This paper aims to propose a signal conditioning circuit design and nonlinear correction method for a linear differential transformer (LVDT) displacement sensor, aiming to achieve high-precision signal extraction and solve the nonlinear problem of its output signal.
The circuit generates a reference signal based on the principle of two-phase orthogonal vector lock-in amplification, and extracts the signal phase difference by combining low-noise amplification, bandpass filtering and phase sensitive detection. Aiming at the nonlinear characteristics of LVDT, an improved Harris hawk optimization (IHHO) algorithm is proposed to optimize the calibration method of Elman neural network (IHHO-Elman).
The experiment shows that the designed circuit structure is simple, does not require phase difference prediction and can extract signals with high accuracy. The IHHO-Elman network outperforms the standard Elman and HHO-Elman networks in terms of convergence speed, average error and error percentage, effectively solving nonlinear problems.
Innovatively combines lock-in amplification technology with IHHO-Elman neural network to provide a new hardware simplified and performance optimized solution for LVDT sensor signal conditioning and nonlinear correction.
