Patented · CMA / CNAS tested
A hollow HDPE tank that becomes the ballast. Carry it up empty at 6.3 kg, fill it on site, and each unit turns into 120 kg of water weight holding the array down.
01 · Product overview
The system is a closed hollow HDPE tank that acts as both the support structure and the counterweight. Water is the only ballast medium — no sand, no gravel, no concrete blocks. Water is available on any site, costs nothing, and drains out when you need to move the array.
Position the tanks. Carry them up empty at 6.3 kg each — two people move a full row without equipment.
Lay the modules. The tank sits under the module joint, so two modules share one unit.
Clamp with one hand. Middle clamps on the shared edge, end clamps on the outer frame, one bolt each.
Fill with water. The array becomes self-weighted. Nothing is fixed to the roof structure.
Adjacent modules share the tank at the seam. A row of 20 modules needs 21 units. This single rule is all your installer needs to order.
| Item | Drilled / concrete ballast | DINGXIU water ballast |
|---|---|---|
| Roof waterproofing | Penetrations cut the membrane, repair is required and warranty claims follow | Zero penetrations. Membrane untouched, existing warranty intact |
| Crew | Certified installers familiar with the rail system | General labour. No system-specific skill |
| Lifting equipment | Crane or hoist for concrete blocks and rails | Manual carry. 6.3 kg per unit, no machinery |
| Later modification | Dismantling damages rails and anchors, effectively scrap | Drain, lift, relocate. Fully reusable |
| Roof load | Concentrated point loads at anchors and block positions | Distributed over the tank footprint, low bearing pressure |
Where it fits: concrete flat roofs · industrial and commercial rooftops · level ground · existing waterproofed membranes where drilling is prohibited · rooftops where a crane cannot reach.
02 · System & specification
| Part | Function |
|---|---|
| Float body | Blow-moulded HDPE hollow tank. Becomes both support and counterweight once filled. |
| Middle clamp | Clamps the adjacent frames of two neighbouring modules on their shared edge. |
| End clamp | Clamps the outer frame of modules at the array edge. |
| Bolt | Fastens clamp to tank. Single-hand tightening, no torque tool required. |
Modules are mounted in portrait. The tank sits under the joint between two adjacent modules, tilted approx. 5°.
| Module size | Modules per MW | Modules per row | Sets per row | Total sets, approx. |
|---|---|---|---|---|
| 550 Wp | ≈ 1 818 | 20 | 21 | 1 920 |
| 580 Wp | ≈ 1 724 | 22 | 23 | 1 815 |
| 620 Wp | ≈ 1 613 | 20 | 21 | 1 710 |
Assumes an ideal rectangle. Real roofs lose rows to setbacks, skylights and walkways, so add 3–5%. We issue an exact layout and quantity count from your roof drawing before you order — free of charge.



03 · Testing & certification
The product is tested by Tianjin Bailun Inspection & Testing, a third-party laboratory holding both CMA and CNAS accreditation, and by the Hebei Gengqiang New Energy laboratory for pull-out strength. Report numbers can be verified with the issuing body.
GB/T 16422.3-2022, UVA-340 lamp, 8 h light at 60 °C + 4 h condensation at 50 °C, 84 cycles. No cracking, no chalking, no significant colour change.
Stored at −40 °C for 168 hours. No damage, no cracks, no surface defects. Suitable for cold-climate winter operation.
Stored at 105 °C for 200 hours. No significant surface defects. Withstands summer rooftop heat.
IEC 61215-2:2016 MQT 11 reference, −40 °C to +85 °C, 200 cycles. No significant surface defects.
Pull-out test including elevated temperature. No tearing and no failure at the clamp. Result: PASS. Report JTTR-PZ-A-04-003.
Equivalent wind speed
≈ 30 m/s108 km/h · Beaufort 11
| Wind speed | Dynamic pressure | Beaufort | Against the tested 0.55 kN/m² |
|---|---|---|---|
| 20 m/s (72 km/h) | 0.245 kN/m² | 8 | Within tested range |
| 25 m/s (90 km/h) | 0.383 kN/m² | 10 | Within tested range |
| 30 m/s (108 km/h) | 0.551 kN/m² | 11 | At the tested limit |
| 35 m/s (126 km/h) | 0.750 kN/m² | 12 | Above tested range — ask us |
| 40 m/s (144 km/h) | 0.980 kN/m² | 12+ | Above tested range — ask us |
Conversion uses q = 0.5 ρ v² at standard air density ρ = 1.225 kg/m³ (15 °C, sea level). Compare this table with the design wind speed your local code requires for the building. Inside the tested range the system fits mechanically; above it, we quote additional units or a different configuration. Final ballast sizing is issued per project.
Virgin HDPE resin inspected on arrival
Blow-moulding parameters monitored inline
Water-tightness and pressure sample tested
Appearance and dimensions checked per unit
04 · Advantages
Water ballast replaces ground screws and expansion bolts. The waterproofing membrane is never opened, so the roof warranty you are responsible for stays intact.
About 12 days per MW against roughly 30 for a drilled system. Grid connection arrives earlier, and earlier connection is the only advantage that shows up directly in the IRR.
Position, lay, clamp, fill. No rail cutting, no alignment jigs, no torque specification. General labour can do it after ten minutes of instruction.
6.3 kg empty. Two people move a whole row by hand — no crane, no hoist, no scaffold, which matters on a finished industrial roof.
Fits any framed module. Only the clamp profile changes with frame thickness, so you are not locked to one module brand.
Drain and lift. Useful for roof maintenance access, array reconfiguration, or relocating the plant — the asset moves with you instead of being scrapped.
HDPE body with a UV stabiliser system, validated by 1000 h UV ageing, −40 °C cold, 105 °C dry heat and 200 thermal cycles.
Tanks nest in pairs at about 6 sets per m³. 197 sets fit a 20GP, 350–400 fit a 40HC. Per-set freight drops accordingly.
| Cost line | Drilled system | Water ballast | Saving per MW |
|---|---|---|---|
| Mounting material | ≈ ¥0.20 /W | ≈ ¥0.20 /W | roughly equal |
| Concrete ballast blocks + crane | ¥30 000–45 000 | none | ¥30 000–45 000 |
| Membrane repair after drilling | ¥10 000–15 000 | none | ¥10 000–15 000 |
| Install labour (18 days shorter) | specialist crew | general labour | ¥15 000–25 000 |
| Typical total | — | — | ¥55 000–85 000 |
Ranges reflect north China labour and material rates in 2026 for a 1 MW distributed rooftop. Replace the unit rates with your own local figures — the structure of the saving does not change: you are buying labour, lifting and waterproofing repair, not steel. In markets where skilled installers are scarce or wages are high, the gap widens.
Lead time. 200 sets ex-works in 10–15 days; full container loads 20–25 days. Tell us the target vessel and we plan production backwards from it.
What we send free. Layout drawing, quantity count from your roof plan, wind calculation basis, datasheet, and the full test reports on request.
05 · Logistics
An empty tank is mostly void, so the whole product is volumetric cargo. Tanks nest in pairs, which is why the per-set freight cost is low even on long routes.
| Vehicle | Capacity |
|---|---|
| 13 m flatbed truck | 450 sets (weighed at loading) |
| 9 m truck | ≈ 235 sets |
| 6.5 m truck | ≈ 140 sets |
| 20GP container | 197 sets (weighed at loading) |
| 40GP container | ≈ 350–360 sets |
| 40HC container | ≈ 400 sets |
Stacked pair ≈ 0.32 m³, single set ≈ 0.16 m³, about 6 sets per m³. A 40HC at 400 sets weighs roughly 2.5 t gross of product — the container cubes out long before it weighs out.
Quantity confirmed, production slot booked
Batch moulded, inspected per unit
Container loaded, documents issued
Remote install guidance on arrival
Roof area, module wattage and rows per array is enough. We return a layout drawing, the exact set count and a quotation — normally within one working day.
Contact

Scan to open WhatsApp
with the message already written
Patent 202630294473.0 · Tested by Tianjin Bailun (CMA / CNAS) report BL9B936C52 · Pull-out JTTR-PZ-A-04-003 · Specifications subject to the product actually supplied · Reports available on request