Research On Temperature Compensation Of SAW-RFID Humidity Sensor Based On Improved AFSA-BP Neural Network

The safe operation of power equipment has strict humidity requirements.When the humidity is too high,the insulation performance of the equipment will be degraded,and even serious faults will occur.When the humidity is too small,the surface of the secondary equipment circuit board is prone to generate static electricity and endanger the electronic components.Safety.Existing humidity testing instruments have problems such as physical connection and external power supply.Therefore,it is necessary to study new humidity sensors and corresponding temperature compensation measures to improve the practicality and accuracy of humidity detection.This paper first describes the characteristics of the two technologies of surface acoustic wave and radio frequency identification,and the label of the acoustic surface RF humidity sensor combining the two technologies.Based on this,the design principle of surface acoustic wave radio frequency identification(SAW-RFID)humidity sensor label is introduced: compare and select the piezoelectric substrate material;model the interdigital transducer(IDT)and analyze the model characteristics.And frequency characteristics,based on the delta model to determine the IDT parameters;briefly describe the model of the reflective grid and determine its parameters;introduce the moisture sensitive material and select the moisture sensitive material with excellent moisture properties;finally designed a small size,performance Excellent printed antenna.Then,the BP neural network with excellent nonlinear fitting performance is proposed to establish the temperature compensation model.In order to overcome the shortcomings of the BP neural network,the artificial fish population with low parameter requirements,good robustness and good global convergence is used.The algorithm(AFSA)optimizes the initial weight threshold of BP neural network;and proposes two improved methods for the problems of AFSA.Finally,the superiority of the improved algorithm is verified by the complex function optimization.Then the temperature compensation model of the humidity sensor is established based on the experimental data of the humidity sensing performance.Some experimental data are used to test the compensation effect of the model.The results show that the improved artificial fish swarm algorithm has a fast optimization speed,and the BP neural network temperature compensation model works well.