TORCH MONITORING SYSTEM AND PROCEDURES

Abstract

A flare monitoring system includes a camera configured to capture one or more images of a burning section of hydrocarbon gas emitted from a flare stack, a storage circuit in which instructions are stored, and a processing circuit configured to execute the instructions to estimate a flow rate of the hydrocarbon gas based on first data corresponding to the one or more images and based on second data corresponding to an inner diameter of the flare stack.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to and the benefit of U.S. Provisional Application No. 63 / 314,247, entitled “EDGE AI BASED FLARE MONITORING TO REDUCE GLOBAL WARMING,” filed February 25, 2022, which is hereby incorporated by reference in its entirety for all purposes. STATE OF THE ART

[0002] This section is intended to introduce the reader to various aspects of the prior art that may be related to various aspects of the present disclosure and are described below. We believe this discussion is helpful in providing the reader with background information that will facilitate a better understanding of the various aspects of the present disclosure. Accordingly, it should be understood that these statements are to be understood in this sense and not as an admission of prior art.

[0003] Facilities that extract, refine, and / or process hydrocarbons (e.g., oil) may flare (e.g., burn) hydrocarbon gases, such as stranded natural gas, for a variety of reasons. The environmental impacts of gas flaring can be limited or reduced through proper operation and maintenance of a flare stack, flare header, and other associated equipment. In other words, the environmental impacts of gas flaring can be exacerbated by improper operations and / or lack of maintenance of the flare stack, flare header, and other associated equipment. For example, if the flow rate of hydrocarbon gas in the flare stack and / or flare header deviates from a desired flow rate, environmental impacts can be increased.

[0004] In conventional designs, a flow rate sensor, such as an ultrasonic flow rate sensor, can be used to monitor the flow rate of the hydrocarbon gas. Unfortunately, flow rate sensors can be expensive and prone to failure. Accordingly, it has been recognized that improved flare monitoring systems and techniques are desirable. SUMMARY

[0005] The following is a summary of certain embodiments disclosed herein. It should be noted that these aspects are presented merely to provide the reader with a brief summary of these particular embodiments, and that these aspects are not intended to limit the scope of this disclosure. Indeed, this disclosure may include a variety of aspects not listed below.

[0006] In one embodiment, a flare monitoring system includes a camera configured to capture one or more images of a burning portion of a hydrocarbon gas emitted from a flare stack, a memory circuit storing instructions, and a processing circuit configured to execute the instructions to estimate a flow rate of the hydrocarbon gas based on first data corresponding to the one or more images and second data corresponding to an inner diameter of the flare stack.

[0007] In another embodiment, instructions are stored on one or more tangible, non-transferable, computer-readable media that, when executed by one or more processors, are configured to cause the one or more processors to receive first data corresponding to one or more camera-captured images of a burning portion of a hydrocarbon gas emitted from a flare stack, receive second data corresponding to an inner diameter of the flare stack, and estimate a flow rate of the hydrocarbon gas through the flare stack or a flare header upstream of the flare stack based on the first data and the second data.

[0008] In another embodiment, a method includes capturing one or more images of a burning portion of a hydrocarbon gas emitted from a flare stack via a camera assembly. The method also includes receiving, via a processing circuit, first data corresponding to one or more images. The method also includes estimating, via the processing circuit, a flow rate of the hydrocarbon gas through the flare stack or a flare header upstream of the flare stack based on the first data and based on second data corresponding to an inner diameter of the flare stack. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Various aspects of this disclosure can be better understood by reading the following detailed description and viewing the drawings in which: Fig. 1 is a schematic view of a system including a gas flare assembly and a flare monitoring assembly in accordance with one aspect of the present disclosure; Fig. 2 is an illustration of an image of a flare taken by a camera of the flare monitoring assembly of Fig. 1, and a reproduction of the image with superimposed rectangular bounding boxes taken by the flare monitoring assembly of Fig. 1, in accordance with one aspect of the present disclosure; Fig. 3 is an illustration of an image of a flare taken by a camera of the flare monitoring assembly of Fig. 1, and a reproduction of the image with superimposed polygon masks taken by the flare monitoring assembly of Fig. 1, in accordance with one aspect of the present disclosure; Fig. 4 is a process flow diagram corresponding to a process performed by the flare monitoring assembly of Fig. 1 is carried out to determine key performance indicators (KPIs), such as a flow rate of gas, according to the gas flare assembly of Fig. 1, in accordance with one aspect of the present disclosure; Fig. 5 is a schematic view of a portion of the flare monitoring assembly of Fig. 1 is, in accordance with one aspect of the present disclosure; Fig. 6 is a flowchart of the process illustrating one or more algorithms used in the process of Fig. 4, in accordance with one aspect of the present disclosure; and Fig. 7 is a representation of the key performance indicators (KPIs) that can be displayed on a graphical user interface (GUI) provided by the flare monitoring assembly of Fig. 1 and the first and second flare stacks of the gas flare assembly of Fig. 1, in accordance with one aspect of the present disclosure. DETAILED DESCRIPTION

[0010] One or more specific embodiments are described below. In an effort to provide a concise description of these embodiments, the specification does not describe all features of an actual implementation. It should be appreciated that, as with any engineering or design project, numerous implementation-specific decisions must be made in developing such an actual implementation to meet the developers' specific goals, such as adhering to system-related and business constraints that may vary from one implementation to another. Furthermore, it should be noted that while such a development effort can be complex and time-consuming, it is a routine design, manufacturing, and fabrication effort for those skilled in the art having the benefit of this disclosure.

[0011] When introducing elements of various embodiments of the present disclosure, the terms "a" and "an" and "the" are intended to refer to one or more of these elements. The terms "comprising," "including," and "having" are intended to be inclusive and mean that there may be additional elements besides those listed. Furthermore, it should be noted that references to "embodiment" or "an embodiment" of the present disclosure are not intended to exclude the existence of additional embodiments that also include the recited features.

[0012] The present disclosure relates to the use of a camera and various processing functions to monitor a flare corresponding to the combustion of a hydrocarbon gas exiting a flare stack. A flare monitoring system includes, for example, a camera, a memory storing instructions, and processing circuitry configured to execute the instructions to perform various functions. The camera may be configured to capture various images (e.g., via a video stream), and the processing circuitry may be configured to estimate a flow rate of the hydrocarbon gas within the flare stack or a flare header downstream of the flare stack based on data indicative of the images and an inner diameter of the flare stack (e.g., a tip of the flare stack).For example, the processing circuitry may execute one or more algorithms to determine the size of the flare (e.g., via rectangular bounding boxes, polygon masks, or both) and then estimate the flow rate of the hydrocarbon gas based on the size of the flare and the inner diameter of the flare stack.

[0013] Other inputs can also be used to estimate the hydrocarbon gas flow rate. For example, the hydrocarbon gas pressure in the flare stack or flare header, the hydrocarbon gas temperature in the flare stack or flare header, the wind speed near the flare stack, the flare angle, the hydrocarbon gas composition, other system-related parameters, or any combination thereof can be used to estimate the hydrocarbon gas flow rate. Various sources corresponding to the inputs described above can be used, such as sensors (e.g., for pressure, temperature, wind speed, or any combination thereof), a historical model (e.g., an autoregressive model (ARIMA) that estimates wind speed based on historical wind speed data), and / or other sources.In addition, at least one additional key performance indicator (KPI) other than the hydrocarbon gas flow rate may be determined by the processing circuit based on one or more outputs of the flare monitoring system.

[0014] By utilizing the features described above, the hydrocarbon gas flow rate can be determined (e.g., within a margin of error) without the need for a conventional flow rate sensor, such as an ultrasonic sensor, which can be expensive and prone to failure. Furthermore, the operation of the flare stack, flare header, or other downstream equipment can be adjusted based on one or more results from the flare monitoring system to limit or otherwise reduce the environmental impacts of hydrocarbon gas flaring. These and other features are described in detail below with reference to the drawings.

[0015] Fig. 1 shows a schematic representation of a system 10 including a gas flare assembly 12 and a flare monitoring assembly 14. The gas flare assembly 12 may include, for example, a flare stack 16, a flare header 18 downstream of the flare stack 16, components 20 of the flare header 18, and additional components 22 downstream of the flare header 18. While only one flare stack 16 is shown in the depicted embodiment, multiple flare stacks 16 may be used in other embodiments.

[0016] The additional components 22, the flare header 18 (e.g., including components 20), and the flare stack 16 can be controlled to direct a hydrocarbon gas 23 (e.g., natural gas) to a tip 24 of the flare stack 16, where the hydrocarbon gas 23 is combusted to produce a flare 26. For example, an igniter 28 and a pilot flame tip 30 can be coordinated to produce a pilot light (or flame) that combusts the hydrocarbon gas 23 upon exiting the tip 24 of the flare stack 16, thus producing the flare 26. It should be noted that the igniter 28 and the pilot flame tip 30, among other components, may be Fig. 1 are shown schematically and may look different and / or be positioned in different locations in certain embodiments of the present disclosure.

[0017] The flare monitoring assembly 14 may include a control device 30 (e.g., controller, control assembly, etc.) having a memory circuit 32 storing instructions and a processing circuit 34 configured to execute the instructions to perform various functions. The control device 30 may also include communication circuitry 36 configured to establish communication between the control device 30 and various other components described below. The control device 30 may also include a display 38. In some embodiments, the display 38 may include a user interface (UI) (e.g., a touchscreen) configured to accept input from a user. In other embodiments, the control device 30 may include a UI separate from the display 38, such as aa keyboard, a mouse, buttons, or a combination thereof. Furthermore, it should be noted that each of the memory circuits 32, the processing circuits 34, and the communication circuits 36 may include one or more components. For example, the processing circuit 34 may correspond to one or more processors.

[0018] The flare monitoring assembly 14 may also include a camera assembly 40 (e.g., one or more cameras, such as one or more RGB cameras), a wind sensor 42, a gas pressure sensor 44, a gas temperature sensor 46, an ambient air sensor 48, and / or a gas composition / density sensor 49. The type of camera(s) used in the camera assembly 40 may be selected based on the desired speed and / or accuracy of processing. For example, the embodiments presented here may quickly estimate the flow rate and / or other KPIs using an inexpensive, low-resolution daylight camera (e.g., 2 megapixels), although more expensive, higher-resolution cameras may also be employed (e.g., for greater accuracy). The sensors 42, 44, 46, 48, 49 are Fig. 1 is shown adjacent to the controller 30 for simplicity, but it should be understood that the sensors 42, 44, 46, 48, 49 may be located at other locations. For example, the wind sensor 42 and the ambient air sensor 48 may be located near the top 24 of the flare stack 16 and / or the flare 26, and the gas pressure sensor 44, the gas temperature sensor 46, and the gas composition / density sensor 49 may be located within the flare stack 16 or the flare header 18. As described in more detail below, in certain embodiments, one or more of the sensors 42, 44, 46, 48, 49 may not be used.

[0019] According to the present disclosure, the controller 30 may be configured to receive a first input 50 corresponding to data indicative of images captured by the camera assembly 40 of the flare 26. In some embodiments, the first input 50 may correspond to data indicative of a video (e.g., with various images therein) captured by the camera assembly 40. The controller 30 may also be configured to receive a second input 52 corresponding to data indicative of an inner diameter 54 of the flare chimney 16 (e.g., near the tip 24 of the flare chimney 16 or elsewhere within the flare chimney 16). In some embodiments, the data indicative of the inner diameter 54 of the flare chimney 16 may be input to the controller 30 (e.g., via the display 38 or the UI).In some embodiments, the controller 30 may also be configured to determine various key performance indicators (KPIs) of the gas flare assembly 12, such as the volumetric flow rate of hydrocarbon gas 23 per second through the flare stack 16, based only on the first input 50 and the second input 52. That is, in certain embodiments, the wind sensor 42, the gas pressure sensor 44, the gas temperature sensor 46, the ambient air sensor 48, and / or the gas composition / density sensor 49 are not used or are not required. However, in certain other embodiments, at least one input corresponding to at least one of the sensors 42, 44, 46, 48, 49, as described below, may be used and may improve the accuracy of one or more of the KPIs corresponding to the gas flare assembly 12 and determined by the flare monitoring assembly 14.

[0020] For example, the controller 30 may receive a third input 54 corresponding to data indicative of a wind speed sensed by the wind sensor 42, a fourth input 56 corresponding to data indicative of a pressure of the hydrocarbon gas 23 in the flare stack 16 sensed by the gas pressure sensor 44, a fifth input 58 corresponding to data indicative of an ambient air density sensed by the ambient air sensor 48, and / or a sixth input 60 corresponding to data indicative of a gas composition and / or gas density sensed by the gas composition / density sensor 49. In some embodiments, one or more of the inputs 56, 58, 60, 62 may be used without the use of the corresponding sensor 44, 46, 48, 49. For example, in certain embodiments, a wind speed estimation module 64 may be employed with or without the use of the wind speed sensor 42.In certain embodiments, the gas composition and / or density may be input to the controller 30 (e.g., via the display 38 or the UI) or otherwise known to the controller 30 without using the gas composition / density sensor 49.

[0021] In general, the controller 30 may use at least the first input 50 (e.g., corresponding to the data indicating the images captured by the camera assembly 40) and the second input 52 (e.g., corresponding to the inner diameter 54 of the flare stack 16) to determine one or more KPIs of the gas flare assembly 12. As described in more detail in later drawings, the controller 30 may, for example, perform various functions to identify, based on the first input 50, one or more rectangular bounding boxes (e.g., via object detection) or one or more polygon masks (e.g., via instance segmentation) corresponding to the flare 26. The object detection or instance segmentation, generally referred to herein as computer vision models and / or edge machine learning (EdgeML) models, may be selected based on the desired processing speed and / or accuracy.In certain embodiments, one or both of these models may incorporate various computer vision features, such as subtracting the background from the image(s) by detecting constant or semi-constant background pixels and subtracting them from the constantly moving fire / smoke pixels in the foreground. Other image augmentation features, such as random shearing, rotation, horizontal flipping, and scaling, may also be used. In addition, a correlation tracking model may be employed to learn correlation filters and predict whether fire (or flames) and / or smoke is present in a current image to improve retrieval performance. These and other features are described in detail below.

[0022] The size of the flare (e.g., flame size, smoke size, size of the flare and smoke, etc.) may be derived from one or more rectangular fillet frames or one or more polygon masks, which may be used in conjunction with the second input 52 to estimate one or more KPIs of the gas flare assembly 12, such as the flow rate of the hydrocarbon gas 23 through the flare stack 16. As described above, the third input 54, the fourth input 56, the fifth input 58, the sixth input 60, the seventh input 62, or any combination thereof may also be used by the controller 30 to estimate one or more KPIs of the gas flare assembly 12. In addition to the flow rate of the hydrocarbon gas 23 through the flare stack 16, the controller 30 may determine other KPIs, such as the flow rate of the hydrocarbon gas 23 through the flare stack 16. E.g. flame area, smoke area, smoke-to-flame ratio, flame-to-smoke ratio, flame angle, etc.It should be noted that according to the present disclosure, the KPIs may be estimated by the controller 30 in real-time or near-real-time (e.g., via a real-time streaming protocol (RTSP) at a desired frame rate). Accordingly, any abnormal KPIs may be handled as described below.

[0023] In certain embodiments, controller 30 may modify the operation of one or more components of gas flare system 12 based on one or more of the estimated KPIs. Controller 30 may modify the operation of components 20 in flare header 18, additional components 22 downstream of flare header 16, or other components (e.g., igniter 28, pilot flame tip 30, etc.) within flare stack 16 or elsewhere in gas flare assembly 12. These components may include, for example, a flashback prevention section, relief lines, a liquid separation drum, water drains, oil drains, a flashback sealing drum, a purge gas section, a gas recovery section, valves, pumps, other traditional gas flare system components, etc.Controller 30 may change the operation of such components in response to one or more KPIs satisfying a predefined relationship to one or more KPI thresholds (e.g., when the one or more KPIs exceed the one or more KPI thresholds). For example, controller 30 may change the operation of such components when the estimated flow rate of the gas exceeds a first threshold, falls below a second threshold, or both.

[0024] Fig. 2 and Fig. 3 are examples of rectangular bounding boxes or polygon masks generated by the flare monitoring assembly 14 Fig. 1 based on images from the camera assembly 40 of the flare monitoring assembly 14. Fig. For example, Figure 2 shows an embodiment of an image 80 of the flare 26 obtained by the camera assembly 40 of the flare monitoring assembly 14 of Fig. 1, and a reproduction 82 of the image 80 with superimposed rectangular bounding boxes 84, 86, which was taken by the flare monitoring assembly 14 of Fig. 1. The first rectangular bounding box 84 may correspond to a flame 88 and the second rectangular bounding box 86 may correspond to the smoke 90. In addition, the rectangular bounding boxes 84, 86 may be determined by the control device 30 from Fig. 1 can be determined via an object identification module 91 that uses, for example, an EfficientDet Dx model. Note that the following discussion of exemplary features of an EfficientDet model is incorporated by reference: MINGXING TAN et al., "EfficientDet: Scalable and Efficient Object Detection," CVPR 2020.

[0025] Fig. Figure 3 shows an embodiment of the image 80 of the flare 26 obtained by the camera assembly 40 of the flare monitoring assembly 14 of Fig. 1, and the reproduction 82 of the image 80 with superimposed polygon masks 92, 94, which were taken by the flare monitoring assembly 14 of Fig. 1. The first polygon mask 92 may correspond to the flame 88, and the second polygon mask 94 may correspond to the smoke 90. Furthermore, the polygon masks 92, 94 may be determined based on an instance segmentation module 96 using, for example, a Mask R-CNN (MRCNN). Note that the following discussion of exemplary features of an MRCNN model is incorporated by reference: KAIMING HE et al., "Mask R-CNN," in CVPR, 2018.

[0026] As described in detail below, one or more of the rectangular bounding boxes 88, 90 or one or more of the polygon masks 92, 94 may be used along with at least one other input (e.g., the inner diameter of the flare stack 16) to determine certain KPIs of the gas flare assembly 12 from Fig. 1, such as the flow rate of the hydrocarbon gas 23 through the flare stack 16.

[0027] Fig. Figure 4 is a flow diagram of the process 100 performed by the flare monitoring assembly 14 of Fig. 1 is carried out to determine key performance indicators (KPIs), such as a flow rate of gas, which is assigned to the gas flare assembly 12 of Fig. 1. In the embodiment shown, the process 100 includes processing a video stream (e.g., corresponding to the first input 50 indicative of images captured by the camera assembly 40 of Fig. 1) via the instance segmentation module 96 (e.g., MRCNN model) to generate masks, such as a first mask corresponding to a flame or fire and a second mask corresponding to smoke. The process 100 also includes calculating (block 102) certain KPIs based on the masks and outputting (block 104) those calculations (e.g., fire length, fire centerline, fire area, fire color, smoke area, smoke color, and other parameters). It should be noted that in certain embodiments, a deep learning model may be used to estimate at least the fire length and / or fire centerline (e.g., based on the images or videos captured by the camera assembly) without using the polygon mask(s) and / or rectangular bounding box(es), and that the various KPIs may be estimated based on the fire length and / or fire centerline.

[0028] The process 100 also includes the calculation (block 106) of the gas flow rate based on the fire length (e.g., generated via block 102) and additional inputs. The additional inputs may include, for example, the estimated wind speed and direction, which are calculated in block 108 via a wind speed prediction model and / or based on wind speed sensor data. Furthermore, the inputs for the calculation (block 106) may include the head pressure and temperature and the chimney characteristics (e.g., the tip inner diameter and / or the chimney type, such as high pressure, low pressure, high temperature, or low temperature). Based on the calculation (block 106), which may use an API 521 model or a reference table, the controller 30 may Fig. 1 the flow rate of the gas over the process 100 of Fig. 4. It should be noted that the following discussion of exemplary characteristics of an API 521 model is incorporated by reference: “Pressure-relieving and Depressuring Systems,” API STANDARD 521, SEVENTH EDITION, JUNE 2020.

[0029] Fig. 5 is a schematic view of an embodiment of a portion 150 of the flare monitoring assembly 14 of Fig. 1. Section 150 of the flare monitoring assembly 14 in Fig. 5 includes, as already described, the camera assembly 40, which may include at least one RBC camera configured to capture images (e.g., via a video transmission) of a flare corresponding to a flare stack. A video capture means 152 may be configured to receive the data corresponding to the images from the camera assembly 40. The video capture means 152 may, for example, be configured to capture various components of the flare monitoring assembly 14, such as the camera assembly 40 and one or more of the Fig. 1. The deep learning model interface 154 may be configured to receive data from the means for video recording 152 and implement or otherwise execute the above-mentioned MRCNN model or EfficientDet Dx model to generate polygon masks corresponding to the flare (e.g., including at least one flame or fire mask and at least one smoke mask) or rectangular bounding boxes corresponding to the flare (e.g., including at least one flame or fire frame and at least one smoke frame). Based on the model(s) described above, the KPIs assigned to the gas flare assembly 12 of Fig. 1, are estimated in block 102 and then output to a message bus in block 104.

[0030] Fig. 6 is a flowchart of process 200 illustrating an embodiment of one or more algorithms that may be used, for example, in process 100 of Fig. 4. The flowchart of the process 200 shows a first portion of the algorithm 202, an API 521 model 204 (e.g., a reference table or graph), and a second portion of the algorithm 206, which are used together to output a volumetric flow rate per second 208 (e.g., of the hydrocarbon gas through the flare stack or flare header). The first portion 202 of the algorithm may, for example, include first data 210 indicating the fire length (or flame length) (y), second data 212 indicating the density of the exiting gases (ρ j ), third data 214, which indicates the density of the ambient air (ρ ∞ ), and fourth data 216, which indicates the inside diameter of the chimney top (d j) are received. This data can be used in the first section of algorithm 202 as follows: Y=y[(dj)ρjρ∞

[0031] The value output by the first section of the algorithm 202 or Y can be used by the API 521 model 204 (e.g., reference table or graph) to locate X, as in Fig. 6. In addition, the second section 206 of the algorithm may receive inputs corresponding to X and the wind speed (µ ∞ ), where: X=μ∞μj

[0032] After applying equation 2 to solve µ j the second section 206 of the algorithm can determine the flow rate per second (V) as in Fig. 6 and described below: V=π(dj2)2μj

[0033] Fig. Figure 7 shows an embodiment of key performance indicators (KPIs) that can be displayed on a graphical user interface (GUI) 300 provided by the flare monitoring assembly 14. Fig. 1 and a first and second flare stack 16a, 16b (e.g. the gas flare assembly 12 of Fig.1). For example, first KPIs 302 for the first flare stack 16a and second KPIs 304 for the first flare stack 16a are shown. The first KPIs 302 include the fire (or flame) area 308, the smoke area 310, the smoke-to-fire (or flame) ratio 312, the fire (or flame) to smoke ratio 314, and the fire (or flame) area 316. Likewise, the second KPIs 304 include the fire (or flame) area 318, the smoke area 320, the smoke-to-fire (or flame) ratio 322, the fire (or flame) to smoke ratio 324, and the fire (or flame) area 326.Also displayed on the GUI 300 are a first graph 328 depicting the flow rate over time, a second graph 330 depicting the fire (or flame) length over time, a third graph 332 relating to the first flare stack 16a and depicting the height and width of the fire (or flame) and smoke over time, and a fourth graph 334 relating to the second flare stack 16b and depicting the height and width of the fire (or flame) and smoke over time. Finally, you may include an embedded video 306 of the first flare stack 16a and / or the second flare stack 16b. The embedded video 306 may be a real-time or near-real-time live feed, or the embedded video 306 may be recorded for display on the GUI 300 at a later time.

[0034] The present disclosure may provide one or more technical effects useful in hydrocarbon gas flaring. For example, the embodiments presented herein may enable improved flare monitoring at a lower cost than conventional configurations and reduce the environmental impacts of hydrocarbon gas flaring.

[0035] Although only certain features and embodiments of the disclosure have been illustrated and described, those skilled in the art may make many modifications and changes, such as variations in the sizes, dimensions, structures, shapes, and proportions of the various elements, values ​​of parameters including temperatures and pressures, mounting arrangements, use of materials, colors, orientations, and so forth, without materially departing from the novel teachings and advantages of the subject matter recited in the claims. The order or sequence of any process or method steps may be varied or rearranged according to alternative embodiments. It is therefore to be understood that the appended claims are intended to cover all modifications and changes that comply with the true spirit of the disclosure.In an effort to succinctly describe the example embodiments, not all features of an actual implementation may have been described, e.g., those not related to the current best mode of carrying out the disclosure or those not related to enabling the claimed disclosure. It should be appreciated that, as with any engineering or design project, numerous implementation-specific decisions may be made in developing such an actual implementation. While such a development effort could be complex and time-consuming, it would be a routine undertaking of design, manufacture, and production for one skilled in the art, armed with the knowledge conferred by this disclosure, without undue experimentation. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] US 63 / 314,247

[0001] Cited non-patent literature

[0000] Title “EDGE AI BASED FLARE MONITORING TO REDUCE GLOBAL WARMING”, published on February 25, 2022

[0001] MINGXING TAN et al., “EfficientDet: Scalable and Efficient Object Detection,” CVPR 2020

[0024] KAIMING HE et al., “Masks R-CNN,” In CVPR, 2018

[0025] Pressure-relieving and Depressuring Systems," API STANDARD 521, SEVENTH EDITION, JUNE 2020

[0028]

Claims

[1] Flare monitoring system, comprising: - a camera configured to capture one or more images of a burning portion of a hydrocarbon gas emitted from a flare stack; - memory circuits storing instructions thereon; and - a processing circuit configured to execute the instructions to estimate based on a flow rate of the hydrocarbon gas: - first data corresponding to the one or more images; and - second data corresponding to an inner diameter of the flare chimney. [2] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to: - based on the first data, to determine one or more rectangular bounding boxes corresponding to the burning portion of the hydrocarbon gas; and - estimate the flow rate based on third data corresponding to one or more sizes of the one or more rectangular bounding boxes. [3] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to: - based on the first data and an object identification model, determine a first rectangular bounding box corresponding to the fire of the burning portion of the hydrocarbon gas and a second rectangular bounding box corresponding to the smoke of the burning portion of the hydrocarbon gas; and - estimate a variety of key performance indicators (KPIs) of a gas flaring system using the flare stack based on the first rectangular bounding box and the second rectangular bounding box. [4] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to: - based on the first data and a masking algorithm, determine one or more polygon masks corresponding to the burning portion of the hydrocarbon gas; and - to estimate the flow rate based on third data corresponding to one or more sizes of the one or more polygon masks. [5] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to: - based on the first data and an instance segmentation model, determine a first polygon mask corresponding to the fire of the burning portion of the hydrocarbon gas and a second polygon mask corresponding to the smoke of the burning portion of the hydrocarbon gas; and - to estimate a variety of key performance indicators (KPIs) of a gas flaring system using the flare stack based on the first polygon mask and the second polygon mask. [6] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to estimate the flow rate based on third data corresponding to the wind speed near the burning portion of the hydrocarbon gas. [7] A flare monitoring system according to claim 6, comprising one or more sensors configured to detect wind speed. [8] The flare monitoring system of claim 6, wherein the processing circuit is configured to execute the instructions to estimate the wind speed based on the following: - an autoregressive model (ARIMA) using historical wind speed data; or - a flare angle corresponding to the burning section of the hydrocarbon gas; or - a combination of the ARIMA and the torch angle. [9] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to estimate the flow rate based on third data corresponding to the following: - a pressure of the hydrocarbon gas in the flare stack; or - a temperature of the hydrocarbon gas in the flare stack; or - a material composition of the hydrocarbon gas within the flare stack; or - a combination of two or more of the factors pressure, temperature or composition of the material. [10] The flare monitoring system of claim 1, wherein the processing circuit is configured to execute the instructions to: - to determine a centerline and a length of the burning section of the hydrocarbon gas based on the initial data and a deep learning model; and - estimate a variety of key performance indicators (KPIs) of a gas flare system using the flare stack based on the length of the fire. [11] One or more tangible, non-transferable, computer-readable media having stored thereon instructions that, when executed by one or more processors, are configured to cause the one or more processors to: - receive first data corresponding to one or more images taken by a camera of a burning portion of a hydrocarbon gas emitted from a flare stack; - to receive second data corresponding to an inner diameter of the flare stack; and - estimate, based on the first data and the second data, a flow rate of the hydrocarbon gas through the flare stack or a flare header upstream of the flare stack. [12] One or more tangible, non-transferable, computer-readable media according to claim 11, wherein the instructions, when executed by the one or more processors, are configured to cause the one or more processors to estimate the flow rate based on third data corresponding to a pressure or a temperature of the hydrocarbon gas within the flare stack or the flare header. [13] One or more tangible, non-transitory, computer-readable media according to claim 11, wherein the instructions, when executed by the one or more processors, are configured to cause the one or more processors to estimate the flow rate based on third data corresponding to a pressure of the hydrocarbon gas in the flare stack or flare header and fourth data corresponding to a temperature of the hydrocarbon gas in the flare stack or flare header. [14] The one or more tangible, non-transferable, computer-readable media of claim 11, wherein the instructions, when executed by the one or more processors, are configured to cause the one or more processors to estimate the flow rate based on third data corresponding to a wind speed near a tip of the flare stack. [15] One or more tangible, non-transitory, computer-readable media according to claim 11, wherein the instructions, when executed by the one or more processors, are configured to cause the one or more processors to: - based on the first data, to determine one or more rectangular bounding boxes corresponding to the burning portion of the hydrocarbon gas; and - estimate the flow rate based on one or more sizes of the one or more rectangular bounding boxes. [16] One or more tangible, non-transitory, computer-readable media according to claim 11, wherein the instructions, when executed by the one or more processors, are configured to cause the one or more processors to: - based on the first data and a masking algorithm, determine one or more polygon masks corresponding to the burning portion of the hydrocarbon gas; and - estimate the flow rate of the hydrocarbon gas based on third data corresponding to one or more sizes of the one or more polygon masks. [17] Method comprising: - capturing one or more images of a burning portion of a hydrocarbon gas emitted from a flare stack via a camera assembly; - receiving first data corresponding to the one or more images via a processing circuit; and - Estimating a flow rate of the hydrocarbon gas through the flare stack or a flare header upstream of the flare stack via the processing circuitry based on the first data and based on second data corresponding to an inner diameter of the flare stack. [18] A method according to claim 17, comprising: - determining, via the processing circuit and based on the first data, one or more rectangular rounding frames corresponding to the burning portion of the hydrocarbon gas; - determining one or more sizes of the one or more rectangular rounding frames via the processing circuit; and - Estimating the flow rate based on third data corresponding to the one or more quantities via the processing circuit. [19] A method according to claim 17, comprising: - determining one or more polygon masks corresponding to the burning portion of the hydrocarbon gas via the processing circuit based on the first data and based on a masking algorithm; - determining one or more sizes corresponding to the one or more polygon masks via the processing circuit; and - Estimating the flow rate via the processing circuitry and based on third data corresponding to the one or more quantities. [20] The method of claim 17, comprising estimating the flow rate based on the processing circuit: - third data corresponding to a pressure of the hydrocarbon gas within the flare stack or flare header; - fourth data corresponding to a temperature of the hydrocarbon gas within the flare stack or flare header; and - fifth data corresponding to a wind speed near a tip of the flare chimney.

Images