Elevator fault diagnosis and early warning method based on multi-mode large model and related equipment
Through the elevator fault diagnosis and early warning method based on multimodal large model, the problems of low accuracy and long time in traditional diagnostic methods are solved, and more accurate and efficient fault diagnosis and early warning are achieved.
Patent Information
- Application Number
- CN202510255624.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-05
- Publication Date
- 2025-06-24
AI Technical Summary
Traditional elevator fault diagnosis methods have problems with low diagnostic accuracy and long diagnosis time, especially when multiple faults occur simultaneously.
The elevator fault diagnosis and early warning method based on multimodal large model is adopted. By obtaining real-time data in different dimensions, marking abnormal data, obtaining fault types and correlations, and generating comprehensive diagnostic results.
Improve the accuracy and efficiency of fault diagnosis, reduce inaccuracy or incomplete problems caused by a single diagnosis, and early warning of users of high failure risk.
Smart Images
Figure CN120191810A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of elevator fault diagnosis, and particularly relates to an elevator fault diagnosis and early warning method and related equipment based on a multimodal large model. Background Art
[0002] With the acceleration of the urbanization process, high-rise buildings are increasing day by day. As a key equipment for vertical transportation, the safety and reliability of elevators are of crucial importance. The traditional elevator maintenance method mainly relies on regular inspections and repairs after faults occur. This method has problems such as high cost, low efficiency, and inability to diagnose faults in a timely manner, and it is difficult to meet the high requirements of modern elevator operation. Therefore, it is of great practical significance to develop a system that can monitor the operating status of elevators in real time, predict potential faults, and perform maintenance in advance.
[0003] In related technologies, by monitoring the operating status of elevators in real time, faults that occur or are potential in elevators are diagnosed, thereby effectively improving the fault diagnosis efficiency and the elevator operation safety rate. However, the diagnosis methods in related technologies mainly perform single diagnosis based on abnormal data. However, once an elevator shows abnormalities, faults often occur in multiple places. Therefore, the single diagnosis method still has the defects of low diagnosis accuracy and long diagnosis time. Summary of the Invention
[0004] To help improve the accuracy and efficiency of fault diagnosis, this application provides an elevator fault diagnosis and early warning method and related equipment based on a multimodal large model.
[0005] In a first aspect, an elevator fault diagnosis and early warning method based on a multimodal large model provided by this application adopts the following technical solutions: An elevator fault diagnosis and early warning method based on a multimodal large model includes: Obtain real-time data of the target elevator in different dimensions and analyze the real-time data; If there is real-time data that does not meet the preset data requirements, mark the corresponding real-time data as abnormal data; Obtain the fault type corresponding to the abnormal data and the first number of items of the abnormal data; If the first number of items is equal to 1, generate a diagnosis result based on the fault type; If the first number of items is greater than 1, obtain the relevance of different faults based on the fault type; Generate a fault diagnosis result based on the relevance; If there is no real-time data that does not meet the preset data requirements, output an early warning plan based on the real-time data.
[0006] By adopting the above technical solution, real-time data in different dimensions is first obtained. If the real-time data does not meet the preset data requirements, it indicates that there is an abnormal situation with the real-time data, that is, there is a fault with the target elevator. Therefore, the real-time data is marked as abnormal data. Then, the fault type corresponding to the abnormal data and the first item number of the abnormal data are obtained. When the first item number is equal to 1, it means there is only 1 item of abnormal data. Therefore, a corresponding diagnosis result can be directly generated based on this item of abnormal data; When the first item number is greater than 1, it means there are multiple items of abnormal data simultaneously. To better diagnose the faults of the target elevator, first, the correlation of different faults is obtained according to the fault types of different abnormal data, and then a comprehensive diagnosis result is generated based on the correlation; If there is no real-time data that meets the preset data requirements, it means there is no abnormal situation with the real-time data, that is, there is no fault with the target elevator currently. Therefore, a corresponding early warning plan is generated based on the real-time data to remind the user to pay timely or key attention to which directions or aspects; When there are multiple items of abnormal data simultaneously, a more comprehensive and accurate comprehensive diagnosis result is generated according to the correlation degree between different fault types, which helps to reduce problems such as inaccurate or incomplete diagnosis results caused by single diagnosis of single data, thereby helping to improve the accuracy of fault diagnosis; At the same time, a more accurate and comprehensive comprehensive diagnosis result helps to reduce the time waste caused by multiple single diagnoses of single data, thereby helping to improve the efficiency of fault diagnosis.
[0007] Optionally, the generating a diagnosis result based on the fault type when the first item number is equal to 1 includes: If the first item number is equal to 1, it is determined whether the second item number corresponding to the fault type is equal to 1; If the second item number is equal to 1, a diagnosis result is generated based on the fault type; If the second item number is greater than 1, the fault times of different fault types are obtained based on historical fault records; The fault ratios corresponding to different fault times are obtained respectively; It is determined whether there is a fault ratio that exceeds the preset ratio threshold; If there is a fault ratio that exceeds the preset ratio threshold, a diagnosis result is generated based on the corresponding fault type.
[0008] Optionally, the generating a diagnosis result based on the corresponding fault type when there is a fault ratio that exceeds the preset ratio threshold includes: If there is a fault ratio that exceeds the preset ratio threshold, the fault ratio is marked as the target ratio; The third item number corresponding to the target ratio is obtained; If the third item number is equal to 1, generate a diagnosis result based on the corresponding fault type; If the third item number is greater than 1, obtain the historical diagnosis records of the target elevator; Based on the historical diagnosis records, obtain the historical development of different fault types; Generate a diagnosis result based on the historical development;
[0009] Optionally, the generating a diagnosis result based on the historical development includes: Based on the historical development, determine whether the fault is an intermediate fault, where the intermediate fault is a maintenance fault that is in progress but not completed; If the fault is the intermediate fault, generate a diagnosis result based on the historical diagnosis situation; If the fault is not the intermediate fault, obtain the unit growth rate of the fault type corresponding to different target ratios based on the historical development; Obtain the target growth rate based on the unit growth rate; Generate a diagnosis result based on the target growth rate;
[0010] Optionally, the relevance includes a first relevance, a second relevance, and a third relevance; the obtaining the relevance of different faults based on the fault type if the first item number is greater than 1 includes: If the first item number is greater than 1, determine whether there is a sub - fault type whose fault type is any other fault type; If there is a sub - fault type whose fault type is any other fault type, mark the corresponding fault type as the first relevance; If there is no sub - fault type whose fault type is any other fault type, determine whether there is a sub - fault type whose fault type is the same comprehensive fault; If there is a sub - fault type whose fault type is the same comprehensive fault, mark the corresponding fault type as the second relevance; If there is no sub - fault type whose fault type is the same comprehensive fault, mark the corresponding fault type as the third relevance.
[0011] Optionally, the generating a fault diagnosis result based on the relevance includes: If the relevance is the first relevance, obtain the upper - level fault type; Obtain the fault diagnosis result based on the upper - level fault type; If the relevance is the second relevance, obtain the comprehensive fault type; Obtain the fault diagnosis result based on the comprehensive fault type; If the relevance is the third relevance, then based on the historical fault records, it is determined whether there is a target record that matches the fault type corresponding to the current state; If the target record exists, a fault diagnosis result is generated based on the target record; Wherein, the upper-level fault type is a fault type including the sub-fault type, and the comprehensive fault type is a fault type jointly composed of multiple fault types.
[0012] Optionally, the step of if the target record exists, generating a fault diagnosis result based on the target record includes: If the target record exists, obtain the diagnosis success rate corresponding to the target record; Obtain the fourth item number of the target record corresponding to the diagnosis success rate exceeding the preset success rate threshold; If the fourth item number is less than 1, a fault diagnosis record is generated based on the fault type; If the fourth item number is equal to 1, a fault diagnosis result is generated based on the target record; If the fourth item number is greater than 1, obtain the diagnosis times and diagnosis scores of different target records, and obtain a comprehensive score based on the diagnosis times and diagnosis scores; Based on the comprehensive score, a fault diagnosis record is generated.
[0013] In a second aspect, the present application also discloses an elevator fault diagnosis and early warning system based on a multi-modal large model, adopting the following technical solution: An elevator fault diagnosis and early warning system based on a multi-modal large model, comprising: A first acquisition module, configured to acquire real-time data of a target elevator in different dimensions and analyze the real-time data; An abnormal marking module, if there is real-time data that does not meet the preset data requirements, the abnormal marking module is configured to mark the corresponding real-time data as abnormal data; A second acquisition module, configured to acquire the fault type corresponding to the abnormal data and the first item number of the abnormal data; A first generation module, if the first item number is equal to 1, the first generation module is configured to generate a diagnosis result based on the fault type; A third acquisition module, if the first item number is greater than 1, the third acquisition module is configured to obtain the relevance of different faults based on the fault type; A second generation module, configured to generate a fault diagnosis result based on the relevance; An early warning module, if the real-time data does not meet the preset data requirements, is used to output an early warning plan based on the real-time data.
[0014] By adopting the above technical solution, real-time data from different dimensions is first obtained. If the real-time data does not meet the preset data requirements, it indicates that there is an abnormal situation with the real-time data, that is, there is a fault with the target elevator. Therefore, the real-time data is marked as abnormal data. Then, the fault type corresponding to the abnormal data and the first item number of the abnormal data are obtained. When the first item number is equal to 1, it means there is only 1 item of abnormal data. Therefore, a corresponding diagnosis result can be directly generated based on this item of abnormal data. When the first item number is greater than 1, it means there are multiple items of abnormal data simultaneously. To better diagnose the faults of the target elevator, first, the correlation of different faults is obtained according to the fault types of different abnormal data, and then a comprehensive diagnosis result is generated based on the correlation. If there is no real-time data that meets the preset data requirements, it means there is no abnormal situation with the real-time data, that is, there is no fault with the target elevator at present. Therefore, a corresponding early warning plan is generated based on the real-time data to remind the user to pay timely or key attention to which directions or aspects. When there are multiple items of abnormal data simultaneously, a more comprehensive and accurate comprehensive diagnosis result is generated according to the correlation degree between different fault types, which helps to reduce problems such as inaccurate or incomplete diagnosis results caused by single diagnosis of single data, thus helping to improve the accuracy of fault diagnosis. At the same time, a more accurate and comprehensive comprehensive diagnosis result helps to reduce the time waste caused by multiple single diagnoses of single data, thus helping to improve the efficiency of fault diagnosis.
[0015] In the third aspect, a computer device provided by the present application adopts the following technical solution: An intelligent terminal includes a memory and a processor. The memory is used to store a computer program that can run on the processor. When the processor loads the computer program, it executes the method of the first aspect.
[0016] By adopting the above technical solution, a computer program is generated based on the method of the first aspect and stored in the memory to be loaded and executed by the processor. Thus, an intelligent terminal is made according to the memory and the processor, which is convenient for users to use.
[0017] In the fourth aspect, a computer-readable storage medium provided by the present application adopts the following technical solution: A computer-readable storage medium stores a computer program. When the computer program is loaded by the processor, it executes the method of the first aspect.
[0018] By adopting the above technical solution, a computer program is generated based on the method of the first aspect and stored in a computer-readable storage medium to be loaded and executed by a processor. Through the computer-readable storage medium, the readability and storage of the computer program are facilitated.
[0019] In summary, the present application includes the following beneficial technical effects: 1. When there are multiple pieces of abnormal data simultaneously, a more comprehensive and accurate comprehensive diagnosis result is generated according to the correlation degree between different fault types, which helps to reduce problems such as inaccurate or incomplete diagnosis results caused by single diagnosis of single data, thereby helping to improve the accuracy of fault diagnosis; at the same time, a more accurate and comprehensive comprehensive diagnosis result helps to reduce the time waste caused by multiple single diagnoses of single data, thereby helping to improve the efficiency of fault diagnosis.
[0020] 2. When the real-time data does not meet the preset data requirements, the real-time data is analyzed to obtain the data with high fault risk, and a corresponding early warning plan is generated based on these data, so as to remind the user to pay attention to these data with high fault risk in time and make corresponding preparations in advance. Description of the Drawings
[0021] Figure 1 is the main flowchart of an elevator fault diagnosis and early warning method based on a multi-modal large model according to an embodiment of the present application; Figure 2 is the flowchart of steps S201 to S206; Figure 3 is the flowchart of steps S301 to S306; Figure 4 is the flowchart of steps S401 to S405; Figure 5 is the flowchart of steps S501 to S505; Figure 6 is the flowchart of steps S601 to S606; Figure 7 is the flowchart of steps S701 to S706; Figure 8 is the module diagram of an elevator fault diagnosis and early warning system based on a multi-modal large model according to an embodiment of the present application.
[0022] Description of the Reference Numerals: 1. First acquisition module; 2. Abnormal marking module; 3. Second acquisition module; 4. First generation module; 5. Third acquisition module; 6. Second generation module; 7. Early warning module. Detailed Embodiments
[0023] In a first aspect, the present application discloses an elevator fault diagnosis and early warning method based on a multi-modal large model.
[0024] Refer to Figure 1 , an elevator fault diagnosis and early warning method based on a multi-modal large model, including steps S101 to S107: Step S101: Obtain real-time data of the target elevator in different dimensions and analyze the real-time data.
[0025] Specifically, in this embodiment, the target elevator is the elevator that needs to be fault-diagnosed or is about to be fault-diagnosed, and the real-time data is the real-time operation data of the target elevator, including operation status data (such as elevator position, operation direction, door opening and closing status, operation speed, etc.), equipment status data (such as equipment code, component status, energy consumption data, etc.), and environmental status data (such as temperature, humidity, and noise level inside the car), etc.
[0026] Step S102: If there is real-time data that does not meet the preset data requirements, mark the corresponding real-time data as abnormal data.
[0027] Specifically, compare the real-time data with the preset data requirements to determine whether the real-time data all meet the corresponding preset data requirements. If not, mark the corresponding real-time data as abnormal data. Among them, abnormal data is the real-time data that does not meet the corresponding preset data requirements, and the preset corresponding requirements are the pre-set judgment criteria for determining whether the real-time data is abnormal. In this embodiment, the preset data requirements vary according to different real-time data.
[0028] Step S103: Obtain the fault type corresponding to the abnormal data and the first item number of the abnormal data.
[0029] Specifically, the fault type is the possible fault type that causes the real-time data to become this abnormal data. In this embodiment, the occurrence of one abnormal data may be caused by various different fault types. For example, a traction system fault or a speed control device fault can cause the elevator's running speed to be abnormal; the first item number is the item number of the abnormal data.
[0030] Step S104: If the first item number is equal to 1, generate a diagnosis result based on the fault type.
[0031] Specifically, if the first item number is equal to 1, it means there is only one abnormal data. Therefore, the corresponding diagnosis result can be directly generated according to the fault type corresponding to this abnormal data. In this embodiment, the diagnosis result is the result of determining what kind of fault exists in the target elevator.
[0032] Step S105: If the first item number is greater than 1, obtain the correlation between different faults based on the fault type.
[0033] Specifically, in this embodiment, the relevance refers to the nature of mutual association or mutual influence existing between different faults, including the first association, the second association, and the third association.
[0034] Step S106: Generate a fault diagnosis result based on the relevance.
[0035] Specifically, in this embodiment, according to different relevance, a corresponding comprehensive diagnosis result is generated, which helps to diagnose the real fault occurring in the target elevator, rather than a single fault diagnosed for a single abnormal data item, thereby helping to improve the accuracy and efficiency of fault diagnosis.
[0036] Step S107: If there is no real-time data that does not meet the preset data requirements, output a warning plan based on the real-time data.
[0037] Specifically, in this embodiment, by analyzing the real-time data, it helps to give a warning to the data with higher risks, thereby helping the user to make preparations and handle them in advance.
[0038] For the elevator fault diagnosis and warning method based on the multi-modal large model provided in this embodiment, first obtain real-time data of different dimensions. If the real-time data does not meet the preset data requirements, it means that the real-time data has an abnormal situation, that is, the target elevator has a fault. Therefore, mark the real-time data as abnormal data. Then obtain the fault type corresponding to the abnormal data and the first item number of the abnormal data. When the first item number is equal to 1, it means that there is only 1 item of abnormal data. Therefore, a corresponding diagnosis result can be directly generated according to this item of abnormal data.
[0039] When the first item number is greater than 1, it means that there are multiple items of abnormal data at the same time. To better diagnose the faults of the target elevator, first obtain the relevance of different faults according to the fault types of different abnormal data, and then generate a comprehensive diagnosis result according to the relevance. If there is no real-time data that meets the preset data requirements, it means that the real-time data has no abnormal situation, that is, the target elevator currently has no fault. Therefore, generate a corresponding warning plan according to the real-time data to remind the user to pay timely or key attention to which directions or aspects.
[0040] When there are multiple items of abnormal data at the same time, a more comprehensive and accurate comprehensive diagnosis result is generated according to the correlation degree between different fault types, which helps to reduce problems such as inaccurate or incomplete diagnosis results caused by single diagnosis of single data items, thereby helping to improve the accuracy of fault diagnosis. At the same time, a more accurate and comprehensive comprehensive diagnosis result helps to reduce the time waste caused by multiple single diagnoses of single data, thereby helping to improve the efficiency of fault diagnosis.
[0041] Reference Figure 2 In one implementation of this embodiment, step S104: if the first item number is equal to 1, generating a diagnosis result based on the fault type includes steps S201 to S206: Step S201: If the first number is equal to 1, determine whether the second number corresponding to the fault type is equal to 1.
[0042] Specifically, the second number, that is, the number of fault types corresponding to a single abnormal data, may be one or more.
[0043] Step S202: If the second number is equal to 1, a diagnosis result is generated based on the fault type.
[0044] Specifically, in this embodiment, if the second number is equal to 1, it means that there is only one fault type that can cause the abnormal data to appear, so the corresponding diagnosis result can be directly generated according to the fault type.
[0045] Step S203: If the second number is greater than 1, the number of faults of different fault types is obtained based on historical fault records.
[0046] Specifically, the historical fault record is a record of fault diagnosis of the target elevator at a historical moment. In this embodiment, all fault diagnosis records after the target elevator is officially put into operation can be selected as the historical fault record; the number of faults is the number of times different fault types cause the occurrence of such abnormal data.
[0047] Step S204: respectively obtain the fault proportions corresponding to different fault times.
[0048] Specifically, in this embodiment, the fault proportion is the percentage of different fault times in all times that cause the abnormal data to appear. For example, the elevator running speed abnormality occurs 10 times in total, of which 3 times are caused by traction system failure, then the fault proportion corresponding to the traction system failure is 30%.
[0049] Step S205: Determine whether the fault ratio exceeds a preset ratio threshold.
[0050] Specifically, the preset proportion threshold is a pre-set judgment standard for determining whether a diagnosis result can be directly generated according to the fault proportion. In this embodiment, the preset proportion threshold can be set to 30% or set to other values according to actual conditions and user needs.
[0051] Step S206: If the existing fault ratio exceeds a preset ratio threshold, a diagnosis result is generated based on the corresponding fault type.
[0052] Specifically, if the failure ratio exceeds the preset ratio threshold, it indicates that the failure ratio is relatively high. The diagnostic result can be directly generated based on the failure ratio. Therefore, the failure type corresponding to the failure ratio exceeding the preset ratio threshold is directly used as the diagnostic result.
[0053] Refer to Figure 3 , in one implementation manner of this embodiment, if the failure ratio exceeds the preset ratio threshold in step S206, generating a diagnostic result based on the corresponding failure type includes steps S301 to S306: Step S301: If the failure ratio exceeds the preset ratio threshold, mark the failure ratio as the target ratio.
[0054] Specifically, in this embodiment, the target ratio is the failure ratio exceeding the preset ratio threshold.
[0055] Step S302: Obtain the third item number corresponding to the target ratio.
[0056] Specifically, the third item number is the quantity of the target ratio, that is, the quantity of the failure ratio exceeding the preset ratio threshold.
[0057] Step S303: If the third item number is equal to 1, generate a diagnostic result based on the corresponding failure type.
[0058] Specifically, in this embodiment, if the third item number is equal to 1, it means that there is only one target ratio. Therefore, the diagnostic result can be directly generated according to the failure type corresponding to the target ratio.
[0059] Step S304: If the third item number is greater than 1, obtain the historical diagnostic record of the target elevator.
[0060] Specifically, in this embodiment, the historical diagnostic record is the record of diagnosing the faults existing in the target elevator at historical moments.
[0061] Step S305: Based on the historical diagnostic record, obtain the historical development of different failure types.
[0062] Specifically, in this embodiment, the historical development is the change situation of the failure type at historical moments, including the quantity change situation and the change speed, etc.
[0063] Step S306: Generate a diagnostic result based on the historical development.
[0064] Refer to Figure 4 , in one implementation manner of this embodiment, step S306 generating a diagnostic result based on the historical development includes steps S401 to S405: Step S401: Based on the historical development situation, determine whether the fault is an intermediate fault, where an intermediate fault is a maintenance fault that is in progress but not yet completed.
[0065] Specifically, in this embodiment, a fault repair is in progress, but the fault repair has not been completed and is still in the test stage or the debugging stage. Therefore, the real-time data corresponding to the target elevator still has anomalies, so this type of fault is determined to be an intermediate fault.
[0066] Step S402: If the fault is an intermediate fault, generate a diagnostic result based on the historical diagnosis situation.
[0067] Specifically, in this embodiment, if the fault is an intermediate fault, then use the original diagnostic result as the current diagnostic result according to the historical diagnosis situation.
[0068] Step S403: If the fault is not an intermediate fault, obtain the unit growth rate of the fault types corresponding to different target ratios based on the historical development situation.
[0069] Specifically, the unit growth rate is the growth rate of the proportion of a certain fault within a unit time. In this embodiment, the unit time can be 1 month or 1 year, or can be set to other times according to actual needs.
[0070] Step S404: Obtain the target growth rate based on the unit growth rate.
[0071] Specifically, in this embodiment, the target growth rate is the unit growth rate with the largest value among all unit growth rates.
[0072] Step S405: Generate a diagnostic result based on the target growth rate.
[0073] Specifically, in this embodiment, select the fault type corresponding to the proportion of the fault corresponding to the target growth rate to generate a diagnostic result.
[0074] Refer to Figure 5 , in one implementation manner of this embodiment, Step S105: If the first number is greater than 1, obtain the relevance of different faults based on the fault type, including Steps S501 to S505: Step S501: If the first number is greater than 1, determine whether there is a sub-fault type whose fault type is any other fault type.
[0075] Specifically, in this embodiment, the sub-fault type is the subordinate fault type. For example, the fault type includes car and door system faults and car structure damage (such as car wall deformation and car bottom loosening). Obviously, the car and door system includes car structure damage, that is to say, car structure damage belongs to the subordinate fault or sub-fault of the car and door system.
[0076] Step S502: If there is a fault type that is a sub-fault type of any other fault type, the corresponding fault type is marked as a first association.
[0077] Specifically, in this embodiment, when there is a fault type that is a sub-fault of another or more fault types, the corresponding fault type is determined to be the first association.
[0078] Step S503: If there is no sub-fault type whose fault type is any other fault type, it is determined whether there is a sub-fault type whose fault type is the same comprehensive fault.
[0079] Specifically, a comprehensive fault is a comprehensive fault, such as an abnormal door opening and closing fault and a speed control and safety protection fault. In this embodiment, if there are multiple fault types that are the same sub-fault type of the comprehensive fault, it is determined that there is a fault type that is a sub-fault type of the same comprehensive fault. For example, the fault type includes two items: door leaf shaking and jamming during door opening and closing, and the door cannot be closed automatically when the door closing button is pressed. Both of these fault types are sub-fault types of the abnormal door opening and closing fault. Therefore, it is determined that there is a fault type that is a sub-fault type of the same comprehensive fault.
[0080] Step S504: If there are fault types that are sub-fault types of the same comprehensive fault, the corresponding fault types are marked as the second association.
[0081] Specifically, in this embodiment, when there are multiple fault types that are simultaneously sub-fault types of one or more comprehensive faults, the corresponding fault types are identified as the second association.
[0082] Step S505: If there is no sub-fault type whose fault type is the same comprehensive fault, the corresponding fault type is marked as the third association.
[0083] Specifically, the third association is an association other than the first association and the second association. In this embodiment, the third association may also be represented as having little or no association.
[0084] Reference Figure 6 In one implementation of this embodiment, step S106 generates a fault diagnosis result based on the correlation, including steps S601 to S606: Step S601: If the correlation is the first correlation, the upper fault type is obtained.
[0085] Specifically, in this embodiment, the upper fault type is a fault type that includes sub-fault types.
[0086] Step S602: Obtain a fault diagnosis result based on the upper fault type.
[0087] Specifically, in this embodiment, if the relevance is the first relevance, a diagnostic result is generated according to the upper-level fault type.
[0088] Step S603: If the relevance is the second relevance, obtain the comprehensive fault type.
[0089] Specifically, in this embodiment, the comprehensive fault type is a fault type composed of multiple fault types.
[0090] Step S604: Based on the comprehensive fault type, obtain the fault diagnosis result.
[0091] Specifically, in this embodiment, if the relevance is the second relevance, a diagnostic result is generated according to the comprehensive fault type.
[0092] Step S605: If the relevance is the third relevance, based on the historical fault records, determine whether there is a target record that matches the fault type corresponding to the current state.
[0093] Specifically, in this embodiment, the target record is a fault diagnosis record that matches the fault type corresponding to the current state.
[0094] Step S606: If there is a target record, generate a fault diagnosis result based on the target record.
[0095] Specifically, in this embodiment, if there is a target record, it means that a fault identical or very similar to the current state has been diagnosed at a historical moment. Therefore, the diagnostic result in the target record can be used as the current diagnostic result.
[0096] Refer to Figure 7 , in one implementation manner of this embodiment, if there is a target record in step S606, generating a fault diagnosis result based on the target record includes steps S701 to S706: Step S701: If there is a target record, obtain the diagnostic success rate corresponding to the target record.
[0097] Specifically, in this embodiment, the diagnostic success rate is the proportion of successfully and accurately diagnosing the faults existing in the target elevator.
[0098] Step S702: Obtain the fourth item number of the target records corresponding to the diagnostic success rate exceeding the preset success rate threshold.
[0099] Specifically, in this embodiment, the fourth item number is the number of target records in all target records whose diagnostic success rate exceeds the preset success rate threshold.
[0100] Step S703: If the fourth item number is less than 1, generate a fault diagnosis record based on the fault type.
[0101] Specifically, in this embodiment, if the fourth item number is less than 1, it means that there is no target record with a diagnostic success rate exceeding the preset success rate threshold. Therefore, the diagnostic record can only be generated based on all the existing fault types currently.
[0102] Step S704: If the fourth item number is equal to 1, generate a fault diagnosis result based on the target record.
[0103] Specifically, in this embodiment, if the fourth item number is equal to 1, it means that there is exactly one target record with a diagnostic success rate exceeding the preset success rate threshold. Therefore, directly generate a fault diagnosis result based on this target record.
[0104] Step S705: If the fourth item number is greater than 1, obtain the diagnosis times and diagnosis scores of different target records, and obtain a comprehensive score based on the diagnosis times and diagnosis scores.
[0105] Specifically, if the fourth item number is greater than 1, it means that there are multiple target records with a diagnostic success rate exceeding the preset success rate threshold. To improve the accuracy of fault diagnosis, further obtain the diagnosis times and diagnosis scores of the target records. In this embodiment, the diagnosis times are the number of times the target records corresponding to different fault types appear; the diagnosis score is the score obtained after each diagnosis, and this score can be set according to the accuracy, efficiency, and complexity of the diagnosis result, or can be actually evaluated by the staff.
[0106] According to the diagnosis times and diagnosis scores, obtain the corresponding comprehensive score, where the comprehensive score can be obtained by adding the products of the diagnosis times and diagnosis scores multiplied by the corresponding coefficients respectively.
[0107] Step S706: Generate a fault diagnosis record based on the comprehensive score.
[0108] In a second aspect, the present application also discloses an elevator fault diagnosis and warning system based on a multi-modal large model.
[0109] Refer to Figure 8 , an elevator fault diagnosis and warning system based on a multi-modal large model, including: The first acquisition module is used to acquire real-time data of different dimensions of the target elevator and analyze the real-time data; The abnormal marking module is used to mark the corresponding real-time data as abnormal data if there is real-time data that does not meet the preset data requirements; The second acquisition module is used to acquire the fault type corresponding to the abnormal data and the first item number of the abnormal data; The first generation module is used to generate a diagnosis result based on the fault type if the first item number is equal to 1; A third acquisition module, if the number of the first items is greater than 1, is configured to acquire the correlation of different faults based on the fault type. A second generation module is configured to generate a fault diagnosis result based on the correlation. An early warning module, if there is no real-time data that does not meet the preset data requirements, is configured to output an early warning plan based on the real-time data.
[0110] In a third aspect, an embodiment of the present application discloses an intelligent terminal, including a memory and a processor. The memory is used to store a computer program that can run on the processor. When the processor loads the computer program, it executes a method for elevator fault diagnosis and early warning based on a multimodal large model in the above embodiment.
[0111] In a fourth aspect, an embodiment of the present application discloses a computer-readable storage medium, and a computer program is stored in the computer-readable storage medium. When the computer program is loaded by the processor, it executes a method for elevator fault diagnosis and early warning based on a multimodal large model in the above embodiment.
[0112] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of the present application shall be covered by the protection scope of the present application.
Claims
1. An elevator fault diagnosis and early warning method based on a multi-modal large model, characterized in that: include: Acquire real-time data of different dimensions of the target elevator and analyze the real-time data; If there is any real-time data that does not meet the preset data requirements, the corresponding real-time data is marked as abnormal data; Acquire a fault type corresponding to the abnormal data and a first number of the abnormal data; If the first number is equal to 1, generating a diagnosis result based on the fault type; If the first number is greater than 1, obtaining correlations between different faults based on the fault types; Based on the correlation, generating a fault diagnosis result; If the real-time data does not meet the preset data requirements, an early warning solution is output based on the real-time data.
2. The elevator fault diagnosis and early warning method based on a multi-modal large model according to claim 1 is characterized in that: If the first number is equal to 1, generating a diagnosis result based on the fault type includes: If the first number is equal to 1, determining whether the second number corresponding to the fault type is equal to 1; If the second number is equal to 1, generating a diagnosis result based on the fault type; If the second number is greater than 1, obtaining the number of faults of different fault types based on historical fault records; Obtain the fault proportions corresponding to different fault times respectively; Determine whether the fault ratio exceeds a preset ratio threshold; If the fault ratio exceeds the preset ratio threshold, a diagnosis result is generated based on the corresponding fault type.
3. The elevator fault diagnosis and early warning method based on a multi-modal large model according to claim 2 is characterized in that: If the fault ratio exceeds the preset ratio threshold, generating a diagnosis result based on the corresponding fault type includes: If the fault ratio exceeds the preset ratio threshold, the fault ratio is marked as the target ratio; Obtain the third number corresponding to the target proportion; If the third number is equal to 1, a diagnosis result is generated based on the corresponding fault type; If the third number is greater than 1, obtaining the historical diagnostic record of the target elevator; Based on the historical diagnosis records, obtaining historical developments of different fault types; Based on the historical development, a diagnosis result is generated.
4. The elevator fault diagnosis and early warning method based on a multi-modal large model according to claim 3 is characterized in that: The generating of the diagnosis result based on the historical development situation includes: Based on the historical development situation, determine whether the fault is an intermediate fault, where the intermediate fault is a maintenance fault that is in progress but not completed; If the fault is the intermediate fault, generating a diagnosis result based on historical diagnosis conditions; If the fault is not the intermediate fault, obtaining unit growth rates of fault types corresponding to different target proportions based on the historical development situation; Based on the unit growth rate, obtaining a target growth rate; Based on the target growth rate, a diagnosis result is generated.
5. The elevator fault diagnosis and early warning method based on multi-modal large model according to claim 1 is characterized in that: The correlation includes a first correlation, a second correlation and a third correlation; if the first number is greater than 1, then based on the fault type, obtaining the correlation of different faults includes: If the first number is greater than 1, determining whether the fault type is a sub-fault type of any other fault type; If there is a fault type that is a sub-fault type of any other fault type, marking the corresponding fault type as the first association; If the fault type does not exist and is a sub-fault type of any other fault type, determine whether the fault type exists and is a sub-fault type of the same comprehensive fault; If the fault type exists as a sub-fault type of the same comprehensive fault, marking the corresponding fault type as the second association; If the fault type does not exist as a sub-fault type of the same comprehensive fault, the corresponding fault type is marked as the third association.
6. The elevator fault diagnosis and early warning method based on multi-modal large model according to claim 5 is characterized in that: The generating of the fault diagnosis result based on the correlation includes: If the correlation is the first correlation, obtaining a higher-level fault type; Based on the upper fault type, obtaining a fault diagnosis result; If the correlation is the second correlation, obtaining a comprehensive fault type; Based on the comprehensive fault type, obtaining a fault diagnosis result; If the correlation is the third correlation, judging whether there is a target record matching the fault type corresponding to the current state based on the historical fault records; If the target record exists, generating a fault diagnosis result based on the target record; The upper fault type is a fault type that includes the sub-fault type, and the comprehensive fault type is a fault type composed of a plurality of the fault types.
7. The elevator fault diagnosis and early warning method based on multi-modal large model according to claim 6 is characterized in that: If the target record exists, generating a fault diagnosis result based on the target record includes: If the target record exists, obtaining the diagnosis success rate corresponding to the target record; Obtaining the fourth number of the target records corresponding to the diagnosis success rate exceeding the preset success rate threshold; If the fourth number is less than 1, generating a fault diagnosis record based on the fault type; If the fourth number is equal to 1, a fault diagnosis result is generated based on the target record; If the fourth number is greater than 1, obtaining the number of diagnoses and the diagnosis scores of the different target records, and obtaining a comprehensive score based on the number of diagnoses and the diagnosis scores; Based on the comprehensive score, a fault diagnosis record is generated.
8. An elevator fault diagnosis and early warning system based on a multi-modal large model, characterized in that: include: A first acquisition module is used to acquire real-time data of different dimensions of a target elevator and analyze the real-time data; An abnormal marking module, if there is real-time data that does not meet the preset data requirements, the abnormal marking module is used to mark the corresponding real-time data as abnormal data; A second acquisition module, used to acquire a fault type corresponding to the abnormal data and a first number of the abnormal data; a first generating module, wherein if the first number is equal to 1, the first generating module is used to generate a diagnosis result based on the fault type; a third acquisition module, wherein if the first number is greater than 1, the third acquisition module is used to acquire the correlation between different faults based on the fault type; A second generating module, used for generating a fault diagnosis result based on the correlation; The early warning module is used to output an early warning plan based on the real-time data if the real-time data does not meet the preset data requirements.
9. An intelligent terminal, comprising a memory and a processor, characterized in that: The memory is used to store a computer program that can be run on the processor, and when the processor loads the computer program, it executes the method according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored therein, characterized in that: When the computer program is loaded by a processor, the method according to any one of claims 1 to 7 is executed.