System and method for acoustic monitoring of a trayed distillation column

Abstract

A method is provided for detecting an operating condition of a vessel including two fluid phases moving past one another, which may include providing a fiber optic cable about an exterior surface of the vessel, receiving time domain data including distributed acoustic sensing measurements from the fiber optic cable at a plurality of locations along the exterior surface of the vessel, determining frequency domain data at each of the plurality of locations based on the time domain data, performing an analysis of the frequency domain data and / or the time domain data using a pre-trained model, and determining an operating condition of the vessel based on the analysis.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of and priority to U.S. patent application Ser. No. 63 / 287,201, entitled "SYSTEMS AND METHODS FOR ACOUSTIC MONITORING OF TRAYED DISTILLATION COLUMNS," filed on December 8, 2021, the entire contents of which are incorporated by reference into this disclosure.

[0002] FIELD OF THE DISCLOSURE This disclosure relates to monitoring of distillation columns, and more particularly to systems and methods for acoustic monitoring of trayed distillation columns. [Background technology]

[0003] A distillation column can be used to separate a mixture of two liquids into its component parts. The distillation column can be filled with the liquid mixture and heated to a temperature higher than the boiling point of one of the components but lower than the boiling point of the other component. This allows the component with the lower boiling point to evaporate into a gas while the component with the higher boiling point remains liquid. The gaseous component is separated from the liquid component and condensed back into a liquid in another vessel, thereby separating the mixture into its component parts.

[0004] A tray distillation column may comprise multiple trays stacked vertically within the column. The mixed liquid may enter the top of the tray distillation column, flow across the length of each tray, and descend through a series of downcomers. When the mixture is heated and one of the components becomes a gas, the gas may rise through the trays to the top of the tray distillation column where it may be separated from the remaining liquid.

[0005] During operation, a tray distillation column may experience many operating conditions that may degrade its performance. Being able to detect these operating conditions may allow engineers to correct them and improve the performance of the tray distillation column. One way to detect the performance of a tray distillation column is to use gamma scanning to generate high resolution information about the distillation column. Gamma scanning can detect the operating conditions of a distillation column, but it only provides a snapshot of a moment in time and cannot be set up to continuously scan the distillation column. It is also expensive and involves radioactive sources, which presents inherent safety risks.

[0006] Temperature and pressure measurements can be made on a tray column while it is in operation. While these measurements can provide some indication of the operating condition of the column, the information obtained is limited and usually does not provide a complete understanding of local performance. Furthermore, temperature and pressure measurements generally require positional sensing elements throughout the column wall of the column. This limits the ability to install sensing elements in an operating column and makes it more expensive to retrofit existing columns. Thus, there is a need for an improved method of detecting the operating condition of a tray column during operation. Summary of the Invention

[0007] In one embodiment, a method for detecting an operating condition of a vessel including two fluid phases moving past one another may include providing a fiber optic cable around an exterior surface of the vessel, receiving time domain data including distributed acoustic sensing measurements from the fiber optic cable at a plurality of locations along the exterior surface of the vessel, determining frequency domain data at each of the plurality of locations based on the time domain data, performing an analysis of the frequency domain data and / or the time domain data using a pre-trained model, and determining an operating condition of the vessel based on the analysis.

[0008] In another embodiment, a method for training a model to detect operating conditions of a vessel including two fluid phases moving past one another may include providing a fiber optic cable around an exterior surface of the vessel; receiving time domain data from the fiber optic cable at a plurality of locations along the exterior surface of the vessel including distributed acoustic sensing measurements at a plurality of time steps, labeling the time domain data at each time step based on the operating condition of the vessel at the time step at which the time domain data was collected; determining frequency domain data associated with each of the operating conditions at each of the plurality of locations based on the time domain data; and training a model to determine the operating conditions of the vessel based on an association between the frequency domain data and / or the time domain data and the labeled operating conditions.

[0009] In another embodiment, a system for determining an operating state of a vessel including two fluid phases moving past one another may include the vessel, a fiber optic cable disposed about an exterior surface of the vessel, and a control unit. The control unit may be configured to receive time domain data including distributed acoustic sensing measurements from the fiber optic cable at a plurality of locations along the exterior surface of the vessel, determine frequency domain data at each of the plurality of locations based on the time domain data, perform an analysis of the frequency domain data and / or the time domain data using a pre-trained model, and determine the operating state of the vessel based on the analysis. [Brief description of the drawings]

[0010] The embodiments set forth in the drawings are illustrative and exemplary in nature and are not intended to be limiting on the scope of the claims. The following detailed description of exemplary embodiments can be understood when read in conjunction with the following drawings, in which like structure is indicated with like reference numerals and in which: [Figure 1] 1 illustrates a schematic depiction of a portion of an exemplary tray distillation column according to one or more embodiments shown and described herein. [Figure 2A]1 depicts an external view of a tray distillation column according to one or more embodiments shown and described herein. [Figure 2B] 2B is a schematic diagram illustrating an interior view of the tray distillation column of FIG. 2A. [Figure 2C] FIG. 2B depicts an exterior view of the tray distillation column of FIG. 2A with fiber optic cables disposed about its surface, according to one or more embodiments shown and described herein; [Figure 2D] 1 illustrates a schematic depiction of a system for performing acoustic monitoring of a tray distillation column, according to one or more embodiments shown and described herein. [Diagram 3] 2E depicts a schematic diagram of the control unit of FIG. 2D according to one or more embodiments shown and described herein. [Figure 4] 4 illustrates a schematic representation of multiple memory modules of the control unit of FIG. 3 in accordance with one or more embodiments shown and described herein. [Diagram 5] 1 depicts an illustration of classification of different driving conditions using a trained model, according to one or more embodiments shown and described herein. [Figure 6] 4 depicts a flowchart of an exemplary method for training the control unit of FIG. 3 to perform acoustic monitoring of a tray distillation column, according to one or more embodiments shown and described herein. [Figure 7] 4 depicts a flowchart of an exemplary method for operating the control unit of FIG. 3 to perform acoustic monitoring of a tray distillation column, according to one or more embodiments shown and described herein. [Figure 8A] 1 depicts an illustration of a regression analysis for determining the amount of jet flooding for sample data. [Figure 8B] 1 depicts an illustration of a regression analysis for determining the amount of downcomer backup for sample data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The embodiments disclosed herein describe systems and methods for acoustic monitoring of a trayed distillation column. Figure 1 shows a portion of an exemplary trayed distillation column 100. In the example of Figure 1, the distillation column 100 includes two trays 102 and 104, and three downcomers 106, 108, and 110. The trays 102 and 104 have sieve holes along their surfaces.

[0012] 1, liquid 112, which is a mixture of two components to be separated, flows down downcomer 106, along tray 102, down downcomer 108, along tray 104, and down downcomer 110. As liquid 112 is heated, one component of the mixture becomes vapor 114, which rises through sieve holes in trays 104, 102 and can be separated from the remaining liquid component of liquid 112. Vapor 114 passes through the sieve holes in trays 102, 104, which prevents liquid 112 from leaking through the sieve holes.

[0013] During operation of the tray distillation column 100, several operating conditions can occur that degrade its performance, such as weeping, foaming, fouling, jet flooding, and downcomer backup. Weeping is when liquid falls or leaks through the sieve holes in the trays 102, 104, instead of being lifted by the vapor 114 and flowing over the sieve holes. Foaming is when the liquid 112 expands and forms bubbles, preventing proper contact between the liquid 112 and the vapor 114. Fouling is when foreign objects obscure passages on the trays, thereby preventing proper flow of the vapor 114 through the trays 102, 104. Jet flooding is when foam entrainment builds up below the trays, preventing proper flow of the vapor 114 through the trays 102, 104. Downcomer backup is when the flow of liquid 112 through the distillation column 100 backs up due to overflow of one or more of the downcomers 106, 108, 110, thereby preventing the proper flow of liquid 112 through the distillation column 100. Other operating conditions that degrade the performance of the tray distillation column 100 can also occur.

[0014] Embodiments disclosed herein relate to systems and methods for detecting different operating conditions of a tray distillation column while the tray distillation column is in operation. In particular, embodiments disclosed herein utilize acoustic monitoring to detect operating conditions of a tray distillation column.

[0015] In an embodiment, a fiber optic cable is wrapped around a tray distillation column. The fiber optic cable can function like multiple microphones placed at different positions along the distillation column by determining strain measurements at each of those positions. Thus, the fiber optic cable can be used as a sensor to detect acoustic sensing measurements at different times (e.g., every 2 meters at a frequency of 20 kHz) at various points along the distillation column. The acoustic sensing measurements can be used to determine the operating state of the distillation column using a pre-trained model using the techniques described herein.

[0016] A model can be trained using the techniques described herein to determine the operating state of a distillation column based on acoustic sensing measurements received by a fiber optic cable wrapped around the distillation column. In one example, a small-scale model of a distillation column (referred to herein as an experimental column) can be used in an experimental environment, for which flow rates and other parameters can be controlled to artificially generate various operating states in the experimental column. Thus, in this example, acoustic sensing measurements can be collected for each artificially induced operating state and labeled accordingly. A model can then be trained to classify and / or determine the degree of the operating state of the experimental column based on the labeled data collected during each of the artificially induced operating states, as disclosed herein. Training a model for an experimental column can be used as a proof of concept.

[0017] In another example, in an actual distillation tower (referred to herein as an industrial tower), it may not be possible to artificially generate operating conditions in the industrial tower. Thus, in this example, acoustic sensing measurements may be collected from the industrial tower while it is operating during normal operation. Over time, it is expected that the industrial tower will naturally experience various operating conditions. Thus, supplemental data (e.g., flow rate, temperature, and pressure data) may be collected from the industrial tower along with the acoustic sensing measurements. The operating condition of the industrial tower may then be determined based on the supplemental data, and the acoustic sensing data may be labeled based on the determined operating condition. A model may be trained to classify and / or determine the degree of the operating condition of the industrial tower based on the labeled data from the industrial tower. Once a model is trained for either the laboratory tower or the industrial tower, the appropriate model may be used to determine the operating condition of the laboratory tower or the industrial tower during real-time operation.

[0018] Figure 2A depicts an exterior view of an exemplary tray column 200 (e.g., a tray distillation column), and Figure 2B depicts a schematic diagram of the internal components of tray column 200. In the embodiment of Figures 2A and 2B, tray column 200 is a laboratory column, as described above. However, in other embodiments, tray column 200 may be an industrial column, as described above.

[0019] 2A, tray tower 200 includes a vapor outlet 202, a liquid inlet 204, a center gas inlet 206, a bottom gas inlet 208, and a liquid outlet 210. Tray tower 200 also includes a number of trays 212a, 212b, 212c, 212d, 212e, 212f, and a number of downcomers 214a, 214b, 214c, 214d, 214e, 214f. However, in other embodiments, tray tower 200 may include any number of inlets, outlets, trays, and downcomers.

[0020] During operation, liquid may be introduced into liquid inlet 204 and may flow across each of the trays and down each of the downcomers to move downward through tray tower 200. The liquid may be heated to evaporate one component of the liquid into a vapor, which may rise upward through the trays of tray tower 200. The embodiments disclosed herein may detect the operating conditions of tray tower 200 in real time while tray tower 200 is operating. Although the embodiments disclosed herein are directed to detecting the operating conditions of a tray distillation tower, the systems and methods disclosed herein may also be used to detect the operating conditions of any vessel that includes two fluid phases moving past each other (e.g., other types of distillation towers, such as packed towers, trickle bed reactors, etc.).

[0021] FIG. 2C illustrates an example of a tray tower 200 with a fiber optic cable 216 wrapped around its exterior. As disclosed herein, the fiber optic cable 216 can act as a sensor to capture acoustic sensing measurements from the tray tower 200. As shown in FIG. 2C, the fiber optic cable 216 can be wrapped around the tray tower 200 for most of its length. A laser pulse can be propagated through the fiber optic cable 216 and the backscattered light can be measured by the sensor. The backscattered light can be affected by axial dynamic distortion or displacement of the optical fiber of the fiber optic cable 216, which can be caused by an acoustic signal output by the tray tower 200. Thus, the measurement of the backscattered light can be indicative of the acoustic signature of the tray tower 200.

[0022] In an embodiment, the amplitude and phase of the acoustic signal from the tray tower 200 may be measured. In the illustrated embodiment, Rayleigh backscattering is measured to determine the amplitude and phase of the acoustic signal. However, in other embodiments, other measurements of backscattered light may be made (e.g., Brillouin backscattering) to determine the acoustic signal from the tray tower 200.

[0023] Experiments have shown that the acoustic signal output by tray tower 200 may vary depending on the operating state of tray tower 200. More specifically, each type of operating state may have a unique acoustic signal that may be measured. Thus, measuring the acoustic signal of tray tower 200 detected by fiber optic cable 216 may provide an indication of the operating state of tray tower 200 as described in further detail herein. Furthermore, by wrapping fiber optic cable 216 around and along the length of tray tower 200, acoustic signals may be detected at different points along tray tower 200. Thus, the location at which a particular operating state occurs may be determined.

[0024] 2D depicts a schematic diagram of a system 220 for performing acoustic monitoring of a trayed distillation column. In the embodiment of FIG. 2D, system 220 includes trayed column 200 of FIGS. 2A-2C, a distributed acoustic sensor (DAS) interrogator 230, an edge server 240, and a control unit 300.

[0025] Tray tower 200 may be either a laboratory tower or an industrial tower as described above. DAS interrogator 230 may receive distributed acoustic sensing measurements from fiber optic cable 216 and transmit the received data to edge server 240. In the illustrated embodiment, edge server 240 is a computing device located in the same facility as tray tower 200. However, in other embodiments, edge server 240 may be located remotely from tray tower 200.

[0026] The edge server 240 can receive acoustic sensing measurements from the DAS interrogator 230 and can perform unsupervised machine learning methods to exploit patterns inherent in the space-time matrix generated by the acoustic sensing measurements and build a low-dimensional representation of those data. Such a representation can then be used to monitor the process and detect changes in this raw unlabeled data set. In particular, after the model is trained by the control unit 300, as described below, the control unit 300 can monitor the tray tower 200 in real time to determine its operating status based on the acoustic sensing measurements from the fiber optic cable 216. However, the fiber optic cable 216 may output a large amount of data (e.g., several TB / hour). Thus, if all the data received by the DAS interrogator 230 is forwarded to the control unit 300, the control unit 300 may not have enough computing resources to process the data within a reasonable time. Thus, the edge server 240 may filter the data sent to the control unit 300.

[0027] In an embodiment, edge server 240 can monitor data received from DAS interrogator 230 and use unsupervised machine learning techniques to analyze the received data. In particular, edge server 240 can determine when the data received from DAS interrogator 230 changes in a manner that suggests the operating state of tray tower 200 has changed. When this occurs, edge server 240 can begin transmitting the data received from DAS interrogator 230 to control unit 300 such that control unit 300 can determine the operating state of tray tower 200 using the techniques described herein.

[0028] In some embodiments, system 220 may not include an edge server 240. In these embodiments, all of the data received by DAS interrogator 230 is forwarded to control unit 300. Additionally, in embodiments where system 220 includes an edge server 240, edge server 240 may be omitted or may forward all of the data received by DAS interrogator 230 during model training by control unit 300. During model training, it may be desirable to utilize as much data as possible to train the model. Thus, it may not be desirable to filter the data sent to control unit 300. Thus, edge server 240 may not filter the data sent to control unit 300 during model training.

[0029] 3 illustrates an exemplary configuration of the control unit 300 of FIG. 2. In the illustrated embodiment, the control unit 300 may be a remote computing device (e.g., a cloud computing device). However, in other embodiments, the control unit 300 may be a computing device co-located with the edge server 240. In some embodiments, the edge server 240 and the control unit 300 may be combined into a single computing device. In the illustrated embodiment, the control unit 300 includes one or more processors 302, a communication path 304, one or more memory modules 306, a data storage component 308, and network interface hardware 310, the details of which are described in the following paragraphs.

[0030] Each of the one or more processors 302 may be any device capable of executing machine-readable and executable instructions. Thus, each of the one or more processors 302 may be a controller, an integrated circuit, a microchip, a computer, or any other physical or cloud-based computing device. Algorithms including trained models, signal pre-processing, and noise reduction methods described below may be executed by the one or more processors 302. The one or more processors 302 are coupled to a communication path 304 that provides signal interconnectivity between various modules of the control unit 300. Thus, the communication path 304 may communicatively couple any number of processors 302 to each other, allowing the modules coupled to the communication path 304 to operate in a distributed computing environment. Specifically, each of the modules may operate as a node that may transmit and / or receive data. As used herein, the term "communicatively coupled" means that the coupled components may exchange data signals with each other, such as, for example, electrical signals through a conductive medium, electromagnetic signals through the air, optical signals through optical waveguides, and the like.

[0031] Thus, the communication path 304 may be formed from any medium capable of transmitting a signal, such as, for example, conductive wires, conductive traces, optical waveguides, etc. In some embodiments, the communication path 304 may facilitate the transmission of wireless signals, such as WiFi, Bluetooth, Near Field Communication (NFC), etc. Furthermore, the communication path 304 may be formed from a combination of media capable of transmitting a signal. In one embodiment, the communication path 304 includes a combination of conductive traces, conductive wires, connectors, and buses that cooperate to enable the transmission of electrical data signals to components such as processors, memories, sensors, input devices, output devices, and communication devices. Additionally, it is noted that the term "signal" refers to waveforms (e.g., electrical, optical, magnetic, mechanical, or electromagnetic), such as DC, AC, sine waves, triangle waves, square waves, vibrations, etc.

[0032] The control unit 300 includes one or more memory modules 306 coupled to the communication path 304. The one or more memory modules 306 may comprise RAM, ROM, flash memory, a hard drive, or any device capable of storing machine-readable and executable instructions such that the machine-readable and executable instructions may be accessed by the one or more processors 302. The machine-readable and executable instructions may include, for example, logic or algorithms written in any programming language of any generation (e.g., first, second, third, fourth, or fifth generation languages), such as machine language that may be directly executed by a processor, or assembly language, object-oriented programming (OOP), scripting languages, microcode, etc. that may be compiled or assembled into machine-readable and executable instructions and stored on the one or more memory modules 306. Alternatively, the machine-readable and executable instructions may be written in a hardware description language (HDL), such as logic implemented via a field-programmable gate array (FPGA) configuration or an application-specific integrated circuit (ASIC), or the like. Thus, the methods described herein may be implemented in any conventional computer programming language, as pre-programmed hardware elements, or as a combination of hardware and software components. The memory module 306 is described in more detail below in conjunction with FIG. 4.

[0033] 3 , the example control unit 300 includes a data storage component 308. The data storage component 308 may store data received from the edge server 240 or data generated by the control unit 300. The data storage component 308 may also store other data used by various components of the control unit 300. In particular, the data storage component 308 may store parameters of a trained model for determining the operating state of the tray tower 200. The model may be referred to herein as the “model.”

[0034] 3, the control unit 300 includes network interface hardware 310 for communicatively coupling the control unit 300 to the fiber optic cable 216. As such, the network interface hardware 310 can transmit data to the tray tower 200 and / or receive data from the edge server 240. The network interface hardware 310 may include wired and / or wireless connections to the tray tower 200 and / or the edge server 240. In other embodiments, the network interface hardware 310 can transmit data to and / or receive data from other computing devices. The network interface hardware 310 may be any device that can be communicatively coupled to the communication path 304 and that can transmit and / or receive data over a network. As such, the network interface hardware 310 can include a communication transceiver for transmitting and / or receiving any wired or wireless communication. For example, the network interface hardware 310 may include antennas, modems, LAN ports, Wi-Fi cards, WiMax cards, mobile communication hardware, short range communication hardware, satellite communication hardware, and / or any wired or wireless hardware for communicating with fiber optic cables 216 and / or other networks and / or devices.

[0035] 4, the one or more memory modules 306 include an acoustic data receiving module 402, a supplemental data receiving module 404, a driving state determination module 406, an ambient noise determination module 408, a frequency domain transform module 410, an ambient noise cancellation module 412, a frequency selection module 414, a model training module 416, a driving state classification module 418, and a driving state quantification module 420. Each of the acoustic data receiving module 402, the supplemental data receiving module 404, the driving state determination module 406, the ambient noise determination module 408, the frequency domain transform module 410, the ambient noise cancellation module 412, the frequency selection module 414, the model training module 416, the driving state classification module 418, and the driving state quantification module 420 may be a program module in the form of an operating system, application program modules, and other program modules stored in the one or more memory modules 306. Such program modules may include, but are not limited to, routines, subroutines, programs, objects, components, data structures, etc. for performing particular tasks or implementing particular data types, as described below.

[0036] The acoustic data receiving module 402 may receive data from the edge server 240, including acoustic sensing measurements captured by the fiber optic cables 216 disposed about the tray tower 200. In particular, the acoustic data receiving module 402 may receive distributed acoustic sensing measurements from multiple locations along the exterior of the tray tower 200. The acoustic data receiving module 402 may receive the data during training of the model or during operation of the tray tower 200 after the model has been trained.

[0037] The acoustic data receiving module 402 may receive time domain data from the fiber optic cable 216 (via the edge server 240). That is, the acoustic data receiving module 402 may receive backscattered data from the fiber optic cable 216 at multiple points along the tray tower 200 and at multiple time points (e.g., every 2 m at a frequency of 10 kHz). The received time domain data may then be converted to frequency domain data by the frequency domain conversion module 410, as described in more detail below. Once the data is received by the acoustic data receiving module 402, the data may be stored in the data storage component 308.

[0038] In embodiments where tray tower 200 is an industrial tower, supplemental data receiving module 404 may receive supplemental data from tray tower 200 other than acoustic sensing measurements detected by fiber optic cable 216. For example, supplemental data receiving module 404 may receive supplemental data from tray tower 200 including flow rate, temperature, and pressure values. These data can be used to determine the operating conditions of tray tower 200 when training a model. The supplemental data can be detected by various sensors that measure various parameters of tray tower 200.

[0039] In an embodiment where the tray tower 200 is an industrial tower, the operating state determination module 406 can determine an operating state of the tray tower 200 based on the supplemental data received by the supplemental data receiving module 404. In an embodiment where the tray tower 200 is a laboratory tower, an operator can artificially cause the tray tower 200 to enter a particular operating state by adjusting the flow rate or other parameters of the tray tower 200. The received acoustic sensing data can then be labeled based on the operating state caused by the operator.

[0040] However, in embodiments where the tray tower 200 is an industrial tower, it may not be possible to have the tray tower 200 experience different operating states on demand. Thus, in these embodiments, the system 220 may allow the tray tower 200 to operate naturally and over time, and it is expected that the tray tower 200 will experience different operating states as it continues to operate. However, in order to label the received acoustically sensed data, it is necessary to determine the operating state of the tray tower 200 at any given time. Thus, the supplemental data receiving module 404 may receive supplemental data from the tray tower 200, and the operating state determination module 406 may determine the operating state of the tray tower 200 based on the supplemental data received by the supplemental data receiving module 404. The acoustically sensed data may then be labeled based on the operating state determined by the operating state determination module 406.

[0041] The ambient noise determination module 408 can determine ambient noise data in the environment in which the tray tower 200 is located. As described above, the fiber optic cable 216 measures acoustic signals from the tray tower 200. However, if ambient noise is present around the tray tower 200, the fiber optic cable 216 may detect the ambient noise that may interfere with the data collected by the fiber optic cable 216. Thus, the ambient noise determination module 408 can determine the ambient noise, which can be subtracted from the data collected by the fiber optic cable 216 to remove this interference, as described in more detail below. In some implementations, the ambient noise determination module 408 can determine the ambient noise based on acoustic sensing data received from multiple locations of the fiber optic cable 216 wrapped around the tray tower 200 in a position where no fluid flows. In other implementations, the ambient noise determination module 408 can determine the ambient noise based on acoustic sensing data received when the tray tower 200 is not operating. The ambient noise determination module 408 can also determine a timestamp when each ambient noise data is determined. Thus, the ambient noise data identified by the ambient noise determination module 408 may be matched with the acoustic data received by the acoustic data receiving module 402 .

[0042] The frequency domain transform module 410 may transform the time domain data received by the acoustic data receiving module 402 into frequency domain data. The frequency domain data may include features used to train a model to detect operating conditions of the tray tower 200. After training, the frequency domain data may include features that are input into the trained model.

[0043] The frequency domain transform module 410 may determine frequency domain data including amplitude and phase data at a plurality of frequency values ​​at each of the locations along the tray tower 200 where data is received by the acoustic data receiving module 402. The frequency domain transform module 410 may utilize a variety of techniques for converting the time domain data to frequency domain data including Fourier transform, fast Fourier transform, short-time Fourier transform, wavelet analysis, and the like. In the illustrated embodiment, the frequency domain transform module 410 determines amplitude and phase data at 12,500 different frequencies ranging from 0 Hz to 12,500 Hz in 1 Hz intervals. However, in other embodiments, the frequency domain transform module 410 may determine frequency data at any number of frequency values.

[0044] The ambient noise removal module 412 can remove the ambient noise identified by the ambient noise determination module 408 from the frequency domain data generated by the frequency domain transformation module 410. In particular, the frequency domain transformation module 410 can transform the acoustic sensing measurements representing the ambient noise determined by the ambient noise determination module 408 into the frequency domain and can remove the transformed ambient noise data from the frequency domain data. Thus, the ambient noise removal module 412 can determine adjusted frequency domain data including the frequency domain data generated by the frequency domain transformation module 410 from which the ambient noise signal has been removed. Thus, the adjusted frequency domain data determined by the ambient noise removal module 412 can remove any interference caused by the ambient noise. In the illustrated embodiment, the ambient noise removal module 412 removes the ambient noise from the frequency domain data output by the frequency domain transformation module 410, but in other embodiments, the ambient noise removal module 412 can remove the ambient noise from the time domain data received by the acoustic data receiving module 402.

[0045] The frequency selection module 414 may be utilized during training to select particular frequencies from among the frequency values ​​for which frequency domain data is obtained by the frequency domain transform module 410. As described above, in the illustrated embodiment, the frequency domain transform module 410 obtains amplitude and phase information for 12,500 different frequency values. However, not all of these frequency values ​​contain significant signals across the operating conditions of the tray tower 200. Thus, the frequency selection module 414 may use regression analysis to select a subset of frequency values ​​that captures the most information associated with the operating conditions. The subset of frequencies selected by the frequency selection module 414 may be referred to herein as condensed frequency domain data.

[0046] In an embodiment, all of the frequency values ​​that are part of the frequency domain data determined by the frequency domain transform module 410 include candidate frequencies (e.g., 12,500 candidate frequencies in the illustrated example). The expert can then review the frequency domain data to determine which candidate frequencies contribute a non-insignificant amount of signal across the driving conditions determined by the driving state determination module 406. The frequency selection module 414 can then be programmed to remove frequency values ​​that have no signal or an insignificant amount of signal across the driving conditions. Thus, the frequency selection module 414 can select a subset of frequency values ​​from among the multiple frequency values ​​determined by the frequency domain transform module 410 that captures the most information associated with the driving conditions.

[0047] The model training module 416 can train a model to determine an operating state of the tray tower 200 as disclosed herein. The model training module 416 can train a model to classify an operating state of the tray tower 200 into one of a plurality of operating states and / or to determine a degree of an operating state of the tray tower 200. In the illustrated embodiment, the model training module 416 can train a model to classify an operating state of the tray tower 200 into one of normal operation, weeping, fouling, and foaming as disclosed herein. Additionally, the model training module 416 can train a model to determine a degree of jet flooding and downcomer backup experienced by the tray tower 200 as disclosed herein.

[0048] In an embodiment, the model training module 416 can train a model to classify the frequency domain data into one of a plurality of driving states using multivariate discriminant analysis. In some examples, the model training module 416 can train a model to classify the frequency domain data and the time domain data into one of a plurality of driving states using multivariate discriminant analysis. In the illustrated example, the model training module 416 trains a model to classify the frequency domain data into one of a plurality of driving states using orthogonal partial least squares discriminant analysis (OPLS-DA). However, in other examples, other types of multivariate analysis (e.g., partial least squares discriminant analysis) may be used to train the model.

[0049] In an embodiment, the model training module 416 can perform the steps of determining the number of principal components, selecting important variables, and estimating parameter values ​​corresponding to the selected variables. In the first step, a number of principal components are selected. In particular, the model training module 416 can determine the principal components of the adjusted frequency domain data output by the ambient noise removal module 412. The model training module 416 may also determine the scores of the principal components for different data sets (e.g., different driving conditions). An expert may then determine the number of principal components to be included in the model. The expert may determine the number of principal components that sufficiently explain the data without overfitting.

[0050] In a second step, the model training module 416 can select important variables using variable importance in projection (VIP). In particular, the model training module 416 can use VIP to determine how many of the frequencies selected by the frequency selection module 414 to include in the model.

[0051] In a third step, the model training module 416 can estimate parameter values ​​corresponding to the selected variables. That is, the model training module 416 can estimate parameter values ​​corresponding to the selected frequencies using the VIP. The three steps described herein may be repeated in an iterative manner to train the model.

[0052] In one embodiment, the model training module 416 can train a model to classify the frequency domain data into specific operating conditions based on the scores of the principal components. For example, FIG. 5 shows exemplary data collected from a tray column 200 under several different operating conditions. In the embodiment of FIG. 5, the data was collected from a laboratory column. The laboratory column used in the embodiment of FIG. 5 uses water and air, which are fed to the top and bottom of the device, respectively. By adjusting the flow rates of the water and air, different operating conditions can be generated within the device.

[0053] In the example of FIG. 5, data was collected from a fiber optic cable wrapped around a device during four different operating states: normal operation, weeping, forming, and water injection only. During each operating state, time domain data from the fiber optic cable was collected. This time domain was converted to frequency domain data, and an OPLS-DA model was built for classification using the techniques described above. FIG. 5 shows a two-dimensional plot of the first two principal components. As can be seen from FIG. 5, each operating state tends to generate frequency domain data with principal components that fall within a certain range. Thus, the model training module 416 can train a model using the frequency domain data collected during different operating states to determine a particular operating state and visualize the results in a two-dimensional plot.

[0054] In addition to training a model to determine a particular operating condition, the model training module 416 can also train a model to determine the extent of a particular operating condition. For example, the model training module 416 can train a model to determine the amount of jet flooding or downcomer backup occurring within the tray tower 200 based on the frequency domain data and / or the time domain data.

[0055] In an embodiment, the acoustic data receiving module 402 can collect acoustic data associated with different degrees of different driving conditions (e.g., jet flooding or downcomer backup). The frequency domain transform module 410 can transform this time domain data into frequency domain data, which can be used training data that the model training module 416 can use to train a model to determine the degree of one or more driving conditions. In the illustrated embodiment, the model training module 416 trains a model to determine the degree of a driving condition using partial least squares (PLS) regression. However, in other embodiments, the model training module 416 may train a model to determine the degree of a driving condition using other algorithms (e.g., other types of regression).

[0056] 8A and 8B show an example prediction of the extent of an operating condition for tray tower 200 using sample data determined by the trained model. In the example of FIG. 8A and FIG. 8B, data was collected from a laboratory tower. In the example of FIG. 8A and FIG. 8B, PLS was used to determine the extent of the operating condition. FIG. 8A shows the predicted amount of jet flooding in tray tower 200 and FIG. 8B shows the predicted amount of downcomer backup in tray tower 200 for the sample data.

[0057] 4 , after model training module 416 has trained the model, control unit 300 can use the trained model to determine in real time an operating state and / or a degree of an operating state of tray tower 200 based on the received acoustic sensing measurements detected by fiber optic cable 216. In particular, as disclosed herein, operating state classification module 418 can determine an operating state of tray tower 200, and operating state quantification module 420 can determine a degree of an operating state of tray tower 200.

[0058] The operating state classification module 418 may classify the condensed frequency domain data determined by the frequency selection module 414 using a pre-trained model. In particular, the operating state classification module 418 may classify the condensed frequency domain data into one of a plurality of operating states using multivariate discriminant analysis. In some embodiments, the operating state classification module 418 may classify the condensed frequency domain data using the condensed frequency domain data and the time domain data received by the acoustic data receiving module 402. In the illustrated embodiment, the operating state classification module 418 may classify the operating state of the tray tower 200 into one of normal operation, weeping, fouling, or forming. However, in other embodiments, the operating state classification module 418 may classify the operating state of the tray tower 200 into other operating states.

[0059] The operating condition quantification module 420 may use the pre-trained model to determine the extent of an operating condition of the tray tower 200. In particular, the operating condition quantification module 420 may use the pre-trained model to determine the extent of an operating condition of the tray tower 200 based on the condensed frequency domain data determined by the frequency selection module 414 and / or the time domain data received by the acoustic data receiving module 402. In the illustrated embodiment, the operating condition quantification module 420 may determine the extent of jet flooding and downcomer backup during operation of the tray tower 200. However, in other embodiments, the operating condition quantification module 420 may determine the extent of other operating conditions of the tray tower 200.

[0060] FIG. 6 depicts a flow chart of an exemplary method for training a model to detect an operating state and / or degree of an operating state of the tray tower 200. In step 600, the acoustic data receiving module 402 receives acoustic data sensing measurements from the fiber optic cable 216 wrapped around the tray tower 200 (via the DAS interrogator 230). This data may include time domain acoustic data collected from multiple points along the tray tower 200. In particular, the time data may be labeled based on the operating state of the tray tower 200 when the data was collected. In an embodiment where the tray tower 200 is a laboratory tower, the data may be labeled based on the operating state caused by an operator by varying a parameter of the tray tower 200. In an embodiment where the tray tower 200 is an industrial tower, the data may be labeled based on the operating state determined by the operating state determination module 406 based on the supplemental data received by the supplemental data receiving module 404.

[0061] After receiving the data, the acoustic data receiving module 402 may store the received data in association with the labeled operating conditions in the data storage component 308. Thus, the control unit 300 may collect acoustic data associated with various operating conditions from the tray tower 200 that may be used to train models as disclosed herein.

[0062] In step 602, the frequency domain transform module 410 transforms the time domain data to the frequency domain to obtain frequency domain data, which includes amplitude and phase data at multiple frequency values ​​at multiple locations along the tray tower 200.

[0063] In step 604, the ambient noise determination module 408 determines an amount of ambient noise associated with the tray tower 200. The ambient noise removal module 412 then removes the ambient noise from the frequency domain data determined by the frequency domain transformation module 410 to obtain conditioned frequency domain data.

[0064] In step 606, the model training module 416 trains a model to classify the conditioned frequency domain data into one of a plurality of driving states using multivariate discriminant analysis. In particular, the model training module 416 trains a model to classify the conditioned frequency domain data into a particular driving state based on the principal components determined using OPLS-DA. In the illustrated embodiment, the plurality of driving states includes at least normal driving, weeping, fouling, and foaming.

[0065] During model training, the frequency selection module 414 may select a number of frequencies from among the frequency domain data to determine condensed frequency domain data. In particular, the frequency selection module 414 may use regression analysis to select a subset of frequency values ​​from among the frequency values ​​determined by the frequency domain transformation module 410 such that each frequency value in the subset of frequency values ​​captures more information associated with the operating state of the tray tower 200 than each frequency value not in the subset of frequency values. That is, the frequency selection module 414 may select a number of frequency values ​​that contain the most information for detecting the operating state of the tray tower 200, as determined by one or more experts.

[0066] In step 608, the model training module 416 trains a model to determine the extent of an operating condition of the tray tower 200, such as jet flooding or downcomer back-up. In an embodiment, the model training module 416 may use regression analysis (e.g., PLS regression) to train a model to determine the extent of the operating condition of the tray tower 200. After the model is trained, the determined parameters of the model may be stored in the data storage component 308, and the trained model may be used to determine the operating condition and / or the extent of the operating condition of the tray tower 200 during operation.

[0067] 7 depicts a flowchart of an exemplary method for using a trained model to determine the operating state and / or degree of operating state of tray tower 200 during operation. At step 700, acoustic data receiving module 402 receives acoustic data sensing measurements from fiber optic cable 216 wrapped around tower 200. This data may include time domain acoustic data collected from multiple points along tray tower 200. In some embodiments, acoustic data receiving module 402 may receive data from edge server 240 after edge server determines, via unsupervised learning techniques, that the data has changed in a manner that suggests the operating state of tray tower 200 may have changed.

[0068] In step 702, the frequency domain transform module 410 transforms the time domain data to the frequency domain to obtain frequency domain data, which includes amplitude and phase data at multiple frequency values ​​at multiple locations along the tray tower 200.

[0069] In step 704, the ambient noise determination module 408 determines an amount of ambient noise associated with the tray tower 200. The ambient noise removal module 412 then removes the ambient noise from the frequency domain data determined by the frequency domain transformation module 410 to obtain conditioned frequency domain data.

[0070] In step 706, the operating state classification module 418 uses the trained model to determine the operating state of the tray tower 200 based on the conditioned frequency domain data output by the ambient noise cancellation module 412. In particular, the frequency selection module 414 condenses the conditioned frequency domain data to include only amplitude and phase data associated with frequency values ​​that contain the most information for determining the operating state of the tray tower 200, thereby obtaining condensed frequency domain data. The operating state classification module 418 then uses the trained model to determine the operating state of the tray tower 200 based on the principal components of the condensed frequency domain data.

[0071] In step 708, the operating condition quantification module 420 uses the trained model based on the condensed frequency domain data and / or time domain data to determine the degree of the operating condition of the tray tower 200. After determining the classification and degree of the operating condition of the tray tower 200, the control unit 300 can output the determined operating condition and the degree of the operating condition to a user.

[0072] It should now be understood that the embodiments described herein are directed to systems and methods for acoustic monitoring of a trayed distillation column. A fiber optic cable can be wrapped around the exterior of the trayed distillation column to measure acoustic signals at various points along the column. A model can be trained to determine the operating state and degree of operating state of the column using labeled data that associates the acoustic data with different operating states.

[0073] The model can be trained to select the frequency values ​​that have the most information for determining different operating states of the distillation column. The model can be trained to classify the frequency domain data into one of a plurality of operating states using multivariate analysis. The model can also be trained to determine the degree of operating state of the distillation column using regression analysis.

[0074] After the model is trained, acoustic data can be collected from a fiber optic cable wrapped around the distillation column while the column is operating in real time. The trained model can then be used to classify and / or determine the degree of operating condition of the tray distillation column based on the acoustic data. Thus, the operating condition of the tray distillation column can be determined at various points along the column during real-time operation of the column.

[0075] It should be noted that the terms "substantially" and "about" may be utilized herein to express the degree of inherent uncertainty that may result from any quantitative comparison, value, measurement, or other representation. These terms are also utilized herein to express the extent to which a quantitative representation may vary from the stated basis without resulting in a change in the basic functionality of the subject matter at issue.

[0076] While particular embodiments have been illustrated and described herein, it should be understood that various other changes and modifications can be made without departing from the spirit and scope of the claimed subject matter. Moreover, although various aspects of the claimed subject matter have been described herein, such aspects need not be utilized in combination. It is therefore intended that the appended claims cover all such changes and modifications that are within the scope of the claimed subject matter.

Claims

1. 1. A method for detecting an operating condition of a vessel containing two fluid phases moving past one another, comprising: providing a fiber optic cable about an exterior surface of the container; receiving time domain data comprising distributed acoustic sensing measurements from the fiber optic cable at a plurality of locations along the exterior surface of the container; determining frequency domain data at each of the plurality of locations based on the time domain data; performing an analysis of the frequency domain data and / or the time domain data using a pre-trained model; and determining the operational state of the vessel based on the analysis.

2. 10. The method of claim 1 , wherein the vessel comprises a tray distillation column.

3. 2. The method of claim 1, wherein performing the analysis includes classifying the frequency domain data and / or the time domain data into one of a plurality of driving states using the pre-trained model.

4. The method of claim 3 , wherein the plurality of operating conditions includes at least normal operation, weeping, fouling, and foaming.

5. The method of claim 3 , further comprising classifying the frequency domain data and / or the time domain data into one of the plurality of driving states using multivariate discriminant analysis.

6. The method of claim 1 , wherein performing the analysis includes quantifying the frequency domain data and / or the time domain data to determine a degree of the operating condition of the vessel.

7. The method of claim 6 , wherein the operating conditions include at least one of jet flooding and downcomer backup.

8. receiving a measurement of ambient noise at a location where the container is located; determining frequency domain data based on the time domain data; removing the ambient noise from the frequency domain data to obtain conditioned frequency domain data; performing the analysis of the conditioned frequency domain data using the pre-trained model; and The method of claim 1 further comprising:

9. 2. The method of claim 1, further comprising determining the frequency domain data by performing a time domain to frequency domain transform of the time domain data at each of the multiple locations to determine amplitude and phase data at multiple frequency values ​​at each of the multiple locations.

10. selecting a subset of frequency values ​​from among the plurality of frequency values ​​to obtain condensed frequency domain data; performing the analysis of the condensed frequency domain data using the pre-trained model; and The method of claim 9 further comprising:

11. 1. A method for training a model to detect an operating state of a vessel containing two fluid phases moving past one another, comprising: providing a fiber optic cable about an exterior surface of the container; receiving time domain data from the fiber optic cable at a plurality of locations along the exterior surface of the vessel, the time domain data including distributed acoustic sensing measurements at a plurality of time steps, the time domain data at each time step being labeled based on the operating state of the vessel at the time step at which the time domain data was collected; determining the frequency domain data associated with each of the driving conditions at each of the plurality of locations based on the time domain data; and training the model to determine the operating state of the vessel based on the association between the frequency domain data and / or the time domain data and the labeled operating states.

12. generating a plurality of operating conditions within the vessel; labeling the time domain data based on the operating conditions induced in the vessel at each time step; The method of claim 11 further comprising:

13. receiving supplemental data associated with the vessel at each time step; determining the operational state of the vessel at each time step based on the supplemental data; and labeling the time domain data based on the driving state determined at each time step; The method of claim 11 further comprising:

14. receiving a measurement of ambient noise at a location where the container is located; determining frequency domain data based on the time domain data; removing the ambient noise from the frequency domain data to obtain conditioned frequency domain data; training the model to determine the operating state of the vessel based on the association between the conditioned frequency domain data and the labeled operating states; The method of claim 11 further comprising:

15. 12. The method of claim 11, further comprising using regression analysis to select a subset of frequency values ​​from among the plurality of frequency values, such that each frequency value of the subset of frequency values ​​captures more information associated with the labeled driving state than each frequency value not in the subset of frequency values.

16. The method of claim 15 , wherein the plurality of operating conditions includes at least normal operation, weeping, fouling, and foaming.

17. The method of claim 11 , further comprising training the model to determine a degree of the driving condition based on the frequency domain data and / or the time domain data.

18. The method of claim 17 , wherein the operating conditions include at least one of jet flooding and downcomer backup.

19. 1. A system for determining an operating state of a vessel containing two fluid phases moving past one another, comprising: The container; a fiber optic cable disposed about an exterior surface of the container; A control unit, receiving time domain data comprising distributed acoustic sensing measurements from the fiber optic cable at a plurality of locations along the exterior surface of the container; determining frequency domain data at each of the plurality of locations based on the time domain data; performing an analysis of the frequency domain data and / or the time domain data using a pre-trained model; a control unit configured to determine the operating state of the vessel based on the analysis.

20. 20. The system of claim 19, wherein the control unit is further configured to use multivariate discriminant analysis to classify the frequency domain data and / or the time domain data into one of a plurality of operating states and / or to determine a degree of the operating state of the vessel.

Images