辫状河三角洲环境下基于本体信息量的单井沉积微相识别方法

Single well sedimentary microfacies identification method based on ontology information in braided river delta environment

辫状河三角洲的沉积相研究是油气储层研究的关键,在石油勘探开发中有重要的意义。探索单井沉积微相的快速自动识别的方法,是一项避免识别结果主观性强的高效且有益地尝试。考虑到测井成本和测井资料的不完整性(工区内所有井的测井资料不一定完整),因此,这里提出了一种基于本体信息量的单井辫状河三角洲沉积微相识别方法。该方法基于现有且成熟的辫状河三角洲沉积相划分体系和岩性及测井相标志,利用无歧义的Web本体语言OWL,建立辫状河三角洲沉积相本体。在本体基础上,对辫状河三角洲下可能发育的每种沉积微相的岩性特征和测井响应特征计算其信息量,然后得到每种沉积微相的四维信息量矢量,最终根据语义距离公式得到识别结果。在东海凹陷X气田的砂质辫状河三角洲的相关数据上进行验证,该方法对缺失一项测井响应特征数据的实验识别结果与已有地质认识的对比平均准确率为91.1%,交叉缺项实验验证平均准确率为83.3%。对单井辫状河三角洲沉积微相识别工作以及其他类型的三角洲沉积体系研究具有一定的参考意义。

Delta Sedimentary facies is the essential to reservoir research,which is of great significance in petroleum exploration and development.Braided river is an important part of the delta system.Exploring the rapid automatic identification method is a useful attempt to avoid the subjective identification of results.Taking into account the incompleteness of logging costs and logging data,it is impossible to use every logging technique for every well in the work area.Therefore,this paper presents a semantic recognition method of single well braided river delta sedimentary microfacies based on ontological information.According to the existing mature and highly recognized braided river delta system and associated sedimentary-facies features,this method uses the unambiguous Web ontology language OWL to establish the braided river sedimentary facies ontology that includes various possible sedimentary microfacies.On the basis of the ontology,we calculate the amount of information for each sedimentary-facies'features that include lithology and logging response characteristics.Logging response characteristics mainly include natural gamma-ray logging,resistivity logging and spontaneous-potential logging.Then four-dimensional information vectors for each deposited microfacies can then be obtained.At last,the result of semantic reasoning can be calculated from information entropy by semantic distance formula.As results,for delta sedimentary microfacies less a log response characteristics feature,the recognition accuracy rate of this method achieves 91.7%on the relevant data of the sandy braided river delta X gas field in the Donghai Depression.Meanwhile,accuracy of crossover experiments achieves 83.3%.The paper has certain reference significance for the recognition of sedimentary microfacies of the single well braided river delta and other types of delta systems.