Optimization Design of the Electromagnetic Flowmeter Excitation Structure Based on GA-BP-GWO Neural Network

This article proposes a back propagation (BP) neural network based on genetic algorithm (GA) optimization and uses the grey wolf optimizer (GWO) to search for the optimal excitation structure of an electromagnetic flowmeter (EFT). The excitation structure of the EFT is simulated by numerical simulations, and the magnetic field distribution is compared with that of the original machine to demonstrate the correctness of the simulations. Then, such numerical simulations are carried out on many excitation structural models, which are constructed using the pole shoe. The bending angle of the excitation coil and the height from the pipe wall are applied as input sets, while the magnetic field strength and uniformity are applied as output sets and then compared with a single hidden layer BP neural network and a double hidden layer BP neural network, radial basis function (RBF) neural network, and adaptive neuro-fuzzy inference system (ANFIS). The results indicate that using GAs for optimizing neural networks yields the best performance. The simulation and modeling of the optimal excitation structure results in the larger-the-better characteristic of magnetic field strength, it improved by 72%. The smaller-the-better characteristic of magnetic field uniformity decreased by 3.7%. The overall satisfaction level increased by 78%.