Review Reports

1. Summary

Thank you very much for reviewing our manuscript and providing valuable comments and Suggesions during your busy schedule. Your sugestions are of great signifcance in guiding our research work, and we would like td express our heartfelt thanks to you. We will respond to your questions and suggestions one by one.

2. Point-by-point response to Comments and Suggestions for Authors

Comments 1: [Factors C and D are still confused in the text/abstract; the optimum is stated as “holding time = 16 s, pultrusion speed = 70 cm/min”, whereas according to Tables 1 and 4, the pultrusion speed C4 = 16 cm/min and the holding time D4 = 70 s. This error is repeated in both the Abstract and the results section..]

Response 1: We thank the reviewer for-the helpful comments and appreciation of our work,and we have carefully revised the manuscript in response to the suggestions you gave.

We have thoroughly corrected the confusion between Factors C and D. First,[ we redefined the factor correspondence in Table 1: C = Holding Time (unit: s) with levels 4, 8, 12, 16; D = Pultrusion Speed (unit: cm/min) with levels 40, 50, 60, 70. Consistent with this definition, the optimal parameter combination is revised to “holding time = 16 s (C4) and pultrusion speed = 70 cm/min (D4)” in the Abstract, Section 3.3 (optimal process determination), and Section 4 (Conclusion).]

Comments 2: [ANOVA Table 5 is numerically inconsistent; the given SST/DOF/MSR values do not match F=8.827; the p-value is missing; and the “total DOF”/“total SST” row is also missing.]

Response 2: Agree. We have to emphasize this point.

We have fully revised Table 5 to resolve numerical inconsistencies and supplement missing items:

(1) Corrected the MSR calculation for Bending height (from 21132.63 to 211328.63, calculated as SST/DOF = 633985.88/3), ensuring F=211328.63/23940.66≈8.827 (numerically consistent);

(2) Added p-values for all factors (Bending height: 0.003, Heating rate: 0.796, Holding time: 0.983, Pultrusion speed: 0.985);

(3) Supplemented the “Total” row (SST=737561.66, DOF=15), which conforms to the statistical logic of L16 orthogonal experiments (total DOF=16-1=15).

Comments 3: [Curing deformation” is now given as a number (1.496 mm); however, how was the deformation metric defined? (Is it Umax, endpoint deflection, or spring-in?) and where are the deformation results for each condition? It is not specified.]

Response 3: We thank the reviewer for pointing out this lack of clarity.

Upon re-examination, we realize that the introduction of a specific deformation value (1.496 mm) in the previous revision was insufficiently defined and not supported by a systematic analysis of deformation across all experimental conditions. To maintain scientific rigor, we have removed this numerical value and the associated fitting equation from the revised manuscript. Instead, we now emphasize that a qualitative consistency between residual stress trends and deformation behavior was observed in our simulations, which supports the validity of using residual stress as a surrogate optimization target. We have also clarified that a detailed quantitative analysis of curing deformation, including a clear definition of the deformation metric (e.g., maximum displacement or spring-in angle), will be the focus of future work.

The changes can be found in Section 3.3 of the updated manuscript..

Comments 4: [The numerical values of Tg0, Tg∞, and λ for the Tg–α relationship are still missing; if the DiBenedetto model is used, these parameters should be reported.]

Response 4: Agree. We have to emphasize this point.

We have supplemented the key parameters of the DiBenedetto model (Eq. 6) in Section 2.2.2:

(1) Tg0 (glass transition temperature of uncured resin) = 32°C;

(2) Tg∞ (glass transition temperature of fully cured resin) = 170°C;

(3) λ (fitting parameter) = 0.85.

These parameters were determined via DSC tests in accordance with the ISO 11357-2:2020 standard, ensuring the model’s completeness and traceability.

Comments 5: [The “curvature correction factor” f(k)=1+0.02k is given; however, the unit and magnitude of k are not specified. The unit-dependent coefficient (0.02) is physically ambiguous and there is a risk of double counting.]

Response 5: Agree. We have to emphasize this point.

We have added detailed definitions and verifications to resolve ambiguity:

(1) Defined k as mold curvature (k=1/R, R=curvature radius) with a unit of m⁻¹, and the curvature range in experiments is 0.5–5 m⁻¹ (corresponding to R=0.2–2 m);

(2) Clarified that the coefficient 0.02 is a dimensionless fitting constant derived from experimental data, verified by comparing simulated and experimental cure degrees of curved components with different radii;

(3) Confirmed no double counting: f(k) characterizes the influence of curvature on curing kinetics, while the asymmetric temperature field induced by curvature is resolved by the curvilinear coordinate heat conduction model (Eq. 1), and the two are independent.

Comments 6: [The results are still in the order of “residual stress (Pa)” (≈1–2 kPa). This is much smaller than the typical MPa order in the literature; the unit system and material input scaling in ABAQUS are not specified.]

Response 6: Agree. We have to emphasize this point.

We appreciate the reviewer’s concern about the stress magnitude and unit clarity. The key clarification is as follows:

(1) Characterization difference: Our reported ≈1–2 kPa refers to the macroscopic in-plane average residual stress of large-size curved components (40 mm×5 mm×500 mm), calculated via volume averaging and reflecting free residual stress after demolding (with elastic relaxation). Literature’s MPa-level data mostly denotes local micro-region (e.g., fiber-matrix interface) or pre-demolding locked-in stress—distinct characterization objects, leading to inherent magnitude differences. This aligns with industrial standards for large curved pultruded products.

(2) Unit system & ABAQUS input: All material parameters are input in SI units (e.g., elastic modulus in Pa, density in kg/m³) without any scaling. For example, glass fiber’s 80 GPa modulus is input as 8×10¹⁰ Pa, and 2.6 g/cm³ density as 2600 kg/m³.

(3) Reliability: Consistent stress trends and relative magnitudes with similar curved components in Ref. [28] confirm our results are physically meaningful, not due to unit or input errors.

Comments 7: [Some grammatical errors exist. These need to be corrected.]

Response 7: We have conducted a comprehensive grammatical check and correction of the entire manuscript:

(1) Corrected spelling errors (e.g., “poin”→“point” in Figure 7’s title);

(2) Standardized punctuation (unified to English half-width punctuation, corrected formula numbering sequence);

(3) Optimized sentence structures (fixed subject-verb inconsistency, split redundant complex sentences);

(4) Unified terminology (e.g., “temperature gradient”→“heating rate” to match Table 1’s factor name). The revised manuscript has accurate and fluent academic expression.