Wavelet neural network method for nonlinear compensation of capacitive weighing sensor

Abstract

The invention discloses a wavelet neural network method for the nonlinear compensation of a capacitive weighing sensor, and the method comprises the following steps: 1, employing the nonlinear compensation principle of the capacitive weighing sensor, wherein the nonlinear compensation principle of the capacitive weighing sensor is mainly based on the basic links shown in figure 1; 2, setting the weight of an input load of the sensor as G, and setting the output voltage as u, wherein u=f(G), and G and u are in a nonlinear relation; 3, setting a relation: y=f1(u)=kG if a compensation link is connected in series behind a sensor. According to the invention, the method determines a nonlinear compensation network of the sensor through wavelet neural network training. The nonlinear compensation principle of the capacitive weighing sensor is introduced for analysis of the topological structure of the network, and a network parameter training and initialization method is given. The result indicates that the method is good in robustness, is high in network training speed, is high in precision, can achieve the online compensation, and is practical in a testing field.

Description

technical field [0001] The invention relates to the technical field of sensors, in particular to a wavelet neural network method for nonlinear compensation of capacitance weighing sensors. Background technique [0002] In the weighing system of the modern electronic scale, the weighing sensor is used to obtain the weight signal, and then the weight signal is sent to the conditioning circuit for conditioning and display, and the weighing result is obtained. At present, resistance strain load cells are widely used in electronic weighing instruments, but limited by the strain limit, the relative change of the resistance of the metal strain wire of the resistive sensor is generally less than 1%, and the resistance of the strain wire is greatly affected by temperature. Big. In contrast, the relative change in capacitance of capacitive sensors can be greater than 100%, so the measurement range is much larger; capacitive sensors generally use metal as electrodes and inorganic mate...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the deficiencies of the prior art, the present invention provides a wavelet neural network method for nonlinear compensation of capacitance load cells, which solves the problem of nonlinearity of capacitance load cells

Images