Loading...

Modeling of Input Fluid Flow Contribution Data from Different Layers of Heterogeneous Reservoir Using Deep Learning Neural Network Algorithm

Salehi, Ali | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57523 (06)
  4. University: Sharif University of Technology
  5. Department: Chemical and Petroleum Engineering
  6. Advisor(s): Masihi, Mohsen; Jamshidi, Saeed
  7. Abstract:
  8. Recognizing the production rate of different fluids such as oil, water and gas from different reservoir layers is very important in planning the appropriate approach for reservoir development. This issue becomes more important when the reservoir is layered. One of the common methods in analyzing the flow of production layers and determining the contribution of each layer in the production rate of each well is the use of the Production Logging Test (PLT). The flow of incoming fluids through different layers is evaluated by considering the real data of a well or by simulating a reservoir model around the well (single well model). Specifically, by using two-phase flow models, temperature models, appropriate relationships for fluid properties, and considering the optimization algorithm or minimizing the error between the measured data and the simulated data, and the inversion of the flow rate evaluation results is done. Another approach in evaluating and obtaining fluid flow rate at any depth can be using deep learning neural network algorithm which is used in this project and using 4 parameters measured from logs (density of fluid mixture, amount of water holdup, the inner diameter of the well and the apparent velocity of the fluid mixture) as well as the two-phase model of Aziz et al., the flow rate of oil, water and gas fluids is estimated separately at each depth. Finally, the efficiency of the model is compared with the output of the Emeraude software, and in general, very little difference is observed between the final results
  9. Keywords:
  10. Production Logging Tools (PLT) ; Two-Phase Flow Modeling ; Optimization Algorithms ; Neural Networks ; Deep Learning ; Layered Reservoir ; Emeraude Software

 Digital Object List

 Bookmark

...see more