摘要
通过传统的钻井水力学模型计算出的井底压力虽然能够反映出井下压力的变化趋势,但不能准确地反映实际测量值的大小。该文基于原有的钻井水力学模型,以及现场实际测得的井底压力,采用BP神经网络技术建立自学习算法,使钻井水力学模型经过实测井底压力数据训练后,计算结果误差达到最小。现场应用效果表明,通过修正后的模型计算结果误差较小,处于安全钻井压力监测误差范围之内,能够满足正常钻井的要求。
Through the conventional drilling hydrodynamic model to calculated the bottom-hole pressure value,it could reflect the variation tendency of the bottom-hole pressure but fails to reveal the actual measured values. Based on drilling Hydrodynamic model and actual bottom-hole pressure value measured on site,this article adopts BP neural network technology to build self-study algorithm so as to make the deviation of the bottom-hole pressure to the minimum after the calculation of the drilling Hydrodynamic model trained by the on-site measured value .The actual application on site shows deviation of the calculation results which are revised by the model was slight and right among the allowed range of deviation under the safe drilling pressure supervision. It could meet the demand of the routine drilling.
出处
《自动化与仪表》
北大核心
2012年第11期9-11,22,共4页
Automation & Instrumentation
关键词
井底压力
BP神经网络
水力学模型
修正
bottom-hole pressure
BP neural network
hydrodynamic model
amending
