关键词: BP神经网络, 随钻测井曲线, 地质导向, 权重系数, 自然伽马

Abstract: 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.

Key words: BP neural network, LWD curves, geosteering, weight cofficient, natural gamma