Method for monitoring tube sheets in heat exchangers

Abstract

Shell-and-tube equipment typically requires periodic maintenance. Described herein is an automated method for tracking the condition of individual tubes during maintenance procedures and recording the condition data for review and analysis. The condition data can optionally be reported in a real-time summary format and / or used to predict time to completion. The method of the present invention helps minimize operational oversight and reduce the cost of performing maintenance procedures in shell-and-tube equipment, including shell-and-tube reactors and heat exchangers.

Description

[Technical Field]

[0001] The present invention relates to a method for monitoring the tube sheet of a heat exchanger. [Background technology]

[0002] Shell-and-tube heat exchangers can have hundreds or even thousands of tubes. Typically, shell-and-tube heat exchangers require regular maintenance, such as cleaning and inspecting individual tubes, to ensure reliable and safe operation. Furthermore, shell-and-tube reactors require periodic catalyst replacement to achieve optimal productivity. Due to the large number of tubes, maintenance work incurs significant labor costs and lengthy process downtime. Therefore, there is a strong economic incentive to perform such work quickly and efficiently. In addition, the installation of catalysts within shell-and-tube reactors requires adherence to precise loading specifications. Failure to properly perform maintenance on each tube within a shell-and-tube reactor can lead to costly process downtime, equipment damage, and shortened catalyst life within the reactor. This specification describes an automated method for tracking the condition of individual tubes during maintenance work and recording condition data for review and analysis. The condition data is reported in a real-time summary format as needed and / or used to predict the time to completion. The method of the present invention helps minimize missed work and reduce the cost of performing maintenance work in shell-and-tube heat exchangers, including shell-and-tube reactors. [Overview of the Initiative] [Means for solving the problem]

[0003] According to one aspect of the present invention, a method for monitoring a tube sheet having a plurality of tube ends arranged in a fixed pattern of rows (R) and columns is provided. This method involves the following steps: (a) Assign a unique identifier to each of the plurality of tube ends; (b) A digital image (D) of at least a portion of the tube sheet is acquired at acquisition time (T); (c) Determine the attribute state of each of the tube ends in the digital image; and (d) Record the data of each tube end in the digital image in the relevant database. The data includes, i. The acquisition time (T), ii. The unique identifier of the tube end, and iii. Attribute status at acquisition time (T) Includes.

[0004] According to another aspect of the present invention, a method for monitoring the status of a shell-and-tube device during maintenance work, comprising the following steps: (a) Assign a unique identifier to each of the tube ends; (b) Select an attribute that has at least two possible states; (c) An initial digital image (Di) of at least a portion of the tube sheet is acquired at acquisition time (Ti); (d) Determine the initial state of the attributes for each of the tube ends in the initial digital image (Di); (e) Creating an initial data record in the relational database for each tube end in the initial digital image (Di) The initial data record includes, i. The initial acquisition time (Ti), ii. The unique identifier of the tube end, and iii. Initial state of attributes at the initial acquisition time (Ti) Includes.

[0005] According to yet another aspect of the present invention, an optical method for monitoring the condition of a shell-and-tube apparatus during maintenance work, comprising the following steps: (a) Assign a unique identifier to each of the tube ends; (b) Arrange at least one digital camera so that at least a part of the tube sheet is within the field of view of at least one digital camera; (c) Place a plurality of colored tube caps on the tube ends, the plurality of tube caps comprising a tube cap having a first color and a tube cap having a second color different from the first color; (d) Acquire an initial digital image (Di) of at least a part of the tube sheet at an acquisition time (Ti); (e) Determine the respective initial colors of the tube ends in the initial digital image (Di); and (f) For each tube end in the initial digital image (Di), create an initial data record in a relational database including, the data record i. the initial acquisition time (Ti), ii. the unique identifier of the tube end, and iii. the initial color of the tube end at the initial acquisition time (Ti) including.

[0006] According to yet another aspect of the present invention, an optical method for monitoring the state of a shell and tube type device (e.g., a reactor) during a granular catalyst loading operation, comprising the following steps: (a) Place a plurality of closing plates on the tube sheet so that all tube ends are covered, each closing plate having a disk recess for installing a colored indicator disk; (b) Assign a unique identifier to each of the plurality of closing plates; (c) Arrange at least one digital camera so that at least a part of the plurality of closing plates is within the field of view of at least one digital camera; (d) Install a plurality of colored indicator disks in the disk recesses, the plurality of colored indicator disks comprising at least one disk having a first color and at least one disk having a second color different from the first color; (e) Acquire the initial digital image (Di) of at least a part of the plurality of closing plates at the initial acquisition time (Ti); (f) Determine the initial color of each of the coloring indicator disks in the initial digital image (Di); and (g) Create an initial data record in the relationship database for each closing plate in the initial digital image (Di) including, the data record i. the initial acquisition time (Ti), ii. the unique identifier of the closing plate, and iii. the initial color of the coloring indicator disk at the initial acquisition time (Ti) is included.

Brief Description of the Drawings

[0007] [Figure 1] FIG. 1 shows a system for monitoring a shell-and-tube type device. [Figure 2A] FIG. 2A shows an exemplary embodiment of a horizontally arranged shell-and-tube heat exchanger. [Figure 2B] FIG. 2B shows one of the tube sheets of the heat exchanger of FIG. 2A. [Figure 3A] FIG. 3A shows another exemplary embodiment of a vertically oriented shell-and-tube heat exchanger. [Figure 3B] FIG. 3B shows one of the tube sheets of the heat exchanger of FIG. 3A. [Figure 4] FIG. 4 shows the visualization of the differential pressure measurement values obtained (or not obtained) in each tube of the heat exchanger of FIG. 3A. [Figure 5] FIG. 5 shows the closing plates applied to the tube sheet of the heat exchanger of FIG. 3A. [Figure 6] FIG. 6 is a schematic diagram of an image capture and data collection process.

Modes for Carrying Out the Invention

[0008] Detailed description of the invention (A) System for monitoring shell-and-tube devices Figure 1 shows a system 100 for monitoring a shell-and-tube device 110. The shell-and-tube device 110, although not necessarily part of the system 100, comprises a hollow shell 112 (partially shown) including a tube sheet 114 attached to its end. The tube sheet 114 has a series of holes 116 defined through it. Tubes 118 are attached to each hole 116 and placed within the hollow shell 112. The shell 112 is shown separated to clearly show the tubes 118. The ends 119 of the tubes 118 are exposed through the holes 116. The tubes 118 and their respective holes / passages may be circular, square, rectangular, etc., as shown.

[0009] The system 100 generally includes an imaging device 120 positioned above the hole 116. The imaging device 120 is configured to view the hole 116, or more generally, to detect the hole 116. As will be described in more detail below, the imaging device 120 may include, for example, a single camera. Alternatively, the imaging device 120 may include multiple imaging devices 120a, 120b for viewing the hole 116 at different angles and viewing angles. The imaging device 120 may be stationary. Alternatively, the imaging device 120 may be mounted on a moving device 122, such as an XYZ translation stage, an XY translation stage, or a vehicle for moving the imaging device 120 relative to the hole 116.

[0010] The imaging device 120 is configured to communicate data regarding the color, state, and / or position (for example) of the tube end 119 to the computer 124. The computer 124 may include an image processor 126, memory 128, a clock 130, programming software 132, and a relational database 134 (part of other features). The processor 126 is configured to analyze the data related to the tube end 119, as will be described later. The computer 124 is connected to a display 140 for displaying the analyzed data, as will be described later. The interconnection between the display 140, the imaging device 120, and the computer 124 may be, for example, wired or wireless.

[0011] Further details and other features related to system 100 and apparatus 110 are described below.

[0012] (B) Shell and tube type apparatus A shell-and-tube device 110 is schematically shown in Figure 1. The shell-and-tube device 110 can constitute part of a heat exchanger, as shown in Figures 2A and 3A. Next, referring to the shell-and-tube heat exchangers 200 and 300 in Figures 2A and 3A, the heat exchangers 200 and 300 generally comprise a shell 112 defining a hollow interior and tubes 118 arranged within the hollow interior.

[0013] As background, shell-and-tube heat exchangers are a common type of heat exchanger used in industry. The name comes from its two main components: one or more heat transfer tubes 118 mounted inside a cylindrical shell 112. The purpose of a shell-and-tube heat exchanger is to transfer heat between two fluids. Each fluid can be a liquid or a gas. In industrial implementations, it is common for at least one of these fluids to be either liquid water or vapor.

[0014] In shell-and-tube heat exchangers 200, 300, one fluid flows inside the tubes 118 (referred to as the "tube-side fluid"), while the other fluid flows outside the tubes 118, but inside the shell 112 (referred to as the "shell-side fluid"). The heat exchanger is configured so that the two fluids do not come into direct contact with each other. Heat is transferred from one fluid to the other by flowing through the walls of the tubes 118 from the tube side to the shell side, or vice versa. To efficiently transfer heat, hundreds to thousands of tubes 118 (collectively called "tube bundles") can be used in a single exchanger.

[0015] The shell-and-tube heat exchangers 200 and 300 also comprise one or more tube sheets, heads, and optionally other components such as baffles, tie rods, spacers, and expansion joints. In particular, tube sheets 114a, 114b, 114c and / or 114d (collectively or individually referred to as tube sheet 114) are attached to the ends of the shell 112. The tube sheet 114 is a plate or forging with planar opposing surfaces and comprises holes 116 into which tubes 118 are inserted. The required thickness of the tube sheet 114 is primarily a function of the operating pressure of the particular shell-and-tube exchanger. The ends of the tubes 118 are secured to the tube sheet 114 by welding or mechanical or hydraulic expansion to prevent mixing of the shell-side fluid with the tube-side fluid.

[0016] The number of tube sheets 114 required depends on the shape of the tube 118. When using a straight tube as shown in Figures 1, 2A, and 3A, two tube sheets 114 may be needed. Alternatively, if the tube 118 is bent into a "U" shape (known as a U-tube), only one tube sheet 114 may be needed.

[0017] The holes 116 in the tube sheet 114 are typically arranged in one of two geometric configurations: triangular or square. The tube sheet 114 utilizes a constant center-to-center distance between adjacent tubes 118, called the "tube pitch." The uniformity of such configurations simplifies the design and construction of the exchanger. A typical tube pitch is 1.25 times the outer diameter of the tube 118. The triangular configuration (see Figure 3B) is often used to obtain high heat transfer and compactness, while the square configuration (see Figure 2B) is generally preferred for maintenance and inspection, where the tube bundle needs to be periodically removed from the shell to clean the outer surface of the tubes.

[0018] Shell-and-tube heat exchangers require heads 220 to accommodate the fluid on the tube side and ensure a desired flow path through the heat exchanger. Typically, each tube sheet 114 has a corresponding head. Heads with a roughly cylindrical shape are called “channels” 222 (see Figure 2A), and heads with a roughly dome shape are called “bonnets” 224 (see Figures 2A and 3A). In some cases, the head may incorporate one or more path baffles 228 (Figure 2A) to guide the fluid flow on the tube side through a particular tube. In these cases, the surface of the tube sheet 114a may further include grooves 230 (Figure 2B) to stabilize the baffles 228 and associated seal gaskets. The heads 220 may be welded in place or attached to the shell 112 using flanges. Flanged bonnets or grooves with removable covers 230 (Figure 2A) are preferred when it is necessary to provide access to the tube sheet 114 and tubes 118 for maintenance and inspection.

[0019] Shell-and-tube heat exchangers 200 and 300 are widely used in industry, for example, in power generation, industrial refrigeration, and petrochemical processing. Shell-and-tube heat exchangers can be installed horizontally (Figure 2A) or vertically (Figure 3A). By convention, in industrial facilities, shell-and-tube heat exchangers are named according to their process function. Typical industrial applications of shell-and-tube heat exchangers include condensers, reboilers, preheaters, boilers, superheaters, quench exchangers, transfer line exchangers (TLEs), evaporators, waste heat boilers, recuperators, cross exchangers, and process heaters. For example, industrial refrigeration systems may consist of both evaporators and condensers, and petrochemical distillation systems may consist of both reboilers and condensers.

[0020] Detailed information regarding shell-and-tube heat exchangers can be found in Perry's Chemical Engineers' Handbook, 6th edition, 2008, particularly in Section 11, Heat-Transfer Equipment, and related Figures 11-1 and 11-2. This handbook is incorporated herein by reference in its entirety for all purposes.

[0021] (C) Another application of shell-and-tube devices The shell-and-tube apparatus 110 can also be integrated into other industrial equipment / process systems, as described below.

[0022] High-strength shell-and-tube heat exchangers with U-tube bundles, such as those disclosed in U.S. Patent No. 4,200,061, are sometimes used as steam generators in nuclear power plants.

[0023] The shell-and-tube apparatus can be incorporated into a gravity-feed film exchanger, such as a gravity-feed film melt crystallizer used for the purification of (meth)acrylic acid.

[0024] Shell-and-tube devices can be incorporated into a reaction system as tightly coupled quench exchangers used to rapidly cool temperature-sensitive products, such as hydrogen cyanide or nitrogen oxides, as they exit the reaction zone, as disclosed in U.S. Patent No. 6,960,333. The aforementioned U.S. patent is incorporated herein by reference in its entirety. Similarly, transfer line exchangers (TLEs) are used to rapidly cool high-temperature process gases as they exit an ethylene furnace.

[0025] In the chemical manufacturing industry, shell-and-tube apparatus 110 is sometimes used as a chemical reactor. Within these so-called “shell-and-tube reactors” (also known as “fixed-bed reactors”), the tube-side fluid typically contains chemical reactants that are converted into one or more chemical products. Generally, commercial-scale shell-and-tube reactors are large apparatuses with 1,000 to 50,000 tubes and tube sheets ranging in diameter from 1 meter to 10 meters. At such scale, the heads of these shell-and-tube reactors can easily enclose a volume large enough for workers to physically enter and work in, and if the shell-and-tube reactor is oriented vertically (as shown in Figure 3A), the upper plane of the upper tube sheet 114c can serve as a de facto “floor” for the enclosed work area.

[0026] In shell-and-tube reactors, one or more granular catalysts are often placed inside the tubes to promote the formation of the desired chemical product. By passing a heat transfer fluid through the shell side of the shell-and-tube reactor, the reaction temperature on the tube side can be precisely controlled, maximizing product yield and extending catalyst life. Temperature control can be further optimized by utilizing unique tube configurations and shell-side baffle designs.

[0027] Chemical transformations carried out in shell-and-tube reactors are either exothermic (reactions that release heat) or endothermic (reactions that absorb heat). For highly exothermic reactions, such as the oxidation of hydrocarbons, it is common to use high-boiling-point fluids such as molten inorganic salts, kerosene, or organic heat transfer fluids (e.g., DOWTHERM®) as the shell fluid. Furthermore, to ensure safe operation at the high temperatures and pressures used in the chemical reaction, custom mechanical designs of the tubes and tube sheets, as well as special structural materials, are typically employed.

[0028] The production of acrylic acid is a well-known example of a commercial hydrocarbon oxidation process using a shell-and-tube reactor. This chemical transformation involves two consecutive exothermic reaction steps in which propylene is first oxidized to the intermediate acrolein, and then acrolein is further oxidized to acrylic acid. Numerous solid mixed metal oxide (MMO) granular catalysts have been developed to facilitate this two-step oxidation process, and the methods for producing these catalysts are well documented in the literature. Inside the reactor, a fixed catalyst bed is assembled by loading one or more granular catalysts into the reactor tubes. As the process gas flows through the tubes, the gas comes into direct contact with the MMO catalyst particles, and the heat of the reaction is transferred through the tube walls to the shell-side coolant.

[0029] Currently, commercial-scale processes for acrylic acid from propylene utilize one of three main configurations of shell-and-tube reactors: tandem reactors, single-shell (SRS) reactors, and single-shell open interstage (SSOI) reactors. As a group, these commercial shell-and-tube reactors may have approximately 12,000 to 22,000 tubes in a single reactor vessel and a production capacity of up to 100 kT / year (220,000,000 pounds / year) of acrylic acid. Some large-scale commercial reactors may have approximately 25,000 to 50,000 tubes in a single reactor vessel and a production capacity of up to 250 kT / year (550,000,000 pounds / year). U.S. Patent No. 9,440,903 describes the configurations and capacities of each of these three reactors for producing acrolein and acrylic acid. This U.S. patent is incorporated herein by reference.

[0030] The production of ethylene oxide is another example of a commercial process using a shell-and-tube apparatus as the reactor. The shell-and-tube apparatus 110 may be provided in the form of a commercial ethylene epoxidation reactor, for example, comprising up to 12,000 tubes. These tubes are typically filled with an epoxidation catalyst, which includes silver and co-catalyst components such as rhenium, tungsten, molybdenum, and chromium, and a coolant is circulated through the shell side of the reactor. See U.S. Patent No. 4,921,681 and U.S. Patent Application Publications 2009 / 0234144 and 2014 / 0135513, which are incorporated herein by reference in their entirety.

[0031] The oxychlorination of ethylene to 1,2-dichloroethane (also known as EDC) is yet another example of a chemical process using a shell-and-tube apparatus. In this process, the tubes in the shell-and-tube apparatus 110 are typically filled with a granular catalyst containing cupric chloride (a so-called "deacon" catalyst), and a coolant is circulated through the shell side of the reactor. In some embodiments, the oxychlorination reaction system may comprise two or more shell-and-tube apparatuses arranged in series. See U.S. Patents 6,180,841, 3,892,816, and 5,905,177, which are each incorporated herein by reference in their entirety.

[0032] In summary, many other commercially important gas-phase catalytic reactions include the conversion of propylene to acrolein and / or acrylic acid (above); the conversion of propane to acrolein and / or acrylic acid; the conversion of glycerin to acrolein and / or acrylic acid; the conversion of t-butanol, isobutene, isobutane, isobutyraldehyde, isobutyric acid, or methyl t-butyl ether to methacrolein and / or methacrylic acid; the conversion of acrolein to acrylic acid; and the conversion of methacrolein to acrylic acid. This process, including conversion to methacrylic acid; conversion from o-xylene or naphthalene to phthalic anhydride; conversion from butadiene or n-butane to maleic anhydride; conversion from indan to anthraquinone; conversion from ethylene to ethylene oxide (above); conversion from propylene to propylene oxide; conversion from isobutene and / or methacrolein to methacrylonitrile; and oxychlorination of ethylene to 1,2-dichloroethane (above), is carried out in a shell-and-tube reactor.

[0033] (D) Shell and tube maintenance Because shell-and-tube systems have numerous tubes 118, completing maintenance and inspection work on each system is extremely time-consuming. Furthermore, tracking the status and progress of maintenance tasks is difficult. Omissions and errors in work also pose significant problems.

[0034] As used herein, the term “work failure” means failure to perform a specific maintenance task on tube 118. For example, a worker may unintentionally skip a tube, resulting in the tube not being cleaned, inspected, or filled with catalyst. The probability of work failures increases with the number of tubes in a shell-and-tube system and with the length of the maintenance work period. Many process owners generally believe that work failures can only be prevented through steps such as (a) continuous monitoring / supervision of workers performing the work, or (b) a 100% inspection after the work is “completed.” The methods of the present invention described herein functionally provide continuous monitoring / supervision of workers performing the work and minimize the need for a 100% inspection.

[0035] In contrast, a "work error" refers to a situation where a task is performed but of insufficient quality, or only partially completed. Examples of work errors include measuring the thickness of a tube wall with an improperly calibrated probe, removing rust only from the first 15 feet of a 20-foot tube, or filling a tube with the wrong type of catalyst. Work errors tend to be relatively unaffected by the number of tubes in a shell-and-tube system. Furthermore, work errors often affect a large number of tubes at once, for example, by filling all the tubes with material supplied from the same, unsuitable pallet of catalyst drums. Addressing work errors in the manner of the present invention improves efficiency and allows for more supervisory resources to be allocated to preventing work errors.

[0036] Numerous maintenance tasks can be performed on the tubes of shell-and-tube systems. These maintenance tasks may involve one or more multi-step tasks, which are typically performed on and repeated for each tube within the shell-and-tube system. Examples of maintenance tasks that can be beneficially monitored using the methods of the present invention include, but are not limited to, the following: a) Examination i. Video inspection of cleanliness and / or mechanical damage ii. Thickness measurement (e.g., eddy current testing) iii. Identification of blocked tubes (e.g., IR detection of low-flow tubes) iv. Identification of organic contamination by reflected ultraviolet light testing b) Cleaning i. Sandblasting ii. CO2 pellet blasting iii. Hydroblast iv. Liquid nitrogen blasting v.perforation vi. Wire brushing vii. Pigging viii. Removal of coke deposits c) Repair i. Re-welding of the weld between the tube and the tube sheet ii. Mechanical stoppers for tubes.

[0037] In the case of shell-and-tube reactors used as reactors, maintenance work may include tasks related to catalyst replacement. Examples of catalyst replacement tasks that can be beneficially monitored using the method of the present invention include, but are not limited to, the following: a) For example, removal of catalyst from the tube using an "air lance", fish tape, or vacuum hose. b) Visually confirm that the tube is empty (i.e., "light check") c) Installation of heat transfer inserts into the tubes, d) Installation of catalytic retainers such as catalytic springs or catalytic clips. e) Loading of ceramic or metal-inert particles into a tube, f) Fill the tube with one or more layers of catalyst. g) Measurement of catalyst bed loss, and h) Measurement of catalyst filling pressure loss (dP).

[0038] (E) Monitoring method for shell-and-tube devices According to one exemplary method for monitoring a shell-and-tube apparatus 110 having multiple tube ends arranged in a fixed pattern of rows (R) and columns (C), this method includes the following general steps: a) Assign a unique identifier to each of the plurality of tube ends, b) A digital image (D) of at least a portion of the tube sheet is acquired at acquisition time (T), c) Determine the attribute state for each of the tube ends in the digital image (D), where the attribute has at least two possible states, d) For each tube end in the digital image (D), data is recorded in a relational database, the data includes the acquisition time (T), the unique identifier of the tube end, and the attribute status at the acquisition time (T), and e) Optionally, create a report using the recorded data stored in the relational database in one or more of the following formats: a table, graph, spreadsheet, or color-coded summary graphic.

[0039] Referring here to the individual steps of an exemplary method using the system 100 shown in Figure 1, the imaging device 120 collects (i.e., acquires) a tube sheet image of the upper surface of the tube sheet 114. The tube sheet 114 may or may not be actively illuminated. The imaging device 120 (e.g., a digital camera) may or may not be initially aligned to a particular row of tube ends 119. When acquiring the image, the imaging device 120 receives light through an aperture representing the state of the tube sheet 114 and converts that light into a set of digital measurements.

[0040] The acquired measurement data is formatted as an array, but is referred to here as a digital image. The digital image of the tube sheet 114 is then transferred to the processor 126 of the computer 124 via Wi-Fi, LAN / PoE (Power over Ethernet) cabling, optical fiber, etc.

[0041] The software 132 of the computer 124 creates a unique tube identifier for each tube end 119 visible in the digital image. First, the image processing software identifies the geometric center of each tube end 119. Next, a unique identifier is assigned to the (x,y) position of each center in the image array. Preferably, the unique identifier for each tube is provided as a set of Cartesian coordinates in (row, column) format, corresponding to the row and column designations used in the manufacturing drawings of the tube sheet. In this way, the software 132 can know which tubes are being displayed in the image array and can uniquely identify each tube.

[0042] The image processing software can determine the position of the geometric center of the tube end 119 by performing the following steps: i. Use the CHT (Circle Hough Transform) function and / or the Canny edge detection algorithm in OpenCV or Matlab® software to identify all regions of geometric interest (i.e., circular tube ends) in the image array. Note that there are specific commands in the software to use these functions. The command returns a variable that represents not only the center coordinates (x, y) of the circle in the array, but also the radius of the circle. ii. Align the image array coordinates with known dimensional data for the tube sheet and map the identified circular tube ends to the tube sheet drawing. Note that this step may be facilitated by using benchmarks within the image to orient the tube sheet drawing. iii. Associate the unique tube identifier (row, column) of each tube in the drawing with the (x,y) position coordinates of the center of each circle in the image array. Since the tube sheet 114 is a stationary component, it generally does not move relative to the image processing device. As a result, the position of each circle center in the array does not change, and this mapping process only needs to be performed once.

[0043] Once software 132 associates the tube ends 119, software 132 then manipulates the image array through various routines using known image processing algorithms such as Canny edge detection, circular Hough transform, and color detection. After each routine, the processed digital data of each tube end 119 (or group of tube ends) is stored in the relational database 134.

[0044] For many of these routines, the processor 126 can analyze only the digital data within a sample window located near the center of each tube end 119, which may be represented, for example, by a 3x3 region containing only 9 pixels. In this way, large areas of the image can be masked out (i.e., ignored) to speed up image processing.

[0045] Next, looking at the various routines, the data in the digital image is processed to determine the attribute details of each tube end 119 in the image. Generally, attributes are features in the image such as shape, color, intensity, and / or texture. Each attribute can generally be described by the presence or absence of one or more specific states. Timestamped data for each tube, including identifiers and details of its attributes, is stored in a relational database 134 (SQL software, etc.) for later analysis. In one embodiment, the timestamp is provided in Julian date format.

[0046] Additional image information, referred to herein as image metadata, may also be stored in the relational database. Image metadata may include GPS coordinates, camera numbers, job descriptions (e.g., "July 2020 inspection"), and / or shell and tube device IDs.

[0047] Workspace parameters can also be stored in a relational database, as will be discussed later. In particular, as mentioned earlier, commercial-scale shell-and-tube reactors may have tube sheets ranging in diameter from 1 to 10 meters. At such scales, the heads of these shell-and-tube reactors can easily enclose a volume large enough for one or more workers to physically enter, creating what is known in industry as a "closed workspace." During maintenance, the environment within such a closed workspace may be controlled to prevent catalyst damage, minimize rust formation in the reactor, and protect workers from potential hazards. Therefore, when performing maintenance, it may be beneficial to measure one or more workspace parameters to better control the environment within the closed workspace.

[0048] For example, climate-controlled air (heated or cooled) can be supplied to a closed workspace to maintain a favorable internal temperature and / or control the relative humidity inside the reactor. In one embodiment, one or more temperature measuring devices can be placed in the ducts of the climate control system and / or within the closed workspace. In another embodiment, one or more Wi-Fi-enabled sensors can be temporarily placed within the closed workspace to continuously monitor the internal relative humidity (%RH). The time-stamped temperature measurements and / or time-stamped %RH measurements are then automatically communicated to a computer 124 via wired or wireless means, stored in a relational database 134, and optionally displayed on a visual display 140.

[0049] In another example, a portable gas analyzer can be used to continuously monitor the air in a closed workspace to detect the presence of hazardous gases (using a so-called "toxic gas detector"), to ensure that sufficient oxygen levels are maintained (using a so-called "oxygen meter"), and / or to monitor for flammability hazards (using a so-called "LEL monitor"). Traditionally, such air monitoring has been performed by an individual known as a "hall watch," and the analyzer's measurement data has been manually recorded on paper log sheets. However, in a preferred embodiment, time-stamped tagged instrument readings from such a gas analyzer may be automatically communicated to a computer 124 via wired or wireless means, recorded in a relational database 134, and optionally displayed on a visual display 140.

[0050] In accordance with safety regulations, it is typically necessary to track the number of workers in a confined workspace and determine their numbers in case of emergency evacuation. Traditionally, this task has also been performed using a "hall watch," which typically involves handwritten log sheets. However, in a preferred embodiment, one or more LiDAR devices, such as a Density Entry Sensor (available from Density Inc. in San Francisco, California, USA), can be mounted above entry and exit points, such as worker passages within reactor heads, to automatically track personnel entering and exiting the workspace. By continuously communicating timestamped entry and exit data to a computer 124 via wired or wireless means, the number of personnel in the workspace during maintenance can be determined in real time. By storing this timestamped workspace occupancy data in a relational database 134, personnel performance metrics, including, for example, personnel efficiency coefficients and downtime, can be calculated.

[0051] Next, referring to Figure 6, one attribute that can be routinely monitored is the intensity of the tube end 119. Specifically, the area of ​​the image within each tube end 119 is evaluated to determine whether the tube end 119 is "dark" or "bright". This attribute can be used, for example, to evaluate the degree of coke accumulation within the tube end 119. If there is a significant accumulation of black coke at the tube entrance, this will appear as a "dark" area in the image of the tube end 119 (see the top tube end 119 in Figure 6). The dark area indicates that the tube 118 requires cleaning or other maintenance. Alternatively, if the coke accumulation is minimal, the tube will still show white inert pellets, and will appear as a "bright" area in the image of the tube end 119 (see the two lower tube ends 119 in Figure 6). Thus, the intensity attribute has two possible states: dark or bright.

[0052] In the example shown in Figure 6, the light energy reflected from three tube ends 119 is collected by the imaging device 120 within the field of view (FOV) of its photodetector. The photodetector measures the intensity of the energy reaching it and represents the individual sensor measurements as digital values. The imaging device 120 transmits these measurements to the image processor 126. Within the processor 126, the collection of related digital measurements, referred to herein as a digital image 133, is stored together as an array. The term “digital image” as used for the purpose of describing Figure 6 is distinct from the conventional term “visual image,” i.e., a photograph from a digital camera or an image on a computer monitor. The digital image 133, which is invisible to the human eye, exists only within the electronic data system. Data visualization software must be used to convert the digital image 133 data into a format suitable for representation as a visual image [meaning a photograph] on a display. In addition to visual images, data visualization software can also be used to present digital image data in one or more summary formats, such as tables, graphs, spreadsheets, or color-coded summary graphics.

[0053] Mathematical operations (commonly known as "image processing") are performed on the above data to provide derived digital images, evaluate the image content (e.g., whether an object is detected within the FOV), and compare multiple digital images to identify changes in object attributes. Note that some image processing can be performed within the circuitry of the imaging device 120 to speed up processing and reduce the amount of data (i.e., required bandwidth) transmitted to the processor 126.

[0054] The processor 126 determines the state of the object attribute ("A1"). In Figure 6, the attribute of the object is "darkness". State 1 (S1) = dark, with intensity measurements between 0 and 4, and State 2 (S2) = light, with intensity measurements between 5 and 9. Thus, the state represents a range of measurements, more specifically, a central value with tolerance. Further details regarding image processing are described in the examples.

[0055] Subsequently, the status data is transferred to the relational database 134 for storage and analysis. The relational database 134 maps attribute statuses (A1) to status values ​​(C1). For example, status 1 (S1) is mapped to status "dirty", and status 2 (S2) is mapped to status "clean". The software in relational database 134 calculates performance metrics for the work (e.g., completion rate, number of out-of-spec items, predicted completion time, timestamp, tube identification "ID").

[0056] Performance metrics are (optionally) transferred to a visual display 140 (such as a digital computer monitor or printer) for real-time reporting. Data visualization software can also be used to visually represent the measurements in a digital image.

[0057] Another attribute that can be monitored in a separate routine is the "texture" of the tube end 119. A "smooth texture" state indicates that no pellets are present at the tube end, while a "rough texture" state indicates that pellets are present at the tube end. This attribute data can be used, for example, to verify that all tubes are properly loaded to the top with inert ceramic balls as intended.

[0058] The above processes and steps may be performed while the apparatus 110 is operating. In one such embodiment, a digital camera may be placed outside the shell-and-tube apparatus to acquire one or more digital images of the tube sheet surface through a well-designed sight glass.

[0059] In yet another routine, system 100 can be used to monitor and track maintenance work performed on the apparatus 110. More specifically, during the process of maintaining the apparatus 110, an operator may place a colored cap (yellow, green, red, etc.) over the inspected tube end 119. The cap may also be referred to as a marker in this specification. Each color is used to indicate a different state of the tube 118, for example, “contains catalyst,” “empty,” or “cleaned.” For example, if an operator confirms that a tube 118 is clean, the operator places a red cap over the tube end 119 of the clean tube 118.

[0060] Once the caps are placed over the tube ends 119, the imaging device 120 is used to collect or acquire images of the capped tube ends 119. The image processing software 132 is then configured to determine which of the possible color state options is applied to the geometric region of interest corresponding to each tube end 119.

[0061] To address the possibility of obstruction of a tube end 119, such as when a worker's tool bucket is set on the tube sheet 114 and covers the group of tubes, one or more “universal” error states may be optionally used in the manner of the present invention. For example, state “U” may be optionally reserved to represent the state “unknown” and assigned to a circular tube end 119 that cannot be detected in the digital image. If the obstruction is later removed and the circular tube end becomes detectable again, the current state may be evaluated and recorded. In another embodiment, a blocked tube end can be addressed by implementing a “hold last” strategy, i.e., by recording the last known state value each time the digital image is processed until the obstruction is removed. Such an approach may further include a corresponding note in the relational database 134 for that group of tubes indicating that the state is “hypothetical”. Optionally, if a detection error occurs, such as obstruction of a tube by the tool bucket, a visual or audible alarm may be triggered to instruct the worker to take corrective action, such as an alert message instructing the worker to remove the obstruction.

[0062] A key advantage of using colored caps is that, using a combination of timestamps and attribute data in the relational database 134, it is possible to compare states in consecutive images and determine when a state change occurred. A change in the state of an attribute is referred to herein as the "behavior" of the attribute. For example, the color behavior (change in color state) of the tube end 119 can be evaluated over a specific period not only to determine when the inspection of the tube 118 was completed, but also to determine the result of that inspection (the state of the tube). Thus, the identified color behavior may be a change from "no cap" to "green cap" at a specific time (e.g., 9:00 a.m.), signaling the point in time when the tube 118 was determined to have passed the mechanical inspection (i.e., a specific time obtained from the image timestamp).

[0063] By using the software of relational database 134 to evaluate the color behavior of all tubes during a specific period, it is possible to (i) generate behavioral metrics such as "number of tubes inspected per hour" or "percentage of tubes that passed inspection," and (ii) predict future behavior such as the remaining time until the inspection work is completed.

[0064] Furthermore, by evaluating all the tubes in this way, the overall condition of the tube sheet 114 at the end of the work can be determined (for example, 98% of the tubes passed inspection), and a database record of the results can be created for future reference.

[0065] The colored caps can also be used for other purposes. In another embodiment, tracking color behavior can be used to monitor progress toward the completion of a dP (pressure drop) measurement task, as described in Example 2.

[0066] Referring to Figure 4, the system 100 can be used as a real-time display interface capable of communicating the status of each tube 118 at a specific time via the display 140. Visualized on the display 140 are differential pressure measurements performed on each tube 118 of the tube sheet 114c of the heat exchanger 300 (example). The display interface may include a representation of the tube sheet 114 using symbols or colors. The display interface may optionally include key performance metrics such as pressure measurements and the percentage of tubes inspected, which are calculated using data records from the relational database 134. The display interface may also include access to relevant information from the relational database 134, such as the device name and a description of the task being performed.

[0067] Furthermore, it may be beneficial to provide workers inside the reactor with one or more portable display devices so that they can monitor the condition of the tubes inside the reactor while performing their work. For example, a worker performing fish taping from below the lower tube sheet can benefit from being able to monitor the behavior of the tube ends in the upper tube sheet in real time. When using such a display device, it is preferable to configure it as a wireless (Wi-Fi) display device. It is also preferable for the display device to have a touchscreen function for ease of use in the field.

[0068] Additional timestamped reactor data can be stored in relational database 134 along with tube data. Examples include temperature, humidity, and O2 concentration inside the reactor head.

[0069] Referring now to Figure 5, once an operator has performed a specific task (e.g., catalyst loading) on ​​a group of tubes 118, the operator can place a block plate 502 across that group of tubes. The block plates 502 fit together in a grid pattern, like puzzle pieces. Such plates are described in U.S. Patent Application Publication 2016 / 220974 (hereinafter, US974), which is incorporated herein by reference in its entirety. US974 teaches the use of colored block plates, where the color of each plate indicates the collective state of all the tubes located beneath it.

[0070] Figure 5 shows a series of 22 slits 502 (labeled P1-P22) applied across different groups of tubes 118 on the tube plate 114c of the heat exchanger 300. Each slit 502 has a removable colored marker 504 applied to it. During use, a particular group of tubes 118 is maintained by the operator, and the slits 502 are manually applied by the operator across the group of tubes 118 to indicate that a particular task (e.g., loading catalyst into the tubes beneath the slit 502) has been completed. The color (or absence) of the marker 504 on the slit 502 indicates the state or condition of those tubes. For example, the absence of a marker may indicate that the tubes beneath the plate have not yet been loaded with catalyst material, while a black marker may indicate that the tubes beneath the plate have been filled with catalyst material. The operator selects the appropriate removable marker 504 to place on the slit 502. The imaging device 120 continuously monitors the color of the markers 504, as well as the color of the caps described above, to determine the progress of a specific task, such as catalyst loading. Instead of tracking the color differences between the markers 504, the shape of the markers 504 may be changed. The imaging device 120 can track the boundaries of each plate 502, so it can know which tube 118 is located under each plate 502.

[0071] (F) Energy transfer While we have discussed the detection of visible light energy so far, the general concepts described above apply to all forms of energy transfer (light, heat, pressure, sound, X-rays, radio waves, electron beams, etc.) and their appropriate purpose-specific detectors.

[0072] If the energy is light reflected from the surface of an object (e.g., wavelengths of light selected from one or more of the visible light spectrum, infrared spectrum, or ultraviolet (UV) spectrum), a photodetector array (e.g., a silicon-based CMOS photodetector containing an array of individual sensors known as pixels) can be used to measure the intensity of the light at one or more of the above wavelengths and create a monochrome (grayscale) digital image or an "RGB" color digital image. Using appropriate data visualization software (e.g., software known as a display driver), the color data can be optionally represented as a visual image on a display device.

[0073] The light source for reflected light may be from the environment known as passive illumination (e.g., sunlight), or it may be from an artificial white light source known as active illumination (e.g., a lamp). The light source may emit wavelengths of light within one or more of the visible light spectrum, the infrared (IR) spectrum, or the ultraviolet (UV) spectrum.

[0074] If the energy is thermal energy radiated from an object (for example, IR radiation with wavelengths of 7.5 μm to 14 μm), a thermal imaging device 120 equipped with a sensor known as a bolometer can be used to create a digital image including temperature values. Using appropriate data visualization software, the temperature data can be arbitrarily represented as a thermographic (visual) image on a display device 140. Note that since infrared energy is emitted / radiated from an object, there is no illumination source of the infrared energy itself.

[0075] When the energy is reflected radio waves (e.g., from a radar system), the resulting digital image includes radio signal recurrence time values ​​representing the distance between a point on an object and a radio wave detector (receiver). When used with the method of the present invention, radar operating in the EHF band (also known as millimeter-wave radar) is preferred. Image acquisition systems based on radar, sonar, lidar, etc., are known as non-contact ranging devices (NRDs) and generally "paint" the surface of an object with a moving energy beam to collect a large number of densely packed recurrence time (distance) measurements. Using (composite) data visualization software, this distance data can optionally be represented as a visual image on a display device 140 (e.g., a weather radar display or a lidar topographic map). By its nature, an NRD requires active "illumination" with energy that can subsequently be reflected.

[0076] (G) Details of the imaging device Further details of the imaging device 120 will be described later. The imaging device 120 may include detectors such as photodetectors or thermal detectors. A photodetector may further include a plurality of light sensors known as image elements or "pixels". Similarly, a thermal detector may include a plurality of thermal sensors known as microbolometers or simply bolometers.

[0077] The most common and preferred embodiment incorporates optical imaging. In the optical imaging embodiment, an imaging device is used that comprises a photodetector and an image processing software package for imaging in visible light. The imaging device 120 may be, for example, a digital camera, an RGB color video camera, or a monochrome camera. The optical system, i.e., the lens, focuses light onto a photodetector (this is a so-called focal plane array or FPA) positioned in the focal plane of the camera to obtain a low-distortion image (i.e., a focused image). Individual sensors (i.e., pixels) within the photodetector convert the light that contacts the photodetector into a digital signal. The digital signal is then sent to an image processor, where the digital signal data is combined to represent a digital image as a mathematical array.

[0078] When a digital image of the tube sheet 114 is acquired, it may contain thousands, or even millions, of digital values, depending on the detector array used. For example, a typical "4K" color digital camera has a CMOS photodetector array with 3840 pixels horizontally and 2160 pixels vertically, resulting in 8,294,400 distinct color measurements. This is commonly referred to in the art as an "8-megapixel array" or simply an "8MP" detector.

[0079] As is well known in the field of digital imaging, the size of the optical system and detector controls how much of the physical world an imaging device can "see," a term known as the field of view (FOV). Detectors are generally configured as a fixed array (grid) of individual sensing elements, and a larger number of sensing elements supports a wider field of view and / or higher resolution. Most commercial photodetectors are implemented as flat arrays built on silicon wafers, meaning that the total number of possible sensing elements is limited by the maximum physical size of the silicon wafer available. Once the maximum array size is reached, only the choice of lenses can affect the resolution and field of view (FOV) width of the imaging device.

[0080] Traditionally, camera lenses are typically expressed in terms of horizontal and vertical FOV angles, while photodetectors are typically expressed in terms of the number of pixels in the horizontal and vertical directions of the detector array. Since the number of pixels in a given photodetector is constant, FOV and image resolution are inversely proportional. That is, the wider the FOV (the larger the image area seen by the detector), the lower the resolution, while the narrower the FOV (the larger the number of pixels per unit of image area), the higher the resolution. Selecting an appropriate detector size (i.e., the total number of pixels) and an appropriate lens FOV is within the capabilities of a person skilled in the art of digital imaging.

[0081] For example, if it is necessary to identify (i.e., resolve) a 6mm spherical catalyst pellet in an image, a person skilled in the art can choose to represent each 2mm x 2mm region with one pixel, so that an image of a single 6mm sphere can be fully represented by a single 3x3 pixel array (a total of 9 pixels). This is known as 500 pixels per meter (PPM) resolution. If the detector array used for image acquisition measures 2560 pixels horizontally x 1440 pixels vertically, the maximum FOV at 500 PPM resolution would be 5120mm x 2880mm (16.8 feet x 9.6 feet). An auxiliary optical system with an appropriate FOV angle would then be selected to provide a sharp image of this region to the focal plane array.

[0082] In some embodiments, multiple detectors can be physically "butted together" to create a multi-element imaging device with an increased number of detection elements to widen the field of view. Such techniques are known, for example, in the field of astronomy for creating wide-field digital telescopes. Unfortunately, such mechanically joined detectors are currently very expensive and difficult to assemble. Therefore, as is known in the art, it is preferable to acquire views of multiple different regions on the surface of a tube sheet and then use image processing software to combine this collection of views into a larger, combined "mosaic" digital image. With image processing software, there is theoretically no limit to the size of a given mosaic digital image array. The acquisition of multiple images can be performed, for example, with multiple cameras, each having an independent FOV, or with a single camera that changes position, such as a pan-tilt-zoom (PTZ) camera.

[0083] In relation to the present invention, a single imaging device, such as the 8-megapixel (8MP) video camera, can be used. The imaging device 120 can be positioned directly above the center of the tube sheet and is inserted, for example, from a top nozzle located in the head of a shell-and-tube apparatus.

[0084] When combining a large tube sheet with a relatively short head, it can be difficult to capture the entire tube sheet surface within the field of view (FOV) of a single camera. This is further complicated by the fact that most photodetector arrays typically utilize a 3:4 or 16:9 aspect ratio, resulting in a difference between the horizontal and vertical FOVs, which further restricts the imageable area.

[0085] Therefore, multiple imaging devices 120a and 120b can also be used in the present invention. The imaging devices can be mounted, for example, on adjustable supports against the inner wall of the container. Furthermore, they may be positioned on the opposite side of the upper tube sheet, facing inward, and approximately 6 feet above the plane of the tube sheet.

[0086] Alternatively, multiple imaging devices may be mounted outward from the center of the vessel. For example, four outward-facing imaging devices can be suspended above the center of the upper tube sheet, spaced 90 degrees apart, and positioned at a downward angle of approximately 15 to 75 degrees relative to the plane of the tube sheet. Such a configuration can be used, for example, in a shell-and-tube reactor with an annular tube sheet layout, in which case the central circular area of ​​the tube sheet does not contain tubes. For reference, an example of an annular tube sheet layout is shown in Figure 1c of U.S. Patent No. 9,440,903.

[0087] The accuracy of the tube identification process and the evaluation of the condition of each tube can also be improved by using multiple views of the same tube sheet area. For example, two or more cameras can be used, each collecting image data of the same tube sheet area from different field of view angles. Then, image processing software can be used to "integrate" the image data from these multiple views, replacing data from obscured views with data from other clear views. In this way, obstacles such as a person standing in front of the camera can be dealt with, and a complete digital image can be obtained. Ultimately, the condition of all tubes can be tracked at all times, as long as at least one camera can detect each tube.

[0088] In a preferred embodiment, image data is collected using at least one RGB-D camera. The RGB-D camera is a hybrid imaging device that includes both an RGB photodetector and a (LiDAR) laser detector, the two detectors of which are internally synchronized to collect image data simultaneously. Processing the synchronized image data generates a so-called "color 3D" image that includes both RGB color data and depth data. The Intel® Realsense™ L515 LiDAR camera (available from Intel Corporation in Santa Clara, California, USA), which includes both a 2MP RGB photodetector and a laser detector operating at 860nm, is an example of a commercially available RGB-D camera suitable for use in the surveillance method of the present invention. If part of the image is obstructed, for example, if the person described above is walking in the camera's field of view, one photodetector can evaluate this as a change in the state of one or more tubes, such as a change in the color of the tube end. However, by collecting image data using an RGB-D camera, both the tube end color and depth data can be evaluated simultaneously. A change in the depth data of the image indicates the presence of one or more obstacles between the camera and the tube sheet. Once detected, measures can be taken to correct the presence of the obstacle, such as using an alternative view (provided by an additional RGB-D camera, for example), logging an error code in a relational database, pausing image processing, or sounding an "obstacle" alarm.

[0089] When using multiple cameras, a common reference point for camera alignment can be established using any tube sheet benchmark. These can be, for example, temporary magnetic markers or permanent marks.

[0090] While still digital cameras can be used to acquire optical images, video cameras are often easier to set up for use with networked computers. Commercially available video cameras typically have built-in functions to transfer image data to an image processor (e.g., a laptop computer) via wireless LAN, LAN / PoE (Power over Ethernet) cabling, or optical fiber. In some embodiments, at least part of the image processing can be performed within the camera's circuitry to speed up processing and reduce the amount of data transferred (and thus reduce the required bandwidth).

[0091] Most video cameras provide a continuous stream of 30 or more frames per second. In the method of the present invention, such high-speed image acquisition is typically far more than necessary. Generally, to monitor most maintenance work, it is sufficient to acquire individual images at a slower rate, such as one image every 30 seconds, one image every 5 minutes, or even one image every hour. Alternatively, a continuous stream of digital images may be acquired (e.g., 30 frames per second), but the image processing software may be configured so that only a portion of this digital image stream is actually processed. For example, in one embodiment, image processing may be performed using only one image (frame) every 15 minutes.

[0092] (H) Software details The software code for performing the image processing steps described herein can be written using various computer programming languages, for example, C++, Python, or MATLAB. The image processing steps employed include filtering, converting pixels between color and grayscale, edge detection (Cany algorithm), circular Hough transform, and converting image data from one color model to another (e.g., RGB to L). * a * b *This may include one or more techniques widely known in the field of digital image processing, such as conversion to 3D, creation of image masks, and color detection. A library of standardized functions has been created to efficiently perform these image processing steps, and is now available for integration into programming code, greatly simplifying the preparation of software routines. OpenCV (Open Source Computer Vision Library: http: / / opencv.org) is one such library of image processing functions, and is now available as open-source software. Originally written in the C++ programming language, so-called "wrappers" are now available that allow OpenCV functions to be used in other programming languages ​​such as Python, Java®, and MATLAB. Image processing described herein can also be performed using proprietary applications such as IMAGE PROCESSING TOOLBOX® and COMPUTER VISION TOOLBOX® (commercially available from The MathWorks, Inc. in Natick, Massachusetts, USA). Alternatively, OpenCV adapted for use with the Python language (also known as OpenCV-Python) can be used for image processing. Furthermore, Python extensions known as "Numerical Python extensions" or "NumPy" can be used to improve the performance of mathematical operations using array data.

[0093] Image processing software such as MATLAB and OpenCV can perform processing using many different color models. As is known in the art, a “color model” is an abstract mathematical representation of color using an ordered list of parameters, which are referred to herein as “channels.” Images can be RGB, HSV, and L * a * b *It can be represented in many different formats corresponding to well-known color models, including. Colors represented in the RGB color model are specified by the intensities of three channels: R (red), G (green), and B (blue), each using values in the range from 0 to 255. RGB is the native format of devices such as video cameras and televisions. Colors represented in the HSV color model are specified by the following three channels: hue representing the dominant wavelength; saturation representing the shade of the color; and value representing the intensity. L * a * b * Colors represented in the color model are specified by the following three channels: L representing the perceived brightness or lightness * ; a representing the color on the axis between red and green * ; b representing the color on the axis between yellow and blue * .

[0094] In contrast to full-color images, grayscale images contain only one channel representing shades of gray. Pixel intensities in this color space are represented by values in the range of 0 to 255, with black having the weakest intensity (value 0) and white having the strongest intensity (value 255). Therefore, the maximum number of states that can be represented by one pixel in grayscale is 256. With only a single channel, image processing in grayscale is much faster and requires fewer computational resources than full color.

[0095] Image processing software further includes color conversion algorithms and can convert an image obtained in one color model (e.g., an RGB image from a video camera) to a different color model. Such conversions are typically performed to simplify processing calculations or to emphasize specific features within a region of interest (ROI). Additionally, the conversion algorithm can convert a color digital image to grayscale. This is often advantageous when searching for high-contrast regions that generally occur along the edges of objects and are an important aspect of object detection algorithms.

Example

[0096] (I) Example Example 1: Tube skimming The vertical shell-and-tube reactor in this embodiment is the second stage of a two-stage tandem reactor system and is used to convert acrolein to acrylic acid. The reactor comprises a tube sheet approximately 7 m (23 feet) in diameter and more than 24,000 catalyst-containing tubes with an inner diameter of 27.2 mm. At the inlet of each of these second-stage tubes is a 35 cm deep upper inert layer containing 5 mm spherical ceramic pellets. This material is bright white when first loaded into the tubes and therefore highly luminous. Over time, carbonaceous deposits (also known as "coke") accumulate in this inert medium layer, changing to brown or black and reducing luminosity. Typically, deposit accumulation is uneven, with some reactor tubes becoming significantly more fouled than others. As the degree of fouling increases, flow through the tubes is restricted, thereby increasing pressure loss through the reactor and degrading reactor performance.

[0097] To address this problem, the reactor is shut down and at least some of the contaminated inert medium in the tubes is replaced with clean inert medium; this maintenance procedure is known as "skimming." This maintenance procedure involves two multi-stage tasks: (1) Remove inert pellets from all heavily soiled tubes using one or more methods such as vacuuming, chipping, polishing, or drilling; (2) Fill any tubes where the upper inert layer is not complete with new, clean inert pellets.

[0098] Before commencing maintenance work, an initial step was taken to assign a unique identifier to each end of the circular tube.

[0099] In some embodiments, this initial step may be performed manually by acquiring a visible light reference image and representing it on a laptop computer using "Image Viewer" software (commercially available from The Mathworks Inc., Natick 01760, Massachusetts, USA). A key feature of Image Viewer is its ability to display the positional values ​​of individual pixels selected by the user, along with their associated color / intensity values. This allows for the manual identification of specific pixels within each tube end and provides a method for associating groups of pixels with appropriate unique tube identifiers. This method is most beneficial when the shell-and-tube apparatus has a relatively small number of tubes.

[0100] However, in this embodiment, this step was performed using software 132 on computer 124. A visible light reference image was acquired and loaded into image processing software in grayscale format. Next, all edges of the region of geometric interest (i.e., circular tube ends) were identified in the image array. This may be done, for example, by using the Canny edge detection algorithm or the Circle-Hough Transform (CHT) algorithm, both of which are well known in the art and available in OpenCV or Matlab software. In a preferred embodiment, the Circle-Hough Transform (CHT) algorithm is applied to position the periphery of the circular tube ends that appear in the image. For example, the OpenCV "HoughCircles" function utilizes the CHT algorithm to detect all circles in the image, providing both the pixel positions that define the circumference of the circles and the pixel positions of the centers of each circle. In this way, a complete list of the center positions of all circles that appear in the image can be obtained. Next, the pixel position coordinates (x, y) of each circle center in the image array are aligned with the dimensional data of the actual tube sheet, and a unique tube identifier is associated with each circle center.

[0101] In a typical shell-and-tube system, only about one-third of the tube sheet area actually constitutes holes (tube ends), while the remaining two-thirds of the tube sheet area constitutes only the plane between the tube ends. Therefore, only about one-third of the data from the imaging device represents measurements within the so-called region of interest (ROI) on the tube sheet. By knowing the position of all tube ends in the image, subsequent processing can be limited to only these circular ROIs, significantly reducing the time required to evaluate each digital image. Those skilled in the field of image processing technology will understand that image processing software can be used to create image "masks" which can then be usefully applied to achieve such optimized image processing.

[0102] In this embodiment, the brightness of the tube end was selected as the attribute to be evaluated during this skimming operation. This selected attribute is defined as having three states (shown in Table 1 below).

[0103] To record the state of the tube sheet at the start of maintenance, an initial digital image of the tube sheet was acquired at time Ti. While a camera with a monochrome or grayscale photodetector would suffice for this task, in this embodiment, a digital camera equipped with an RGB (visible light spectrum) photodetector was used to collect color measurement data from the tube sheet.

[0104] Subsequently, the resulting initial digital image (Di) of the tube sheet was transferred from the camera to an image processor. Using OpenCV image processing software, the RGB pixel data of the initial digital image was converted to grayscale, removing the hue-related channels. This effectively reduced the color measurements to an array of luminance values ​​ranging from 0 to 255, where 0 represents the lowest luminance value and pure black, 255 represents the highest luminance value and pure white, and the intermediate values ​​from 1 to 254 represent various shades of gray. Next, the attribute state of each tube was determined by evaluating the average value of the grayscale pixels within each circular tube end and assigning one of three state values ​​according to an appropriate range of average luminance values:

[0105] [Table 1]

[0106] Subsequently, attribute status data for each tube end was transferred to a relational database, and a multi-field database record was created for each tube. The database record includes a timestamp representing the time (Ti) when the initial digital image (Di) was acquired, a unique tube identifier, and an assigned attribute status value. A lookup table in the relational database was used to map attribute statuses to specific tube statuses, and these statuses were included in each database record.

[0107] Next, a relational database was used to perform an initial analysis and report on the condition of the tube sheets (e.g., determining the total number of contaminated and clean tubes present at the start of the operation). Furthermore, process duration data could be used to predict the time required to complete the skimming operation. In this embodiment, approximately 6,000 tubes were found to be contaminated and were identified as requiring inert replacement. Based on a historical average cleaning time of 5 minutes per tube (process duration of 5 minutes) and an available team of 10 workers (performing 10 cleaning cycles simultaneously), the estimated time required for this operation was approximately 50 hours.

[0108] Once maintenance work began, the condition of the tube sheet was monitored by acquiring additional digital images of the tube sheet every 10 minutes. Approximately 20 washes were performed at each 10-minute interval. Similar to the initial digital images, the image processor converted subsequent images to grayscale and evaluated the attribute status of each tube end. Subsequently, the condition of the tubes within the tube sheet was recorded and reported periodically using a relational database. In this manner, monitoring of the tube sheet continued until the maintenance work was completed.

[0109] While this specific example demonstrates the application of the method of the present invention to a shell-and-tube chemical reactor, those skilled in the art will readily envision applying similar techniques to other shell-and-tube apparatuses, such as evaluating brightness at the tube ends of a multi-pass horizontal heat exchanger during the removal of mineral scale or polymeric solids.

[0110] Example 2: Measurement of pressure loss In a two-stage SSOI-type shell-and-tube reactor, a new catalyst packing was packed into the tubes of the lower stage. This stage of the reactor consisted of a circular tube sheet with a diameter of 6,430 mm (20.9 feet) and 22,000 seamless carbon steel tubes with a total length of 3,750 mm (12.3 feet). The tubes had an inner diameter of 25.4 mm (1 inch) and were arranged in a 60-degree triangular pattern at a pitch of 38 mm (1.5 inches). Each of these tubes was packed with a two-layer catalyst packing consisting of approximately 1 m (39 inches) of 7 mm x 9 mm cylindrical catalyst pellets and approximately 2.5 m (98 inches) of 5 mm x 7 mm cylindrical catalyst pellets.

[0111] After filling all the tubes in the lower reaction stage with catalyst, it was necessary to evaluate the catalyst packing density in each tube by measuring the differential pressure (dP) passing through each tube. This dP measurement process typically takes more than 24 hours.

[0112] In this embodiment, differential pressure (dP) measurement was performed using multiple pneumatic back pressure measuring devices. Single-tube and multi-tube dP measuring devices of this type are well known in the field of catalyst packing, and various embodiments are described, for example, in U.S. Patent No. 6,694,802 and WO02074428(A2) and DE3935636A1, respectively, which are incorporated herein by reference. The specific device used in this embodiment was a single-tube dP wand consisting of a 0.0625-inch flow orifice and equipped with a 60 psig dry air supply to ensure that an ultrasonic airflow was achieved for accurate measurement. During the dP measurement, the circular tube ends were temporarily color-coded to clearly indicate which tubes did not meet the required pressure drop standards and therefore required corrective action.

[0113] The attribute selected to be evaluated during this differential pressure (dP) measurement process was the color of the tube end. This attribute was defined as having four color states (shown in Table 2 below). Those skilled in the art of catalyst packing will know that there are various methods for temporarily coloring the circular tube ends within a tube sheet.

[0114] For example, in one embodiment, multiple standard #5 size tapered laboratory stoppers can be inserted into the end of a 25.4 mm diameter tube. These stoppers are made of a general-purpose material and can be readily purchased from laboratory equipment suppliers in many different colors, including red, green, black, white, and blue.

[0115] In another embodiment, a 25mm x 25mm (1 inch x 1 inch) hand-cut square of colored adhesive tape, such as Duck Tape® brand colored duct tape (commercially available as rolls in various colors from Shurtape Technologies, LLC. at 44011 Avon, Ohio, USA), can be temporarily placed across the end of the tube.

[0116] In another embodiment, multiple tube marking devices, as disclosed in U.S. Patent No. 8,063,778, can be installed at the end of the tube.

[0117] In this example, multiple CAPLUGS® T-series tapered plugs (commercially available in multiple colors from Protective Industries, Inc., Buffalo, New York, USA) were used to mark the ends of circular tubes. To minimize variations in the hue / intensity of the caps, it is preferable to use such uniformly colored commercially available plastic caps. This simplifies the task of distinguishing specific color states. As taught in U.S. Patent No. 2,580,762A, the shape of these devices allows them to function as caps or plugs. The said U.S. Patent is incorporated herein by reference. In the art, it is common to simply refer to them as “caps,” and this convention is followed herein. In this specific example, four independent colors of T-12X type caps [material code: PE-LD01] with manufacturer color indications: RED002, GRN002, BLU003, and YEL002 were selected to provide the four required color states (see Table 2). The colors of these caps were determined to be easily distinguishable by the imaging device 120 and image software 132.

[0118] Before commencing dP measurements, an initial step was performed in which a unique identifier was assigned to each end of the circular tubes. Furthermore, yellow T12-X CAPLUGS® ("caps") were attached to the tube ends 119 of all unmeasured catalyst tubes in the reactor.

[0119] A pair of AIDA Model UHD-100A RGB digital cameras (AIDA Imaging Corporation, 91797 West Covina, California, USA) Aida Imaging, Inc. (www.Aidaimaging.com) (hereinafter referred to as Camera 1 (e.g., Imaging Device 120a) and Camera 2 (e.g., Imaging Device 120b)) were positioned close to the inner wall of the reactor head, on the opposite side of the tube sheet. Each camera was equipped with an 8MP color photodetector with 4096 pixels horizontally x 2160 pixels vertically, giving an imaging device resolution of 500 pixels / meter (PPM); thus, each pixel in the detector array represented a 2mm x 2mm area of ​​the tube sheet surface. Camera 1 was positioned along the south wall of the reactor head, and its field of view constituted half of the tube sheet surface, corresponding to the northern hemisphere of the tube sheet, while Camera 2 was positioned along the north wall, and its field of view constituted the other half of the tube sheet surface, corresponding to the southern hemisphere of the tube sheet. Existing benchmarks on the tube sheet surface, originally installed during the manufacture of the shell-and-tube reactor, were conveniently used as reference points for the proper positioning of the two cameras.

[0120] Initial digital images of the tube sheet were acquired from each camera at the same time (Ti) to record the state of the tube sheet at the start of maintenance work. The resulting pair of initial digital images were then transferred from the cameras to an image processor, and the images from camera 1 and camera 2 were integrated (e.g., using software tools from the Python data analysis library pandas) to create an integrated initial digital image of the entire tube sheet surface, containing color data from over 16 million pixels. Using OpenCV image processing software, the RGB pixel data of the integrated initial digital image was converted to HSV color format for evaluation. Next, the HSV color values ​​for each tube end were determined by calculating the average color values ​​of a 7x7 (49 pixel) sample window concentrically placed within each circular tube end. Then, one of four color states was assigned to each tube end according to an appropriate range of HSV color mean values:

[0121] [Table 2]

[0122] In this embodiment, SQL Server 2019 (Microsoft Corp., Redmond, Washington, USA) was used as the relational database software. When using OpenCV-python for image processing, as in this embodiment, the Microsoft "pymssql" driver can be used to facilitate the transfer of color attribute state data for each tube end between the image processing software and the relational database. Similar to the previous embodiment, a multi-field database record was created for each tube, and this record contains a timestamp representing the time (Ti) when the initial digital image was acquired, a unique tube identifier, and an assigned color state value. Additionally, a lookup table in the relational database was used to associate attribute states with specific tube states, and these states were also included in each database record.

[0123] Subsequently, this relational database was used to perform an initial analysis and report on the condition of the tube sheets, for example, by determining the total number of unmeasured tubes present at the start of the work (22,000 in this case). This initial result was valuable because it captured all 22,000 tubes in the acquired images and positively confirmed that yellow caps were actually attached to all tube ends.

[0124] Subsequently, dP measurement was performed. Immediately before measuring the dP of the tube, the yellow cap was removed. The end of the dP stick was inserted into the end of the tube, and a constant flow rate of air was blown into the tube. The back pressure generated by the catalyst inside the tube was displayed on the dP stick's display to compare with the acceptable dP value (± tolerance). Optionally, the exact dP value could also be recorded electronically. A new cap was immediately attached to the tube, and the color of the new cap indicated the dP measurement result. Green indicated an acceptable dP (within the tolerance range of 6.26 psig to 7.34 psig), red indicated an unacceptably high dP (above 7.34 psig), and blue indicated an unacceptably low dP (below 6.26 psig). All tubes on the tube sheet 114 were measured, and these steps were repeated for additional tubes 118 until the corresponding tube ends 119 were marked.

[0125] Once the process was underway, the state of the tube sheet during maintenance was monitored by simultaneously acquiring further digital images of the tube sheet from cameras 1 and 2 at 15-minute intervals. These digital images were also transferred to an image processor to evaluate the color attributes of each tube end 119 and assign a corresponding color state value. Since the ranges of hue values ​​("H" in Table 2) for each state did not overlap, the "S" and "V" channels were unnecessary, and the "H" channel could be used exclusively as the criterion for assigning color state values. Similar to the initial digital images, a timestamped database record for each tube was continuously added to the relational database in the SQL server, allowing for the continuous calculation of performance metrics such as the total number of unmeasured tubes, the percentage of out-of-spec tubes, and the estimated completion time of the work.

[0126] Visual display software, including the Delphi graphical user interface (UI) package (commercially available from Idera, Inc. in Houston, Texas, USA), was used to query an SQL database and generate a continuously updating interactive display of the tube sheet on a touchscreen computer monitor (see Figure 4). In another embodiment, the visual display could use colors that match assigned color states, rather than the (black and white) pattern shown in Figure 4.

[0127] In this manner, continuous monitoring of the tube sheet was carried out until all tubes on the tube sheet were measured and marked. Upon completion of this dP measurement work, 98.9% of the tubes were determined to be within the acceptable dP range, indicating that a uniform catalyst density had been achieved. Subsequently, corrective actions were taken in a separate procedure for the few tubes that fell outside the acceptable dP range (red / blue caps).

[0128] This embodiment illustrates that by monitoring the color behavior of these tube ends 119, it is possible to track (i) the percentage of measurements completed in real time, (ii) the number of remaining measurements to be performed (completion rate), and (iii) the number of tubes that are out of tolerance and require correction. Such real-time monitoring would be extremely difficult to perform if each of the 22,000 tubes in the reactor had to be repeatedly counted manually during operation.

[0129] This specific example illustrates the application of the present invention's method to a shell-and-tube chemical reactor, but many other embodiments exist in which the color of the tube ends can be used as a selective attribute for tracking maintenance work on other shell-and-tube equipment. For example, the present invention's method could be used to track the progress of visual inspections of large horizontal steam condensers in power plants. In such condensers, minerals such as calcium carbonate and magnesium silicate are known to accumulate on the tube sides, significantly hindering heat transfer. Microbiological fouling may also be present, which can slow heat transfer and induce serious lower deposit corrosion. In this embodiment, three color states (green, white, and red) are defined, and the tube ends are temporarily colored by attaching caps of the colors corresponding to the defined states. Digital image acquisition and image processing are carried out in much the same manner as described in the above embodiment, and the tube states mapped to these states are: green = clean tube, white = tube with only scale, and red = tube with biofilm. The inspection results can be used not only to formulate cleaning work plans but also to provide valuable feedback on the performance of mineral scale and bio-growth inhibitor systems currently used in the plant.

[0130] Example 3: Catalyst loading In the shell-and-tube reactor of this embodiment, several different catalyst replacement operations were required, including (but not limited to) removal of spent catalyst, tube cleaning, catalyst refilling, and shutdown checks. Multi-tube catalyst refilling is a specific catalyst replacement operation in this embodiment. The purpose of this operation was to uniformly fill each tube of the reactor with a 4,600 mm (15-foot) layer of spherical granular catalyst with a diameter of 5 mm.

[0131] The shell-and-tube reactor in this embodiment had an upper horizontal tube sheet with a diameter of 5,517 mm (18.1 feet) and contained more than 22,000 seamless carbon steel tubes. The tubes had an inner diameter of 22.3 mm (0.878 inches) and were arranged vertically, with the upper end of each tube attached to the upper tube sheet by circumferential welding. The tubes were arranged on the tube sheet in a 60-degree triangular pattern, and the pitch of the tube sheet was 34 mm (1.34 inches). The top head of this reactor was removable, allowing easy access to the upper horizontal tube sheet for maintenance work. By removing the top head, ambient illumination (active illumination) could be used for image acquisition.

[0132] In this embodiment, multi-tube catalyst loading was carried out using a multi-tube loader (MTL) of the type described in U.S. Patent Application Publication No. 2016 / 0220974 (A1) (incorporated herein by reference). The largest capacity MTL used in this embodiment was capable of simultaneously loading 120 tubes with particulate catalyst. As taught in the US974 application, multiple so-called tube sheet "seal plates" 502, illustrated in Figure 5, were used to orient the MTL onto the tube sheet during the multi-tube catalyst loading operation. While beneficial to the loading process, the use of these tube sheet seal plates 502 generally obscures the circular tube ends. Therefore, the state of the shell-and-tube reactor was determined by monitoring the attributes of the seal plates 502 themselves, rather than the attributes of the individual tube ends 119.

[0133] Before starting the multi-tube catalyst loading process, more than 200 sealing plates (P1, P2, P3, ...) were placed on the shell-and-tube reactor tube sheet to cover all of the ends of the circular tubes, forming a grid pattern of the type schematically shown in Figure 5. Note that Figure 5 is a simplified representation of the tube sheet of Example 3, and includes only 224 tubes (118) and 22 sealing plates (502).

[0134] As schematically shown in Figure 5, multiple shapes and sizes of seal plates may be used, but the number of tube ends 119 covered by each individual plate will differ. Also, as taught in US974, the seal plates may have different colors, and the selected color of each individual seal plate is used as a control step in the filling process. In another embodiment, all seal plates may be a single color, and the control step function can be performed by marking the top surface of the seal plates with, for example, numbers, text, or symbols. Such markings may be permanent features of the seal plates, or they may be temporarily attached to the surface of the plates to minimize the total number of seal plates required (e.g., using magnetic labels, adhesive tape, or dry-erase marking pens).

[0135] In this embodiment, each block plate 502 was fabricated from a white opaque (polymethyl methacrylate) acrylic sheet having a "P95" matte surface finish to minimize glare. Each block plate has a single circular recess 504 on its upper surface suitable for receiving a colored indicator disc with a diameter of 38 mm (1.5 inches). In some embodiments, it may be beneficial to permanently mark the outer periphery of each recess with any high-contrast marking, such as a black circle with a line width of 3 mm (0.1 inches) or more. The recess may further have any concentric through-hole less than 38 mm in diameter (e.g., a 19 mm or 0.75 inch hole, not shown) to facilitate the removal of the mounted indicator disc. As shown in Figure 5, the circular recess 504 is indicated by a black circle in the left corner of each plate. While the specific position is not critical, it is preferable that the circular recess 504 be positioned consistently on each block plate so as to simplify image processing.

[0136] The colored indicator discs are also preferably made from a matte, opaque acrylic sheet and are provided in multiple colors suitable for performing control process functions. Thus, in this embodiment, the selected attribute to be evaluated during this multi-tube catalyst loading operation is the color of the indicator disc placed on each block plate, and this attribute is defined as having four color states (shown in Table 3 below). In this example, white indicator discs were initially placed on all block plates.

[0137] In this example, an OAK-1 digital camera (available from Luxonis Holding Corp (www.store.opencv.ai) in Westminster, Colorado, USA) was selected for image acquisition. The OAK-1 camera features a Sony IMX378 CMOS color photodetector measuring 4056 pixels horizontally x 3040 pixels vertically (12MP), capable of imaging the entire top surface of the reactor tube sheet at a resolution of 500 pixels / meter (PPM). Thus, each pixel in the detector array can represent a 2mm x 2mm area of ​​the tube sheet surface. The OAK-1 camera further features an optical system with an 81-degree horizontal FOV and a 68.8-degree vertical FOV. Those skilled in the art can determine, using simple trigonometry, that the camera needs to be positioned above the tube sheet at a vertical distance of approximately 4030mm (13.2 feet) from the geometric center in order to image the entire tube sheet within the available FOV.

[0138] Before commencing maintenance work, a reference image of the sealing plates in their designated positions on the tube sheet was acquired, and the center point of each colored indicator disc was positioned within the image array. Then, the coordinates of the center point of each colored indicator disc were used to represent the position of each sealing plate within the image array, and a unique identifier (represented as P1, P2, P3, etc. in the diagram) was assigned to each sealing plate at this center point position. Furthermore, an image mask was created to simplify image processing.

[0139] An initial digital image of the block plate was acquired at time Ti to record the reactor state at the start of catalyst loading. Subsequently, the obtained initial digital image of the block plate (Di) was transferred from the camera to the image processor in native RGB format.

[0140] It should be noted that uncontrolled fluctuations in ambient lighting conditions (passive illumination) can negatively affect the quality of the digital image of the tube sheet, potentially making image processing tasks such as edge detection more difficult. Under low-light conditions, the use of auxiliary illumination (active illumination) may be beneficial. Conversely, under high-light conditions where part of the tube sheet is exposed to full sunlight, some pixels in the photodetector may become saturated and lose their ability to properly measure color data. In such cases, changing the camera position to obtain different viewing angles of the tube sheet may resolve the problem.

[0141] In some embodiments, optical filters can be used in combination with camera lenses to improve image quality during the acquisition process. For example, tinted glass photographic filters can be used to enhance color differences, or polarizing filters can be used to reduce glare that may obscure image details.

[0142] In this example, we first use the OpenCV function cv2.BGR2LAB to convert a color digital image from RGB format to L in the image processor. * a * b * Convert to color format. L * a * b * In terms of format, the intensity of the light is L * While it is captured within a channel (luminance value), a * Channel and b * The channel (saturation value) is relatively unaffected by the intensity of the lighting. * Channel and b * By using only the channel, the colors of the indicator discs can be clearly distinguished under a wide range of tube sheet lighting.

[0143] L * a * b * In the final format, the color values ​​of each color indicator disk were determined by calculating the average color value of 9x9 (81 pixel) sample windows arranged concentrically on each color indicator disk.

[0144] Next, the color state of each cover plate is a * and b * The average value of only the color channels was evaluated and determined by assigning one of four color state values ​​according to the range in Table 3 below. Those skilled in image processing techniques will know a * and b * While the official color model values ​​for this range are thought to be in the range of (-128) to (+128), OpenCV uses adjusted values ​​in the range of 0 to 255 instead. For reference, the conversion formula is shown at the bottom of Table 3.

[0145] [Table 3]

[0146] Subsequently, the color data for each blocker was transferred to a relational database, creating a multi-field database record for each blocker. The database recorded a timestamp representing the time (Ti) when the initial digital image (Di) was acquired; a unique identifier for the blocker; and an assigned color state value. Then, using lookup tables in the relational database, the color states were mapped to specific tube states, and these tube states were also included in each database record.

[0147] Since each seal plate can cover a different number of tubes, the relational database further includes a lookup table for seal plate size data. This data can be used to associate the number of tube ends covered by each seal plate with the unique identifier of that seal plate. In this way, the accuracy of the tube count for each color state can be improved. Table 4 shows an example of such a lookup table for the tube sheet shown in Figure 5.

[0148] [Table 4]

[0149] After acquiring and processing the initial image, the following catalyst loading process was carried out: a) Remove one of the cover plates to expose the end of the circular tube underneath the plate. b) Place a multi-tube loader (MTL) across the exposed tube ends. c) Add catalyst to MTL d) Fill the tube with catalyst. e) Remove the MTL from the exposed end of the tube. f) Change the colored indicator disc of the cover plate. g) Replace the plate so that it covers the filled tube.

[0150] Then, after a certain period of time, and possibly while MTL is still performing steps a) to g), do the following: h) Remove one of the plates to expose the tube end 119 of the filled tube located beneath the plate. i) Check / correct the length of the catalyst layer in each tube by adding or removing catalyst particles. j) Replace the indicator disc inside the board. k) Replace the plate so that it covers the confirmed tube.

[0151] As the process progressed, the state of the tube sheets during the catalyst filling operation was monitored by acquiring subsequent digital images after each seal plate was replaced (i.e., each time process g or k was completed). As described above, these digital images were also transferred to an image processor, where the color of each indicator disk was evaluated and a corresponding color state value was assigned. Similar to the initial digital images, a timestamped database record of each seal plate was continuously added to the relational database, enabling the continuous calculation of performance metrics such as the total number of filled tubes and the estimated time to complete the operation.

[0152] Example 4: Installation of catalytic converter retainer In this embodiment, the monitored task was the installation of catalyst retainers at the lower tube ends of a vertical shell-and-tube reactor equipped with 25.4 mm (1 inch) tubing. As is known in the art, catalyst retainers are used to support catalyst loading within each tube, and each retainer must be installed at the same constant vertical distance from the lower tube sheet. Achieving the precise height of the installed retainers is crucial in controlling the length of all catalyst layers subsequently loaded.

[0153] Figure 1E of U.S. Patent No. 9,440,903 (as incorporated herein by reference) illustrates a specific “catalyst clip” used as a catalyst retainer in this embodiment, although other catalyst retainers may be used. As is known in the art, specific tools are used to assist in the proper mounting of these catalyst clips. However, because a typical commercially available shell-and-tube reactor contains thousands of tubes, it is not uncommon for at least some of the catalyst clips to be installed at the wrong height or at an improper angle (i.e., not horizontal). In some cases, a clip may not be attached to a given tube, or the clip may come loose during loading.

[0154] In this embodiment, it was desired that all catalyst clips be positioned at a height of 12.7 mm (0.50 inches) to 19.1 mm (0.75 inches) above the lower tube sheet. Therefore, the attribute selected for evaluation in this embodiment was the clip installation depth within the tube, measured relative to the bottom plane of the lower tube sheet. This attribute was defined as having four numerical states (shown in Table 5 below).

[0155] In this embodiment, the selected imaging device was a non-contact ranging device (NRD) rather than a digital camera. Specifically, the NRD is a LiDAR device equipped with at least one laser operating at wavelengths of 800 nm to 1600 nm. A commercially available Density Entry Sensor (available from Density Inc., San Francisco, California, USA) can be repurposed for other purposes by the method of the present invention and is an example of such a LiDAR device. As development of NRD systems for use in autonomous vehicles progresses, many excellent low-cost LiDAR devices operating at wavelengths of 905 nm and 1550 nm are commercially available.

[0156] In a preferred embodiment, digital images can be acquired using at least one Velarray M1600 solid-state LiDAR device (available from Velodyne Lidar, Inc., San Jose, California, USA). Furthermore, MATLAB software including the "velodynelidar" interface is even more preferably used for image processing and, if necessary, for visualization of the associated point cloud.

[0157] As described in the previous example, a unique identifier was first assigned to the tube end, and then a digital image including the tube end within the lower tube sheet was acquired.

[0158] In this embodiment, the measurements from the imaging device were return time values, which were converted into desired relative depth measurements by image processing software. As will be apparent to those skilled in the art, measuring the distance between the LiDAR device and the bottom surface of the lower tube sheet is also necessary in order to properly calculate the relative installation depth.

[0159] After calculation, the relative installation depth measurements were evaluated according to the range in Table 5, and the attribute state values ​​were determined. Next, the appropriate tube state was associated with each attribute state using a relational database:

[0160] [Table 5]

[0161] After initiating the catalyst clip installation process, the condition of the tube sheet was continuously monitored by acquiring and processing additional digital images of the tube sheet every 5 minutes until the catalyst clip installation process was completed. In some embodiments, visual display software can be used to display the collected LiDAR return time data measurements as a so-called "point cloud" image on a video display screen, but this is not an essential requirement for carrying out the method of the present invention.

[0162] Similar to the initial digital images, timestamped database records for each lower tube end were continuously added to the relational database, allowing for the calculation of performance metrics such as the total number of catalyst clips installed and the estimated time to completion. Furthermore, real-time monitoring of the installation process enabled prompt corrective action to be taken whenever improper installation techniques were detected, avoiding hours of undesirable rework.

[0163] Although the present invention has been described with respect to at least one embodiment, the present invention can be further modified within the spirit and scope of the invention. Accordingly, this application is intended to cover any variation, use, or adaptation of the present invention using its general principles. Furthermore, this application is intended to cover departures from the present invention that are within the scope of prior art or common practice and fall within the limitations of the claims.

Claims

1. A method for monitoring a tube sheet having a plurality of tube ends arranged in a fixed pattern of rows (R) and columns (C), comprising the following steps: (a) Assign a unique identifier to each of the plurality of tube ends, (b) A digital image (D) of at least a portion of the tube sheet is acquired at acquisition time (T), (c) Determine the state of the attribute for each of the tube ends in the digital image, wherein the attribute has at least two possible states; and (d) Record the data of each tube end in the digital image in the relevant database. The data includes, i. The acquisition time (T), ii. A unique identifier for the tube end, and iii. Attribute status at acquisition time (T) The attribute includes at least one of the following: a) Visual appearance of the marker installed inside or on the end of the tube, b) Brightness of the granular material inside the tube end, c) Texture of the granular material inside the end of the tube, d) The frequency of the infrared light emitted from the end of the tube, e) The frequency of ultraviolet light emitted from the end of the tube, and f) Frequency of ultraviolet light reflected from the end of the tube The method, including the method described above.

2. The method according to claim 1, wherein the unique identifier assigned to each of the plurality of tube ends is a set of Cartesian coordinates in (row, column) format.

3. The method according to claim 1, wherein steps (b) to (d) are performed multiple times.

4. The method according to claim 1, further comprising irradiating the tube sheet with at least one light source that emits wavelengths of light within the visible light spectrum, the infrared spectrum, or the ultraviolet spectrum.

5. The method according to claim 1, further comprising the step of positioning the at least one digital camera such that at least a portion of the tube sheet is within the field of view of the at least one digital camera, and performing the acquisition step using the at least one digital camera.

6. The method according to claim 5, wherein the at least one digital camera detects a wavelength of light selected from one or more of the visible light spectrum, the infrared spectrum, or the ultraviolet (UV) spectrum.

7. To acquire the digital image (D), at least one non-contact distance measuring device (NRD) is used, and the method further includes positioning the at least one NRD such that at least a portion of the tube sheet falls within the measurement range of the at least one NRD, and the acquisition step is i. Collect multiple distance measurements using at least one NRD, and ii. Representing at least a portion of the plurality of distance measurements as the digital image (D). The method according to claim 1, including the method described in claim 1.

8. The method according to claim 7, wherein the at least one NRD is selected from the group consisting of a radar device, a sonar device, a laser scanning (LiDAR) device, and an electron beam device.

9. The method according to claim 1, wherein the tube sheet constitutes part of a heat exchanger, and the heat exchanger is one of a condenser, reboiler, preheater, boiler, superheater, quench exchanger, transfer line exchanger (TLE), evaporator, waste heat boiler, recuperator, cross exchanger, and process heater.

10. The method according to claim 1, wherein the tube sheet constitutes part of a reaction system for producing either hydrogen cyanide or nitrogen oxides.

11. The method according to claim 1, wherein the tube sheet constitutes a part of the reactor.

12. The method according to claim 1, further comprising the step of creating one or more of a table, graph, spreadsheet, and color-coded summary graphic using at least a portion of the data recorded in the relational database.

13. A method for monitoring the condition of a shell-and-tube apparatus, which includes a tube sheet having multiple tube ends arranged in a fixed pattern of rows and columns, during maintenance work, comprising the following steps: (a) Assign a unique identifier to each of the tube ends; (b) Select an attribute that has at least two possible states; (c) An initial digital image (Di) of at least a portion of the tube sheet is acquired at an initial acquisition time (Ti); (d) Determine the initial state of the attributes for each of the tube ends in the initial digital image (Di); (e) Creating an initial data record in the relational database for each tube end in the initial digital image (Di). The initial data record includes, i. The initial acquisition time (Ti), ii. A unique identifier for the tube end, and iii. Initial state of attributes at the initial acquisition time (Ti) The attribute includes at least one of the following: a) Visual appearance of the marker installed inside or on the end of the tube, b) Brightness of the granular material inside the tube end, c) Texture of the granular material inside the end of the tube, d) The frequency of the infrared light emitted from the end of the tube, e) The frequency of ultraviolet light emitted from the end of the tube, and f) Frequency of ultraviolet light reflected from the end of the tube The method, including the method described above.

14. The method of claim 13, further comprising the following steps: (f) A digital image (Dx) of at least a portion of the tube sheet is acquired at a later acquisition time (Tx) (where Tx > Ti), (g) Determine the subsequent state of the attributes of each tube end in the subsequent digital image (Dx), (h) For each tube end in the aforementioned digital image (Dx), a subsequent data record is created in the relational database, where the aforementioned subsequent data record is: i. The acquisition time (Tx) after the above, ii. Unique identifier of the tube end, iii. The state of the attribute after the acquisition time (Tx) mentioned above. including, and (i) Repeat steps (f) to (h) until the maintenance work is completed.

15. The aforementioned shell-and-tube apparatus is used to carry out a chemical transformation, and the chemical transformation is performed i. Conversion of propylene to acrolein and / or acrylic acid; ii. Conversion of propane to acrolein and / or acrylic acid; iii. Conversion of glycerin to acrolein and / or acrylic acid; iv. Conversion of t-butanol, isobutene, isobutane, isobutyraldehyde, isobutyric acid, or methyl t-butyl ether to methacrolein and / or methacrylic acid; v. Conversion of acrolein to acrylic acid; vi. Conversion of methacrolein to methacrylic acid; vii. Conversion of o-xylene or naphthalene to phthalic anhydride; viiii. Conversion of butadiene to maleic anhydride; ix. Conversion of n-butane to maleic anhydride; x. Conversion from indane to anthraquinone; xi. Conversion of ethylene to ethylene oxide; and xi. Conversion of propylene to propylene oxide The method according to claim 13, selected from the group comprising:

16. The method according to claim 13, wherein the shell-and-tube apparatus is used to carry out the oxychlorination of ethylene to 1,2-dichloroethane (EDC).

17. The method according to claim 13, further comprising the step of creating one or more of a table, graph, spreadsheet, and color-coded summary graphic using one or more data records stored in the relational database.

18. The method according to claim 13, further comprising the step of creating performance metrics for maintenance work, wherein the step of creating performance metrics includes calculating the performance metrics and displaying them in a table, graph, spreadsheet or color-coded summary graphic.

19. An optical method for monitoring the condition of a shell-and-tube apparatus, which includes a tube sheet having multiple tube ends arranged in a fixed row and column pattern, during maintenance work, comprising the following steps: (a) Assign a unique identifier to each of the tube ends, (b) Position the at least one digital camera such that at least a portion of the tube sheet is within the field of view of the at least one digital camera, (c) A plurality of colored tube caps are arranged on the end of the tube, and the plurality of tube caps include a tube cap having a first color and a tube cap having a second color different from the first color. (d) An initial digital image (Di) of at least a portion of the tube sheet is acquired at an initial acquisition time (Ti), (e) Determine the initial color of each of the tube ends in the initial digital image (Di), and (f) Create an initial data record in the relational database for each tube end in the initial digital image (Di). The initial data record includes, i. The initial acquisition time (Ti), ii. A unique identifier for the tube end, and iii. Initial color of the tube end at the initial acquisition time (Ti) The method, including the method described above.

20. The method according to claim 19, further comprising the following steps: (g) Remove a colored tube cap from one or more of the multiple tube ends, and / or attach a colored tube cap to one or more of the multiple tube ends. (h) A digital image (Dx) of at least a portion of the tube sheet is acquired at a later acquisition time (Tx) (where Tx > Ti), (i) Determine the color of each of the tube ends in the subsequent digital image (Dx), (j) For each of the tube ends in the subsequent digital image (Dx), a subsequent data record is created in the relational database, where the subsequent data record is: i. The acquisition time (Tx) after the above, ii. A unique identifier for the tube end, and iii. The color of the tube end at the subsequent acquisition time (Tx) including, and (k) Repeat steps (g) to (j) until the maintenance work is completed.

21. The method according to claim 19, further comprising the step of creating one or more of a table, graph, spreadsheet, and color-coded summary graphic using one or more data records stored in the relational database.

22. The method according to claim 21, further comprising transmitting one or more of the tables, graphs, spreadsheets and color-coded summary graphics to at least one display.

23. The method according to claim 19, further comprising predicting the time until the completion of the maintenance work.

24. A method for monitoring the state of a shell-and-tube apparatus including a tube sheet having multiple tube ends arranged in a fixed pattern of rows (R) and columns (C) during granular catalyst loading, comprising the following steps: (a) A plurality of sealing plates are arranged on the tube sheet such that all of the plurality of tube ends are covered, and each of the plurality of sealing plates is provided with a disc recess for installing a color indicator disc. (b) Assign a unique identifier to each of the plurality of blocking plates, (c) A plurality of color indicator discs are placed in the disc recess, and the plurality of color indicator discs each comprises at least one disc having a first color and at least one disc having a second color different from the first color. (d) Position the at least one digital camera such that at least a portion of the plurality of blocking plates is within the field of view of the at least one digital camera, (e) Initial digital images (Di) of at least some of the plurality of blocking plates are acquired at an initial acquisition time (Ti), (f) Determine the initial color of each of the plurality of colored indicator discs in the initial digital image (Di), and (g) Creating an initial data record in the relational database for each of the multiple blocking plates in the initial digital image (Di). The data record includes, i. The initial acquisition time (Ti), ii. The unique identifier of the aforementioned sealing plate, and iii. Initial color of the colored indicator disc at the initial acquisition time (Ti) The method, including the method described above.

25. The method according to claim 24, wherein the unique identifier assigned to each of the plurality of tube ends is a set of Cartesian coordinates in (row, column) format.

26. The method according to claim 24, wherein steps (e) to (g) are performed multiple times.

27. ​​The method according to claim 24, further comprising irradiating the tube sheet with at least one light source that emits wavelengths of light in the visible light spectrum, the infrared spectrum, or the ultraviolet spectrum.

28. The method according to claim 24, wherein the at least one digital camera detects a wavelength of light selected from one or more of the visible light spectrum, the infrared spectrum, or the ultraviolet (UV) spectrum.

29. The method according to claim 24, wherein the tube sheet constitutes part of a heat exchanger, and the heat exchanger is one of a condenser, reboiler, preheater, boiler, superheater, quench exchanger, transfer line exchanger (TLE), evaporator, waste heat boiler, recuperator, cross exchanger, and process heater.

30. The method according to claim 24, wherein the tube sheet constitutes part of a reaction system for producing hydrogen cyanide or nitrogen oxides.

31. The method according to claim 24, wherein the tube sheet constitutes a part of the reactor.

32. The method according to claim 24, further comprising the step of creating one or more of a table, graph, spreadsheet and color-coded summary graphic using at least a portion of the data recorded in the relational database.

33. The shell-and-tube apparatus is used to carry out a chemical transformation, and the chemical transformation is i. Conversion of propylene to acrolein and / or acrylic acid; ii. Conversion of propane to acrolein and / or acrylic acid; iii. Conversion of glycerin to acrolein and / or acrylic acid; iv. Conversion of t-butanol, isobutene, isobutane, isobutyraldehyde, isobutyric acid, or methyl t-butyl ether to methacrolein and / or methacrylic acid; v. Conversion of acrolein to acrylic acid; vi. Conversion of methacrolein to methacrylic acid; vii. Conversion of o-xylene or naphthalene to phthalic anhydride; viiii. Conversion of butadiene to maleic anhydride; ix. Conversion of n-butane to maleic anhydride; x. Conversion from indane to anthraquinone; xi. Conversion of ethylene to ethylene oxide; and xi. Conversion of propylene to propylene oxide The method according to claim 24, selected from the group comprising:

34. The method according to claim 24, wherein the shell-and-tube apparatus is used to carry out the oxychlorination of ethylene to 1,2-dichloroethane (EDC).

35. The method according to claim 24, further comprising the step of creating one or more of a table, graph, spreadsheet and color-coded summary graphic using one or more data records stored in the relational database.

36. The method according to claim 24, further comprising the step of creating performance metrics for maintenance work, wherein the step of creating performance metrics includes calculating the performance metrics and displaying them in a table, graph, spreadsheet or color-coded summary graphic.

37. The method according to claim 24, further comprising predicting the time until the completion of maintenance work.

38. The method of claim 24, further comprising the following steps: (h) Attaching a colored indicator disc to one or more of the disc recesses, and / or removing a colored indicator disc from one or more of the disc recesses, (i) A digital image (Dx) of at least a portion of the plurality of blocking plates is acquired at a later acquisition time (Tx) (where Tx > Ti), (j) Determine the color of each of the colored indicator discs in the subsequent digital image (Dx), (k) For each of the blocking plates in the subsequent digital image (Dx) mentioned above, a subsequent data record is created in the relational database, where the subsequent data record is: i. The acquisition time (Tx) after the above, ii. The unique identifier of the aforementioned sealing plate, and iii. The color of the colored indicator disc after the acquisition time (Tx) mentioned above. including, and (l) Repeat steps (h) to (k) until the granular catalyst filling operation is completed.

39. The method according to claim 24, wherein the shell-and-tube apparatus is a shell-and-tube reactor, a tandem reactor, a single-tube reactor, or a single-shell open-stage interstage reactor (SSOI).

40. The method according to claim 38, further comprising using one or more data records stored in the relational database to create one or more of a table, graph, spreadsheet, and color-coded summary graphic.

41. The method according to claim 38, further comprising predicting the time until the completion of the granular catalyst filling operation.

42. Next step: Measure one or more workspace parameters, The measured values ​​of the workspace parameters are recorded in the relational database, and Optionally, the measured values ​​of the workspace parameters may be displayed on a visual display. The method according to claim 24, further comprising:

Images