High-temperature protection, structure optimization, and damage detection for missile-borne electronic devices

Abstract

In this study, the high-temperature performance of hypersonic missile radomes was investigated through a combination of numerical simulations, experiments, and machine learning-based damage detection. Two- and three-dimensional steady-state and transient heat conduction models were developed in MATLAB and COMSOL to examine the effects of different materials (ceramic, air and copper), filler configurations, and geometric shapes (cylindrical vs. conical) on radome insulation. Results indicated that introducing an air gap significantly reduced the peak temperature in the metallic layer (by up to 10–15 %), while shape optimization (e.g., cylindrical structures) further improved thermal uniformity. High-temperature damage simulations were performed using a Huffman Damage Coefficient, confirming that an air interlayer markedly decreased the damage area from 10 to 15 % to about 2–5 %. Laboratory tests employing infrared thermography validated the numerical predictions, showing consistency in temperature trends and structural integrity under ∼1000 °C heating. Furthermore, machine learning techniques (ResNet50) were applied to classify and detect microscopic damage, achieving a 95 % accuracy. These findings offer a robust theoretical and experimental basis for designing high-performance radomes and provide guidance for future integrated approaches to thermal protection in hypersonic missile systems.

Keywords

• Thermal insulation
• Radome design
• COMSOL simulation
• Machine learning