Thickness Optimization Design of Heat-protective Clothing Based on LMNNSA Method

The performance of the heat-protective clothing is critical for protecting the safety of workers in the high temperature environment. How to design a good quality and comfortable heat-protective clothing is a challenge problem. To solve this problem, this paper proposes a novel data-driven optimization method integrating Levenberg-Marquardt back propagation neural network (LMNN) and simulated annealing algorithm (SA), named as LMNNSA, to design the thickness of each layer for the heat-protective clothing. Firstly, the heat transfer model of the environment-clothing-skin system is established to obtain the temperature distribution of each layer for the heat-protective clothing. Secondly, the LMNN method is presented to analyze the relationship among the temperature of any point in the heat-protective clothing, its related position in each layer, and the working time. Thirdly, the SA optimization algorithm is used to acquire the optimal solution of the fabric thickness when reaching the minimal temperature of the skin surface. Finally, the feasibility and effectiveness of the proposed LMNNSA method is demonstrated by comparing the changes of skin temperature before and after the improvement of the thickness. This study provides a guideline for designing the fabric thickness of heat-protective clothing.