This paper proposes a comprehensive electromagnetic-noise optimization design method based on the forgotten least squares method for the electromagnetic performance and vibration-noise synergistic optimization problems faced by Fe-based soft magnetic composite core reactors in high frequency applications. The nonlinear magnetization model of the core material is constructed through experiments, and the magnetic-mechanical coupling kinetic equations are established by combining Maxwell’s system of equations and mechanical vibration theory. The least squares method incorporating the forgetting factor is used to identify the parameters of the twoinertia mechanical system, and adaptive filtering is realized with the help of parameter sensitivity analysis. Simulation experimental results show that the sound insulation using the scheme of this paper can be up to 14.6dB(A), and the total sound pressure level of measurement points B and C at 1m outside the reactor cabinet after optimization is reduced by 10.2dB(A) and 11.5dB(A) respectively compared with the pre-optimization period, and its comprehensive noise reduction effect reaches more than 10.9dB(A), while the temperature rise is controlled within a reasonable range, which verifies the validity of the methodology and the applicability of engineering.
