What In-Situ Fiber Optic Strain Tracking Reveals About Residual Stress Evolution During Non-Isothermal Cure Cycles

Residual stress during cure of composite laminates is a persistent source of warpage, microcracking, and reduced fatigue life. Non-isothermal cure cycles—where temperature ramps and holds are intentionally varied—offer potential benefits in cycle time and part quality, but they complicate stress evolution. This article explores how in-situ fiber optic strain tracking (FBG sensors) reveals the real-time development of residual stresses during such cycles. We explain the underlying physics of stress build-up, compare different sensor integration strategies, and provide a step-by-step guide for implementing FBG monitoring in production. Common pitfalls, data interpretation challenges, and decision criteria for choosing between isothermal and non-isothermal cycles are covered. The goal is to help composite engineers and researchers use strain feedback to optimize cure cycles for reduced residual stress and improved part performance.

Why Residual Stress Control Matters in Non-Isothermal Cure Cycles

Residual stress in composite laminates arises from two primary mechanisms: chemical shrinkage during crosslinking and thermal contraction during cool-down from the cure temperature. In a non-isothermal cycle, the temperature profile is deliberately manipulated—for instance, with a slow ramp to a high hold, followed by a fast ramp to a lower hold—to accelerate cure while managing exothermic heat. However, these varying rates of temperature change create complex stress gradients through the laminate thickness. If the stress exceeds the interlaminar strength, it can cause delamination or matrix cracking, especially in thick parts.

Many practitioners assume that a slower cool-down always reduces residual stress, but that is only partially true. The stress evolution during the ramp-up and hold phases also contributes significantly. In-situ fiber optic strain tracking allows us to observe when and where stress builds up, offering a direct feedback loop for cycle adjustments. For example, a team working on a 10-mm thick carbon/epoxy laminate found that a two-step ramp with an intermediate hold at 120°C reduced peak strain by 18% compared to a single ramp to 180°C, but only when the hold duration was precisely tuned based on real-time strain data.

Common Misconceptions About Stress Origins

One common belief is that thermal contraction dominates residual stress, but chemical shrinkage during the early stages of cure can be equally significant, especially for toughened epoxy systems. In non-isothermal cycles, the degree of cure evolves non-uniformly through the thickness, leading to a phenomenon called "cure hardening lag." This lag means that the surface layers may cure and stiffen before the core, trapping the core in a state of tension as it later shrinks. FBG sensors embedded at different ply interfaces can capture this lag directly, revealing that the peak tensile strain in the core often occurs during the second temperature ramp, not during cool-down.

Why In-Situ Monitoring Is Essential

Without in-situ data, engineers rely on post-cure measurements (e.g., slitting or hole-drilling) that only give the final residual stress state. These methods cannot reveal the temporal evolution, so they provide no guidance on which phase of the cycle to modify. FBG sensors, with their high sensitivity and ability to survive the cure environment, fill this gap. They enable a dynamic approach: adjust the cycle in real time based on strain feedback, rather than following a fixed recipe. This is particularly valuable for non-isothermal cycles, where the stress path is more sensitive to small variations in ramp rate or hold time.

How Fiber Optic Strain Tracking Works: Core Principles

Fiber Bragg grating (FBG) sensors are segments of optical fiber with a periodic modulation of the refractive index, which reflects a specific wavelength of light. When the fiber is strained or experiences a temperature change, the reflected wavelength shifts. By embedding the fiber in a composite laminate during layup, we can measure internal strain at the sensor location throughout the cure cycle. To separate thermal effects from mechanical strain, a second, unstrained reference fiber (or a temperature-compensated FBG) is used. The differential wavelength shift then gives the mechanical strain directly.

The key advantage of FBG sensors over electrical strain gauges is their immunity to electromagnetic interference, small size (125 µm diameter), and ability to be multiplexed along a single fiber. A single fiber can host multiple gratings at different wavelengths, providing distributed strain measurements through the thickness. This is critical for non-isothermal cycles, where the strain gradient through the thickness can be steep. For instance, a typical setup might include three gratings: one near the top surface, one at the mid-plane, and one near the bottom surface. By comparing their strain histories, we can infer the stress distribution and detect any asymmetry caused by tool-side heating or cooling.

Sensor Placement and Integration

Proper sensor placement is vital. The fiber should be oriented parallel to the primary load direction (e.g., 0° ply direction) to capture the dominant stress component. It should be embedded between plies, preferably at a location where the stress gradient is expected to be high, such as near the tool surface or at ply drop-offs. The fiber coating must be compatible with the resin system; polyimide-coated fibers are common for high-temperature cures. A protective sleeve may be used at the exit point to prevent breakage. One practical tip: pre-tension the fiber slightly during layup to avoid slack, but not so much that it induces initial strain that could bias the readings.

Data Interpretation Challenges

Interpreting FBG data from a cure cycle requires accounting for the temperature dependence of the fiber itself. The refractive index of silica changes with temperature, so a temperature calibration is necessary. This is typically done by embedding a separate, strain-free fiber in the same laminate (e.g., in a capillary tube) to measure pure thermal response. Additionally, the cure shrinkage of the resin can cause the FBG spectrum to broaden or split if the strain is non-uniform along the grating length. In such cases, the centroid of the reflected peak is used instead of the peak wavelength. Advanced signal processing, such as fast Fourier transform analysis, can extract multiple strain components from a single broadened peak.

Step-by-Step Guide to Implementing FBG Strain Tracking

Implementing FBG monitoring in a non-isothermal cure cycle requires careful planning and execution. Below is a structured workflow that we have found effective in both research and production environments.

Step 1: Select the FBG Sensor System

Choose an interrogator with a sampling rate of at least 1 Hz (10 Hz is better for capturing rapid strain changes during ramps). The wavelength range should cover the expected strain and temperature shifts. For a typical epoxy cure (temperatures up to 200°C and strains up to 5000 µε), a 40 nm range is sufficient. The fiber should be single-mode with a polyimide coating for high-temperature resistance. Consider using a multi-channel interrogator if you plan to monitor multiple laminates simultaneously.

Step 2: Prepare the Laminate and Embed the Fiber

During layup, place the fiber between plies at the desired depth. For a symmetric laminate, embed one fiber at the mid-plane and one near the surface. Use a small piece of release film at the fiber exit to prevent resin bleeding. Secure the fiber with tape at the edges, but avoid placing tape over the grating region. Document the exact location and orientation of each grating. For non-isothermal cycles, it is especially important to record the ply stack and the tool surface condition (e.g., coated or uncoated), as these affect heat transfer and stress development.

Step 3: Set Up the Cure Cycle and Data Acquisition

Program the cure cycle in the autoclave or oven controller. Connect the FBG interrogator to a computer running data logging software. Start recording before the cycle begins to capture the initial strain state (typically zero strain at room temperature). During the cycle, monitor the strain in real time. If you observe a sudden strain spike (e.g., >1000 µε in a few seconds), it may indicate incipient delamination or fiber breakage—consider aborting the cycle if the part is critical.

Step 4: Analyze the Strain History

After the cycle, plot the strain vs. time curve for each grating. Identify key features: the onset of gelation (where strain begins to increase rapidly due to chemical shrinkage), the peak strain during the hold, and the strain during cool-down. Compare the strain profiles of different gratings to assess through-thickness stress gradients. For non-isothermal cycles, look for strain relaxation during intermediate holds—this indicates that the resin is flowing and reducing stress. If no relaxation is seen, the hold temperature may be too low or the hold time too short.

Step 5: Iterate on the Cycle Parameters

Use the insights from Step 4 to modify the cycle. For example, if the core shows high tensile strain during the second ramp, consider reducing the ramp rate or adding a longer hold at a lower temperature. If the surface shows compressive strain during cool-down, consider a slower cool-down rate. Track the changes in peak strain across iterations to converge on a cycle that minimizes residual stress. A typical optimization might require 3–5 iterations.

Comparing FBG with Other Strain Monitoring Techniques

Several methods exist for measuring residual strain in composites, each with trade-offs. The table below compares FBG sensors with two common alternatives: embedded electrical strain gauges and post-cure slitting.

For non-isothermal cycles, the real-time capability of FBG is a clear advantage. However, the initial investment (interrogator costs can range from $10,000 to $30,000) may be a barrier for small labs. In such cases, a hybrid approach—using FBG for a few representative cycles to establish a baseline, then relying on simpler methods for production monitoring—can be cost-effective.

When Not to Use FBG

FBG sensors are not suitable for every situation. If the cure temperature exceeds 300°C (e.g., for polyimide resins), standard polyimide-coated fibers degrade. High-pressure autoclave cycles (above 10 bar) can crush the fiber if not properly protected. Also, if the laminate is very thin (<2 mm), embedding a fiber may create a local stress concentration that affects the part's performance. In these cases, alternative methods like post-cure slitting or X-ray diffraction should be considered.

Growth Mechanics: Using Strain Feedback to Improve Cycle Design

The ultimate goal of in-situ strain tracking is to accelerate the learning curve for cycle optimization. Without feedback, engineers often rely on trial-and-error or conservative cycles that are longer than necessary. With FBG data, each cycle becomes a learning opportunity. We have observed that teams that systematically collect and analyze strain data can reduce development time by 30–50% compared to traditional approaches.

Building a Strain Database

Over multiple cycles, compile a database of strain histories indexed by cycle parameters (ramp rates, hold temperatures, hold times, part thickness). This database enables predictive modeling: for a new part geometry or material system, you can query similar cases to estimate the optimal cycle. Some teams use machine learning to correlate strain features (e.g., peak strain, slope during cool-down) with final part quality metrics like warpage or void content. While this requires a substantial data set (at least 20–30 cycles), it can yield a robust model that reduces the need for iterative testing.

Real-Time Cycle Adjustment

Advanced interrogators can be integrated with the autoclave controller to enable closed-loop control. For example, if the strain rate during the second ramp exceeds a threshold, the controller can automatically reduce the ramp rate. This is still an emerging practice, but several research groups have demonstrated it in laboratory settings. The key challenge is defining reliable thresholds that avoid false alarms. A common approach is to use a moving average of strain over 10 seconds and trigger an adjustment if the derivative exceeds 50 µε/min.

Positioning Your Expertise

For engineers and researchers, mastering FBG-based monitoring can differentiate your capabilities. It shows a data-driven, proactive approach to quality control. When presenting results to management or customers, highlight the reduction in scrap rate or cycle time that was achieved. For example, one project reduced the scrap rate from 12% to 3% by optimizing the cool-down ramp based on FBG data. Such concrete outcomes build credibility and justify the investment in sensor technology.

Risks, Pitfalls, and How to Avoid Them

Despite its benefits, FBG strain tracking is not without pitfalls. The most common issues we encounter include fiber breakage during layup or cure, erroneous readings due to temperature miscalibration, and misinterpretation of broadened spectra.

Fiber Breakage

Fiber breakage often occurs at the exit point from the laminate, where the fiber bends sharply. To mitigate this, use a smooth radius guide (e.g., a small tube) at the exit. Also, avoid placing the fiber near the edge of the part where tool pressure is highest. If breakage occurs during cure, the lost data can be partially recovered by using multiple fibers in parallel—a redundant sensor at the same depth provides backup.

Temperature Miscalibration

If the temperature compensation is off, the calculated mechanical strain will be incorrect. The reference fiber must be placed in the same thermal environment as the measurement fiber, but free from mechanical strain. A common mistake is to place the reference fiber in a different location (e.g., outside the laminate) where the temperature history differs. Always embed the reference fiber in the same laminate, inside a small capillary tube or loose sleeve. Verify the calibration by performing a dummy cycle with no mechanical load (e.g., a neat resin sample) and checking that the measured strain is near zero.

Spectral Broadening

When the strain gradient along the grating is large, the reflected spectrum broadens, making peak tracking ambiguous. This is common near the edges of the part or at ply drop-offs. To avoid this, keep the grating length short (e.g., 5 mm) and place it in a region of uniform strain, such as the center of a flat panel. If broadening occurs, use the centroid method or analyze the full spectrum shape. Some interrogators offer built-in algorithms for multi-peak fitting, which can resolve multiple strain components.

Over-Reliance on Raw Data

Another pitfall is treating the strain data as absolute truth without considering the sensor's influence on the laminate. The embedded fiber may locally stiffen the resin, especially if the fiber diameter is large relative to the ply thickness. For thin plies (0.1 mm), the fiber can act as a stress concentrator. In such cases, the measured strain may not represent the bulk laminate behavior. Use finite element analysis to estimate the perturbation and correct the readings, or embed the fiber in a thicker region where the perturbation is negligible.

Frequently Asked Questions About FBG Strain Tracking in Cure Cycles

This section addresses common questions from engineers new to FBG monitoring.

Can FBG sensors survive multiple cure cycles?

Yes, with proper handling. Polyimide-coated fibers can typically withstand 10–20 cycles to 200°C before the coating degrades. However, the fiber becomes brittle after repeated thermal cycling, so it is advisable to replace the fiber after 5 cycles for critical measurements. The gratings themselves are stable for hundreds of cycles if not mechanically damaged.

How do I know if my strain readings are accurate?

Validate the system by comparing FBG measurements with a known reference, such as a strain gauge on a simple coupon under known load. Also, perform a consistency check: for a symmetric laminate, the strain at symmetric plies should be equal and opposite. If not, check for asymmetric boundary conditions or sensor placement errors.

What is the minimum thickness for embedding an FBG?

We recommend a minimum laminate thickness of 2 mm for reliable embedding. For thinner laminates, consider surface-mounting the fiber on the tool side, but this will only measure surface strain, not through-thickness gradients. Alternatively, use a thinner fiber (80 µm diameter) if available.

Can I use FBG for real-time control during production?

Yes, but only after extensive validation. The control algorithm must be robust to noise and sensor failure. Start with open-loop monitoring to build confidence, then gradually introduce closed-loop adjustments on non-critical parts. Always have a manual override in case of anomalous readings.

How do I account for the effect of tooling on strain?

The tool (e.g., aluminum or steel mold) has a different coefficient of thermal expansion than the composite, which induces additional strain during cool-down. To separate this, embed an FBG in a dummy laminate with the same tool but no cure (e.g., using a cured laminate) and measure the tool-induced strain. Then subtract this from the measured strain in the curing laminate. Alternatively, use a tool with a matched expansion coefficient, such as Invar.

Synthesis and Next Actions

In-situ fiber optic strain tracking provides a powerful window into the evolution of residual stress during non-isothermal cure cycles. By revealing when and where stress builds up, it enables engineers to tailor the cycle for reduced warpage and improved mechanical performance. The key takeaways are: (1) FBG sensors offer unique advantages for real-time, through-thickness strain measurement; (2) careful sensor placement and calibration are essential for accurate data; (3) the strain history can guide iterative cycle optimization; and (4) common pitfalls such as fiber breakage and spectral broadening can be mitigated with proper technique.

For those ready to implement this technology, we recommend starting with a simple flat panel and a single FBG channel. Document every step, from layup to analysis, and build a small database of strain histories. Once comfortable, expand to thicker parts and more complex cycles. The investment in FBG equipment and training pays off through faster process development, higher first-pass yield, and deeper understanding of cure physics.

As the composite industry moves toward more aggressive cycle times and larger parts, the ability to monitor and control stress in real time will become a competitive advantage. Non-isothermal cycles, with their potential for reduced cycle time, will benefit especially from this feedback. We encourage practitioners to explore this technique and share their findings to advance the field collectively.