An artificial intelligence-based hydropower station bridge machine fault detection method

Abstract

The application discloses an artificial intelligence-based hydropower station bridge machine fault detection method, and relates to the technical field of fault detection, comprising the following steps: collecting hydropower station bridge machine operation data and performing feature extraction on the hydropower station bridge machine operation data to obtain equipment operation state feature data; performing fault mode recognition according to the equipment operation state feature data to obtain a fault recognition result; performing fault trend analysis according to the fault recognition result to obtain a trend analysis result; performing hydropower station bridge machine health assessment according to the trend analysis result to obtain a health assessment result; and performing maintenance strategy formulation according to the health assessment result to obtain a maintenance strategy. The application can identify the potential development trend of equipment faults in advance, more accurately capture complex fault types, improve the universality and adaptability, make the prediction of the development trend of faults more accurate, improve the adaptability of the model under complex operation conditions, and effectively improve the health management level of the hydropower station bridge machine.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, specifically to a fault detection method for bridge cranes in hydropower stations based on artificial intelligence. Background Technology

[0002] As a crucial piece of equipment in hydropower station operation, bridge cranes undertake critical tasks such as equipment hoisting, inspection, and maintenance. Their stability directly impacts the safe operation of the hydropower station. However, due to prolonged exposure to high loads, high humidity, and complex environments, bridge cranes are highly susceptible to structural fatigue, mechanical wear, and electrical faults. Therefore, efficient and accurate fault detection and health assessment of bridge cranes can effectively improve equipment safety and maintenance efficiency, and reduce economic losses caused by unexpected failures.

[0003] Existing technologies for fault mode recognition have shortcomings: current fault mode recognition methods typically use models based on a single feature signal and rely solely on static features, making it difficult to accurately capture complex fault types, identify potential equipment faults, comprehensively characterize fault features, and affect the accuracy of fault mode recognition.

[0004] Existing technologies for fault trend analysis have shortcomings: existing trend analysis methods usually rely on time series analysis, such as ARIMA and LSTM. These methods mainly rely on historical data for fitting and lack modeling of fault mode transition probabilities, making it difficult to accurately predict fault development trends. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an artificial intelligence-based method for detecting bridge crane faults in hydropower stations, thereby solving the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a method for detecting faults in bridge cranes at hydropower stations based on artificial intelligence, comprising the following steps: S1. Collect the operating data of the bridge crane at the hydropower station and extract features from the operating data of the bridge crane at the hydropower station to obtain equipment operating status feature data; S2. Based on the equipment operating status characteristic data, perform fault mode identification to obtain fault identification results; S3. Perform fault trend analysis based on the fault identification results to obtain the trend analysis results; S4. Conduct a health assessment of the bridge cranes in the hydropower station based on the trend analysis results, and obtain the health assessment results; S5. Based on the health assessment results, a maintenance strategy is formulated to obtain the maintenance strategy.

[0008] In one embodiment of the present invention, the fault mode identification in S2 includes: Based on the acquired equipment operating status characteristic data, a fault mode recognition model is used for analysis and identification to determine whether a fault exists and its type.

[0009] In one embodiment of the present invention, the fault mode recognition model includes: ; in: The fault identification value at time t; The i-th feature data collected by the sensor at time t; : Weight coefficients of the i-th feature data in fault mode recognition; Weighting coefficient for time-series rate of change; : The time-series rate of change of the i-th feature data; Types of feature data.

[0010] In one embodiment of the present invention, the weighting coefficients for fault mode identification include: ; in: :feature Information entropy; :feature Information entropy; : Index of the type of feature data, indicating that it is not of the same type as another feature data.

[0011] In one embodiment of the present invention, the weighting coefficient of the time-series rate of change includes: ; ; in: :feature The rate of change at time t; :feature The rate of change at time t; The characteristic change rate is The probability of failure mode F occurring under the given conditions; The sum of probabilities of failure mode F occurring; Under fault mode F, the characteristic change rate is The probability of; : The prior probability of failure mode F; The characteristic change rate in all samples is The probability of.

[0012] In one embodiment of the present invention, the fault trend analysis in S3 includes: Based on the obtained fault identification results, a fault trend analysis model is used to analyze and predict the fault development trend.

[0013] In one embodiment of the present invention, the fault trend analysis model includes: ; in: The predicted fault trend at time t+Δt; : State transition probability; The rate of change of fault state.

[0014] In one embodiment of the present invention, the state transition probability includes: ; in: Historical data sequence weights; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; The number of historical data sequences; The number of fault states; : Index of fault state types.

[0015] In one embodiment of the present invention, the historical data sequence weights include: ; in: Initial weights; Attenuation factor; : The time interval of historical data sequence k relative to the current moment.

[0016] In one embodiment of the present invention, it further includes: Maintenance of the bridge cranes in the hydropower station is carried out according to the maintenance strategy. The maintenance operations include replacing damaged parts, repairing the motors, and conducting structural inspections on the bridge frame. The maintenance process is monitored in real time to ensure that all faults are dealt with. After the maintenance is completed, the equipment is inspected to confirm that it has been restored to normal working condition.

[0017] The embodiments of this invention have the following beneficial effects: This artificial intelligence-based fault detection method for hydropower station bridge cranes, through a fault mode recognition model, combines data from multiple sensors and considers the rate of change over time, dynamically adjusting weight coefficients. This enables early identification of potential fault development trends, more accurately capturing complex fault types, and providing more information for subsequent fault prediction and maintenance decisions, thus improving versatility and adaptability. Through a fault trend analysis model, combining historical fault states, current state change trends, and environmental factors, a dynamic evolution law of fault modes is established, comprehensively evaluating the transition probability between faults. This makes the prediction of fault development trends more accurate, improves the model's adaptability under complex operating conditions, and effectively enhances the health management level of hydropower station bridge cranes. Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating an artificial intelligence-based fault detection method for bridge cranes in hydropower stations proposed in this invention. Figure 2 This is a flowchart illustrating the fault mode recognition model of a fault detection method for bridge cranes in hydropower stations based on artificial intelligence, as proposed in this invention. Figure 3 This is a flowchart illustrating the fault trend analysis model of a fault detection method for bridge cranes in hydropower stations based on artificial intelligence, as proposed in this invention. Figure 4 This is a flowchart illustrating the health assessment model of a fault detection method for bridge cranes in hydropower stations based on artificial intelligence, as proposed in this invention. Detailed Implementation

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] Reference Figures 1-4 This embodiment provides an artificial intelligence-based method for detecting bridge crane faults in hydropower stations, including the following steps: S1. Collect the operating data of the bridge crane at the hydropower station and extract features from the operating data of the bridge crane at the hydropower station to obtain equipment operating status feature data.

[0022] In this embodiment, the data collection and feature extraction process includes: Real-time operational data of the bridge cranes at the hydropower station is collected by deploying sensors, including vibration sensors, temperature sensors, current sensors, strain sensors, and optical cameras. Each sensor monitors different operating parameters of the bridge crane, such as motor temperature, current fluctuations, bearing vibration, wire rope tension, and bridge frame deformation. The collected data is processed, including noise filtering, data smoothing, and missing value imputation. Fourier transform is used to perform frequency domain analysis on the vibration signals, extracting features such as vibration amplitude and frequency components. For current signals, wavelet transform is used to extract instantaneous energy and spectral features. For other sensor data, such as temperature and strain data, the data are transformed into statistical features closely related to the health status of the equipment, such as mean, variance, and peak value.

[0023] S2. Based on the equipment operating status characteristic data, perform fault mode identification to obtain fault identification results.

[0024] In this embodiment, fault mode identification includes: Based on the acquired equipment operating status characteristic data, a fault mode recognition model is used for analysis and identification to determine whether a fault exists and its type.

[0025] Furthermore, the fault mode recognition model includes: ; in: The fault identification value at time t represents the fault mode identification result of the equipment, indicating the fault state of the equipment. The i-th type of feature data collected by the sensor at time t includes the temporal characteristics of vibration, temperature, tension, and current. : Weight coefficients of the i-th feature data in fault mode recognition; Weighting coefficient for time-series rate of change; : The time-series change rate of the i-th feature data, describing its dynamic changes; Types of feature data.

[0026] This model describes how to identify fault modes using equipment operating status characteristic data.

[0027] Traditional fault mode identification (FMO) models rely on single-feature signals and static features, making them inaccurate in capturing complex fault types and identifying potential equipment failures. This model, however, combines data from multiple sensors to comprehensively assess equipment fault modes, enabling it to capture complex fault types more accurately. Furthermore, it considers time-series change rates, allowing it to identify potential development trends of equipment failures in advance and dynamically adjust weighting coefficients to provide more information for subsequent fault prediction and maintenance decisions. This model exhibits high versatility and adaptability.

[0028] The steps for using this model include: Data Acquisition: Obtain the operating status characteristic data of various devices from step S1. Data such as vibration, temperature, tension, and current are input into the model. Weight calculation: Calculate the weight coefficients for fault mode recognition based on the input data. Weighting coefficients for time-series change rates This ensures that the signals that best reflect the fault have a higher weight in the model, and dynamically captures the changing trends of the equipment status; Fault Mode Recognition: Based on the input data and calculated weight coefficients, fault mode recognition is performed to obtain a fault identification value. This is used to identify the current fault mode of the device.

[0029] Furthermore, the weighting coefficients for the fault mode identification include: ; in: :feature Information entropy ,in, Features Value The probability, Features The number of possible values ​​for the feature, in actual detection, is... The value of is continuously changing. To facilitate the calculation of information entropy, the continuous value interval is divided into several discrete intervals, each interval being called a discrete state. Representation of features In the first Within a discrete interval, For the index of the discrete interval; :feature Information entropy; : Index of the type of feature data, indicating that it is not of the same type as another feature data.

[0030] Furthermore, the weighting coefficients for the time-series rate of change include: ; ; in: :feature The rate of change at time t; :feature The rate of change at time t; The characteristic change rate is The probability of failure mode F occurring at that time; The sum of probabilities of failure mode F occurring; Under fault mode F, the characteristic change rate is The probability is obtained by statistically analyzing the distribution of the rate of change values ​​in historical data. The prior probability of failure mode F is the statistical likelihood of failure mode F occurring within a certain time period. It is calculated by statistically analyzing a large amount of historical operating data or failure records to determine the ratio of the number of times failure mode F occurs to the total number of failures. The ratio obtained is the prior probability. The characteristic change rate in all samples is The probability is obtained by statistically analyzing the distribution of the rate of change values ​​in historical data.

[0031] S3. Perform fault trend analysis based on the fault identification results to obtain the trend analysis results.

[0032] In this embodiment, the fault trend analysis includes: Based on the obtained fault identification results, a fault trend analysis model is used to analyze and predict the fault development trend so that countermeasures can be taken in advance.

[0033] Furthermore, the fault trend analysis model includes: ; in: The predicted fault trend at time t+Δt; State transition probability: represents the probability of a device transitioning from a fault state to a fault state. Evolved to another state The possibility; The rate of change of fault state.

[0034] This model describes how to predict equipment failure trends based on historical failure data and time-series variation characteristics.

[0035] Traditional fault trend analysis typically relies on time series models, and trend analysis often depends on fixed weight parameters. This model method uses state transition probabilities and dynamically adjusts the trend prediction direction based on historical states to improve the accuracy of trend prediction and reduce misjudgments.

[0036] The use of the above model includes: Data input: Obtain the equipment fault identification value through step S2. Understand the current fault status of the equipment; Probability calculation: based on the obtained fault identification values ​​of the equipment. Calculate state transition probability To improve the accuracy of trend prediction; Trend analysis: Based on the input data and the calculated state transition probabilities Calculate fault trend prediction value ,like Compared with the current fault identification value After comparison, A larger value indicates a worsening fault condition, requiring early warning. A smaller value indicates that the fault condition is stable or has improved.

[0037] Furthermore, the state transition probabilities include: ; in: Historical data sequence weights; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; The number of historical data sequences; The number of fault state types; Index of fault status types; Based on the weights of the historical data sequences and the number of transitions between equipment fault states, the denominator is calculated. This denominator is the product of the weight of each historical data sequence and the number of times the equipment fault state transitioned to any other state for that historical data sequence. The total is then calculated. Summing the results of the multiplications yields the denominator. Then, the numerator is calculated by multiplying the weight of each historical data sequence by the number of times the equipment fault state transitioned to a specific state for that historical data sequence. The total is then calculated. The sum of the results of multiplying the products gives the numerator, and the numerator divided by the denominator gives the state transition probability.

[0038] Furthermore, the historical data sequence weights include: ; in: Initial weights; The decay factor, typically between 0 and 1, controls the decay rate. : The time interval of historical data sequence k relative to the current moment.

[0039] S4. Conduct a health assessment of the bridge cranes in the hydropower station based on the trend analysis results, and obtain the health assessment results.

[0040] In this embodiment, the health assessment includes: Based on the obtained trend analysis results, a health assessment model is used to comprehensively analyze the trends of the status and failures of various equipment, conduct a health assessment of the bridge cranes in the hydropower station, and obtain a comprehensive health score for the equipment.

[0041] Furthermore, the health assessment model includes: ; in: The health assessment results of the equipment at time t; Fault status weight, reflecting the current severity of the equipment fault; Environmental risk factors comprehensively consider the impact of external environmental factors (such as temperature, humidity, vibration, etc.) on equipment health; , , Weighting coefficients, learned from historical data, are used to balance the contributions of different factors to health assessment results.

[0042] This model describes how to conduct equipment health assessments based on the equipment's current status and historical trends, combined with multi-dimensional environmental factors.

[0043] Traditional health assessments often only consider the equipment's failure status, focusing on a single dimension. This model, however, combines failure modes, trend predictions, and environmental risk factors to provide a more comprehensive health assessment. It can more accurately evaluate the equipment's health status, identify potential risks in advance, and dynamically adjust weights based on the equipment's failure status and trends, thus improving the model's adaptability.

[0044] The use of this model includes: Data collection: Real-time data of the hydropower station bridge crane is acquired through steps S1, S2, and S3, including sensor signals, fault identification values, and fault trend prediction values. wait; Parameter calculation: Based on the acquired data, calculate the fault state weight at the current moment. and environmental risk factors ; Health assessment: Based on the acquired data and calculated parameters, the data is input into the model for calculation to obtain the health assessment results of the equipment. .

[0045] Furthermore, the fault state weights include: ; in: The weight of each failure mode reflects the degree of impact of that failure mode on the health of the equipment.

[0046] Furthermore, the environmental risk factors include: ; in: : The measured value of the uth environmental factor at time t, including temperature, humidity, vibration, etc.; The number of environmental factors that affect equipment health; The weighting coefficients of environmental factors reflect the degree of impact of each environmental factor on equipment health.

[0047] S5. Based on the health assessment results, a maintenance strategy is formulated to obtain the maintenance strategy.

[0048] In this embodiment, the maintenance strategy formulation includes: After obtaining the health assessment results, the equipment's health score is used to determine whether emergency repairs or preventative maintenance are needed. Maintenance decisions are then made, taking into account various factors, including the importance of the equipment, the time required for repairs, and the equipment's failure modes, to formulate appropriate maintenance strategies and find the best maintenance solution.

[0049] Step S5 is followed by: Maintenance of the bridge cranes in the hydropower station is carried out according to the maintenance strategy. The maintenance operations include replacing damaged parts (including bearings, gears, wire ropes, etc.), repairing motors, and structural inspection of the bridge frame. A computerized maintenance management system (existing technology) is used for operation scheduling and recording, and the maintenance process is monitored in real time to help operation and maintenance personnel effectively manage equipment maintenance tasks, ensure that all faults are handled, and inspect the equipment after maintenance to confirm whether the equipment has been restored to normal working condition.

[0050] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0051] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0052] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for fault detection of bridge cranes in hydropower stations based on artificial intelligence, characterized in that, Includes the following steps: S1. Collect the operating data of the bridge crane at the hydropower station and extract features from the operating data of the bridge crane at the hydropower station to obtain equipment operating status feature data; S2. Based on the equipment operating status characteristic data, perform fault mode identification to obtain fault identification results; S3. Perform fault trend analysis based on the fault identification results to obtain the trend analysis results; S4. Conduct a health assessment of the bridge cranes in the hydropower station based on the trend analysis results, and obtain the health assessment results; S5. Based on the health assessment results, a maintenance strategy is formulated to obtain the maintenance strategy.

2. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The fault mode identification in S2 includes: Based on the acquired equipment operating status characteristic data, a fault mode recognition model is used for analysis and identification to determine whether a fault exists and its type.

3. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 2, characterized in that, The fault mode identification model includes: ； in: The fault identification value at time t; The i-th feature data collected by the sensor at time t; : Weight coefficients of the i-th feature data in fault mode recognition; Weighting coefficient for time-series rate of change; : The time-series rate of change of the i-th feature data; Types of feature data.

4. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 3, characterized in that, The weighting coefficients for the fault mode identification include: ； in: :feature Information entropy; :feature Information entropy; : Index of the type of feature data, indicating that it is not of the same type as another feature data.

5. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 3, characterized in that, The weighting coefficients for the time-series rate of change include: ； ； in: :feature The rate of change at time t; :feature The rate of change at time t; The characteristic change rate is The probability of failure mode F occurring under the given conditions; The sum of probabilities of failure mode F occurring; Under fault mode F, the characteristic change rate is The probability of; : The prior probability of failure mode F; The characteristic change rate in all samples is The probability of.

6. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The obstacle trend analysis in S3 includes: Based on the obtained fault identification results, a fault trend analysis model is used to analyze and predict the fault development trend.

7. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The fault trend analysis model includes: ； in: The predicted fault trend at time t+Δt; : State transition probability; The rate of change of fault state.

8. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The state transition probabilities include: ； in: Historical data sequence weights; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; In the historical data sequence k, the fault state of the equipment ranges from state k to k. Transition to another state The number of times; The number of historical data sequences; The number of fault states; : Index of fault state types.

9. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The weights of the historical data sequence include: ； in: Initial weights; Attenuation factor; : The time interval of historical data sequence k relative to the current moment.

10. The method for fault detection of bridge cranes in hydropower stations based on artificial intelligence according to claim 1, characterized in that, The method also includes: Maintenance of the bridge cranes in the hydropower station is carried out according to the maintenance strategy. The maintenance operations include replacing damaged parts, repairing the motors, and conducting structural inspections on the bridge frame. The maintenance process is monitored in real time to ensure that all faults are dealt with. After the maintenance is completed, the equipment is inspected to confirm that it has been restored to normal working condition.

Images