power range, grid flexibility, thermodynamic analysis, renewable energy integration