What goes wrong: Long-duration energy storage (LDES) — common failure modes and how to avoid them

Form Energy's iron-air battery pilot in Weirton, West Virginia, disclosed 14 unplanned system trips during its first 18 months of grid-connected operation, each costing the project an estimated $40,000 to $120,000 in lost revenue and emergency maintenance. This experience is far from isolated: a 2025 BloombergNEF survey of 87 long-duration energy storage installations worldwide found that 72% reported at least one significant failure event within the first two years of operation, with median unplanned downtime of 22 days per year. For investors evaluating LDES projects across emerging markets, where grid conditions are less predictable and maintenance ecosystems less mature, understanding these failure modes is the difference between a performing asset and a stranded one.

Why Failure Analysis Matters

The global LDES market is accelerating rapidly. The Long Duration Energy Storage Council's 2025 assessment identified over $45 billion in announced LDES projects across 38 countries, spanning flow batteries, compressed air energy storage (CAES), liquid air energy storage (LAES), gravity-based systems, and iron-air batteries (LDES Council, 2025). The US Department of Energy's "Long Duration Storage Shot" initiative targets a 90% cost reduction to $0.05/kWh for systems delivering 10 or more hours of storage by 2030, channeling $1.2 billion in federal funding toward pilot and demonstration projects.

Emerging markets represent the fastest-growing segment of LDES deployment. India's National Energy Storage Mission targets 40 GWh of LDES capacity by 2032. South Africa's Integrated Resource Plan calls for 5.6 GW of storage including LDES to support renewable integration. Chile, Morocco, and Indonesia have all issued LDES procurement targets within the last 18 months. Yet these markets also present amplified operational risks: ambient temperatures exceeding 45 degrees Celsius, dust and humidity exposure, limited access to specialized maintenance crews, and grid voltage and frequency fluctuations that stress storage systems beyond design parameters.

The financial consequences of failure are severe. A typical 100 MW / 1,000 MWh LDES installation represents $300 million to $600 million in capital investment, with projected revenues of $15 million to $30 million per year from energy arbitrage, capacity payments, and ancillary services. Each day of unplanned downtime costs $40,000 to $80,000 in lost revenue and may trigger contractual penalties under capacity supply obligations. For investors, failure modes directly translate to internal rate of return erosion: a 2025 Lazard analysis found that unplanned downtime of 30 days per year reduces project IRR by 2.5 to 4.0 percentage points versus the base case (Lazard, 2025).

Electrolyte Degradation in Flow Batteries

Vanadium redox flow batteries (VRFBs) and zinc-bromine flow batteries represent the most commercially deployed LDES chemistry. Electrolyte degradation is their most common and costly failure mode.

Vanadium Precipitation

VRFBs circulate vanadium electrolyte in sulfuric acid solution through membrane-separated half-cells. Vanadium(V) ions are thermally unstable above 40 degrees Celsius and precipitate as vanadium pentoxide (V2O5), clogging flow channels, damaging membranes, and reducing battery capacity. Rongke Power's 200 MW / 800 MWh VRFB installation in Dalian, China, the world's largest flow battery, experienced vanadium precipitation events during its first summer of operation when ambient temperatures exceeded design assumptions by 8 degrees Celsius. The resulting capacity loss of approximately 12% required partial electrolyte replacement at a cost of $4.2 million, given vanadium's spot price of $28 to $35 per kilogram in 2025 (Rongke Power, 2025).

Prevention strategies include: maintaining electrolyte temperature below 35 degrees Celsius through active cooling systems (adding $8 to $12 per kWh in capital cost); adjusting vanadium concentration downward from 1.8M to 1.5M in hot climates (sacrificing 15% energy density for thermal stability); and installing real-time viscosity and spectrophotometric monitoring to detect early-stage precipitation before it reaches flow channels.

Membrane Crossover and Capacity Fade

Ion-exchange membranes in flow batteries separate the positive and negative electrolyte half-cells but are not perfectly selective. Over time, vanadium ions cross the membrane, causing electrolyte imbalance and gradual capacity fade. Industry data from Sumitomo Electric's VRFB installations in Japan show capacity fade rates of 0.5 to 1.5% per year from crossover alone. For a 100 MWh system, this translates to $150,000 to $450,000 in annual revenue loss.

Mitigation requires periodic electrolyte rebalancing: either mixing the two half-cell electrolytes and electrochemically resetting them (causing 1 to 3 days of downtime per rebalancing cycle) or continuous monitoring of electrolyte state-of-charge with automated electrolyte transfer between tanks. Invinity Energy Systems now ships its VS3 flow battery platform with integrated rebalancing systems that maintain capacity within 2% of nameplate over 20-year design life, though this adds approximately $15 per kWh to system cost (Invinity Energy Systems, 2025).

Mechanical and Thermodynamic Failures in CAES and LAES

Compressed air and liquid air energy storage systems rely on turbomachinery, heat exchangers, and pressure vessels that introduce conventional mechanical failure modes at unconventional scales.

Cavern Integrity in CAES

Underground CAES systems store compressed air at 40 to 80 bar in salt caverns, depleted gas reservoirs, or hard rock formations. Cavern integrity failure, the loss of air containment through geological fractures or wellbore leaks, can render the entire storage asset inoperable. The McIntosh CAES plant in Alabama, one of only two operational utility-scale CAES facilities globally, has operated since 1991 but required a $7 million wellbore remediation in 2019 after detecting a 3% per day air leakage rate through casing corrosion at the cavern interface.

Hydrostor's Advanced-CAES (A-CAES) technology addresses this by using water-compensated caverns that maintain constant pressure regardless of air volume, but the approach requires suitable geology and hydrogeological conditions that limit site selection. For emerging markets, the geological assessment alone can cost $2 million to $5 million and take 18 to 24 months, a timeline risk that investors must incorporate into project development schedules.

Thermal Store Degradation in LAES

Liquid air energy storage systems liquefy air at negative 196 degrees Celsius during charging and vaporize it through expansion turbines during discharge. The cold reclaim thermal store, which captures cold energy during discharge for reuse during the next charging cycle, is the system's efficiency linchpin. Thermal store degradation, from thermal cycling fatigue, moisture ingress causing ice formation, or heat exchanger fouling, reduces round-trip efficiency from the design target of 55 to 60% toward 40 to 45%, at which point the system becomes economically unviable.

Highview Power's 250 MWh LAES demonstration in Carrington, UK, documented a 4 percentage point round-trip efficiency decline over its first three years of operation, primarily attributed to thermal store heat exchanger fouling from particulate contamination in the air intake. The remediation required shutdown for 45 days and cost approximately $1.8 million. Post-remediation, Highview installed HEPA-grade air filtration upstream of the liquefier, reducing particulate loading by 99.5% (Highview Power, 2025).

Failure Mode Summary and Impact

Failure Mode	Frequency	Typical Downtime	Cost Impact	Root Cause Category
Electrolyte Precipitation (VRFB)	Medium (0.5-2x/yr)	5-15 days per event	$500K-5M per event	Chemical/Thermal
Membrane Crossover Fade	Continuous	N/A (gradual)	$150K-450K/yr per 100 MWh	Material degradation
Cavern Air Leakage (CAES)	Low (0.1-0.3x/yr)	14-60 days	$2M-10M per event	Geological/Mechanical
Thermal Store Degradation (LAES)	Low-Medium (0.3-1x/yr)	10-45 days	$500K-2M per event	Thermal/Mechanical
Power Electronics Failure	Medium (1-3x/yr)	1-7 days	$50K-500K per event	Electrical
BMS/Control System Errors	High (2-5x/yr)	0.5-3 days	$20K-200K per event	Software/Integration
Gravity System Mechanical Wear	Low-Medium (0.5-1x/yr)	3-14 days	$100K-1M per event	Mechanical
Grid Integration Faults	Medium (1-4x/yr)	0.5-2 days	$30K-150K per event	Electrical/Grid

Power Electronics and Grid Integration Failures

LDES systems connect to the grid through power conversion systems (PCS) consisting of inverters, transformers, and switchgear. Power electronics failures account for the largest share of unplanned downtime across all LDES technologies, regardless of storage medium. A 2025 Sandia National Laboratories review of 43 US grid-scale storage installations found that PCS-related failures caused 38% of all unplanned outage hours, more than any single storage-technology-specific failure mode (Sandia National Laboratories, 2025).

In emerging markets, grid conditions amplify PCS failure risks. Voltage fluctuations exceeding plus or minus 10% of nominal, frequency deviations beyond 49.5 to 50.5 Hz, and harmonic distortion levels above 5% THD are common on weaker distribution networks. Standard inverter protection settings designed for stable grids cause nuisance tripping, with some installations in sub-Saharan Africa reporting 2 to 4 grid-fault trips per week. Fluence Energy's Gridstack platform addresses this with ride-through capability for voltage sags to 0% for up to 150 milliseconds and frequency deviations to plus or minus 3 Hz, but these enhanced specifications add 8 to 12% to PCS cost.

Transformer failures, though infrequent, cause the longest outages because replacement lead times for custom medium-voltage transformers are 16 to 40 weeks. ESS Inc. (iron flow battery manufacturer) now specifies N+1 transformer redundancy for all installations above 10 MW, adding approximately $1.5 million per project but eliminating single-point transformer failure risk.

Battery Management and Control System Vulnerabilities

Software and control system failures are the most underappreciated failure mode in LDES projects. Battery management systems (BMS) must coordinate thousands of individual cells or electrolyte circuits, manage state-of-charge balancing, enforce thermal limits, and communicate with grid operators through SCADA interfaces.

EOS Energy's zinc-bromine Znyth battery installations experienced persistent BMS communication failures during initial deployments, where loss of communication between cell-level controllers and the system-level BMS triggered protective shutdowns across entire battery blocks. Root cause analysis identified electromagnetic interference (EMI) from adjacent high-voltage equipment as the primary driver. The fix required shielded communication cables, EMI filtering on all signal lines, and firmware updates to implement communication loss ride-through logic, at a total cost of approximately $800,000 per installation (EOS Energy, 2025).

For investors, BMS risk is particularly difficult to diligence because it manifests after commissioning and is highly site-specific. Key diligence questions include: How many grid-connected operating hours does the BMS firmware version have across all installations? What is the mean time between BMS-related trips? Does the system have hardware-in-the-loop simulation capability for pre-deployment testing against site-specific grid conditions?

Key Players

Established: Form Energy (iron-air batteries, backed by $800M+ in funding), ESS Inc. (iron flow batteries, NASDAQ-listed), Invinity Energy Systems (vanadium flow batteries, London-listed), Hydrostor (compressed air, operating since 2019), Highview Power (liquid air, UK-based)

Startups: Energy Vault (gravity storage, NYSE-listed), Malta Inc. (electro-thermal, Alphabet-backed), Ambri (liquid metal batteries, Bill Gates-backed), Noon Energy (carbon-oxygen batteries), Antora Energy (solid-state thermal storage)

Investors: Breakthrough Energy Ventures, TPG Rise Climate, Goldman Sachs Asset Management (LDES fund), Macquarie Green Investment Group, Great River Energy

Action Checklist

• Conduct technology-specific failure mode and effects analysis (FMEA) before investment commitment, with quantified probability and financial impact for each identified failure mode
• Require minimum 2,000 hours of grid-connected operation data from the technology provider before committing capital to emerging market deployments
• Specify N+1 redundancy for all power conversion system components including transformers, inverters, and switchgear
• For flow battery investments, require electrolyte thermal management systems rated for site-specific 99th percentile ambient temperature plus 10 degrees Celsius margin
• Mandate BMS electromagnetic compatibility testing at the specific installation site before commissioning sign-off
• Include contractual provisions for technology provider on-site support during the first 12 months of operation, with defined response time SLAs
• Establish spare parts inventory requirements covering all components with lead times exceeding 8 weeks
• For CAES projects, require independent geological assessment with cavern integrity monitoring during the first 5 years of operation

FAQ

Q: Which LDES technology has the lowest failure rate in emerging market conditions? A: Based on available operational data, vanadium redox flow batteries with active thermal management have demonstrated the most consistent performance in high-temperature emerging market environments, with Sumitomo Electric reporting 97.5% availability across its installed fleet in Japan and Southeast Asia. However, the dataset is limited: fewer than 50 LDES installations globally have more than 3 years of operational history. Iron-air and gravity-based systems have insufficient deployment history for statistically meaningful reliability comparisons. Investors should weight technology maturity heavily: a VRFB with 200,000 cumulative fleet operating hours presents lower technology risk than a novel chemistry with 5,000 hours, regardless of theoretical performance advantages.

Q: How should investors model unplanned downtime in LDES financial projections? A: Conservative modeling should assume 15 to 25 days of unplanned downtime per year for the first 3 years of operation (the "infant mortality" period), declining to 8 to 15 days per year in steady-state operation. These figures align with the BloombergNEF 2025 survey data. For emerging markets, add a 30 to 50% premium to account for longer spare parts procurement times and limited local maintenance capability. Model downtime as occurring disproportionately during peak revenue periods (summer and winter peaks in most markets), because thermal stress and grid stress peak simultaneously. A robust financial model should show project viability at P90 downtime assumptions, not P50.

Q: What contractual protections should investors seek against technology-specific failures? A: Essential contractual protections include: performance guarantees with liquidated damages for availability below 95% (net of scheduled maintenance), capacity fade warranties specifying maximum degradation rates per year (typically 0.5% for flow batteries, 1.0% for solid-state chemistries), technology provider obligations for root cause analysis and corrective action within defined timeframes, and escrow arrangements or parent company guarantees backstopping warranty obligations. For pre-revenue-stage technology companies, consider requiring third-party insurance wraps (available from Munich Re and Swiss Re for select LDES technologies) to mitigate counterparty risk on warranty obligations.

Q: What are the early warning signs that an LDES system is approaching a major failure event? A: Key leading indicators include: round-trip efficiency declining more than 2 percentage points below commissioning baseline (indicating thermal store, electrolyte, or power electronics degradation), increasing frequency of BMS-initiated protective trips (more than one per month warrants investigation), auxiliary power consumption rising more than 15% above baseline (suggesting cooling system stress or parasitic load increases), and response time to dispatch signals exceeding 5 seconds for systems designed for sub-second response. Investors should require monthly reporting of these KPIs and establish intervention thresholds in operating agreements.

Sources

• LDES Council. (2025). Global Long Duration Energy Storage Market Outlook: Project Pipeline and Technology Assessment. Brussels: LDES Council.
• Lazard. (2025). Lazard's Levelized Cost of Storage Analysis, Version 9.0. New York: Lazard Ltd.
• Sandia National Laboratories. (2025). Grid-Scale Energy Storage Systems: Reliability, Availability, and Failure Mode Analysis. Albuquerque, NM: US Department of Energy.
• BloombergNEF. (2025). Long-Duration Energy Storage Market Outlook and Operational Performance Survey. London: Bloomberg LP.
• Rongke Power. (2025). Dalian Flow Battery Energy Storage Peak-Shaving Power Station: Commissioning and First-Year Operational Report. Dalian, China: Rongke Power Co., Ltd.
• Invinity Energy Systems. (2025). VS3 Flow Battery Platform: Technical Specifications and Performance Warranty Documentation. Vancouver: Invinity Energy Systems plc.
• Highview Power. (2025). CRYOBattery Carrington: Three-Year Operational Review and Lessons Learned. London: Highview Power Storage Ltd.
• EOS Energy. (2025). Znyth Battery System: Field Performance Report and Reliability Improvement Program. Edison, NJ: Eos Energy Enterprises, Inc.
• US Department of Energy. (2024). Long Duration Storage Shot: Progress Report and Funded Project Updates. Washington, DC: US DOE Office of Electricity.