Research On The Liquid Level Prediction And The Optimization Frequency Of Stroke Of Submersible Plunger Pump

Oil was the most important strategic resource and industrial raw material in the world.Efficient exploitation of underground oil was the focus of scientific and technological workers.The situation of crude oil lifting has been greatly changed,since of the emergence of new pumping units such as submersible plunger pump.It has subverted the inefficient working mode of traditional oil lifting machinery fundamentally,especially the rod-free pumping unit led by submersible plunger pump had came out.Its characteristics of high operating efficiency,high controllability and low production and maintenance cost were of great significance to energy developers all over the world.The submersible linear motor was the power source of the submersible plunger pump to lift the underground crude oil.The ground control cabinet controlled its up and down reciprocating operation through the submersible cable,and could adjust the running times at any time.In order to develop the characteristics of the simple and easy control of the submersible plunger pump further,on the basis of mastering the operation principle of the submersible linear motor,this thesis put forward the optimal control strategy of the submersible plunger pump based on the dynamic liquid level height adjustment of the running stroke,based on the prediction research of the dynamic liquid level time series.In order to obtain a large number of data to update the sample database,a downhole information acquisition system was designed.In order to solve the above key technical problems,this thesis mainly carried out the following research work.First of all,aiming at the problems of high cost,high error and high delay of obtaining liquid level by manual measurement;the soft measurement method was complex and easy to over fit,which could only give liquid level height,and could not get liquid level changing law,a liquid level prediction method based on extreme learning machine was proposed.And the method of optimizing its parameters by particle swarm optimization algorithm was given,so as to predict the law of liquid level changing during the future period of time.Through comparative analysis,the prediction model designed in this thesis had a high prediction accuracy and could meet the requirements of the optimal stroke planning of the submersible plunger pump.Secondly,on the basis of the study of dynamic liquid level prediction,this thesis present the optimal impulse calculation method based on the dynamic programming method,which sought the optimal solution in the impulse total matching space,however the calculation was too large.In order to apply this method to the actual production environment,a new optimization method based on fuzzy matching was proposed.On the premise that the optimization effect was close to the global optimization,the calculation amount of the algorithm was greatly reduced,making its practical application possible.Then,based on the theory of the optimization method of the flushing times proposed in this thesis,an example of the actual oil well data was given.Through the simulation experiment,the accuracy of the prediction of the driven liquid level height and the effectiveness of the optimization method of the flushing times were verified.Finally,in order to get more abundant data of liquid level,make the prediction result more accurate,the prediction model has the ability of real-time updating,and avoid the problem of increasing the failure rate due to the laying of new communication cables in oil wells,this thesis designed a downhole information acquisition system based on power line carrier,and carried out the field test.