순환신경망을 이용한 미측정 정두압 데이터 예측
Prediction of Missing Tubing Head Pressure Using Recurrent Neural Network
1
조회
0
다운로드
6
인용
0
참고문헌
키워드
초록
이 연구에서는 딥러닝의 일종인 순환신경망(Recurrent Neural Network, RNN) 기법을 사용하여 생산정의 일부구간의 누락된 압력자료를 예측하는 방법을 제시하였다. 장단기 기억(Long Short-Term Memory, LSTM) 셀을 채택한 RNN은 시계열 데이터에 대한 예측 성능이 우수한 방법이다. 생산정의 정저와 정두에 설치된 게이지를 통해 측정된 압력자료는 게이지 오류, 노이즈, 아웃라이어, 미작동 등 다양한 원인으로 비정상적인 결과를 제공하는 경우가 빈번하다. 이 연구에서는 이러한 현장자료의 전처리를 통해 주어진 자료의 보정을 선행한 후, LSTM-RNN을 통해 누락된 정두압력을 예측하였다. 이 결과를 PipeSim을 사용하여 생산관 모델링을 통해 도출한 정두압력과 비교하여LSTM-RNN의 적용성을 분석하였다.
Abstract
This study suggests a method to restore missing pressure data in production wells using a recurrent neural network (RNN). The use of RNNs with long short-term memory (LSTM) cells is considered to be an efficient method for predicting time series data. Pressure data measured by gauges at the bottomhole and wellhead often provide abnormal values due to various reasons such as gauge malfunction, noise, outliers, and non-operations. In this study, pre-treatment of the abnormal measurement data was carried out, and then missing wellhead pressure data were restored through the LSTM-RNN method. The applicability of the LSTM-RNN was investigated in comparison with the pressures estimated by the production well model of PipeSim.
참고문헌 (23)
- Ahn, S., "Data-driven inverse modeling with a pre-trained neural network at heterogeneous channel reservoirs", Journal of Petroleum Science and Engineering, 170, pp.785~796 (2018)
- Ashena, R., "Bottom hole pressure estimation using evolved neural networks by real coded ant colony optimization and genetic algorithm", Journal of Petroleum Science and Engineering, 77, pp.375~385 (2011)
- Beggs, H., "Production optimization using nodal analysis", pp.88~ (1991), Oil and Gas Climate Initiative Publications
- Bowie, B., "Machine learning applied to optimize Duvernay well performance", SPE Canada Unconventional Resources Conference, pp.1~24 (1898), SPE
- Carvajal, G., "Intelligent digital oil and gas fields", pp.75~100 (2018), Gulf professional publishing
- Ehrahimi, A., "A robust model for computing pressure drop in vertical multiphase flow", Journal of Natural Gas Science and Engineering, 26, pp.1306~1316 (2015)
- "Excelsior" (2019)[URL]
- Guyaguler, B., "Optimization of well placement in a gulf mexico waterflooding project", SPE Reservoir Evaluation and Engineering, 5(3), pp.229~236 (2002)
- Hochreiter, S., "Long short-term memory", Neural Computation, 9(8), pp.1725~1780 (1997)
- Jacobs, T., "Automated drilling technologies showing promise", Journal of Petroleum Technology, 67(6), pp.50~55 (2015)
- Jahanandish, I., "Predicting bottomhole pressure in vertical multiphase flowing wells using artificial neural networks", Journal of Petroleum Science and Engineering, 75, pp.336~342 (2011)
- Jang, I., "Well-placement optimisation using sequential artificial neural networks", Energy Exploration & Exploitation, 36(3), pp.433~449 (2018)
- Kotezhekov, V., "Development of automatic system for decline analysis", SPE Russian Petroleum Technology Conference, pp.1~11 (1877), SPE
- Lee, K., "Feature extraction using a deep learning algorithm for uncertainty quantification of channelized reservoirs", Journal of Petroleum Technology, 171, pp.1007~1022 (2018)
- Min, B., "Optimal well placement based on artificial neural network incorporating the productivity potential", Energy Sources, Part A : Recovery, Utilization, and Environmental Effects, 33(18), pp.1726~1738 (2011)
- Osman, E., "An artificial neural network model for predicting bottomhole flowing pressure in vertical multiphase flow", SPE Middle East Oil and Gas Show and Conference, pp.1~11 (2005), SPE
- Sagheer, A., "Time series forecasting of petroleum production using deep LSTM recurrent networks", Neurocomputing, 323, pp.203~213 (2019)
- Sayyafzadeh, M., "Reducing the computation time of well placement optimisation problems using self-adaptive metamodeling", Journal of Petroleum Technology, 151, pp.143~158 (2017)
- Spesivtsev, P., "Study of the transient terrain-induced and severe slugging problems by use of the drift-flux model", SPE Journal, 22(5), pp.1570~1584 (2017)
- "Tensorflow" (2019)[URL]
- Tian, C., "Applying machine learning techniques to interpret flow rate, pressure and temperature data from permanent downhole gauges", SPE Reservoir Evaluation and Engineering, 22(2), pp.386~401 (2019)
- Unrau, S., "Machine learning algorithms applied to detection of well controle events", SPE Kingdom of Saudi Arabia Annual Technical Symposium and Exhibition, pp.1~10 (1881), SPE
- Vyas, A., "Modeling early time rate decline in unconventional reservoirs using machine learning techniques", Abu Dhabi International Petroleum Exhibition & Conference, pp.1~23 (1882), SPE