Consistency matters more than the advertised reuse percentage
Comparing virgin and reclaimed sand does not mean contrasting a clean material with an inevitably poor one. The task is to establish what each stream contributes to the moulding system, what remains from its previous use and what variation it introduces over successive cycles. A useful decision concerns an identified casting, binder and machine, with reproducible results.
The Ministry of Industry describes Tanger-Tétouan-Al Hoceïma as a region with a substantial industrial contribution, including automotive activities 1. This context gives local relevance to controlling moulding materials, without itself supplying an inventory of foundries or sand consumption. This dossier addresses production managers and buyers who need to document a technical comparison.
Virgin sand provides a purchasing reference. Reclaimed sand provides recovered material whose properties also result from moulding, pouring, shakeout and return-stream treatment. Qualification must follow that history. A certificate for the initial material does not completely describe it after several journeys through the circuit.
1. Define what “reclaimed” means at the installation
Start by drawing the actual circuit: mould preparation, pouring, cooling, shakeout, metal removal, lump breaking, dust removal, treatment and storage. Mark the places where core sands, coatings or special aggregates can enter the return stream. Reclaimed must mean a specific outlet from that circuit, rather than any sand recovered from the floor.
The literature distinguishes, among other approaches, mechanical cleaning of grains and thermal reclamation of organic residues. Ghormley and colleagues compare several technologies at an American foundry, considering purchases, transport and treatment 2. Their economic conclusion cannot be transferred directly to Tangier: distances, energy, waste and processes differ.
On the material record, identify the reclamation process, monitored settings, original binder, sand type and stream separation. Material that has passed through a reclamation machine is not automatically compatible with every binder. The machine name does not replace measurements on its output.
2. Separate return families before comparing them
A homogeneous return stream is easier to characterise than a variable mixture. Distinguish moulds and cores by binder, aggregate and added products. Special sands entering a silica matrix can alter its composition and grading. Coating waste, metal fragments and fines should not become invisible contributions to the mixture.
Separation does not necessarily require another installation. It begins with identifying containers, assigning circuits and establishing a storage rule. Check that the proposed organisation remains possible during a production-series change, shutdown or cleaning. These are often the moments when normally separate streams become mixed.
Define a rule for material whose origin has been lost. Reintroducing it without examination can reduce immediate cost while removing trial traceability. A comparison must make it possible to identify what was tested. If the return composition changes during the campaign, create a new identification instead of artificially retaining the same number.
3. Sample the stream actually being used
Sampling should represent the sand sent to the mixer, rather than a handful taken from the top of a container. Define the lot, covered duration and sampling locations. ISO 14488 addresses representative sampling of particulate materials; its public catalogue notice specifically identifies the connection with particle-size distribution 4.
Prepare separate portions for grading, chemical tests and laboratory mixtures. Identify storage conditions, particularly where sand moisture or temperature affects preparation. Retain a reference portion. It allows the material to be examined again if a moulding result becomes difficult to interpret.
Arrange several samples from independent production runs. Several measurements on one bag improve knowledge of that bag; they do not demonstrate circuit consistency over a month. In the report, separate analytical repeats, preparation repeats and independent lots. This distinction avoids giving a short qualification excessive apparent precision.
4. Compare the complete curve rather than one index
A fineness index summarises a distribution. Two sands can have a similar index with different proportions across the sieves. For substitution, retain the masses in each size class, fine fraction, coarse material and method used. The curve provides a better discussion basis than the commercial designation alone.
Then compare virgin sand, reclaimed sand and the proposed mixtures. Reclamation can modify grains or remove particular fractions; storage can also cause segregation. A curve measured immediately after treatment does not necessarily describe material withdrawn from a silo hours later. If the results differ, examine the route between those locations.
Do not immediately correct a drift by adding virgin sand. First investigate whether it comes from treatment, sampling, withdrawal or a change in returns. Addition may hide the cause without removing it. For the general material-qualification framework, also consult our silica-sand page.
5. Interpret residues within the binder system
Loss on ignition helps monitor a mass change under defined heating conditions. It is not a universal measurement of remaining resin alone. Sample preparation, temperature, duration and other constituents affect the result. Compare results only when obtained using a compatible, properly identified procedure.
A study of furfuryl-bonded sands examines thermal reclamation, grain surfaces and reuse 3. It illustrates the value of characterising treatment and the material after treatment. Its experimental parameters do not establish an optimal temperature for every circuit or an acceptance limit applicable to every binder.
Depending on the process, the laboratory can add the relevant characteristics of the surface and its reaction with the binder. Ask the binder-system supplier which tests are interpretable in this application. A measurement intended for one chemical system should not become a universal criterion for another. The dossier should connect each measurement to a concrete decision.
6. Build a readable mixture matrix
A useful campaign includes a controlled reference, characterised reclaimed sand and mixtures with known proportions. Do not simultaneously introduce a new binder, dosage and sand if the purpose is to isolate the return-stream effect. If adjustment is necessary, distinguish the substitution trial clearly from the reformulation trial.
Express proportions on a defined basis, particularly a dry basis where moisture varies. Record the masses actually weighed. The virgin-to-reclaimed ratio in the recipe must be distinguished from reclamation mass yield. These indicators answer different questions: mixture composition on one hand and recovery from the treated stream on the other.
For every mixture, retain preparation conditions, the time between mixing and testing, storage conditions and the compaction or specimen-manufacturing method. A strength result becomes meaningful within this context. The comparison should allow the original preparation to be reconstructed rather than relying on an operator’s recollection.
7. Test the mould before drawing conclusions about the casting
Choose prepared-sand tests according to the process: strength needed for handling, curing development, mould integrity, permeability or another property actually used to control the line. Do not try to maximise every result independently. Increased strength may accompany a dosage change whose consequences appear only during pouring.
Then use a trial geometry representative of production difficulties. A simple specimen does not reproduce every demand on a thin core, deep pocket or heavily heated area. Laboratory tests support selection and understanding; the trial casting checks integration of the mixture into the actual process.
Retain pouring parameters and shakeout observations. Also assess cleaning, residual sand and defects using a consistent method. A casting accepted after substantially longer cleaning can reveal an important operational change, even if its dimensional inspection is satisfactory.
8. Monitor the circuit over several cycles
The behaviour of a mixture prepared once does not necessarily describe a circulating system. Successive returns change stock composition. During industrial introduction, identify the initial proportion, material already in the circuit and locations where virgin material actually enters.
Establish a sampling schedule connected to stock-renewal stages rather than an administrative calendar alone. If a result changes, compare it with series changes, shutdowns, silo level and treatment settings. This interpretation can reveal alternating streams that a weekly average would conceal.
Before starting, define the conditions for continuing, correcting and returning to the previous situation. Being able to return to a controlled recipe makes the trial easier and limits improvisation. Retain a material reference and a recovery procedure, especially when lots are mixed in the same storage.
9. Close the material balance
Under sufficiently stable operation, incoming virgin sand compensates for losses and circuit outputs, among other things. However, stock variation can make a short balance misleading. Write the balance with inputs, outputs and inventory change. The difference between two weighings should not automatically be attributed to reclamation.
Identify collected dust, rejects, material removed with castings and handling losses separately where measurable. State which quantities remain estimates. This table helps distinguish a genuine treatment improvement from losses shifted to another location. It also permits campaigns of different durations to be compared.
Cost should include virgin material, treatment, transport, maintenance and observed manufacturing effects. The American life-cycle study shows that distances can change the environmental value of the scenarios 2. A sound local approach uses the data for the circuit studied, without promising automatic savings from a reuse percentage.
10. Turn the results into an operating rule
Validation should identify the material, circuit, binder, recipes and products covered. Record the chosen limits, monitoring frequency and actions associated with drift. A vague statement such as “compliant reclaimed sand” does not allow the next team to reproduce the decision.
Identify events requiring reassessment: a binder change, new core-sand stream, substantial reclamation modification, new supplier or persistent quality deviation. These modify the qualified system. Tracking them is more useful than permanent acceptance based on an old trial.
The target is consistent production with a justified amount of virgin sand, known returns and understood losses. A higher reuse proportion can be valuable, but it must remain compatible with quality and operation. The final dossier should explain why the mixture was selected and under what conditions that conclusion remains valid.
11. Interpreting drift without assigning its cause too quickly
Consider a campaign in which specimen strength decreases while the declared reclaimed-sand proportion remains unchanged. This is a reasoning example, not a result measured by Minerallum. The first step is to check lot identification and actual masses. An unchanged control-panel setting does not guarantee an unchanged physical proportion if an extractor’s output has varied.
Next examine the preparation method. Were the tests performed after the same delay? Was the sand stored under the same conditions? Did the laboratory change specimen preparation or equipment? A difference introduced by measurement can be mistaken for a material difference. Retained portions and records make it possible to repeat a controlled comparison at this stage.
If this initial check does not explain the deviation, compare the results with grading curves and return-stream information. A new core production series may have introduced a different stream, while a dust-removal adjustment may have changed the fine fraction. These explanations are hypotheses to test. They should not be presented as confirmed causes on the basis of coinciding dates alone.
The next trial should isolate one hypothesis. For example, prepare a mixture from retained portions, using the same masses and protocol, and compare it with the recent preparation. Keep the binder and conditions identical where possible. If several changes are necessary, record them and restrict the conclusion to that new combination. The result should reduce uncertainty rather than introduce additional variables.
Finally, divide the decision into what has been demonstrated and what still requires production monitoring. An effect reproduced in controlled preparation may justify action on the material. An effect observed only on the line also requires consideration of industrial conditions. This discipline avoids rejecting a sand because of a dosing problem or accepting returns whose drift was merely compensated by additional binder. It makes qualification useful beyond the first delivery.
Archive the conclusion, retained portions and supporting data together. For the next lot, the team should be able to retrieve the selected recipe, monitoring points and changes made since the trial quickly. A traceable decision is easier to reassess than an oral opinion. This organisation also reduces unnecessary discussion between purchasing, laboratory and production: each function works with the same definition of the material, process and expected result.
Specify the retention period and location of reference portions in the archive. A reference that is listed but cannot be found does not allow a trial to be repeated. Also associate each results file with its corresponding lot and protocol. This connection avoids comparing a recent analysis with a recipe that had already been modified.