Implementation of network-computing and NN based remote real-time oil well monitoring system

This paper introduces a real-time oil well monitoring system based on network computing and Neural Network (NN) technologies. In the system, some enterprise computing techniques, such as Browser/Server web mode, JMS, Message-oriented Middleware (MOM) and Java Applet are employed. In addition, GPRS wireless communication is used to achieve remote transmission of oil well data. This scheme adopts Java Applet that operates at client side (Web browser) to receive messages "pushed" by server through JMS (Java Message Service). In this way, server and client are able to communicate in time so that client side can reflect the real-time oil well data and the fault diagnosis result. In remote monitoring center, the fault diagnosis station takes the responsibility for the fault detection and diagnosis of oil pumping units by means of Neural Networks and Evolutionary computation. This solution accomplishes network share of oil well information, improves the efficiency of system development. The system has already been applied successfully to an oil field, and has got the anticipated results.