Fault diagnosis method for helical gear reducer under intelligent monitoring

Abstract

The application provides a fault diagnosis method for helical gear reducers under intelligent monitoring, and relates to the technical field of fault diagnosis. Real-time operation parameters of an inclined stream conveyor are acquired to evaluate the load state of the helical gear reducer. A first monitoring frequency and a first number of monitoring indexes are configured based on the load scale. A first sensing device is activated, and monitoring is performed according to the first monitoring frequency to obtain a first fault risk coefficient. If the first fault risk coefficient exceeds a first fault risk threshold, a sensing monitoring array is activated to obtain a second fault risk coefficient. If the second fault risk coefficient exceeds a second fault risk threshold, an audible-light warning device is activated to issue a warning signal for fault warning. The technical problem that the prior art lacks real-time monitoring and intelligent diagnosis means for helical gear reducers, resulting in low fault diagnosis efficiency and accuracy, is solved. The technical effect of real-time monitoring of the operation state of the helical gear reducer and improvement of the fault diagnosis efficiency and accuracy is achieved.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and specifically to a fault diagnosis method for helical gear reducers under intelligent monitoring. Background Technology

[0002] With the rapid development of the modern warehousing and logistics industry, inclined conveyor belts, as key equipment, play a crucial role in the transportation and sorting of goods. Among them, the helical gear reducer, as the core drive component of the inclined conveyor belt, directly affects the conveyor belt's efficiency and operational stability. However, in current technology, the maintenance of helical gear reducers mainly relies on regular manual inspections or post-failure repairs, lacking real-time monitoring and intelligent diagnostic methods for their operational status. This passive maintenance approach not only fails to meet the real-time requirements of efficient operation in logistics systems but also makes it difficult to detect potential faults in a timely and accurate manner, seriously affecting the overall efficiency of the logistics system.

[0003] Existing technologies suffer from the lack of real-time monitoring and intelligent diagnostic methods for helical gear reducers, resulting in low efficiency and accuracy in fault diagnosis. Summary of the Invention

[0004] This application provides a fault diagnosis method for helical gear reducers under intelligent monitoring, which solves the technical problem that the lack of real-time monitoring and intelligent diagnosis means for helical gear reducers in the prior art leads to low efficiency and accuracy of fault diagnosis.

[0005] In view of the above problems, this application provides a method for fault diagnosis of helical gear reducers under intelligent monitoring. The method includes: acquiring real-time operating parameters of an inclined logistics conveyor belt through fixed-point monitoring; evaluating the load status of the helical gear reducer based on the real-time operating parameters to determine the load scale, wherein the helical gear reducer is a drive component of the inclined logistics conveyor belt; configuring a first monitoring frequency and a first number of monitoring indicators based on the load scale; filtering from a predetermined monitoring indicator library based on the first number of monitoring indicators to determine the first monitoring indicators; activating a first sensing device based on the first monitoring indicator; monitoring the helical gear reducer of the inclined logistics conveyor belt at the first monitoring frequency; and obtaining a first fault risk coefficient based on the analysis of the first monitoring data; if the first fault risk coefficient exceeds a first fault risk threshold, activating a sensor monitoring array; comprehensively monitoring the helical gear reducer at a second monitoring frequency; and obtaining a second fault risk coefficient based on the analysis of the second monitoring data; if the second fault risk coefficient exceeds a second fault risk threshold, activating an audible and visual early warning device to issue an early warning signal for a fault warning of the helical gear reducer of the inclined logistics conveyor belt, wherein the second fault risk threshold is greater than the first fault risk threshold.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The method provided in this application acquires real-time operating parameters of an inclined logistics conveyor belt through fixed-point monitoring, evaluates the load status of a helical gear reducer based on these parameters, and determines the load scale, wherein the helical gear reducer is the drive component of the inclined logistics conveyor belt; configures a first monitoring frequency and a first number of monitoring indicators based on the load scale, and selects the first monitoring indicators from a predetermined monitoring indicator library based on the first number of monitoring indicators; activates a first sensing device based on the first monitoring indicator, and monitors the helical gear reducer of the inclined logistics conveyor belt according to the first monitoring frequency, and obtains a first fault risk coefficient based on the first monitoring data analysis; if the first fault risk coefficient exceeds a first fault risk threshold, activates a sensor monitoring array, and performs comprehensive monitoring of the helical gear reducer according to a second monitoring frequency, and obtains a second fault risk coefficient based on the second monitoring data analysis; if the second fault risk coefficient exceeds a second fault risk threshold, activates an audible and visual early warning device to issue an early warning signal, providing a fault warning for the helical gear reducer of the inclined logistics conveyor belt, wherein the second fault risk threshold is greater than the first fault risk threshold. This achieves the technical effect of real-time monitoring of the helical gear reducer's operating status and improving the efficiency and accuracy of fault diagnosis. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A flowchart illustrating the fault diagnosis method for helical gear reducers under intelligent monitoring provided in this application;

[0010] Figure 2 A flowchart illustrating the process of determining the first monitoring index in the intelligent monitoring-based helical gear reducer fault diagnosis method provided in this application. Detailed Implementation

[0011] This application provides a fault diagnosis method for helical gear reducers under intelligent monitoring, which addresses the technical problem of low fault diagnosis efficiency and accuracy in existing technologies due to the lack of real-time monitoring and intelligent diagnostic methods for helical gear reducers. It achieves the technical effect of real-time monitoring of the helical gear reducer's operating status, thereby improving the efficiency and accuracy of fault diagnosis.

[0012] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0013] like Figure 1 As shown, this application provides a fault diagnosis method for helical gear reducers under intelligent monitoring, the method comprising:

[0014] The real-time operating parameters of the inclined logistics conveyor belt are obtained by fixed-point monitoring. The load status of the helical gear reducer is evaluated based on the real-time operating parameters to determine the load scale. The helical gear reducer is the driving component of the inclined logistics conveyor belt.

[0015] Specifically, fixed-point monitoring devices, such as high-definition cameras, are deployed at key locations on the inclined conveyor belt in warehousing logistics, such as near the drive rollers and load-bearing areas. The placement of these monitoring devices is optimized through experiments to ensure comprehensive coverage of the conveyor belt's transport area, avoiding blind spots. These fixed-point monitoring devices continuously acquire real-time operating parameters of the inclined conveyor belt, including cargo load overflow, cargo load uniformity, and cargo transport uniformity. Based on these real-time operating parameters, the load status of the helical gear reducer is evaluated to determine the current load magnitude, which reflects the size and nature of the helical gear reducer's working load. As the driving component of the inclined logistics conveyor belt, the performance of the gear reducer directly affects the belt's stability and efficiency. Real-time load status evaluation allows for timely detection of abnormal helical gear reducer loads, warning of potential failure risks, and prevention of equipment damage and conveyor interruptions.

[0016] Furthermore, the system acquires real-time operating parameters of the inclined logistics conveyor belt through fixed-point monitoring, and evaluates the load status of the helical gear reducer based on these parameters to determine the load scale. This includes: acquiring monitoring videos of the inclined logistics conveyor belt at fixed points; analyzing the monitoring videos to obtain real-time operating parameters, including cargo load overflow, cargo load uniformity, and cargo conveying uniformity; evaluating the load status of the helical gear reducer based on these parameters to generate a load scale, wherein the load scale is positively correlated with cargo load overflow and negatively correlated with cargo load uniformity and cargo conveying uniformity.

[0017] Specifically, high-performance camera devices, such as high-definition cameras, are deployed at key locations on the inclined conveyor belt to ensure panoramic capture of the goods transport process. These cameras utilize high-definition resolution and motion capture technology to record the dynamic position, distribution, and transport speed of the goods in real time. Fixed-point monitoring collects data from fixed locations, ensuring the continuity and comparability of the video data. Monitoring videos of the inclined logistics conveyor belt are acquired through fixed-point monitoring devices. Then, based on the video frames, the boundary overflow of goods on the inclined logistics conveyor belt is analyzed, as well as the uniformity of the goods load distribution on the conveyor belt. Furthermore, the uniformity of the goods conveying speed is analyzed through trajectory tracking and motion analysis to obtain real-time operating parameters. These real-time operating parameters include goods load overflow degree, goods load uniformity, and goods conveying uniformity. Goods load overflow degree refers to the boundary overflow of goods on the conveyor belt, used to measure the degree of overload risk and reflect the potential load pressure exerted by the goods on the helical gear reducer. Goods load uniformity describes the uniformity of goods distribution on the conveyor belt; a lower uniformity indicates localized load concentration and a higher risk of equipment damage. Goods conveying uniformity refers to the uniformity of goods conveying speed; a low uniformity indicates poor operation or abnormal mechanical condition of the conveyor belt. Then, based on the goods load overflow degree, goods load uniformity, and goods conveying uniformity, the load status of the helical gear reducer is comprehensively evaluated, and a load scale is generated. The load scale is positively correlated with the goods load overflow degree; that is, the greater the goods load overflow degree, the larger the load scale of the reducer. The load distribution is negatively correlated with the uniformity of cargo load and the uniformity of cargo conveying, indicating that the more uniform the load distribution and the more stable the conveying speed, the smaller the load scale of the reducer. Optionally, the load scale can be generated through a load state evaluation formula, as follows: Load scale = W1·overflow degree + W2·(1-load uniformity) + W3·(1-conveyor uniformity), where W1, W2, and W3 are weighting coefficients, preset according to actual needs, expert experience, and historical data. Comprehensive load state evaluation using real-time operating parameters obtained through fixed-point monitoring can accurately reflect the real-time load pressure of the helical gear reducer, improve fault prediction accuracy, and ensure the stable operation of the logistics conveyor belt.

[0018] Based on the load scale, configure the first monitoring frequency and the number of first monitoring indicators, and select the first monitoring indicators from the predetermined monitoring indicator library according to the number of first monitoring indicators.

[0019] Furthermore, configuring a first monitoring frequency and a first number of monitoring indicators based on the load scale includes: acquiring the component attributes of the helical gear reducer, wherein the component attributes include gear width, gear helix angle, gear module, and number of gear teeth; evaluating and obtaining a gear robustness coefficient based on the gear width, gear helix angle, gear module, and number of gear teeth; multiplying the ratio of the gear robustness coefficient to the standard gear robustness coefficient by the load scale to obtain a corrected load scale; setting the ratio of the corrected load scale to the historical maximum load scale as the real-time monitoring scale; multiplying the real-time monitoring scale by a second monitoring frequency and rounding it to obtain a first monitoring frequency; and multiplying the real-time monitoring scale by the number of monitoring indicators in a predetermined monitoring indicator library and rounding it to obtain a first number of monitoring indicators.

[0020] Specifically, the component properties of a helical gear reducer directly determine its load-bearing capacity under different load conditions. These component properties include gear width, gear helix angle, gear module, and number of teeth. Gear width represents the lateral dimension of the gear; a larger width indicates stronger load-bearing capacity. Gear helix angle describes the gear meshing angle, affecting transmission smoothness and load-bearing capacity. Gear module is a baseline parameter for gear dimensions, reflecting the gear's structural strength; more teeth result in higher transmission accuracy and stronger load-bearing capacity. Furthermore, based on these component properties, a gear robustness coefficient is calculated to quantify the gear's strength under load. A higher robustness coefficient indicates better compressive and shear resistance under load conditions. The formula for calculating the gear robustness coefficient is: Ks = k1·W + k2·tan(α) + k3·m·z, where Ks is the gear robustness coefficient, k1, k2, and k3 are weighting coefficients reflecting the contribution of each parameter to the gear's load-bearing capacity, W is the gear width, α is the gear helix angle, m is the gear module, and z is the number of teeth. Then, the ratio of the gear robustness coefficient to the standard gear robustness coefficient is multiplied by the load scale to obtain the corrected load scale. The corrected load scale is the value obtained after adjusting the load scale, taking into account the difference between the actual load-bearing capacity of the helical gear reducer and the industry standard reference value. The real-time monitoring scale is determined by the ratio of the corrected load scale to the historical maximum load scale. The real-time monitoring scale is used to dynamically adjust the monitoring frequency. The formula is: Real-time monitoring scale = Corrected load scale / Historical maximum load scale. The real-time monitoring scale reflects the proportion of the current load relative to the equipment's historical operating limit. The larger the value, the closer the current load is to the equipment's maximum capacity, requiring a higher monitoring frequency and more monitoring indicators. Further, the real-time monitoring scale is multiplied by the second monitoring frequency and rounded to obtain the first monitoring frequency. The second monitoring frequency is the preset maximum monitoring frequency, which can be set as the maximum monitoring frequency required by the equipment under high load conditions. The first monitoring frequency refers to the equipment monitoring frequency dynamically calculated based on the real-time monitoring scale, used to adjust the time interval or period for the monitoring device to collect data from the helical gear reducer. Simultaneously, the real-time monitoring scale is multiplied by the number of monitoring indicators in the predetermined monitoring indicator library and rounded to obtain the first number of monitoring indicators. This predetermined monitoring indicator library contains multiple monitoring parameters used to comprehensively assess the reducer's condition, including speed monitoring (input / output speed), temperature, noise, vibration, etc. By dynamically adjusting the number of monitoring indicators based on the real-time monitoring scale, it ensures that monitoring indicators are increased during high loads and decreased during low loads, optimizing the allocation of monitoring resources. By dynamically configuring the first monitoring frequency and the first number of monitoring indicators according to the load scale, a balance between resource utilization efficiency and monitoring coverage is improved, achieving intelligent monitoring of the helical gear reducer.

[0021] In one embodiment, such as Figure 2As shown, the process of selecting a first monitoring indicator from a predetermined monitoring indicator database based on the number of the first monitoring indicators includes: acquiring several monitoring indicators from the predetermined monitoring indicator database, wherein the monitoring indicators include at least rotational speed, temperature, noise, and vibration; combining the regional environmental characteristics of the inclined logistics conveyor belt, analyzing the monitoring difficulty and environmental interference of the several monitoring indicators respectively, and determining several monitoring simplicity coefficients and several monitoring accuracies; calculating several indicator fit degrees based on the weighted calculation of the several monitoring simplicity coefficients and several monitoring accuracies, and sorting the several monitoring indicators according to the indicator fit degrees from large to small to generate a monitoring indicator sequence; selecting the monitoring indicator with the highest number of the first monitoring indicators from the monitoring indicator sequence and setting it as the first monitoring indicator.

[0022] Specifically, the pre-defined monitoring index database is a pre-set database containing various monitoring indicators affecting the operating status of helical gear reducers. Multiple monitoring indicators are obtained from this database, including at least speed, temperature, noise, and vibration. Speed ​​is used to assess the smoothness of the input / output shaft operation; temperature reflects the heat loss of gears and bearings; noise monitors the meshing state and mechanical friction; and vibration reflects the stability and abnormal characteristics of the equipment operation. Different environmental characteristics directly affect the ease and accuracy of monitoring. For example, if the noise level in the factory production area where the conveyor belt is located is high, the noise monitoring indicator will be significantly affected, thus impacting its accuracy. A comprehensive environmental identification of the area where the inclined logistics conveyor belt is located is necessary to obtain the specific environmental characteristics of the area, such as temperature, humidity, noise level, and vibration level. Then, for each monitoring indicator, such as rotational speed, temperature, noise, and vibration, the monitoring difficulty in the current environment is assessed based on the adaptability of the monitoring equipment and the complexity of data acquisition. Multiple monitoring ease coefficients are assigned; the higher the ease coefficient, the lower the difficulty. Monitoring ease coefficient = 1 - difficulty score. For example, noise signals require noise reduction processing, increasing complexity, while rotational speed monitoring is directly read from a standard sensor, resulting in low difficulty. Simultaneously, the interference of the environment on monitoring accuracy is analyzed, such as the direct impact of noise on noise monitoring, obtaining several monitoring accuracy levels. Greater environmental interference results in lower monitoring accuracy, quantified using the following formula: Monitoring accuracy = 1 - interference score. Finally, several indicator fit degrees are calculated by weighting the several monitoring ease coefficients and several monitoring accuracy levels. Indicator fit degree = W4 * monitoring ease coefficient + W5 * monitoring accuracy, where W4 and W5 are weights adjusted according to the importance of the indicator. After calculating the suitability of multiple monitoring indicators, a monitoring indicator sequence is generated by sorting them from highest to lowest suitability. Indicators with high suitability are selected first. This monitoring indicator sequence is based on the applicability and accuracy of the monitoring indicators in the current environment, ensuring the effectiveness and reliability of the selected indicators. Finally, the first N indicators (the number of the first monitoring indicators) are selected from the monitoring indicator sequence and designated as the first monitoring indicators. By selecting the first N indicators from the monitoring indicator sequence, where N is the number of the first monitoring indicators, it is ensured that the selected indicators meet the monitoring requirements and can achieve effective monitoring in the current environment. By comprehensively considering the ease of use of the monitoring indicators and the characteristics of environmental interference, the most suitable combination of indicators for the current environment and monitoring requirements is selected, thereby ensuring the effectiveness and accuracy of the monitoring work and providing accurate and reliable data support for the monitoring of helical gear reducers on inclined logistics conveyor belts.

[0023] The first sensing device is activated according to the first monitoring index, and the helical gear reducer of the inclined logistics conveyor belt is monitored according to the first monitoring frequency. The first fault risk coefficient is obtained based on the analysis of the first monitoring data.

[0024] Specifically, a first sensing device is activated based on a determined first monitoring indicator. This first sensing device corresponds to the first monitoring indicator and is used to collect real-time data on the reducer's operating status. For example, for speed, a speed sensor installed on the input / output shaft is activated; for temperature, a temperature sensor installed in the gearbox or bearing is activated; for vibration, an acceleration sensor installed on the helical gear reducer housing is activated; and for noise, a sound level meter near the helical gear reducer is activated. After dynamically activating the corresponding sensing device based on the selected first monitoring indicator, the sensing device monitors the helical gear reducer on the inclined conveyor belt at a determined first monitoring frequency, obtains first monitoring data, analyzes the first monitoring data, calculates the probability of failure under the current operating state, and outputs a first failure risk coefficient. This first failure risk coefficient represents the probability of a failure occurring under the current operating state; the higher the coefficient value, the closer the equipment's operating state is to a failure state. When the first failure risk coefficient is less than a first failure risk threshold, it indicates that the equipment is operating normally, and the existing monitoring frequency is maintained. The first failure risk threshold is a pre-set risk assessment critical value determined based on the safe operation requirements of the helical gear reducer and historical failure experience. If the threshold is exceeded, a high-intensity monitoring mode is activated, further enabling the sensor monitoring array for comprehensive diagnosis. By dynamically activating the corresponding sensors and adjusting the monitoring frequency and equipment as needed, resource utilization efficiency is improved, over- or under-monitoring is avoided, and fault diagnosis efficiency and accuracy are enhanced.

[0025] If the first fault risk coefficient exceeds the first fault risk threshold, the sensor monitoring array is activated, and the helical gear reducer is fully monitored according to the second monitoring frequency. The second fault risk coefficient is obtained based on the analysis of the second monitoring data.

[0026] Specifically, when the first fault risk coefficient is greater than or equal to the first fault risk threshold, it indicates an abnormality in the helical gear reducer. At this point, the monitoring mode is upgraded, the sensor monitoring array is activated, and data is collected according to the second monitoring frequency. The second fault risk coefficient is then calculated through comprehensive analysis. The sensor monitoring array contains various types of sensors, each corresponding to a specific monitoring indicator. The array covers all key parameters of the helical gear reducer's operation, providing comprehensive status information. The array is dynamically activated based on trigger conditions, switching from a mode that only monitors a subset of indicators to full-coverage monitoring. For example, in addition to speed, temperature, and vibration, noise and pressure are added to ensure data comprehensiveness. The second monitoring frequency is a predetermined maximum monitoring frequency, capable of capturing subtle anomalies within a short period. For example, if the first monitoring frequency is 5 times per second, the second monitoring frequency is increased to 10 times per second. Each sensor in the sensor monitoring array synchronously collects real-time data according to the second monitoring frequency, forming a multi-dimensional equipment operating status dataset, i.e., the second monitoring data. This second monitoring data is then analyzed to obtain the second fault risk coefficient. Preliminary monitoring using the first sensing device and the first monitoring frequency results in fewer monitoring indicators, high adaptability, and high analysis efficiency. Real-time monitoring can promptly detect anomalies, and upgrades can be performed when anomalies are detected to achieve comprehensive fault analysis, thereby improving the efficiency and accuracy of fault diagnosis for helical gear reducers.

[0027] If the second fault risk coefficient exceeds the second fault risk threshold, the audible and visual warning device is activated to issue a warning signal to provide a fault warning for the helical gear reducer of the inclined logistics conveyor belt, wherein the second fault risk threshold is greater than the first fault risk threshold.

[0028] Specifically, the second fault risk coefficient is a fault probability value calculated based on comprehensive monitoring data, representing the degree of fault risk in the current state of the equipment. The second fault risk threshold is a preset, higher-level fault risk judgment standard; if it is greater than the first fault risk threshold, it indicates that the equipment is approaching or has reached a dangerous operating state. When the first fault risk coefficient is less than or equal to the second fault risk threshold, continuous monitoring is maintained. Based on the current monitoring data, it is compared with monitoring data over a period of time to assess the changing trend of fault risk. If the risk coefficient remains stable or decreases, it indicates that the equipment status is approaching normal, and the monitoring frequency can be appropriately reduced. When the first fault risk coefficient is greater than or equal to the second fault risk threshold, it is determined that the helical gear reducer has a high fault risk, and the audible and visual warning device is immediately activated to issue a warning signal, providing a fault warning for the helical gear reducer on the inclined logistics conveyor belt. The audible and visual warning device consists of a high-decibel alarm and a high-brightness LED light. The sound signal is used to quickly attract the attention of on-site operators, while the light signal facilitates visual perception at long distances and in complex environments. Upon activation of the audible and visual early warning device, an early warning report is generated based on the current secondary fault risk coefficient and relevant operating parameters. This report is then sent to the equipment maintenance team and management personnel via various methods, such as SMS, email, or mobile app, ensuring timely response. The early warning report includes risk details, specifically the secondary fault risk coefficient and the exact values ​​of abnormal indicators. Through tiered threshold fault early warning systems, it is possible to ensure that no fault risk is overlooked while conserving resources, providing crucial support for the intelligent monitoring and safety management of helical gear reducers.

[0029] Furthermore, monitoring the helical gear reducer of the inclined logistics conveyor belt according to the first monitoring frequency, and analyzing the first monitoring data to obtain a first fault risk coefficient, includes: monitoring the helical gear reducer of the inclined logistics conveyor belt according to the first monitoring frequency to obtain first monitoring data; calling the first fault diagnosis plugin of the first monitoring index to predict the fault probability of the first monitoring data, outputting a first predicted fault probability, which is set as the first fault risk coefficient, wherein the first fault diagnosis plugin is constructed based on random forest and trained to convergence using sample first monitoring data and sample first fault probability.

[0030] Specifically, the first sensing device monitors and collects data on the helical gear reducer of the inclined logistics conveyor belt at a first monitoring frequency, obtaining multiple state index data of the helical gear reducer to form the first monitoring data. Based on machine learning technology, a random forest is used to construct a first fault diagnosis plugin, which is used to predict the fault probability of the first monitoring data. Random forest is an ensemble learning algorithm that predicts by constructing multiple decision trees and integrating the judgment results of multi-dimensional data, and has high accuracy and stability. The construction process of the first fault diagnosis plugin is as follows: a large amount of sample data is obtained, including sample first monitoring data and sample point fault probabilities. Sample first monitoring data includes monitoring data such as speed, temperature, and vibration under different load conditions. The sample first fault probability is obtained by statistically analyzing the proportion of fault occurrences in the sample monitoring data within a predetermined historical time interval, resulting in a fault probability label. Through supervised learning of the sample data, the decision tree parameters are optimized until the model converges, forming the first fault diagnosis plugin, which has efficient and accurate fault probability prediction capabilities. The first fault diagnosis plugin for the first monitoring indicator is invoked. Real-time collected monitoring data is input into the plugin, and the trained model is used to calculate the fault probability of the current operating state. The output is a first predicted fault probability, which is then used as a quantitative indicator of the current fault risk of the helical gear reducer: the first fault risk coefficient. The coefficient ranges from 0 to 1; a higher coefficient indicates a higher fault risk in the current operating state of the helical gear reducer. By invoking the first fault diagnosis plugin for the first monitoring indicator to predict the fault probability of the monitoring data and outputting the predicted fault probability, the accuracy and efficiency of fault diagnosis can be improved.

[0031] Furthermore, if the first fault risk coefficient exceeds the first fault risk threshold, the sensor monitoring array is activated, and the helical gear reducer is comprehensively monitored according to the second monitoring frequency. Based on the analysis of the second monitoring data, the second fault risk coefficient is obtained, including: if the first fault risk coefficient exceeds the first fault risk threshold, activating the sensor monitoring array, wherein the sensor monitoring array includes several sensing devices, and the sensing devices correspond to monitoring indicators; comprehensively monitoring the helical gear reducer according to the second monitoring frequency through the sensor monitoring array, and outputting the second monitoring data; pre-constructing a comprehensive fault diagnosis channel, inputting the second monitoring data into the comprehensive fault diagnosis channel for prediction, and outputting the second fault risk coefficient.

[0032] Furthermore, a comprehensive fault diagnosis channel is pre-constructed, including: collecting sample monitoring datasets based on the operation logs of similar helical gear reducers using inclined logistics conveyor belts, constrained by several monitoring indicators, and statistically analyzing the fault proportion of different sample monitoring data within a predetermined historical time zone, setting it as the sample fault risk coefficient, thus obtaining a sample fault risk coefficient set; using the sample monitoring dataset and the sample fault risk coefficient set as training data, and dividing them into K equal parts to obtain K training sets, where K is an integer greater than 9; using the K training sets to perform supervised training on a random forest until convergence, obtaining K fully trained comprehensive fault diagnosis plugins, which are then combined to obtain the comprehensive fault diagnosis channel.

[0033] Specifically, when the first fault risk coefficient is greater than or equal to the preset first fault risk threshold, it indicates that the helical gear reducer may have a fault risk. At this time, the sensor monitoring array is activated. The sensor monitoring array contains multiple sensors corresponding to different monitoring indicators, covering the key operating parameters of the helical gear reducer. The sensor monitoring array includes several sensors, and the sensors correspond to the monitoring indicators. Through the sensor monitoring array, multiple monitoring data are collected synchronously according to the second monitoring frequency to form comprehensive second monitoring data.

[0034] The system retrieves the operating logs of similar helical gear reducers on an inclined conveyor belt. These logs record operational data under various conditions, including monitoring system data, equipment maintenance logs, and fault data. Several monitoring indicators that directly or indirectly reflect the operating status of the helical gear reducers are selected as constraints, such as speed, temperature, vibration, and noise. A sample monitoring dataset is collected, containing operating records of the helical gear reducers under different time periods and conditions, including speed fluctuation range, temperature trend, vibration spectrum characteristics, and noise levels. Within a predetermined historical time zone, such as the past year, the probability of fault occurrence for each sample monitoring data point is statistically analyzed to obtain a sample fault risk coefficient. This coefficient reflects the probability or frequency of fault occurrence in each sample data point. The sample fault risk coefficient is calculated as: (Number of faults in the corresponding time period / Total number of observations). For example, if a sample experiences 10 faults in 100 operating records, its fault risk coefficient is 0.1. The fault risk coefficients for all sample monitoring data are calculated to form a sample fault risk coefficient set. The sample monitoring dataset and sample failure risk coefficient set are used as training data. The sample order of the sample monitoring dataset and sample failure risk coefficient set is randomly shuffled, and the data is divided into K equal parts, where K is an integer greater than 9. A random forest algorithm is used for supervised training on the K training sets. The K training sets are input into the random forest model, including the input feature sample monitoring data and the target value sample failure risk coefficient. In each training iteration, multiple decision trees are constructed through random sampling and feature selection of the sample data. By adjusting the number, depth, and splitting strategy of the decision trees, the convergence and accuracy of the random forest are ensured. Hyperparameters are optimized until the model converges, i.e., the model performance reaches a stable state. In each of the K training iterations, a high-precision random forest model is obtained, resulting in K fully trained comprehensive fault diagnosis plugins. All K trained plugins are combined to form a comprehensive fault diagnosis channel. This comprehensive fault diagnosis channel covers multi-dimensional indicators and data characteristics, enabling analysis of the operating status of the helical gear reducer of the inclined logistics conveyor under different loads and environmental conditions, improving the efficiency and accuracy of fault diagnosis. After obtaining the comprehensive fault diagnosis channel, the second monitoring data is input into the comprehensive fault diagnosis channel for prediction to obtain a second fault risk coefficient. The second fault risk coefficient is calculated based on comprehensive data from the comprehensive fault diagnosis channel and can accurately reflect the current comprehensive fault risk of the equipment. Through the ensemble learning capability of random forest, the comprehensive fault diagnosis channel can extract key features from complex multidimensional monitoring data and accurately predict the fault risk of the helical gear reducer.

[0035] Furthermore, the second monitoring data is input into the comprehensive fault diagnosis channel for prediction, and the second fault risk coefficient is output. This includes: calculating the deviation between the first fault risk coefficient and the first fault risk threshold, and setting it as the first deviation coefficient; multiplying the ratio of the first deviation coefficient to the maximum deviation coefficient by K and rounding it down to obtain Q, where the maximum deviation coefficient is the deviation between the second fault risk threshold and the first fault risk threshold; randomly selecting Q comprehensive fault diagnosis plugins from the K comprehensive fault diagnosis plugins in the comprehensive fault diagnosis channel, inputting the second monitoring data into the Q comprehensive fault diagnosis plugins, outputting Q fault risk coefficients, and calculating the average to obtain the second fault risk coefficient.

[0036] Specifically, the deviation between the first fault risk coefficient and the first fault risk threshold is calculated to obtain the first deviation coefficient, which reflects the degree to which the first fault risk coefficient exceeds the first threshold. The first deviation coefficient = first fault risk coefficient - first fault risk threshold. The deviation between the second fault risk threshold and the first fault risk threshold is calculated to obtain the maximum deviation coefficient, which = second fault risk threshold - first fault risk threshold. Then, the ratio of the first deviation coefficient to the maximum deviation coefficient is multiplied by K and rounded to obtain the number of diagnostic plugins selected, Q, i.e., Q = (first deviation coefficient / maximum deviation coefficient) × K (rounded down), where K is the total number of diagnostic plugins in the comprehensive fault diagnosis channel. The larger the deviation, the more plugins are selected. The comprehensive fault diagnosis channel contains K comprehensive fault diagnosis plugins, each based on a random forest model, capable of independently analyzing the second monitoring data. Q comprehensive fault diagnosis plugins are randomly selected from the K plugins in the comprehensive fault diagnosis channel for the current analysis. Random selection improves the robustness of the diagnosis and avoids the overfitting risk that may arise from fixed plugin selection. The second monitoring data is input into Q selected comprehensive fault diagnosis plugins. Each plugin outputs a fault risk coefficient, resulting in Q fault risk coefficients. The average of these Q coefficients is then calculated to obtain the second fault risk coefficient. This second fault risk coefficient represents the comprehensive diagnostic result, accurately and reliably reflecting the probability of failure of the helical gear reducer under the current condition. The number of diagnostic plugins is adjusted using a first deviation coefficient to ensure analysis accuracy under high-risk conditions while avoiding resource waste under low-risk conditions. Through multi-plugin comprehensive analysis, the characteristics of different models are utilized to reduce the error of a single model, improving the reliability and accuracy of helical gear reducer fault diagnosis, shortening downtime, and ultimately enhancing the overall efficiency of the logistics system.

[0037] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0038] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A fault diagnosis method for helical gear reducers under intelligent monitoring, characterized in that, The methods include: The real-time operating parameters of the inclined logistics conveyor belt are obtained by fixed-point monitoring. The load status of the helical gear reducer is evaluated based on the real-time operating parameters to determine the load scale. The helical gear reducer is the driving component of the inclined logistics conveyor belt. Based on the load scale, configure the first monitoring frequency and the number of first monitoring indicators, and filter from a predetermined monitoring indicator library according to the number of first monitoring indicators to determine the first monitoring indicators; The first sensing device is activated according to the first monitoring index, and the helical gear reducer of the inclined logistics conveyor belt is monitored according to the first monitoring frequency. The first fault risk coefficient is obtained based on the analysis of the first monitoring data. If the first fault risk coefficient exceeds the first fault risk threshold, the sensor monitoring array is activated, and the helical gear reducer is fully monitored according to the second monitoring frequency. The second fault risk coefficient is obtained based on the analysis of the second monitoring data. If the second fault risk coefficient exceeds the second fault risk threshold, the audible and visual warning device is activated to issue a warning signal to provide a fault warning for the helical gear reducer of the inclined logistics conveyor belt, wherein the second fault risk threshold is greater than the first fault risk threshold. If the first fault risk coefficient exceeds the first fault risk threshold, the sensor monitoring array is activated to perform comprehensive monitoring of the helical gear reducer according to the second monitoring frequency. Based on the analysis of the second monitoring data, the second fault risk coefficient is obtained, including: If the first fault risk coefficient exceeds the first fault risk threshold, the sensor monitoring array is activated, wherein the sensor monitoring array includes several sensing devices, and the sensing devices and monitoring indicators correspond to each other. The helical gear reducer is comprehensively monitored according to the second monitoring frequency through the sensor monitoring array, and the second monitoring data is output. A comprehensive fault diagnosis channel is pre-constructed, the second monitoring data is input into the comprehensive fault diagnosis channel for prediction, and the second fault risk coefficient is output. A comprehensive fault diagnosis channel is pre-built, including: Based on the operation logs of similar helical gear reducers using inclined logistics conveyor belts, and constrained by several monitoring indicators, sample monitoring datasets are collected. The proportion of failures in different sample monitoring data within a predetermined historical time zone is statistically analyzed and set as the sample failure risk coefficient, thus obtaining the sample failure risk coefficient set. The sample monitoring dataset and sample failure risk coefficient set are used as training data and divided into K equal parts to obtain K training sets, where K is an integer greater than 9; The random forest is trained under supervision using the K training sets until convergence, resulting in K fully trained comprehensive fault diagnosis plugins, which are then combined to obtain the comprehensive fault diagnosis channel. Real-time operating parameters of the inclined logistics conveyor belt are obtained through fixed-point monitoring. Based on these real-time operating parameters, the load condition of the helical gear reducer is evaluated to determine the load scale, including: The monitoring video of the inclined logistics conveyor belt is acquired at fixed points, and real-time operating parameters are obtained by analyzing the monitoring video. The real-time operating parameters include cargo load overflow, cargo load uniformity, and cargo conveying uniformity. The load state of the helical gear reducer is evaluated according to the cargo load overflow degree, the cargo load uniformity and the cargo conveying uniformity, and a load scale is generated, wherein the load scale is positively correlated with the cargo load overflow degree, and is negatively correlated with the cargo load uniformity and the cargo conveying uniformity; The first monitoring frequency and the first monitoring index quantity are configured based on the load scale, including: A component attribute of the helical gear reducer is acquired, wherein the component attribute includes a gear width, a gear inclination, a gear modulus and a gear tooth number; A gear solid coefficient is evaluated based on the gear width, the gear inclination, the gear modulus and the gear tooth number, and a ratio of the gear solid coefficient to a standard gear solid coefficient is multiplied by the load scale to obtain a corrected load scale; A ratio of the corrected load scale to a historical maximum load scale is set as a real-time monitoring scale, and the real-time monitoring scale is multiplied by a second monitoring frequency to obtain the first monitoring frequency after rounding; The real-time monitoring scale is multiplied by a monitoring index quantity of a predetermined monitoring index library to obtain the first monitoring index quantity after rounding.

2. The method of claim 1, wherein, The first monitoring index is determined by screening in the predetermined monitoring index library according to the first monitoring index quantity, including: A plurality of monitoring indexes of the predetermined monitoring index library are acquired, wherein the monitoring indexes at least include a rotating speed, a temperature, a noise and a vibration; In combination with regional environmental features of the inclined logistics conveying belt, monitoring difficulty and environmental interference of the plurality of monitoring indexes are analyzed respectively to determine a plurality of monitoring simplicity coefficients and a plurality of monitoring accuracies; A plurality of index adaptation degrees are calculated by weighted calculation based on the plurality of monitoring simplicity coefficients and the plurality of monitoring accuracies, and the plurality of monitoring indexes are sorted in descending order of the index adaptation degrees to generate a monitoring index sequence; Monitoring indexes of the first monitoring index quantity in the monitoring index sequence are selected as the first monitoring indexes.

3. The method of claim 2, wherein, The helical gear reducer of the inclined logistics conveying belt is monitored at the first monitoring frequency, and a first fault risk coefficient is obtained by analyzing first monitoring data, including: The helical gear reducer of the inclined logistics conveying belt is monitored at the first monitoring frequency to obtain first monitoring data; A first fault diagnosis plug-in of the first monitoring index is called to predict a fault probability of the first monitoring data, and a first predicted fault probability is output as the first fault risk coefficient, wherein the first fault diagnosis plug-in is constructed based on a random forest and is trained to convergence by using sample first monitoring data and sample first fault probability.

4. The method of claim 1, wherein, The second monitoring data are input into the comprehensive fault diagnosis channel for prediction, and the second fault risk coefficient is output, including: A deviation value of the first fault risk coefficient and the first fault risk threshold value is calculated as a first deviation coefficient; A ratio of the first deviation coefficient to a maximum deviation coefficient is multiplied by K to obtain Q, wherein the maximum deviation coefficient is a deviation value of the second fault risk threshold value and the first fault risk threshold value; Randomly selecting Q comprehensive fault diagnosis plug-ins from the K comprehensive fault diagnosis plug-ins of the comprehensive fault diagnosis channel, and inputting the second monitoring data into the Q comprehensive fault diagnosis plug-ins to output Q fault risk coefficients, and obtaining the second fault risk coefficient after mean calculation.

Images