Drilling costs of ultra-deepwell is the significant part of development investment,and accurate prediction of drilling costs plays an important role in reasonable budgeting and overall control of development cost.In o...Drilling costs of ultra-deepwell is the significant part of development investment,and accurate prediction of drilling costs plays an important role in reasonable budgeting and overall control of development cost.In order to improve the prediction accuracy of ultra-deep well drilling costs,the item and the dominant factors of drilling costs in Tarim oilfield are analyzed.Then,those factors of drilling costs are separated into categorical variables and numerous variables.Finally,a BP neural networkmodel with drilling costs as the output is established,and hyper-parameters(initial weights and bias)of the BP neural network is optimized by genetic algorithm(GA).Through training and validation of themodel,a reliable prediction model of ultra-deep well drilling costs is achieved.The average relative error between prediction and actual values is 3.26%.Compared with other models,the root mean square error is reduced by 25.38%.The prediction results of the proposed model are reliable,and the model is efficient,which can provide supporting for the drilling costs control and budget planning of ultra-deep wells.展开更多
Multilateral wells promise cost savings to oil and fields as they have the potential to reduce overall drilling distances and minimize the number of slots required for the surface facility managing the well.However,dr...Multilateral wells promise cost savings to oil and fields as they have the potential to reduce overall drilling distances and minimize the number of slots required for the surface facility managing the well.However,drilling a multilateral well does not always increase the flow rate when compared to two single-horizontal wells due to competition in production inside the mother-bore.Here,a holistic approach is proposed to find the optimum balance between single and multilateral wells in an offshore oil development.In so doing,the integrated approach finds the highest Net Present Value(NPV)configuration of the field considering drilling,subsurface,production and financial analysis.The model employs stochastic perturbation and Markov Chain Monte-Carlo methods to solve the global maximising-NPV problem.In addition,a combination of Mixed-Integer Linear Programming(MILP),an improved Dijkstra algorithm and a Levenberg-Marquardt optimiser is proposed to solve the rate allocation problem.With the outcome from this analysis,the model suggests the optimum development including number of multilateral and single horizontal wells that would result in the highest NPV.The results demonstrate the potential for modelling to find the optimal use of petroleum facilities and to assist with planning and decision making.展开更多
基金supported by the Science and Technology Innovation Foundation of CNPC“Multiscale Flow Law and Flow Field Coupling Study of Tight Sandstone Gas Reservoir”(2016D-5007-0208)13th Five-Year National Major Project“Multistage Fracturing Effect and Production of Fuling Shale Gas HorizontalWell Law Analysis Research”(2016ZX05060-009).
文摘Drilling costs of ultra-deepwell is the significant part of development investment,and accurate prediction of drilling costs plays an important role in reasonable budgeting and overall control of development cost.In order to improve the prediction accuracy of ultra-deep well drilling costs,the item and the dominant factors of drilling costs in Tarim oilfield are analyzed.Then,those factors of drilling costs are separated into categorical variables and numerous variables.Finally,a BP neural networkmodel with drilling costs as the output is established,and hyper-parameters(initial weights and bias)of the BP neural network is optimized by genetic algorithm(GA).Through training and validation of themodel,a reliable prediction model of ultra-deep well drilling costs is achieved.The average relative error between prediction and actual values is 3.26%.Compared with other models,the root mean square error is reduced by 25.38%.The prediction results of the proposed model are reliable,and the model is efficient,which can provide supporting for the drilling costs control and budget planning of ultra-deep wells.
文摘Multilateral wells promise cost savings to oil and fields as they have the potential to reduce overall drilling distances and minimize the number of slots required for the surface facility managing the well.However,drilling a multilateral well does not always increase the flow rate when compared to two single-horizontal wells due to competition in production inside the mother-bore.Here,a holistic approach is proposed to find the optimum balance between single and multilateral wells in an offshore oil development.In so doing,the integrated approach finds the highest Net Present Value(NPV)configuration of the field considering drilling,subsurface,production and financial analysis.The model employs stochastic perturbation and Markov Chain Monte-Carlo methods to solve the global maximising-NPV problem.In addition,a combination of Mixed-Integer Linear Programming(MILP),an improved Dijkstra algorithm and a Levenberg-Marquardt optimiser is proposed to solve the rate allocation problem.With the outcome from this analysis,the model suggests the optimum development including number of multilateral and single horizontal wells that would result in the highest NPV.The results demonstrate the potential for modelling to find the optimal use of petroleum facilities and to assist with planning and decision making.
