Advanced Machine Learning Methods for Production Data Pattern Recognition

Abstract Despite long production histories, operators of mature oilfields sometimes struggle to account for reservoir compartmentalization. Geological-led workflows do not adequately honor legacy production data since inherent bias is introduced into the process of allocating production by interpreted flow units. This paper details the application of machine learning methods to identify possible reservoir compartments based on legacy production data recorded from individual well completions. We propose an experimental data-driven workflow to rapidly generate multiple scenarios of connected volumes in the subsurface. The workflow is premised upon the logic that well completions draining the same connected reservoir space can exhibit similar production characteristics (rate declines, GOR trends and pressures). We show how the specific challenges of digitized legacy data are solved using outlier detection for error checking and Kalman smoothing imputation for missing data in the structural time series model. Finally, we compare the subsurface grouping of completions obtained by applying unsupervised pattern recognition with Hierarchal clustering. Application of this workflow results in multiple possible scenarios for defining reservoir compartments based on production data trends only. The method is powerful in that, it provides interpretations that are independent of subsurface scenarios generated by more traditional workflows. We demonstrate the potential to integrate interpretations generated from more conventional workflows to increase the robustness of the overall subsurface model. We have leveraged the power of machine learning methods to classify more than forty (40) well completions into discrete reservoir compartments using production characteristics only. This effort would be extremely difficult, or otherwise unreliable given the inherent limitations of human spatial, temporal, and cognitive abilities.

Download Full-text

Improving stock trading decisions based on pattern recognition using machine learning technology

PLoS ONE ◽

10.1371/journal.pone.0255558 ◽

2021 ◽

Vol 16 (8) ◽

pp. e0255558

Author(s):

Yaohu Lin ◽

Shancun Liu ◽

Haijun Yang ◽

Harris Wu ◽

Bingbing Jiang

Keyword(s):

Machine Learning ◽

Pattern Recognition ◽

Investment Strategy ◽

Training Data ◽

Stock Trading ◽

Learning Methods ◽

Data Set ◽

Empirical Results ◽

Machine Learning Methods ◽

Trading Decisions

PRML, a novel candlestick pattern recognition model using machine learning methods, is proposed to improve stock trading decisions. Four popular machine learning methods and 11 different features types are applied to all possible combinations of daily patterns to start the pattern recognition schedule. Different time windows from one to ten days are used to detect the prediction effect at different periods. An investment strategy is constructed according to the identified candlestick patterns and suitable time window. We deploy PRML for the forecast of all Chinese market stocks from Jan 1, 2000 until Oct 30, 2020. Among them, the data from Jan 1, 2000 to Dec 31, 2014 is used as the training data set, and the data set from Jan 1, 2015 to Oct 30, 2020 is used to verify the forecasting effect. Empirical results show that the two-day candlestick patterns after filtering have the best prediction effect when forecasting one day ahead; these patterns obtain an average annual return, an annual Sharpe ratio, and an information ratio as high as 36.73%, 0.81, and 2.37, respectively. After screening, three-day candlestick patterns also present a beneficial effect when forecasting one day ahead in that these patterns show stable characteristics. Two other popular machine learning methods, multilayer perceptron network and long short-term memory neural networks, are applied to the pattern recognition framework to evaluate the dependency of the prediction model. A transaction cost of 0.2% is considered on the two-day patterns predicting one day ahead, thus confirming the profitability. Empirical results show that applying different machine learning methods to two-day and three-day patterns for one-day-ahead forecasts can be profitable.

Download Full-text

Benchmarking machine learning methods for comprehensive chemical fingerprinting and pattern recognition

Journal of Chromatography A ◽

10.1016/j.chroma.2019.02.027 ◽

2019 ◽

Vol 1595 ◽

pp. 158-167 ◽

Cited By ~ 13

Author(s):

Stephen E. Reichenbach ◽

Claudia A. Zini ◽

Karine P. Nicolli ◽

Juliane E. Welke ◽

Chiara Cordero ◽

...

Keyword(s):

Machine Learning ◽

Pattern Recognition ◽

Learning Methods ◽

Machine Learning Methods ◽

Chemical Fingerprinting

Download Full-text

Potential for Vertical Heterogeneity Prediction in Reservoir Basing on Machine Learning Methods

Geofluids ◽

10.1155/2020/3713525 ◽

2020 ◽

Vol 2020 ◽

pp. 1-12 ◽

Cited By ~ 1

Author(s):

Hongqing Song ◽

Shuyi Du ◽

Ruifei Wang ◽

Jiulong Wang ◽

Yuhe Wang ◽

...

Keyword(s):

Machine Learning ◽

Neural Networks ◽

Deep Neural Networks ◽

Rapid Development ◽

Petroleum Industry ◽

Production Data ◽

Learning Methods ◽

Machine Learning Methods ◽

Vertical Heterogeneity ◽

Dynamic Production

With the rapid development of computer technology, some machine learning methods have begun to gradually integrate into the petroleum industry and have achieved some achievements, whether in conventional or unconventional reservoirs. This paper presents an alternative method to predict vertical heterogeneity of the reservoir utilizing various deep neural networks basing on dynamic production data. A numerical simulation technique was adopted to obtain the required dataset, which contains dynamic production data calculated under different heterogeneous reservoir conditions. Machine learning models were established through deep neural networks, which learn and capture the characteristics better between dynamic production data and reservoir heterogeneity, so as to invert the vertical permeability. On the basis of model validation, the results show that machine learning methods have excellent performance in predicting heterogeneity with the RMSE of 12.71 mD, which effectively estimated the permeability of the entire reservoir. Moreover, the overall AARD of the predictive result obtained by the CNN method was controlled at 11.51%, revealing the highest accuracy compared with BP and LSTM neural networks. And the permeability contrast, an important parameter to characterize heterogeneity, can be predicted precisely as well, with a derivation of below 10%. This study proposed a potential for vertical heterogeneity prediction in reservoir basing on machine learning methods.

Download Full-text

SUPPORT VECTOR MACHINES UNTUK MENYELESAIKAN MASALAH KLASIFIKASI PADA PENGENALAN POLA

Jurnal Poli-Teknologi ◽

10.32722/pt.v18i2.1432 ◽

2019 ◽

Vol 18 (2) ◽

Author(s):

Abdul Azis Abdillah

Keyword(s):

Machine Learning ◽

Pattern Recognition ◽

Support Vector Machines ◽

Support Vector ◽

Classification Problems ◽

Learning Methods ◽

Learning Pattern ◽

Machine Learning Methods ◽

Vector Machines ◽

Learning Machine

ABSTRACTSupport Vector Machines (SVM) are known as the latest machine learning (machine learning) methods to solve classification problems in pattern recognition. This paper discusses the use of SVM in solving problems in pattern recognition. An example of the problem given in this paper contains a collection of data on Any Linearly Separable Datase, Any dataset with Noise, and Real datasets.Key words: machine learning, pattern recognition, SVMABSTRAKSupport Vector Machines (SVM) dikenal sebagai metode machine learning (pembelajaran mesin) paling mutakhir untuk menyelesaikan masalah klasifikasi pada pengenalan pola. Tulisan ini bertujuan untuk membahas penggunaan SVM dalam memecahkan masalah klasifikasi pada pengenalan pola. Contoh masalah yang diberikan pada tulisan ini meliputi klasifikasi data pada Sembarang Linearly Separable Dataset, Sembarang Dataset dengan Noise, dan Real dataset.Kata kunci : klasifikasi, pengenalan pola, SVM

Download Full-text