Optimization of the sensitivity of the magnetoimpedance sensor of small magnetic fields by methods of sequential approximation and particle swarm

The use of multiparametric optimization of an unknown discrete function in the development of applied solutions for physical systems is considered. Such optimization is practically implemented in real time using modern data transfer protocols at high speed with continuously increasing computing power. Optimization of the sensitivity of a modern magnetic sensor based on high-frequency magnetoimpedance in ferromagnetic microwires is studied as an applied problem. Iterative methods of a global maximum search—successive approximation and particle swarm algorithms—have been used for this optimization. The output signal of the sensor depends non-linearly on both the internal magnetic properties of the microwire and the excitation mode, which requires a certain calibration to establish optimal excitation parameters. Using an automated installation, sensor output signals for various excitation parameters and external magnetic fields were measured. The results of the search for the global maximum of sensor sensitivity by the successive approximation method and the particle swarm algorithm were presented. It was established that the particle swarm algorithm turned out to be more effective and precise than the successive approximation method.. With various excitation parameters, the particle swarm algorithm always determined the maximum sensitivity of the sensor when varying the three basic parameters of the excitation signal: frequency, amplitude, and the constant component. The results obtained will be applied in the development of highly sensitive intelligent magnetic sensors and systems based on them.