A High-Density Electrical Inversion Model Coupled with Neural Network Algorithm

The high-density electrical method technology holds significant importance in the exploration of geological hazards in coal mines. In recent years, a nonlinear inversion method exemplified by the BP (Backpropagation) neural network has gained widespread prevalence in the inversion of high-density electrical methods. Addressing issues such as susceptibility to local minima, slow convergence, and subpar inversion accuracy encountered in high-density electrical inversion using the BP neural network method, this study integrates the GA (Genetic Algorithm) with the BP neural network for the nonlinear inversion of high-density electrical methods. This algorithm initially establishes the structure of the neural network, randomly generates the initial weights and thresholds of the neural network to create a population of chromosomes, and then carries out chromosome replication, crossover, and mutation. The optimal value yielded by the genetic algorithm is employed as the initial weight and threshold of the BP neural network, thereby improving its performance in the inversion of high-density electrical data. Through the inversion calculations of two sets of models, it becomes evidently clear that the GA-BP algorithm demonstrates superiority and feasibility in the two-dimensional nonlinear inversion of high- density electrical methods.