Research On The Working Conditon Diagnosis And Optimization Technology Of The Screw Pump Well

At present, the screw pump is widely used in exploiting crude oil with high viscosity, high sediment concentration or high gas content. And the quantity of the screw pump well is increased year by year. But some problems have arisen in the process of application, including several faults like rod parting, rod-tube eccentric wear, badly abrasion between stator and rotor, and unsuitable working parameters, low operating efficiency. Through investigation and analysis, it is considered that feature extraction and working condition diagnosis method of the screw pump well remain to be completed. The unreasonable design proposal and working parameter cause the oil well productivity not to be able to obtain full play, and create the waste of system resources and energy. Therefore, in order to diagnose working condition of the screw pump well exactly, to improve the operating rate, to prolong pump inspection period and to reduce the energy consumption, this article has developed some researches on optimized operation and working condition diagnosis of the ground-driven screw pump well.Through theory analysis and contrastive arrangement to massive actual production data, status parameters (output, working fluid level, electrical machine power input) that can respond the working status correctly are chosen to be the proper vector of the working condition diagnosis. Several kinds of common working condition modes of the screw pump well have been summarized, taking as the target vector of working condition diagnosis. Using the pattern recognition ability of the neural network, the working condition diagnosis mathematical model has been established, adopting the improved BP neural network. And through preference to the network's architecture, good convergent effect has been received.Basing on the theory of oil well's supply and exhausted compatibility, through researches on the oil well inflow characteristics, wellbore pipe flow characteristics, subsurface screw pump characteristics and their joint feature, taking maximum system efficiency of the screw pump well as objective function, the optimized design as well as the efficient operation mathematical model of screw pump production system has been established. Thus, it can carry on the optimization design to the screw pump production system and can make the pump operating efficiently by adjusting pumping parameter when the working condition changes.Using the joint program of VB and Matlab, software about the diagnosis and optimization of the screw pump well has been developed. Using the computed result of the optimization software to instruct actual production, it has made good effects. Regarding the screw pump well to be put into production, appropriate pump type, rotate speed and other parameters have been designed through the software. And after using these parameters to instruct production, the screw pump wells work in stable status and can meet output request. Regarding to the screw pump wells that have already been put into production, pumping parameters with optimal efficiency has been sought. Using these values to adjust theses pumps, the system efficiency is increased by 15%. On the other hand, diagnoses are made by extracting proper vector from the screw pump wells and the results are in accordance with actual working condition mode. The results show that method about the diagnosis and optimization of the screw pump well proposed in this thesis is correct. The researches done in this article have provided the reference for the scientific management to the screw pump well, and have great important theoretic significance and applicable value to increase the system efficiency of screw pump well as well as the effect of energy saving.