Table of Content

10 December 2020, Volume 42 Issue 6

Previous Issue    Next Issue

A Special Issue on Artificial Intelligence Technology & Application in Oil and Gas Fields

A Review of the Application Progress of Machine Learning in Oil and Gas Industry

MIN Chao, DAI Boren, ZHANG Xinhui, DU Jianping

2020, 42(6):  1-15.  DOI: 10.11885/j.issn.1674-5086.2020.06.05.03

Asbtract ( 852 )   HTML ( 72)   PDF (1129KB) ( 1359 )  

References | Related Articles | Metrics

With the rise of deep learning in recent years, machine learning has been further developed in the oil and gas field. However, due to the particularity and complexity of the oil and gas industry, there is no training sample base suitable for deep learning, nor a targeted model establishment and selection method system. In addition, the uninterpretability of methods such as deep learning leads to the high dependence of learning models on the environment, which restricts the popularization and application of machine learning in the oil and gas industry. Starting from the development stage of machine learning, this paper introduces the major breakthroughs and problems in the application of machine learning in various fields of oil and gas industry. Then, suggestions are given on the processing methods and sample building of different types of data in the oil and gas industry, and how to carry out model adaptability analysis, etc. Finally, the development potential and research direction of machine learning in oil and gas artificial intelligence are proposed.

Seismic Inversion Experiments Based on Deep Learning Algorithm Using Different Datasets

HUANG Xuri, DAI Yue, XU Yungui, TANG Jing

2020, 42(6):  16-25.  DOI: 10.11885/j.issn.1674-5086.2020.06.02.02

Asbtract ( 402 )   HTML ( 36)   PDF (1082KB) ( 286 )  

References | Related Articles | Metrics

Recently deep learning of artificial intelligence demonstrates certain advantages for seismic processing, interpretation and inversion. Previous studies show that the combination of deep learning and seismic inversion could generate more robust results than traditional methods. The deep learning technique could achieve results of high resolution which is critical for reservoir development. This paper investigates the effects of different training datasets used in seismic inversion based on Convolutional Neural Network (CNN) by designing a reservoir model and its corresponding seismic response. The result shows that the prediction accuracy of this Neural Network increases with the increase of training datasets size in a certain range. The relationship of inversion quality and ratio between the entire datasets and the training data is demonstrated. In addition, different levels of noises in seismic are tested for CNN training. The results demonstrate the generalization and anti-noise ability of the designed CNN.

The Relationship Between Topography and Distribution of Slope in T₂³ Top Surface of Gaoyou Sag

TANG Jun, DUAN Yuying, DUAN Hongliang, LIU Dan, ZHU Fengfan

2020, 42(6):  26-34.  DOI: 10.11885/j.issn.1674-5086.2020.05.22.02

Asbtract ( 149 )   HTML ( 0)   PDF (1062KB) ( 147 )  

References | Related Articles | Metrics

At present, the main GIS technique for geological structure analysis in oil field is three-dimensional modeling, but few methods of slope analysis for geological structure analysis. Two slope statistical indexes are extracted by means of mean variable point analysis and area elevation integral:Hypsometric Integral (HI) and depth range. The 17×17 (pixel) window is used to make statistics and mapping, and the relationship between spatial distribution and topography is analyzed. The results of research show that the depth variation amplitude of T₂³ top surface increases gradually from the northwest to the southeast. The biggest value of depth variation is 700 m in the northwest and 2 198 m in the southeast, which indicates the topography of T₂³ top surface in southeast is more complex than that in northwest. The mean of HI is 0.5, indicating the top surface of T₂³ top surface is in the prime of the development of groundwater erosion topography. The spatial distribution of HI indicates the southeast with larger depth and more complex terrain is basically in the aging stage of erosion terrain development. The transitional zone of depth and terrain complexity from high to low is in the prime. Areas of less depth are in their infancy and are characterized by strong drainage branching.

Data-driven Drilling Acceleration in Bohai XX Block

LIU Zhaonian, ZHAO Ying, SUN Ting

2020, 42(6):  35-41.  DOI: 10.11885/j.issn.1674-5086.2020.06.09.02

Asbtract ( 231 )   HTML ( 6)   PDF (1181KB) ( 143 )  

References | Related Articles | Metrics

With the increase of drilling operating depth, formation conditions and well structure become complicated, resulting in increasement in drilling investment. In order to improve the drilling efficiency and reduce the drilling cost, this paper used the logging data obtained in the drilling process, combined with the neural network and GA, and finds the optimal ROP and its corresponding drilling parameters, and ensured the efficient drilling operations. Collect 8 000 sets of data for 2 stratigraphic sections of different wells in a block in the Bohai Area (4 000 sets per layer), and train the machine learning model separately for each stratum. Taking one of the layers as an example, 3 900 sets of drilling parameters are taken as inputs, and the corresponding 3 900 sets of ROP are used as output to train the BP neural network; then the remaining 100 sets of drilling parameters are taken as input data. The obtained neural network is used to predict 100 sets of ROP. Finally, 4 000 sets of drilling parameters are used as the population individuals in the genetic algorithm, and the predicted ROP is used as one important parameter in the GA-the individual fitness value, the optimal ROP and its corresponding drilling parameters are derived by GA. The method makes full use of the data of the oilfield site, and obtains machine learning models suitable for different strata in the Bohai Area, which improves the ROP and realizes the drilling speed.

Application of Cross Correlation Function Method in Locating Perforation Depth

LI Jin, LIU Yuhai, ZHANG Jian, WANG Jiang, ZHANG Yiling

2020, 42(6):  42-48.  DOI: 10.11885/j.issn.1674-5086.2020.06.24.01

Asbtract ( 231 )   HTML ( 5)   PDF (510KB) ( 111 )  

References | Related Articles | Metrics

Perforation depth correction technology for oil and gas wells is a prerequisite to ensure accurate perforation operations for oil and gas well completion. Perforating operations use the natural radioactive gamma correction method of oil and gas well reservoirs and the traditional depth correction method of tubing and casing joints. With the prolonged oil and gas production time, the downhole gamma radioactivity is weakened and the tubing and casing joints are damaged. The two signals of the depth calibration measurement are often obscured by downhole noise, and the effective signal characteristics are not obvious. Directly relying on the traditional amplitude identification method will bring about large perforation accuracy errors and even cause accidental perforation. It cannot meet the demand for precise deep perforation of thin reservoirs. Based on the calculation principle of the signal cross-correlation function method, the standard hoop signal and gamma signal are cross-correlated with the corresponding signals collected by the perforation depth calibration to suppress background noise, highlight the effective signal, and achieve the purpose of accurate perforation depth calibration. Field application shows that the method is simple in depth calibration technology, can meet the purpose of real-time calibration and perforation completion in one field operation, and has certain practical application value.

Establishment of a New Method for Viscosity Calculation of Crude Oil by Data Analysis

ZHANG Lijun, WANG Shuai, ZHENG Wei

2020, 42(6):  49-55.  DOI: 10.11885/j.issn.1674-5086.2020.06.08.02

Asbtract ( 184 )   HTML ( 3)   PDF (492KB) ( 82 )  

References | Related Articles | Metrics

Aiming at the problem that there is little or no sampling of formation fluid in heavy oil fields in the stage of offshore exploration and evaluation, through the statistics of a large number of sampled heavy oil fluid in bohai sea area, data matrix analysis was used to explore the main control factors affecting formation crude oil viscosity and the relationship between different parameters and formation crude oil viscosity. The empirical formula of formation crude oil viscosity and other parameters is established by nonlinear regression, and the value chart of formation crude oil viscosity and surface density under different parameters is also established. The mean error between the viscosity of formation crude oil calculated by the new empirical formula and the measured viscosity of formation crude oil is 5.6%; the empirical formula and chart are applied to quality control of the fluid properties of a certain offshore heavy oil, and the results of the two methods are equivalent, indicating the reliability of the methods. The viscosity formula and chart of formation crude oil based on data driven can solve the problem of little or no sampling of formation fluid in offshore heavy oil field, and provide important reference for the oilfield development plan, development mode selection, recovery rate prediction, platform facilities design and crude oil transportation.

A Study on the Optimization of Fracturing Operation Parameters Based on PCA-BNN

TAN Chaodong, HE Jiayuan, ZHOU Tong, LIU Jiankang, SONG Wenrong

2020, 42(6):  56-62.  DOI: 10.11885/j.issn.1674-5086.2020.05.12.05

Asbtract ( 292 )   HTML ( 7)   PDF (448KB) ( 303 )  

References | Related Articles | Metrics

Domestic and foreign scholars have carried out the research of shale gas fracturing production prediction and fracturing parameter optimization based on machine learning on the premise of a large number of foreign shale fracturing sample data. With the continuous development of F Gas Field in recent years, a large number of data from fracturing operation, production dynamic and interpretation results have been accumulated. In view of the fact that these data are not fully utilized in the design of fracturing operation parameters at present, the Bayesian neural network model is established to optimize the fracturing operation parameters by using history data of the fracturing operation parameters and reservoir physical parameters from 200 wells. The reservoir physical parameters, completion parameters and fracturing operation parameters which have an impact on the fracturing effect are selected. The correlation of 11 parameters is analyzed by Pearson correlation coefficient method. Principal Component Analysis (PCA) is used for further dimension reduction. The principal components are used as input parameters of Bayesian Neural Network model. The validity period is used as output parameter. Bayesian method is introduced to adjust regularization coefficient adaptively to avoid neural network overfitting. And then, a three-layer Bayesian neural network prediction model is generated. The model is trained by using 90% of the data of 200 wells as training set and 10% as test set. The experimental results show that the mean relative error of the model prediction results after training is within 5%, which can be used to optimize the fracturing operation parameters.

Knowledge Mining for Oilfield Development Index Prediction Model Using Deep Learning

ZHONG Yihua, WANG Shuning, LUO Lan, YANG Jinlian, YUE Yongpeng

2020, 42(6):  63-74.  DOI: 10.11885/j.issn.1674-5086.2020.05.11.02

Asbtract ( 236 )   HTML ( 5)   PDF (618KB) ( 211 )  

References | Related Articles | Metrics

The changing characteristics of oilfield development index are regarded as the important basis of oilfield development planning, oilfield exploitation evaluation, oilfield development scheme design and adjustment, decision management problems of oilfield development risk prediction and early warning, etc. For one of the unsolved bottleneck problem of building intelligent oilfield, i.e. the problem on knowledge mining of selecting prediction method and model of oilfield development indexes intelligent prediction system, based on the massive data of oilfield development, this paper uses the convolutional neural network and cyclic neural network of deep learning to extract the characteristics and knowledge reflecting the development dynamic of oilfield. On this basis, combining the model base and knowledge base of oilfield development index prediction, a knowledge mining method to select the optimal prediction model of oilfield development index is proposed through the input information and dynamic characteristics index of oilfield development, the model base and the knowledge base of oilfield development index prediction by using the joint extraction method of entity and relationship of deep learning. The simulation example of conceptual design shows that the proposed knowledge mining process may realize autonomous obtaining an appropriate prediction model of oilfield development index as long as inputting relevant information of oilfield development.

Research and Application of Oil Multi-peak Model Based on Machine Learning

HUANG Cheng, PAN Wenjin

2020, 42(6):  75-81.  DOI: 10.11885/j.issn.1674-5086.2020.05.13.01

Asbtract ( 279 )   HTML ( 2)   PDF (385KB) ( 154 )  

References | Related Articles | Metrics

In the actual process of oilfield development, affected by the production of new blocks, the adjustment of development plans and the "t three mining" measures, the annual output data will show multi-peak form. As the classical Hubbert, HCZ and other models cannot directly fit the multi-peak data sequence, the multi-peak prediction model of oilfield production based on machine learning is studied. Based on the Hubbert model, the piecewise least squares fitting is performed for multi-peak data sequence, the penalty term controlling the number of segments is introduced into the fitting error function. Using dynamic programming algorithm, the multi-peak Hubbert prediction model for the optimal segment is automatically obtained. The model is applied to the actual oilfield production data, and the prediction results achieve the expected purpose. This paper presents a method to build a multi-peak prediction model of oilfield production through automatic optimal segmental linear regression learning. In practical application, it has the advantages of simple modeling and strong adaptability.

A Study on Oil Well Production Prediction Based on Time Series Dynamic Analysis

YANG Yang, CHENG Yuefei, QIAO Ying, LIU Jiong

2020, 42(6):  82-88.  DOI: 10.11885/j.issn.1674-5086.2020.05.22.04

Asbtract ( 383 )   HTML ( 9)   PDF (598KB) ( 276 )  

References | Related Articles | Metrics

The effect of oil well production prediction method currently used is not ideal. This study aims at dynamic prediction of oil production, using time series analysis combined with residual error correction. We build an ARIMA (Autoregressive integrated moving average) model with the ability of time series dynamic analysis, to predict the initial value and the real residual oil well production; the residual error was corrected by constructing the (SVM) Support Vector Machine time series prediction model to obtain the predicted value of oil well production combination. The LSTM (Long Short-Term Memory) model is compared with the above methods. The experimental results show that the average relative error rates of the combined prediction model and the LSTM model are 9.81% and 32.44% respectively. The conclusion is that the combined model prediction is more accurate, and provides an effective method for the dynamic prediction of oil well production, which can be used as a fast and real-time auxiliary basis for oil well production planning and has practical value.

Prediction of Annual Increase of Oil Production Based on GM (1, 1)Neural Network Combined Optimization

LIU Haohan, YAN Yongqin, MIN Lingyuan, YUE Ping, YIN Yanling

2020, 42(6):  89-96.  DOI: 10.11885/j.issn.1674-5086.2020.06.05.01

Asbtract ( 172 )   HTML ( 0)   PDF (461KB) ( 89 )  

References | Related Articles | Metrics

Increasing oil production of old wells has become an inevitable choice to stabilize production and reduce development costs of oilfield block development. In view of the limitation of polynomial regression prediction, the fact that the grey theory cannot reflect the characteristics of influence factors, and the neural network needs more data and is less sensitive to data, this paper establishes an optimal control model, combining the high precision forecasting of grey theory with the neural network. Taking the actual measures to increase oil production in an oilfield block from 2011 to 2018 as an example, by confirming the influence factors of annual oil increment, a new optimal control grey neural network model is established, which is used to predict the annual oil increment with different measures. Compared with polynomial regression prediction, GM(1, 1) prediction and BP neural network prediction, the results show that the new model has better simulation effect and higher prediction precision. The prediction accuracy of the annual oil increment with the new method is 97.34% in 2018. The grey neural network model based on optimal control can be an artificial intelligence model to predict the annual oil increment with different measures, which provides a new idea for accurately predicting of oil increment with different measures and decision-making of oilfield development.

Maintenance Decision of High-risk Equipment of Offshore Platform in Integrity Management

TANG Yang, YANG Xin, JING Jiajia, ZHANG Zhidong, YANG Yan

2020, 42(6):  97-106.  DOI: 10.11885/j.issn.1674-5086.2020.05.23.01

Asbtract ( 226 )   HTML ( 7)   PDF (660KB) ( 91 )  

References | Related Articles | Metrics

In the maintenance management of offshore platform equipment, the importance and classification of high-risk equipment are unclear, and the maintenance methods are unreasonable, resulting in equipment "maintenance excess" or "maintenance lack", high maintenance costs, and even leading to major accidents and economic losses. With comprehensive consideration of the characteristics of the offshore platform equipment, failure mode and impact and management requirements, etc., determine its 10 important factors, formulate 10 important evaluation indicators and scoring standards, and achieve qualitative description and quantitative conversion. AHP and Monte Carlo simulation methods are used to propose an evaluation method for the importance of offshore platform equipment, which quantitatively evaluates the importance of equipment and effectively eliminates the subjective impact of scoring. Based on the percentage of the cumulative frequency of importance, the equipment is divided into three categories, and the maintenance logic decision diagram is established by classification, which realizes the equipment maintenance decision based on the importance. Taking the well control system as an example, the importance of each equipment was quantitatively evaluated and scientifically classified, and then their respective maintenance methods were determined, which effectively verified the feasibility of the method. The research can provide technical support for oilfield enterprises to establish a scientific and complete equipment integrity management system.

Fault Diagnosis of Electric Submersible Pump Based on Principal Component Analysis

SUI Xianfu, PENG Long, HAN Guoqing, FAN Baitao, YU Jifei

2020, 42(6):  107-114.  DOI: 10.11885/j.issn.1674-5086.2020.06.10.01

Asbtract ( 261 )   HTML ( 5)   PDF (1105KB) ( 214 )  

References | Related Articles | Metrics

Electric Submersible Pump (ESP) is currently widely employed to help enhance production for non-flowing well with high production and high water cut well. However, ESP failures are common in the oil industry, and these failures lead to production disruptions, resulting in significant economic loses. The purpose of this paper is to evaluate Principal Component Analysis (PCA) as an unsupervised machine learning technique to detect the cause of ESP failures and estimate the remaining life of ESP. The data provided by the ESP system are usually closely correlated; principal component analysis utilizes these data to extract eigenvalues and create new space, then reevaluate ESP system with multiple new principal components. Hortlin square statistical algorithm and the square error analysis algorithm are used to establish a PCA diagnostic model. This model was successfully applied in the Bohai oilfield to diagnose the ESP performance real-time. The most responsible decision variable for the potential ESP failures are determined according to the order of contribution. Also, the PCA diagnostic model was able to determine the time at which the ESP failures occurs. This paper demonstrates that the application of PCA method preforms well in monitoring ESP operations and predicts the impending ESP failures.

Granular Computing for Pumping Well Parameter Optimization

ZHANG Hengru, ZHU Kelin, XU Yuanyuan, QIAO Ying

2020, 42(6):  115-123.  DOI: 10.11885/j.issn.1674-5086.2020.05.29.02

Asbtract ( 172 )   HTML ( 6)   PDF (529KB) ( 143 )  

References | Related Articles | Metrics

Aiming at the problem of optimization of pumping unit parameters in oil and gas production, a granular computing method is proposed based on the data of pumping unit production control and maintenance measures. In this paper, machine learning methods such as granular computing, cost-sensitive rough sets and recommendation systems are used. Firstly, a decision tree is employed to build a multi-granular fusion model based on the time, space, and business level of the pumping well data. Secondly, a suitable evaluation model is defined for the pumping well business with the cost-sensitive rough set. Finally, a recommended model of core parameters and maintenance measures is designed based on field-aware factorization machine under cost constraints. Based on real oil-gas production data, we have designed comparative experiments with different granularities. We adjusted the production parameters of the pumping well from coarse granule to fine granule, and found that the recommendation accuracy for optimizing the core parameters of production is gradually increased first and then gradually decreased. We can conclude that in parameter optimization, appropriate granularity selection is required.

Active Learning Method for Abnormal Operating Conditions of Natural Gas Gathering System

FANG Yu, CAO Xuemei, LI Binqian, MIN Fan, QIAO Ying

2020, 42(6):  124-132.  DOI: 10.11885/j.issn.1674-5086.2020.05.12.08

Asbtract ( 237 )   HTML ( 6)   PDF (1678KB) ( 144 )  

References | Related Articles | Metrics

Various abnormal operating conditions in natural gas gathering system pose a threat to safe production. This paper proposes an intelligent processing system model for abnormal operating conditions. The abnormal operating conditions classification prediction module of the model adopts the active learning method, which can classify the abnormal type in real time and accurately, and provide a basis for the system to recommend appropriate processing schemes to experts. Firstly, use the SCADA system to monitor data in real time and perform abnormal conditions warning. Secondly, we use the active learning algorithm to classify the early warning of abnormal operating conditions. The classification results provide support for constructing the abnormal working condition inference engine, and then implement intelligent decision-making assistance. The experimental results show that the proposed method can save the cost of experts, identify the types of abnormal operating conditions, and propose a reasonable solution.

Study and Application of Multi-parameter Early Warning Model for Gas Wells

WANG Hao, ZHANG Lifu, LUO Hao

2020, 42(6):  133-140.  DOI: 10.11885/j.issn.1674-5086.2020.06.12.01

Asbtract ( 188 )   HTML ( 3)   PDF (989KB) ( 104 )  

References | Related Articles | Metrics

SF Gas Field started the construction of digital field in 2016, completed the information collection and deployment of stations and gas wells, and realized real-time data upload and station visualization. However, in data application, the effective alarm rate relying on the fixed threshold alarm mode is low, which cannot automatically prompt abnormal working conditions. It needs manual auxiliary judgment, which takes a long time and has low accuracy. In order to realize the intelligent analysis of abnormal data and automatic alarm grading, and improve the work efficiency and production efficiency under the condition of informatization, the intelligent improvement plan was launched in 2018, the main production parameters of the gas well are calculated by means of a custom statistical method, and the corresponding algorithm is formed to judge whether there is any abnormal situation according to the calculation results. By combining the multi-parameter warning information, the multiparameter joint warning model is formed, and the working condition experience database is matched. According to the preset value, the abnormal situation and disposal opinions are pushed to realize the joint warning. This new management method of informationized gas field ensures the timeliness of abnormal diagnosis and production disposal in gas wells and well stations, and comprehensively improves the efficiency of abnormal management.

Design of Data Acquisition and Transmission Scheme for Remote and Inefficient Gas Wells Based on LPWAN Technology

ZHAO Yong, TANG Jia, HAN Tao, WU Yuxin

2020, 42(6):  141-148.  DOI: 10.11885/j.issn.1674-5086.2020.06.10.03

Asbtract ( 164 )   HTML ( 3)   PDF (1261KB) ( 74 )  

References | Related Articles | Metrics

To achieve the purpose of digital transformation and intelligent development of oil and gas enterprises under the background of international low oil price, improving production efficiently is an important strategy for the sustainable development of these enterprises. The acquisition and transmission of gas well production data is an effective method to early warning dynamic analysis and reduce labor intensity. The traditional scheme of data acquisition and transmission is of high cost, long construction period and high equipment power consumption, which is not a good for remote and inefficient gas wells. In this paper, through the research of Low Power WAN (LPWAN) technology, we summarized the technical advantages, and by combining with the production characteristics of remote and inefficient gas wells, the environment and information requirements, we accomplished an economic and reliable data transmission scheme based on Lora and NB-IoT heterogeneous networking. The scheme uses sensors to collect production data of gas well and transmit it to acquisition and transmission module through LoRa technology. After analysis, it is transferred to cloud server through NB-IoT technology. The control center can realize remote monitoring by visiting cloud server. According to the test and analysis of the scheme in this paper, the results show that the scheme can realize the data acquisition and transmission of remote and inefficient gas wells with high efficiency, low cost and low power consumption, which provides an effective supplement to the traditional information construction scheme.

Application of LoRa Technology in Oilfield Data Acquisition System

MEI Dacheng, CHEN Yaping, HE Jingqi, YUE Lin

2020, 42(6):  149-156.  DOI: 10.11885/j.issn.1674-5086.2020.05.25.03

Asbtract ( 210 )   HTML ( 3)   PDF (596KB) ( 145 )  

References | Related Articles | Metrics

In view of the complex networking and ultra-high power consumption of existing network communication technologies in complex petroleum environments, LoRa, a low-power, ultra-long-range wireless communication technology, is used to improve the accuracy and real-time performance of existing network communication technologies and reduce the difficulty of networking in the intelligent oilfield data acquisition systems. According to the research of LoRaWAN wireless standard protocol and polling time slot allocation algorithm, a wireless data acquisition method based on LoRa technology for differential time slot allocation strategy was designed. The improved method was simulated on OPNET 14.5 platform. Through simulation, the throughput of the ad hoc network using this method is significantly improved compared with the fixed time slot allocation strategy, which effectively reduces the data reporting period of important data nodes, increases the utilization of network channels, and improve digital construction of oilfields.

Question Answering System for Drilling Safety Based on Tri-BiLSTM-CNN

WANG Bing, ZHENG Yamei, CHEN Maoke, GAO Lingyun

2020, 42(6):  157-164.  DOI: 10.11885/j.issn.1674-5086.2020.05.12.06

Asbtract ( 200 )   HTML ( 1)   PDF (450KB) ( 164 )  

References | Related Articles | Metrics

The FAQ question answering system in a specific field usually has the following three problems:(1) how to effectively represent sentences semantically; (2) how to effectively match sentences semantically; (3) how to segment domain words. To solve the above three problems, a deep learning model based on Triplet BiLSTM-CNN is proposed. Firstly, the bidirectional long-term memory network and convolutional neural network are combined to construct the network model, which makes full use of the advantages of BiSLTM in processing the serialized data and the advantages of CNN in capturing local features. Then, the Triplet parallel structure is used to match sentences. Finally, character vector is used instead of word vector to avoid the influence of segmentation error on the model. The experimental results on real data sets in the field of drilling safety show that Triplet BiLSTM-CNN model can better vectorize sentence semantics and significantly improve the accuracy of sentence similarity calculation, and the effect is significantly better than that of CNN and LSTM. The model is applied to the FAQ question answering system in the field of drilling safety, which can effectively reduce the labor cost, and is of great significance and application value to improve the efficiency and quality of drilling work.

Research of Extraction on Petroleum Unstructured Information Based on Named Entity Recognition

ZHONG Yuan, LIU Xiaorong, WANG Jie, CHEN Yan, ZHANG Tai

2020, 42(6):  165-173.  DOI: 10.11885/j.issn.1674-5086.2020.05.12.01

Asbtract ( 227 )   HTML ( 6)   PDF (643KB) ( 152 )  

References | Related Articles | Metrics

With the acceleration of the construction of "intelligent oilfield", it is of great significance to build an intelligent analysis system for mass oil data. However, as a result of the dynamic text data generated in oilfield production process is often of unstructured and various types. Extracting the crucial information for analysis becomes a popular area of research, and information extraction needs high-quality entities to support. In this paper, we propose an unstructured text information extraction method based on NER (Named Entity Recognition) according to the particular problem. Feature extraction of oil corpus is carried out by Bidirectional Long Short-Term Memory (Bi-LSTM) network model, and combines Conditional Random Field (CRF) as classifier. Bi-LSTM+CRF method is used to construct a high-precision NER model to extract named entities from unstructured texts in petroleum industry. The experimental results on the text data set of well workover treatment show that this method has a higher precision and recall rate than other state-of-art methods.

Research on MDR Concept System Model and Ontology Representation Standardization

YUAN Jingshu, LI Hongqi

2020, 42(6):  174-180.  DOI: 10.11885/j.issn.1674-5086.2020.05.22.03

Asbtract ( 219 )   HTML ( 0)   PDF (1464KB) ( 94 )  

References | Related Articles | Metrics

With the in-depth application of big data and artificial intelligence in various fields, the standardization of knowledge representation has been mentioned as an important agenda. In order to realize the standardization of knowledge representation, this paper introduces the concept of "Three Worlds" model, that is, the real world, the conceptual world and the computer world. According to this model, things in real world can be abstracted concept in concept system model in concept world. Finally, acoording to concrete standards, the concept system model can be tanslated into information model in computer world, so as to realize knowledge representation. In order to standardize the representation of conceptual system model, series standards of ISO/IEC 11179 MDR is studied, the concept system registration meta-model specification is analyzed, and a seven tuple model of MDR concept system is constructed. In addition, the representation method of the relationship between concepts is defined, which lays the foundation for the representation of conceptual system model. In order to transform the concept system model registered in MDR into OWL ontology representation model in the computer world, the rules of mapping MDR concept system to OWL ontology is defined in the paper. Finally, the concept system registration prototype system was designed and developed to verify the feasibility of the MDR concept system and ontology representation. The research results provide a methodology for the standardization of domain knowledge representation.

Research on Improvement of Cloud Virtual Machine Memory Migration Algorithm for Oil and Gas Exploration and Development

QIAO Ying, WU Yijia, YANG Xuhua

2020, 42(6):  181-186.  DOI: 10.11885/j.issn.1674-5086.2020.05.28.02

Asbtract ( 171 )   HTML ( 4)   PDF (490KB) ( 66 )  

References | Related Articles | Metrics

The dynamic migration of virtual machines is of great significance to the cloud platform security construction of digital oil fields. The paper discusses the virtual machine dynamic migration mechanism in the key technology of cloud computingvirtualization technology. Aiming at the problem of too long downtime of the pre-copy migration method, based on the pre-copy method, this paper proposes an identification-based dynamic migration method, which can effectively reduce the migration time and improve the migration efficiency. First, a counter is added to record the number of dirty pages to determine whether it is frequently changed. Then, a window value is set. When the number of frequently changed dirty pages reaches the window value, all dirty pages are sent to the target server. This method avoids the problem of excessive virtual machine downtime caused by too many pages during the last transfer. Experimental results show that this method is better than the pre-copy method in migration time and data transfer volume, and improves the efficiency of virtual machine migration.

A Quantitative Study of International Petroleum Trade Relations Based on Complex Networks

ZHANG Songlin, LIU Haohan, YIN Hu, LI Fen

2020, 42(6):  187-196.  DOI: 10.11885/j.issn.1674-5086.2020.06.05.02

Asbtract ( 162 )   HTML ( 5)   PDF (1449KB) ( 98 )  

References | Related Articles | Metrics

Aiming at the problem of how to evaluate the international petroleum trade relationship more accurately, a network and structural characteristic index quantitative analysis of the relationship is carried out. In this study, the complex network theory was used for network modeling, the structural balance theory was used for quantitative analysis, and the key indexes such as calculated clustering coefficient and degree of balance were used for annual benchmark comparison. The conclusion is that with the increase of the number of traders, especially the number of importing countries, the clustering coefficient of the network becomes larger, the balance degree value increases, the balance of the entire trade network becomes worse, and the trade relationship becomes more tense. This research adopts network and structured way to evaluate the relationship, which is a beneficial attempt to advance the research of oil trade relationship from qualitative to quantitative.