Research And Implementation On Pumping On-line Automatic Diagnostic System

Oil production site in general has a wide geographical distribution, and the working environment is bad. It is time and effort consuming to search information of the oil-field. With development and wide application of Computer Science, the mode of oil-field production can be improved. And because of the complicated down-hole environment, underground faults are difficult to detect. So it is very important to develop a set of pumping on-line automatic diagnostic system, which means a lot to the energy-saving and benefit- increasing.This paper design a set of pumping wells on-line automatic diagnosis system. This system realizes function of data on-line monitoring, measurement and statistics, energy saving analysis and alarming. This system is based on ASP.NET and Sql Server database design and can timely and accurately provide basis for decision making.As is known, dynamometer card is the important basis of down-hole fault diagnosis. This paper mainly studied the down-hole fault diagnosis, and put forward a method based on two-dimensional Fourier transform and BP neural network. According to Gibbs equation, this paper made finite difference program which can compute the down-hole dynamometer. This process remove the effect of pumping rod and many other factors. Furthermore, changing the down-hole dynamometer to gray matrix, and transforming the gray matrix with two-dimensional Fourier transform. Finally, use the transformed coefficients of gray matrix as the characteristics of the dynamometer. By using BP neural network, this paper designed a classifiers, and giving three underground fault examples which has proved the feasibility of this method. Automatic diagnosis system has been applied in the oil-field, the proposed dynamometer identification method also been verified. Next step is to make future observe and improvement in practice.