Recycled-Plastic Injection Moulding: When Cavity Sensors Help

Cavity sensors can improve quality control in recycled-plastic injection moulding when local pressure, temperature or flow information explains defects that machine measurements miss. Their value depends on a verified link between the signal and part quality, followed by a reliable inspection or control response that justifies tool modification and maintenance.

Material variation and machine-side control

Recycled-plastic injection moulding creates a quality problem when a stable machine recipe produces changing parts as material batches change. In-cavity sensors are useful when information from inside the mould helps explain or control variation that machine measurements leave unresolved. The investment question is which quality characteristic needs better observation, how a sensor signal relates to that characteristic, and whether the resulting intervention pays for the tool modification and its maintenance.

Viscosity describes resistance to flow. Changes in this property can alter how material fills the cavity even when the programmed settings remain unchanged. A 2026 study documented by TU Delft addresses inconsistent properties across post-consumer recycled thermoplastic batches and tests adaptive cavity-pressure control. Its experiment used a plate mould, virgin polypropylene and two recycled batches over 50 cycles. The reported mean part-mass deviation was 0.21% from the reference. That is evidence for a particular mass-control experiment, not proof of dimensional, mechanical or surface quality across production tools.

The existing machine deserves an explicit baseline. ENGEL describes its iQ weight control software as comparing each injection pressure curve with a reference and adjusting the injection volume; its extended system also uses flow time. These are manufacturer descriptions of commercial functions. Their significance for selection is that material variation can already be addressed through machine-side control. A cavity-sensor proposal should demonstrate additional useful information under the same material, tooling and acceptance conditions, rather than assume that more instrumentation automatically produces better parts.

Conventional cavity sensors and thin-film arrays

Conventional in-cavity pressure and temperature sensors offer another established option. Kistler’s application guidance distinguishes direct contact sensors, indirect pressure measurement behind an ejector and contactless measurement behind the cavity wall. It relates sensor location to the defect being monitored: a sensor close to a critical feature answers a different question from one near the gate. This suggests defining the relevant defect before selecting hardware. An incomplete fill at the end of a cavity and variation in packing require different evidence, even if both eventually produce rejected parts.

Thin-film sensing adds a different form of spatial observation. Fraunhofer IST describes a tool insert with 13 temperature-measuring points deposited on its surface. Their arrangement captures the advance of the flow front, meaning the leading boundary of the incoming polymer. The institute also describes processing the signals on an edge device, a computer located close to the machine, and presenting a quality indication before the mould opens. Those functions should be distinguished from evidence that the system automatically corrects all causes of defective parts.

The Fraunhofer programme entry for Eurosensors 2026 identifies a thin-film temperature-array application specifically for recycled-material injection moulding. It describes real-time monitoring of the flow front. The entry does not provide a quantified industrial scrap reduction, a validated false-acceptance rate or long-term performance across recyclate suppliers. Monitoring, predicting whether a part will pass inspection, and changing the process automatically are separate capabilities. A proposal should state which of them is available and what evidence supports it.

Linking sensor signals to accepted part quality

Qualification should connect each cycle’s signal to a measured outcome. A practical trial can retain independent measurements of weight, dimensions, surface defects and any mechanical property required by the customer, while recording the material batch and machine conditions. If a prediction model is involved, evaluate it on material and production runs excluded from model fitting. Count defective parts accepted by the system separately from sound parts rejected by it. Those two errors have different costs, and a single overall accuracy percentage can obscure their commercial significance.

The proposed control response also needs a defined limit. A system that identifies an unusual cycle may trigger inspection or sorting without changing machine settings. A system authorised to adjust the process needs a validated range of permitted changes and evidence that an improvement in one characteristic does not degrade another. Maintaining part weight, for example, is insufficient acceptance evidence where a customer also specifies dimensions or strength. These are qualification requirements inferred from the difference between a process signal and the product characteristics being sold.

The financial case should be built for the relevant family of tools. Include the instrumented insert, wiring, electronics, integration, calibration, maintenance, spare parts and the time required to establish a dependable relationship between signals and defects. Compare these costs with measured reductions in rejected parts, inspection effort or production interruption using the same accounting boundary. A thin-film array is a plausible candidate when the location and timing of flow contain valuable information that simpler measurements miss. A production commitment requires a trial showing that this additional information changes quality decisions reliably over the intended material and operating range.