地层原油物性参数大数据处理方法研究

The Research on Data Processing Method of Physical Property Parameter of Formation Crude Oil

生产测井数据解释过程中遇到测量数据不完整的生产井时,解释人员往往会主观加上人为因素,凭借经验进行解释,其结论是否合理需要进一步研究.文章提出基于生产测井大数据平台,提出一种流体物性参数深度学习处理方法,很好的解决了前述问题,科学合理的完善了生产测井数据中不足之处.由于生产测井解释中关键参数流体物性值计算不准确,直接导致生产测井解释精度偏低,不能解决生产实践问题.传统的计算方法,是利用在特定环境中分析出的公式来计算流体物性值,这些公式多有一定局限性、不能通用.研究流体物性参数深度学习大数据处理方法,设计新的计算模型,能综合考虑不同地区的井况特色,结合深度学习算法自适应调整模式,形成一套科学合理的算法系统弥补测量中的不足,达到准确计算地层原油物性参数值的目的.通过对生产测井数据的验证,其计算精度达到95%,满足测井解释的生产需求,能提高产液剖面解释精度,为提高油气增产提供有力的科学依据.

When the production logging data interpretation process is encountered in the production wells with incomplete measurement data, it is often subjective and human factors, and the conclusion is reasonable and need to be further studied. Based on the data platform of production logging, this paper puts forward a kind of fluid physical parameters deep learning processing method, which is a good solution to the above problems. Because of the inaccurate calculation of the key parameters in the production logging interpretation, the accuracy of production logging interpretation is low, which cannot solve the problem of production practice. The traditional calculation method is to use the formula to calculate the physical property of the fluid in a particular environment, which has some limitations and cannot be used in general.Study on fluid physical parameters and depth study of large data processing method, a new model to calculate the design, can consider different regions of the well condition characteristics, combined with the depth of learning algorithm based on adaptive adjustment mode, formed a set of scientific and reasonable algorithm system for measurement, the accurate calculation of formation crude oil physical property parameter value is reached. Through the verification of production logging data, the calculation precision can reach 95%,which can improve the interpretation accuracy of production fluid profile, and provide a scientific basis for the improvement of oil and gas production.