HELLO CHINA

INDUSTRY CLUSTER · New materials

Hepu photovoltaic rolled glass

A loss of photovoltaic-glass transmission or distorted pattern can lead back through low-iron sand impurities, furnace atmosphere, rolling rolls, annealing stress and coating uniformity. Hepu is read along that continuous chain.

Formation and regional foundations

The park brings silica feed, continuous furnaces and rolled-glass lines close to the port, while pilot and smart-factory work adds coating research and process optimisation. Furnaces cannot stop freely, so energy, safety and quality share the production plan.

When photovoltaic rolled glass loses transmission, distorts its pattern or bows, the cause may predate finishing. Low-iron sand impurities, furnace atmosphere, roll temperature and speed, annealing stress and coating uniformity can each leave a mark, while a continuous furnace cannot stop for every question.

Core geography: A continuous Hepu Tieshan East Port manufacturing corridor from low-iron sand treatment and melting through rolling, annealing, cutting, coating and photovoltaic-glass inspection

Six hot and cold transitions from sand source to module lamination

At which temperature window one ripple first forms

The hot end locates bubbles and pattern by continuous time; the cold end retains stress, edge and coating results by pane position, joined at the furnace shift.

  • Low-iron silica-sand lot — Low-iron sand is sampled by stockpile zone and loading lot, with chemistry, size, moisture and magnetic debris defining feed variation.
  • Glass batch and furnace shift — Batch weighing and charging cadence align with flame, furnace pressure and pull curves, allowing cold-end bubble positions to return to melting time.
  • Hot rolled-glass ribbon — The rolled ribbon retains thickness, texture and roll angle across its width, separating once-per-revolution ripples from gradual transverse variation.
  • Glass temperature band in the lehr — Top and bottom temperatures, edge airflow and ribbon speed remain together through annealing, while polarised stress and bow locate cooling imbalance.
  • Cut and ground glass edge — Cutting and edging parameters follow pane orientation into tempering, keeping edge images, furnace position and fragmentation as one specimen set.
  • Coated rolled-glass surface — Measurements on the coated face cover pattern peaks, valleys and a transverse grid for transmission, thickness and contact angle, mapping local voids to nozzle coordinates.

01 · Low-iron sand controls iron and size first

Feed colour follows through to transmission.

Sand receipt records chemistry, particle size, moisture and mine lot. Different heights and loading positions in a sand stockpile can concentrate dark heavy minerals, allowing an average assay to hide local variation. Moisture gained during outdoor storage changes actual batch mass, while iron picked up in transport must be distinguished from impurity at the mine.

Once low-iron sand enters the hot melt, it cannot be separated again, and a small pocket of iron-rich heavy mineral may create a local colour streak rather than uniform transmission loss. A passing daily average can dilute an anomaly from one stockpile height, truck, or loading position. Sand sampling therefore retains spatial and vehicle identity so a streak on the ribbon can return through time to the particular feed.

Sample multiple points and retain raw sand for retesting when glass colour shifts. Take incremental samples by stockpile zone and loading lot, dry them, and measure chemistry, size distribution, and magnetic debris before sealing retained sand. When glass colour shifts, trace the charging time to the original bag instead of resampling an already blended inventory.

  • Average low-iron sand chemistry may miss local heavy minerals, so retained multi-point samples determine whether colour can be traced.

02 · Batching and melting jointly determine bubbles and colour

Weighing, homogenisation and flame belong together.

Batch weighing, mixing, and charging cadence share a timeline with flame length, furnace pressure, and glass pull rate. Material hanging in weigh bins, belt residue, and feed moisture alter the actual recipe, while a production-load change moves combustion zones and melting rhythm. When bubbles or stones reach the cold end, their cross-width position and estimated melting time identify the relevant charge and shift.

A stone in glass may be unmelted batch or eroded refractory, and the two differ in chemistry, boundary morphology, and usual location. Counting cold-end defects alone loses the sequence among charging cadence, melting temperature, and furnace-pressure movement, while colour can also shift when atmosphere changes the iron state. A representative stone retains its original position and receives sectional analysis before the investigation turns to feed or furnace structure.

Longitudinal position at the cold end is converted back to melting time, while the cross-width coordinate is mapped to charging and combustion zones; clustered defects trigger retrieval of weighing, flame, pressure, and pull traces. Stones are sealed with their original coordinates for microscopic chemistry that distinguishes unmelted feed from furnace debris. Batching cadence or furnace structure is changed only when position, time, and section evidence agree.

  • A stone section distinguishes unmelted batch from refractory erosion; defect counts alone cannot reveal origin.

03 · Rolling rolls write pattern into hot glass

Roll temperature, gap and speed determine texture.

Rolling records upper and lower roll temperatures, gap, speed and glass temperature. Surface temperature of upper and lower rolls varies across the width, while cooling flow and deposits change local thermal expansion. Saving pattern pitch, transverse thickness, and ribbon speed together helps determine whether a defect follows roll circumference or glass supply.

Roll contamination, thermal deformation or speed variation creates periodic texture and cross-width thickness change. A pattern recurring once per roll revolution usually points to surface condition or eccentricity, whereas a slow transverse thickness gradient may come from gap and melt-temperature distribution. Calling both rolling defects lets maintenance and process adjustments interfere with each other.

Thickness and pattern are measured continuously at both edges and the centre of the ribbon, with roll angle, speed, and pull changes marked so a periodic defect can be tied to circumference. Furnace load and pull remain common before and after cleaning or replacing a roll, with at least two full revolutions observed to prevent a brief speed shift from resembling a repair. A phase that follows the roll supports local grinding; one that does not sends review to temperature and draw.

  • A texture repeating with roll circumference points to the surface, while transverse thickness also requires gap and melt-temperature checks.

04 · Annealing releases thermal stress gradually

Cooling too fast locks risk inside glass.

The annealing curve covers top, bottom and width and records speed and environment. Lehr zones do not cool the top and bottom surfaces symmetrically, and ribbon edges are more exposed to ambient airflow. A small speed change gives the same temperature setting a different residence time, redistributing residual stress across the pane.

A target average at the centre of the lehr does not prove that both edges cooled at the same rate; hidden tensile stress may leave no immediate crack yet release during cutting or tempering. Longitudinal and transverse bow direction can also reveal imbalance between upper and lower surfaces or left and right cooling zones. Stress assessment combines full-width polarised distribution, bow, and post-cut edge response rather than relying on one centre temperature.

Fixed positions across the ribbon are measured for polarised stress, longitudinal and transverse bow, and edge response after cutting, with each result linked to its lehr cooling zone. After one zone is adjusted, a complete ribbon transit passes through the lehr before resampling and transition glass remains separately identified. Old and new distributions are compared only after that full travel to determine whether the cooling correction created a stable state.

  • A passing centre temperature can still leave edge tension, and cutting response exposes the hidden distribution.

05 · Cutting, edging and tempering manage the pane edge

A small edge chip can grow after heat treatment.

Cutting records wheel type and wear, cooling, pressure, and crack direction; edging adds wheel grit, feed, chamfer, and microchip size. Original pane orientation and the identity of every edge remain attached to the tempering position because load density and placement also change heating uniformity. If breakage follows, the investigation can separate a pre-furnace edge injury from the thermal field of a particular furnace zone instead of seeing only fragments.

Breakage after tempering does not prove that the crack began inside the furnace; microdamage from cutting, edging, or handling may propagate only when heated. Fragment size describes the final breakup, but without pre-furnace edge images and orientation marks it cannot separate a mechanical origin from uneven heating. Fixed segments on all four edges are photographed, and mirror region, radial marks, and furnace position are reviewed together after a break.

All four edges are photographed by segment before tempering with cutter, edging wheel, chamfer, and cooling settings retained, then bow and fragmentation are mapped by furnace position. An anomaly that follows one edge segment across furnace locations sends review to cutting and edging, while concentration in one furnace zone sends it to thermal field and loading. The two maps choose the corrective owner, preventing mechanical rework and thermal calibration from substituting for one another.

  • The moment of tempering breakage is not the crack origin; pre-furnace edge images separate mechanical and thermal causes.

06 · Coating must cover a textured surface uniformly

Film thickness on a plane does not represent grooves.

Coating evaluation measures transmission, thickness and surface across width and pattern peaks and valleys. Pattern peaks, slopes, and valleys receive different coating flux, and cleaning liquid can remain inside grooves. One average thickness across the pane cannot reveal spray boundaries, conveyor drift, or local shadowing caused by the textured surface.

Average transmission can hide missed or thick areas that later affect optics and bonding. Even when average transmission passes, a local coating void can form an optical patch or weak bond in a module, while excess thickness changes colour and abrasion. A flat witness on the same line still cannot represent coverage over the textured pane.

Coated glass is cut on a cross-width grid through peaks, valleys, and normal regions for transmission, contact angle, and change after abrasion, with uncoated same-lot glass retained. Every cell carries original face, cleaning nozzle, coating nozzle, and conveyor position so a damp-heat anomaly maps back to machine coordinates. Movement with conveyor alignment points first to centring, while a fixed cross-width band points toward local flow or nozzle condition.

  • A flat witness cannot represent patterned valleys, so coating uniformity must be measured across texture and pane width.

07 · Module lamination tests the final glass interface

Size, bow and surface jointly affect encapsulation.

Module trials record glass size, cleaning, encapsulant, lamination and edge sealing. Pane bow changes contact order during lamination, while cleaning residue and coating surface energy affect encapsulant wetting. Edge seal, busbar height, and vacuum-bag pressure can amplify problems that were not visible on a single glass sheet.

A pane can pass individual tests yet create bubbles or delamination through bow or surface energy. Passing pane transmission and dimensions proves only separate incoming attributes. Lamination bubbles may arise from geometric contact or interfacial volatiles, and without preserved bubble locations, duties of glass, encapsulant, and equipment become indistinguishable. Bubble geometry and location strengthen that separation.

Compare laminates from different pane positions and inspect transmission and adhesion after damp heat. Make mini-modules from centre and edge panes using one encapsulant lot and lamination programme, preserving original orientation. After damp heat, measure transmission, peel, and bubble distribution to determine whether ribbon-position differences persist in the module.

  • Even after individual pane tests pass, module lamination can expose bubbles or delamination from bow and surface energy.

08 · A furnace transition needs a new quality baseline

Repair, sand and draw changes can alter distributions.

After furnace repair, feed change or major adjustment, full-width and time-series sampling restarts. Furnace repair replaces refractory and changes flow inside the melter, a new sand source introduces another particle and impurity distribution, and a higher pull rate shortens melting and annealing time. First glass may therefore carry several changes rather than continuing the old campaign.

At the start of a new campaign, furnace heat storage, roll surface, and annealing condition have changed, so the prior average can dismiss early shifts in colour, bubbles, thickness, or stress as random variation. Transition glass receives a separate lot identity through cutting, coating, and storage and remains apart from stable output. Transition status ends only after full-width data from several consecutive shifts establish the new baseline, not when material is prematurely mixed into routine inventory.

Identify product status separately until a new baseline and representative module retest are complete. Give transition output separate identities and collect full-width optical, thickness, stress, and defect maps each shift until a sustained window stabilises. After the plant baseline is reached, coat and laminate representative panes to confirm that cold-end conformity extends to final use.

  • First glass after repair or feed change needs a new baseline because the previous campaign average no longer represents current distributions.

Sources and reading notes

Planned furnaces, fired lines, trial production and stable output are four distinct states. Architectural float glass, power plants and module capability stay outside rolled-glass manufacturing, and every test belongs to a furnace campaign and glass lot.

  • Hepu photovoltaic rolled glass: National recognition year is 2024.
  • Hepu photovoltaic rolled glass: List validity is 3 years.
  • Photovoltaic-glass projects in investment material: 6 projects.
  • Year of pilot and coating research material: 2025.

Related companies

These companies have a locally relevant industry role supported by public information. Order is not a ranking, inclusion does not imply authorization, partnership or endorsement, and Hello China does not participate in or guarantee transactions.

A company may request a free correction or removal through the formal channel if its record is inaccurate or it does not wish to be shown. Do not send business licences, authorization letters or other sensitive files through insecure channels.

Sources

Use the original publishers to check dates, scope and definitions; public material is not a transaction guarantee.