Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: Ph.D. Dissertation
- Language: Farsi
- Document No: 44358 (06)
- University: Sharif University of Technology
- Department: Chemical and Petroleum Engineering
- Advisor(s): Soltanieh, Mohammad; Farhadpour, Farhad Ali
- Abstract:
- Permanent Downhole Gauge (PDG) has been employed since 1990s to acquire more information and better access to well and control production. PDG provides huge amount of pressure transient data during well test operation and production. However, such data are encountering with various noise and outliers. The noise and outliers can cause misleading and misinterpretation results. Therefore, denoising and outlier removal are two imperative steps before pressure transient analysis. The conventional denoising methods are mostly based on wavelet thresholding methods that preserve mean of data and have good performance when noise distribution is Gaussian distribution. However, noise distribution sometime is not following the ideal Gaussian distribution. Moreover, pressure data are often infected by outliers; therefore conventional methods are not able to capture such noises and outliers. One of the objectives of this research is to develop a reliable denoising procedure to capture noises and outliers simultaneously from pressure transient data and independently from data and noise distributions. Various procedures are selected and tested by using a set of synthetic noisy data. Among the available denoising methods, Local Balance of Fit (LBF) is demonstrating the best performance and is used for denoising in well test analysis applications. Contrary to wavelet thresholding methods, which already are preserving mean of the data and usually is influenced by the outliers, LBF method is working based on preserving median of the data, which is more robust than wavelet thresholding methods in the presence of outliers. After denoising and outlier removal, the data are used for reservoir characterization. However, huge amount of pressure data provided by PDGs can be used for characterizing reservoir heterogeneity. Many software packages have been developed for pressure transient analysis and parameter estimations. The problem now is to develop a well test procedure to acquire the required information from pressure transient data and able to capture reservoir heterogeneity in radial and/or vertical directions. Numerical well test models are providing flexible procedures for radially and vertically heterogeneous reservoirs. In radially heterogeneous reservoirs; there is a major uncertainty of non-uniqueness of solutions. Controlling reservoir boundaries and use of porosity-permeability correlation are two major factors that can effectively reduce non-unique solutions. In addition, logarithmic zonal discretization offers an appropriate and equally sensitive environment for all variables in radially heterogeneous zones to be searched effectively by nonlinear regression routines. As a result, rock properties and interior boundaries of zones can be identified together by such routines. In addition, contrary to analytical solution with limited application ranges, the numerical procedure can provide solutions under various boundary conditions. In vertically heterogeneous reservoirs, a new coupled wellbore-reservoir approach is introduced. The explicit inclusion of the wellbore flow model provides a medium for hydraulic communication between the different reservoir layers and also captures the wellbore storage effect directly. This means that different reservoir layers can flow to the well according to their flow capacity and pressure profile inside the wellbore. The significance of this new coupled approach is in its ability to capture both flow dynamics inside the wellbore and the pressure transient response of the reservoir. It offers an opportunity for developing a reliable numerical well test procedure, which reduces the need for multiple tests and is capable of delivering the parameters for the different layers simultaneously
- Keywords:
- Well Testing ; Noise Removing ; Outlier Removal ; Local Balance of Fit (LBF) ; Coupled Wellbore-Reservoir Model ; Radial Heterogeneity ; Vertical Heterogeneity
Digital Object List
-
محتواي کتاب
- view
Bookmark