Method and system for reservoir monitoring

Abstract

The invention provides a method of improving the accuracy of a reservoir model. The method comprising obtaining tracer test data associated with the reservoir and constructing a model of the reservoir. The method comprising determining a model tracer residence time distribution data set for the reservoir and calculating at least one moment of the model tracer residence time distribution data set. The method comprising calculating at least one moment of residence time distribution data set from the tracer test data and comparing the at least one moment of the measurement data set and the at least one moment of the model data set to assess the accuracy of the model.

Description

[0001] Method and System for Reservoir Monitoring

[0002] The present invention relates to the field of hydrocarbon reservoir monitoring and more specifically hydrocarbon reservoir monitoring using tracers. Aspects of the invention include a method of simulating the transport of tracers in a reservoir with one or more wells. Another aspect relates to the interpretation of tracer data by use of residence time distribution (RTD) analysis.

[0003] Background to the invention

[0004] The efficient recovery of hydrocarbons from a reservoir is a difficult and complex process which reguires an understanding of the flow conditions of the hydrocarbons in the reservoir and connected wells.

[0005] Tracer technology has been previously used to label fluids so that their migration through the reservoir and connected well is tracked to better understand reservoir fluid flow pathways and reservoir characteristics. Accurate tracer flow analysis can be difficult due to the different volumetric and phase behaviour of tracers and fluids as they travel through the reservoir. Reservoir simulation models have been used in the past to predict the flow of fluids through reservoirs.

[0006] Reservoir modelling may use tracer response information to understand and predict fluid flow and transport processes in the reservoir. The purpose of the model is to provide forecasts that may guide decisions on production and development of the reservoir. The accuracy of the model and other simulation technigues affects the ability to predict characteristics or conditions of the reservoir. The accuracy of reservoir modelling has presented numerous challenges and increased difficulty.

[0007] Summary of the invention

[0008] It is amongst the aims and objects of the invention to provide a method and system for enhanced history matching of a reservoir model.

[0009] It is another object of an aspect of the invention to provide a method and system for improving the accuracy of comparing simulated data with measured data. It is another object of the present invention to provide a method of assessing the grid size accuracy of a reservoir model.

[0010] Further aims and objects of the invention will become apparent from reading the following description.

[0011] According to a first aspect of the invention there is provided a method of improving the accuracy of a reservoir model comprising: obtaining tracer test data associated with a reservoir; calculating, estimating and / or determining a model tracer data set for the reservoir; calculating at least one spatial or temporal moment of model tracer data set; calculating at least one spatial or temporal moment of measured tracer data set from the tracer test data; comparing the at least one spatial or temporal moment of the measurement data set and the at least one spatial or temporal moment of the model data set to assess the accuracy of the model.

[0012] The method may comprise obtaining tracer test data from historical tracer injection and / or production data associated with production from the reservoir. The method may comprise calculating, estimating and / or determining a model tracer data set for the reservoir based on a first set of parameters values.

[0013] The method may comprise obtaining tracer test data from historical tests previously conducted. The historical tracer data may be stored in a database. The method may comprise performing a tracer test to generate the tracer test data.

[0014] The method may comprise constructing a model of the reservoir. The method may comprise constructing a model of the reservoir from measured physical data. The method may comprise constructing the model from historical measured tracer data. The method may comprise constructing the model from historical physical data.

[0015] The method may comprise adjusting one or more parameter and / or one or more parameter value in the model until a desired misfit threshold is reached. The method may comprise adjusting one or more parameter and / or one or more parameter value in a first set of parameters until a desired misfit threshold is reached. The method may comprise iteratively varying at least one a value of at least one parameter and / or at least one parameter in the model until a misfit or difference between at least one moment of the measurement data set and the at least one moment of the model data set is reduced.

[0016] The method may comprise iteratively varying at least one a value of at least one parameter in the model until a misfit or difference between the at least one moment of the measurement data set and the at least one moment of the model data set is reduced to a desired level.

[0017] The tracer test data may comprise data from one or more tracers. The tracer test data may comprise data from two or more tracers. The method may comprise analysing data from one or more tracers separately. The method may comprise analysing data from two or more tracers separately. The method may comprise analysing the combined data from the two or more tracers. The method may comprise modelling data from each of the one or more tracers separately. The method may comprise modelling data from each of the two or more tracers separately. The method may comprise modelling data from the combined data of the two or more tracers.

[0018] The method may comprise iterative tuning of the model until the model moment data best matches the measured moment data.

[0019] The model tracer data set may be at least one tracer concentration as a function of time. The model tracer data set may be tracer concentrations as a function of time. The model tracer data set may be modelled tracer concentrations as a function of time. The model tracer data set may be modelled produced tracer concentrations as a function of time. The model tracer data set may be a tracer concentration time series.

[0020] The measured tracer data set may be at least one tracer concentration as a function of time. The measured tracer data set may be tracer concentrations as a function of time. The measured tracer data set may be measured tracer concentrations as a function of time. The measured tracer data set may be measured produced tracer concentrations as a function of time The measured tracer data set may be a tracer concentration time series. The model tracer data set may be a modelled residence time distribution. The measured tracer data set may be a residence time distribution. The one or more tracer may be a water tracer, an oil tracer and / or a gas tracer. The model tracer data set may be a tracer concentration as a function of space. The model tracer data set may be a tracer concentration measured at distinct positions in space. The measured tracer data set may be a tracer concentration as a function of space. The measured tracer data set may be a tracer concentration measured at distinct positions in space. The model tracer data set may be a modelled spatial distribution of tracer. The measured tracer data set may be a spatial distribution of tracer.

[0021] The reservoir model may be an interwell model, a partitioning inter-well tracer test model and / or a single well chemical tracer test model.

[0022] The method may be a computer-implemented method. The method may be a computer- implemented history matching method. The method may comprise storing the measurement data to a database. The method may comprise storing the model data to a database.

[0023] The at least one moment may be selected from a zero order moment, a first order moment and / or a second order moment. The at least one order moment may be more than two orders.

[0024] The measured residence time distribution data set may be obtained by injecting tracer into a reservoir and monitoring tracer concentrations produced from the reservoir as a function of time and / or space. The modelled residence time distribution data set may be obtained by modelling an injection of tracer into a model reservoir and modelling tracer concentrations produced from the reservoir as a function of time and / or space.

[0025] The method may comprise creating a tracer curve from the measured tracer concentration as a function of time. The method may comprise creating a tracer curve from measured tracer concentrations produced as a function of time. The method may comprise calculating an area under the tracer curve to calculate the at least one moment of measured tracer data set.

[0026] The method may comprise creating a tracer curve from the modelled tracer concentration as a function of time. The method may comprise creating a tracer curve from modelled produced tracer concentrations as a function of time. The method may comprise calculating an area under the tracer curve to calculate the at least one moment of modelled tracer data set. The method may comprise comparing the at least one moment of the measurement data set and the at least one moment of the model data set to assess the accuracy of the model.

[0027] The method may comprise calculating a swept volume of a tracer in the reservoir from the first order moment of the tracer data set. The method may comprise calculating a modelled swept volume of a modelled tracer in the model reservoir from the first order moment of the modelled tracer data set. The method may comprise comparing a first order moment of the modelled tracer data set with a first order moment of the measured tracer data set to assess the accuracy of the swept volume model.

[0028] The method may comprise characterizing the reservoir based on the model. The method may comprise controlling the reservoir based on the model. The method may comprise controlling injection and / or injection rates into the reservoir based on the model. The method may comprise controlling production and / or production rates from the reservoir based on the model. The method may comprise characterising well to well communications. The method may comprise characterising well to reservoir communication. The method may comprise characterising reservoir to well communication. The method may comprise characterising fluid dynamics, residence times and / or heterogeneities in the reservoir. The method may comprise characterising fluid dynamics, residence times and / or heterogeneities in at least one well.

[0029] The reservoir may be in fluid communication with at least one well. The reservoir may be in fluid communication with two or more wells. The reservoir may be in fluid communication with at least one injector well. The reservoir may be in fluid communication with two or more injector wells. The reservoir may be in fluid communication with a plurality of injector wells. The reservoir may be in fluid communication with at least one producer well. The reservoir may be in fluid communication with at least two producer wells. The reservoir may be in fluid communication with a plurality of producer wells.

[0030] The method may comprise analysing the tuned model to determine at least one characteristic of the reservoir selected from the group comprising residence time, spatial distribution of tracer mass, oil saturation, relative permeability, porosity, flow paths, flow rate, water breakthrough time. According to a second aspect of the invention there is provided a method of improving the accuracy of a reservoir model comprising: obtaining tracer test data associated with the reservoir; constructing a model of the reservoir, calculating, estimating and / or determining a model tracer residence time distribution data set for the reservoir; calculating at least one moment of the model tracer residence time distribution data set; calculating at least one moment of residence time distribution data set from the tracer test data; comparing the at least one moment of the measurement data set and the at least one moment of the model data set to assess the accuracy of the model.

[0031] The method may comprise calculating, estimating and / or determining a comparison value between the at least one moment of the residence time distribution data from the tracer test data set and the at least one moment of the residence time distribution data from the model data set. The method may comprise calculating, estimating and / or determining a model tracer data set for the reservoir based on a first set of parameters values.

[0032] The method may comprise performing a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range. The method may comprise performing a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is reduced. The method may comprise performing a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until a best fit is reached between the measured data and model data.

[0033] Embodiments of the second aspect of the invention may include one or more features of the first aspect of the invention or its embodiments, or vice versa.

[0034] According to a third aspect of the invention there is provided a method of improving hydrocarbon reservoir monitoring data, the method comprising: providing tracer test data associated with a reservoir; obtaining tracer residence time distribution data for the reservoir; calculating at least one moment of the residence time distribution data from the measurement data set; generating a model of the reservoir; generating a modelled residence time distribution data set for the reservoir; calculating at least one moment of the residence time distribution data from the model data set; comparing the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0035] The method may comprise comparing the at least one moment from the measurement data set and the at least one moment from the measurement data set to assess the accuracy of the model.

[0036] The method may comprise measuring the tracer residence time distribution data for a reservoir under a first set of parameters. The method may comprise generating a modelled residence time distribution data set for the first set of parameters of the reservoir.

[0037] Embodiments of the third aspect of the invention may include one or more features of the first or second aspects of the invention or their embodiments, or vice versa.

[0038] According to a fourth aspect of the invention there is provided a method of assessing a reservoir model comprising: providing tracer test data associated with a reservoir; obtaining tracer residence time distribution data for the reservoir; calculating at least one moment of the residence time distribution data from the measurement data set; generating a model of the reservoir wherein the model has a selected grid size; generating a modelled residence time distribution data set for the reservoir; calculating at least one moment of the residence time distribution data from the model data set; comparing the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set. The method may comprise comparing the at least one moment from the measurement data set and the at least one moment from the measurement data set to assess the accuracy of the model and the suitability of the of the model grid size.

[0039] The method may comprise adjusting the model grid size until a desired misfit threshold is reached. The method may comprise adjusting the model grid size increasing the model grid size or reducing the model grid size. The method may comprise selecting a model grid size based on the compared moment data. The method may comprise selecting a model grid size which maintains a desired accuracy threshold with the tracer test data based on the compared moment data.

[0040] The method may comprise calculating, estimating and / or determining the lowest model grid size resolution (most coarse size) that maintains a desired accuracy threshold with the tracer test data.

[0041] The at least one moment may be selected from a zero order moment, a first order moment, a second order moment. The at least one order moment may be more than two orders.

[0042] The measured residence time distribution data set may be obtained by injecting at least one tracer into a reservoir and monitoring the tracer concentration produced from the reservoir as a function of time. The measured residence time distribution data set may be obtained by injecting tracer into a reservoir and monitoring at least one tracer concentration produced from the reservoir as a function of time. The modelled residence time distribution data set may be obtained by modelling an injection of tracer into a model reservoir and modelling the tracer concentration produced from the reservoir as a function of time. The modelled residence time distribution data set may be obtained by modelling an injection of tracer into a model reservoir and modelling at least one tracer concentration produced from the reservoir as a function of time.

[0043] The at least one tracer may be an artificial tracer and / or a natural tracer. The at least one tracer may be selected from the group comprising chemical, fluorescent, phosphorescent, DNA, isotope, isotope signature, temperature, stable isotope, radioactive isotope, of elements constituting a part of a tracer molecule. The at least one tracer may comprise stable or radioactive isotopes of elements constituting a part of a tracer molecule. The tracer material may be a solid, liquid or gas. The at least one tracer may be a naturally occurring difference in the composition of a fluid. The at least one tracer may be a naturally occurring difference in the temperature of a fluid. The at least one tracer may be a natural geochemical. The at least one tracer may be a naturally occurring ion. The at least one tracer may be chemical difference in the composition of a fluid. The at least one tracer may be naturally occurring differences in the composition or temperature of fluids. The at least one natural tracer may be a naturally occurring difference in the composition of a fluid. The at least one natural tracer may be a naturally occurring difference in the temperature of a fluid. The at least one natural tracer may be a natural geochemical. The at least one natural tracer may be a naturally occurring ion. The at least one natural tracer may be chemical difference in the composition of a fluid. The at least one natural tracer may be naturally occurring differences in the composition or temperature of fluids.

[0044] The measured residence time distribution data set may be obtained by use of naturally occurring differences in fluid(s) compositions or temperature as “natural tracers”. The measured residence time distribution data set may be obtained by injecting natural tracer into a reservoir and / or monitoring at least one natural tracer produced from the reservoir as a function of time. The modelled residence time distribution data set may be obtained by modelling an injection of natural tracer into a model reservoir and / or modelling at least one natural tracer concentration produced from the reservoir as a function of time.

[0045] The measured residence time distribution data set may be obtained by identification of at least one variation of a natural tracer in a reservoir and / or monitoring at least one natural tracer produced from the reservoir as a function of time. The modelled residence time distribution data set may be obtained by modelling movement of at least one natural tracer introduced into a model reservoir and / or modelling at least one variation of a natural tracer produced from the reservoir as a function of time.

[0046] The method may comprise creating a tracer curve from the measured tracer concentration as a function of time. The method may comprise creating a tracer curve from measured tracer concentrations produced as a function of time. The method may comprise calculating an area under the tracer curve to calculate the at least one moment of measured tracer data set. The method may comprise creating a tracer curve from the modelled tracer concentration as a function of time. The method may comprise calculating an area under the tracer curve to calculate the at least one moment of modelled tracer data set.

[0047] The method may comprise calculating a swept volume of a tracer in the reservoir from the first order moment of the tracer data set. The method may comprise calculating a modelled swept volume of a modelled tracer in the model reservoir from the first order moment of the modelled tracer data set. The method may comprise comparing a first order moment of the modelled tracer data set with a first order moment of the measured tracer data set to assess the accuracy of the swept volume model at the selected grid size.

[0048] Embodiments of the fourth aspect of the invention may include one or more of any of features of the first to third aspects of the invention or their embodiments, or vice versa.

[0049] According to a fifth aspect of the invention there is provided a system for assessing and / or tuning a reservoir model comprising: at least one processor configured to generate a model tracer data set and calculate at least one moment of the model tracer data from the model tracer data set; wherein the processor is configured to calculate at least one moment of the measured tracer data from a measured tracer data set; and wherein the processor is configured to compare the at least one moment of the measured tracer data from the measurement data set and the at least one moment of the model data from the model tracer data set.

[0050] The at least one processor may be configured to determine a comparison value between the at least one moment of the measured tracer data from the measurement data set and the at least one moment of the model data from the model tracer data set.

[0051] The system may assess and / or tune a reservoir model based on the comparison of the at least one moment of the measured tracer data from the measurement data set and the at least one moment of the model data from the model tracer data set. The system may comprise assesses and / or tune a reservoir model based on the comparison value.

[0052] The at least one processor may be configured to perform at least one tracer flow simulation to generate a model tracer data set. The at least one processor may be configured to calculate at least one moment of the residence time distribution data from the model tracer data set. The at least one processor may be configured to calculate at least one moment of the residence time distribution data from the measured tracer data set.

[0053] The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range.

[0054] The model may be used to optimise or control production. The model may be used to characterise the reservoir. The at least one processor may be configured to perform injection and / or production optimisation tasks using the tuned model. The injection and / or production optimisation tasks may include predicting optimum injection rates for one or more injectors. The production optimisation tasks may include predicting optimum fluid production rates for one or more producers. The at least one processor may be a computer processor.

[0055] Embodiments of the fifth aspect of the invention may include one or more of any of features of the first to fourth aspects of the invention or their embodiments, or vice versa.

[0056] According to a sixth aspect of the invention there is provided a system for assessing and / or tuning a reservoir model comprising: at least one processor configured to generate a model tracer data set and calculate at least one moment of the residence time distribution data from the model tracer data set; wherein the processor is configured to calculate at least one moment of the residence time distribution data from a measured tracer data set; and wherein the processor is configured to compare the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0057] The at least one processor may be configured to determine a comparison value between the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0058] The system may assess and / or tune a reservoir model based on the comparison of the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0059] The system may comprise assesses and / or tune a reservoir model based on the comparison value.

[0060] The at least one processor may be configured to perform at least one tracer flow simulation to generate a model tracer data set and / or calculate at least one moment of the residence time distribution data from the model tracer data set.

[0061] The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range.

[0062] The model may be used to optimise or control production. The model may be used to characterise the reservoir. The at least one processor may be configured to perform injection and / or production optimisation tasks using the tuned model. The injection and / or production optimisation tasks may include predicting optimum injection rates for one or more injectors. The production optimisation tasks may include predicting optimum fluid production rates for one or more producers.

[0063] Embodiments of the sixth aspect of the invention may include one or more of any of features of the first to fifth aspects of the invention or their embodiments, or vice versa.

[0064] According to a seventh aspect of the invention there is provided a computer-readable medium, comprising statements, instructions and / or code configured to direct at least one processor to: generate a model tracer data set and calculate a one moment of the model tracer data from the model tracer data set; calculate one moment of the measured tracer data from a measured tracer data set; and compare the at least one moment of the measured tracer data from the measurement data set with the at least one moment of the model data from the model tracer data set.

[0065] The at least one processor may be configured to determine a comparison value of the at least one moment of the measured tracer data from the measurement data set compared to the at least one moment of the model data from the model tracer data set.

[0066] The comparison value may be a misfit value. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is reduced.

[0067] Embodiments of the seventh aspect of the invention may include one or more of any of features of the first to sixth aspects of the invention or their embodiments, or vice versa.

[0068] According to an eighth aspect of the invention there is provided a computer-readable medium, comprising statements, instructions and / or code configured to direct at least one processor to: generate a model tracer data set and calculate at least one moment of the residence time distribution data from the model tracer data set; calculate at least one moment of the residence time distribution data from a measured tracer data set; and compare the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0069] The at least one processor may be configured to determine a comparison value between the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set. The comparison value may be a misfit value. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range. The at least one processor may be configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is reduced.

[0070] Embodiments of the eighth aspect of the invention may include one or more of any of features of the first to seventh aspects of the invention or their embodiments, or vice versa.

[0071] According to a ninth aspect of the invention there is provided a method of assessing phase-related information of a reservoir model comprising: obtaining tracer test data associated with phases of the reservoir; calculating, estimating and / or determining a model tracer data set for the reservoir based on a first set of parameters values; calculating at least one moment of model tracer data set; calculating at least one moment of measured tracer data set from the tracer test data; comparing the at least one moment of the measurement data set and the at least one moment of the modelled data set to assess the accuracy of the model.

[0072] The method may comprise tracer test data associated with fluid or solid phases of the reservoir. The method may comprise tracer test data associated with fluid or solid phases of the reservoir.

[0073] The method may comprise calculating at least one spatial or temporal moment of model tracer data set. The method may comprise calculating at least one spatial or temporal moment of measured tracer data set from the tracer test data.

[0074] The method may comprise injecting at least one phase-partitioning tracer with at least one passive tracer into the reservoir. The method may comprise injecting at least one partitioning tracer with at least one passive tracer concurrently.

[0075] The method may comprise calculating at least one moment of measured retention time of the at least one partitioning tracer and at least one passive tracer. The method may comprise calculating at least one moment of model retention time of the at least one partitioning tracer and at least one passive tracer.

[0076] Embodiments of the ninth aspect of the invention may include one or more of any of features of the first to eighth aspects of the invention or their embodiments, or vice versa.

[0077] According to a tenth aspect of the invention there is provided a method of assessing a single well chemical tracer test model comprising: obtaining tracer test data associated with the reservoir; calculating, estimating and / or determining a model tracer data set for the reservoir based on a first set of parameters values; calculating at least one moment of model tracer data set; calculating at least one moment of measured tracer data set from the tracer test data; comparing the at least one moment of the measurement data set and the at least one moment of the modelled data set to assess the accuracy of the model.

[0078] The method may comprise injecting at least one partitioning tracer with at least one passive tracer into the reservoir. The method may comprise injecting at least one reactive partitioning tracer with at least one passive tracer into the reservoir. The method may comprise injecting at least one partitioning tracer with at least one passive tracer concurrently.

[0079] The method may comprise calculating at least one moment of measured retention time of the at least one partitioning tracer and at least one passive tracer. The method may comprise calculating at least one moment of model retention time of the at least one partitioning tracer and at least one passive tracer.

[0080] The method may comprise analysing the tuned model to determine at least one characteristic of the reservoir including oil saturation and / or relative permeability.

[0081] Embodiments of the tenth aspect of the invention may include one or more of any of features of the first to ninth aspects of the invention or their embodiments, or vice versa.

[0082] According to an eleventh aspect of the invention there is provided a method of assessing a reservoir model comprising: calculating at least one moment of the residence time distribution data from tracer measurement data set; calculating at least one moment of a residence time distribution data from a model data set; comparing the at least one moment of the residence time distribution data from the measurement data set with the at least one moment of the residence time distribution data from the model data set to assess the accuracy of the model.

[0083] Embodiments of the eleventh aspect of the invention may include one or more of any of features of the first to tenth aspects of the invention or their embodiments, or vice versa.

[0084] According to a twelfth aspect of the invention there is provided a method of monitoring a reservoir, the method comprising: injecting at least one tracer into a reservoir; detecting concentrations of the at least one tracer in fluid produced from the reservoir; constructing a model of the reservoir and assessing the model according to any of the first, second, third, fourth, ninth, tenth or eleventh aspects of the invention.

[0085] The reservoir may be subterranean reservoir. The reservoir may be a hydrocarbon reservoir.

[0086] Embodiments of the twelfth aspect of the invention may include one or more of any of features of the first to eleventh aspects of the invention or their embodiments, or vice versa.

[0087] According to a thirteenth aspect of the invention there is provided a system for monitoring a hydrocarbon reservoir, the system comprising: at least one tracer; an injection device configured to inject the at least one tracer into an injection well and / or reservoir; a sampling device configured to sample produced fluid; at least one processor configured to generate a model tracer data set and calculate at least one moment of the residence time distribution data from the model tracer data set; wherein the processor is configured to calculate at least one moment of the residence time distribution data from a measured tracer data set; and wherein the processor is configured to compare the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0088] The at least one processor may be configured to determine a comparison value between the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

[0089] The sampling device may be configured to collected samples of the produced fluid at known sampling times. The sampling device may be configured to collected samples for later analysis. The sampling device may be configured to detect and / or measure the concentration tracer in real time. The sampling device may be configured to generate a measured tracer concentration data set.

[0090] Embodiments of the thirteenth aspect of the invention may include one or more of any of features of the first to twelfth aspects of the invention or their embodiments, or vice versa.

[0091] Brief description of the drawings

[0092] There will now be described, by way of example only, various embodiments of the invention with reference to the following drawings (like reference numerals referring to like features) in which:

[0093] Figure 1 is a simplified sectional diagram through an injector well, reservoir and a producer well in accordance with an aspect of the invention;

[0094] Figure 2A is a simplified diagram of illustrating the concept of residence time distribution of a tracer transported from an injector well through a reservoir to a producer.

[0095] Figure 2B is a graph of a residence time distribution showing the number of particles as a function of time for the tracer residence time distribution of Figure 2A; Figure 3A is a simplified diagram of illustrating the concept of residence time distribution of a tracer transported from an injector well through a reservoir with high heterogeneity to a producer.

[0096] Figure 3B is a graph of a residence time distribution showing the number of particles as a function of time for the tracer residence time distribution of Figure 3A.

[0097] Figure 4A is a tracer concentration curve over time showing model response curves and measured data where the zero order moment from the RTD is a function of the model.

[0098] Figure 4B is a tracer concentration curve over time showing model response curves and measured data where the zero order moment from the RTD is a function of the measured data.

[0099] Figure 5A is a representation of the connection magnitude between the injector and the producer wells based on the zero and first order moments from the RTD as a function of the model data.

[0100] Figure 5B is a representation of the connection magnitude between the injector and the producer wells based on the zero and first order moments from the RTD as a function of the measured data.

[0101] Figure 6 is a tracer concentration curve over time showing model response curves using a fine grid size and a coarse grid size.

[0102] Detailed of preferred embodiments

[0103] Figure 1 shows a schematic sectional view of an injector-producer system 10 which comprise an injector well 12 and a producer well 14 in communication with a reservoir 16. In this example the tracer release apparatus 18 is arranged, installed or positioned at surface at the injector well 12. It will be appreciated that a tracer release apparatus may alternately or additionally be arranged, installed or positioned at least one location downhole.

[0104] In use, the tracer release apparatus 18 is actuated to inject tracer into the reservoir. The released tracer migrates through the reservoir and is carried with the fluids produced in one or more producer wells. One producer well is shown Figure 1. The tracer type and concentration may be measured in the producer well or at the surface of the producer well and the time delay between tracer release in the injector well and arrival and / or the concentration levels measured at the producer well may provide well flow conditions and / or volume in the injector well, reservoir and / or producer well. Analysis may also provide information on the inflow profile in the producer well. The movement of tracer through the reservoir may be modelled and measured tracer data may be compared with simulated data to improve the model.

[0105] Residence Time Distribution (RTD) is the distribution of times used by a population of tracer particles to travel through a medium. The tracers represent elements of fluid that travel through different paths, and that therefore use different amounts of time to pass through a medium. Interpretation of tracer data by use of residence time distribution (RTD) analysis may also be used to quantify sweep volume and the magnitude of connections in oil and gas reservoirs.

[0106] RTD analysis involves injecting known tracers into a reservoir and transforming the tracer concentration time series, measured at production wells to a distribution of residence times for the tracers. As an example water tracers which passively follow the flow of water can be used to quantify the significance of individual injector-producer connections as well as the pore volume swept by water from the injector to producers. The injected water tracer tags the injected water and follows this as it travels through different paths in the reservoir. It will be appreciated that in other applications oil tracers may be used to trace oil flow and / or gas tracer may be used to trace gas flow.

[0107] Figures 2A and 3A are illustrations of two different examples showing water tracer migration through two different reservoirs. As shown in Figure 2A and 2B water takes different paths from the injector 112, 112a, through the reservoir 116, 116a to the producer 114, 114a shown as arrows “A” in Figure 2A and arrows “B” in Figure 3A respectively. The time it takes to arrive at a given producer taking a specific path is denoted as the residence time for that particular path. For a collection of tracer particles injected, the different paths taken yields a distribution of residence times. The distribution of residence times is closely related to a concentration curve at a given producer. Figure 2B and 3B are simple graphs 120, 120a showing the distribution of residence times based on the time it takes for the particles to pass through the reservoir. If the water is traced with a water tracer the distribution of residence times can be quantified by multiplying the sampled concentration (mass / volume) and volume produced during the sampling period (rate multiplied by sample period). The residence time distribution is affected by the flow - if a heterogeneity is present in the well as shown in Figure 3A it may enhance fluid velocity in part of the reservoir as shown as “C” in Figure 3A. As a result the distribution will be shifted towards smaller residence times. This will yield a smaller average residence time.

[0108] The distribution, E(t), of these times is called the residence time distribution, of the fluid in the system. E(t) is defined from produced tracer concentrations, C(t), production rate, and injected tracer amount, M, as:

[0109] E(t)=C(t) <2p(t) / M (1)

[0110] The unit of E is the inverse of the time unit. Important information about the geometry and flow in a system can be obtained from moments of the residence time distribution. They are defined as

[0111] The zero-order temporal moment of the residence time distributions (m0= f E(t)dt) give the fraction of water, originating from a given injector, produced in the producer. The first order moment m1= f t - E(t)dt is closely related to average residence time «t) =

[0112] The swept pore volumes can be determined once the average residence time and the fraction of water produced from the injector are known. For example the fraction m0corresponds with the amount injected fluid that migrates to a given producer. If m0= 0.15, this means that 15% of the injected water travels towards the producer. The time it takes for the water to arrive is given from (t) this corresponds to the time it takes to “fill” the pore space with this water. Since m0give the fraction of water towards the producer, we know that Qt ■ m0is the rate at which this volume fills up which correspond to a swept pore volume: = Qi •mo • (t).

[0113] The zero moment represents the relative amount of tracer produced in production well and the first moment represents the average residence time for the tracers between the injection well and a producer. In a modelling system a time series of measured data is compared with a time series of simulated data. A misfit (squared difference between measured and simulated time series) is calculated. By systematically performing simulations where reservoir parameters are guessed and re-guessed this misfit may be reduced.

[0114] The misfit may be defined as: where f(t, a) is the model function of time defined by model parameters a, and ytare the measured response at times ti.

[0115] This misfit definition can be generalized to multiple spatial dimensions as well as time and f (t) can be a multidimensional function of space and time. The values of f at specific points in space and time depend on the set of parameters a and history matching is the process of finding the values of a that gives the best possible representation of the measured data.

[0116] In practice, this is done by minimizing the misfit with respect to the model parameters, i.e. by solving the equation:

[0117] However, by calculating the moments of the RTD the best representation of the reservoir information contained in the tracer curves may be obtained. A misfit that compares the RTD moments of the simulated and measured tracer curve instead of the curves themselves provides a more accurate comparison of the data.

[0118] This is achieved by calculating the moments m0, m1, m2.. from the model using Eq. (2) and comparing to the moments calculated using the measured data.

[0119] Figure 4A and 4B each show tracer concentration curves over time for the same tracer data. Each curve has five measured data points (circles) and a model curve response matched to the data for a given set of parameters (dotted line). Figure 4A is an example of a model approach where the misfit is calculated as the summed squared difference between the measured data and the model response. Error uncertainty between the model and measured data is illustrated with vertical lines in Figure 4A and a matching process using a standard approach would be defined by the effort to reduce these distances to a value as close to zero as possible. The grey areas under the model curve in Figure 4A corresponds to the zero order moment from the model (m0).

[0120] In Figure 4B the misfit is defined as the difference between the moments of the residence time distributions calculated from the model response vs. the measured data. The grey areas under the model curve in Figure 4B corresponds to the zero order moment from the measured data

[0121] Comparing these two representations shown in Figures 4A and 4B enables an assessment of the quality of the fit between model data and measured data using the zero-order moment of the RTD. In the same manner the first order moment, second order moment etc. can be included in the matching process.

[0122] The tracer data can be analysed to assess the volumetric sweep to quantify the amount of fluid flowing from injectors to producers. It gives an indication of magnitude of the injector producer connections. The first order moment from a model can be compared to the first order moment from tracer data. These first order moments can be used to find swept volume from an injector to a producer from tracer data and model data, which in turn can be illustrated as a volume or area of a given size.

[0123] The magnitude of the injector producer connections may be quantified by the zero-order moment calculated from measured tracer data and calculated from simulated tracer response.

[0124] Figure 5A and 5B are representations of the zero and first order moments for model (200, 200a) and measured data (200a) for tracer flow from an injector (212, 212a) to a producer (214, 214a). The representations (200, 200a) comprise an arrow (A, A’) and a surrounding elliptical area (E, E’). The width of each arrow represents the magnitude of injectorproducer connection by the zero-order moment and the size of the elliptical area represents the value of the swept area from the first order moment. By comparing the representations of these moments from a model shown in Figure 5A and the moments obtained from measured data in Figure 5B the misfit between the model and measured data is shown. Although the representation shown in Figures 5A and 5B are shown as 2- Dimensional representations it will be appreciated that the reservoir data may be represented in a 3 Dimensional representation.

[0125] By analysing a misfit that compares the RTD moments of the simulated and measured data, numerical smearing may be mitigated. A misfit that compares the RTD moments of the simulated and measured data may be less impacted by numerical error and inconsistency. Numerical smearing in computer models originates from the discretization of space and time and is a practical problem for all numerical reservoir simulation models.

[0126] Tracer curves that are significantly stretched for coarse grids vs. finer grids may result in numerical smearing. The effect is purely numerical, i.e. for the same set of parameters the simulated result are significantly distorted depending on an arbitrarily set discretization size.

[0127] Figure 6 shows a model curve based on an injected tracer amount of 10kg and a production rate of 300m3 / day. Figure 6 shows a model data curve plotted on a fine grid shown as a dotted line and model data curve plotted on a coarse grid shown as a continuous line. The coarse grid model is identical to the fine grid model, except for grid size. The measured data points as shown as dots. A misfit based on a first comparison approach given by Eq. (4) would be significantly different for the coarse grid than for the fine grid. However, using a misfit based on the difference of the moments of the RTD, the fine grid and coarse grid would both be considered to be close to the data, unaffected by the artificial grid size parameter. The location of the zero and first order moments from the RTD of the fine grid and coarse grid results are much less affected by the impact of the grid size.

[0128] Specifically, for the results given from the fine grid and coarse grid models in Figure 6 the zero-order moments are m0= 0.21 for both the fine grid and coarse grid model curves. This was calculated by transforming the concentration curves from the fine and coarse grid models to residence time distributions by multiplying the concentration and the production rate (Q = 300 m3 / day) and dividing by the injected tracer amount (M = 10 kg). In a similar manner (using the general formula defining the moments, Eq. 2) the first order moment is given as m}= 0.53 months for the fine grid model and as m}= 0.67 months for the coarse grid model. In terms of average residence time «T) = nto / we find 2.5 months and 3 months for the fine grid model and coarse grid model, respectively. This is seen to correspond well with the average time spent by the tracer particles by visual inspection of Figure 6.

[0129] The five measured concentration values plotted as points in the Figure 6 are given in Table 1 , together with a calculation of the corresponding zero and first order moments from the data points. These moments are found using the formulas given in Table 1.

[0130] The contribution from each datapoint is given in column 5 and 6 of Table 1 and the summation of these columns is given in the last line of the table. The zero-order moment from the data is thus mo' = 0. 21 and the first order moment is = 0.49.

[0131] Table 1. Measured data points and calculation of zero and first order moments from the datapoints.

[0132] A first observation from the calculation of the moments using the two model curves is that the difference between the fine and coarse models is negligible for the zero-order moment and moderate (30%) for the first order moment.

[0133] Similarly, the difference between the moments obtained from the measured data and from the two models is negligible for the zero-order moment. For the first order moment the difference between data and model is about 8% for the fine-grid model and about 30% for the coarse-grid model. The conclusion from an assisted history matching process comparing the at least one moment from the measurement data set and the at least one moment from the model data set is that coarse grid model is a good model which successfully reproduces the main features of the data.

[0134] In comparison, using a conventional misfit approach, defined through Equation (3). By calculation of the squared differences between the measured data and each of the models we find an average difference in misfit of 75% between the measured data and each of the models. This leads to the conclusion that the fine grid model has a misfit that is 60% smaller than the coarse model which would suggest that the fine grid model should be used. The simulation time increase by use of fine models vs. coarse models is typically significant. Reaching a correct model conclusion from a coarse grid model rather than a fine grid model is thus a significant advantage. By providing an assisted history matching process of comparing the at least one moment from the measurement data set and the at least one moment from the model data set provides an improved assessment of the suitability of each of the grid models.

[0135] The assisted history matching process may also be beneficial in calculating, estimating and / or determining the best possible values of the retention times of partitioning and passive tracer in a partitioning inter-well tracer test (PITT), where a partitioning tracer is injected co-currently with an ideal tracer.

[0136] For negligible oil flow rates, compared to the water flow rates, the oil saturation is given by:

[0137] S = _tp~tw_

[0138] ° tp+ tw(K - l)Wwhere tpand tware retention times for the partitioning and passive tracers, respectively, and K = colcwis the oil / water partition coefficient of the partitioning tracer, that can be determined in in lab experiments. Using this methodology requires determination of values of tpand tw, typically using analytical or simulation models. Comparing the at least one moment from the measurement data set and the at least one moment from the model data set provides an improved assessment of the best possible values of the retention times of the partitioning and passive tracer in a PITT. The assisted history matching process may also be beneficial in a single well chemical tracer test (SWCTT). In these tests several tracers are injected - one of which is a reactive partitioning tracer than forms a secondary non-partitioning tracer during a shut-in period, following the initial injection into a near well volume. A common approach for SWCTTs is to compare measured tracer data to model curves in order find physical parameters such as the oil saturation or relative permeability. The methodology described above would be suited for the single well chemical tracer test, in a manner similar to the PITT described above.

[0139] The invention provides a method of improving the accuracy of a reservoir model. The method comprises obtaining tracer test data associated with a reservoir and calculating, estimating and / or determining a model tracer data set for the reservoir based on a first set of parameters values. The method comprises calculating at least one moment of model tracer data set and calculating at least one moment of measured tracer data set from the tracer test data. The method also comprises comparing the at least one moment of the measurement data set and the at least one moment of the model data set to assess the accuracy of the model.

[0140] Throughout the specification, unless the context demands otherwise, the terms 'comprise' or 'include', or variations such as 'comprises' or 'comprising', 'includes' or 'including' will be understood to imply the inclusion of a stated integer or group of integers, but not the exclusion of any other integer or group of integers. Furthermore, relative terms such as “up”, “down”, “top”, “bottom”, “upper”, “lower”, “upward”, “downward”, “horizontal”, “vertical”, “and the like are used herein to indicate directions and locations as they apply to the appended drawings and will not be construed as limiting the invention and features thereof to particular arrangements or orientations.

[0141] The foregoing description of the invention has been presented for the purposes of illustration and description and is not intended to be exhaustive or to limit the invention to the precise form disclosed. The described embodiments were chosen and described in order to best explain the principles of the invention and its practical application to thereby enable others skilled in the art to best utilise the invention in various embodiments and with various modifications as are suited to the particular use contemplated. Therefore, further modifications or improvements may be incorporated without departing from the scope of the invention as defined by the appended claims.

Claims

Claims1. A method of improving the accuracy of a reservoir model comprising: obtaining tracer test data associated with the reservoir; constructing a model of the reservoir, determining a model tracer residence time distribution data set for the reservoir; calculating at least one moment of the model tracer residence time distribution data set; calculating at least one moment of residence time distribution data set from the tracer test data; comparing the at least one moment of the measurement data set and the at least one moment of the model data set to assess the accuracy of the model.

2. The method according to claim 1 comprising obtaining tracer test data from historical tracer injection and / or production data associated with production from the reservoir or performing at least one tracer test to generate tracer test data.

3. The method according to claim 1 or 2 comprising adjusting one or more parameter values in the model until a desired misfit threshold is reached.

4. The method according to any preceding claim comprising iterative tuning of the model until the model moment data best matches the measured moment data.

5. The method according to any preceding claim wherein the tracer test data comprises data from one or more tracers.

6. The method according to any preceding claim wherein the measured tracer data set is a tracer concentration as a function of time and / or space and the model tracer data set is a tracer concentration as a function of time and / or space.

7. The method according to any preceding claim wherein the reservoir model is an interwell model, a partitioning inter-well tracer test model and / or a single well chemical tracer test model.

8. The method according to any preceding claim wherein the method is a computer- implemented history matching method.

9. The method according to any preceding claim wherein the at least one moment is selected from a zero order moment, a first order moment and / or a second order moment.

10. The method according to any preceding claim comprising obtaining the measured residence time distribution data set by injecting tracer into a reservoir and monitoring the tracer concentration produced from the reservoir as a function of time and / or space and obtaining the modelled residence time distribution data set by modelling an injection of tracer into a model reservoir and modelling the tracer concentration produced from the reservoir as a function of time and / or space.

11. The method according to any preceding claim comprising calculating the at least one moment of measured tracer data set from an area under a tracer curve of measured tracer concentration as a function of time.

12. The method according to any preceding claim comprising calculating the at least one moment of modelled tracer data set from an area under a tracer curve of modelled tracer concentration as a function of time.

13. The method according to any preceding claim comprising comparing the at least one moment of the measurement data set and the at least one moment of the measurement data set to assess the accuracy of the model.

14. The method according to any preceding claim comprising calculating a swept volume of a tracer in the reservoir from the first order moment of the tracer data set and calculating a modelled swept volume of a modelled tracer in the model reservoir from the first order moment of the modelled tracer data set.

15. The method according to any preceding claim comprising comparing a first order moment of the modelled tracer data set with a first order moment of the measured tracer data set to assess the accuracy of the swept volume model.

16. The method according to any preceding claim comprising characterizing and / or controlling the reservoir based on the model.

17. The method according to any preceding claim comprising analysing the tuned model to determine at least one characteristic of the reservoir selected from the group comprising residence time, spatial distribution of tracer mass, oil saturation, relative permeability, porosity, flow paths, flow rate, water breakthrough time.

18. The method according to any preceding claim comprising assessing the reservoir model comprising: generating a model of the reservoir wherein the model has a selected grid size; generating a modelled residence time distribution data set for the reservoir; calculating at least one moment of the residence time distribution data from the model data set; comparing the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

19. The method according to claim 18 comprising comparing the at least one moment from the measurement data set and the at least one moment from the measurement data set to assess the accuracy of the model and the suitability of the of the model grid size.

20. The method according to claim 18 or 19 comprising adjusting the model grid size until a desired misfit threshold is reached.

21. The method according to claim 20 comprising adjusting the model grid size by increasing the model grid size or reducing the model grid size.

22. A system for assessing and / or tuning a reservoir model comprising: at least one processor configured to generate a model tracer data set and calculate at least one moment of the residence time distribution data from the model tracer data set; wherein the processor is configured to calculate at least one moment of the residence time distribution data from a measured tracer data set; and wherein the processor is configured to determine a comparison value between the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

23. The system according to claim 22 wherein the at least one processor is configured to perform at least one tracer flow simulation to generate a model tracer data set and / or calculate at least one moment of the residence time distribution data from the model tracer data set.

24. The system according to claim 22 or claim 23 wherein the at least one processor is configured to perform a history matching procedure to iteratively adjust at least one parameter and / or at least one parameter value in the model until the resulting comparison value is within a target range.

25. A computer-readable medium, comprising statements, instructions and / or code configured to direct at least one processor to: generate a model tracer data set and calculate at least one moment of the residence time distribution data from the model tracer data set; calculate at least one moment of the residence time distribution data from a measured tracer data set; and determine a comparison value between the at least one moment of the residence time distribution data from the measurement data set and the at least one moment of the residence time distribution data from the model data set.

Images