Compressed-Air Energy Storage: What the Compressor Train Has to Do

On 12 September 2026 the two 350 MW units of Huaneng's Jintan salt-cavern compressed-air energy storage phase two completed an integrated start, at 700 MW total capacity and with the project stating 100% domestic core equipment (1). The compression train is the half of the plant that turns electricity into stored pressure. Its discharge pressure is not chosen from a catalogue — it is set by a cavern about a kilometre underground holding air at more than 130 bar.

Why this is not a process-compressor duty

Four things separate a storage compressor from a process machine.

The discharge pressure is set by the store. The machine is pushing air into a fixed volume of rock, salt or artificial cavern, so its discharge pressure is whatever the store is at when it arrives. The Jintan caverns run above 130 bar (2). At the Jiayuguan artificial-cavern station the first injection test reached a maximum of 12.63 MPa (3).

The machine's hours are decided by the grid. Jintan is designed for eight hours of charging and four hours of discharge, with about 330 charge-discharge cycles a year (4). A process compressor of comparable size normally runs near continuously at a steady duty.

The heat is a product, not a waste stream. The plants are non-supplementary-fired: no fuel is burned to reheat the air at discharge. The compression heat therefore has to be captured, stored and returned to the air before it reaches the turbine, or the round-trip efficiency collapses. At Jintan that heat goes into sixteen hot-water tanks of 3,500 m³ each, 56,000 m³ in total. The tanks are pressurised, which raises the boiling point of the stored water from 100 °C to 180 °C (2).

The load range is wide. The compressor has to work across the whole charge cycle rather than at a design point, and it has to start and stop with the grid.

The configuration

The compression side is built as a cluster rather than a single machine. At Jintan, the contractor's account of the integrated start records the A-line and B-line compressor trains being brought up together for the first time — four machines started in one attempt — followed by eight continuous hours of air injection into the phase-one cavern (4). ShenYang Blower Works, the compressor supplier for both Jintan phases, states that the project forms its large compression cluster from multiple compressor trains combined in series and parallel (5). Xiangtan Electric supplied eight large brushless-excitation synchronous motors to the compressor supplier as drive equipment, in three frame sizes (6). On the generation side, Dongfang Electric's turbine works supplied a 350 MW air turbine-generator set on a single shaft with triple reheat and multiple cylinders, configured for frequent grid peak-shaving duty (7).

The technical route, and where it came from

The machine class that solved this duty in China is the axial-plus-centrifugal combined train. The problem it addresses is specific: a storage plant needs, in one train, both the very large volume flow at low pressure ratio that a cavern charge starts with and the higher ratio at the top of the charge, at a pressure that ends above 100 bar.

Shaangu's combined axial-and-centrifugal scheme is the reference for the route. The company states the scheme holds 26 granted patents, that the combined compressor train reaches a combined efficiency of 90% with a design life above 40 years, and that the blading is fully adjustable on the stationary side to suit a storage plant's frequent start-stop operation (8). In 2026 the company's "key technology and application of high-efficiency compressor trains for 300 MW-class non-supplementary-fired compressed-air energy storage" was appraised jointly by the China Machinery Industry Federation and the China General Machinery Industry Association, and its own half-year report records the appraisal conclusion: overall international advanced level, with the axial-centrifugal series coupling technology and the train's start-stop speed at international leading level (9)(10).

The commercial record behind that: Shaangu states it has supplied compressor trains for more than ten compressed-air storage demonstration projects covering the 10 MW to 350 MW series, and holds the leading market share in the segment. The named projects include the Yingcheng 300 MW non-supplementary-fired project, described as the world's first of its kind, and the Zhuozi 3×350 MW project in Inner Mongolia (8). A state news agency account adds that the Yingcheng project, which the company supplied with its core compressor train, is credited with three world records — unit power, storage scale and conversion efficiency — with key core equipment and deep-space utilisation products 100% domestically supplied (10).

Where the route goes without salt

The salt-cavern route is limited by geology. The artificial-cavern route removes that limit, and its first full-scale Chinese application is at Jiayuguan in Gansu. The station is described as the world's first 300 MW-class artificial-cavern compressed-air storage plant; it reduced storage depth from the conventional 600 m to 110 m, and its No. 4 storage cavern completed a first injection test at a maximum pressure of 12.63 MPa, described as setting three records for the largest cross-section, the largest volume and the highest pressure in an artificial-cavern store. All core compressor trains for the project were supplied by Shaangu, and the plant is to deliver 594 GWh a year of green electricity (3).

The boundary of the compressed-air route

Six limits belong alongside the record.

The appraisal wording is a peer appraisal, not a measured field record. "Overall international advanced level, with axial-centrifugal series coupling and start-stop speed at international leading level" is the conclusion of a joint appraisal by two industry bodies (9). It is a technical judgement by peers, not an independently measured operating result.

The 90% figure is the compressor train's, not the plant's. The 90% combined efficiency and the 40-year design life are the supplier's own figures for the compressor train (8). The plant-level conversion efficiency reported for Jintan phase two is above 70% (11), and the phase-one plant's charge-discharge thermal efficiency is reported above 60% (11). These are three different levels and must not be presented as one number.

An integrated start is not commercial operation. Huaneng states that pre-grid-connection testing and system commissioning remain (1). No operating hours, availability figure or round-trip efficiency from Jintan phase two has been published, because the plant has not run a commercial cycle yet.

The phase-one operating record is at 60 MW. Phase one has run continuously for more than 1,500 days, has taken part in more than 960 grid peak-shaving events and has delivered more than 650 GWh, and entered commercial operation in May 2022 (2). All of that validates the route at 60 MW. It does not demonstrate the durability of a 350 MW train.

The cost position is not settled. One industry account puts the life-cycle cost of electricity for compressed-air storage at about 0.2 to 0.3 CNY per kWh, and frames the next phase for the industry as a cost fight rather than a technology fight (8). The route is proven; it is not yet cheap.

And the storage itself is the gating item. Salt-cavern siting is geology-bound and the artificial-cavern route is a rock-engineering project before it is a machinery project. The compressor cannot fix a site that does not have the geology, which is why the artificial-cavern demonstration matters more to the route's reach than any further gain in machine efficiency.

What the trend looks like

Two movements are visible in the same quarter. The machine side is consolidating on combined axial-and-centrifugal trains with adjustable blading, because that is what a wide load range on a fixed store requires. And the storage side is moving off salt, because salt is where the geology happens to be. In the first quarter of 2026, compressed-air storage under construction in China exceeded 5 GW with more than 36 GW at the planning stage (2) — a build-out rate that the salt-cavern resource alone cannot host.

What this article does not cover

Note that this article does not select a compressor train for a storage plant, does not calculate stage count, power, mass flow or round-trip efficiency, does not size a cavern or a thermal store, does not give prices, and does not compare suppliers. It describes the duty a storage compressor performs, the configuration and machine class that Chinese plants use for it, the projects that carry the record, and the limits of what has been demonstrated.

Related on this site

• Compressing CO2: The Pressure Is Set by the Phase, Not by the Distance
• Diaphragm vs Piston vs Screw Compressors: A Comparison Table

Sources