How to Assess Solar Module Hail Risk: What Qualification Proves (IEC 61215-2 MQT 17 Uses Ice Balls of a Minimum 25 mm Diameter and Is Described as 'Sufficient' for Generic Qualification Where Severe Hail Is Unlikely, per the IEC TS 63397 Introduction; Test Results Are Not a Lifetime Prediction), What the Severe-Hail Tiers Add (IEC TS 63397:2022 HMQT 17 and Hail Maps; FM 4478 Classes 2–4 and VSH, the Latter at 72–79 J per Impact; Kiwa PVEL's 35–55 mm HSS, Where Most 2.0/2.0 mm Glass Modules Break at 45 mm and 43% at 40 mm), What Mitigation the Evidence Supports (Thicker Front Glass Roughly Twice as Resistant per VDE Americas; Stow Angle Scaling with cos²θ, a 75° Stow Modelled to Cut Farm Damage up to 80%), the Chinese Baseline (GB/T 9535.1-2025 Effective 1 May 2026; 32 Convective Processes and 55 mm Hail in 2024 per the CMA; Hail Defined as >5 mm in a PICC Clause) and the Loss Profile Relayed from Insurers (6% of Claims, >72% of Losses)

Short answer. A module's IEC 61215 certificate says little about hail beyond a minimum: the introduction to IEC TS 63397 states that the hail test MQT 17 of IEC 61215-2:2021 uses spherical ice balls of a minimum 25 mm diameter (optionally larger), that it may cause invisible internal damage detectable by the Gate No. 2 power measurement only if relatively large, and that it is sufficient for applications where modules are unlikely to see severe hail — which is why a standardised higher-stress test was 'considered useful'. IEC 61215-1 itself says test results are not to be construed as a quantitative prediction of module lifetime. The tiers above the minimum are IEC TS 63397:2022 (additional requirements for hail-prone deployments, with a hail test HMQT 17 and informative hail maps), FM Approvals' 4478 standard (hail classes 2, 3, 4 and VSH, the latter at 72–79 J per impact, tested per ANSI/FM 4473 with no cracking under 10× magnification), and laboratory sequences such as Kiwa PVEL's 35–55 mm Hail Stress Sequence, where the majority of 'standard' 2.0 mm/2.0 mm glass–glass modules break at 45 mm and 43% break at 40 mm, while thicker-glass designs are significantly less susceptible and unbroken modules show under 2% power loss. The mitigation evidence points two ways: glass — RETC/VDE Americas describe a 3.2 mm front-glass/backsheet package as roughly twice as hail-resistant as a 2.0 mm dual-glass bifacial module, with wide uncertainty bars — and angle — effective impact energy scales with cos²θ, so 70 J becomes 17.5 J at a 60° stow, and a 2026 finite-element study finds a 75° stow can cut farm-wide damage by up to 80% against 30–40% for thicker glass alone. In China the qualification standard GB/T 9535.1-2025 takes effect on 1 May 2026, the CMA's 2024 bulletin recorded 32 regional convective processes with hail up to 55 mm and, per the relevant departments, 14.57 billion yuan of wind-and-hail losses, and a PICC household PV clause covers hail defined as ice of more than 5 mm diameter. The loss profile that drives all this is relayed second-hand from insurers: hail as 6% of claims but more than 72% of losses by value (kWh Analytics, via RETC), or 1.4% of claims and 54% of costs (GCube, via IEA PVPS). This guide quotes no test clause text, Chinese hail-damage case, utility-scale policy or tracker specification.

1. What the qualification test proves

CNWattHub reading: a certificate to IEC 61215 or GB/T 9535.1 is a floor, not a hail rating. The questions for the supplier are whether the hail test was run at the minimum 25 mm or at a larger optional size, and whether the post-test power measurement is on file — because the TS introduction says the damage that matters may be invisible.

2. What the severe-hail tiers add

• IEC TS 63397:2022 (edition 1.0, 14 December 2022; July 2023 corrigendum included) defines additional testing requirements for modules deployed where hailstone impacts cause higher stress beyond the IEC 61215 series, to assist module selection in higher-risk regions and to provide design tools; its contents include a hail test HMQT 17 (clause 5.10) and informative hail maps (Annex A).
• FM 4478 (March 2025) requires hail-damage-resistance testing per ANSI/FM 4473, with Class 2 the minimum for certification; the VSH rating uses a Class 4-size ice ball at increased speed to reach 72–79 J (53–58 ft-lb) per impact; after impact the module must show no cracking, splitting, misaligned surfaces or rupture under 10× magnification; ratings are Class 2, 3, 4 or VSH.
• Kiwa PVEL HSS (2026 Scorecard dataset) uses lab-made ice balls of 35–55 mm, controlling shape, density and velocity; most 2.0/2.0 mm glass–glass modules break from 45 mm hail and 43% from 40 mm, glass near edges and junction-box holes being more likely to break than glass in the centre; 3.2 mm glass/backsheet, 2.5/2.5 mm and 3.2/2.0 mm designs are significantly less susceptible; unbroken BOMs show under 2% power loss; almost 350 PQP modules were tested.
• RETC's Hail Resiliency Curve Test shoots progressively larger ice balls at progressively higher speeds and energies until the glass breaks.

CNWattHub reading: three test philosophies coexist — a pass/fail at a declared severity (IEC TS 63397, FM classes), a stepped survey across sizes (PVEL), and a test-to-failure curve (RETC). A buyer comparing 'hail-hardened' claims should ask which philosophy produced the number, because a Class 4 pass, a 45 mm survival rate and a Weibull curve are not interchangeable.

3. What mitigation the evidence supports

CNWattHub arithmetic (on the tagged relation): with effective energy = KE × cos²θ, a 45 mm stone that would deliver its full energy to a flat module delivers 25% of it at 60° and about 6.7% at 75° (cos² 75° ≈ 0.067) — which is the physical reason the simulation study ranks angle control above glass thickness. The caveat is operational: the stow has to be reached before the hail arrives, which is why the Task 13 slides pair the angle with a 1–2 minute stow time and alerting tools.

4. The Chinese baseline

• Climatology. The CMA's 2024 bulletin records 32 regional severe-convective processes in 2024 (five-year average 42); per the relevant departments, 7.992 million person-affections of wind-and-hail disasters, 1,147.4 thousand hectares of crops affected, 88 deaths and 14.57 billion yuan of direct losses, with overall convective losses lighter than the 2019–2023 average; late-March hail of 20–40 mm in Jiangxi, Fujian and Zhejiang; on 15–17 April hail over 20 mm or giant hail in Sichuan, Guizhou, Jiangxi and Guangxi with a maximum diameter of 55 mm.
• Insurance definition. PICC's household solar PV equipment clause lists hail among covered perils (with lightning, rainstorm, flood, gale, tornado, typhoon, hurricane, sandstorm, blizzard and ice) and defines hail as solid precipitation of ice blocks or balls from strongly convective cumulonimbus with a diameter greater than 5 mm and a hard core.
• Standard. GB/T 9535.1-2025 effective 1 May 2026 (text not read).

CNWattHub reading: the 55 mm maximum the CMA reports for April 2024 sits above the 45 mm level that RETC calls the design case for Texas and at the top of the 35–55 mm range PVEL tests — so the Chinese hail climatology is not benign everywhere, even though the national bulletin frames 2024's convective losses as lighter than average and does not break out PV losses. The household clause is cited for its definition only; a utility-scale policy's hail terms were not located at source.

5. The loss profile, second-hand

• RETC relays kWh Analytics' estimate that hail is 6% of solar project insurance claims but more than 72% of losses by dollar amount, with hail losses since 2018 exceeding $600 million in Texas alone, and says Texas utility-scale farms must withstand ≥45 mm events likely within an asset's life.
• The IEA PVPS Task 13 slides cite GCube for hail as 1.4% of claims but 54% of incurred costs (2018–2023), and report a 2023 Nebraska storm destroying a 5 MW plant in under ten minutes and a 2024 Texas storm destroying a 350 MW plant.

CNWattHub reading: the two claim-share pairs (6%/72% and 1.4%/54%) reach this guide through different relaying sources — kWh Analytics' figure via an RETC article, GCube's via an IEA PVPS presentation — and neither underlying dataset's definitions, scope or basis of measurement was checked here, so the statistical bases are unverified and the two pairs cannot be compared directly or used to explain each other. Each is cited only as its relaying source reports it.

6. An assessment sequence built from the sources

1. Site hazard. Local hail climatology (China: CMA bulletins; 55 mm reported in April 2024) and, where available, a probabilistic PML/AAL assessment.
2. Certificate scope. IEC 61215 pass at the 25 mm minimum or a larger optional size; post-test power data; the GB/T 9535.1-2025 text itself was not read.
3. Severe-hail evidence. IEC TS 63397 HMQT 17, FM class or VSH, PVEL HSS size survived, or a RETC curve — and which philosophy produced it.
4. Glass package. 2.0/2.0 mm versus thicker front glass; PVEL breakage rates by design.
5. Stow capability. Angle (≥55°, 75° modelled), time to stow (1–2 minutes), trigger and alerting.
6. Insurance wording. Hail named as a peril; its definition (e.g., >5 mm).
7. Residual. Invisible damage after a storm — power measurement, not visual inspection alone.

Scope limits

This guide reads the IEC webstore abstracts of IEC 61215-1:2021, IEC 61215-2:2021 and IEC TS 63397:2022 and the TS 63397 preview introduction (test clauses not read), the GB/T 9535.1-2025 record (text not read), FM 4478 (March 2025), Kiwa PVEL's HSS page, an RETC article, an IEA PVPS Task 13 presentation, one open-access simulation study, the CMA's 2024 climate bulletin and one PICC household clause. Insurance claim shares are third-party estimates relayed by a laboratory and a presentation. It quotes no Chinese hail-damage case, utility-scale policy, module datasheet hail rating or tracker specification. Not engineering, procurement, insurance or investment advice.