State monitoring device and program related to reciprocal motion device

Abstract

To provide a state monitoring device and program related to a reciprocal motion device, which is able to understand an internal state of the reciprocal motion device using digital twin.SOLUTION: A state monitoring device 40 related to a reciprocal motion device 1 generates data for virtual components equivalent to vibration data related to damage and / or wear of real components inside a casing of the reciprocal motion device 1 using digital twin of the reciprocal motion device 1; trains a learning model using correspondence between the generated vibration data of the virtual components and the amount of wear and / or damage of the real parts; acquires vibration data of an actual device from the reciprocal motion device 1 being in operation; estimates a wear and / or damage state of the real components from the vibration data of the actual device acquired using the learning model; and outputs the estimated wear and / or damage state of the real components.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a status monitoring device and a program for a reciprocating device. [Background technology]

[0002] Reciprocating compressors are widely used to compress air and other gases in chemical plants, machinery parts manufacturing factories, food manufacturing factories, and other facilities. There have been cases of unexpected damage to this type of reciprocating compressor during operation, resulting in significant damage, such as shutdowns. The causes of such damage are believed to be loosening of the crosshead and piston rod connections of the reciprocating compressor, wear of sliding parts such as piston rings, rider rings, and piston rod packings, and a decrease in the rigidity of the base. Therefore, it would be desirable to have a system that can monitor damage to the reciprocating device, but reciprocating compressors are covered by a casing during operation and operate at high speeds, making it difficult to measure and understand the repeated loads generated inside the reciprocating compressor and the acceleration fluctuations of each moving part using an actual machine. Furthermore, unlike typical rotating equipment, reciprocating devices do not generate characteristic vibrations, so simply measuring vibrations does not allow one to grasp the internal state.

[0003] Therefore, a condition monitoring system for motion equipment has been disclosed that introduces a mathematical model that corresponds to the dynamic characteristics of the system and changes in various parameters, and by measuring the vibration of the casing or the vibration of the part to be monitored, can determine abnormalities in the motion equipment and the cause of the failure through mathematical analysis, and does not require large-scale equipment (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6444885 Summary of the Invention [Problem to be solved by the invention]

[0005] The technology disclosed in the aforementioned Patent Document 1 can be used to identify abnormalities in exercise equipment and the causes of their failure. However, it cannot determine, for example, the amount of wear on the sliding parts of the exercise equipment or the extent of damage to the parts. In recent years, digital twins, which create models of real objects in a virtual space and enable simulations in that space, have come into widespread use.

[0006] Therefore, an object of the present disclosure is to provide a status monitoring device and program for a reciprocating motion device that can grasp the internal status of the reciprocating motion device using a digital twin. [Means for solving the problem]

[0007] The present disclosure solves the above problems and achieves the object by the following solutions. The first disclosure is a condition monitoring device for a reciprocating motion device, comprising: a data generation unit that generates data for a virtual part using a digital twin of the reciprocating motion device, the data being equivalent to vibration data related to damage and / or wear of a real part inside a casing of the reciprocating motion device; a learning unit that learns a learning model using a correspondence between the vibration data of the virtual part generated by the data generation unit and the amount of wear and / or degree of damage of the real part; a vibration data acquisition unit that acquires vibration data of the real machine from the reciprocating motion device while it is in operation; a condition estimation unit that uses the learning model to estimate the wear state and / or damage state of the real part from the vibration data of the real machine acquired by the vibration data acquisition unit; and a condition output unit that outputs the wear state and / or damage state of the real part estimated by the condition estimation unit. The second disclosure is a condition monitoring device in which, in the condition monitoring device of the first disclosure, the data generation unit generates waveform data relating to vibrations at a monitoring position of the casing that is close to a reciprocating part inside the casing as vibration data of the virtual part, and the vibration data acquisition unit acquires waveform data relating to vibrations at the monitoring position of the reciprocating device in operation as vibration data of the actual machine. A third disclosure is the condition monitoring device of the second disclosure, wherein the waveform data is data obtained by filtering vibrations in a specific frequency range centered on the natural frequency of the reciprocating device and / or the actual part. A fourth disclosure is a condition monitoring device according to the second or third disclosure, wherein the data generation unit generates the waveform data generated by one reciprocating motion of the reciprocating motion unit as vibration data of the virtual part, and the vibration data acquisition unit acquires the waveform data generated by one or more reciprocating motions of the reciprocating motion device during operation as vibration data of the actual machine. A fifth disclosure is a condition monitoring device in which, in any one of the first to fourth disclosures, the condition output unit outputs the wear state and / or damage state of the actual part estimated by the condition estimation unit, together with the matching rate when estimated by the condition estimation unit. A sixth disclosure is a condition monitoring device according to any one of the first to fifth disclosures, comprising a life prediction unit that predicts the life of the actual part based on the wear and / or damage state of the actual part estimated by the condition estimation unit, a usage limit value that is a limit value for the amount of wear and / or degree of damage in the actual part, and the usage time of the actual part. The seventh disclosure is a condition monitoring device according to any one of the first to sixth disclosures, which is equipped with a digital twin creation unit that generates a physical model from design data of the reciprocating motion device in operation, and creates the digital twin by setting parameters to be identified for the generated physical model using vibration data of the actual reciprocating motion device in operation. An eighth disclosure is a program for causing a computer to function as any one of the condition monitoring devices according to the first to seventh disclosures. The ninth disclosure is a condition monitoring system including a condition monitoring device that monitors the condition of a reciprocating motion device, and a vibration sensor provided at a monitoring position of a casing in the reciprocating motion device near a reciprocating part inside the casing, wherein the condition monitoring device includes a data generation unit that generates vibration data for a virtual part using a digital twin of the reciprocating motion device, the vibration data being vibration data related to damage and / or wear of a real part inside the casing, the vibration data being equivalent to the vibration data related to vibrations at the monitoring position, and a data generation unit that generates a digital twin of the reciprocating motion device based on the vibration data of the virtual part generated by the data generation unit and the amount of wear and / or damage of the real part. The condition monitoring system comprises a learning unit that learns a learning model using a correspondence between the degree of damage and the degree of scratch; a vibration data acquisition unit that acquires vibration data of the actual machine from the reciprocating motion device while it is in operation using the vibration sensor; a condition estimation unit that uses the learning model to estimate the wear state and / or damage state of the actual part from the vibration data of the actual machine acquired by the vibration data acquisition unit; and a condition output unit that outputs the wear state and / or damage state of the actual part estimated by the condition estimation unit, wherein the vibration sensor detects vibrations at the monitoring position on the reciprocating motion device while it is in operation and transmits the vibration data of the actual machine to the condition monitoring device. [Effects of the Invention]

[0008] According to the present disclosure, it is possible to provide a status monitoring device and a program for a reciprocating motion device that can grasp the internal status of the reciprocating motion device using a digital twin. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is an overall configuration diagram of a status monitoring system according to an embodiment of the present invention; [Figure 2] 1 is a diagram showing a longitudinal sectional front view of an example of a reciprocating motion device according to an embodiment of the present invention. [Figure 3] 1 is a functional block diagram of a state monitoring device according to an embodiment of the present invention; [Figure 4] 10 is a flowchart showing a digital twin construction process for the state monitoring device according to the present embodiment. [Figure 5] FIG. 1 is a diagram illustrating an example of a digital twin constructed by the state monitoring device according to the present embodiment. [Figure 6] 4 is a flowchart showing a learning process of the status monitoring device according to the present embodiment. [Figure 7] 4 is a flowchart showing a state monitoring process of the state monitoring device according to the present embodiment. [Figure 8] FIG. 10 is a diagram for explaining a verification example using the state monitoring device according to the present embodiment. [Figure 9] FIG. 10 is a diagram showing an example of waveform data in a verification example using the state monitoring device according to the present embodiment. [Figure 10] FIG. 10 is a table showing characteristic parts of a verification example using the state monitoring device according to the present embodiment. [Figure 11] FIG. 10 is a graph showing characteristic parts in a verification example using the state monitoring device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. Note that this is merely an example, and the technical scope of the present disclosure is not limited to this example. (Embodiment) [Condition monitoring system 1000] FIG. 1 is a diagram showing the overall configuration of a condition monitoring system 1000 according to this embodiment. FIG. 2 is a diagram showing a vertical cross-sectional front view of an example of the reciprocating motion device 1 according to this embodiment.

[0011] 1 includes a reciprocating device 1, a condition monitoring device 40, and a vibration sensor 60. The reciprocating device 1 is a device whose condition is to be monitored, and in this example, a reciprocating compressor used to compress hydrogen in a chemical plant will be described. The condition monitoring system 1000 constructs (creates) a digital twin 101 (described later) of the reciprocating motion device 1 in the condition monitoring device 40. The condition monitoring device 40 is an information processing device, such as a personal computer (PC) or a server. The condition monitoring system 1000 then uses the digital twin 101 of the reciprocating motion device 1 constructed by the condition monitoring device 40 to estimate and output the wear and / or damage states of the actual parts of the reciprocating motion device 1.

[0012] [Reciprocating motion device 1] Here, the configuration of the reciprocating motion device 1 to be monitored will be described. As shown in Figure 2, the reciprocating device 1 consists of a crank section 2, an intermediate connection section 3, and a piston-cylinder section 4, and the crankcase 5 of the crank section 2, the distance piece 6 of the intermediate connection section 3, and the cylinder 7 of the piston-cylinder section 4 are integrated to form a casing 8.

[0013] The crank section 2 is provided with a crankshaft 9 which is rotatably supported in approximately the center of the crankcase 5 so as to face in the front-to-rear direction, and which has two crankpins 9a (only one of which is shown) at two eccentric portions which are offset in the front-to-rear position and which are 180° from each other; a pair of left and right connecting rods 11 whose bases are rotatably fitted onto each crankpin 9a via crankpin metals 10, and whose tip ends extend outward in the left-to-right directions from the crankshaft 9; and a crosshead 14 which is rotatably connected to the tip of each connecting rod 11 by a crosshead pin 12 and a crosshead pin metal 13 which face in the front-to-rear direction, and which slides in the left-to-right direction within the crankcase 5.

[0014] The crank section 2 has a nearly symmetrical structure with the crankshaft 9 at the center, and the intermediate connection section 3 and the piston-cylinder section 4 are connected to the outside on the left and right, respectively. However, in the following explanation, for ease of understanding, only the structure to the right of the crankshaft 9 will be explained.

[0015] The base of the connecting rod 11 is attached to the crank pin 9a by joining the two split parts with a bolt 15. A crosshead shoe 16 is provided on the outer periphery of the crosshead 14 to reduce sliding resistance with the crankcase 5.

[0016] In the crank section 2, the crankshaft 9 is rotated by an electric motor (not shown) provided outside the crankcase 5, causing the crosshead 14 to move linearly back and forth in the left and right direction via the connecting rod 11. The part where the crankshaft 9, connecting rod 11, and crosshead 14 move in series is hereinafter also referred to as the reciprocating part 1a (see FIG. 1).

[0017] The crosshead 14 is connected to the left end of a piston rod 17 that passes through the distance piece 6 of the intermediate connection portion 3 and reaches the inside of the cylinder 7 of the piston-cylinder portion 4, facing in the left-right direction.

[0018] An oil wiper ring 18 is provided at the penetration portion of the piston rod 17 at the boundary between the crankcase 5 and the distance piece 6, and an intermediate gland packing 19 is provided at the penetration portion of the piston rod 17 in the middle part of the distance piece 6.

[0019] A gland packing 20 is provided at the penetration portion of the piston rod 17 at the connection portion between the cylinder 7 of the piston-cylinder portion 4 and the distance piece 6. Furthermore, the cylinder 7 is provided with a number of coolant passages 21, and the cylinder 7 is cooled by the coolant passing through these passages.

[0020] A piston 24, which has a piston ring 22 and a rider ring 23 attached to its outer periphery, is connected to the right end of the piston rod 17 that has advanced into the cylinder 7. This piston 24 reciprocates left and right within a cylinder liner 25 provided on the inner surface of the cylinder 7, compressing the gas (hydrogen) drawn in from a gas inlet 26 via an intake valve 27 and discharging it from a gas outlet 29 via an exhaust valve 28. A suction valve open type unloader 30 is connected to the suction valve 27 . The above is a rough structure of the reciprocating motion device 1 whose state is to be monitored.

[0021] In such a reciprocating device 1, vibrations are generated from the reciprocating section 1a described above and the sliding parts of each section, for example, the bearing portion of the crankshaft 9 in the crankcase 5, the rotating part between the crankpin 9a and the crankpin metal 10, the rotating part between the crosshead pin 12 and the crosshead pin metal 13, the sliding part between the crosshead shoe 16 and the crankcase 5, the sliding parts between the oil wiper ring 18, the intermediate gland packing 19, and the gland packing 20 and the piston rod 17, and the sliding parts between the outer surface of the piston 24, the piston ring 22, and the rider ring 23 and the cylinder liner 25.

[0022] In order to check the wear and / or damage state of each actual part due to vibrations generated from each part of the reciprocating motion device 1 as described above, a digital twin 101 of the reciprocating motion device 1 is created in advance in the condition monitoring device 40. The condition monitoring device 40 then generates vibration data of various states of the virtual part using the created digital twin 101, and inputs the correspondence between the amount of wear and / or the degree of damage of the real part and the generated vibration data into a learning model for learning. Here, the virtual part refers to a part of the digital twin 101.

[0023] Furthermore, a vibration sensor 60 is installed in a part of the casing 8 of the reciprocating motion device 1, at a position below the crankcase 5, more specifically at a position (monitoring position) below the crosshead 14. Then, using vibration data of the reciprocating motion device 1 in operation acquired from the vibration sensor 60 and the learning model, the condition monitoring device 40 estimates and outputs the wear and / or damage state of the actual part, including its position.

[0024] [Status monitoring device 40] Next, the function of the state monitor 40 will be described. FIG. 3 is a functional block diagram of the state monitoring device 40 according to this embodiment. As shown in FIG. 3, the status monitoring device 40 includes a control unit 41, a storage unit 50, an input unit 56, a display unit 57, and a communication unit 59. The control unit 41 is a central processing unit (CPU) that controls the entire status monitoring device 40. The control unit 41 appropriately reads and executes the operating system (OS) and application programs stored in the storage unit 50, thereby cooperating with the above-mentioned hardware and executing various functions.

[0025] The control unit 41 includes a digital twin creation unit 42, a simulation data generation unit 43, a learning unit 44, an actual machine data acquisition unit 45, a state estimation unit 46, a state output unit 47, and a life prediction unit 48. The digital twin creation unit 42 generates a physical model from the design data of the reciprocating motion device 1 in operation, and sets identifying parameters for the generated physical model using vibration data from the reciprocating motion device 1 in operation to create a digital twin 101. The digital twin creation unit 42 then stores the created digital twin 101 in the digital twin storage unit 52.

[0026] The simulation data generation unit 43 functions as a data generation unit. Using the digital twin 101 of the reciprocating motion device 1 created by the digital twin creation unit 42 and stored in the digital twin storage unit 52, the simulation data generation unit 43 generates data for virtual parts that is equivalent to vibration data related to damage and / or wear of real parts inside the casing of the reciprocating motion device 1.

[0027] Here, the vibration data generated by the simulation data generating unit 43 is waveform data related to vibration. The waveform data is data obtained by filtering vibration within a specific frequency range centered on the natural frequency of the reciprocating motion device 1 and / or the actual part. The specific frequency range is, for example, a range determined by calculating the first to third bending natural frequencies of the crosshead 14-piston 24 system (see FIG. 1, etc.), and specifically, for example, a range of 100 Hz to 1000 Hz. This range is determined from an impact test of the reciprocating motion device 1 and a simulation using the digital twin 101, and is a range determined from values ​​obtained when the rotation angle conditions of the crankshaft 9 are, for example, 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, etc.

[0028] If vibration data is simply used, it is not possible to read a specific waveform from the waveform data of the vibration, and the characteristic parts are unknown. Therefore, as described above, by filtering the vibration in a specific frequency range centered on the natural frequency of the reciprocating motion device 1 and / or the actual part, it is possible to obtain characteristic waveform data. Filtering vibrations in such a specific frequency range is useful in a reciprocating device 1 in which slight vibrations occur not only due to rotation in the bearings but also when the crankshaft 9 makes one rotation.

[0029] The learning unit 44 learns a learning model using the correspondence between the vibration data of the virtual part in various states (wear, cracks, etc.) generated by the simulation data generation unit 43 and the amount of wear and / or the degree of damage of the real part caused by the vibration data. Here, the data to be learned, "the amount of wear and / or the degree of damage of the real part caused by the vibration data," uses the amount of wear and / or the degree of damage of the virtual part of the digital twin 101. In addition, the learning unit 44 may learn using data of real phenomena. Then, the learning unit 44 stores the learned learning model in the learning model storage unit 53.

[0030] The actual machine data acquisition unit 45 functions as a vibration data acquisition unit. The actual machine data acquisition unit 45 acquires vibration data of the reciprocating motion device 1 in operation as vibration data of the actual machine. More specifically, the actual machine data acquisition unit 45 acquires waveform data relating to vibration as vibration data from a vibration sensor 60 provided on the reciprocating motion device 1 in operation. The waveform data is data obtained by filtering vibrations in a specific frequency range centered on the natural frequency of the reciprocating motion device 1 and / or the actual part.

[0031] The state estimation unit 46 uses the learning model stored in the learning model storage unit 53 to estimate the wear state and / or damage state of the actual part from the vibration data of the actual machine acquired by the actual machine data acquisition unit 45. The state output unit 47 outputs the wear state and / or damage state of the actual part estimated by the state estimation unit 46 to the display unit 57. Life prediction unit 48 predicts the life of the real part based on the wear and / or damage state of the real part estimated by state estimation unit 46, a service limit value which is a limit value for the amount of wear and / or the degree of damage in the real part, and the usage time of the real part, and outputs the predicted life to display unit 57. Here, the service limit value of the part may be predetermined for each part and input at the time of life prediction, or may be stored in storage unit 50, for example. Furthermore, the usage time may be input at the time of life prediction, or may be obtained from reciprocating motion device 1 in cooperation with the real device, reciprocating motion device 1.

[0032] The storage unit 50 is a storage area such as a hard disk or semiconductor memory element for storing programs, data, etc. required for the control unit 41 to execute various processes. The memory unit 50 includes a program memory unit 51, a digital twin memory unit 52, and a learning model memory unit 53.

[0033] The program storage unit 51 is a storage area that stores various programs. The program storage unit 51 includes a digital twin creation program 51a as a control program that executes the digital twin creation unit 42. The program storage unit 51 also includes a state monitoring program 51b as a control program that executes each of the functional units, namely, the simulation data generation unit 43, the learning unit 44, the actual machine data acquisition unit 45, the state estimation unit 46, the state output unit 47, and the life prediction unit 48. The program storage unit 51 may also store a program for each functional unit executed by the control unit 41.

[0034] The digital twin storage unit 52 is a storage area that stores the digital twin 101 of the reciprocating motion device 1 created by the digital twin creation unit 42. The learning model storage unit 53 is a storage area that stores the learning model learned by the learning unit 44.

[0035] The input unit 56 receives input from the user and may include a keyboard, a pointing device, and the like. The display unit 57 displays a screen for accepting data input from the user and a screen showing the results of arithmetic processing by the computer, and includes a display device such as a liquid crystal display (LCD). The input unit 56 and the display unit 57 may be integrated into a touch panel display. The communication unit 59 is a network adapter that enables the condition monitoring device 40 to connect to another processing system or storage device via a private or public network, and may include a modem, a cable modem, and an Ethernet adapter.

[0036] [Vibration sensor 60] 1 is a detection device that detects vibrations of the reciprocating device 1. As described above, the vibration sensor 60 is provided at a monitoring position on the casing 8 near the reciprocating unit 1a inside the casing 8 of the reciprocating device 1. It is not possible to provide the vibration sensor 60 inside the casing 8 of the reciprocating device 1 from the viewpoint of safety, such as explosion-proof restrictions. For this reason, the vibration sensor 60 is installed at a monitoring position on the casing 8 near the reciprocating unit 1a, where stronger vibrations can be detected. A detection signal from the vibration sensor 60 is transmitted to the condition monitor 40 via a communication line.

[0037] 〔others〕 Although not shown, a rotation detection sensor for detecting the rotation of the crankshaft 9 of the reciprocating device 1 is provided on or near the crankshaft 9. The rotation detection sensor may be one that detects rotation using a laser or the like, or may be an eddy current sensor or the like, as long as it detects the rotation of the crankshaft 9.

[0038] [Processing Description] Next, the processing in the state monitor 40 will be described with reference to a flowchart. FIG. 4 is a flowchart showing the digital twin construction process of the state monitoring device 40 according to this embodiment. FIG. 5 is a diagram showing an example of a digital twin 101 constructed by the state monitoring device 40 according to this embodiment. First, when creating the digital twin 101 of the reciprocating motion device 1, in step S (hereinafter, "step S" will be simply referred to as "S") 11 of Figure 4, the digital twin creation unit 42 launches the digital twin creation program 51a for creating the digital twin.

[0039] In S12, the digital twin creation unit 42 performs a read process to read design data of the actual reciprocating motion device 1. The design data is, for example, design drawing data or 3D CAD (3-dimensional Computer Aided Design) data. In S13, the digital twin creation unit 42 creates a physical model of the reciprocating motion device 1 based on the design data. In S14, the digital twin creation unit 42 inputs conditions into the program and sets clearances, etc. Here, the conditions to be input include materials, constraint conditions, friction conditions, etc.

[0040] In S15, the digital twin creation unit 42 performs a simulation using the physical model set in the processing of S14. In S16, the digital twin creation unit 42 compares the data obtained when the reciprocating motion device 1 is in operation with the simulation results. In S17, the digital twin creation unit 42 sets identification parameters based on the comparison results and performs adjustments to make the parameters the same as the actual machine. In S18, the digital twin creation unit 42 sets the adjusted model as the digital twin 101 and stores the constructed digital twin 101 in the digital twin storage unit 52. Thereafter, the digital twin creation unit 42 ends this processing.

[0041] An example of a digital twin 101 constructed by the above process is shown in Figure 5. The digital twin 101 is a virtual device that exists in the virtual space of a computer and has the same functions as the actual reciprocating motion device 1. The digital twin 101 is a virtual device corresponding to one reciprocating motion device 1. Therefore, it is necessary to construct a digital twin 101 for each reciprocating motion device 1. In this case, the digital twin 101 sets clearances and the like for the physical model so that it is in the same state as the target reciprocating motion device 1.

[0042] In the following explanation, when it is necessary to explain each part of the digital twin 101 that corresponds to each part of the reciprocating motion device 1, the symbols for each part of the digital twin 101 will be explained using three-digit symbols with the last two digits being the same as those for the reciprocating motion device 1 and the prefix 1.

[0043] Next, the learning process will be described. FIG. 6 is a flowchart showing the learning process of the state monitor 40 according to this embodiment. 6, the simulation data generation unit 43 generates vibration data related to damage and / or wear of a part using the digital twin 101. Here, the simulation data generation unit 43 generates vibration data for various patterns that may occur as damage and / or wear of a part. In S22, the simulation data generating unit 43 filters the generated vibration data in a specific frequency range centered on the natural frequency of the reciprocating motion device 1 and / or the actual part.

[0044] In S23, the learning unit 44 causes the learning model to learn the association between the waveform data, which is the vibration data after filtering, and the amount of wear and / or the degree of damage of the part. In S24, the learning unit 44 stores the learned learning model in the learning model storage unit 53. As a result, the learning model stored in the learning model storage unit 53 is learned by associating vibration waveform data with the amount of wear and / or the degree of damage of the part. Therefore, when vibration waveform data is input to the learning model, the part and the amount of wear and / or the degree of damage of the part are output according to the matching rate.

[0045] Next, the status monitoring process will be described. FIG. 7 is a flowchart showing the state monitoring process of the state monitor 40 according to this embodiment. This process is performed whenever it is desired to diagnose the state of the actual reciprocating motion device 1. In S31 of FIG. 7, the actual machine data acquisition unit 45 acquires, via the vibration sensor 60, vibration data of the reciprocating motion device 1 while it is in operation. In S32, the actual machine data acquisition unit 45 obtains waveform data by filtering the acquired vibration data in the same specific frequency range as that used when generating the learning model.

[0046] In S33, the state estimation unit 46 inputs the filtered vibration data into the learning model in the learning model storage unit 53. In S34, the condition estimation unit 46 acquires the wear state and / or the degree of damage of the part output by the learning model. The acquired data is the wear state and / or the degree of damage of the part, and may be associated with a matching rate. In S35, the status output unit 47 outputs the wear state and / or the degree of damage to the display unit 57. At that time, the status output unit 47 may also output the match rate.

[0047] In S36, the life prediction unit 48 receives input of the service limit value and usage time of the part. If the life prediction is not to be performed, the process is not performed and the process ends. In S37, the life prediction unit 48 predicts the life of the received part based on the usage limit value and usage time of the part, and the current state of wear and / or the degree of damage. In S38, the life prediction unit 48 outputs the predicted life of the part to the display unit 57. Thereafter, the control unit 41 ends this process.

[0048] [Example of validation data] Next, a verification example will be described in which it was verified that the state estimation using the digital twin 101 described above does not deviate from the state of the actual reciprocating motion device 1. 8 to 11 are diagrams relating to a verification example using the state monitoring device 40 according to this embodiment. This verification examined the differences between the actual machine and the digital twin when the diameter of the crosshead pin 12 changes due to wear. The crosshead pin 12a shown in Fig. 8(A) is a normal crosshead pin with no wear or damage, while the crosshead pin 12b shown in Fig. 8(B) is a crosshead pin with a worn diameter. The pin size 12a-1, which is the diameter of the crosshead pin 12a, is 35.0 mm, and the pin size 12b-1, which is the diameter of the crosshead pin 12b, is 34.8 mm.

[0049] FIG. 8(C) shows a graph 70 that indicates the vibration as acceleration when the crankshaft 9 rotates once and the reciprocating part 1a moves back and forth once. The acceleration changes depending on whether the crosshead pin 12 is normal or abnormal due to wear or other reasons. Graph 71 is an enlarged view of the graph before and after the circled portion 70a in graph 70. The difference 71a between the normal and abnormal cases in graph 71 allows prediction of the wear amount and the lifespan of the part to be planned for replacement. The waveform data shown in graph 70 has been filtered.

[0050] FIG. 9 shows the difference in acceleration between the actual machine and the digital twin depending on whether the crosshead pin 12 is worn or not. FIG. 9(A) shows a graph 81 indicating the change in acceleration when the crankshaft 9 rotates once and the reciprocating part 1a makes one reciprocating motion. FIG. 9(B) shows a graph 82 indicating the change in acceleration when the crankshaft 109 rotates once and the reciprocating part 101a of the digital twin 101 makes one reciprocating motion. In both cases, the graph shows the acceleration of the crosshead pin 12a, which is a normal crosshead pin 12, and the acceleration of the worn crosshead pin 12b. Also, during one rotation of the crankshaft 9, there are acceleration peaks in the first and second halves, and these positions are indicated by rectangles on the graph.

[0051] In addition to what is shown in graphs 81 and 82, Figure 10(A) shows table 85, which shows the rotation angles of the first half peak and second half peak for acceleration among the rotation angles of the crankshaft 109 of the digital twin 101 and the rotation angles of the crankshaft 9 of the actual reciprocating motion device 1 when the pin size of the crosshead pin 12 is different. 11A and 11B show graphs of the difference in rotation angle at the peak for each pin size. Fig. 11A shows graph 91, which shows the difference in rotation angle at the first half of the peak for different pin sizes of the crosshead pin 12, and Fig. 11B shows graph 92, which shows the difference in rotation angle at the second half of the peak for different pin sizes of the crosshead pin 12.

[0052] As shown in Table 85 in Figure 10(A) and the graphs in Figure 11, the difference in rotation angle between the first and second peaks of acceleration between the digital twin 101 and the actual reciprocating motion device 1 is small, indicating that predictions based on rotation angle using the digital twin 101 are highly accurate.

[0053] On the other hand, in addition to what is shown in graphs 81 and 82, Figure 10(B) shows table 86, which shows the acceleration of the first half peak and the second half peak of the acceleration of the digital twin 101 and the actual reciprocating motion device 1 when the pin size of the crosshead pin 12 is different. According to table 86, there is a difference in the amount of acceleration of the first half peak and the second half peak between the digital twin 101 and the actual reciprocating motion device 1, which shows that prediction based on acceleration using the digital twin 101 is not useful.

[0054] As described above, the status monitoring system 1000 of this embodiment has the following advantages. (1) The condition monitoring device 40 includes a simulation data generation unit 43 that generates data for virtual parts using the digital twin 101 of the reciprocating motion device 1, the data being equivalent to vibration data relating to damage and / or wear of real parts inside the casing 8 of the reciprocating motion device 1; a learning unit 44 that learns a learning model using a correspondence between the vibration data of the virtual parts generated by the simulation data generation unit 43 and the amount of wear and / or damage of the real parts; a real machine data acquisition unit 45 that acquires vibration data of the real machine from the reciprocating motion device 1 in operation; a condition estimation unit 46 that uses the learning model to estimate the wear and / or damage state of the real parts from the vibration data of the real machine acquired by the real machine data acquisition unit 45; and a condition output unit 47 that outputs the wear and / or damage state of the real parts estimated by the condition estimation unit 46.

[0055] Therefore, it is possible to estimate the wear and / or damage state of the actual parts from the vibration data of the actual machine using the digital twin 101 of the reciprocating motion device 1. As a result, it is useful because it is possible to estimate the wear and / or damage state of the actual parts in the reciprocating motion device 1, the interior of which cannot be grasped from the outside.

[0056] (2) The simulation data generation unit 43 of the condition monitoring device 40 generates waveform data relating to vibrations at a monitoring position of the casing near the reciprocating part 1a inside the casing 8 as vibration data of a virtual part, and the actual machine data acquisition unit 45 acquires waveform data relating to vibrations at a monitoring position of the reciprocating machine 1 in operation as vibration data of the actual machine. Therefore, the digital twin 101 can estimate the wear and / or damage state of the actual part using the vibration waveform data of the actual part.

[0057] (3) The waveform data is data obtained by filtering vibrations within a specific frequency range centered on the natural frequency of the reciprocating motion device 1 and / or the actual part. Therefore, the wear and / or damage state of the actual part can be estimated using data that is easier to understand.

[0058] (4) The simulation data generation unit 43 of the condition monitoring device 40 generates waveform data generated by one reciprocating motion of the reciprocating motion unit 1a as vibration data of the virtual part, and the actual machine data acquisition unit 45 acquires waveform data generated by one or more reciprocating motions of the reciprocating motion device 1 in operation as vibration data of the actual machine. Therefore, it is only necessary to generate waveform data generated by one reciprocating motion of the reciprocating motion unit 1a, and processing is easy.

[0059] (5) The state output unit 47 of the state monitoring device 40 is configured to output the wear state and / or damage state of the actual part estimated by the state estimation unit 46 together with the coincidence rate when estimated by the state estimation unit 46. Therefore, it is possible to output various possibilities, and by outputting them in descending order of the matching rate, for example, it is possible to grasp the wear and / or damage state of the actual part that is most likely to occur based on the output order.

[0060] (6) The condition monitoring device 40 is provided with a life prediction unit 48 that predicts the life of the actual part based on the wear and / or damage state of the actual part estimated by the condition estimation unit 46, a usage limit value that is a limit value for the amount of wear and / or the degree of damage in the actual part, and the usage time of the actual part. Therefore, the estimated wear and / or damage state of the actual part can be used to predict the life of the actual part.

[0061] (7) The condition monitoring device 40 is provided with a digital twin creation unit 42 that generates a physical model from the design data of the reciprocating motion device 1 in operation, and sets parameters to be identified for the generated physical model using vibration data from the actual reciprocating motion device 1 in operation, thereby creating a digital twin 101. Therefore, the digital twin 101 of the reciprocating motion device 1 created in the condition monitoring device 40 can be used to estimate the wear and / or damage state of the actual part.

[0062] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments. Furthermore, the effects described in the embodiments are merely a list of the most favorable effects resulting from the present disclosure, and the effects of the present disclosure are not limited to those described in the embodiments. The above-described embodiments and the modified embodiments described below can also be used in appropriate combinations, but detailed description thereof will be omitted.

[0063] (Variations) (1) In the present embodiment, a reciprocating compressor has been described as an example of a reciprocating device, but the present invention is not limited to this. Other reciprocating devices may also be used. This is particularly useful for reciprocating devices that are covered by a casing during operation and operate at high speeds, making it difficult to measure and understand the repeated loads generated inside the reciprocating device and the acceleration fluctuations of each moving part using an actual device. (2) In this embodiment, the verification example has been described focusing on the wear state, but it is not limited to this. Damage such as peeling and loosening can also be assumed, and even in such a state, it can be estimated using the method described in this embodiment. [Explanation of symbols]

[0064] 1 Reciprocating device 8 Casing 9. Crankshaft 11 Connecting rod 12, 12a, 12b Crosshead pin 14 Crosshead 17 Piston rod 24 pistons 40 Condition monitoring device 41 Control Unit 42 Digital Twin Creation Department 43 Simulation data generation unit 44 Learning Department 45 Actual machine data acquisition section 46 State estimation unit 47 Status output section 48 Life Prediction Section 50 Storage section 51a Digital Twin Creation Program 51b Condition Monitoring Program 52 Digital Twin Memory Unit 53 Learning model memory unit 56 Input section 57 Display section 60 Vibration Sensor 101 Digital Twin 1000 Condition Monitoring System

Claims

1. a data generation unit that generates data for a virtual part using a digital twin of the reciprocating motion device, the data being equivalent to vibration data relating to damage and / or wear of an actual part disposed inside a casing of the reciprocating motion device; a learning unit that learns a learning model using the correspondence between the vibration data of the virtual part generated by the data generation unit and the amount of wear and / or the degree of damage of the real part; a vibration data acquisition unit that acquires vibration data of the actual machine from the reciprocating motion device during operation; a state estimation unit that estimates a wear state and / or a damage state of the actual part from the vibration data of the actual machine acquired by the vibration data acquisition unit using the learning model; a state output unit that outputs the wear state and / or damage state of the actual part estimated by the state estimation unit; A condition monitoring device for a reciprocating motion device, comprising:

2. The condition monitoring device according to claim 1, the data generating unit generates waveform data relating to vibration at a monitoring position of the casing near a reciprocating part provided inside the casing as vibration data of the virtual part; The vibration data acquisition unit acquires waveform data relating to vibrations at the monitoring position of the reciprocating device during operation as vibration data of the actual device.

3. 3. The condition monitoring device according to claim 2, The waveform data is data obtained by filtering vibrations in a specific frequency range centered on the natural frequency of the reciprocating device and / or the actual part.

4. 3. The condition monitoring device according to claim 2, the data generating unit generates the waveform data generated by one reciprocating motion of the reciprocating unit as vibration data of the virtual part; The vibration data acquisition unit acquires the waveform data generated during one or more reciprocating motions of the reciprocating motion device while in operation as vibration data of the actual machine.

5. The condition monitoring device according to claim 1, The state output unit outputs the wear state and / or damage state of the actual part estimated by the state estimation unit together with a coincidence rate when estimated by the state estimation unit.

6. The condition monitoring device according to claim 1, a life prediction unit that predicts a life of the actual part based on the wear state and / or damage state of the actual part estimated by the condition estimation unit, a usage limit value that is a limit value for the amount of wear and / or the degree of damage in the actual part, and a usage time of the actual part.

7. The condition monitoring device according to claim 1, A condition monitoring device comprising a digital twin creation unit that creates a physical model from design data of the reciprocating motion device while it is in operation, and creates the digital twin by setting identification parameters for the generated physical model using vibration data of the actual reciprocating motion device while it is in operation.

8. a computer for monitoring the status of a reciprocating motion device, a data generating means for generating data for a virtual part using a digital twin of the reciprocating motion device, the data being equivalent to vibration data relating to damage and / or wear of an actual part disposed inside a casing of the reciprocating motion device; a learning means for learning a learning model using the correspondence between the vibration data of the virtual part generated by the data generating means and the amount of wear and / or the degree of damage of the real part; a vibration data acquisition means for acquiring vibration data of the actual machine from the reciprocating device during operation; a state estimation means for estimating a wear state and / or a damage state of the actual part from the vibration data of the actual machine acquired by the vibration data acquisition means using the learning model; a state output means for outputting the wear state and / or damage state of the actual part estimated by the state estimation means; A program to function as a

9. a status monitoring device that monitors the status of the reciprocating device; a vibration sensor provided at a monitoring position of the casing in the vicinity of a reciprocating portion provided inside the casing of the reciprocating device; A condition monitoring system comprising: The condition monitoring device is a data generation unit that generates vibration data related to damage and / or wear of a real part inside the casing, the data being equivalent to the vibration data related to vibrations at the monitoring position, for a virtual part using a digital twin of the reciprocating motion device; a learning unit that learns a learning model using the correspondence between the vibration data of the virtual part generated by the data generation unit and the amount of wear and / or the degree of damage of the real part; a vibration data acquisition unit that acquires vibration data of the actual reciprocating motion device from the vibration sensor while the reciprocating motion device is in operation; a state estimation unit that estimates a wear state and / or a damage state of the actual part from the vibration data of the actual machine acquired by the vibration data acquisition unit using the learning model; a state output unit that outputs the wear state and / or damage state of the actual part estimated by the state estimation unit; Equipped with The vibration sensor detects vibrations at the monitoring position of the reciprocating device during operation and transmits vibration data of the actual device to the condition monitoring device.

Images