Replacing some raw materials requires an oxide balance
Adding cullet does not mean replacing sand kilogram for kilogram with an equivalent quantity. Recycled glass already contributes several glass components. Sand is mainly a silica source with its own secondary constituents. Batch calculations must therefore separate received mass, dry material and each oxide contribution.
Kénitra has a documented industrial context, including Dicastal’s Atlantic Free Zone operation described by the Ministry of Industry in 2019 1. That reference concerns aluminium automotive components. It locates the industrial territory without establishing any particular glass production. This dossier addresses teams preparing glass qualification or a project and seeking a traceable calculation method.
The Glass Packaging Institute presents container glass as a mixture combining sand, materials supplying other oxides and cullet 2. The balance converts these materials into comparable contributions, then checks that composition and process operation meet the intended product requirements.
1. Define the calculation boundary
Choose what your reference mass represents: molten glass, saleable glass, dry batch or received materials. These bases are not equivalent. A cullet proportion stated for the introduced batch cannot be compared directly with a proportion calculated against the glass obtained.
Specify the period and physical balance boundaries. If internal rejects return to the furnace, decide how to account for them. This return belongs to internal circulation; it must not be counted repeatedly as an external input, creating artificially high consumption.
Write units in column headings. Mass in tonnes, composition in percent and a dimensionless coefficient can coexist, but their use must be explicit. A spreadsheet correct in arithmetic can still be wrong if its data describe different boundaries.
2. Separate internal and external cullet
Internal cullet comes from the production being considered. External cullet has a different origin and history. Even when appearances are similar, retain separate categories to examine composition, preparation and traceability.
Identify glass families, colours and available controls. The phrase “recycled glass” does not prove a single composition. External material must be qualified for the formulation concerned. An estimate from a previous supplier is not automatically applicable to a new stream.
Khaimovich’s article abstract specifically addresses accounting for cullet oxide composition when calculating glass batches 3. The useful principle is to calculate contributions from the analysed material; the simple equations developed here illustrate that balance without reproducing the article’s algorithm.
3. Harmonise the analytical basis
An analysis should state whether results concern received material, dry material or material normalised after a particular preparation. Keep laboratory indications. Normalising an oxide sum to 100% can conceal useful information about moisture or other losses when used alone.
Prepare one column per material and one row per relevant component. For each value, add batch, date, method and basis. An empty cell means absent information; it must not become zero by default. Unknown composition in a secondary input may prevent closing the balance.
The NIST SRM 1830 certificate illustrates glass composition expressed as mass fractions with associated uncertainties 4. This reference material supports analytical checks; its composition is not a mandatory recipe for all production. Its methodological value is to show that values must remain connected with their measurement quality.
4. Correct moisture before dosing
If h is water’s mass fraction relative to received mass, dry mass = received mass × (1 − h). Conversely, received mass for a given dry quantity = dry mass / (1 − h). This expression assumes that the moisture measurement actually uses that basis.
Moisture reported against dry mass has another definition. It must not be inserted directly into the same formula. Request the definition before calculating and record it. An incorrectly configured automatic correction can systematically shift contributions even when the dosing equipment weighs correctly.
Record the sampling method and location. A dry stockpile surface and its interior are not necessarily equivalent. For more on control, consult our guide to silica sand moisture.
5. Calculate the cullet’s silica contribution
On a dry basis, the cullet’s SiO₂ contribution is cullet mass × cullet fraction, where the fraction is SiO₂’s mass fraction. A percentage must be divided by 100 before multiplication. This check prevents an error by a factor of one hundred.
If several cullet streams are used, add their separate contributions. The correct average content is mass-weighted rather than a simple average of analyses. Two streams with very different masses must not carry identical weight in the calculation.
Apply this logic to every retained component. Cullet does not supply silica alone: its other oxides enter the balance. A formula built only around SiO₂ may satisfy that row while leaving other constituents outside their expected ranges.
6. Determine the remaining sand requirement
In a simplified model without silica loss, the remaining SiO₂ requirement is the glass target minus contributions from cullet and other materials. Dry sand mass is this requirement divided by the sand’s SiO₂ fraction. Identify the model as simplified: real process losses and reactions are addressed subsequently.
Before using the formula, check that other silica-bearing materials are included. Feldspar or another mineral source can contribute to several rows simultaneously. Ignoring those contributions and supplying the whole requirement with sand produces an inconsistent formulation.
A negative calculated requirement does not mean introducing negative sand mass. It indicates that selected contributions exceed the target in this model. Review proportions, compositions or constraints. The spreadsheet should flag this situation rather than silently substitute zero.
7. An entirely hypothetical arithmetic example
Consider only a fictional target of 100 kg of glass containing 70 kg of SiO₂. Assume 20 kg of dry cullet at 70% SiO₂, no other silica source and no SiO₂ loss. The cullet contributes 14 kg; the remaining requirement is therefore 56 kg of SiO₂.
With hypothetical sand containing 98% SiO₂ on a dry basis, the mass needed for this row alone is 56 / 0.98 = approximately 57.14 kg. If fictional moisture on received mass is 2%, the corresponding received mass becomes 57.14 / 0.98 = approximately 58.31 kg.
These numbers describe neither a Minerallum product nor a validated glass formulation. They make the difference between silica contribution, dry mass and received mass visible. Other oxides, remaining materials and furnace behaviour are not resolved by this example. It therefore cannot directly serve as a production recipe.
8. Close the other balance rows
After silica, examine contributions from the other constituents expected in the glass. Each material may supply several components and each component may come from several materials. This interdependence often requires solving the complete system rather than adjusting materials individually without revisiting earlier rows.
In a coefficient table, each column describes contributions from one unit of material. Required masses multiply those columns to reach targets. Check constraints: positive masses, material availability and acceptable ranges. A mathematical solution does not guarantee industrial feasibility.
If several solutions exist, explain selection criteria: consistency, quality, cost, furnace constraints and documented availability. If no solution exists, first check analytical bases and missing data before concluding that a material cannot be used.
9. Distinguish oxide contributions from introduced masses
A raw material is not always entirely retained in glass in its received form. Calculations must use coefficients corresponding to its composition and the transformations considered. Do not assume that total dry batch mass becomes exactly the same glass mass.
Establish coefficients using data validated by the formulation team. A conversion calculated for a pure substance does not automatically represent an impure or moist industrial material. Actual contents and secondary components must be included on their documented basis.
Separate calculated transformation losses, measured handling losses and unexplained differences. Combining them into a single “yield” can conceal a weighing error or composition drift. An increasing discrepancy deserves investigation before an empirical recipe correction.
10. Do not confuse chemical consistency with clean cullet
A satisfactory average analysis does not exclude isolated contaminants. The Glass Packaging Institute describes cullet sorting and preparation to remove foreign materials 2. The chemical balance must therefore be accompanied by physical quality records appropriate to the stream used.
Retain sorting criteria, grading, observations and batch control results. A rare contaminant may be absent from a small analytical portion. Composition analysis and checks for unwanted materials answer complementary questions, without either automatically replacing the other.
Also document storage after sorting. Contact with foreign materials can undo proper preparation. Internal returns must retain identification when they originate from different compositions or colours. Stream traceability makes variation observed in glass easier to explain.
11. Check weighing and actual transfer
The calculation describes planned quantities. Operation must demonstrate quantities actually weighed, transferred and introduced. Identify dosing equipment, its checks and changes in settings. A correct setting is not proof of mass actually delivered.
Examine residues, recovery and material paths. A prepared batch may lose some constituent during transfer or be distributed differently between sequences. Records must connect the analytical sample with the location where it represents the process.
When changing cullet proportion, compare preparation and operating conditions with the reference. If other parameters change simultaneously, attribute the result to the investigated configuration. Avoid turning improved finished product into isolated proof of the sand’s effect.
12. Test sensitivity before purchasing
Repeat the balance using observed or allowed composition variations for each material. Vary one identified factor at a time to understand its effect, then examine relevant combinations. Bounds must come from data or a specification rather than invented precision.
A small content change may require different mass corrections depending on cullet proportion. High uncertainty in an important stream therefore deserves more attention than a final decimal place in a secondary input. Retain results that actually influence the decision.
Chemical sensitivity does not resolve grading or thermal operation. It identifies where measurements should improve or material requirements should tighten. Industrial trials then complement the calculation with quality criteria defined for the glass produced.
13. Organise a spreadsheet that detects its errors
Reserve one area for input data and another for results. In the inputs, identify every source and protect units: a composition expressed as percent should not be replaceable with a fraction without the calculation flagging it. Keep dry and received masses in different columns.
Add straightforward checks: sum of proportions, negative values, moisture outside its definition, missing analytical references and deviations from targets. These checks do not validate the formulation, but they prevent several silent errors. Unknown information should trigger a question rather than be replaced with an assumed value to complete the table.
When checking a change, keep the previous version and compare calculated contributions component by component. Total mass difference alone does not explain composition difference. If a new cullet analysis increases a secondary oxide, identify which material correction answers that variation and how it affects other rows.
Have the model reviewed using an independent calculation on a deliberately simple example. Check division by one hundred, moisture conversions and contribution addition in particular. A complex sheet can be tested with a case where only one material supplies the component investigated, then with two known contributions. Determine the expected result before observing the spreadsheet’s answer.
Finally, associate every issued formula with a version of the analytical records. A recipe updated with new contents but sent under its old identifier creates operational ambiguity. Validation should specify which version is authorised, from which batches and for which glass composition. This traceability distinguishes a calculation error from correct use of the wrong file.
14. Distinguish material balance from environmental claims
The JRC’s glass BREF presentation includes efficient material use, energy consumption and residue recycling within its scope 5. These subjects are connected, but a material balance alone does not demonstrate a quantified energy saving.
An environmental comparison requires comparable boundaries, period, measurements and function. Transport, preparation, cullet quality and furnace operation may enter the study. Increasing recycled content does not justify a universal saving claim for an unmeasured installation.
The useful conclusion is a calculated formulation, identified materials, explicit assumptions and a validation programme. Our silica sand page positions the mineral family; this balance explains how to connect its contribution with cullet. For a glassmaking decision, both levels must remain connected to analyses of the batches actually used.