| Autor: |
Vladimir Krichevsky, Artur Aslanyan, Rushana Farakhova, Andrey Akimov, Emil Gadelshin, Ilnur Yamalov, Vladimir Ovcharov, Danila Gulyaev |
| Rok vydání: |
2020 |
| Předmět: |
|
| Zdroj: |
Day 4 Thu, November 05, 2020. |
| DOI: |
10.4043/30101-ms |
| Popis: |
The massive industry digitalization creates huge data banks which require dedicated data processing techniques. A good example of such a massive data bank is the long-term pressure records of Permanent Downhole Gauges (PDG) which became very popular in the last 20 years and currently cover thousands of wells in Company RN. Many data processing techniques have been applied to interpret the PDG data, both single-well (IPR, RTA[1]) and multi-well (CRM [2] - [5] and various statistical correlation models). The ability of any methodology to predict the pressure response to rate variations and/or rate response to pressure variations can be easily tested via numerical modelling of synthetic fields or via comparison with the actual field production history. This paper presents a Multi-well Retrospective Testing (MRT, see Appendix A and [6] - [9]) methodology of PDG data analysis which is based on the Multi-well Deconvolution (MDCV, see Appendix B and [10] - [20]) and the results of its blind testing against synthetic and real fields. The key idea of the MDCV is to find a reference transient pressure response (called UTR) to the unit-rate production in the same well (specifically called DTR) or offset wells (specifically called CTR) and then use convolution to predict pressure response to arbitrary rate history with an account of cross-well interference. The MRT analysis is using the reconstructed UTRs (DTRs and CTRs) to predict the pressure/rates and reconstruct the past formation pressure history, productivity index history, cross-well interference history and reservoir properties like potential and dynamic drainage volumes and transmissibility. The results of the MRT blind testing have concluded that MRT could be recommended as an efficient tool to estimate the current and predict the future formation pressure without production deferment caused by temporary shut-down for pressure build up. It showed the ability to accurately reconstruct the past formation pressure history and productivity index. It also reconstructs the well-by-well cross-well interference and reservoir properties around and between the wells. The blind-test also revealed limitations of the method and the way to diagnose the trust of the MRT predictions. Engineers are now considering using MRT in Company RN as a part of the selection/justification package for the new wells drilling, conversions, workovers, production optimization and selection of surveillance candidates. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
