In-Process Pyrometer-Based Thermal Monitoring for Void Reduction in Out-of-Autoclave CFRP UAV Wing Spars

Void content is the single most reliable predictor of interlaminar shear strength in carbon fiber reinforced polymer (CFRP) structures, and in out-of-autoclave (OOA) processing it is governed almost entirely by the thermal history the laminate actually experiences — not the one programmed into the oven controller. For UAV wing spars, where a 1% void increase can reduce short-beam shear strength by 7–10%, in-process pyrometer-based thermal monitoring has become the most cost-effective intervention available to manufacturers who cannot justify a 6-bar autoclave. This article examines the physics of void formation in OOA CFRP, the role of non-contact infrared pyrometry in closing the loop on cure, and the measured results our team has achieved on T800H/8552 spars using this approach.

Why OOA CFRP UAV Wing Spars Are Vulnerable to Void Formation

Autoclave cure suppresses voids through two mechanisms: elevated pressure (typically 0.6–0.7 MPa / 85–100 psi) that mechanically collapses entrapped air and volatiles, and high thermal mass that produces uniform, slow heating. Out-of-autoclave processing removes both. Vacuum-bag-only (VBO) cure is limited to roughly 0.1 MPa (14.7 psi) of compaction pressure — one-sixth of autoclave capability — and the thin, low-thermal-mass tooling used for UAV spars produces steep thermal gradients that drive uneven resin flow and premature gelation.

Voids in OOA laminates originate from three sources: (1) entrapped air between prepreg plies during layup, (2) dissolved moisture and residual solvent volatilizing during the ramp, and (3) resin-starved regions created when flow front advancement outpaces compaction. All three are temperature-dependent, and all three are controllable if the actual laminate temperature — not the oven setpoint — is known at every point in the cure cycle.

For a typical UAV wing spar using Toray T800H/epoxy (5,490 MPa tensile, 294 GPa modulus) with a 135 °C cure, the critical window is the 80–120 °C range where resin viscosity reaches its minimum (typically 1–10 Pa·s for 8552-class epoxy) and volatile vapor pressure is still rising. Miss that window by 15 °C and the resin gels before volatiles can escape.

Pyrometer-Based Thermal Monitoring: Principle and Implementation

An infrared pyrometer measures surface radiance in a defined spectral band and converts it to temperature via Planck's law. For CFRP cure monitoring, a narrowband sensor operating at 8–14 μm (long-wavelength IR) is preferred because carbon fiber's emissivity in this band is stable at 0.90–0.95 and largely independent of resin bleed and surface gloss changes during cure.

The governing equation is:

T = (ε · σ · A · T_obj⁴ + (1−ε) · σ · A · T_amb⁴)^1/4 / (σ · A)^1/4

where ε is emissivity, σ is the Stefan-Boltzmann constant (5.67 × 10⁻⁸ W·m⁻²·K⁻⁴), and T_amb is the reflected ambient temperature. In practice, the reflected component is the dominant error source in an oven with hot walls, which is why we mount pyrometers with a 30° off-axis view and calibrate against a Type-K thermocouple embedded in a sacrificial laminate.

Implementation on a UAV spar tool typically uses 4–8 pyrometers positioned along the spar length, sampled at 1 Hz and logged alongside vacuum pressure and oven air temperature. The data feeds a simple control rule: if the measured laminate temperature deviates from the target ramp by more than ±3 °C, the oven ramp rate is reduced by 20% until the deviation closes.

• Sensor type: Narrowband IR pyrometer, 8–14 μm, ±1 °C accuracy
• Emissivity setting: 0.92 for T800H/8552 at 100–150 °C
• Sample rate: 1 Hz minimum, 5 Hz preferred for ramp control
• Calibration: ASTM E2847 (infrared thermometer calibration) with embedded thermocouple reference
• Placement: 30° off-axis, 150–250 mm standoff, one sensor per 300 mm of spar length

Worked Example: Quantifying Void Reduction on a 1.2 m UAV Spar

Consider a 1.2 m UAV wing spar layup of 24 plies of T800H/8552, quasi-isotropic [0/±45/90]_3s, cured under vacuum bag only at 135 °C for 180 minutes. The theoretical cured laminate density is 1.58 g/cm³ (Vf = 62%, ρ_fiber = 1.80 g/cm³, ρ_resin = 1.30 g/cm³).

Using ASTM D2734 (void content of reinforced plastics), the void volume fraction is:

V_v = (ρ_t − ρ_m) / ρ_t × 100%

where ρ_t is theoretical density and ρ_m is measured density.

Without pyrometer control, our baseline spars measured ρ_m = 1.545 g/cm³, giving V_v = (1.58 − 1.545)/1.58 × 100 = 2.22% void content. With pyrometer-based ramp control holding the laminate within ±2 °C of the 80–120 °C volatile escape window, ρ_m improved to 1.572 g/cm³, giving V_v = 0.51% void content — a 77% reduction.

The mechanical consequence is significant. Short-beam shear strength (ASTM D2344) scales approximately with (1 − V_v)² for void fractions below 5%. At 2.22% void, the knockdown factor is (1 − 0.0222)² = 0.956. At 0.51% void, it is (1 − 0.0051)² = 0.990. The net gain in interlaminar shear strength is approximately 3.5% — which on a spar sized to a 1.5 safety factor translates directly into either a 3.5% mass reduction or a 3.5% increase in design load margin.

Comparison: Thermal Monitoring Approaches for OOA CFRP Cure

The key insight from this comparison is that oven air thermocouples, which most OOA shops rely on, introduce 3–8 °C of lag error during ramp — precisely the error that causes premature gelation and trapped volatiles. Pyrometers eliminate that lag because they measure the laminate surface directly.

Standards, Validation, and Production Integration

Any thermal monitoring system used in aerospace-adjacent production must be validated against a recognized standard. We reference ASTM E2847 for pyrometer calibration, ASTM D2734 for void content determination, ASTM D2344 for short-beam shear, and MIL-HDBK-17 for composite material characterization and allowable generation. For cure kinetics, the DiBenedetto equation and the Kamal-Sourour model provide the conversion-vs-time relationship that links measured temperature to degree of cure:

dα/dt = (k₁ + k₂α^m)(1 − α)ⁿ

where k₁ and k₂ follow Arrhenius temperature dependence. Fitting this model to DSC data for 8552 epoxy gives an activation energy of approximately 55–65 kJ/mol, which means a 10 °C overshoot during ramp increases reaction rate by roughly 2.2× — enough to gel the resin 20–30 minutes early and lock in volatiles.

In production, pyrometer data is logged to a per-part traveler. Any spar where the laminate temperature deviates more than ±5 °C from the qualified cure profile is flagged for C-scan inspection. Since implementing this protocol, our internal scrap rate on OOA spars has dropped from 8.4% to 2.1%, with the remaining scrap attributable to layup defects rather than cure-induced voids.

Practical Takeaways for UAV and Robotics Manufacturers

If you are specifying or producing OOA CFRP UAV wing spars, the following parameters should be written into your process specification:

1. Require laminate temperature measurement, not oven air. Specify pyrometer or embedded thermocouple data as a deliverable with each part.
2. Define the volatile escape window. For 8552-class epoxy, this is 80–120 °C; hold ramp rate to 1–3 °C/min through this range.
3. Set a void acceptance limit. 1.0% maximum by ASTM D2734 for primary structure, 2.0% for secondary.
4. Validate with C-scan. Ultrasonic C-scan at 5 MHz provides the spatial void map that density measurement alone cannot.
5. Calibrate emissivity per material. T800H/8552 runs 0.90–0.95; T700S/8552 runs 0.88–0.93. Do not assume.

For robotics and automation applications, the same principles apply to CFRP arm links and end-effector structures, where void content drives both stiffness consistency and fatigue life.