WO2025174652A1

WO2025174652A1 - Methods and systems for characterizing cuttings and cavings in drilling operations

Info

Publication number: WO2025174652A1
Application number: PCT/US2025/014869
Authority: WO
Inventors: Eric Van Oort; Pradeepkumar Ashok; Abraham C. MONTES; Santiago CALLERIO; Sebastian PREZ
Original assignee: University of Texas System; University of Texas at Austin
Current assignee: University of Texas System; University of Texas at Austin
Priority date: 2024-02-15
Filing date: 2025-02-06
Publication date: 2025-08-21
Anticipated expiration: 2026-08-15

Abstract

Aspects of the present disclosure provide systems and methods for characterization of cuttings and cavings in drilling operations. A method for assessing wellbore conditions includes collecting a sample comprising rock fragments and mud from a drilling fluid, determining a rock fragment size distribution, determining a rock fragment shape distribution, determining a raw volume of the rock fragments, determining a volume correction factor using the rock fragment size distribution and the rock fragment shape distribution, determining an adjusted volume of the rock fragments from the raw volume using the volume correction factor, and assessing a wellbore condition based on the adjusted volume of the rock fragments and/or the rock fragment shape distribution.

Description

METHODS AND SYSTEMS FOR CHARACTERIZING CUTTINGS AND CAVINGS IN DRILLING OPERATIONS

BACKGROUND

Field

[0001] Embodiments of the present disclosure relate to methods and systems for characterization of cuttings and cavings in wellbore operations.

Description of the Related Art

[0002] In drilling operations, rock fragments retrieved from drilling operations are often used for wellbore assessment. These rock fragments often consist of cuttings and cavings. Cuttings are fragments of rock that have been cut away by drill bits during drilling operations. On the other hand, cavings are fragments of rock that have been retrieved that have not been cut away by drill bits during drilling operations.

[0003] The cuttings and cavings retrieved from the wellbores are characterized to assess wellbore cleaning sufficiency and wellbore stability. Characterization of cuttings and cavings can include identifying cavings, identifying cuttings, and determining the volumes of cuttings and cavings.

[0004] Conventionally, these cuttings and cavings are manually characterized. Manual characterization lends itself to subjectivity which may lead to bias in evaluation and assessment of wellbore conditions. Manual characterization is also not done in real-time. That is, there is a delay between when the cuttings and cavings are collected and when they are manually characterized. The bias evaluation and delayed assessment may hinder identification of problems in the wellbore, thereby allowing the wellbore conditions to deteriorate.

[0005] The failure to properly assess wellbore cleaning sufficiency and wellbore stability may lead to undesirable events that generate non-productive time (NPT) in drilling operations, such as stuck pipe events.

[0006] Thus, there is a need for improved methods and systems for characterization of cuttings and cavings in drilling operations. SUMMARY

[0007] Aspects of the present disclosure provide systems and methods for characterization of cuttings and cavings collected from drilling operations.

[0008] A method for assessing wellbore conditions, including: collecting a sample comprising rock fragments from a drilling fluid, determining a rock fragment size distribution, determining a rock fragment shape distribution, determining a raw volume of the rock fragments, determining a volume correction factor using the rock fragment size distribution and the rock fragment shape distribution, determining an adjusted volume of the rock fragments from the raw volume using the volume correction factor, and assessing a wellbore condition based on the adjusted volume of the rock fragments and/or the rock fragment shape distribution.

[0009] A method for determining a volume of rock fragments collected from a drilling fluid, including: capturing an image of a sample of the returning fluid, identifying mud in the sample using the image, segmenting the image of the sample to identify individual rock fragments in the sample, determining a size distribution of the rock fragments based on the segmented image, determining a shape distribution of the rock fragments based on the segmented image, and determining a volume of the rock fragments using a volume correction factor, wherein the volume correction factor is determined using the size distribution of the rock fragments and the shape distribution of the rock fragments.

[0010] A method for calculating a volume of cavings in a sample of rock fragments collected from a drilling fluid, including: determining a cavings size distribution and a cavings shape distribution of the sample, determining a raw volume of the cavings, calculating a cavings volume correction factor using the cavings size distribution and cavings shape distribution, and scaling the raw volume of the cavings using the cavings volume correction factor.

[0011] A method for assessing a wellbore condition, including: collecting a sample comprising rock fragments and mud from a drilling fluid, characterizing the plurality of rock fragments in the sample, determining a calculated true volume of the plurality of rock fragments, and assessing the wellbore condition based on the calculated true volume of the plurality of rock fragments. Characterizing the plurality of rock fragments in the sample includes: capturing a 2D image of the sample and a 3D point cloud of the sample, identifying mud in the sample based on the 2D image, preprocessing the 2D image and the 3D point cloud, segmenting the 2D image and the 3D point cloud to identify each of the plurality of rock fragments, determining the morphological characteristics of each of the plurality of rock fragments based on the segmented 2D image and the segmented 3D point cloud, classifying the rock fragments based on the morphological characteristics of the rock fragments, determining a size distribution of the plurality of rock fragments based on the morphological characteristics of each of the plurality of rock fragments, and determining a shape distribution of the plurality of rock fragments based on the morphological characteristics of each of the plurality of rock fragments. Determining a calculated true volume of the plurality of rock fragments, includes: determining a raw volume of the plurality of rock fragments based on the 3D point cloud, calculating a volume correction factor based on a stacking factor and the shape distribution of the plurality of rock fragments, and scaling the raw volume by the volume correction factor.

BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The appended figures illustrate only exemplary embodiments and are therefore not to be considered limiting of the scope of the disclosure, as the disclosure may admit to other equally effective embodiments.

[0013] The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

[0014] FIGs. 1 A-1 B illustrate a cuttings and cavings detection system, according to one or more embodiments of the present disclosure.

[0015] FIG. 2 illustrates a method for wellbore assessment using a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure [0016J FIG. 3 illustrates a method for characterizing cuttings and cavings in a sample of rock fragments from drilling operations, according to one or more embodiments.

[0017] FIG. 4 illustrates 2D images of samples of rock fragments from drilling operations that are classified by their mud contents, according to one or more embodiments of the present disclosure.

[0018] FIG. 5A illustrates an unsegmented 2D image of a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure.

[0019] FIG. 5B illustrates 2D image of a sample of rock fragments from drilling operations segmented by the cuttings model, according to one or more embodiments of the present disclosure.

[0020] FIG. 6A illustrates a 2D image of a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure.

[0021] FIG. 6B illustrates a 2D image of a sample of rock fragments from drilling operations segmented with the cavings model, according to one or more embodiments of the present disclosure.

[0022] FIG. 7 illustrates morphological characteristics of cuttings and cavings in a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure.

[0023] FIG. 8 illustrates classifications of cuttings and cavings by morphology based on the segmented 2D image of a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure.

[0024] FIG. 9 illustrates a method for calculating a true volume of the cuttings and cavings in a sample of rock fragments from drilling operations, according to one or more embodiments of the present disclosure. [0025J FIG. 10A illustrates the relationship between the volume correction factor, the shape of the cuttings and cavings, and the stacking level factor according to one or more embodiments of the present disclosure.

[0026] FIG. 10B illustrates the error rate of the calculated true volume and the empirically determined true volume of the cuttings and cavings in a sample of rock fragments from drilling operations, according to one or more embodiments.

[0027] FIG. 11 illustrates exemplary wellbore condition assessments, according to one or more embodiments.

[0028] To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures. It is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without further recitation.

DETAILED DESCRIPTION

[0029] Aspects of the present disclosure provide systems and methods for characterization of cuttings and cavings in drilling operations.

[0030] FIGs. 1A-1 B illustrate a cuttings and cavings detection system 100 according to some embodiments. The cuttings and cavings detection system 100 includes a skid 101 , an enclosure 102, a conveyor belt 103, a 3D laser scanner 104, a high-definition camera 105, a sensor computer 111 , and a remote host 112.

[0031] The system 100 is not site-specific and can be used at a variety of sites where drilling operations take place. The system 100 is also Class 1 Div. 1 certified for use at drilling sites.

[0032] The skid 101 extracts rock fragments 107 from drilling fluid used in drilling operations. A sample 106 of those rock fragments 107 is then provided to the system 100. The rock fragments 107 include cuttings 109 and cavings 110. In some embodiments, the sample 106 contains some mud 108 (i.e. the rock fragments 107 are wetted with mud 108). The rock fragments 107 may be originated from a variety of rocks, including, but not limited to, basalt, limestone, sandstone, and shale. The skid 101 supplies the sample 106 to the conveyor belt 103 of the system 100. The conveyor belt 103 is suitable for use in hazardous environments and supplies sample 106 to and through the enclosure 102. The sample 106 is supplied at a constant speed and is processed by the high-definition camera 105 and the 3D laser scanner 104. The constant speed allows for accuracy in analyzing the images taken by the high-definition camera 105 and the data from the 3D laser scanner 104.

[0033] The enclosure 102 houses the 3D laser scanner 104 and the high-definition camera 105. The enclosure 102 also prevents sunlight from interfering with the 3D laser scanner 104 and the high-definition camera 105 and prevents unwanted splashing of the sample 106 by encapsulating the conveyor belt 103, the 3D laser scanner 104, and the high-definition camera 105.

[0034] The 3D laser scanner 104 scans the sample 106 as the sample 106 is moved through the system 100 and generates 3D point cloud data of the sample 106. The high-definition camera 105 takes 2D images of the sample 106 as the sample 106 moves through the system 100. The 3D point cloud data and 2D images are used for cuttings 109 and cavings 110 characterization and wellbore condition assessment.

[0035] The sensor computer 111 may be located at the worksite. However, in some embodiments, the sensor computer 111 may be located remotely and in wireless communication with the remainder of the system 100. The sensor computer 111 is configured to receive and process data from the high definition camera 105 and the laser scanner 104. For example, the sensor computer 111 is configured to calculate a raw volume of rock fragments 107 in a sample 106 passing through the system 100 using 3D point cloud data captured by the laser scanner 104. The sensor computer 111 may be configured to calibrate the system 100 for the speed and control the speed of the conveyor belt 103. For example, the sensor computer 111 change the speed or correct for irregularities in the conveyor belt 103 during data collection by high definition camera 105, the laser scanner 104, and/or the sensor computer 111. The irregularities may include, for example, bumps and valleys on the surface of the belt 103.

[0036] FIG. 1 B illustrates the remote host 112. The remote host 112 is in communication (e.g., wireless communication) with the sensor computer 111. The remote host 112 is configured to identify mud 108 in the sample 106, identify segment rock fragments 107 in the sample, characterize the morphologies of rock fragments 107, correct the raw volume generated by the sensor computer 111 to find a true volume of rock fragments 107, simulate rock fragments 107 transport, and assess risk to the wellbore based on these steps. These steps will be discussed in further detail later in the descriptions of FIGs. 2-10B.

[0037] The remote host 112 may comprise a multitude of modules for doing independent calculations and determinations based on requirements of each step. For instance, the remote host 112 may comprise separate modules for characterizing morphologies of the rock fragments 107, calculating volume of the rock fragments 107 and assessing risk, handling input and output data, processing sensor data, and processing transport simulation data. These modules may be further divided based on sub-steps (such as sub-step 1211 of operation 1200 discussed later with regards to FIG. 2) and based on data used in those steps (such as 3D point cloud data vs. 2D images).

[0038] FIG. 2 illustrates an exemplary method 1000 of assessing wellbore conditions.

[0039] At operation 1100, a sample 106 of rock fragments 107 is taken from drilling operations. In some embodiments, the sample 106 also contains mud 108. At operation 1200, the rock fragments 107, including cuttings 109 and cavings 110, of the sample 106 are characterized. In this example, the size and morphology of the cuttings

109 and cavings 110 from the sample 106 are determined. At operation 1300, the true volumes of the cuttings 109 and cavings 110 from the sample 106 are estimated. The true volumes are based on the size and morphology of the cuttings 109 and cavings

110 determined at operation 1200. At operation 1400, the wellbore conditions are assessed. The wellbore is assessed using the size and morphology of the cuttings 109 and cavings 110 determined at operation 1200 and the true volume calculation determined at operation 1300.

[0040] FIG. 3 illustrates an exemplary embodiment of the characterization operation 1200 of the method 1000. With reference to the example of FIG. 1 , at step 1201 of the characterization operation 1200, 2D images and 3D point clouds of the drilling sample

106 are captured using a detection system, such as detection system 100, having a high-definition camera 105 and a 3D laser scanner 104 of FIG. 1. The rock fragments

107 in the sample 106 may include cuttings 109, cavings 110, and mud 108.

[0041] The samples 106 being analyzed may vary in a number of ways. The rock fragments 107 in the samples 106 may have different spatial distributions, ranging from highly sparse to stacked with rock fragments 107. The samples 106 may also have different levels of mud 108 content, ranging from excessively wet with mud to dry rock fragments 107. The samples 106 may also contain rock fragments 107 and mud 108 with different colors, ranging from a high contrast between color of rock fragments 107 and mud 108 to low contrast between colors of rock fragments 107 and mud 108. The samples 106 may also contain variations in uniformity of the rock fragments 107 in the sample 106.

[0042] At step 1211 of the characterization operation 1200, the generated 2D images are used to identify mud 108 and differentiate mud 108 from the rock fragments 107 in the sample 106. To differentiate between mud 108 and rock fragments 107, the sample 106 is classified by its mud 108 content using the 2D images. Possible sample classifications include an empty belt, wet sample with mud 108 and no rock fragments 107, a dry sample with rock fragments 107, or a wet sample with rock fragments 107 and mud 108. Examples of the sample classifications can be seen in FIG. 4.

[0043] The classification may be performed using a variety of models such as random forest (“RF”) classifier using a pretrained VCG convolutional neural network (“CNN”) model as a backbone to compress the information contained in the images. In use, the CNN model achieved highly accurate classification results, such as over 80% accuracy, or over 90% accuracy. In validation testing, the CNN model accurately identified 100% of cases involving dry fragments 107, accurately identified 98% of cases involving excessively wet samples with fragments 107 and mud 108, and accurately identified 95% of cases of a wet sample with no fragments 107. While the use of the RF model achieved highly accurate results, other suitable models may be used, such as any suitable classifier, including logistic regression models. While the use of a CNN backbone is disclosed, other suitable models may be used such as other types of feature extraction models, including a transformer.

[0044] Turning back to FIG. 3, after isolating the rock fragments 107 from the mud 108 in the 2D images at step 1211 , the resulting 2D images are preprocessed at step 1212 for segmentation. Preprocessing of the 2D images may include denoising, histogram equalization, image thresholding, and morphological optimization. In one example, a genetic algorithm is used to determine the correct configuration for the preprocessing techniques described above. The preprocessing step 1212 may be performed to accentuate the sections of images corresponding to individual rock fragments 107 and deemphasize regions not of interest (such as empty sections of the belt 103, mud stains on the belt, and metallic components of the belt). In this respect, the preprocessing step 1212 may avoid erroneous segmentation in the following segmentation step 1213.

[0045] At step 1213 of the characterization operation 1200, the cuttings 109 and cavings 110 of the rock fragments 107 in the 2D images are segmented. In one example, segmentation uses two models to process the 2D images to identify individual cuttings 109 and cavings 110. Segmentation of the cuttings 109 and cavings 110 is complicated because of the diverse arrangements, varying mud 108 content, and variations in uniformity of the cuttings 109, cavings 110, and other solids.

[0046] The first model (hereinafter the “cuttings model”) segments the cuttings 109 from the sample 106. Suitable cuttings models may include, but are not limited to, a Region-based Convolutional Neural Network (R-CNN), Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN), Watershed, and Convolutional Autoencoders. The cuttings model identifies the individual cuttings 109 in the sample without erroneously identifying spurious fragments. The cuttings model also identifies cuttings within the sample sufficient to extract a faithful size distribution of the cuttings 109.

[0047] The second model (hereinafter the “cavings model”) is used to segment the cavings 110. An exemplary cavings model is a Feature Pyramid Network, but may include other suitable cavings models. The FPN was used as the cavings model because it performs panoptic segmentation (/.e. it segments the entire image) providing data to calculate cuttings-to-cavings ratios. In one example, a FPN model could be used as the cuttings model as well. The cavings model segments the 2D image to identify the parts of the 2D image corresponding to cavings 110, and the parts that correspond to other portions, such as the belt 103, the frame of the belt, the cuttings 109, and the mud 108. The segmentation may be accomplished even in samples where the cuttings 109 are not uniform in size, shape, or color, and the cavings 110 are not uniform in size, shape, or color. The cavings model identifies the individual cavings 110 in the sample without erroneously identifying spurious fragments. The cavings model also identifies cavings 110 within the sample 106 sufficient to extract a faithful size distribution of the cavings 110.

[0048] Resulting from step 1213 are a first segmented 2D image identifying the individual cuttings 109 using the cuttings model and a second segmented 2D image identifying the individual cavings 110 using the cavings model. From the segmented 2D image identifying the cuttings 109 and the segmented image identifying the cavings 110, a cuttings-to-cavings ratio can be determined. The cuttings-to-cavings ratio is determined by taking the ratio of the cavings areas to cuttings areas in the 2D image as a surrogate of volumetric ratio. That is, the FPN model generates masks (e.g. images) where only one element (e.g. cuttings, cavings, belt, etc.) of the sample 106 is highlighted per mask and then these masks are used to calculate the area ratios of the cuttings and the cavings. FIG. 5A illustrates an unsegmented 2D image of a sample 106, and FIG. 5B illustrates a segmented 2D image of the sample 106 using the cuttings model. FIG. 6A illustrates an unsegmented 2D image of the sample 106, and 6B illustrates a segmented 2D image using the cavings model and differentiating between cuttings 109 and cavings 110.

[0049] Turning back to FIG. 3, at step 1214 of the characterization operation 1200, the angularity and circularity of the rock fragments 107 are characterized based on the segmented 2D images from step 1213. Step 1214 includes using the first segmented 2D image generated by the cuttings model to determine the angularity and circularity of the cuttings 109 in the sample 106 and includes using the second segmented 2D image generated by the cavings model to determine the angularity and circularity of the cavings 110 in the sample 106.

[0050] The angularity of the rock fragments 107 relates to the roundness of the rock fragments 107, and the circularity relates to the eccentricity of the rock fragments 107. In one example, step 1214 includes using a best-fit ellipse on the segmented 2D images to determine a length and a width of each individual cutting 109 and each individual caving 110. An example of a 2D image including a determination of angularity and circularity of the rock fragments 107 can be seen in FIG. 7. The angularity of the cuttings 109 and cavings 110 may be determined from the empty area in the ellipse. The circularity of the cuttings 109 and cavings 110 may be determined from the eccentricity of the ellipse. In this manner, step 1214 provides a data-set including the length, widths, angularity, and circularity for each rock fragment 107 of the sample 106.

[0051] Turning back to FIG. 3, at step 1215 of the characterization operation 1200, the cuttings 109 and cavings 110 are classified based on their morphology.

[0052] The cuttings 109 are classified into a class of angular or round based on their calculated angularity and circularity determined in step 1214. The cavings 110 are classified into classes including angular, tabular, splintery (i.e. elongated), and round based on angularity and circularity determined in step 1214 and an ensemblebased classifier. In another example, the FPN is used to classify the cavings by compressing the image into features and maps those features to the intended classes. FIG. 8 illustrates a segmented 2D image with the cavings 110 classified by their morphology.

[0053] Resulting from step 1215 is a shape distribution of the cuttings 109 in the sample. The shape distribution of the cuttings 109 is a dataset indicating the percentage of the cuttings 109 classified as round and angular. Also resulting from step 1215 is a shape distribution of the cavings 110 in the sample. The shape distribution of the cavings 110 is a dataset indicating the percentage of the cavings 110 classified as angular, round, tabular, elongated, and reworked. [0054J Turning back to FIG. 3, at step 1216 the size distribution of the cuttings 109 is determined. The size distribution of the cuttings 109 is a dataset indicating the length distribution and width distribution of the cuttings 109 of the sample 106. Also determined at step 1216 is the size distribution of the cavings 110. The size distribution of the cavings 110 is a dataset indicating the length distribution and width distribution of the cavings 110 of the sample 106. It is contemplated that one or both of steps 1215 and 1216 may be combined with step 1214.

[0055] Turning back to FIG. 3, the 3D point cloud data of the sample 106 are used in the characterization operation 1200. In one example, the 3D point cloud data can be processed in tandem with the 2D images of the sample 106. In some embodiments, only the 2D images or the 3D point clouds are used in the characterization operation 1200 of the method 1000. In some embodiments, the characterization from the 2D images may be compared to the characterization from the 3D point cloud data.

[0056] At step 1222, the 3D point cloud data is preprocessed for segmentation. Similar to step 1212, preprocessing may include denoising, histogram equalization, image thresholding, and morphological optimization. Similar techniques are used to preprocess the 3D point cloud data that are used to preprocess the 2D images of the sample in step 1212. The preprocessing step 1222 is important to avoid erroneous segmentation in the segmentation step 1223.

[0057] At step 1223, the preprocessed 3D point cloud data is segmented. Similar techniques are used to segment the sample 106 using the 3D point cloud data as are used to segment the sample 106 with the 2D images in step 1213. In some embodiments, segmentation of 3D point clouds may be more comprehensive than the 2D images. Because the color in 3D point clouds is determined by height instead of reflection of certain wavelengths, 3D point cloud data can provide better contrast and facilitating border detection. Also, 3D images can better determine and exclude mud stains due to their height differentiation capabilities.

[0058] At step 1224, the segmented 3D point cloud data is used to determine the angularity and circularity of the cuttings 109 and cavings 110 of the sample 106. Similar techniques are used to determine the angularity and circularity of the cuttings 109 and cavings 110 as were used in step 1214.

[0059] At step 1225, the cuttings 109 and cavings 110 are classified based on their morphologies and the segmented 3D point cloud data. Similar techniques and classifications are used with the 3D point cloud data as are used with the 2D image in step 1215.

[0060] At step 1226, the individual size distributions of the cuttings 109 and cavings 110 are determined based on the segmented 3D point cloud data. Similar techniques are used to differentiate between cuttings 109 and cavings 110 in the sample using the 3D point cloud data as are used in step 1216.

[0061] At step 1202, the combined cuttings size distributions are determined and the combined cavings size distributions are determined. The combined cutting size distributions are based on the cutting size distributions determined at step 1216 and the cutting size distributions determined at step 1226. In one example, the combined cuttings size distributions include a weighted sum of the cutting width distributions determined at steps 1216 and 1226 and a weighted sum of the cutting length distributions as determined at steps 1216 and 1226. In one or more examples where only 2D images are used in the operation 1200 or where only 3D point cloud is used in the operation 1200, step 1202 is not necessary as there is nothing to combine.

[0062] The combined caving size distributions are based on the cavings size distributions determined at step 1216 and the cavings size distributions determined at step 1226. The combined cavings size distributions include a weighted sum of the cavings width distributions determined at steps 1216 and 1226 and a weighted sum of the cavings length distributions as determined at steps 1216 and 1226.

[0063] The weight given to the data generated by the 2D image and the weight given to the data generated by the 3D point cloud data depends on the mud content of the sample. For example, if the sample 106 has more mud content, then more weight is given to the 3D point cloud data, and if a sample 106 has less mud content, then more weight is given to the 2D image data. [0064J Resulting from step 1202 is a combined cuttings width distribution, a combined cavings width distribution, a combined cuttings length distribution, and a combined cavings length distribution.

[0065] At step 1203, the combined cuttings shape distribution and combined cavings shape distribution are determined.

[0066] The combined cutting shape distribution is based on the cutting shape distribution determined at step 1215 and the cutting shape distribution determined at step 1225. The combined cutting shape distribution is a weighted sum of the cutting shape distribution determined at step 1215 and the cutting shape distribution determined at step 1225, wherein the weight given to each is based on the mud content in the sample. In one or more examples where only 2D images are used in the method 1000 or where only 3D point clouds are used in the method 1000, step 1202 is not necessary as there is nothing to combine

[0067] The combined caving shape distribution is based on the caving shape distribution determined at step 1215 and the caving shape distribution determined at step 1225. The combined caving shape distribution is a weighted sum of the caving shape distribution determined at step 1215 and the caving shape distribution determined at step 1225, wherein the weight given to each is based on the mud content in the sample. For example, if the sample 106 has more mud content, then more weight is given to the 3D point cloud data, and if the sample 106 has less mud content, then more weight is given to the 2D image data.

[0068] Resulting from step 1203 is a combined cuttings shape distribution and a combined cavings shape distribution.

[0069] The combined cuttings shape distribution and combined cavings shape distribution determined at step 1203 and the combined cuttings size distributions and the combined cavings size distributions determined at step 1202 are used in operation 1300 of the method 1000 for calculating a true volume of the cuttings 109 and a true volume of the cavings 110 in the sample. [0070J FIG. 9 illustrates the true volume calculation operation 1300 of the method 1000 wherein a calculated true volume of the cuttings 109 and a true volume of the cavings 110 is determined, according to some embodiments.

[0071] At step 1301 , the 3D point cloud retrieved from step 1201 of the characterization operation 1200 is reconstructed to create a 3D profile of the rock fragments 107. At step 1302, the reconstructed 3D profile generated at step 1301 is integrated along the belt movement axis. At step 1303, a raw volume of the rock fragments 107 is determined based on the integrated 3D profile generated at step 1302.

[0072] The raw volume that is generated by the integrated point cloud at step 1303 is the raw total volume including the volume of cuttings 109 and cavings 110 in the sample 106. The volume of the cuttings 109 and the volume of the cavings 110 are differentiated for analysis purposes. The cuttings-to-cavings ratio determined previously is used to calculate how much of that raw volume is a volume of cuttings 109 and how much of that volume is a volume of cavings 110. Thus, a raw volume of the cuttings 109 and a raw volume of the cavings 110 is determined at step 1303.

[0073] The raw volume of the cuttings 109 and the raw volume of the cavings 110 do not take into account void space when the sample 106 includes stacked cuttings 109 and cavings 110. The void space represents the volume of the spaces in between the cuttings 109 and cavings 110 when the cuttings 109 and cavings 110 are stacked atop each other in a sample 106. In one embodiment, to determine a calculated true volume of the cuttings 109 and a calculated true volume of the cavings 110 without the void space, volume correction factors are used.

[0074] The volume correction factors are determined at step 1304. A first volume correction factor is determined for cuttings 109 for use with the cuttings raw volume determined at step 1303, and a second volume correction factor is determined for cavings 110 for use with the cavings raw volume determined at step 1303.

[0075] A volume correction factor is a percentage that is used to scale the raw volumes to account for the void space. [0076J The volume correction factors are determined using a best-fit line generated from empirical data collection. In one embodiment, the empirical data collection involves empirically determining volume correction factors, empirically determining size distributions, and empirically determining shape distributions. The empirically determined volume correction factor is then plotted against the empirically determined size distribution and the empirically determined shape distribution. The resulting plot is given a best fit line, which can be used to estimate a volume correction factor based on size distribution and shape distribution.

[0077] Therefore, in order to determine a cuttings volume correction factor, step 1303 includes using the combined cutting size distributions determined at step 1202 and the combined cuttings shape distribution determined at step 1203 in conjunction with the best-fit line to determine the cuttings volume correction factor.

[0078] Further, in order to determine a cavings volume correction factor, step 1303 includes using the combined caving size distributions determined at step 1202 and the combined cavings shape distribution determined at step 1203 in conjunction with the best-fit line to determine the cavings volume correction factor.

[0079] In some embodiments, the combined size distributions are converted into a stacking factor before use with the best-fit line. The stacking level factor is defined as the mean of the sample height divided by the median size of the rock fragment 107 determined in step 1202 of the characterization operation 1200. An equation for the stacking level factor can be seen reproduced below:

[0080] In these embodiments, the cutting stacking level factor and the combined cutting shape distribution determined at step 1203 are used to determine the cutting volume correction factor. Also, the cavings stacking level factor and the combined cavings shape distribution determined at step 1203 are used to determine the cavings volume correction factor. Generally, the larger the stacking level factor, the larger the volume correction factor. Also generally, the more angular the cuttings 109 and cavings 110 are in the sample, the larger the volume correction factor. FIG. 10A illustrates the relationship between the volume correction factors, the combined shape distributions, and the stacking level factor.

[0081] At step 1305, a calculated true volume of the cuttings and a calculated true volume of the cavings are determined by using their respective volume correction factor determined at step 1304 to scale their respective raw volumes determined at step 1303. In this respect, the calculated true volume is an adjusted volume of the raw volume based on the volume correction factor.

[0082] In validation testing, the calculated true volume of the cuttings 109 and the calculated true volume of the cavings 110 produced an error range of +/-10% when compared against empirically determined true volumes of cuttings 109 and empirically determined true volumes of cavings 110. FIG. 10B illustrates the error rate of the calculated true volume and the empirically determined true volume of the cuttings and cavings.

[0083] At operation 1400, the condition of the wellbore from which the sample 106 was retrieved is assessed. The wellbore condition may be assessed based on one or more of the cuttings 109 size distribution determined at step 1216, the cuttings 109 size distribution determined at step 1226, the cavings 110 size distribution determined at step 1216, the cavings 110 size distribution determined at step 1226, the cuttings 109 shape distribution determined at step 1215, the cuttings 109 shape distribution determined at step 1225, the cavings 110 shape distribution determined at step 1215, the cavings 110 shape distribution determined at step 1225, the combined cuttings 109 size distribution determined at step 1202, the combined cavings 110 size distribution determined at step 1202, the combined cuttings 109 shape distribution determined at step 1203, the combined cavings 110 shape distribution determined at step 1203, the true volumes calculated at step 1305, and/or any combination thereof. In some embodiments, the true volume is also compared with an expected volume from a cuttings transport model that is determined at step 1401 of the operation 1400.

[0084] Wellbore conditions assessed include cleaning sufficiency and wellbore instability. The condition of the wellbore can be assessed by determining the contents of rock fragments 107 in the samples 106 including characterizing the size, shape, and volumes of cuttings 109 and cavings 110 both individually and as a whole. For instance, angular cavings 110 are generated from shear failure of the borehole wall and are typically associated with a lack of mud 108 density. Splintery cavings 110 stem from tensile failure in near-balance or underbalance conditions while drilling low permeability formations and are typically associated with a lack of mud 108 density. These types of early or real-time assessments lead to quick remedial action, which can minimize undesirable events, such as stuck pipes, that generate NPT. Exemplary assessments based on data collected using method 1000 can be seen in FIG. 11 .

[0085] In some embodiments, the sensor computer 111 and the remote host 112 disclosed herein may include a central processing unit (CPU), a memory, and support circuits. The sensor computer 111 is configured to take action in response to measured data received from the 3D laser scanner 104 and the high-definition camera 105, and in response to other inputs such as user inputs. The remote host 112 is configured to take action in response to measured data received from the 3D laser scanner 104 and the high-definition camera 105, and in response to other inputs such as user inputs and inputs from the sensor computer 111. The CPU is a general purpose computer processor configured for use in an industrial setting for monitoring and controlling a detection system and operations related thereto. The memory described herein may include random access memory, read only memory, floppy or hard disk drive, or other suitable forms of digital storage, local or remote. The support circuits are conventionally coupled to the CPU and comprise cache, clock circuits, input output subsystems, power supplies, and the like, and combinations thereof. Software instructions (program) and data can be coded and stored within the memory for instructing a processor within the CPU. A software program readable by the CPU in the sensor computer 111 , includes code, which when executed by the processor, takes action relating to operating the detection system 100 and detecting using the detection system 100. The program will include instructions that are used to control the various hardware and electrical components within the detection system 100 to perform the various tasks used to implement the operational schemes described herein. [0086] Above described are one or more embodiments of the present disclosure. Additionally, the present disclosure incorporates herein by reference in its entirety, the attached article titled “Automatic Determination of Cuttings and Cavings Properties for Hole Cleaning and Wellbore Stability Assessment using a Laser-Based Sensor”.

[0087] It is contemplated that any one or more elements or features of any one disclosed embodiment or example may be beneficially incorporated in any one or more other non-mutually exclusive embodiments or examples. While the foregoing is directed to embodiments of the present disclosure, other and further embodiments of the disclosure may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.

[0088] It will be appreciated by those skilled in the art that the preceding embodiments are exemplary and not limiting. It is intended that all modifications, permutations, enhancements, equivalents, and improvements thereto that are apparent to those skilled in the art upon a reading of the specification and a study of the drawings are included within the scope of the disclosure. It is therefore intended that the following appended claims may include all such modifications, permutations, enhancements, equivalents, and improvements. The present disclosure also contemplates that one or more aspects of the embodiments described herein may be substituted in for one or more of the other aspects described. The scope of the disclosure is determined by the claims that follow.

[0089] The following claims are not intended to be limited to the aspects shown herein, but are to be accorded the full scope consistent with the language of the claims. Within a claim, reference to an element in the singular is not intended to mean “one and only one” unless specifically so stated, but rather “one or more.” Unless specifically stated otherwise, the term “some” refers to one or more. No claim element is to be construed under the provisions of 35 U.S.C. §112(f) unless the element is expressly recited using the phrase “means for”. All structural and functional equivalents to the elements of the various aspects described throughout this disclosure that are known or later come to be known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims. Moreover, nothing disclosed herein is intended to be dedicated to the public regardless of whether such disclosure is explicitly recited in the claims.

Claims

WHAT IS CLAIMED IS:

1 . A method for assessing wellbore conditions, comprising: collecting a sample comprising rock fragments from a drilling fluid; determining a rock fragment size distribution; determining a rock fragment shape distribution; determining a raw volume of the rock fragments; determining a volume correction factor using the rock fragment size distribution and the rock fragment shape distribution; determining an adjusted volume of the rock fragments from the raw volume using the volume correction factor; and assessing a wellbore condition based on the adjusted volume of the rock fragments and/or the rock fragment shape distribution.

2. The method of claim 1 , wherein the rock fragments comprise one or more of cavings and cuttings.

3. The method of claims 1 or 2, further comprising: capturing a 2D image of the sample and a 3D point cloud of the sample; and segmenting the 2D image and the 3D point cloud, wherein each rock fragment of the sample is identified in the 2D image and the 3D point cloud.

4. The method of any of claims 1 -3, wherein determining the rock fragment size distribution comprises determining a distribution of lengths of the rock fragments and a distribution of widths of the rock fragments.

5. The method of any of claims 1 -4, wherein determining the rock fragment shape distribution comprises classifying the rock fragments by morphology.

6. The method of any of claims 1 -5, wherein determining the adjusted volume of the rock fragments further comprises determining a stacking factor using the rock fragment size distribution.

7. A method for determining a volume of rock fragments collected from a drilling fluid, comprising: capturing an image of a sample of the rock fragments ; identifying mud in the sample using the image; segmenting the image of the sample to identify individual rock fragments in the sample; determining a size distribution of the rock fragments based on the segmented image; determining a shape distribution of the rock fragments based on the segmented image; and determining a volume of the rock fragments using a volume correction factor, wherein the volume correction factor is determined using the size distribution of the rock fragments and the shape distribution of the rock fragments.

8. The method of claim 7 further comprising preprocessing the image before segmenting the image.

9. The method of claims 7 or 8, wherein the rock fragments comprise one or more of cuttings and cavings.

10. A method for calculating a volume of cavings in a sample of rock fragments collected from a drilling fluid, comprising: determining a cavings size distribution and a cavings shape distribution of the sample; determining a raw volume of the cavings; calculating a cavings volume correction factor using the cavings size distribution and cavings shape distribution; and scaling the raw volume of the cavings using the cavings volume correction factor.

11 . The method of claim 10 further comprising: capturing a 2D image of the sample and a 3D point cloud of the sample; identifying mud in the sample using the 2D image of the sample; and segmenting the 2D image and the 3D point cloud of the sample to identify individual cavings in the sample.

12. The method of claim 11 further comprises differentiating the cavings from cuttings in the sample based on the segmented 2D image and the segmented 3D point cloud.

13. The method of any of claims 10-12, wherein determining the caving size distribution comprises determining a length distribution of the cavings and a width distribution of the cavings.

14. The method of any of claims 10-13, wherein determining the cavings shape distribution comprises classifying the cavings based on morphological characteristics.

15. The method of claim 14, wherein the morphological characteristics of the cavings comprise circularity or angularity.

16. The method of claims 14 or 15, wherein classifications of the cavings comprise angular, round, tabular, elongated, or reworked.

17. The method of any of claims 10-16, wherein the volume correction factor is determined using a stacking factor and the cavings shape distribution.

18. The method of claim 17, wherein the stacking factor is determined using the cavings size distribution.

19. A method for assessing a wellbore condition comprising: collecting a sample comprising a plurality of rock fragments and mud; characterizing the plurality of rock fragments in the sample, comprising: capturing a 2D image of the sample and a 3D point cloud of the sample; identifying mud in the sample based on the 2D image; preprocessing the 2D image and the 3D point cloud; segmenting the 2D image and the 3D point cloud to identify each of the plurality of rock fragments; determining morphological characteristics of each of the plurality of rock fragments based on the segmented 2D image and the segmented 3D point cloud; classifying the plurality of rock fragments based on the morphological characteristics of the plurality of rock fragments; determining a size distribution of the plurality of rock fragments based on the morphological characteristics of each of the plurality of rock fragments; and determining a shape distribution of the plurality of rock fragments based on the morphological characteristics of each of the plurality of rock fragments; determining a calculated true volume of the plurality of rock fragments, comprising: determining a raw volume of the plurality of rock fragments based on the 3D point cloud; calculating a volume correction factor based on a stacking factor and the shape distribution of the plurality of rock fragments; and scaling the raw volume using the volume correction factor; and assessing the wellbore condition based on the calculated true volume of the plurality of rock fragments.

20. The method of claim 19, wherein the plurality of rock fragments comprise cuttings, cavings, or both.