Research On Sensor Fault Diagnosis Of Cigar Tobacco Leaf Production IoT System

In recent years,chinese cigar consumption market has developed rapidly,and the demand for high-quality cigar leaves is increasing.At present,the production of cigar leaves in China is still in its infancy,and the quality of tobacco leaves produced is low,which is difficult to meet the needs of the industry.Most domestic cigars use imported tobacco leaves as the main raw material,which greatly limits the development of chinese cigar industry.The Io T technology has been widely used in the field of agriculture.By setting up the Io T monitoring system for cigar leaf production,the real-time monitoring of the growth environment and growth of cigar leaves can be realized,which can provide real-time data support for the scientific management decision of cigar leaf production,thus promoting the development of cigar industry.However,Io T sensing devices are prone to various faults,especially sensors of sensing devices,in harsh working environments such as high temperature,high humidity and strong ultraviolet radiation.As an important part of the Io T monitoring system for cigar leaf production,the probability of sensor aging or failure is greatly increased,and a series of problems such as sensor data drift,deviation and even complete failure occur.In this paper,the sensor of the Io T monitoring system for cigar leaf production is taken as the research object,and the scientific sensor fault diagnosis method is explored to achieve the purpose of effective diagnosis of sensor faults.This is of great significance to ensure the stability and reliability of the overall operation of the cigar tobacco leaf production Io T monitoring system.The specific research contents are as follows :(1)Development of cigar tobacco leaf production Io T monitoring systemThe Io T monitoring system for cigar leaf production is designed and implemented,which provides equipment and data basis for subsequent sensor fault diagnosis research.Based on the six-domain model of the Io T,the overall architecture design of the Io T monitoring system for cigar leaf production was completed,and the Io T monitoring system for cigar leaf production including sensing processing terminal and wireless transmission terminal was further developed.After the coordinated development of system hardware and software,the field installation and debugging of the Io T monitoring system for cigar tobacco leaf production were realized.Under the normal working conditions of the Io T system,the fault of the Io T system is studied and analyzed.(2)Research on sensor fault detection model based on STW-PCAThe purpose of sensor fault detection is to detect the fault of sensor in the Io T monitoring system of cigar leaf production in time through analysis.Aiming at the problem that sensor data is easy to produce abnormal data such as data loss and outliers due to environmental factors or human factors,a sensor abnormal data recognition and repair algorithm based on improved local outlier factor algorithm(LOF)and grey model GM(1,1)is proposed.Aiming at the problem that the high dimension of sensor data affects the efficiency of sensor fault detection,a principal component analysis(PCA)method based on adjacent sliding time window(STW)is proposed,and a sensor fault detection model based on STW-PCA is constructed.The experimental results show that the STW-PCA model can effectively detect the occurrence of sensor deviation fault and drift fault in the Io T monitoring system of cigar tobacco leaf production,and the fault detection rate is above 80 %.(3)Research on sensor fault diagnosis modelAiming at the problem of low accuracy and low efficiency of sensor fault diagnosis,a sensor fault diagnosis model based on BP neural network is proposed to make up for the low accuracy of traditional contribution rate analysis.Aiming at the problem that the BP neural network model is easy to fall into local extremum and the training rate is slow,an improved eagle optimization algorithm(IAO)is proposed.The improved eagle optimization algorithm is used to optimize the BP neural network model,and a sensor fault diagnosis model based on IAO-BP neural network is constructed.The experimental results show that the fault diagnosis rate of IAO-BP neural network model is 94.25 %.(4)Design and implementation of sensor fault diagnosis systemCombined with the sensor fault diagnosis requirements in the process of cigar tobacco leaf production,the sensor fault diagnosis system of cigar tobacco leaf production is developed,and the server-side host computer software,mobile phone APP software and Web page system are developed.The program scripts of the sensor fault detection model and the fault diagnosis model are embedded into the Web page end system,and the sensor fault diagnosis of the Io T monitoring system for cigar leaf production is realized,which significantly improves the stability and reliability of the overall operation of the Io T monitoring system for cigar leaf production.