This study aims to improve the manufacturing reliability of glass fiber-reinforced polymer (GRP) composite structures for submarine applications by quantitatively evaluating micro-voids generated during manufacturing and by relating the evaluation...
This study aims to improve the manufacturing reliability of glass fiber-reinforced polymer (GRP) composite structures for submarine applications by quantitatively evaluating micro-voids generated during manufacturing and by relating the evaluation results to mechanical performance and process conditions. Micro-voids, resin impregnation defects, and interfacial discontinuities formed during liquid composite molding (LCM) can reduce the mechanical performance and long-term reliability of submarine composite structures. Therefore, reliable micro-void evaluation and process-based defect control are essential for improving the manufacturing quality of GRP composite structures. GRP composite specimens were fabricated under different processing conditions. Internal micro-voids were evaluated using Ultrasonic C-scan testing, optical microscopy-based Image Analysis, and micro-computed tomography (Micro-CT). Ultrasonic attenuation was used as a wide-area non-destructive indicator of internal defects, whereas Image Analysis and Micro-CT were used to examine local two-dimensional void morphology, cross-sectional void content, and three-dimensional void distribution. The results confirmed that each evaluation method provides complementary information depending on its measurement principle, analysis range, and representativeness. The relationship between micro-void evaluation results and mechanical performance was analyzed using tensile and flexural tests. Higher internal defect levels were generally associated with lower mechanical strength. In particular, Ultrasonic C-scan-based attenuation showed a relatively strong negative correlation with tensile strength, with a correlation coefficient of approximately -0.703. This indicates that ultrasonic attenuation can be used as a representative non-destructive quality indicator related to the mechanical reliability of GRP composite structures. Process optimization was performed using ultrasonic attenuation as the response variable. Factorial Design analysis showed that defoamer concentration and mixing time significantly affected ultrasonic attenuation, whereas mixing speed and interaction terms were not significant within the investigated range. Based on the first-order regression model, the Steepest Ascent Method was applied in the attenuation-reduction direction to identify a promising process region near 0.045 wt% defoamer concentration and 222 min mixing time. Response Surface Methodology (RSM) with Central Composite Design (CCD) was then applied around this region. The model-based predicted optimum condition was determined to be approximately 0.049 wt% defoamer concentration and 232 min mixing time, with a predicted ultrasonic attenuation of approximately 21.0%. The predicted condition should not be interpreted as a final verified processing condition unless an additional confirmation experiment is performed. Nevertheless, the results demonstrate that Ultrasonic C-scan-based attenuation can serve as a practical non-destructive quality indicator and that defoamer concentration and mixing time are key process variables for reducing micro-void-related attenuation in submarine GRP composite structures. This study provides an integrated framework linking micro-void evaluation, mechanical performance correlation, and DOE-based process optimization for improving the manufacturing reliability of submarine composite structures.