Research On Oil Well Control Method Based On Moving Fluid Level Interval Prediction

Most of the oil wells in China are in the middle and late stages of exploitation,and the phenomenon of imbalance of oil well supply and discharge,or even "empty pumping" is quite common.The dynamic fluid level is the most direct parameter to reflect the well's fluid supply capacity,and a reasonable dynamic fluid level height is essential to improve production management and optimize production control.During the normal exploitation of oil wells,the down pumping depth will not change,but the reasonable dynamic fluid level interval of oil wells will change with the oil recovery process.Therefore,real-time monitoring of the dynamic fluid level of oil wells and maintaining the dynamic fluid level within a reasonable range is the key to ensure the normal production of oil fields and improve the oil recovery rate.In this paper,we first fill in the missing data caused by environmental or human factors in the production process,and then model the filled data for soft measurement of dynamic fluid level.In view of the problem that the fluctuation of dynamic level prediction output leads to uncertainty of control and frequent adjustment,the interval prediction method is introduced into the dynamic level prediction based on the process of whether the dynamic level is in a reasonable interval in the field production process.To address the problem of high sampling repetition rate of the traditional Bootstrap method,an improved Bootstrap method is proposed to extract the output data required by the Bootstrap algorithm in the process of training the soft measurement model of dynamic fluid level by Ada Boost integrated learning algorithm,replacing the traditional resampling process,and finally obtaining the Bootstrap interval prediction sample set.Secondly,we establish the relationship between the dynamic fluid level and the number of strokes to realize the real-time adjustment of the dynamic fluid level of the pumping unit,and develop a well control strategy to control the dynamic fluid level within a reasonable range by the obtained dynamic fluid level prediction interval.