Gearbox fault prediction method and system

Abstract

The invention relates to a gearbox fault prediction method and system, and belongs to the technical field of fault detection, and the method comprises the steps: obtaining the operation state data of a gearbox in real time, and carrying out the time-frequency domain analysis of the operation state data, and obtaining the time-frequency data; when the time-frequency data meets the triggering condition, predicting the time-frequency data based on the trained-completely neural network model, and determining a fault time node; based on a knowledge graph and a reasoning algorithm, performing fault type reasoning on the time-frequency data to obtain a fault type; and determining a fault prediction result based on the fault time node and the fault type. The fault time is predicted through the neural network model, the fault type is predicted through the knowledge graph, and real-time and comprehensive detection and fault prediction of the ship gearbox are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of fault detection, and in particular to a gearbox fault prediction method and system. Background Art

[0002] The ship gearbox is an important part of the ship's power transmission system, responsible for transmitting the engine power to the propulsion system. Under long-term high-load operation and complex marine environments, the gearbox is prone to wear, breakage, fatigue and other faults. These faults will not only reduce the transmission efficiency, but may also cause ship shutdown or even serious safety accidents.

[0003] Traditional maintenance methods usually rely on regular inspections, which cannot monitor the health status of the gearbox in real time, which may result in failures not being discovered in time, increasing downtime and repair costs. Summary of the invention

[0004] In view of this, it is necessary to provide a gearbox fault prediction method and system to solve the problem that the traditional maintenance method in the prior art usually relies on regular maintenance, which can easily lead to failure to detect faults in time, increase downtime and maintenance costs.

[0005] In order to solve the above problems, the present application provides a gearbox fault prediction method, comprising: Acquire the operating status data of the gearbox in real time, and perform time-frequency domain analysis on the operating status data to obtain time-frequency data; When the time-frequency data meets the trigger condition, the time-frequency data is predicted based on the well-trained neural network model to determine the fault time node; Based on the knowledge graph and the reasoning algorithm, the fault type is inferred on the time-frequency data to obtain the fault type; A fault prediction result is determined based on the fault time node and the fault type.

[0006] In some possible implementations, the operating status data includes vibration, temperature, and acoustic data.

[0007] In some possible implementations, the operating status data is collected by sensors arranged on gearbox measuring points, and the gearbox measuring points include gear meshing points, bearings, transmission shafts, motors, and planetary gearboxes.

[0008] In some possible implementations, the operating status data is acquired through a sliding window technology.

[0009] In some possible implementations, after acquiring the operating status data of the gearbox in real time, the method further includes: The operating status data is subjected to data denoising, smoothing and normalization processing.

[0010] In some possible implementations, performing time-frequency domain analysis on the operating status data to obtain time-frequency data includes: Based on short-time Fourier transform and wavelet transform technology, the operating status data is analyzed in time and frequency domain to obtain time and frequency data.

[0011] In some possible implementations, the trigger condition is that the time-frequency data undergoes a sudden change or the monitoring time of the time-frequency data is at a periodic detection time node.

[0012] In some possible implementations, the fully trained neural network model adopts an LSTM model.

[0013] In some possible implementations, the knowledge graph is constructed through the following steps: Obtain historical fault data; Based on the historical fault data, determining nodes, the nodes including key components of the gearbox, component status data, fault types, operating conditions, and historical fault cases; Based on the historical fault data and the nodes, determining relationships between nodes, the relationships including component-fault relationships, fault-sensor signal relationships, operating condition-fault relationships, and historical case-fault relationships; A knowledge graph is constructed based on the nodes and the relationships.

[0014] The present application also provides a gearbox fault prediction system, comprising: A data acquisition module is used to acquire the operating status data of the gearbox in real time, and perform time-frequency domain analysis on the operating status data to obtain time-frequency data; A fault time prediction module is used to predict the time-frequency data based on a well-trained neural network model to determine the fault time node when the time-frequency data meets the trigger condition; A fault type prediction module is used to perform fault type inference on the time-frequency data based on the knowledge graph and the inference algorithm to obtain the fault type when the time-frequency data meets the trigger condition; The fault result generating module is used to determine the fault prediction result based on the fault time node and the fault type.

[0015] The beneficial effects of the present application are as follows: the gearbox fault prediction method provided by the present application detects the operating status of the gearbox in real time, and when the trigger conditions are met, uses a neural network model to predict the fault time, and uses a knowledge graph to predict the fault type, thereby achieving real-time and comprehensive detection and fault prediction of the ship gearbox, and improving the safety and reliability of equipment operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic flow chart of an embodiment of a gearbox fault prediction method provided in the present application; Figure 2 A schematic diagram of an embodiment of the knowledge graph construction method provided in this application; Figure 3 A schematic diagram of the structure of an embodiment of the gearbox fault prediction system provided in the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0018] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts may be implemented out of order, and steps without logical contextual relationships may be reversed in order or implemented simultaneously. In addition, those skilled in the art may add one or more other operations to the flowcharts, or may remove one or more operations from the flowcharts, under the guidance of the content of this application. Some of the block diagrams shown in the accompanying drawings are functional entities, and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor systems and / or microcontroller systems.

[0019] The descriptions of "first", "second", etc. involved in the embodiments of the present application are only used to describe the implicit purpose, and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" and "second" may explicitly or implicitly include at least one of the features. "And / or" describes the association relationship of the associated objects, indicating that three relationships may exist, for example: A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone.

[0020] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0021] The present application provides a gearbox fault prediction method, which is described below.

[0022] Figure 1 A schematic diagram of an embodiment of a gearbox fault prediction method provided in this application is shown in FIG. Figure 1 As shown, the gearbox fault prediction method includes: S101, acquiring the operating status data of the gearbox in real time, and performing time-frequency domain analysis on the operating status data to obtain time-frequency data; It should be noted that short-time Fourier transform and wavelet transform techniques can be used for time-frequency domain analysis.

[0023] S102, when the time-frequency data meets the triggering condition, the time-frequency data is predicted based on the well-trained neural network model to determine the fault time node; It should be noted that the triggering condition is that the time-frequency data undergoes a sudden change or the monitoring time of the time-frequency data is at a periodic detection time node, and the LSTM model is used as the fully trained neural network model.

[0024] S103, based on the knowledge graph and the reasoning algorithm, the fault type is inferred for the time-frequency data to obtain the fault type; It should be noted that compared with the traditional method based on a single data source, the knowledge graph can integrate multi-source heterogeneous data to achieve more accurate fault type prediction. Through the knowledge graph, the causal relationship and reasoning path between nodes can be clearly displayed to help users understand the prediction results and causes of failures, rather than just the "black box" prediction output. The reasoning algorithm can use GCN-DistMult.

[0025] S104: Determine a fault prediction result based on the fault time node and the fault type.

[0026] Compared with the existing technology, this application detects the operating status of the gearbox in real time, uses a neural network model to predict the fault time when the trigger conditions are met, and uses a knowledge graph to predict the fault type, thereby achieving real-time and comprehensive detection and fault prediction of the ship gearbox, and improving the safety and reliability of equipment operation.

[0027] In order to better obtain the operating status data, in some embodiments, the operating status data is collected by sensors arranged on the gearbox measuring points. In order to reasonably arrange the sensors and ensure comprehensive coverage of the operating status of the gearbox, the gearbox measuring points include gear meshing points, bearings, drive shafts, motors and planetary gearboxes, etc. The sensors use high-frequency vibration sensors, temperature sensors and acoustic sensors, and the operating status data include vibration, temperature and acoustic data. At the same time, a streaming data processing platform (such as Apache Kafka) is used to process real-time data through sliding window technology to generate a time series data stream containing vibration, temperature and acoustic signals, i.e., operating status data.

[0028] In order to improve data quality, in some embodiments, after the operating status data of the gearbox is acquired in real time, the operating status data is preprocessed. Specifically, the method further includes: Perform data denoising, smoothing and normalization on the operating status data.

[0029] In order to better construct the knowledge graph, in some embodiments, such as Figure 2 As shown in the figure, the knowledge graph is constructed through the following steps: S201, obtaining historical fault data; It should be noted that, specifically, real-time operating data is collected from various sensors of the ship gearbox (vibration, temperature, acoustics, etc.), analyzed in combination with historical failure cases, and the raw data is cleaned and classified through a preprocessing module.

[0030] S202, determining nodes based on historical fault data, where the nodes include key components of the gearbox, component status data, fault type, operating conditions, and historical fault cases; It should be noted that, in this application, the key components of the gearbox include gears, bearings, seals, transmission systems, etc.; component status data include characteristic parameters of temperature, vibration, and sound signals, which are collected from various sensors to reflect the health status of the components (such as vibration amplitude, frequency characteristics, etc.); fault types include common fault types such as gear wear, tooth breakage, bearing fatigue, and lubrication failure; operating conditions include speed, torque, and load conditions, which represent the actual working environment of the equipment; historical failure cases include records of previous failures, such as failure time, cause, impact, and repair measures.

[0031] S203, based on the historical fault data and nodes, determining the relationship between nodes, the relationship including component-fault relationship, fault-sensor signal relationship, operating condition-fault relationship and historical case-fault relationship; It should be noted that in the present application, component-fault relationships include "gearbox->occurrence->wear fault", "bearing->association->fatigue failure"; fault-sensor signal relationships include "wear->triggering->abnormal vibration signal", "overheating->triggering->temperature rise"; operating condition-fault relationships include "high load->possible triggering->gear breakage", "high temperature->acceleration->lubrication failure"; historical case-fault relationships include "Case A->triggering->tooth breakage", "Case B->association->abnormal vibration".

[0032] S204. Build a knowledge graph based on nodes and relationships.

[0033] It should be noted that when using the graph database Neo4j for storage and management, the relationship between nodes is stored through entity-relationship triples (such as node-relationship-node).

[0034] In order to better implement a gearbox fault prediction method in an embodiment of the present application, based on a gearbox fault prediction method, correspondingly, Figure 3 As shown, the embodiment of the present application further provides a gearbox fault prediction system 300, comprising: The data acquisition module 301 is used to acquire the operating status data of the gearbox in real time, and perform time-frequency domain analysis on the operating status data to obtain time-frequency data; The fault time prediction module 302 is used to predict the time-frequency data based on the well-trained neural network model and determine the fault time node when the time-frequency data meets the trigger condition; The fault type prediction module 303 is used to perform fault type inference on the time-frequency data based on the knowledge graph and the inference algorithm to obtain the fault type when the time-frequency data meets the trigger condition; The fault result generating module 304 is used to determine the fault prediction result based on the fault time node and the fault type.

[0035] In order to better apply the fault prediction results, in some embodiments, the gearbox fault prediction system also includes a maintenance decision and management module, which is used to receive the fault prediction results, including the predicted fault time node and type, and dynamically adjust the maintenance plan based on the predicted fault severity by applying preset maintenance decision rules; at the same time, detailed maintenance tasks are generated according to the decision rules, including maintenance time nodes, required parts and equipment, maintenance steps, etc., and are automatically recorded in the maintenance log.

[0036] In order to better apply the fault prediction results, in some embodiments, the gearbox fault prediction system also includes a user interface and an alarm system module. The user interface and the alarm system module mainly link the previous information (including operating status data, time-frequency data and fault prediction results) together using corresponding interfaces and communication methods. The displayed information includes: 1) Real-time curve of sensor data; operating status of key components such as gearboxes and bearings, with normal, warning, and fault status marked; 2) Display the predicted fault type and occurrence timeline; 3) Prediction information displays possible fault information and shows maintenance time and suggestions; 4) Detailed tasks of the maintenance plan, showing a list of maintenance tasks to be performed; 5) Provide historical maintenance record query, showing completed maintenance tasks, details and results; 6) When a fault is predicted to occur, a real-time alarm prompt will pop up at the top or edge of the interface, with different levels of severity depending on the time approaching; 7) Historical maintenance record download window; Among them, QNetworkAccessManager is used to communicate the information of the output end with Qt over the network, the above information is passed in, and then QSqlDatabase is used to store and manage the data information, and finally QtDesigner is used to design the interface to display the above content.

[0037] Therefore, the gearbox fault prediction system includes data acquisition, data preprocessing, feature extraction, real-time data stream processing, fault prediction, maintenance decision and management, and user interface and early warning system modules. Through the collaborative work of these modules, the system forms a comprehensive intelligent maintenance solution system platform to achieve comprehensive monitoring, fault prediction and intelligent maintenance of ship gearboxes.

[0038] The gearbox fault prediction system 300 provided in the above embodiment can implement the technical solution described in the above gearbox fault prediction method embodiment. The specific implementation principles of the above units can refer to the corresponding contents in the above gearbox fault prediction method embodiment, which will not be repeated here.

[0039] The above is a detailed introduction to a gearbox fault prediction method provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0040] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.

Claims

1. A gearbox fault prediction method, characterized in that: include: Acquire the operating status data of the gearbox in real time, and perform time-frequency domain analysis on the operating status data to obtain time-frequency data; When the time-frequency data meets the trigger condition, the time-frequency data is predicted based on the well-trained neural network model to determine the fault time node; Based on the knowledge graph and the reasoning algorithm, the fault type is inferred on the time-frequency data to obtain the fault type; A fault prediction result is determined based on the fault time node and the fault type.

2. The gearbox fault prediction method according to claim 1, characterized in that: The operating status data includes vibration, temperature and acoustic data.

3. The gearbox fault prediction method according to claim 1, characterized in that: The operating status data is collected and acquired by sensors arranged on gearbox measuring points, and the gearbox measuring points include gear meshing points, bearings, transmission shafts, motors and planetary gearboxes.

4. The gearbox fault prediction method according to claim 1, characterized in that: The operating status data is obtained through a sliding window technology.

5. The gearbox fault prediction method according to claim 1, characterized in that: After obtaining the operating status data of the gearbox in real time, the method further includes: The operating status data is subjected to data denoising, smoothing and normalization processing.

6. The gearbox fault prediction method according to claim 1, characterized in that: Performing time-frequency domain analysis on the operating status data to obtain time-frequency data includes: Based on short-time Fourier transform and wavelet transform technology, the operating status data is analyzed in time and frequency domain to obtain time and frequency data.

7. The gearbox fault prediction method according to claim 1, characterized in that: The trigger condition is that the time-frequency data undergoes a sudden change or the monitoring time of the time-frequency data is at a periodic detection time node.

8. The gearbox fault prediction method according to claim 1, characterized in that: The fully trained neural network model adopts the LSTM model.

9. The gearbox fault prediction method according to claim 1, characterized in that: The knowledge graph is constructed through the following steps: Obtain historical fault data; Based on the historical fault data, determining nodes, the nodes including key components of the gearbox, component status data, fault types, operating conditions, and historical fault cases; Based on the historical fault data and the nodes, determining relationships between nodes, the relationships including component-fault relationships, fault-sensor signal relationships, operating condition-fault relationships, and historical case-fault relationships; A knowledge graph is constructed based on the nodes and the relationships.

10. A gearbox fault prediction system, characterized in that: include: A data acquisition module is used to acquire the operating status data of the gearbox in real time, and perform time-frequency domain analysis on the operating status data to obtain time-frequency data; A fault time prediction module is used to predict the time-frequency data based on a well-trained neural network model to determine the fault time node when the time-frequency data meets the trigger condition; A fault type prediction module is used to perform fault type inference on the time-frequency data based on the knowledge graph and the inference algorithm to obtain the fault type when the time-frequency data meets the trigger condition; The fault result generating module is used to determine the fault prediction result based on the fault time node and the fault type.

Images