Adaptive modified ant colony optimization algorithm for global temperature perception of the underground tunnel fire
Bin Sun
Abstract
Due to complexity and diversity of the real tunnel structures and fire scenes, there exist few effective methods for temperature prediction of the underground tunnel fire. Hence, here a data-driven adaptive modified ant colony optimization (ACO) algorithm is developed to predict the global temperature field of the underground tunnel fire. The algorithm is not limited to the special tunnel structure and fire scene, in which temperature of the underground tunnel fire can be predicted based solely on some sensor data. Meanwhile, to solve the problem that the choice of the model parameters is usually case by case in the traditional ACO algorithms, the model parameters can be adjusted and determined adaptively in the improved algorithm. A numerical example of full scale fire test of an underground tunnel is used to verify the ability and effectiveness of the developed algorithm. The results show that the maximal error is less than 10%. The developed algorithm can be used to predict global temperature with respect to time of the underground tunnel fire, which is easy to be used in engineering application due to its advantages.