基于BP神经网络模型的随钻测井曲线预测

Prediction of LWD curves based on BP neural network model

随着剩余油气资源的勘探开发难度不断增加,随钻地质导向仪器的应用越来越广泛。但在实际钻井过程中,随钻地质导向仪器中测量地质参数的传感器与钻头之间存在一定的距离,导致所测地质参数与正钻地层间具有一定的延迟,影响了地质导向仪器的导向效果。利用BP神经网络的自组织、自学习、非线性动态处理等优势,通过自然伽马反映地层岩性,而地层岩性与钻压、排量、钻时等钻井工程参数具有一定关联,以待钻井周围已钻邻井资料作为样本进行学习和训练,建立了BP神经网络预测模型。利用当前正钻井随钻地质导向仪器已测量井段的参数对BP神经网络的权重系数进行精细调整,最后采用该BP神经网络模型及当前钻井工程参数对钻头正钻地层的自然伽马参数进行预测。实例应用表明,该方法预测计算得到的自然伽马数据较为准确,与实测自然伽马曲线相比具有较高的相似性,可以作为仪器测点零长井段地质参数的预测方法,为实际钻井过程中地质导向提供一定的参考。

With the increasing difficulty of exploration and development of remaining oil and gas resources,the application of geological steering instrument while drilling is more and more extensive.However,in the actual drilling process,there is a certain distance between the sensor measuring geological parameters in the geological steering instrument while drilling and the drill bit,which leads to a certain delay between the measured geological parameters and the formation being drilled,and affects the steering effect of the geological steering instrument.This paper makes use of the advantages of BP neural network such as self-organization,self-learning and nonlinear dynamic processing,as well as the natural gamma response to the formation lithology,which has a certain relationship with the drilling engineering parameters such as WOB,displacement and drilling time.It uses the drilled temporary well materials around the well to be drilled as samples for learning and training,and establishes an initial BP neural network prediction model.Then,the initial BP neural network is fine adjusted by using the parameters of the measured well section of the current geosteering instrument while drilling.Finally,the natural gamma parameters of the formation being drilled by the bit are predicted by using the BP neural network model and the current drilling engineering parameters.The example application shows that the natural gamma parameters predicted and calculated by this method are more accurate,and have a higher similarity with the measured natural gamma curves.It can be used as a prediction method for the geological parameters of zero length well section at the instrument measuring point,and provide a certain reference for geological steering in the actual drilling process.