With the development of information technology,a large number of product quality data in the entire manufacturing process is accumulated,but it is not explored and used effectively.The traditional product quality pred...With the development of information technology,a large number of product quality data in the entire manufacturing process is accumulated,but it is not explored and used effectively.The traditional product quality prediction models have many disadvantages,such as high complexity and low accuracy.To overcome the above problems,we propose an optimized data equalization method to pre-process dataset and design a simple but effective product quality prediction model:radial basis function model optimized by the firefly algorithm with Levy flight mechanism(RBFFALM).First,the new data equalization method is introduced to pre-process the dataset,which reduces the dimension of the data,removes redundant features,and improves the data distribution.Then the RBFFALFM is used to predict product quality.Comprehensive expe riments conducted on real-world product quality datasets validate that the new model RBFFALFM combining with the new data pre-processing method outperforms other previous me thods on predicting product quality.展开更多
In the early time of oilfield development, insufficient production data and unclear understanding of oil production presented a challenge to reservoir engineers in devising effective development plans. To address this...In the early time of oilfield development, insufficient production data and unclear understanding of oil production presented a challenge to reservoir engineers in devising effective development plans. To address this challenge, this study proposes a method using data mining technology to search for similar oil fields and predict well productivity. A query system of 135 analogy parameters is established based on geological and reservoir engineering research, and the weight values of these parameters are calculated using a data algorithm to establish an analogy system. The fuzzy matter-element algorithm is then used to calculate the similarity between oil fields, with fields having similarity greater than 70% identified as similar oil fields. Using similar oil fields as sample data, 8 important factors affecting well productivity are identified using the Pearson coefficient and mean decrease impurity(MDI) method. To establish productivity prediction models, linear regression(LR), random forest regression(RF), support vector regression(SVR), backpropagation(BP), extreme gradient boosting(XGBoost), and light gradient boosting machine(Light GBM) algorithms are used. Their performance is evaluated using the coefficient of determination(R^(2)), explained variance score(EV), mean squared error(MSE), and mean absolute error(MAE) metrics. The Light GBM model is selected to predict the productivity of 30 wells in the PL field with an average error of only 6.31%, which significantly improves the accuracy of the productivity prediction and meets the application requirements in the field. Finally, a software platform integrating data query,oil field analogy, productivity prediction, and knowledge base is established to identify patterns in massive reservoir development data and provide valuable technical references for new reservoir development.展开更多
Considering the phase behaviors in condensate gas reservoirs and the oil-gas two-phase linear flow and boundary-dominated flow in the reservoir,a method for predicting the relationship between oil saturation and press...Considering the phase behaviors in condensate gas reservoirs and the oil-gas two-phase linear flow and boundary-dominated flow in the reservoir,a method for predicting the relationship between oil saturation and pressure in the full-path of tight condensate gas well is proposed,and a model for predicting the transient production from tight condensate gas wells with multiphase flow is established.The research indicates that the relationship curve between condensate oil saturation and pressure is crucial for calculating the pseudo-pressure.In the early stage of production or in areas far from the wellbore with high reservoir pressure,the condensate oil saturation can be calculated using early-stage production dynamic data through material balance models.In the late stage of production or in areas close to the wellbore with low reservoir pressure,the condensate oil saturation can be calculated using the data of constant composition expansion test.In the middle stages of production or when reservoir pressure is at an intermediate level,the data obtained from the previous two stages can be interpolated to form a complete full-path relationship curve between oil saturation and pressure.Through simulation and field application,the new method is verified to be reliable and practical.It can be applied for prediction of middle-stage and late-stage production of tight condensate gas wells and assessment of single-well recoverable reserves.展开更多
Reservoir identification and production prediction are two of the most important tasks in petroleum exploration and development.Machine learning(ML)methods are used for petroleum-related studies,but have not been appl...Reservoir identification and production prediction are two of the most important tasks in petroleum exploration and development.Machine learning(ML)methods are used for petroleum-related studies,but have not been applied to reservoir identification and production prediction based on reservoir identification.Production forecasting studies are typically based on overall reservoir thickness and lack accuracy when reservoirs contain a water or dry layer without oil production.In this paper,a systematic ML method was developed using classification models for reservoir identification,and regression models for production prediction.The production models are based on the reservoir identification results.To realize the reservoir identification,seven optimized ML methods were used:four typical single ML methods and three ensemble ML methods.These methods classify the reservoir into five types of layers:water,dry and three levels of oil(I oil layer,II oil layer,III oil layer).The validation and test results of these seven optimized ML methods suggest the three ensemble methods perform better than the four single ML methods in reservoir identification.The XGBoost produced the model with the highest accuracy;up to 99%.The effective thickness of I and II oil layers determined during the reservoir identification was fed into the models for predicting production.Effective thickness considers the distribution of the water and the oil resulting in a more reasonable production prediction compared to predictions based on the overall reservoir thickness.To validate the superiority of the ML methods,reference models using overall reservoir thickness were built for comparison.The models based on effective thickness outperformed the reference models in every evaluation metric.The prediction accuracy of the ML models using effective thickness were 10%higher than that of reference model.Without the personal error or data distortion existing in traditional methods,this novel system realizes rapid analysis of data while reducing the time required to resolve reservoir classification and production prediction challenges.The ML models using the effective thickness obtained from reservoir identification were more accurate when predicting oil production compared to previous studies which use overall reservoir thickness.展开更多
Accurate prediction ofmonthly oil and gas production is essential for oil enterprises tomake reasonable production plans,avoid blind investment and realize sustainable development.Traditional oil well production trend...Accurate prediction ofmonthly oil and gas production is essential for oil enterprises tomake reasonable production plans,avoid blind investment and realize sustainable development.Traditional oil well production trend prediction methods are based on years of oil field production experience and expertise,and the application conditions are very demanding.With the rapid development of artificial intelligence technology,big data analysis methods are gradually applied in various sub-fields of the oil and gas reservoir development.Based on the data-driven artificial intelligence algorithmGradient BoostingDecision Tree(GBDT),this paper predicts the initial single-layer production by considering geological data,fluid PVT data and well data.The results show that the GBDT algorithm prediction model has great accuracy,significantly improving efficiency and strong universal applicability.The GBDTmethod trained in this paper can predict production,which is helpful for well site optimization,perforation layer optimization and engineering parameter optimization and has guiding significance for oilfield development.展开更多
In order to overcome the deficiencies of current methods for the prediction of the productivity of shale gas hor-izontal wells after fracturing,a new sophisticated approach is proposed in this study.This new model stem...In order to overcome the deficiencies of current methods for the prediction of the productivity of shale gas hor-izontal wells after fracturing,a new sophisticated approach is proposed in this study.This new model stems from the combination several techniques,namely,artificial neural network(ANN),particle swarm optimization(PSO),Imperialist Competitive Algorithms(ICA),and Ant Clony Optimization(ACO).These are properly implemented by using the geological and engineering parameters collected from 317 wells.The results show that the optimum PSO-ANN model has a high accuracy,obtaining a R2 of 0.847 on the testing.The partial dependence plots(PDP)indicate that liquid consumption intensity and the proportion of quartz sand are the two most sensitive factors affecting the model’s performance.展开更多
Accurately predicting the production rate and estimated ultimate recovery(EUR)of shale oil wells is vital for efficient shale oil development.Although numerical simulations provide accurate predictions,their high time...Accurately predicting the production rate and estimated ultimate recovery(EUR)of shale oil wells is vital for efficient shale oil development.Although numerical simulations provide accurate predictions,their high time,data,and labor demands call for a swifter,yet precise,method.This study introduces the DuongeCNNeLSTM(D-C-L)model,which integrates a convolutional neural network(CNN)with a long short-term memory(LSTM)network and is grounded on the empirical Duong model for physical constraints.Compared to traditional approaches,the D-C-L model demonstrates superior precision,efficiency,and cost-effectiveness in predicting shale oil production.展开更多
Based on an analysis of the limitations of conventional production component methods for natural gas development planning,this study proposes a new one that uses life cycle models for the trend fitting and prediction ...Based on an analysis of the limitations of conventional production component methods for natural gas development planning,this study proposes a new one that uses life cycle models for the trend fitting and prediction of production.In this new method,the annual production of old and new wells is predicted by year first and then is summed up to yield the production for the planning period.It shows that the changes in the production of old wells in old blocks can be fitted and predicted using the vapor pressure model(VPM),with precision of 80%e95%,which is 6.6%e13.2%higher than that of other life cycle models.Furthermore,a new production prediction process and method for new wells have been established based on this life cycle model to predict the production of medium-to-shallow gas reservoirs in western Sichuan Basin,with predication error of production rate in 2021 and 2022 being 6%and 3%respectively.The new method can be used to guide the medium-and long-term planning or annual scheme preparation for gas development.It is also applicable to planning for large single gas blocks that require continuous infill drilling and adjustment to improve gas recovery.展开更多
Hydrocarbon production in oil and gas fields generally progresses through stages of production ramp-up,plateau(peak),and decline during field development,with the whole process primarily modeled and forecasted using l...Hydrocarbon production in oil and gas fields generally progresses through stages of production ramp-up,plateau(peak),and decline during field development,with the whole process primarily modeled and forecasted using lifecycle models.SINOPEC's conventional gas reservoirs are dominated by carbonates,low-permeability tight sandstone,condensate,volcanic rocks,and medium-to-high-permeability sandstone.This study identifies the optimal production forecasting models by comparing the fitting coefficients of different models and calculating the relative errors in technically recoverable reserves.To improve forecast precision,it suggests substituting exponential smoothing method-derived predictions for anomalous data caused by subjective influences like market dynamics and maintenance activities.The preferred models for carbonate gas reservoir production forecasts are the generalized Weng's,Beta,Class-I generalized mathematical,and Hu-Chen models.The Vapor pressure and Beta models are optimal for forecasting the annual productivity of wells(APW)from gas-bearing low-permeability tight sandstone reservoirs.The Wang-Li,Beta,and Yu QT tb models are apt for moderate-to-small-reserves,single low-permeability tight sandstone gas reservoirs.The Rayleigh,Hu-Chen,and generalized Weng's models are suitable for condensate gas reservoirs.For medium-to-high-permeability sandstone gas reservoirs,the lognormal,generalized gamma,and Beta models are recommended.展开更多
A deep learning method for predicting oil field production at ultra-high water cut stage from the existing oil field production data was presented,and the experimental verification and application effect analysis were...A deep learning method for predicting oil field production at ultra-high water cut stage from the existing oil field production data was presented,and the experimental verification and application effect analysis were carried out.Since the traditional Fully Connected Neural Network(FCNN)is incapable of preserving the correlation of time series data,the Long Short-Term Memory(LSTM)network,which is a kind of Recurrent Neural Network(RNN),was utilized to establish a model for oil field production prediction.By this model,oil field production can be predicted from the relationship between oil production index and its influencing factors and the trend and correlation of oil production over time.Production data of a medium and high permeability sandstone oilfield in China developed by water flooding was used to predict its production at ultra-high water cut stage,and the results were compared with the results from the traditional FCNN and water drive characteristic curves.The LSTM based on deep learning has higher precision,and gives more accurate production prediction for complex time series in oil field production.The LSTM model was used to predict the monthly oil production of another two oil fields.The prediction results are good,which verifies the versatility of the method.展开更多
Since the oil production of single well in water flooding reservoir varies greatly and is hard to predict, an oil production prediction method of single well based on temporal convolutional network(TCN) is proposed an...Since the oil production of single well in water flooding reservoir varies greatly and is hard to predict, an oil production prediction method of single well based on temporal convolutional network(TCN) is proposed and verified. This method is started from data processing, the correspondence between water injectors and oil producers is determined according to the influence radius of the water injectors, the influence degree of a water injector on an oil producer in the month concerned is added as a model feature, and a Random Forest(RF) model is built to fill the dynamic data of water flooding. The single well history is divided into 4 stages according to its water cut, that is, low water cut, middle water cut, high water cut and extra-high water cut stages. In each stage, a TCN based prediction model is established, hyperparameters of the model are optimized by the Sparrow Search Algorithm(SSA). Finally, the models of the 4 stages are integrated into one whole-life model of the well for production prediction. The application of this method in Daqing Oilfield, NE China shows that:(1) Compared with conventional data processing methods, the data obtained by this processing method are more close to the actual production, and the data set obtained is more authentic and complete.(2) The TCN model has higher prediction accuracy than other 11 models such as Long Short Term Memory(LSTM).(3) Compared with the conventional full-life-cycle models, the model of integrated stages can significantly reduce the error of production prediction.展开更多
In accordance with the fracturing and producing mechanism in coalbed methane well, and combining the knowledge of fluid mechanics, linear elastic fracture mechanics, thermal transfer, computing mathematics and softwar...In accordance with the fracturing and producing mechanism in coalbed methane well, and combining the knowledge of fluid mechanics, linear elastic fracture mechanics, thermal transfer, computing mathematics and software engineering, the three-dimensional hydraulic fracture propagating and dynamical production predicting models for coalbed methane well is put forward. The fracture propagation model takes the variation of rock mechanical properties and in-situ stress distribution into consideration. The dynamic performance prediction model takes the gas production mechanism into consideration. With these models, a three-dimensional hydraulic fracturing optimum design software for coalbed methane well is developed, and its practicality and reliability have been proved by ex-ample computation.展开更多
In order to accurately predict the productivity of herringbone multilateral well,a new productivity prediction model was founded.And based on this model,orthogonal test and multiple factor variance analysis were appli...In order to accurately predict the productivity of herringbone multilateral well,a new productivity prediction model was founded.And based on this model,orthogonal test and multiple factor variance analysis were applied to study optimization design of herringbone multilateral well.According to the characteristics of herringbone multilateral well,by using pressure superposition and mirror image reflection theory,the coupled model of herringbone multilateral well was developed on the basis of a three-dimensional pseudo-pressure distribution model for horizontal wells.The model was formulated in consideration of friction loss,acceleration loss of the wellbore and mixed loss at the confluence of main wellbore and branched one.After mathematical simulation on productivity of the herringbone multilateral well with the coupled model,the effects of well configuration on productivity were analyzed.The results show that lateral number is the most important factor,length of main wellbore and length of branched wellbore are the secondary ones,angle between main and branched one has the least influence.展开更多
Increasing the production and utilization of shale gas is of great significance for building a clean and low-carbon energy system.Sharp decline of gas production has been widely observed in shale gas reservoirs.How to...Increasing the production and utilization of shale gas is of great significance for building a clean and low-carbon energy system.Sharp decline of gas production has been widely observed in shale gas reservoirs.How to forecast shale gas production is still challenging due to complex fracture networks,dynamic fracture properties,frac hits,complicated multiphase flow,and multi-scale flow as well as data quality and uncertainty.This work develops an integrated framework for evaluating shale gas well production based on data-driven models.Firstly,a comprehensive dominated-factor system has been established,including geological,drilling,fracturing,and production factors.Data processing and visualization are required to ensure data quality and determine final data set.A shale gas production evaluation model is developed to evaluate shale gas production levels.Finally,the random forest algorithm is used to forecast shale gas production.The prediction accuracy of shale gas production level is higher than 95%based on the shale gas reservoirs in China.Forty-one wells are randomly selected to predict cumulative gas production using the optimal regression model.The proposed shale gas production evaluation frame-work overcomes too many assumptions of analytical or semi-analytical models and avoids huge computation cost and poor generalization for numerical modelling.展开更多
This paper presents the development of numerical prediction products(NPP) correction and display system(NPPCDS) for rapid and effective post-processing and displaying of the T213 NPP(numerical prediction products of t...This paper presents the development of numerical prediction products(NPP) correction and display system(NPPCDS) for rapid and effective post-processing and displaying of the T213 NPP(numerical prediction products of the medium range numerical weather prediction spectral model T213L31) through instant correction method. The NPPCDS consists of two modules: an automatic correction module and a graphical display module. The automatic correction module automatically corrects the T213 NPP at regularly scheduled time intervals, while the graphical display module interacts with users to display the T213 NPP and its correction results. The system helps forecasters extract the most relevant information at a quick glance without extensive post-processing. It is simple, easy to use, and computationally efficient, and has been running stably at Huludao Meteorological Bureau in Liaoning Province of China for the past three years. Because of its low computational costs, it is particularly useful for meteorological departments that lack advanced computing capacity and still need to make short-range weather forecasting.展开更多
There has been increased interest in quantifying the manure production of livestock, primarily driven by public authorities, who aim to evaluate the environmental impact of livestock production, but also at the farm l...There has been increased interest in quantifying the manure production of livestock, primarily driven by public authorities, who aim to evaluate the environmental impact of livestock production, but also at the farm level, to manage manure storage and availability of fertilizer for crop production. Moreover, current manure production estimates from intensively reared beef calves are higher than actual production due to changes in farming systems, advances in animal genetics and feed efficiency. This study aims to redefine and update manure production estimates in intensively reared beef calves to predict manure production as a policy and planning tool, as there are no current models available. A trial was conducted to collect data on manure production during the growing-finishing period (243 d) of 54 Limousine calves (from 346.7 to 674.0 kg live weight, LW). Such data were used to develop two models to predict manure excretion: (1) a complex mechanistic model (CompM), and (2) a simplified empirical model (SimpM). Both models were evaluated against an independent dataset including a total of 4,692 animals on 31 farms and 5 breeds. Results from CompM require interpretation because the model does not output a single value but a range of manure production (minimum, medium and maximum), and would therefore be more suitable for professional use. The SimpM could be considered simple, reliable, and versatile for predicting manure excretion at farm level. SimpM could be refined and improved by including data from other studies on beef cattle with distinct characteristics and management.展开更多
Using the meteorological data during 1971- 2013 and lemon growth and yield data during 2003- 2013 in Anyue,the suitability problem of lemon growth and correlation problem between meteorological factors and lemon growt...Using the meteorological data during 1971- 2013 and lemon growth and yield data during 2003- 2013 in Anyue,the suitability problem of lemon growth and correlation problem between meteorological factors and lemon growth in Anyue area were studied. According to relevance between the selected meteorological factors and yield of lemon,meteorological prediction model of lemon yield was established in Anyue,and the prediction accuracy was higher. The research had certain guiding significance for management work of lemon production in Anyue area.展开更多
Reservoirs characterized by high temperature,high-pressure,medium high cementation strength,low porosity,and low permeability,in general,are not affected by sand production issues.Since 2009,however,it is known that c...Reservoirs characterized by high temperature,high-pressure,medium high cementation strength,low porosity,and low permeability,in general,are not affected by sand production issues.Since 2009,however,it is known that cases exists where sand is present and may represent a significant technical problem(e.g.,the the Dina II condensate gas field).In the present study,the main factors affecting sand production in this type of reservoir are considered(mechanical properties,stress fields,production system,completion method and gas flow pattern changes during the production process).On this basis,a new liquid-solid coupled porous elasto-plastic 3D sand production model is introduced and validated through comparison with effective sand production data.The related prediction errors are found to be within 15%,which represents the necessary prerequisite for the utilization of such a model for the elaboration of sand prevention measures.展开更多
Production prediction is an important factor influencing the realization of an intelligent agricultural supply chain.In an Internet of Things(IoT)environment,accurate yield prediction is one of the prerequisites for a...Production prediction is an important factor influencing the realization of an intelligent agricultural supply chain.In an Internet of Things(IoT)environment,accurate yield prediction is one of the prerequisites for achieving an efficient response in an intelligent agricultural supply chain.As an example,this study applied a conventional prediction method and deep learning prediction model to predict the yield of a characteristic regional fruit(the Shatian pomelo)in a comparative study.The root means square error(RMSE)values of regression analysis,exponential smoothing,grey prediction,grey neural network,support vector regression(SVR),and long short-term memory(LSTM)neural network methods were 53.715,6.707,18.440,1.580,and 1.436,respectively.Among these,the mean square error(MSE)values of the grey neural network,SVR,and LSTM neural network methods were 2.4979,31.652,and 2.0618,respectively;and theirRvalues were 0.99905,0.94,and 0.94501,respectively.The results demonstrated that the RMSE of the deep learning model is generally lower than that of a traditional prediction model,and the prediction results are more accurate.The prediction performance of the grey neural network was shown to be superior to that of SVR,and LSTM neural network,based on the comparison of parameters.展开更多
This study presents an avant-garde approach for predicting and optimizing production in tight reservoirs,employing a dual-medium unsteady seepage model specifically fashioned for volumetrically fractured horizontal we...This study presents an avant-garde approach for predicting and optimizing production in tight reservoirs,employing a dual-medium unsteady seepage model specifically fashioned for volumetrically fractured horizontal wells.Traditional models often fail to fully capture the complex dynamics associated with these unconventional reservoirs.In a significant departure from these models,our approach incorporates an initiation pressure gradient and a discrete fracture seepage network,providing a more realistic representation of the seepage process.The model also integrates an enhanced fluid-solid interaction,which allows for a more comprehensive understanding of the fluid-structure interactions in the reservoir.This is achieved through the incorporation of improved permeability and stress coupling,leading to more precise predictions of reservoir behavior.The numerical solutions derived from the model are obtained through the sophisticated finite element method,ensuring high accuracy and computational efficiency.To ensure the model’s reliability and accuracy,the outcomes were tested against a real-world case,with results demonstrating strong alignment.A key revelation from the study is the significant difference between uncoupled and fully coupled volumetrically fractured horizontal wells,challenging conventional wisdom in the field.Additionally,the study delves into the effects of stress,fracture length,and fracture number on reservoir production,contributing valuable insights for the design and optimization of tight reservoirs.The findings from this study have the potential to revolutionize the field of tight reservoir prediction and management,offering significant advancements in petroleum engineering.The proposed approach brings forth a more nuanced understanding of tight reservoir systems and opens up new avenues for optimizing reservoir management and production.展开更多
基金supported by the National Science and Technology Innovation 2030 Next-Generation Artifical Intelligence Major Project(2018AAA0101801)the National Natural Science Foundation of China(72271188)。
文摘With the development of information technology,a large number of product quality data in the entire manufacturing process is accumulated,but it is not explored and used effectively.The traditional product quality prediction models have many disadvantages,such as high complexity and low accuracy.To overcome the above problems,we propose an optimized data equalization method to pre-process dataset and design a simple but effective product quality prediction model:radial basis function model optimized by the firefly algorithm with Levy flight mechanism(RBFFALM).First,the new data equalization method is introduced to pre-process the dataset,which reduces the dimension of the data,removes redundant features,and improves the data distribution.Then the RBFFALFM is used to predict product quality.Comprehensive expe riments conducted on real-world product quality datasets validate that the new model RBFFALFM combining with the new data pre-processing method outperforms other previous me thods on predicting product quality.
基金supported by the National Natural Science Fund of China (No.52104049)the Science Foundation of China University of Petroleum,Beijing (No.2462022BJRC004)。
文摘In the early time of oilfield development, insufficient production data and unclear understanding of oil production presented a challenge to reservoir engineers in devising effective development plans. To address this challenge, this study proposes a method using data mining technology to search for similar oil fields and predict well productivity. A query system of 135 analogy parameters is established based on geological and reservoir engineering research, and the weight values of these parameters are calculated using a data algorithm to establish an analogy system. The fuzzy matter-element algorithm is then used to calculate the similarity between oil fields, with fields having similarity greater than 70% identified as similar oil fields. Using similar oil fields as sample data, 8 important factors affecting well productivity are identified using the Pearson coefficient and mean decrease impurity(MDI) method. To establish productivity prediction models, linear regression(LR), random forest regression(RF), support vector regression(SVR), backpropagation(BP), extreme gradient boosting(XGBoost), and light gradient boosting machine(Light GBM) algorithms are used. Their performance is evaluated using the coefficient of determination(R^(2)), explained variance score(EV), mean squared error(MSE), and mean absolute error(MAE) metrics. The Light GBM model is selected to predict the productivity of 30 wells in the PL field with an average error of only 6.31%, which significantly improves the accuracy of the productivity prediction and meets the application requirements in the field. Finally, a software platform integrating data query,oil field analogy, productivity prediction, and knowledge base is established to identify patterns in massive reservoir development data and provide valuable technical references for new reservoir development.
基金Supported by National Natural Science Foundation of China(52104049)Young Elite Scientist Sponsorship Program by BAST(BYESS2023262)Science Foundation of China University of Petroleum,Beijing(2462022BJRC004).
文摘Considering the phase behaviors in condensate gas reservoirs and the oil-gas two-phase linear flow and boundary-dominated flow in the reservoir,a method for predicting the relationship between oil saturation and pressure in the full-path of tight condensate gas well is proposed,and a model for predicting the transient production from tight condensate gas wells with multiphase flow is established.The research indicates that the relationship curve between condensate oil saturation and pressure is crucial for calculating the pseudo-pressure.In the early stage of production or in areas far from the wellbore with high reservoir pressure,the condensate oil saturation can be calculated using early-stage production dynamic data through material balance models.In the late stage of production or in areas close to the wellbore with low reservoir pressure,the condensate oil saturation can be calculated using the data of constant composition expansion test.In the middle stages of production or when reservoir pressure is at an intermediate level,the data obtained from the previous two stages can be interpolated to form a complete full-path relationship curve between oil saturation and pressure.Through simulation and field application,the new method is verified to be reliable and practical.It can be applied for prediction of middle-stage and late-stage production of tight condensate gas wells and assessment of single-well recoverable reserves.
文摘Reservoir identification and production prediction are two of the most important tasks in petroleum exploration and development.Machine learning(ML)methods are used for petroleum-related studies,but have not been applied to reservoir identification and production prediction based on reservoir identification.Production forecasting studies are typically based on overall reservoir thickness and lack accuracy when reservoirs contain a water or dry layer without oil production.In this paper,a systematic ML method was developed using classification models for reservoir identification,and regression models for production prediction.The production models are based on the reservoir identification results.To realize the reservoir identification,seven optimized ML methods were used:four typical single ML methods and three ensemble ML methods.These methods classify the reservoir into five types of layers:water,dry and three levels of oil(I oil layer,II oil layer,III oil layer).The validation and test results of these seven optimized ML methods suggest the three ensemble methods perform better than the four single ML methods in reservoir identification.The XGBoost produced the model with the highest accuracy;up to 99%.The effective thickness of I and II oil layers determined during the reservoir identification was fed into the models for predicting production.Effective thickness considers the distribution of the water and the oil resulting in a more reasonable production prediction compared to predictions based on the overall reservoir thickness.To validate the superiority of the ML methods,reference models using overall reservoir thickness were built for comparison.The models based on effective thickness outperformed the reference models in every evaluation metric.The prediction accuracy of the ML models using effective thickness were 10%higher than that of reference model.Without the personal error or data distortion existing in traditional methods,this novel system realizes rapid analysis of data while reducing the time required to resolve reservoir classification and production prediction challenges.The ML models using the effective thickness obtained from reservoir identification were more accurate when predicting oil production compared to previous studies which use overall reservoir thickness.
文摘Accurate prediction ofmonthly oil and gas production is essential for oil enterprises tomake reasonable production plans,avoid blind investment and realize sustainable development.Traditional oil well production trend prediction methods are based on years of oil field production experience and expertise,and the application conditions are very demanding.With the rapid development of artificial intelligence technology,big data analysis methods are gradually applied in various sub-fields of the oil and gas reservoir development.Based on the data-driven artificial intelligence algorithmGradient BoostingDecision Tree(GBDT),this paper predicts the initial single-layer production by considering geological data,fluid PVT data and well data.The results show that the GBDT algorithm prediction model has great accuracy,significantly improving efficiency and strong universal applicability.The GBDTmethod trained in this paper can predict production,which is helpful for well site optimization,perforation layer optimization and engineering parameter optimization and has guiding significance for oilfield development.
基金This study was financially supported by China United Coalbed Methane Corporation,Ltd.(ZZGSSALFGR2021-581),Bin Li received the grant.
文摘In order to overcome the deficiencies of current methods for the prediction of the productivity of shale gas hor-izontal wells after fracturing,a new sophisticated approach is proposed in this study.This new model stems from the combination several techniques,namely,artificial neural network(ANN),particle swarm optimization(PSO),Imperialist Competitive Algorithms(ICA),and Ant Clony Optimization(ACO).These are properly implemented by using the geological and engineering parameters collected from 317 wells.The results show that the optimum PSO-ANN model has a high accuracy,obtaining a R2 of 0.847 on the testing.The partial dependence plots(PDP)indicate that liquid consumption intensity and the proportion of quartz sand are the two most sensitive factors affecting the model’s performance.
基金funded by the National Natural Science Foundation of China(No.51974356).
文摘Accurately predicting the production rate and estimated ultimate recovery(EUR)of shale oil wells is vital for efficient shale oil development.Although numerical simulations provide accurate predictions,their high time,data,and labor demands call for a swifter,yet precise,method.This study introduces the DuongeCNNeLSTM(D-C-L)model,which integrates a convolutional neural network(CNN)with a long short-term memory(LSTM)network and is grounded on the empirical Duong model for physical constraints.Compared to traditional approaches,the D-C-L model demonstrates superior precision,efficiency,and cost-effectiveness in predicting shale oil production.
基金funded by the project entitled Technical Countermeasures for the Quantitative Characterization and Adjustment of Residual Gas in Tight Sandstone Gas Reservoirs of the Daniudi Gas Field(P20065-1)organized by the Science&Technology R&D Department of Sinopec.
文摘Based on an analysis of the limitations of conventional production component methods for natural gas development planning,this study proposes a new one that uses life cycle models for the trend fitting and prediction of production.In this new method,the annual production of old and new wells is predicted by year first and then is summed up to yield the production for the planning period.It shows that the changes in the production of old wells in old blocks can be fitted and predicted using the vapor pressure model(VPM),with precision of 80%e95%,which is 6.6%e13.2%higher than that of other life cycle models.Furthermore,a new production prediction process and method for new wells have been established based on this life cycle model to predict the production of medium-to-shallow gas reservoirs in western Sichuan Basin,with predication error of production rate in 2021 and 2022 being 6%and 3%respectively.The new method can be used to guide the medium-and long-term planning or annual scheme preparation for gas development.It is also applicable to planning for large single gas blocks that require continuous infill drilling and adjustment to improve gas recovery.
基金funded by the project entitled Technical Countermeasures for the Quantitative Characterization and Adjustment of Residual Gas in Tight Sandstone Gas Reservoirs of the Daniudi Gas Field(P20065-1)organized by the Science&Technology R&D Department of SINOPEC.
文摘Hydrocarbon production in oil and gas fields generally progresses through stages of production ramp-up,plateau(peak),and decline during field development,with the whole process primarily modeled and forecasted using lifecycle models.SINOPEC's conventional gas reservoirs are dominated by carbonates,low-permeability tight sandstone,condensate,volcanic rocks,and medium-to-high-permeability sandstone.This study identifies the optimal production forecasting models by comparing the fitting coefficients of different models and calculating the relative errors in technically recoverable reserves.To improve forecast precision,it suggests substituting exponential smoothing method-derived predictions for anomalous data caused by subjective influences like market dynamics and maintenance activities.The preferred models for carbonate gas reservoir production forecasts are the generalized Weng's,Beta,Class-I generalized mathematical,and Hu-Chen models.The Vapor pressure and Beta models are optimal for forecasting the annual productivity of wells(APW)from gas-bearing low-permeability tight sandstone reservoirs.The Wang-Li,Beta,and Yu QT tb models are apt for moderate-to-small-reserves,single low-permeability tight sandstone gas reservoirs.The Rayleigh,Hu-Chen,and generalized Weng's models are suitable for condensate gas reservoirs.For medium-to-high-permeability sandstone gas reservoirs,the lognormal,generalized gamma,and Beta models are recommended.
基金Supported by China National Science and Technology Major Project(2016ZX05016-006)
文摘A deep learning method for predicting oil field production at ultra-high water cut stage from the existing oil field production data was presented,and the experimental verification and application effect analysis were carried out.Since the traditional Fully Connected Neural Network(FCNN)is incapable of preserving the correlation of time series data,the Long Short-Term Memory(LSTM)network,which is a kind of Recurrent Neural Network(RNN),was utilized to establish a model for oil field production prediction.By this model,oil field production can be predicted from the relationship between oil production index and its influencing factors and the trend and correlation of oil production over time.Production data of a medium and high permeability sandstone oilfield in China developed by water flooding was used to predict its production at ultra-high water cut stage,and the results were compared with the results from the traditional FCNN and water drive characteristic curves.The LSTM based on deep learning has higher precision,and gives more accurate production prediction for complex time series in oil field production.The LSTM model was used to predict the monthly oil production of another two oil fields.The prediction results are good,which verifies the versatility of the method.
基金Major Unified Construction Project of Petro China(2019-40210-000020-02)。
文摘Since the oil production of single well in water flooding reservoir varies greatly and is hard to predict, an oil production prediction method of single well based on temporal convolutional network(TCN) is proposed and verified. This method is started from data processing, the correspondence between water injectors and oil producers is determined according to the influence radius of the water injectors, the influence degree of a water injector on an oil producer in the month concerned is added as a model feature, and a Random Forest(RF) model is built to fill the dynamic data of water flooding. The single well history is divided into 4 stages according to its water cut, that is, low water cut, middle water cut, high water cut and extra-high water cut stages. In each stage, a TCN based prediction model is established, hyperparameters of the model are optimized by the Sparrow Search Algorithm(SSA). Finally, the models of the 4 stages are integrated into one whole-life model of the well for production prediction. The application of this method in Daqing Oilfield, NE China shows that:(1) Compared with conventional data processing methods, the data obtained by this processing method are more close to the actual production, and the data set obtained is more authentic and complete.(2) The TCN model has higher prediction accuracy than other 11 models such as Long Short Term Memory(LSTM).(3) Compared with the conventional full-life-cycle models, the model of integrated stages can significantly reduce the error of production prediction.
文摘In accordance with the fracturing and producing mechanism in coalbed methane well, and combining the knowledge of fluid mechanics, linear elastic fracture mechanics, thermal transfer, computing mathematics and software engineering, the three-dimensional hydraulic fracture propagating and dynamical production predicting models for coalbed methane well is put forward. The fracture propagation model takes the variation of rock mechanical properties and in-situ stress distribution into consideration. The dynamic performance prediction model takes the gas production mechanism into consideration. With these models, a three-dimensional hydraulic fracturing optimum design software for coalbed methane well is developed, and its practicality and reliability have been proved by ex-ample computation.
基金Project(12521044) supported by Scientific and Technological Research Program of Heilongjiang Provincial Education Department,China
文摘In order to accurately predict the productivity of herringbone multilateral well,a new productivity prediction model was founded.And based on this model,orthogonal test and multiple factor variance analysis were applied to study optimization design of herringbone multilateral well.According to the characteristics of herringbone multilateral well,by using pressure superposition and mirror image reflection theory,the coupled model of herringbone multilateral well was developed on the basis of a three-dimensional pseudo-pressure distribution model for horizontal wells.The model was formulated in consideration of friction loss,acceleration loss of the wellbore and mixed loss at the confluence of main wellbore and branched one.After mathematical simulation on productivity of the herringbone multilateral well with the coupled model,the effects of well configuration on productivity were analyzed.The results show that lateral number is the most important factor,length of main wellbore and length of branched wellbore are the secondary ones,angle between main and branched one has the least influence.
基金funded by National Natural Science Foundation of China(52004238)China Postdoctoral Science Foundation(2019M663561).
文摘Increasing the production and utilization of shale gas is of great significance for building a clean and low-carbon energy system.Sharp decline of gas production has been widely observed in shale gas reservoirs.How to forecast shale gas production is still challenging due to complex fracture networks,dynamic fracture properties,frac hits,complicated multiphase flow,and multi-scale flow as well as data quality and uncertainty.This work develops an integrated framework for evaluating shale gas well production based on data-driven models.Firstly,a comprehensive dominated-factor system has been established,including geological,drilling,fracturing,and production factors.Data processing and visualization are required to ensure data quality and determine final data set.A shale gas production evaluation model is developed to evaluate shale gas production levels.Finally,the random forest algorithm is used to forecast shale gas production.The prediction accuracy of shale gas production level is higher than 95%based on the shale gas reservoirs in China.Forty-one wells are randomly selected to predict cumulative gas production using the optimal regression model.The proposed shale gas production evaluation frame-work overcomes too many assumptions of analytical or semi-analytical models and avoids huge computation cost and poor generalization for numerical modelling.
基金Under the auspices of National Natural Science Foundation of China(No.91125010)
文摘This paper presents the development of numerical prediction products(NPP) correction and display system(NPPCDS) for rapid and effective post-processing and displaying of the T213 NPP(numerical prediction products of the medium range numerical weather prediction spectral model T213L31) through instant correction method. The NPPCDS consists of two modules: an automatic correction module and a graphical display module. The automatic correction module automatically corrects the T213 NPP at regularly scheduled time intervals, while the graphical display module interacts with users to display the T213 NPP and its correction results. The system helps forecasters extract the most relevant information at a quick glance without extensive post-processing. It is simple, easy to use, and computationally efficient, and has been running stably at Huludao Meteorological Bureau in Liaoning Province of China for the past three years. Because of its low computational costs, it is particularly useful for meteorological departments that lack advanced computing capacity and still need to make short-range weather forecasting.
文摘There has been increased interest in quantifying the manure production of livestock, primarily driven by public authorities, who aim to evaluate the environmental impact of livestock production, but also at the farm level, to manage manure storage and availability of fertilizer for crop production. Moreover, current manure production estimates from intensively reared beef calves are higher than actual production due to changes in farming systems, advances in animal genetics and feed efficiency. This study aims to redefine and update manure production estimates in intensively reared beef calves to predict manure production as a policy and planning tool, as there are no current models available. A trial was conducted to collect data on manure production during the growing-finishing period (243 d) of 54 Limousine calves (from 346.7 to 674.0 kg live weight, LW). Such data were used to develop two models to predict manure excretion: (1) a complex mechanistic model (CompM), and (2) a simplified empirical model (SimpM). Both models were evaluated against an independent dataset including a total of 4,692 animals on 31 farms and 5 breeds. Results from CompM require interpretation because the model does not output a single value but a range of manure production (minimum, medium and maximum), and would therefore be more suitable for professional use. The SimpM could be considered simple, reliable, and versatile for predicting manure excretion at farm level. SimpM could be refined and improved by including data from other studies on beef cattle with distinct characteristics and management.
基金Supported by Government Science Research Item of Anyue County,China(2013-17)
文摘Using the meteorological data during 1971- 2013 and lemon growth and yield data during 2003- 2013 in Anyue,the suitability problem of lemon growth and correlation problem between meteorological factors and lemon growth in Anyue area were studied. According to relevance between the selected meteorological factors and yield of lemon,meteorological prediction model of lemon yield was established in Anyue,and the prediction accuracy was higher. The research had certain guiding significance for management work of lemon production in Anyue area.
基金This study has been supported by the Major Science and Technology Project“Comprehensive Research of Exploration Matching and Application of New Technology”(2016ZX5051-3)of CNPC.
文摘Reservoirs characterized by high temperature,high-pressure,medium high cementation strength,low porosity,and low permeability,in general,are not affected by sand production issues.Since 2009,however,it is known that cases exists where sand is present and may represent a significant technical problem(e.g.,the the Dina II condensate gas field).In the present study,the main factors affecting sand production in this type of reservoir are considered(mechanical properties,stress fields,production system,completion method and gas flow pattern changes during the production process).On this basis,a new liquid-solid coupled porous elasto-plastic 3D sand production model is introduced and validated through comparison with effective sand production data.The related prediction errors are found to be within 15%,which represents the necessary prerequisite for the utilization of such a model for the elaboration of sand prevention measures.
基金This work was supported by the 2021‘Cultivation plan for thousands of young andmiddle-aged backbone teachers in Guangxi Colleges and universities’by the Project of Humanities and Social Sciences in‘Research on Collaborative Integration of Logistics Service Supply Chain under High-QualityDevelopmentGoals’(2021QGRW044)In addition,the studywas supported by the 2019 National Social Science Project in‘Research on the Integration of Transnational Supply Chains under the Belt and Road Initiative(19BJY184)’This paper was also supported by Guangxi Philosophy and Social Science Planning Office Project:Research on the DynamicMechanism and Model Innovation of the Cross-border Integration Growth of Guangxi Logistics Enterprises(18BGL010).
文摘Production prediction is an important factor influencing the realization of an intelligent agricultural supply chain.In an Internet of Things(IoT)environment,accurate yield prediction is one of the prerequisites for achieving an efficient response in an intelligent agricultural supply chain.As an example,this study applied a conventional prediction method and deep learning prediction model to predict the yield of a characteristic regional fruit(the Shatian pomelo)in a comparative study.The root means square error(RMSE)values of regression analysis,exponential smoothing,grey prediction,grey neural network,support vector regression(SVR),and long short-term memory(LSTM)neural network methods were 53.715,6.707,18.440,1.580,and 1.436,respectively.Among these,the mean square error(MSE)values of the grey neural network,SVR,and LSTM neural network methods were 2.4979,31.652,and 2.0618,respectively;and theirRvalues were 0.99905,0.94,and 0.94501,respectively.The results demonstrated that the RMSE of the deep learning model is generally lower than that of a traditional prediction model,and the prediction results are more accurate.The prediction performance of the grey neural network was shown to be superior to that of SVR,and LSTM neural network,based on the comparison of parameters.
文摘This study presents an avant-garde approach for predicting and optimizing production in tight reservoirs,employing a dual-medium unsteady seepage model specifically fashioned for volumetrically fractured horizontal wells.Traditional models often fail to fully capture the complex dynamics associated with these unconventional reservoirs.In a significant departure from these models,our approach incorporates an initiation pressure gradient and a discrete fracture seepage network,providing a more realistic representation of the seepage process.The model also integrates an enhanced fluid-solid interaction,which allows for a more comprehensive understanding of the fluid-structure interactions in the reservoir.This is achieved through the incorporation of improved permeability and stress coupling,leading to more precise predictions of reservoir behavior.The numerical solutions derived from the model are obtained through the sophisticated finite element method,ensuring high accuracy and computational efficiency.To ensure the model’s reliability and accuracy,the outcomes were tested against a real-world case,with results demonstrating strong alignment.A key revelation from the study is the significant difference between uncoupled and fully coupled volumetrically fractured horizontal wells,challenging conventional wisdom in the field.Additionally,the study delves into the effects of stress,fracture length,and fracture number on reservoir production,contributing valuable insights for the design and optimization of tight reservoirs.The findings from this study have the potential to revolutionize the field of tight reservoir prediction and management,offering significant advancements in petroleum engineering.The proposed approach brings forth a more nuanced understanding of tight reservoir systems and opens up new avenues for optimizing reservoir management and production.
