Design Of Remote Monitoring And Fault Warning System For Layered Water Injection

Layered water injection technology is one of the important links of oil production in China,which can make the pressure between reservoirs stable and improve the oil recovery rate.The monitoring and automatic management of layered water injection data can significantly improve the water injection effect.Therefore,the establishment of layered water injection remote monitoring system will provide real-time detailed water injection data for the layered water injection remote control management center,which can be analyzed and controlled according to the accurate water injection data.In the process of layered water injection,there are formation changes,string leakage and packer failure,which seriously affect the efficiency of water injection.Therefore,the application of fault early warning technology in layered water injection monitoring process can improve the safety of the system.When the system fails,it can accurately and quickly predict and alarm the fault,eliminate the fault in time,and reduce the loss.According to the actual situation of layered water injection wells,this paper analyzes the requirements of the whole water injection remote monitoring system,designs the system structure of layered water injection remote monitoring system according to the actual development environment,determines the remote communication scheme,designs its database structure according to the characteristics of layered water injection monitoring data,and adopts the C/S architecture,and uses QT language environment to control the server and the system The client is designed separately.Then,the causes of nozzle failure,string failure and leakage,packer failure and formation change are analyzed,and a layered water injection well fault early warning model based on support vector regression machine is established.Firstly,the water injection data is determined as the parameter of early warning research,and the data mining method is used to preprocess the historical data of water injection production.Finally,the support vector regression method is adopted The support vector regression method is used to establish the early warning model of water injection well fault,which completes the real-time monitoring of production data and the realtime early warning of stratified water injection process.Finally,the system is applied in the actual layered water injection process,and the function and performance tests are carried out respectively.In the function test,the functions of personnel management,data display and remote control are mainly tested,and then the CPU and memory utilization and average response time of the server under the running state of the software are tested by LoadRunner 11 in the system performance test.