Informatics Publishing Limited

Hongmei Liu ¹, Xia Meng ², Wei He ³, Rong Li ¹

Affiliations
1 Department of Information Technology, Beijing Vocational College of Agriculture, Beijing 102442, CN
2 Architectural Design and Research Institute of Tsinghua University Co. Ltd., Beijing 100 084, CN
3 Department of Civil Engineering, Tsinghua University, Beijing 100 084, CN

Abstract

For weak rock mass with notable rheological property, instability is mostly caused by flowing deformation. As the basis for the design of roadway supporting structure, the rheological parameters of surrounding rocks are of great importance. Unfortunately, the rheological parameters obtained from indoor tests often fail to reflect the geological defects in a large research area due to the impact from constraints of sampling representativeness, sampling disturbance and testing technical level. What is worse, field tests are time-consuming, unrepeatable and costly. To solve these problems, this paper conducts inversion of the rheological parameters of surrounding rocks based on the BP neural network. Taking a mine roadway as an example and considering the vault subsidence data in the entrance section, the author applies FLAC^3D in numerical simulation, adopts BP neural network for network learning and sample training, and performs displacement inversion of the rheological parameters of the surrounding rocks in the section.

Keywords

Numerical Simulation, Creep, BP Neural Network.

Full Text

     

Shu Yan ¹, Jianhua Yao ², Ji Li ¹, Xu Zhu ³, Chao Wang ⁴, Wei He ¹, Shizhong Ma ¹

Affiliations
1 Northeast Petroleum University, Fazhan Road 99, Daqing 163 000, CN
2 Daqing Oilfield Sanwei Engineering Inspection Co.Ltd.,Yinhu Road 3, Daqing 163 000, CN
3 No.1 Oil Production Company, Daqing Oilfield Co. Ltd., Zhongqi Road 34, Daqing 163 000, CN
4 No.2 Oil Production Company, Daqing Oilfield Co. Ltd., Qiangxi Road 15, Daqing 163 000, CN

Abstract

Based on the dense well pattern data of PI2a unit of Putaohua oil layer in Sazhong development area of Daqing oilfield as well as modern sediment and outcrop, the paper attempts to determine point bar configuration parameters through a combination of quantitative methods to establish a three-dimensional point bar configuration model. On the basis of this model, the paper proposed a parameter-constrained three-dimensional configuration model of lateral accretion interbeds via spatial translation according to the planar trajectory of interbeds, so as to maximally restore the geological structure and ascertain the anisotropism and residual oil inside point bars. The results show that there are 15 interbeds in the study area with an azimuth of 320° and an obliquity range of 3.5°-13.1°. The interbeds average an interval of 9.5m, and each of them stretches 25m on average. The profile of interbeds is concave, which is slow, steep, slow from the bottom to the top, and the surface takes the shape of crescent. The profile of lateral accretion bodies is characterized by a echelon-like superimposed structure. The interior of point bars shows strong anisotropism due to interbeds, with residual oil being scarce at the bottom, abundant in the middle and extremely abundant at the top. Given the distribution pattern, the paper put forward suggestions for future residual oil potential tapping.

Keywords

Configuration Modelling, Anisotropism, Configuration Model, Configuration Parameter, Point Bar.

Full Text