Fault Diagnosis System Of The Wireless Measurement System For Greenhouse Environment

With the development of information technology,the Internet of Things(IOT)technology has been developed rapidly,and the application of the technology of agricultural IOT in the facility agriculture is becoming more and more extensive,and the future is bright.The development of sensor technology,wireless network,micro-computer technology,Internet,which promote the monitoring and management of greenhouse environmental information further.A large number of sensors for the main greenhouse environment monitoring of greenhouse environment monitoring system in facility agriculture has been deployed,but it is with high humidity and high temperature working environment in facility agriculture,and as a result of measurement and control system for sensor fault of the greenhouse environment is very frequent.To collect environmental information by sensors in facility agriculture is the foundation of greenhouse monitoring system.However,the abnormal data of sensor fault can cause serious harm to the accuracy of the greenhouse environment monitoring and the reliability of the environmental control.Therefore,the research of fault diagnosis based on wireless measurement and control system of greenhouse environment has important economic significance and engineering application value.In this paper,based on the facility agriculture IOT system GHIOT developed by Jiangsu University,the sensor fault diagnosis method based on spatial and temporal correlation was proposed for the abnormal environment parameters of sensor nodes,and the network fault diagnosis was carried out to solve the problem of data and time asynchronous in the data transmission between the intelligent gateway and the server or the data acquisition unit.The fault diagnosis function module was added to the GHIOT system,and to verify the results of the system.The result of the system is proved to be effective.The main research work includes the following aspects:(1)Research on spatial-temporal correlation of greenhouse environmental parameters.According to the characteristics of the greenhouse environment changes slowly and the strong coupling of environmental parameters of adjacent nodes,the research on spatial-temporal correlation analysis was carried out.Based on the analysis of spatial-temporal correlation of greenhouse environment to predict the data of sensor nodes,to predict the temporal correlation of environment data by the first-order autoregressive time series prediction algorithm et al.To spatial correlation prediction of environmental parameters,and to predict the spatial correlation of homogeneous sensors based on the information of neighbor node parameters,and to predict the spatial correlation of heterogeneous sensor based on regression model.The verification results show that the temporal correlation of first-order autoregressive prediction method for temporal prediction algorithm works best,and its variance is 1.439.The spatial correlation prediction variance of homogeneous and heterogeneous sensors was 1.493 and 1.883.The prediction results can effectively reflect the change of greenhouse environment.(2)Sensor fault identification based on the comparison of spatial-temporal information.In order to judge the accuracy of the sensor data in the greenhouse environment measurement and control system,this paper proposed a sensor fault identification method based on the comparison of node information.Firstly,the method is a sensor fault detection method that based on the PCA,and it is to achieve the sensor system fault detection by monitoring statistics T2 and SPE changes.When the system detects a fault moment,the sensor nodes were to realize the different sensor fault identification using the comparison of node information based on spatial-temporal characteristics,and to comparing the effects with the different methods,comparison of node information based on temporal scale,spatial scale and spatial-temporal scale,for multi-sensor fault identification.Verification results show that the sensor fault detection method based on PCA can effectively realize the preliminary fault detection of the sensor system,and its fault detection rate reached 90.23%.The sensor fault identification method based on the comparison of node information,that takes the time and spatial scale into consideration,can effectively achieve the specific fault sensor positioning.Compared to the traditional recognition methods of sensor fault,find the fault recognition accuracy of sensor fault identification method mentioned in this paper is 95.14%,it can effectively ensure the efficiency of fault diagnosis and improve the accuracy of fault diagnosis,and it can reduce the false alarm rate,and it is with reliability and accuracy.(3)Multi-sensor data fusion and reconstruction of greenhouse environment based on spatial-temporal correlation.In order to meet the robustness requirements of the sensor data of greenhouse environment monitoring system,this paper proposed a multi-sensor data fusion and reconstruction algorithm based on spatial-temporal correlation.The prediction values,the spatial and temporal correlation,were used for data fusion based on the improved support function algorithm,and validated the data fusion results.According to the fault characteristics and the control effect,the sensor fault was judged,and the sensor fault data was reconstructed.Verification results show that the input variables,the spatial-temporal prediction,using the improved support function of the data fusion algorithm can effectively assign a dynamic weighted value for the predicted value of spatial-temporal correlation,and the value of fault data of sensor nodes average Correct Diagnosis Rate(CDR)is 96.72%,and the average False Alarm Rate(FAR)is 3.48%,the abnormal data RMSE of data fusion is 1.07,its data fusion results are better than the average value algorithm and the traditional support function algorithm.The optimal estimates value,generated by data fusion,can factually reflect the change of greenhouse environment with reliability and accuracy.(4)Greenhouse Wireless Sensor Network fault diagnosis.According to the sensor network in the process of running time,number of abnormal situation,for example,time asynchronous and data asynchronous exceptions,this paper proposed a fault diagnosis method of greenhouse wireless sensor network,the method utilized reference broadcast synchronization mechanism,based on JSON cross platform transmission and data acquisition unit read command again to realize time synchronization and data synchronization of the wireless monitoring system of intelligent gateway and the Server,the data acquisition unit realize the network fault diagnosis and recovery.The results show that the detection and fault diagnosis method is proposed in this paper can effectively realize the greenhouse environment wireless monitoring system network fault testing,processing,judgment and fault recovery.The accuracy of fault recovery of time asynchronous is 83.34%,and the accuracy of fault recovery of data asynchronous is 98.57%,and it can realizes the time synchronization between the intelligent gateway and the Server,and it can realizes the data synchronization between the intelligent gateway and the server,and the data acquisition unit.The fault diagnosis effect obviously and the reliability high.(5)Systems implementation test.The fault diagnosis system of the wireless monitoring system of greenhouse environment was integrated,and the system deployment and running test was carried out in Lishui botanical science base of JAAS.The fault detection,identification,diagnosis and recovery of abnormal data based on spatial-temporal information are presented.Aiming at the network fault of the wireless measurement and control system of greenhouse environment,the data asynchronous and time asynchronous fault detection and fault recovery are carried out.The further verification in performance of the system including: detection,sensor node abnormal data sensor node environment parameter information identification,diagnosis and data of sensor fault data recovery,network fault recovery,network fault time and data synchronization.The verification results show that the fault diagnosis system of the wireless monitoring system can realize the monitoring for greenhouse environment,and it has a good stability by run the system for testing for a long time,and the fault diagnosis effect is obvious,and the reliability of the system is improved.