Study On Optimization Method Of Development Adjustment In Waterflood Oilfield Based On Data-Driven

Currently,most of China’s waterflooded oilfields have entered the stage of ultra-high water cut,and they often use development adjustment methods such as adjusting production allocation and injection volume,fracturing to improve efficiency,and side drilling of old wells to utilize remaining oil and improve development results.The understanding of reservoir utilization conditions,such as injection-production well correlation,waterflood wave range and intensity,and remaining oil distribution,is the basis for development adjustment.However,there are three difficulties in the analysis of reservoir dynamic conditions and the evaluation of adjustment plans:(1)There is a large difficulty in processing dynamic injection-production data: due to equipment factors,human factors,and random factors,dynamic data from injection-production wells contains a large amount of noise,and there are many wells with a long production history,so manually filtering data,handling bad data,noise reduction,and standardization are time-consuming and labor-intensive.(2)The operability of the injection-production relationship estimation method is poor.The current conventional methods for analyzing dynamic data,numerical simulation,and streamline simulation related to injectionproduction relationships involve complex models and calculations,with a lot of data,difficult fitting,and a lot of time consumption.(3)The optimization efficiency of oilfield development adjustment scheme is low.The work of reservoir numerical simulation is too large and too dependent on geological data.Relying on expert experience cannot be quantitatively evaluated,resulting in poor universality of adjustment and optimization design methods,making it difficult to promote and apply them.Therefore,this paper introduces data-driven optimization methods for the development adjustment of waterflooded oilfields based on reservoir engineering and seepage mechanics theories,including morphological filtering for noise reduction,extended Kalman filter simulation,and biomimetic intelligent optimization algorithms.To address the problem of processing abnormal values in injection-production dynamic data,this paper analyzes the characteristics of dynamic data and studies methods for classifying,identifying,and processing abnormal values.A dynamic data denoising filter based on mathematical morphology filtering method is established.The features of dynamic data that can be captured are optimized,and adaptive variable-scale structural elements are formed,with an accumulated denoising amount that is more than twice that of conventional structural elements.This filter can quickly identify and denoise abnormal values in injection-production dynamic data.To optimize the injection production response function and improve the reliability of injection production relationship estimation results,this paper proposes an injectionproduction correlation estimation method based on the extended Kalman filtering algorithm.By analyzing the response characteristics of injection and production,a response function between injection and production wells is proposed that includes the attenuation and temporality of flow transmission.This function can be calculated using dynamic data such as injection-production well production or water injection/liquid production index to improve the applicability of water flooding in the later stages.The injection-production correlation estimation method was applied to a water flooding development block in the Bohai Oilfield.The estimated liquid production was compared with the actual value,and the maximum error was 7.74%,with an average error of 0.61%.In order to solve the problem that the current development adjustment scheme cannot be quantitatively evaluated,a method for optimizing adjustment plans based on the principle of superposition and biomimetic intelligence algorithms is proposed.The biomimetic intelligence optimization algorithm is used to quickly search for optimal flow channels,and then the reservoir geological parameters and dynamic parameters of injection and production wells are used to calculate the flow and energy superposition of injection wells on production wells to optimize the development plan.The optimized deployment plan for the sidetracked wells is consistent with the numerical simulation results of the reservoir.This method avoids the complex calculation of historical fitting and achieves fast and quantitative evaluation of the water flooding development effect and various development adjustment plans for different well networks and geological conditions.The data-driven water-flooded oilfield development adjustment and optimization method proposed in this article can be used for well pattern adjustment,fracturing and water plugging optimization,and side-drilling horizontal well deployment optimization.It provides optimization design for the deployment of complex structure wells for residual oil recovery in Daqing and Bohai oilfield.After searching for dominant flow channels and conducting potential superposition calculations,the optimal direction and length of the complex structure wells are selected.The optimal solution is consistent with the calculation results of numerical simulation of oil production increase and input-output ratio.It also provides wellbore design for well drilling operations adjustment.The reservoir pressure prediction is performed using the injection-production correlation analysis method.The injection pressure of related water injection wells is reduced by50%,and the average error between the predicted and measured pressure values of the adjusted wells is 10%.