WO2019216870A1

WO2019216870A1 - Reducing uncertainties in petrophysical measurements for reserves evaluations

Info

Publication number: WO2019216870A1
Application number: PCT/US2018/031349
Authority: WO
Inventors: Sandeep Mukherjee; Bhaskar Bikash SARMAH; Syed Muhammad Farrukh HAMZA
Original assignee: Halliburton Energy Services, Inc.
Priority date: 2018-05-07
Filing date: 2018-05-07
Publication date: 2019-11-14

Abstract

A method for generating a reconstructed log based on a formation resistivity and a geochemical property includes lowering a well tool into a wellbore, the well tool comprising a first pair of electrodes used to generate a macroresistivity measurement and a second pair of electrodes used to generate a microresistivity measurement. The method also includes receiving the geochemical property of a subsurface formation around the wellbore, generating a compositional range based on the geochemical property, and determining the formation resistivity based on the macroresistivity measurement and the microresistivity measurement. The method also includes generating the reconstructed log, the reconstructed log comprising at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

Description

REDUCING UNCERTAINTIES IN PETROPHYSICAU MEASUREMENTS FOR

RESERVES EVACUATIONS

BACKGROUND

[0001] The disclosure generally relates to the field of measuring petrophysical properties of a formation, and more particularly to reducing uncertainties in petrophysical measurements.

[0002] Reservoir potentials, such as the volume of hydrocarbons, as predicted by a reserves estimation, and their associated confidence levels (e.g., 90%, 50%, 10%, etc.) depends on formation properties and their corresponding probability distributions. Such formation properties can include the height of a hydrocarbon zone, oil-water-gas saturations, specific hydrocarbon saturations, net-to-gross ratio, porosity, etc. Formation logs (i.e., sets of

measurements/calculations and their corresponding depth of measurement) provide data for determining these formation properties and the probability distributions associated with these factors. Such formation logs can include resistivity logs, neutron logs, and various other information provided by tools lowered into a wellbore.

[0003] Well logs can have significant measurement uncertainty at troubled zones due to unfavorable measuring conditions at the troubled zones. These uncertainties result in erroneous calculations/estimations for determining formation properties, their corresponding probability distributions, and amounts of hydrocarbons in a formation. Increasing the accuracy of the formation data can reduce the error of the formation measurements and their associated probability distribution functions. Doing so will provide more accurate information useful for determining zones for placements of horizontal wellbores, an optimal well treatment operation, an amount of hydrocarbon expected to be extracted from the formation, whether or not to proceed with a drilling operation, etc.

BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Embodiments of the disclosure may be better understood by referencing the accompanying drawings.

[0005] FIG. 1 depicts a diagram of a wireline well logging system and the underlying formation, according to some embodiments.

[0006] FIG. 2 depicts a flowchart of operations to determine a blended resistivity log, according to some embodiments.

[0007] FIGS. 3-4 depicts flowcharts of operations to determine formation properties and hydrocarbon amounts, according to some embodiments.

[0008] FIG. 5 depicts a flowchart of operations to determine formation properties and hydrocarbon amounts using an artificial neural network, according to some embodiments. [0009] FIG. 6 depicts example zones of a resistivity log, according to some embodiments.

[0010] FIG. 7 depicts example zones of a geochemical log, according to some embodiments.

[0011] FIG. 8 depicts example zones of a set of reconstructed logs, according to some embodiments.

[0012] FIG. 9 depicts a diagram of an artificial neural network being used to determine formation properties, according to some embodiments.

[0013] FIG. 10 depicts an example drilling system, according to some embodiments.

[0014] FIG. 11 depicts an example computer system with a formation property predictor, according to some embodiments.

DESCRIPTION

[0015] The description that follows includes example systems, methods, techniques, and program flows that embody embodiments of the disclosure. However, it is understood that this disclosure may be practiced without these specific details. For instance, this disclosure refers to determining a reconstructed neutron porosity log in illustrative examples. Aspects of this disclosure can be also applied to other logs such as a reconstructed sonic log. In other instances, well-known instruction instances, protocols, structures and techniques have not been shown in detail in order not to obfuscate the description.

[0016] Various embodiments minimize uncertainty in the petrophysical measurements to improve reserves estimations. Conventional reserve estimations generally do not account for the uncertainty of measurements and the errors that are imposed on the absolute estimation of subsurface hydrocarbon reserves. For example, in cases where wellbore conditions are less than ideal to challenging or reservoir characterization is affected by thinly bed pays, the acquired petrophysical data may not ideally represent the subsurface reservoir conditions. The swings in absolute value of measurements and the deviations in calculated values of the reservoir properties in challenging wellbores as well as the linear volumetric summation of responses in zones of thinly laminated reservoir sections can result in erroneous probability distribution functions in deriving critical petrophysical parameters such as porosity, hydrocarbon saturation and original hydrocarbon in place.

[0017] Various embodiments include operations to determine formation properties (e.g., mechanical, chemical, petrophysical, etc.) and reserve estimates based on well logs (e.g., open- hole well logs, cased-hole well logs, etc.). During such operations, well logs such as resistivity logs and neutron porosity logs can be acquired and combined with their corresponding probability distributions to determine formation properties. Additionally, some embodiments can include measurements to determine formation composition and/or predict water saturation based on geochemical and electrical properties. Geochemical properties and water saturation predictions can increase accuracy when determining formation properties, reserve estimates, and their corresponding probability distributions. Such embodiments provide a means of evaluating a reservoir and making operational decisions on drilling location, well treatments options, production rate, etc. For example, formation properties such as porosity and saturation can be used to predict the amount of oil and gas in a reservoir and determine an optimal drilling location. If one or more geochemical properties and/or microresistivity measurements are used, formation properties and their corresponding probability distributions can be determined with increased accuracy and reduced bias. This greater accuracy and reduced bias can likewise increase the accuracy of reserve estimates.

[0018] Some embodiments include operations for generating blended resistivity logs using microresistivity logs and macroresistivity logs. A blended resistivity log can be partitioned into one or more high-accuracy zones and one or more target zones based on exceeding a threshold or changing in disproportionate variance with other logs. In some embodiments, a target zone is a continuous length in the borehole. Operations can enhance the accuracy of measurements of a target zone by using one or more values in a reconstructed log in the target zone, where a reconstructed log is a well log that has been generated based on two or more types of measurement.

[0019] Various embodiments can include using one or more geochemical properties from the well to generate a compositional range of the formation. Geochemical properties can include chemical information such as the presence or quantity of materials (e.g., atomic elements, molecules, compounds, minerals, phases of matter, etc.). For example, based on geochemical analysis that shows the presence of quartz, carbonates, and shales, a three-dimensional compositional range of chemical combinations including quartz, carbonates, and shales can be generated. The use of the geochemical properties or their corresponding compositional ranges increases the accuracy of reconstructed logs.

[0020] Various embodiments include generating a reconstructed density log using the blended resistivity logs, one or more geochemical properties and fluid saturation ranges. The reconstructed density log can be used to generate a reconstructed neutron porosity log. The reconstructed density and the reconstructed neutron porosity log can be used to generate a model for formation properties. In some embodiments, a statistical and/or artificial neural network training operation can train the model using the high-accuracy zones of a set of well logs to determine formation properties for both high-accuracy zones and target zones. The trained model can be used to determine formation properties for the target zones. These trained models can provide both a prediction for the values of formation properties as well as their corresponding probability distributions. The formation properties as well as their corresponding probability distributions can be used to control well operations. The formation properties as well as their corresponding probability distributions can also be used to determine the volume of the formation (Gross Rock Volume) as well as specific reserve estimates for a formation. Such reserve estimates can include specific confidence levels for the minimum hydrocarbon in place in a formation, such as a 90% confidence level (P90), 50% confidence level (P50), and/or 10% confidence level (P10).

Example Well System

[0021] FIG. 1 depicts a diagram of a wireline well logging system and the underlying formation, according to some embodiments. In particular, FIG. 1 depicts a wireline logging system 100, where a tool or probe that includes a "transmit coil" and one or more "receive coils" is fed through a structure to measure electromagnetic properties such as resistivity and conductivity. Alternatively, the transmit coil can be used as a receive coil in other system configurations. Additionally, the wireline logging tool can include equipment to measure neutron porosity, acoustic properties, and other properties of the formation and/or fluid in the formation.

[0022] As shown, the wireline logging system 100 includes a well tool 102 disposed in a wellbore 104 which penetrates a formation 106. The wellbore 104 has a wellbore wall 120 that can be exposed to the well tool 102. The well tool 102 is connected by a wireline cable 112 to a well-logging truck 108 located at the rig site. The wireline cable 112 provides support and/or power to the well tool 102 and is used to transmit measurements taken from the device to the well-logging truck 108. The well-logging truck 108 contains a computer 110 that receives the measurements, preferably stores the measurements, and uses the measurements to evaluate the formation 106. The measurements can be used to determine formation properties such as permeability, porosity, oil saturation, net pay, etc.

[0023] The well tool 102 can include a first pair of electrodes that are separated by a distance greater than or equal to a defined threshold, wherein the first pair of electrodes are used to measures a first resistance measurement. The well tool can include a second pair of electrodes that are separated by a distance less than the defined threshold and/or less than the distance separating the first pair of electrodes. The defined threshold can be values less than 12 inches, such as 1 inch, 0.5 inches, 2 inches, etc. For example, if the defined distance is 6 inches, the first pair of electrodes can be separated by a distance of 12 inches and the second pair of electrodes can be separated by a distance of 0.2 inches. As another example, the distance separating the second pair of electrodes can be less or equal to 1 inch while the distance separating the first pair of electrodes can be greater than or equal to 12 inches. The well tool 102 is shown to include a single housing, wherein the well tool 102 includes both the first pair of electrodes and second pair of electrodes in the same housing. An alternative well tool can include a plurality of other tools, sub-tools, and/or components with their own housing. For example, the well tool can include a macroresistivity induction tool having its own housing and a microresistivity induction tool having its own housing, wherein the macroresistivity induction tool includes the first pair of electrodes and the microresistivity induction tool includes the second pair of electrodes. The second pair of electrodes can be part of a plurality of pairs of electrodes that are each less than the defined threshold apart to increase the accuracy of microresistivity measurements. For example, a microresistivity induction tool that is part of a well tool can include 6 pairs of electrodes that are each 0.3 inches apart.

[0024] The wireline logging system 100 provide support for the tool, as well as enabling communication between the processors on the surface outside the wellbore 104 and providing a power supply. The wireline logging system 100 can include any downhole conveyance such as wire, cable, e-line, slickline, braided line, metallic wire, non-metallic wire, or composite wire, single or multiple strands, as well as tubing, coiled tubing, joint tubing, pipe, or other tubular, combinations thereof, and the like. The wireline logging system 100 can include fiber optic cabling for carrying out communications. The optical cable can be provided internal or external of the wireline logging system 100. The wireline logging system 100 is sufficiently strong and flexible to tether the well tool 102 through the wellbore 104, while also permitting

communication through the wireline logging system 100 to surface tools unit, such as the well logging truck 108.

Example Operations

[0025] The flowcharts are provided to aid in understanding the illustrations and are not to be used to limit scope of the claims. The flowcharts depict example operations that can vary within the scope of the claims. Additional operations may be performed; fewer operations may be performed; the operations may be performed in parallel; and the operations may be performed in a different order. For example, the operations depicted in blocks 302 and 304 can be performed in parallel or concurrently. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by program code. The program code may be provided to a processor of a general-purpose computer, special purpose computer, or other programmable machine or apparatus.

[0026] FIG. 2 depicts a flowchart of operations to determine a blended resistivity log, according to some embodiments. Operations of the flowchart 200 begin at block 204.

[0027] At block 204, macroresistivity measurements are determined. The macroresistivity measurements can be determined based on the measurements made by a macroresistivity induction tool and can include resistivity measurements at resolutions greater than a defined threshold. The defined threshold can have the same length as the defined threshold described above with reference to Figure 1. For example, a macroresistivity log can have a resolution of 12 inches, wherein each measurement provides an estimated resistivity for a 12-inch length. In some embodiments, a plurality of macroresistivity measurements with their associated measurement depths can be combined to form a macroresistivity log.

[0028] At block 206, microresistivity measurements are determined. The microresistivity measurements can be determined based on the measurements of a microresistivity induction tool, and can provide resistivity measurements at resolutions less than or equal to the defined threshold (as described above). For example, a microresistivity log can have a resolution of 0.2 inches, wherein each measurement provides an estimated resistivity for a 0.2-inch length. In some embodiments, a plurality of microresistivity measurements with their associated measurement depths can be combined to form a microresistivity log.

[0029] At block 210, a blended resistivity log is generated based on the microresistivity measurements and the macroresistivity measurements. In some embodiments, the blended resistivity log can include the microresistivity measurements rescaled using the maximum and minimum of the macroresistivity measurements. Before rescaling, a microresistivity

measurement can be provided as instrument-specific values and can be rescaled based on the maximum and minimum values of a macroresistivity measurement. In some embodiments, the microresistivity measurements can be re-scaled using a polynomial fitting method. For example, a maximum and minimum values of a macroresistivity log can be 2000 ohms and 20 ohms respectively, and the maximum value, minimum value, and example value of a microresistivity log can be 100 ohms, 50 ohms, and 75 ohms, respectively. Using a square polynomial fitting method, the respective maximum, minimum, and example value of the rescaled log can be 2000 ohms, 20 ohms, and 515 ohms. Each of the re-scaled microresistivity measurements can be combined to form a blended resistivity log. [0030] FIGS. 3-4 depicts flowcharts of operations to determine formation properties and hydrocarbon amounts, according to some embodiments. Operations of the flowchart 300 begin at block 302.

[0031] At block 302, the geochemical properties and blended resistivity data are determined. With respect to FIG. 2, the blended resistivity data can be determined using the operations described above for blocks 204-210. While the following discloses receiving and using multiple geochemical properties, it should be understood that a single geochemical property can be received and used during log reconstruction. One or more geochemical properties can be determined based on one or more measurements received from a sensor in the well tool, cuttings analysis, or surface materials testing. Based on the presence/amount of the elements and other materials measured from the samples, a mineral identification operation or device can predict the identity and amount of the minerals present in the formation based on the geochemical properties. For example, dry mineral samples from the formation can be analyzed using visual analysis, microscopy, and gamma ray spectroscopy to determine compositional fractions of silica, aluminum, calcium, iron etc. As another example, a sensor in the well tool can test a borehole wall to determine the presence and compositional fraction of minerals.

[0032] At block 304, a compositional range of materials (compositional range) and formation composition are generated based on the geochemical properties. The compositional range of materials for a formation can include the number and proportional amounts of minerals, mineral phases, and other materials for the formation for a depth of the formation. The compositional range can be determined based on the geochemical properties. For example, the geochemical properties can suggest mineral phase distributions that includes predicting that the compositional range is 25-35% quartz, 40-60% ilite, and 10-30% calcite, wherein the center of each distribution is 30%, 50%, and 20%, respectively. In some embodiments, the formation composition can be based on the center of each distribution. Alternatively, the formation composition can be assigned to be other possible compositions in the compositional range of materials. For example, the formation composition can be 25% quartz, 60% ilite, and 15% calcite. In some embodiments, the compositional range of different depths can be combined to generate a log of compositional ranges.

[0033] At block 306, a first reconstructed density log for high-accuracy zones is determined based on the formation composition. In some embodiments, high-accuracy zones can be determined as depth ranges in the wellbore based on a reconstructed fluid saturation value. The reconstructed fluid saturation can be determined based on geochemical properties and resistivity data. For example, zones with low estimated clay content (e.g., less than 50% by volume or by weight) and having resistivity values in the lesser range of the resistivity distribution (e.g., in the lesser half, lesser third, lesser 10%, etc.) can be determined to have a very high water saturation, wherein a very high water saturation can be water saturation greater than 0.8. The high-accuracy zones can be determined to be based on the zones with very high water saturation. A first reconstructed density can be determined using Equation 1, where Pbuik is ^a reconstructed density, Pmatrix is ^a matrix density that can be determined from the compositional range as determined based on the geochemical properties, and p^_iuid is a fluid density:

Pbuik Pmatrix T P fluid (1)

[0034] The fluid density of a hydrocarbon bearing formation can be determined in part by density of water and the density of a hydrocarbon, where p_water is the water density, and Phydrocarbon is the hydrocarbon density:

P fluid Pwater t Phydrocarbon (¾

[0035] For very high water saturations (e.g., water saturations approaching unity), the fluid density can be simplified to be the density of water. Then, in the case where all densities are normalized to the density of water, the bulk density of the high-accuracy zones can be determined using Equation 3:

Pbuik Pmatrix 1 ( ) [0036] The very high water saturation zones can have resistivity responses that are proportional to the volume of mineral phases and the volume of water determined by reconstructed porosity values. A reconstructed porosity value, or a weight on the reconstructed porosity value, can be determined by the formation composition. For example, for a formation zone with a compositional range comprising 20% of a material with a porosity of 0.1 and 80% of a material with a porosity of 0.4, the reconstructed porosity can be 0.34. The reconstructed porosity values can be combined to generate a reconstructed porosity log. Based on the reconstructed porosity log and a known resistivity of the water in the formation, a resistivity calculation operation can be used to generate a reconstructed resistivity log for the high-accuracy zones.

[0037] In some embodiments, high-accuracy zones can be filtered using caliper

measurements. A caliper measurement can be used to determine the distance a sensor of a well tool is from a borehole wall. The caliper measurement can be used with a threshold to flag a depth as inaccurate for resistivity measurements and/or other measurements by the sensor. In some embodiments, the caliper measurement and/or flagged depths can be used to determine target zones and/or re-classify zones previously labeled as high-accuracy to no longer be labeled as high-accuracy.

[0038] At block 308, a determination is made of whether the reconstructed resistivity log matches the blended resistivity log within a first threshold at the high-accuracy zones. A first threshold can be a predetermined value, such as 5%, 3%, 2%, etc. In some embodiments, the reconstructed resistivity log can be determined to match the blended resistivity log at the high- accuracy zones when the values in the reconstructed resistivity log is within the first threshold range of the blended resistivity log for the high-accuracy zones. For example, in the case where the first threshold is 2%, a reconstructed resistivity can be determined to match the blended resistivity at the high-accuracy zones when the reconstructed resistivity is 101.5 Ohms and the reconstructed resistivity is 100 Ohms. In some embodiments, a reconstructed resistivity log can match with a blended resistivity log when each reconstructed resistivity value matches with their corresponding blended resistivity log. Alternatively, a reconstructed resistivity log can be determined to match with a blended resistivity log at the high-accuracy zones if a summed error value is less than the first threshold. For example, the reconstructed resistivity log can be determined to match with a blended resistivity log when the sum of the squares of differences between the reconstructed resistivity log and the blended resistivity log are less than a first threshold. If the reconstructed resistivity is determined to match the blended resistivity, then operations of the flowchart 300 can proceed to block 314. Otherwise, operation of the flowchart 300 can proceed to block 310.

[0039] At block 310, the reconstructed porosity log is updated based on the compositional range. The formation porosity can fall within the porosity boundaries determined by the compositional range corresponding with the formation depth of the porosity boundary. For example, for a formation zone at a depth between 5000 feet and 5500 feet with a compositional range comprising 20-80% of a material with a porosity of 0.1 and 20-80% of a material with a porosity of 0.4, the reconstructed porosity can vary between the boundary values 0.16 to 0.34. In some embodiments, the formation porosity can be updated based on a difference minimization algorithm between the reconstructed resistivity and the blended resistivity.

[0040] At block 316, a second reconstructed density log, density probability distribution, and a second reconstructed resistivity log can be generated based on the compositional range and water saturation log at the target zones. The second reconstructed density is a reconstructed bulk density that can be generated based on a reconstructed fluid density. The reconstructed fluid density can be predicted based on Equation 4, where p^_iuid is the reconstructed fluid density, f is a weighting factor, S_w is a water saturation, p_water is the density of water, and Phydrocarbon is a hydrocarbon density:

[0041] In some embodiments, the value of p_water

Phydrocarbon can be known, and S_w can be varied to determine a reconstructed value of p luid - A reconstructed fluid density allows the use of Equation 1 to determine a second reconstructed density value. A second reconstructed density log at the target zones can be generated based on combining the second reconstructed density values at their corresponding well depths in the target zones. Using known resistivity algorithms, each value of the second reconstructed resistivity log at the target zones can be generated based on the values of the second reconstructed density log and water saturation log. The second reconstructed resistivity values can be combined with the first reconstructed resistivity values to generate a second reconstructed density log that includes values that correspond with both the target zones and high-accuracy zones. In some embodiments, the second reconstructed density log and density probability distribution function can be determined using a probabilistic model. For example, a reconstructed density in a first target zone can be determined by determining the likelihood of that reconstructed density to produce a resistivity value equal to the blended resistivity data for that zone based on a combination of materials that would fall within the compositional range of the formation.

[0042] At block 320, a determination is made of whether the second reconstructed resistivity log matches the blended resistivity within a second threshold. A second threshold can be a predetermined value, such as 4%, 2%, 1%, etc. The second threshold can be less than the first threshold. In some embodiments, the second reconstructed resistivity log can be determined to match the blended resistivity log using operations similar to the operations described for block 308, wherein the second threshold would substitute for the first threshold. In some embodiments, a reconstructed resistivity log can match with a blended resistivity log when each reconstructed resistivity value matches with their corresponding blended resistivity log within the second threshold. If the reconstructed resistivity is determined to match the blended resistivity within the second threshold at the target zones, then operations of the flowchart 300 can proceed to transition point A, continues at transition point A in the flowchart 400 of FIG. 4. Otherwise, operation of the flowchart 300 can proceed to block 324.

[0043] At block 324, the water saturation log is updated based on the second reconstructed resistivity log. The water saturation log can be updated by updating one or more water saturation values in the water saturation log. The water saturation log can be determined based on the updated reconstructed resistivity log using an optimization scheme. For example, for a formation zone at a depth between 5000 feet and 5500 feet, a resistivity log can limit a water saturation value to a range between 0.2-0.4. Based on this range, the water saturation in the zone can be updated from an initial value to a different value between 0.2-0.4 using an optimization scheme (e.g., Bayesian optimization, gradient descent, etc.).

[0044] Operations of the flowchart 400 of FIG. 4 are now described. From transition point A, operations of the flowchart 400 continue at block 424.

[0045] At block 424, the water saturation log is updated based on the second reconstructed density log. An algorithmic and/or probabilistic method can use the blended resistivity log, reconstructed density log, and reconstructed porosity log to update the water saturation. For example, for a formation zone at a depth between 5000 feet and 5500 feet, a density log can limit a water saturation value to a range between 0.2-0.4. Based on this range, the water saturation in the zone can be updated from an initial value to a different value between 0.2-0.4 using an optimization scheme.

[0046] At block 444, a reconstructed neutron porosity log and third reconstructed resistivity log are updated based on the reconstructed density log and an updated water saturation log. The neutron porosity values can be part of the measurements in a neutron porosity log acquired by a well tool. An algorithmic and/or probabilistic method can use the blended resistivity, the second reconstructed density, and water saturation (which can have been updated at block 424) to update the reconstructed neutron porosity. The reconstructed neutron porosities at different depths can be combined to generate a reconstructed neutron porosity log.

[0047] At block 448, a determination is made of whether the third reconstructed resistivity log matches the blended resistivity log within a third threshold range. A third threshold can be a predetermined value, such as 3.5%, 2%, 1%, etc. The third threshold can be less than the second threshold. In some embodiments, the third reconstructed resistivity log can be determined to match the blended resistivity log using operations similar to the operations described for block 308, wherein the second threshold would substitute for the first threshold. In some embodiments, a reconstructed resistivity log can match with a blended resistivity log when each reconstructed resistivity value matches with their corresponding blended resistivity log within the third threshold. If the reconstructed resistivity log is determined to match the blended resistivity log within the second threshold at the target zones, then operations of the flowchart 400 can proceed to block 452. Otherwise, operation of the flowchart 400 can proceed to block 424.

[0048] At block 452, formation properties and hydrocarbon volumes can be determined based on the third reconstructed resistivity, other logs for the well, and probability distributions associated with the logs. The formation properties and hydrocarbon volumes can be determined based on a formation model. The formation model can be a one-dimensional or multi dimensional model that associates a set of formation properties with a set of coordinates. The formation model can be generated based on a probabilistic error minimization method comparing the third reconstructed resistivity log and other well logs such as the updated water saturation log, reconstructed neutron porosity log, blended resistivity log, the second reconstructed density, etc. with the high-accuracy zones in the well. The operation can include using the corresponding probability distributions for the values in each log. During generation of a formation model, the high-accuracy measurement log zones can be used to calibrate various weighting factors associated with each of the logs used to generate the formation model. Once generated, the formation model can be used to determine various formation properties, including formation properties indicative of an estimated hydrocarbon recovery from the subsurface formation. For example, the formation model can be used to determine the mineral composition and fluid volumetries (i.e., the volume of hydrocarbons present) for a specific a zone and/or the entire formation. The formation model can also be used to determine any associated probability distributions for these formation properties, which can be used to determine reserve estimates for the formation.

[0049] FIG. 5 depicts a flowchart of operations to determine formation properties and hydrocarbon amounts using an artificial neural network, according to some embodiments.

Operations of the flowchart 500 begin at block 504.

[0050] At block 504, formation properties for high-accuracy zones are determined. With reference to FIG. 3, the high-accuracy zones can be determined based on whether the water saturation is greater than a water saturation threshold using an operation similar to the operations described for block 306. Alternatively, or in addition, high-accuracy zones can be determined based on how correlated the logs of a zone are with respect to each other. For example, derivatives of the resistivity, neutron porosity, and caliper logs can be correlated with each other to determine whether dramatic changes are sensor errors or are corroborated with data from other logs from the well.

[0051] At block 506, an artificial neural network is trained to determine the reconstructed logs and hydrocarbon volumetries at the high-accuracy zones based on corresponding measurements from the high-accuracy zones, highest resolution measurements, and/or regional analogues. The artificial neural network can be trained based on reconstructed logs. The reconstructed logs can be generated based on measurements taken from the well such as caliper logs, microresistivity logs, macroresistivity logs, and geochemical properties to accurately reconstruct formation property values such as formation porosity and hydrocarbon content. Alternatively, or in addition, regional analogues can be used. For example, the artificial neural network can be trained to provide predictions for a target well using formation log data from a well that shares the same formation with the target well.

[0052] At block 508, a probability distribution of formation properties and formation reserve are determined using the trained artificial neural network. In some embodiments, the trained artificial neural network can be used to determine formation properties based on the

measurements taken from the well at the target zones. For example, an artificial neural network trained on high-accuracy data from a formation at a depth of 5000-5500 feet can be used to determine values and probability distribution for formation properties at a zone of interest from 5500-6000 feet. The values of the formation properties and corresponding probability distributions can be used by the artificial neural network to determine information related to reserve estimates, such as the confidence level for a minimum reserve estimate.

Example Data

[0053] FIG. 6 depicts example zones of a resistivity log, according to some embodiments. Each of plots 604-612 show a data corresponding with a depth that is in the range of 3092 feet (FT) to 3132 FT, and each resistivity plot is logarithmic along the x-axis and ranges from 0.02 to 2000 ohms. Plot 604 depicts a macroresistivity measurement, which includes peaks at the depth 605 and the depth 606, which correspond with depths of 3091 feet and 3101 feet, respectively. Plot 608 depicts a microresistivity measurement, which do not show the same peak positions or values at the depths of 3091 or depth 3101. With respect to FIG. 2, the microresistivity measurement depicted in plot 608 can be rescaled according to a maximum and minimum of plot 604 to generate the blended resistivity plot 612 using operations similar to those described for block 210.

[0054] FIG. 7 depicts example zones of a geochemical log, according to some embodiments. Fig. 7 includes graphical representations of data regarding the detected presence of various chemical elements and a machine-driven interpretation of the data regarding mineral fractions. The elemental plots include the silica plot 704, aluminum plot 708, calcium plot 712, iron plot 716, potassium plot 720, magnesium plot 724, and sulfur plot 728. The geochemical properties also includes mineral identity plot 730, wherein the relative horizontal length at any depth represents the volumetric fraction of a mineral. For example, at a depth of 6050 feet, the mineral composition can include kerogen region 734, Mg-Chlorite region 732, Ilite region 736, quartz region 738, calcite region 740, and dolomite region 742. Based on the detected amounts of silica, aluminium, calcium, iron, potassium, and sulfur, as well as the lack of other minerals of interest such as magnesium, a computer can provide a prediction that the formation composition at a depth of 6050 feet would 10% kerogen, 5% magnesium-chlorite, 50% ilite, 25% quartz, and 10% dolomite. In other embodiments, different formation compositions can be determined, based on different geochemical properties and formation identification algorithms.

[0055] FIG. 8 depicts example zones of a set of reconstructed logs, according to some embodiments. The example logs include the gamma log 812 and the flag value 814, wherein the flag value 814 can indicate zones where measurements can be unfavorable based on caliper measurements. For example, at depths where the flag value 814 exceeds a threshold value, reconstructed logs such as the blended resistivity log 822 depicted in the blended resistivity plot 820 can be used to compensate for unfavorable measurement conditions. The density plot 830 includes both the initial density log 832 and the reconstructed density log 834. The density plot 840 includes both the initial neutron porosity log 842 and the reconstructed neutron porosity log 844, Both the density plot 830 and density plot 840 can show significant differences between their initial log and the reconstructed log at zones indicated by the flag value 814 to be unfavorable.

Example Artificial Neural Network

[0056] FIG. 9 depicts a diagram of an artificial neural network being used to determine formation properties, according to some embodiments. The artificial neural network 900 can accept initial well data 904 as inputs, such as initial density logs, initial gamma logs, initial neutron porosity logs, and initial geochemical logs. For example, with reference to FIG. 2 and FIG. 5, a blended resistivity log generated at block 214 can be provided as one of the initial well data 904 as described for block 504. Alternatively, or in addition, high-resolution measurements from other zones and regional analogues can be included in the initial log data 904. The initial log data 904 can be provided to the hidden layers 908, which can be further altered with one or more values from the teaching inputs 916.

[0057] The outputs from each of the hidden layers 912 can be combined to provide the formation property output 920, which can include values such as a reconstructed porosity, reconstructed porosity logs, reconstructed resistivity logs, reconstructed water saturation, a set of weights associated with each formation log used to determine a formation model, formation composition logs, etc. With respect to FIG. 5, the artificial neural network 900 can be trained using data from the high-accuracy zones as described for block 506 to determine formation properties and their corresponding probability distributions in the target zones as described for block 508. In some embodiments, a part of the input data and training data can be withheld from a training step to test the efficacy of the trained artificial neural network (e.g., using an operation similar or identical to the K-Fold Cross Validation technique).

Example Drilling System

[0058] Instead of performing operations herein using a wireline tool, some embodiments can be incorporated as part of a measurement or logging while drilling configuration. For example, FIG. 10 depicts an example drilling system, according to some embodiments. FIG. 10 depicts a drilling system 1000. The drilling system 1000 includes a drilling rig 1001 located at the surface

1002 of a borehole 1003. The initial location of the borehole 1003 and various operational parameters (e.g. drilling speed, weight on bit, drilling fluid pump rate, drilling direction, drilling fluid composition) for drilling can be selected based on the formation properties and/or volume of hydrocarbon in the formation (as described above). For example, location of the borehole

1003 can be selected to target the high-hydrocarbon zone 1060. The drill string 1004 can be operated for drilling the borehole 1003 through the subsurface formation 1032 with the bottomhole assembly (BHA).

[0059] The BHA includes a drill bit 1030 at the downhole end of the drill string 1004. The direction of the drill bit 1030 is changed so that the borehole 103 enters and/or remains in the high-hydrocarbon zone 1060, wherein the high-hydrocarbon zone 1060 is determined based on formation properties determined based on geochemical properties and variable water saturation (as described herein). The BHA and the drill bit 1030 can be coupled to computing system 1050, which can operate the drill bit 1030 as well as received data based on the sensors attached to the BHA. The drill bit 1030 can be operated to create the borehole 1003 by penetrating the surface 1002 and subsurface formation 1032. In some embodiments, a drilling plan can call for the drill bit 1030 to slow drilling when within a range of the high-hydrocarbon zone 1060. By increasing the accuracy of formation property determination, the drill bit 1030 can more safely and efficiently penetrate the high-hydrocarbon zone 1060. For example, sensors on the BHA can transmit a signal to the computing system 1050 that the drill bit is near a high-hydrocarbon zone, and the computing system can slow or redirect the drill bit 1030.

Example Computer Device

[0060] FIG. 11 depicts an example computer system with a formation property predictor, according to some embodiments. A computer device 1100 includes a processor 1101 (possibly including multiple processors, multiple cores, multiple nodes, and/or implementing multi threading, etc.). The computer device 1100 includes a memory 1107. The memory 1107 can be system memory (e.g., one or more of cache, SRAM, DRAM, zero capacitor RAM, Twin Transistor RAM, eDRAM, EDO RAM, DDR RAM, EEPROM, NRAM, RRAM, SONOS, PRAM, etc.) or any one or more of the above already described possible realizations of machine- readable media. The computer device 1100 also includes a bus 1103 (e.g., PCI, ISA, PCI- Express, HyperTransport® bus, InfiniBand® bus, NuBus, etc.) and a network interface 1105 (e.g., a Fiber Channel interface, an Ethernet interface, an internet small computer system interface, SONET interface, wireless interface, etc.).

[0061] The computer device 1100 includes a formation property predictor 1111. The formation property predictor 1111 can perform one or more operations described above. For example, the formation property predictor 1111 can generate blended resistivity values based on well geochemical properties. Additionally, the formation property predictor 1111 can determine formation properties based on the well geochemical properties.

[0062] Any one of the previously described functionalities can be partially (or entirely) implemented in hardware and/or on the processor 1101. For example, the functionality can be implemented with an application specific integrated circuit, in logic implemented in the processor 1101, in a co-processor on a peripheral device or card, etc. Further, realizations can include fewer or additional components not illustrated in Figure 11 (e.g., video cards, audio cards, additional network interfaces, peripheral devices, etc.). The processor 1101 and the network interface 1105 are coupled to the bus 1103. Although illustrated as being coupled to the bus 1103, the memory 1107 can be coupled to the processor 1101. The computer device 1100 can be device at the surface and/or integrated into component(s) in the wellbore. For example, with reference to FIG. 1, the computer device 1100 can be incorporated in the computer 110 and/or a computer at a remote location.

[0063] As will be appreciated, aspects of the disclosure can be embodied as a system, method or program code/instructions stored in one or more machine-readable media.

Accordingly, aspects can take the form of hardware, software (including firmware, resident software, micro-code, etc.), or a combination of software and hardware aspects that can all generally be referred to herein as a“circuit,”“module” or“system.” The functionality presented as individual modules/units in the example illustrations can be organized differently in accordance with any one of platform (operating system and/or hardware), application ecosystem, interfaces, programmer preferences, programming language, administrator preferences, etc.

[0064] Any combination of one or more machine readable medium(s) can be utilized. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable storage medium can be, for example, but not limited to, a system, apparatus, or device, that employs any one of or combination of electronic, magnetic, optical, electromagnetic, infrared, or semiconductor technology to store program code. More specific examples (a non-exhaustive list) of the machine-readable storage medium would include the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a machine-readable storage medium can be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable storage medium is not a machine-readable signal medium.

[0065] A machine-readable signal medium can include a propagated data signal with machine readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated signal can take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A machine-readable signal medium can be any machine readable medium that is not a machine-readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0066] Program code embodied on a machine-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0067] Computer program code for carrying out operations for aspects of the disclosure can be written in any combination of one or more programming languages, including an object oriented programming language such as the Java® programming language, C++ or the like; a dynamic programming language such as Python; a scripting language such as Perl programming language or PowerShell script language; and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on a stand-alone machine, can execute in a distributed manner across multiple machines, and can execute on one machine while providing results and or accepting input on another machine.

Variations

[0068] The program code/instructions can also be stored in a machine-readable medium that can direct a machine to function in a particular manner, such that the instructions stored in the machine-readable medium produce an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

[0069] As will be appreciated, aspects of the disclosure may be embodied as a system, method or program code/instructions stored in one or more machine-readable media. Accordingly, aspects may take the form of hardware, software (including firmware, resident software, micro-code, etc.), or a combination of software and hardware aspects that may all generally be referred to herein as a“circuit,”“module” or“system.” The functionality presented as individual modules/units in the example illustrations can be organized differently in accordance with any one of platform (operating system and/or hardware), application ecosystem, interfaces, programmer preferences, programming language, administrator preferences, etc.

[0070] Any combination of one or more machine readable medium(s) may be utilized. The machine readable medium may be a machine readable signal medium or a machine readable storage medium. A machine readable storage medium may be, for example, but not limited to, a system, apparatus, or device, that employs any one of or combination of electronic, magnetic, optical, electromagnetic, infrared, or semiconductor technology to store program code. More specific examples (a non-exhaustive list) of the machine readable storage medium would include the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a machine readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine readable storage medium is not a machine readable signal medium.

[0071] A machine readable signal medium may include a propagated data signal with machine readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A machine readable signal medium may be any machine readable medium that is not a machine readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0072] Plural instances may be provided for components, operations or structures described herein as a single instance. Finally, boundaries between various components, operations and data stores are somewhat arbitrary, and particular operations are illustrated in the context of specific illustrative configurations. Other allocations of functionality are envisioned and may fall within the scope of the disclosure. In general, structures and functionality presented as separate components in the example configurations may be implemented as a combined structure or component. Similarly, structures and functionality presented as a single component may be implemented as separate components. These and other variations, modifications, additions, and improvements may fall within the scope of the disclosure.

[0073] Use of the phrase“at least one of’ preceding a list with the conjunction“and” should not be treated as an exclusive list and should not be construed as a list of categories with one item from each category, unless specifically stated otherwise. A clause that recites“at least one of A, B, and C” can be infringed with only one of the listed items, multiple of the listed items, and one or more of the items in the list and another item not listed.

Example Embodiments

[0074] Example embodiments include the following:

[0075] Embodiment 1 : A method for generating a reconstructed log based on a formation resistivity and a geochemical property, the method comprising: lowering a well tool into a wellbore, wherein the well tool comprises, a first pair of electrodes used to generate a macroresistivity measurement, wherein a first distance between the first pair of electrodes is greater than a defined distance, and a second pair of electrodes used to generate a

microresistivity measurement, wherein a second distance between the second pair of electrodes is less than the defined distance; receiving the geochemical property of a subsurface formation around the wellbore; generating a compositional range based on the geochemical property; determining the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and generating the reconstructed log, wherein the reconstructed log comprises at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

[0076] Embodiment 2: The method of Embodiment 1, further comprising determining a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

[0077] Embodiment 3: The method of Embodiments 1 or 2, wherein generating the reconstructed log based on the geochemical property and the formation resistivity comprises: determining a formation composition based on the geochemical property; and generating the reconstructed log of the density of the subsurface formation based on the formation composition.

[0078] Embodiment 4: The method of any of Embodiments 1-3, wherein determining the formation resistivity comprises rescaling the microresistivity measurement based on a maximum of the macroresistivity measurement and a minimum of the macroresistivity measurement.

[0079] Embodiment 5: The method of any of Embodiments 1-4, further comprising: training an artificial neural network on a high-accuracy zone of the wellbore; and determining a formation property using the artificial neural network, wherein the formation property is determined based on the reconstructed log.

[0080] Embodiment 6: The method of any of Embodiments 1-5, further comprising determining a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation.

[0081] Embodiment 7: The method of any of Embodiments 1-6, further comprising determining whether hydrocarbons are present at a depth based on the formation property of the subsurface formation.

[0082] Embodiment 8: One or more non-transitory machine-readable media comprising program code for generating a reconstructed log based on a formation resistivity and a geochemical property, the program code to: lower a well tool into a wellbore, wherein the well tool comprises, a first pair of electrodes used to generate a macroresistivity measurement, wherein a first distance between the first pair of electrodes is greater than a defined distance, and a second pair of electrodes used to generate a microresistivity measurement, wherein a second distance between the second pair of electrodes is less than the defined distance; receive the geochemical property of a subsurface formation around the wellbore; generate a compositional range based on the geochemical property; determine the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and generate the reconstructed log, wherein the reconstructed log comprises at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

[0083] Embodiment 9: The one or more non-transitory machine-readable media of

Embodiment 8, further comprising program code to determine a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

[0084] Embodiment 10: The one or more non-transitory machine-readable media of

Embodiments 8 or 9, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code to: determining a formation composition based on the geochemical property; and generating the reconstructed log of the density of the subsurface formation based on the formation composition.

[0085] Embodiment 11 : The one or more non-transitory machine-readable media of any of Embodiments 8-10, wherein the program code to determine the formation resistivity comprises program code to rescale the microresistivity measurement based on a maximum of the macroresistivity measurement and a minimum of the macroresistivity measurement. [0086] Embodiment 12: The one or more non-transitory machine-readable media of any of Embodiments 8-11, further comprising program code to: training an artificial neural network on a high-accuracy zone of the wellbore; and determining a formation property using the artificial neural network, wherein the formation property is determined based on the reconstructed log.

[0087] Embodiment 13: The one or more non-transitory machine-readable media of any of Embodiments 8-12, further comprising program code to determine a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation

[0088] Embodiment 14: The one or more non-transitory machine-readable media of any of Embodiments 8-13, further comprising program code to determine whether hydrocarbons are present at a depth based on the formation property of the subsurface formation.

[0089] Embodiment 15: A system for generating a reconstructed log based on a formation resistivity and a geochemical property, the system comprising: a well tool comprising, a first pair of electrodes used to generate a macroresistivity measurement, wherein a first distance between the first pair of electrodes is greater than a defined distance, and a second pair of electrodes used to generate a microresistivity measurement, wherein a second distance between the second pair of electrodes is less than the defined distance; a processor; and a machine-readable medium having program code executable by the processor to cause the processor to: receive the geochemical property of a subsurface formation around a wellbore; generate a compositional range based on the geochemical property; determine the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and generate the reconstructed log, wherein the reconstructed log comprises at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

[0090] Embodiment 16: The system of Embodiment 15, wherein the machine-readable medium comprises program code executable by the processor to determine a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

[0091] Embodiment 17: The system of Embodiments 15 or 16, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code to: determining a formation composition based on the geochemical property; and generating the reconstructed log of the density of the subsurface formation based on the formation composition. [0092] Embodiment 18: The system of any of Embodiments 15-17, wherein the program code to determine the formation resistivity comprises program code to rescale the

microresistivity measurement based on a maximum of the macroresistivity measurement, and a minimum of the macroresistivity measurement.

[0093] Embodiment 19: The system of any of Embodiments 15-18, further comprising program code to determine a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation.

[0094] Embodiment 20: The system of any of Embodiments 15-19, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code executable by the processor to update a water saturation.

Claims

WHAT IS CLAIMED IS:

1. A method for generating a reconstructed log based on a formation resistivity and a

geochemical property, the method comprising:

lowering a well tool into a wellbore, wherein the well tool comprises,

a first pair of electrodes used to generate a macroresistivity measurement, wherein a first distance between the first pair of electrodes is greater than a defined distance, and

a second pair of electrodes used to generate a microresistivity measurement, wherein a second distance between the second pair of electrodes is less than the defined distance;

receiving the geochemical property of a subsurface formation around the wellbore;

generating a compositional range based on the geochemical property;

determining the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and

generating the reconstructed log, wherein the reconstructed log comprises at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

2. The method of claim 1, further comprising determining a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

3. The method of claim 1, wherein generating the reconstructed log based on the geochemical property and the formation resistivity comprises:

determining a formation composition based on the geochemical property; and

generating the reconstructed log of the density of the subsurface formation based on the formation composition.

4. The method of claim 1, wherein determining the formation resistivity comprises rescaling the microresistivity measurement based on a maximum of the macroresistivity measurement and a minimum of the macroresistivity measurement.

5. The method of claim 1, further comprising:

training an artificial neural network on a high-accuracy zone of the wellbore; and determining a formation property using the artificial neural network, wherein the formation property is determined based on the reconstructed log.

6. The method of claim 1, further comprising determining a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation.

7. The method of claim 6, further comprising determining whether hydrocarbons are present at a depth based on the formation property of the subsurface formation.

8. One or more non-transitory machine-readable media comprising program code for generating a reconstructed log based on a formation resistivity and a geochemical property, the program code to:

lower a well tool into a wellbore, wherein the well tool comprises,

receive the geochemical property of a subsurface formation around the wellbore;

generate a compositional range based on the geochemical property;

determine the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and

generate the reconstructed log, wherein the reconstructed log comprises at least one of a porosity and a density of the subsurface formation based on the compositional range and the formation resistivity, wherein the compositional range defines at least one of a boundary of values and an initial value.

9. The one or more non-transitory machine-readable media of claim 8, further comprising program code to determine a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

10. The one or more non-transitory machine-readable media of claim 8, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code to:

determining a formation composition based on the geochemical property; and generating the reconstructed log of the density of the subsurface formation based on the formation composition.

11. The one or more non-transitory machine-readable media of claim 8, wherein the program code to determine the formation resistivity comprises program code to rescale the

microresistivity measurement based on a maximum of the macroresistivity measurement and a minimum of the macroresistivity measurement.

12. The one or more non-transitory machine-readable media of claim 8, further comprising program code to:

training an artificial neural network on a high-accuracy zone of the wellbore; and determining a formation property using the artificial neural network, wherein the

formation property is determined based on the reconstructed log.

13. The one or more non-transitory machine-readable media of claim 8, further comprising program code to determine a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation.

14. The one or more non-transitory machine-readable media of claim 13, further comprising program code to determine whether hydrocarbons are present at a depth based on the formation property of the subsurface formation.

15. A system for generating a reconstructed log based on a formation resistivity and a geochemical property, the system comprising:

a well tool comprising,

a first pair of electrodes used to generate a macroresistivity measurement, wherein a first distance between the first pair of electrodes is greater than a defined distance, and a second pair of electrodes used to generate a microresistivity measurement, wherein a second distance between the second pair of electrodes is less than the defined distance;

a processor; and

a machine-readable medium having program code executable by the processor to cause the processor to:

receive the geochemical property of a subsurface formation around a wellbore; generate a compositional range based on the geochemical property; determine the formation resistivity based on the macroresistivity measurement and the microresistivity measurement; and

16. The system of claim 15, wherein the machine-readable medium comprises program code executable by the processor to determine a probability distribution for a formation property, wherein the formation property is determined based on the reconstructed log.

17. The system of claim 15, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code to:

determine a formation composition based on the geochemical property; and

generate the reconstructed log of the density of the subsurface formation based on the formation composition.

18. The system of claim 15, wherein the program code to determine the formation resistivity comprises program code to rescale the microresistivity measurement based on a maximum of the macroresistivity measurement, and a minimum of the macroresistivity measurement.

19. The system of claim 15, further comprising program code to determine a formation property based on the reconstructed log, wherein the formation property is indicative of an estimated hydrocarbon recovery from the subsurface formation.

20. The system of claim 15, wherein the program code to generate the reconstructed log based on the geochemical property and the formation resistivity comprises program code executable by the processor to update a water saturation.