Method and system for comprehensively diagnosing faults in rotating machinery - Patents.com
The method integrates multiple diagnostic techniques to quantify rotating machinery defects, providing accurate defect severity assessment and optimizing maintenance through weighted evaluations.
Patent Information
- Application Number
- JP2023540867
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-12-31
- Filing Date
- 2021-12-29
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2041-12-29
AI Technical Summary
Existing diagnostic systems for rotating machinery provide qualitative evaluations independently, lacking a method to automatically quantify equipment status by integrating multiple diagnostic techniques.
A method and system that simultaneously apply various diagnostic techniques, including machine learning, to determine a defect level based on vibration signals, frequency analysis, and operating information, applying weighted values to quantify defect severity.
Enables quantitative evaluation of equipment condition, accurately determining defect progression and maintenance needs, thereby optimizing maintenance schedules.
Smart Images

Figure 0007579455000002 
Figure 0007579455000003 
Figure 0007579455000004
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a method and system for diagnosing faults in rotating machinery, and more particularly to a method and system for diagnosing faults in rotating machinery that simultaneously coordinates a variety of direct and indirect diagnostic techniques. [Background technology]
[0002] In general, a diagnostic system for diagnosing the state of a rotating machine can monitor trends in equipment vibrations and operating variables. In addition, the diagnostic system can change a monitoring period depending on whether or not there is an abnormality in the rotating machine, and can predict the state of the equipment by analyzing the transition of such fluctuations.
[0003] A variety of direct and indirect diagnostic techniques are used to diagnose the condition of rotating machinery. For example, the diagnostic system may monitor equipment by dividing fault frequency bands into fine sections. Or, for example, the diagnostic system may automatically diagnose equipment based on fault characteristics and / or equipment information using verified diagnostic rules. Or, for example, the diagnostic system may diagnose equipment by utilizing machine learning that extracts features based on multiple data and realizes a classification model through learning. Or, for example, the diagnostic system may group similar equipment to compare equipment with each other and diagnose the equipment. Or, for example, the diagnostic system may diagnose using operation information.
[0004] Here, each diagnostic technique independently outputs a result for the condition of the rotating machine, and the result is also a qualitative assessment. Summary of the Invention [Problem to be solved by the invention]
[0005] An object of the present disclosure is to provide a rotating machinery fault diagnosis method and system that automatically quantifies equipment condition by simultaneously and coordinately executing multiple diagnostic techniques based on information obtained from the rotating machinery. [Means for solving the problem]
[0006] The method for diagnosing a defect in a rotating machine according to the present disclosure includes a step of determining a defect level based on data diagnosing the condition of the rotating machine, the data diagnosing the condition of the rotating machine including at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a defect of the rotating machine, or an overall vibration value of the rotating machine, a step of applying a weighting value to the defect level based on at least one of information regarding a defect in the condition history data of the rotating machine or whether or not an alarm has occurred associated with operation information of the rotating machine, and a step of determining a defect severity of the rotating machine based on the defect level to which the weighting value has been applied.
[0007] For example, the condition history data of the rotating machine may include information related to the maintenance history of the rotating machine and the like, and defects in the condition history data of the rotating machine may be the most frequent defects in the like.
[0008] For example, an alarm is generated when a monitored item related to the operation information of the rotating machine exceeds a preset reference value.
[0009] For example, the rotary machine operating information may include at least one of a pump flow rate associated with the rotary machine, a pre-stage pressure associated with the rotary machine, or a fluid temperature associated with the rotary machine.
[0010] For example, a weighted value may be added to the defect level based on the correspondence between the most frequent defects in similar equipment associated with the rotating machine and the defect condition of the rotating machine associated with the defect level.
[0011] For example, a weight value may be added to the defect level based on the occurrence of an alarm associated with the operation information of the rotating machine.
[0012] For example, based on a discrepancy between the most frequent defect in similar equipment associated with the rotating machine and the defect status of the rotating machine associated with the defect level, it is determined whether an alarm associated with the operating information of the rotating machine will be generated.
[0013] For example, a first fault value for the rotating machine is diagnosed by machine learning based on a feature vector associated with a vibration signal of the rotating machine. The machine learning determines whether or not the rotating machine has a fault. The machine learning is performed based on a feature vector associated with a vibration signal of the rotating machine.
[0014] For example, a second fault value may be diagnosed based on a frequency associated with a fault in the rotating machine and the first fault value.
[0015] For example, the third fault value may be diagnosed based on the overall vibration value of the rotating machine and the second fault value.
[0016] For example, a fault level of the rotating machine may be determined based on at least one of the first fault value, the second fault value, or the third fault value.
[0017] For example, based on the absence of a fault in the rotating machine, the first fault value may be determined to be zero and the fault severity may be determined to be zero.
[0018] For example, based on the presence of a defect in the rotating machine, the first defect value may be determined based on a total sample associated with the rotating machine and a defect sample associated with the rotating machine.
[0019] For example, the second fault value may be determined to be a preset first value based on a frequency associated with a fault in the rotating machine being within a preset range.
[0020] For example, the second defect value may be determined to be the first defect value and the defect level may be determined to be the first defect value based on a frequency associated with a defect in the rotating machine being outside a preset range.
[0021] For example, the third defect value may be determined to be the second defect value and the defect level may be determined to be the second defect value based on the overall vibration value of the rotating machine being less than a first critical value.
[0022] For example, based on the overall vibration value of the rotating machine being greater than a first critical value, the third defect value may be determined to a predetermined second value, and the defect level may be determined to the predetermined second value.
[0023] For example, based on the overall vibration value of the rotating machine being greater than a second critical value, the third defect value may be determined to a preset third value, and the defect level may be determined to the preset third value.
[0024] Meanwhile, the rotating machine defect diagnosis system according to the present disclosure determines a defect level based on data diagnosing the condition of the rotating machine, where the data diagnosing the condition of the rotating machine includes at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a defect of the rotating machine, or an overall vibration value of the rotating machine, and applies a weighting value to the defect level based on at least one of information regarding a defect in the condition history data of the rotating machine or whether or not an alarm has occurred regarding operation information of the rotating machine, and determines the defect severity of the rotating machine based on the defect level to which the weighting value has been applied.
[0025] Meanwhile, the calculation processor of the rotating machine defect diagnosis system according to the present disclosure determines a defect level based on data diagnosing the condition of the rotating machine, where the data diagnosing the condition of the rotating machine includes at least one of a feature vector associated with the vibration signal of the rotating machine, a frequency associated with a defect of the rotating machine, or an overall vibration value of the rotating machine, and applies a weighting value to the defect level based on at least one of information regarding a defect in the condition history data of the rotating machine or whether or not an alarm has occurred related to the operation information of the rotating machine, and determines the defect severity of the rotating machine based on the defect level to which the weighting value has been applied. Effect of the Invention
[0026] The rotating machinery defect diagnosis method and system disclosed herein can quantitatively evaluate minute changes in the condition of the equipment, and by utilizing the evaluation result value (e.g., defect severity), it is possible to accurately confirm the degree of progression of the equipment defect and more accurately determine the maintenance timing, lifespan, etc. of the equipment.
[0027] The technical effects of the present disclosure as described above are not limited to the effects mentioned above, and other technical effects not mentioned will be clearly understood by those skilled in the art from the following description. [Brief description of the drawings]
[0028] [Figure 1] 1 is a configuration diagram showing a fault diagnosis system for a rotating machine according to an embodiment of the present disclosure; [Diagram 2] FIG. 2 is a configuration diagram showing a calculation processor of the defect diagnosis system according to an embodiment of the present disclosure. [Diagram 3] 4 is a flow chart illustrating a method for comprehensively diagnosing faults on a rotating machine according to one embodiment of the present disclosure. [Figure 4] 4 is a flow chart illustrating steps for calculating a fault level for a rotating machine according to one embodiment of the present disclosure. [Diagram 5]4 is a flow chart illustrating steps for applying weights based on fault levels for a rotating machine according to one embodiment of the present disclosure. [Figure 6] 4 is a flowchart illustrating a method for deriving an overall fault severity based on diagnostic results for a rotating machine according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0029] Hereinafter, the embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present embodiments are not limited to the embodiments disclosed below, and may be realized in various forms, and the present embodiments are provided only to fully disclose the present invention and to fully inform those skilled in the art of the scope of the invention. The shapes of elements in the drawings may be exaggerated for clearer explanation, and elements represented by the same reference numerals in the drawings refer to the same elements.
[0030] Fig. 1 is a configuration diagram showing a fault diagnosis system for a rotating machine according to an embodiment of the present disclosure. Fig. 2 is a configuration diagram showing a calculation processor of the fault diagnosis system according to an embodiment of the present disclosure.
[0031] Here, the rotary machine 10 may be various rotary devices such as a pump, a compressor, a fan, etc. However, this is for the purpose of explaining the present disclosure, and the type of the rotary machine 10 is not limited.
[0032] Meanwhile, the defect diagnosis system 100 acquires data from the rotating machine 10, constructs the data, and when a defect diagnosis of the rotating machine 10 is required again, can perform an automatic predictive diagnosis using the constructed data. In addition, the defect diagnosis system 100 outputs diagnosed defect values, allowing an inspector to intuitively grasp the presence or absence of an abnormality in the rotating machine 10 and determine the timing of replacement and maintenance.
[0033] First, the rotating machine 10 may be equipped with sensors for acquiring various information including operation information of the rotating machine 10. The sensors are linked to the fault diagnosis system 100 so that the acquired data is provided to the fault diagnosis system 100. However, this is for the purpose of explaining the present disclosure, and it will be made clear that data related to the rotating machine 10 may not be acquired by sensors but may be acquired directly by an operator and input to the predictive diagnosis system.
[0034] The fault diagnosis system 100 includes a storage unit 110 in which data provided from the rotating machine 10 is stored, an arithmetic processor 120 that performs predictive diagnosis based on data acquired from the rotating machine 10, and an output unit 130 that displays fault information. Here, the arithmetic processor 120 includes an arithmetic device provided in a computer, software for arithmetic, and a computer language for arithmetic, and the arithmetic processor 120 can execute the following processes.
[0035] Hereinafter, a method for diagnosing a fault in a rotating machine according to various embodiments of the present disclosure will be described in detail, with the above-mentioned components being given the same reference numerals and detailed descriptions thereof being omitted.
[0036] For example, the fault diagnosis system 100 monitors trends in equipment vibrations and operating variables related to the equipment. The fault diagnosis system 100 can change the monitoring period depending on whether or not there is an abnormality in the equipment. The fault diagnosis system 100 analyzes fluctuation trends, and the fault diagnosis system 100 can predict the equipment status from the analyzed fluctuation trends. The fault diagnosis system 100 sets the monitoring target and the monitoring period. The fault diagnosis system 100 monitors vibration trends for each equipment point. The fault diagnosis system 100 monitors operating variables during the same time period.
[0037] For example, the defect diagnosis system 100 can diagnose defects based on narrow-band diagnosis. That is, the defect diagnosis system 100 monitors equipment by dividing the defect frequency bands into smaller bands for each piece of equipment. The defect diagnosis system 100 predicts not only the presence or absence of a defect but also the type of defect. The defect diagnosis system 100 derives the defect frequency bands of each piece of equipment. The defect diagnosis system 100 sets the tolerance range for each defect frequency band. For example, the defect diagnosis system 100 sets an alert if the defect is within 2σ from the reference value, and sets a fault if the defect is within 3σ from the reference value. The defect diagnosis system 100 diagnoses the defect frequency by period. Such narrow-band diagnosis techniques are effective in detecting early-stage defects.
[0038] For example, the fault diagnosis system 100 can diagnose faults based on rule-based diagnosis. That is, the fault diagnosis system 100 automatically diagnoses equipment based on fault characteristics and / or equipment information. The fault diagnosis system 100 can implement verified diagnostic rules with logic in the form of a decision tree. When data is input, the fault diagnosis system 100 can automatically derive expert-level diagnostic results. Such rule-based diagnosis uses verified diagnostic rules to increase the reliability of diagnostic results and allows the process and content of diagnostic results to be tracked.
[0039] For example, the defect diagnosis system 100 can diagnose defects by comparing equipment of the same type. That is, the defect diagnosis system 100 groups equipment of the same type, and the defect diagnosis system 100 can diagnose defects by comparing the equipment of the same type with each other. The defect diagnosis system 100 derives defects that occur frequently for each type of equipment. The defect diagnosis system 100 groups equipment of the same type that performs the same function into equipment of the same type. By deriving defects that occur frequently for each type of equipment of the same type, it is possible to secure vulnerable parts in advance and to prepare for emergencies in the event of a sudden equipment failure. The defect diagnosis system 100 optimizes maintenance cycles by reflecting the characteristics of equipment of the same type.
[0040] For example, the defect diagnosis system 100 can diagnose defects through machine learning. That is, the defect diagnosis system 100 diagnoses defects through artificial intelligence that utilizes a large amount of data. The defect diagnosis system 100 extracts features from various data, and the defect diagnosis system 100 realizes a classification model through learning. For example, the defect diagnosis system 100 can determine a classification model based on normality, abnormality, and the type of defect. A small number of diagnosis models based on such machine learning can be applied to various equipment.
[0041] FIG. 3 is a flow chart illustrating a method for comprehensively diagnosing faults on a rotating machine according to one embodiment of the present disclosure.
[0042] 3, in step S310, the fault diagnosis system 100 determines a fault level based on data obtained by diagnosing the condition of the rotating machine. In step S330, the fault diagnosis system 100 applies a weighting value to the fault level based on information about a fault in the condition history data of the rotating machine or on the occurrence or non-occurrence of an alarm related to the operation information of the rotating machine. In step S350, the fault diagnosis system 100 determines the fault severity of the rotating machine based on the fault level to which the weighting value has been applied.
[0043] For example, the condition history data of the rotating machine includes a maintenance history of the rotating machine and similar equipment information related to the rotating machine. A defect in the condition history data of the rotating machine may be a defect that occurs most frequently in the similar equipment.
[0044] For example, an alarm may be generated based on a monitoring item associated with operation information of a rotating machine exceeding a preset reference value. The operation information of the rotating machine may include at least one of a flow rate of a pump associated with the rotating machine, a front-end pressure associated with the rotating machine, or a fluid temperature associated with the rotating machine.
[0045] For example, the fault diagnosis system 100 may add a weight to the fault level based on a match between the most frequent fault in the same equipment related to the rotating machine and the fault state of the rotating machine related to the fault level. Alternatively, the fault diagnosis system 100 may add a weight to the fault level based on the occurrence of an alarm related to the operation information of the rotating machine. For example, the fault diagnosis system 100 may determine whether an alarm related to the operation information of the rotating machine has occurred based on a match between the most frequent fault in the same equipment related to the rotating machine and the fault state of the rotating machine related to the fault level.
[0046] For example, the data obtained by diagnosing the condition of the rotating machine from the fault diagnosis system 100 includes at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a defect of the rotating machine, or an overall vibration value of the rotating machine. Here, machine learning may be performed based on the feature vector associated with the vibration signal of the rotating machine. For example, the fault diagnosis system 100 determines the first defect value to be 0 based on the absence of a defect in the rotating machine. In this case, the fault diagnosis system 100 determines the defect level to be 0 based on the first defect value being 0. Alternatively, the fault diagnosis system 100 may determine the first defect value based on an overall sample associated with the rotating machine and a defect sample associated with the rotating machine based on the presence of a defect in the rotating machine.
[0047] For example, the fault diagnosis system 100 determines the second fault value to be the first preset value based on the frequency associated with the fault of the rotating machine being within a preset range. Alternatively, the fault diagnosis system 100 determines the second fault value to be the first fault value based on the frequency associated with the fault of the rotating machine being outside of a preset range. In this case, the fault diagnosis system 100 determines the fault level to be the first fault value.
[0048] For example, the fault diagnosis system 100 determines the third fault value to be the second fault value based on the overall vibration value of the rotating machine being smaller than the first critical value. At this time, the fault diagnosis system 100 determines the fault level to be the second fault value. Alternatively, the fault diagnosis system 100 determines the third fault value to be a second value set in advance based on the overall vibration value of the rotating machine being larger than the first critical value. At this time, the fault diagnosis system 100 determines the fault level to be the second value set in advance. Alternatively, the fault diagnosis system 100 determines the third fault value to be a third value set in advance based on the overall vibration value of the rotating machine being larger than the second critical value. At this time, the fault diagnosis system 100 determines the fault level to be the third value set in advance.
[0049] FIG. 4 is a flow chart illustrating steps for calculating a fault level for a rotating machine according to one embodiment of the present disclosure.
[0050] First, the fault diagnosis system 100 can receive data obtained by diagnosing the condition of the rotating machine and perform machine learning based on the data obtained by diagnosing the condition of the rotating machine.
[0051] 4, in step S410, the fault diagnosis system 100 determines whether a fault has occurred in the rotating machine based on the results of the machine learning. For example, if a fault has occurred in the rotating machine, the fault diagnosis system 100 may calculate a proportion of samples that indicate a fault. For example, the proportion of samples that indicate a fault may be a proportion of fault samples associated with the rotating machine to all samples associated with the rotating machine. The fault diagnosis system 100 determines a first fault value based on the proportion of samples that indicate a fault. Alternatively, for example, if no fault has occurred in the rotating machine, the fault diagnosis system 100 may determine the first fault value to be 0.
[0052] In step S430, the fault diagnosis system 100 determines whether an alarm is generated for a frequency associated with a defect of the rotating machine. In step S410, if the first fault value is determined to be a value other than 0, the fault diagnosis system 100 may determine whether an alarm is generated for a frequency associated with a defect of the rotating machine. For example, an alarm is generated for a frequency associated with a defect of the rotating machine when the frequency associated with the defect of the rotating machine is within a preset range. For example, an alarm is not generated for a frequency associated with a defect of the rotating machine when the frequency associated with the defect of the rotating machine is outside a preset range. In the case where an alarm is generated for a frequency associated with a defect of the rotating machine, the fault diagnosis system 100 determines a second fault value to be a first value set in advance. Alternatively, in the case where an alarm is not generated for a frequency associated with a defect of the rotating machine, the fault diagnosis system 100 determines the second fault value to be the first fault value.
[0053] In step S450, the fault diagnosis system 100 determines whether the overall vibration value of the rotating machine exceeds an allowable standard. In step S430, when an alarm for a frequency associated with a defect of the rotating machine occurs, the fault diagnosis system 100 determines whether the overall vibration value of the rotating machine exceeds an allowable standard. For example, the fault diagnosis system 100 determines the third fault value to be the second fault value based on the overall vibration value of the rotating machine being smaller than the first critical value. Alternatively, the fault diagnosis system 100 determines the third fault value to be a second value set in advance based on the overall vibration value of the rotating machine being larger than the first critical value. Alternatively, the fault diagnosis system 100 determines the third fault value to be a third value set in advance based on the overall vibration value of the rotating machine being larger than the second critical value. For example, the second critical value may be a value larger than the first critical value.
[0054] In step S470, the fault diagnosis system 100 determines a fault level for the rotating machine. For example, in step S430, if an alarm for a frequency associated with a fault in the rotating machine does not occur, the fault diagnosis system 100 determines the second fault value to be the first fault value and the fault level to be the first fault value. For example, in step S450, if the overall vibration value of the rotating machine is smaller than the first critical value, the fault diagnosis system 100 determines the third fault value to be the second fault value and the fault level to be the second fault value. For example, in step S450, if the overall vibration value of the rotating machine is larger than the first critical value, the fault diagnosis system 100 determines the third fault value to be the first value set in advance and the fault level to be the third fault value. For example, in step S450, if the overall vibration value of the rotating machine is larger than the second critical value, the fault diagnosis system 100 determines the third fault value to be the second value set in advance and the fault level to be the third fault value.
[0055] In step S490, the fault diagnosis system 100 determines that the rotating machine is in a normal state. For example, if it is diagnosed in step S410 that no fault has occurred in the rotating machine based on machine learning, the fault diagnosis system 100 may determine that the rotating machine is in a normal state. For example, the normal state of the rotating machine means that the fault level is 0.
[0056] For example, if the fault diagnosis system 100 determines that the rotating machine is in a normal state, steps S430 and S450 may be omitted. For example, if no alarms are generated for the frequency associated with a fault in the rotating machine, step S450 may be omitted.
[0057] FIG. 5 is a flow chart illustrating steps for applying weights based on fault levels to a rotating machine according to one embodiment of the present disclosure.
[0058] 5, in step S510, the fault diagnosis system 100 determines whether a fault in the condition history data of the rotating machine matches a fault state of the rotating machine associated with a fault level. For example, the fault diagnosis system 100 determines whether a fault state of the rotating machine associated with a fault level matches a fault that occurs most frequently in a similar equipment associated with the rotating machine. Here, the condition history data of the rotating machine may include a maintenance history of the rotating machine and information associated with the similar equipment.
[0059] In step S530, the fault diagnosis system 100 determines whether an alarm associated with the operation information is generated. For example, when a defect on the state history data of the rotating machine does not match a defect state of the rotating machine associated with a defect level, the fault diagnosis system 100 may determine whether an alarm associated with the operation information is generated. For example, when a defect on the state history data of the rotating machine does not match a defect state of the rotating machine associated with a defect level, the fault diagnosis system 100 may determine whether a monitoring item associated with the operation information of the rotating machine exceeds a preset reference value. That is, the fault diagnosis system 100 may generate an alarm based on a monitoring item associated with the operation information of the rotating machine exceeding a preset reference value. Here, the operation information of the rotating machine includes at least one of a flow rate of a pump associated with the rotating machine, a front-to-rear stage pressure associated with the rotating machine, or a fluid temperature associated with the rotating machine.
[0060] In step S550, the fault diagnosis system 100 may not apply a weighting value to the fault level. For example, if the fault state of the rotating machine associated with the fault level matches the most frequent fault in the same type of equipment associated with the rotating machine, the fault diagnosis system 100 may apply a weighting value to the fault level. For example, if an alarm related to the operation information of the rotating machine occurs, the fault diagnosis system 100 applies a weighting value to the fault level.
[0061] In step S570, the fault diagnosis system 100 applies a weighting value to the fault level. For example, if the most frequent fault in a homogeneous equipment related to the rotating machine and the fault state of the rotating machine related to the fault level are inconsistent, the fault diagnosis system 100 may not apply a weighting value to the fault level. For example, if an alarm associated with the operation information of the rotating machine does not occur, the fault diagnosis system 100 may apply a weighting value to the fault level. For example, if the most frequent fault in a homogeneous equipment related to the rotating machine and the fault state of the rotating machine related to the fault level are inconsistent and an alarm associated with the operation information of the rotating machine does not occur, the fault diagnosis system 100 may apply a weighting value to the fault level.
[0062] For example, if the fault condition of the rotating machine associated with the fault level coincides with the most frequent fault in the same equipment associated with the rotating machine, steps S530 and S550 may be omitted.
[0063] FIG. 6 is a flow chart illustrating a method for deriving an overall fault severity based on diagnostic results for a rotating machine according to one embodiment of the present disclosure.
[0064] 6, the severity calculation program may be a program that automatically quantifies the state of equipment from a micro defect to a transient defect (e.g., defect level (hereinafter, DL) 1 to DL3). For example, a defect diagnosis system may include the severity calculation program.
[0065] In order to comprehensively consider and quantify the state of the equipment from various aspects, the diagnosis results from each diagnosis technique can be collectively inquired to the fault diagnosis system, and the diagnosis results can be input to the fault diagnosis system.
[0066] DL1 is a step for evaluating whether an equipment (e.g., a rotating machine) has a micro defect or no symptoms. For example, the fault diagnosis system queries machine learning results for the equipment or performs machine learning based on data related to the equipment. Thereafter, if the machine learning results indicate a defect, the fault diagnosis system can calculate a proportion of samples that indicate a defect to determine a first DL value. For example, the proportion of samples that indicate a defect may be a proportion of defective samples related to the rotating machine to all samples related to the rotating machine. For example, the proportion of samples that indicate a defect is expressed by the following Equation 1.
[0067]
number
[0068] For example, if the machine learning results indicate no fault, the fault diagnosis system can determine the DL value to be 0 and the equipment status to be normal.
[0069] For example, since vibration signals have unique characteristics depending on the equipment state, the defect diagnosis system can calculate a feature vector that can well express each characteristic using machine learning diagnosis techniques. The defect diagnosis system can classify by features that well represent the minimized distance between features of the same state and the maximized distance between features of different states, and can region by defect state. The defect diagnosis system can learn features by state (e.g. (normal, by defect type)) for a large amount of previous data, and can classify regions by state. At this time, when new data is input, the defect diagnosis system can predict the equipment state according to the region where the data is input. Therefore, since subjective human intervention is minimized, it becomes possible to make an objective judgment on defects without preconceptions.
[0070] DL2 may be a step for evaluating the severity of the defect. The defect diagnosis system may inquire whether an alarm occurs at the defect association frequency. For example, whether an alarm occurs at the defect association frequency is determined based on whether the frequency associated with the defect of the rotating machine is within a preset range or outside the preset range. That is, for example, if the frequency associated with the defect of the rotating machine is within a preset range, an alarm may occur. Alternatively, if the frequency associated with the defect of the rotating machine is outside the preset range, an alarm may not occur. At this time, if an alarm occurs for the defect association frequency, the defect diagnosis system may determine the second DL value to be 0.4. Here, 0.4 may be a preset value or may be set to a different value according to various embodiments of the present disclosure. Alternatively, if an alarm does not occur for the defect association frequency, the defect diagnosis system may determine the first DL value to be the final DL value, i.e., the second DL value.
[0071] For example, narrowband frequency diagnosis technique is a method to monitor and evaluate the frequency of the region of interest by dividing the frequency region, unlike evaluating the entire frequency region as one energy value. In other words, since various defects that occur in equipment induce amplitude changes in specific frequency regions, the defect diagnosis system can monitor the equipment by dividing the frequency region of interest with parameters and setting an allowable range. Therefore, the defect diagnosis system can obtain information about defects by frequency region of interest and identify the cause of the defects.
[0072] DL3 is a step of evaluating the equipment as being in a warning or dangerous state. For example, if the overall vibration value exceeds an allowable standard for an alert (e.g., the first critical value), the fault diagnosis system may determine the third DL value to be 0.6. Here, 0.6 is a preset value and may be set to a different value according to various embodiments of the present disclosure. For example, if the overall vibration value exceeds an allowable standard for a fault (e.g., the second critical value), the fault diagnosis system may determine the third DL value to be 0.8. For example, if the overall vibration value does not exceed an allowable standard (e.g., the first critical value), the fault diagnosis system may determine the second DL value to be the final DL value, i.e., the third DL value.
[0073] For example, a diagnosis technique using overall vibration values can be a method of evaluating the overall vibration values output from equipment based on limit or allowable values according to international standards or recommendations from the equipment manufacturer. The fault diagnosis system can classify equipment into types, capacities, support structures, etc., and apply evaluation standards appropriate for the relevant equipment. Management standards such as international standard vibration standards (ISO APi, etc.) are continually revised to improve the validity of the standards, so the fault diagnosis system can evaluate defects based on the revised management standards. Therefore, when diagnosing defects for a state in which an abnormal value of the allowable standard occurs, the fault diagnosis system can diagnose the defect more accurately.
[0074] The first added weight 1 is a step for calculating an additional weight to the DL value. For example, the defect diagnosis system can refer to and / or determine the equipment status history data by linking the equipment maintenance history and the same type of equipment. At this time, if the most frequently occurring defect matches the diagnosis result of the current equipment, the defect diagnosis system can add a weight to the DL value. For example, if the most frequently occurring defect matches the current equipment status, the defect diagnosis system applies a weight by multiplying the DL value by 1.1. For example, if the most frequently occurring defect does not match the current equipment status, the defect diagnosis system does not apply a weight to the DL value.
[0075] For example, a technique for comparing and diagnosing similar equipment may be a technique that uses equipment status history data linking the maintenance history and similar equipment. That is, the defect diagnosis system can add additional severity when the most frequently occurring defect in the similar equipment matches the current equipment status. For example, the defect diagnosis system can reclassify the similar equipment diagnosed as having the most frequently occurring defect, and can utilize at least one of narrowband frequency information and machine learning information of the corresponding similar equipment for predictive diagnosis of equipment defects. Therefore, the defect diagnosis system can centrally monitor the characteristic values of the most frequently occurring defects, thereby minimizing the number of monitoring targets.
[0076] The second added weight 2 is a step for calculating an additional weight to the DL value. For example, the second added weight is considered when the first added weight is not applied. That is, the fault diagnosis system can consider the second added weight when the most frequently occurring fault in the same type of equipment does not match the current equipment state. For example, the fault diagnosis system can generate an alarm or inquire if an alarm has occurred when a monitoring item related to the PMS (power management system) operation information exceeds the allowable standard. When an alarm occurs, the fault diagnosis system can apply a weight by multiplying the DL value by 1.1. For example, when a monitoring item related to the PMS (power management system) operation information does not exceed the allowable standard, the fault diagnosis system does not apply a weight to the DL value.
[0077] For example, a fault diagnosis technique using operation information can utilize operation information that affects the equipment. For example, the operation information that affects the equipment can include a pump flow rate associated with the equipment, upstream and downstream pressures associated with the equipment, and fluid temperatures associated with the equipment. The reliability of the diagnosis results can be improved by correlation analysis linking vibration characteristics and operation information.
[0078] The fault diagnosis system can evaluate the finally calculated DL value as the fault severity that quantitatively indicates the equipment condition.
[0079] In this way, the rotating machinery defect diagnosis method and system disclosed herein can quantitatively evaluate minute changes in the equipment condition, and by utilizing the evaluation result value (severity), the extent to which the equipment defect has progressed can be accurately determined, and the maintenance timing, lifespan, etc. of the equipment condition can be more accurately determined.
[0080] The embodiment of the present disclosure described above and shown in the drawings should not be construed as limiting the technical subject matter of the present disclosure. The scope of protection of the present disclosure is limited only by the matters described in the claims, and those with ordinary knowledge in the technical field of the present disclosure can improve and modify the technical subject matter of the present disclosure in various forms. Therefore, as long as such improvements and modifications are obvious to those with ordinary knowledge, they will belong to the scope of protection of the present disclosure. [Explanation of symbols]
[0081] 10: Rotating machinery 100: Rotating machinery fault diagnosis system 110: Storage area 120: Calculation processor 130: Output section
Claims
1. Although the defect level is determined based on data obtained by diagnosing the condition of the rotating machine, the data diagnosing the condition of the rotating machine includes at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a fault of the rotating machine, or an overall vibration value of the rotating machine; applying a weight to the defect level based on at least one of information regarding a defect on the state history data of the rotating machine or the occurrence or non-occurrence of an alarm associated with operation information of the rotating machine; and determining a fault severity of the rotating machine based on the fault levels to which the weights have been applied; The condition history data of the rotating machine includes a maintenance history of the rotating machine and information related to similar equipment; and A method for diagnosing a fault in a rotating machine, wherein the fault in the condition history data of the rotating machine is the fault that occurs most frequently in the same type of equipment.
2. 2. The method of claim 1, further comprising the step of generating an alarm when a monitoring item related to operation information of the rotating machine exceeds a preset reference value.
3. 2. The method for diagnosing faults in a rotating machine according to claim 1, wherein the operating information of the rotating machine includes at least one of a pump flow rate associated with the rotating machine, an upstream or downstream pressure associated with the rotating machine, or a fluid temperature associated with the rotating machine.
4. The step of applying weights to the defect levels comprises:
2. The method for diagnosing a fault in a rotating machine as described in claim 1, further comprising the step of adding a weighted value to the fault level based on a match between the most frequent fault in similar equipment associated with the rotating machine and the fault condition of the rotating machine associated with the fault level.
5. The step of applying weights to the defect levels comprises:
2. The method for diagnosing a fault in a rotating machine according to claim 1, further comprising the step of adding a weight value to the fault level based on the occurrence of an alarm related to operation information of the rotating machine.
6. 2. A method for diagnosing a fault in a rotating machine as described in claim 1, wherein whether or not an alarm has been generated in association with the operation information of the rotating machine is determined based on a discrepancy between the most frequent fault in similar equipment related to the rotating machine and the fault state of the rotating machine related to the fault level.
7. The step of determining defect severity comprises: diagnosing a first fault value for the rotating machine through machine learning based on a feature vector associated with a vibration signal of the rotating machine; diagnosing a second fault value based on a frequency associated with a fault in the rotating machine and the first fault value; diagnosing a third fault value based on the overall vibration value of the rotating machine and the second fault value; and The method of claim 1 , further comprising determining a fault level of the rotating machine based on at least one of the first fault value, the second fault value, or the third fault value.
8. The step of diagnosing the first defect value comprises: The method for diagnosing a fault in a rotating machine according to claim 7 , further comprising a step of determining whether or not the rotating machine has a fault by the machine learning.
9. Upon the existence of a defect in the rotating machine, the first defect value is determined based on a total sample associated with the rotating machine and a defect sample associated with the rotating machine; and 9. The method of claim 8, wherein the second fault value is determined to be a predetermined first value based on a frequency associated with a fault in the rotating machine being within a predetermined range.
10. The third defect value is determined to be the second defect value based on the overall vibration value of the rotating machine being less than a first critical value; and The method for diagnosing a fault in a rotating machine according to claim 9 , wherein the fault level is determined to be the second fault value.
11. The third defect value is determined to be a second preset value based on the overall vibration value of the rotating machine being greater than a first critical value; and The method for diagnosing a fault in a rotating machine according to claim 9 , wherein the fault level is determined to be the second preset value.
12. The third defect value is determined to be a third preset value based on the overall vibration value of the rotating machine being greater than a second critical value; and The method for diagnosing a fault in a rotating machine according to claim 9 , wherein the fault level is determined to be the preset third value.
13. Although the defect level is determined based on data obtained by diagnosing the condition of the rotating machine, the data diagnosing the condition of the rotating machine includes at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a fault of the rotating machine, or an overall vibration value of the rotating machine; applying a weighting value to the defect level based on at least one of information regarding defects in the state history data of the rotating machine or the occurrence or non-occurrence of an alarm associated with operation information of the rotating machine; and determining a fault severity for the rotating machine based on the fault levels to which the weights have been applied; The condition history data of the rotating machine includes a maintenance history of the rotating machine and information related to similar equipment; and A fault diagnosis system for a rotating machine, wherein a fault in the condition history data of the rotating machine is a fault that occurs most frequently in the same type of equipment.
14. Although the defect level is determined based on data obtained by diagnosing the condition of the rotating machine, the data diagnosing the condition of the rotating machine includes at least one of a feature vector associated with a vibration signal of the rotating machine, a frequency associated with a fault of the rotating machine, or an overall vibration value of the rotating machine; applying a weighting value to the defect level based on at least one of information regarding defects in the state history data of the rotating machine or the occurrence or non-occurrence of an alarm associated with operation information of the rotating machine; and determining a fault severity for the rotating machine based on the fault levels to which the weights have been applied; The condition history data of the rotating machine includes a maintenance history of the rotating machine and information related to similar equipment; and A calculation processor of a fault diagnosis system for a rotating machine, wherein the fault in the condition history data of the rotating machine is the fault that occurs most frequently in the same type of equipment.
Citation Information
Patent Citations
Method of diagnosing anomalies in low speed rotational bearing of elevator
JP2013224853A
State monitoring system and wind power generator
JP2017173321A
State monitoring device, state monitoring system and state monitoring method
JP2018179977A
Mold transformer diagnose system using severity assessment
KR1020130129679A
Plant defect diagnostic system using vibration characteristics
KR1020150058928A
