Folding propulsion fan fault prediction method based on multi-source monitoring data processing

Abstract

The invention discloses a folding propulsion fan fault prediction method based on multi-source monitoring data processing, which belongs to the field of fault detection, and comprises the following steps: arranging a first sensing monitoring module on a folding mechanism of a folding fan; a second sensing monitoring module is arranged on the spiral propeller, and the spiral propeller is connected with the folding fan; acquiring a first multi-source sensing monitoring data set and a second multi-source sensing monitoring data set according to the first sensing monitoring module and the second sensing monitoring module, inputting the first multi-source sensing monitoring data set and the second multi-source sensing monitoring data set into a lagging fault prediction model, respectively performing fault probability prediction, and outputting a first fault probability and a second fault probability; according to the first fault probability and the second fault probability, a jamming early warning signal is generated, and a folding mechanism and / or a spiral propeller are / is controlled to carry out jamming fault processing. By performing multi-source sensing monitoring and fault probability prediction on the folding mechanism and the spiral propeller, the technical effects of improving the fault prediction accuracy of the folding propulsion fan and realizing accurate fault positioning are achieved.

Description

Technical Field

[0001] The invention relates to the field of fault detection, and in particular to a folding propulsion fan fault prediction method based on multi-source monitoring data processing. Background Art

[0002] The folding propulsion fan realizes the folding and unfolding of the fan through the folding mechanism, and provides propulsion through the propeller propeller. However, the structure of the folding propulsion fan is relatively complex, involving the coordinated work of the folding mechanism and the propeller propeller. There is a certain risk of failure in actual use, especially the jamming failure, which may be caused by the failure of the folding mechanism itself or by the abnormality of the propeller. If the possibility of failure cannot be accurately predicted and the cause of the failure cannot be located in time, it will be difficult to implement effective fault prevention and fault handling.

[0003] At present, in order to monitor the faults of folding propulsion fans, a unified fault detection scheme is adopted, that is, the operation of the folding mechanism and the propeller as a whole is monitored, and fault detection is performed based on the collected monitoring data. However, this fault detection scheme has limitations. First, although the folding mechanism and the propeller work together, there are differences in their structures and functions. Using the same sensor and monitoring strategy, it is difficult to fully reflect their respective operating states and it is easy to miss fault characteristics. Secondly, due to the lack of targeted monitoring, the correlation between the obtained data and the fault characteristics is not strong, which affects the fault location. Therefore, the fault prediction accuracy of the folding propulsion fan in the prior art is not high, and the specific fault cannot be effectively located. Summary of the invention

[0004] The present application provides a folding propulsion fan fault prediction method based on multi-source monitoring data processing, aiming to solve the technical problems in the prior art that the folding propulsion fan fault prediction accuracy is low and the specific fault cannot be effectively located.

[0005] The present application discloses a method for predicting a folding propulsion fan fault based on multi-source monitoring data processing, comprising: setting a first sensor monitoring module, the first sensor monitoring module being arranged on a folding mechanism of the folding fan, the first sensor monitoring module comprising a stress sensor, a vibration sensor and a displacement sensor; setting a second sensor monitoring module, the second sensor monitoring module being arranged on a propeller, the propeller being connected to the folding fan and being used to propel the folding fan to perform a folding operation, the second sensor monitoring module comprising a speed sensor, a vibration sensor and an operation sensor; acquiring a first multi-source sensor monitoring data set and a second multi-source sensor monitoring data set according to the first sensor monitoring module and the second sensor monitoring module; inputting the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set into a jamming fault prediction model, respectively performing fault probability predictions according to the jamming fault prediction model, and outputting corresponding first fault probability and second fault probability; generating a jamming warning signal according to the first fault probability and the second fault probability, and using the jamming warning signal to control the folding mechanism and / or the propeller to perform jamming fault processing.

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

[0007] Due to the use of a first sensor monitoring module, the first sensor monitoring module is arranged on the folding mechanism of the folding fan, the first sensor monitoring module includes a stress sensor, a vibration sensor and a displacement sensor, and by arranging various types of sensors on the folding mechanism, stress, vibration, displacement and other signals are comprehensively collected to reflect the operating status and fault characteristics of the folding mechanism; a second sensor monitoring module is arranged, the second sensor monitoring module is arranged on the screw propeller, the screw propeller is connected to the folding fan, and is used to propel the folding fan to perform a folding operation, the second sensor monitoring module includes a speed sensor, a vibration sensor and an operation sensor, and by arranging various types of sensors on the screw propeller, speed, vibration, operation and other signals are comprehensively collected to reflect the working status and abnormal characteristics of the propeller; according to the first sensor monitoring module and the second sensor monitoring module, a first multi-source sensor monitoring data set and a second multi-source sensor monitoring data set are obtained, and the folding mechanism and the screw propeller are respectively obtained through the first sensor monitoring module and the second sensor monitoring module. The multi-source monitoring data of the folding mechanism and the propeller are used to form a targeted data set to provide data support for subsequent fault analysis; the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set are input into the jam fault prediction model, and the fault probability prediction is performed respectively according to the jam fault prediction model, and the corresponding first fault probability and second fault probability are output to improve the targetedness and accuracy of the fault prediction, and the respective fault probabilities of the folding mechanism and the propeller are obtained to provide a basis for fault location; according to the first fault probability and the second fault probability, a jam warning signal is generated, and the jam warning signal is used to control the folding mechanism and / or the propeller to handle the jam fault, and the technical solution of improving the timeliness and effectiveness of fault handling solves the technical problems in the prior art that the fault prediction accuracy of the folding propulsion fan is not high and the specific fault cannot be effectively located. By performing multi-source sensor monitoring and fault probability prediction on the folding mechanism and the propeller respectively, the technical effect of improving the fault prediction accuracy of the folding propulsion fan and realizing accurate fault location is achieved.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flow chart of a folding propulsion fan fault prediction method based on multi-source monitoring data processing is provided for an embodiment of the present application;

[0010] Figure 2 A flow chart of generating accompanying detection results in a folding propulsion fan fault prediction method based on multi-source monitoring data processing is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0011] The overall idea of ​​the technical solution provided by this application is as follows:

[0012] An embodiment of the present application provides a folding propulsion fan fault prediction method based on multi-source monitoring data processing. By respectively setting multi-source sensor monitoring modules on the folding mechanism and the propeller, data of the folding process and the propulsion process are obtained to form a comprehensive and targeted monitoring data set. These monitoring data are input into a jamming fault prediction model for analysis to obtain the respective failure probabilities of the folding mechanism and the propeller. According to the difference in failure probabilities between the two, an accurate jamming warning signal is generated to guide fault location and disposal.

[0013] Specifically, firstly, by arranging stress, vibration and displacement sensors on the folding mechanism, and arranging speed, vibration and operation sensors on the propeller, the multi-source state data of the folding mechanism and the propeller are collected respectively, which overcomes the deficiency of the unified fault detection scheme in the prior art. Then, the multi-source monitoring data sets of the folding mechanism and the propeller are respectively input into the pre-trained jamming fault prediction model, and the respective fault characteristics and laws are fully explored to obtain more accurate and targeted fault probability prediction results, which makes up for the defect of lack of prediction means in the prior art. Afterwards, by comparing the fault probabilities of the folding mechanism and the propeller, the main source of the fault is judged, and the corresponding jamming warning signal is generated to achieve accurate positioning of the fault and targeted treatment, thereby improving the timeliness of fault prevention and treatment. Compared with the prior art, the present invention fully considers the structural and functional differences between the folding mechanism and the propeller, improves the pertinence and accuracy of fault monitoring and prediction; at the same time, a special fault prediction model is introduced to achieve the transition from passive detection to active prevention, and improve the timeliness of fault warning and the effectiveness of treatment.

[0014] After introducing the basic principles of the present application, the non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the drawings in the specification.

[0015] like Figure 1 As shown, the embodiment of the present application provides a folding propulsion fan fault prediction method based on multi-source monitoring data processing, the method comprising:

[0016] S1: Setting a first sensor monitoring module, wherein the first sensor monitoring module is set on a folding mechanism of a folding fan, and the first sensor monitoring module includes a stress sensor, a vibration sensor and a displacement sensor.

[0017] Specifically, in order to achieve fault prediction of the folding propulsion fan, a first sensor monitoring module is provided. The first sensor monitoring module is installed on the folding mechanism of the folding fan and is used to collect the operating status data of the folding mechanism in real time.

[0018] The first sensor monitoring module includes multiple types of sensors to comprehensively monitor the status of the folding mechanism. Among them, the stress sensor is used to measure the stress changes of the folding mechanism during operation, and can promptly detect structural damage caused by abnormal stress; the vibration sensor is used to detect the vibration of the folding mechanism. Abnormal vibration usually means that the mechanism is not running smoothly and there may be hidden dangers of failure; the displacement sensor is used to monitor the displacement changes of the folding mechanism to determine whether the folding operation is in place.

[0019] Through the coordinated monitoring of these sensors, the first sensor monitoring module can comprehensively collect multi-source heterogeneous data of the folding mechanism, providing a reliable data basis for subsequent fault prediction analysis. As an important data source for folding fan fault prediction, the first sensor monitoring module set on the folding mechanism can effectively capture various sign data during the folding process, providing data support for accurate fault warning.

[0020] S2: Setting a second sensor monitoring module, wherein the second sensor monitoring module is set on a screw propeller, and the screw propeller is connected to the folding fan for propelling the folding fan to perform a folding operation, and the second sensor monitoring module includes a speed sensor, a vibration sensor and an operation sensor.

[0021] Specifically, while the first sensor monitoring module is provided, a second sensor monitoring module is provided to monitor the operating status of the propeller. The propeller is another component of the folding fan, and its function is to provide power for the folding operation of the folding fan. The second sensor monitoring module cooperates with the first sensor monitoring module to jointly realize comprehensive monitoring of the folding propulsion fan.

[0022] The second sensor monitoring module is installed on the propeller and is equipped with a variety of sensors, including speed sensors, vibration sensors and operation sensors. Among them, the speed sensor is responsible for collecting the speed data of the propeller in real time. By analyzing the changes in the speed, the possible performance degradation of the propeller can be found; the vibration sensor is used to monitor the vibration level of the propeller. Abnormal vibration often indicates that there are some faults in the mechanical structure inside the propeller; the operation sensor can continuously monitor the operation stability of the propeller. When the operation is abnormal, it means that the load of the propeller is abnormal.

[0023] By setting a second sensor monitoring module on the propeller, the real-time working status of the propeller can be fully grasped, and data reflecting its health level can be obtained. These data, combined with the monitoring data of the folding mechanism, can more accurately judge the overall operating condition of the folding propulsion fan, providing a reliable basis for timely detection and warning of faults.

[0024] S3: Acquire a first multi-source sensor monitoring data set and a second multi-source sensor monitoring data set according to the first sensor monitoring module and the second sensor monitoring module.

[0025] Specifically, the multi-source monitoring data of the folding mechanism and the propeller are acquired by using the first sensor monitoring module and the second sensor monitoring module respectively, and the data are sorted into two independent data sets to prepare data for subsequent fault prediction.

[0026] Among them, the data collected by the first sensor monitoring module is classified as the first multi-source sensor monitoring data set, which contains heterogeneous data of various monitoring points of the folding mechanism, including stress sensor data, vibration sensor data, and displacement sensor data. These data reflect the operating status of the folding mechanism from different angles and contain fault symptom information. Similarly, the data collected by the second sensor monitoring module is classified as the second multi-source sensor monitoring data set, which includes data of various monitoring indicators of the propeller, such as speed data, vibration data, and current data. These data comprehensively describe the working condition of the propeller and contain fault precursor information.

[0027] By acquiring the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set, comprehensive data collection and collation of the folding propulsion fan is achieved, laying a data foundation for fault prediction and fault location.

[0028] S4: Input the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set into a jamming fault prediction model, perform fault probability predictions according to the jamming fault prediction model, and output corresponding first fault probabilities and second fault probabilities.

[0029] Specifically, the jamming fault prediction model is a machine learning model trained based on a large amount of historical fault data. The model can automatically extract features related to jamming faults from the input monitoring data and calculate the fault probability based on the degree of matching between the features and the preset fault mode. The first fault probability indicates the possibility of predicting a jamming fault in the folding mechanism based on the first multi-source sensor monitoring data set, and the second fault probability indicates the possibility of predicting a jamming fault in the propeller based on the second multi-source sensor monitoring data set.

[0030] The monitoring data is analyzed and fault prediction is carried out through the Caton fault prediction model, and the advantages of machine learning technology in data processing and pattern recognition are used to improve the automation and accuracy of fault prediction. By calculating the first fault probability and the second fault probability, the failure risk of key components of the folding propulsion fan is quantitatively evaluated, the accuracy of folding propulsion fan fault prediction is improved, and accurate fault location is achieved, providing a quantitative basis for subsequent fault warning and maintenance, which helps to improve the safety and reliability of the equipment.

[0031] S5: Generate a jam warning signal according to the first failure probability and the second failure probability, and use the jam warning signal to control the folding mechanism and / or the propeller to handle the jam failure.

[0032] Specifically, when the first failure probability and / or the second failure probability exceeds the preset failure probability threshold, it indicates that the folding mechanism and / or the propeller have a high risk of jamming failure, and corresponding preventive or disposal measures need to be taken. At this time, a jamming warning signal is automatically generated, which contains the fault risk level, fault source information, etc. Then, the jamming warning signal is sent to relevant personnel in a variety of ways, such as displaying it on the monitoring interface, sending text messages or email notifications, etc., so as to promptly remind the staff to pay attention to the equipment status.

[0033] At the same time, the jam warning signal directly triggers the fault handling operation of the folding mechanism and / or the propeller. According to the preset control strategy, the operating parameters of the equipment are automatically adjusted, such as reducing the speed of the propeller, increasing the driving torque of the folding mechanism, etc., to reduce the impact of the jam fault. In some cases, the operation of the folding mechanism and / or the propeller is directly stopped to avoid further deterioration of the fault, and maintenance personnel are promptly arranged to inspect and repair it.

[0034] Through the jam warning signal, active warning and intervention are carried out before the fault occurs, which greatly improves the operating safety of the folding propulsion fan, realizes closed-loop management from fault prediction to fault prevention and control, and minimizes the impact of jam fault on equipment operation.

[0035] Furthermore, the embodiment of the present application also includes:

[0036] Determine whether the first fault probability and the second fault probability are greater than a preset fault probability; if the first fault probability and the second fault probability are both greater than the preset fault probability, use an accompanying detection module to perform accompanying fault detection on the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to obtain accompanying detection results; identify the folding mechanism and the screw propeller according to the accompanying detection results, and generate a jam warning signal.

[0037] In a preferred embodiment, first, the first fault probability and the second fault probability are judged to determine whether they both exceed the preset fault probability. The preset fault probability is set according to the design specifications, operating conditions and other factors of the folding propulsion fan, indicating the upper limit of the risk of a stuck fault. When the first fault probability and the second fault probability are both greater than the preset fault probability, it means that the folding mechanism and the propeller both have a high risk of stuck faults, which require further analysis and processing. At this time, the accompanying detection module is started to analyze the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to realize the accompanying fault detection function. The accompanying fault refers to a secondary fault caused by correlation or jointness at the same time as the main fault occurs. Through the comprehensive analysis of the two data sets, the accompanying detection module can find out whether there is a correlation or accompanying relationship between the folding mechanism and the propeller, and generate corresponding accompanying detection results, thereby identifying the primary and secondary causes and propagation paths of the stuck fault, and providing a basis for generating an accurate stuck warning signal. For example, when the folding mechanism jams, the vibration of the propeller will intensify, causing problems such as bearing wear. The accompanying detection module comprehensively analyzes the two data sets and finds that when the folding mechanism jams, the vibration frequency of the propeller increases significantly and is synchronized with the jamming frequency of the folding mechanism, indicating that there is a fault transmission relationship between the two.

[0038] Subsequently, based on the results of the accompanying detection, it is determined whether the main source of the jamming fault is the folding mechanism or the propeller, as well as the influence relationship between the two. This fine-grained fault identification helps maintenance personnel to carry out fault analysis and disposal work more targeted, improving the efficiency and accuracy of troubleshooting. At the same time, the generated jamming warning signal will also contain more fault attributes and related information, laying the foundation for accurate warning and intelligent processing of faults. For example, if the accompanying detection finds that the folding mechanism is the main source of faults, and the propeller has a concomitant fault, it will be marked in the jamming warning signal: "The folding mechanism has a jamming fault, and the fault risk is high; at the same time, the propeller is detected to have a fault transmission, which requires synchronous maintenance." In this way, the jamming warning signal not only indicates the main faulty components, but also reveals the propagation and impact of the fault, providing maintenance personnel with richer and more comprehensive information for timely disposal.

[0039] Through the accompanying fault detection and identification mechanism, when the first fault probability and the second fault probability are both high, by analyzing the correlation between the two multi-source sensor monitoring data sets, the causes and impacts of the jamming fault can be understood more comprehensively and deeply, and an early warning signal containing rich information can be generated, thereby achieving refined fault management and control of the folding propulsion fan.

[0040] Furthermore, the embodiment of the present application also includes:

[0041] The device attributes of the folding mechanism and the screw propeller are identified according to the accompanying detection results, wherein the identification information includes main fault identification information and accompanying fault identification information of the accompanying fault caused by the main fault device.

[0042] In a feasible implementation, in order to further clarify the source and impact of the jamming failure, the device attributes of the folding mechanism and the propeller are identified.

[0043] First, the accompanying detection results given by the accompanying detection module are used to judge and identify the fault attributes of the folding mechanism and the propeller. The identification content mainly includes: first, the main fault identification information, which is used to indicate the main source equipment of the fault; second, the accompanying fault identification information, which is used to indicate the equipment affected by the main fault equipment and causing joint faults.

[0044] By analyzing the accompanying detection results, the different roles and influence relationships of the folding mechanism and the propeller in the jamming failure were found. For example, the accompanying detection results showed that the folding mechanism had abnormal stress fluctuations and displacement mutations before the failure occurred, and after the failure occurred, its vibration level was much higher than that of the propeller, which indicated that the folding mechanism was the main source of the jamming failure. Therefore, the "main fault" mark was added to the folding mechanism; at the same time, the accompanying detection found that after the folding mechanism failed, the speed and current of the propeller also fluctuated significantly, and had a certain correlation and synchronization with the fault signal of the folding mechanism, indicating that although the propeller did not directly cause the failure, it was affected by the failure of the folding mechanism and had a certain degree of joint abnormality. Therefore, the "accompanying fault" mark was added to the propeller.

[0045] Through the device attribute identification, the primary and secondary fault devices and their associations can be clearly indicated in the jam warning signal, providing more accurate and targeted information for subsequent fault diagnosis and maintenance. For example, the jam warning signal is: "Main fault device: folding mechanism, fault type: jam, fault risk: high; accompanying fault device: propeller, fault type: speed fluctuation, fault risk: medium". In this way, maintenance personnel can carry out targeted fault handling, give priority to the main fault device, and take into account the accompanying fault devices affected by it, thereby improving the efficiency and economy of fault handling.

[0046] Furthermore, the embodiment of the present application also includes:

[0047] A jam warning signal is generated according to a device that identifies the main fault identification information, including: determining whether the folding mechanism and the screw propeller are both identified as the main fault identification information, and if the folding mechanism and the screw propeller are both identified as the main fault identification information, obtaining a first jam warning signal and a second jam warning signal; integrating the first jam warning signal and the second jam warning signal into a jam warning signal to perform jam fault processing on the folding mechanism and the screw propeller; if the folding mechanism and the screw propeller are identified as the main fault identification information and the accompanying fault identification information, a jam warning signal is generated according to the device that identifies the main fault identification information.

[0048] In a preferred embodiment, first, determine whether the folding mechanism and the screw propeller are both identified as the main fault identification information. If the folding mechanism and the screw propeller are both identified as the main fault device, it means that the two devices are the starting point of the jamming fault, each causing the fault, and there may be some coupling amplification effect between them. In this case, independent jamming warning signals are generated for the folding mechanism and the screw propeller, namely the first jamming warning signal and the second jamming warning signal. These two warning signals respectively contain the fault risk level, fault type, recommended operation and other contents of their respective devices. For example, the first jamming warning signal is "the risk of folding mechanism failure is high, it is recommended to stop the machine for maintenance immediately", and the second jamming warning signal is "the risk of screw propeller failure is high, it is recommended to reduce the speed and check the lubrication condition".

[0049] After the first jam warning signal and the second jam warning signal are generated, the two jam warning signals are integrated into a unified jam warning signal to comprehensively guide the jam fault handling of the folding mechanism and the propeller. The integrated jam warning signal not only contains the independent fault information of the two devices, but also reflects the mutual influence and coupling relationship between the folding mechanism and the propeller, so that maintenance personnel can formulate a collaborative disposal plan. For example, the integrated jam warning signal is "Warning: The folding mechanism and the propeller are jammed at the same time, the failure risk is high, and there is mutual stimulation between the failures of the two devices. It is recommended to coordinate investigation and repair."

[0050] If the folding mechanism and the propeller are identified as the main fault device and the accompanying fault device respectively, it means that the jamming fault is mainly caused by the main fault device, and the accompanying fault device is only affected by it and has a related abnormality. In this case, a targeted jamming warning signal is generated first based on the identification information of the main fault device. For example, if the folding mechanism is identified as the main fault device, the generated jamming warning signal should focus on the fault diagnosis and treatment of the folding mechanism, while taking into account the abnormal condition of the propeller. For example, the jamming warning signal is "the folding mechanism has a jamming fault, the risk of failure is high, and it is recommended to check and repair it immediately; at the same time, the propeller has a speed fluctuation, which is affected by the folding mechanism failure. It is recommended to reduce the speed and monitor at any time."

[0051] By making full use of the device attribute identification information, the generation strategy and content composition of the jam warning signal are dynamically adjusted to make it more consistent with the actual fault status and impact relationship of the device. This adaptive signal generation mechanism can provide a more accurate, complete and targeted decision-making basis for fault handling, help to quickly control the fault situation, coordinate equipment repair, and minimize fault downtime and losses.

[0052] Further, such as Figure 2 As shown, the embodiment of the present application also includes:

[0053] The accompanying detection module is used to perform abnormal data identification on the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to obtain a first abnormal sensor data set and a second abnormal sensor data set; the abnormal time series accompaniment, abnormal data accompaniment and abnormal pattern accompaniment of the first abnormal sensor data set and the second abnormal sensor data set are identified; and an accompanying detection result is generated based on the abnormal time series accompaniment, abnormal data accompaniment and abnormal pattern accompaniment.

[0054] In a preferred embodiment, first, the accompanying detection module comprehensively scans and analyzes the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to identify potential abnormal data therein. Abnormal data refers to sensor readings or characteristic values ​​that deviate from normal working conditions and exceed a predetermined threshold range. For example, when the stress data of the folding mechanism suddenly surges and exceeds the design strength threshold, or the speed data of the propeller drops sharply to near zero, it is determined to be abnormal data. Through abnormal data identification, key information reflecting equipment failure or abnormality is extracted from a large amount of monitoring data, laying the foundation for subsequent analysis. The results of the identification will be saved in the first abnormal sensor data set and the second abnormal sensor data set, respectively, for targeted processing.

[0055] Next, the accompanying detection module will conduct an in-depth comparative analysis of the first abnormal sensor data set and the second abnormal sensor data set to identify three possible accompanying characteristics: abnormal time sequence accompanying, abnormal data accompanying and abnormal pattern accompanying. Among them, abnormal time sequence accompanying refers to the time sequence of abnormal occurrence of the two data sets. The device corresponding to the data set that appears abnormal first is the main fault source, and the device corresponding to the data set that appears abnormal later is the accompanying fault that is affected. For example, it is found that the abnormal stress data of the folding mechanism appears before the abnormal speed of the propeller, which indicates that the folding mechanism is the root cause of the fault and the abnormality of the propeller is caused by it. Abnormal data accompanying refers to whether the abnormal data of the folding mechanism and the propeller are consistent in change trend and amplitude. If the abnormal data of the folding mechanism and the propeller change synchronously in the time dimension and show similar increase and decrease trends and change amplitudes, it means that there is fault transmission or mutual influence between the two. Abnormal pattern accompaniment refers to whether the morphological characteristics of abnormal data in the time or frequency domain are similar. Signal processing and pattern recognition techniques are used to analyze the morphological characteristics of abnormal data, such as periodicity, symmetry, peak distribution, etc. If the abnormal data of the folding mechanism and the screw propeller show similar pattern characteristics, such as the same oscillation period or attenuation law, it is inferred that their anomalies originate from a common failure mechanism or mutual coupling.

[0056] Afterwards, the accompanying detection module integrates the recognition results of abnormal time sequence accompanying, abnormal data accompanying and abnormal pattern accompanying to generate the final accompanying detection result. The accompanying detection result describes the correlation and influence relationship between the first abnormal sensor data set and the second abnormal sensor data set, including the abnormal sequence of the two data sets, the abnormal correlation strength level, the correlation coefficient or similarity index of the abnormal data, etc. For example, the accompanying detection result is "the folding mechanism abnormality precedes the propeller abnormality by 0.5s, the correlation coefficient of the abnormal data of the two devices is 0.8, the abnormal pattern similarity is 75%, and it is comprehensively determined that there are significant accompanying fault characteristics."

[0057] By making full use of the abnormal information and correlation features contained in the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set, a complete accompanying fault detection mechanism is constructed. Starting from multiple dimensions such as time, amplitude, and shape, the abnormal data relationship between devices is comprehensively characterized. The main fault and accompanying fault can be identified quickly and accurately, revealing the complex fault transmission and coupling process.

[0058] Furthermore, the embodiment of the present application also includes:

[0059] According to the abnormal time sequence accompaniment, the device that fails first in the abnormal time sequence window is pre-identified as the main fault device, and the device that fails later is pre-identified as the accompanying fault device; according to the abnormal data accompaniment and the abnormal pattern accompaniment, identification is performed to generate dual constraints, wherein the dual constraints include a preset proportion and a preset similarity; if any of the dual constraints is not met, the pre-identification result is optimized.

[0060] In a feasible implementation, first, the folding mechanism and the screw propeller are preliminarily identified as the primary and secondary faults by using the abnormal time sequence concomitancy. Specifically, an abnormal time sequence window is set, and the abnormal data falling into the window at the same time are analyzed. The device that first appears abnormal data is pre-identified as the primary fault device, and the device that appears abnormal data later is pre-identified as the accompanying fault device. For example, if it is found that the abnormal stress data of the folding mechanism appears in the first half of the abnormal time sequence window, and the abnormal speed data of the screw propeller appears in the second half, it is preliminarily determined that the folding mechanism is the primary fault device and the screw propeller is the accompanying fault device. At the same time, according to the abnormal data concomitancy and the abnormal mode concomitancy, dual constraints are generated, including the preset proportion and the preset similarity, as a quantitative indicator for judging the accuracy of the pre-identification result. Among them, the preset proportion refers to the number of associated data in the abnormal data concomitancy reaching a certain proportion. If the number of synchronously changing data points of the abnormal data of the two devices in the time dimension accounts for less than the preset value (such as 80%) of the total number of abnormal data points, it is considered that the abnormal data concomitancy is not significant enough, and the pre-identification result needs to be optimized. The preset similarity refers to the key characteristic parameters in the abnormal pattern accompaniment, such as correlation coefficient, similarity, etc. If the correlation coefficient of the abnormal data of two devices is less than the preset value (such as 0.6), or the similarity of the abnormal pattern is lower than the preset value (such as 70%), it is considered that the abnormal pattern accompaniment is not strong enough, and the pre-identification result also needs to be optimized.

[0061] Then, the dual constraints are checked. If any of the constraints in abnormal data concomitant and abnormal pattern concomitant is not satisfied, it means that the current primary and secondary fault pre-identification results are not reliable enough and need to be optimized and adjusted. The adjustment strategy can be to exchange the identification of the primary and secondary fault devices, or to identify both devices as the primary fault device (when the abnormal data and patterns of the two devices appear almost synchronously). Then, the optimized identification result is used as the final concomitant detection result to guide subsequent fault handling.

[0062] Based on the analysis of abnormal time series accompaniment, the abnormal data accompaniment and abnormal pattern accompaniment are further integrated, and dual constraints are introduced for quantitative evaluation, thereby achieving adaptive optimization of the preliminary primary and secondary fault pre-identification results and improving the accuracy of fault location.

[0063] Furthermore, the embodiment of the present application also includes:

[0064] Among them, the abnormal data accompaniment is identified by abnormal proportion, including identifying the proportion of the accompanying data volume in the total abnormal data volume to obtain the abnormal proportion; the abnormal pattern accompaniment is identified by accompanying pattern similarity to obtain pattern similarity; the abnormal proportion and the pattern similarity are used as input information of the double constraint conditions, and compared with the preset proportion and the preset similarity. If the abnormal proportion is greater than the preset proportion, and the pattern similarity is greater than the preset similarity, the pre-identification result is output.

[0065] In a preferred embodiment, first, the abnormal data concomitancy is quantitatively evaluated through the abnormal proportion, and the degree of correlation between the abnormal data of the folding mechanism and the screw propeller is measured. Specifically, the number of data points of the abnormal data of the two devices that change synchronously in the time dimension is counted, and the proportion of the data points to the total number of abnormal data points of each device is calculated, which is defined as the abnormal proportion. The higher the abnormal proportion, the more synchronous the abnormal data changes of the two devices, and the greater the possibility of concomitant failures between them. For example, if the number of synchronous change points of abnormal data of the folding mechanism and the screw propeller account for 85% and 90% of the total amount of their respective abnormal data, respectively, the abnormal proportions can be identified as 0.85 and 0.90, respectively. These two higher proportion values ​​indicate significant abnormal data concomitancy.

[0066] At the same time, pattern similarity is used to quantify the similarity of abnormal pattern co-occurrence, describing the degree of consistency of the morphological features of abnormal data in the time or frequency domain using pattern similarity. By extracting features and matching patterns on abnormal data, a similarity coefficient between 0 and 1 is calculated. The closer the coefficient is to 1, the closer the two abnormal patterns are, and the higher the possibility that they may originate from a common fault mechanism or mutual coupling. For example, if the abnormal data of the folding mechanism and the propeller have an obvious resonance peak in the spectrum after time-frequency analysis, and the peak frequency difference is less than 5%, it is considered that the abnormal pattern similarity is very high, and the similarity coefficient can reach more than 0.95, which further supports the judgment that there is a co-occurrence fault between the two.

[0067] Afterwards, the obtained abnormal proportion and the obtained pattern similarity are used as two quantitative indicators of the dual constraint conditions, and are compared with the preset proportion and preset similarity, so as to verify the pre-identification results of the primary and secondary faulty equipment. Only when the abnormal proportion and pattern similarity simultaneously meet or exceed the preset proportion and preset similarity, the current pre-identification result is considered credible and output as the final concomitant fault detection result; otherwise, the pre-identification result is optimized and adjusted until the dual constraint conditions are met.

[0068] Furthermore, the embodiment of the present application also includes:

[0069] Acquire a first abnormal sensor data set and a second abnormal sensor data set; establish a first fault prediction model and a second fault prediction model, wherein the first fault prediction model includes a preset jamming fault sample corresponding to the folding mechanism, and the second fault prediction model includes a preset jamming fault sample corresponding to the propeller; predict the first abnormal sensor data set and the second abnormal sensor data set according to the first fault prediction model and the second fault prediction model, respectively, and output a first fault probability and a second fault probability.

[0070] In a preferred embodiment, first, abnormal data related to the jamming fault is further extracted from the acquired first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to form the first abnormal sensor data set and the second abnormal sensor data set. For example, by setting a threshold condition, stress pulse signals with abnormal amplitudes, vibration data with abnormal frequencies, etc. are identified from the monitoring data of the stress, vibration, displacement, etc. of the folding mechanism, and are included in the first abnormal sensor data set; similarly, speed curves with abnormal changes, over-limit current data, etc. are extracted from the monitoring data of the speed, current, temperature, etc. of the propeller, and are included in the second abnormal sensor data set.

[0071] Next, independent jamming fault prediction models, namely the first fault prediction model and the second fault prediction model, are established for the folding mechanism and the propeller respectively based on the corresponding historical fault sample data. Specifically, the first fault prediction model uses the sensor monitoring data of the folding mechanism when it has jammed in the past period of time as training samples, and is trained and optimized through machine learning algorithms (such as support vector machines, decision trees, etc.) to grasp the characteristic patterns and laws of the jamming faults of the folding mechanism; similarly, the second fault prediction model needs to be trained using the historical jamming fault data corresponding to the propeller to learn the abnormal data characteristics when the propeller jams. These two models reflect the failure mechanisms of the folding mechanism and the propeller, respectively, and are the basis for accurate predictions.

[0072] Afterwards, the first fault prediction model and the second fault prediction model are used to perform fault prediction on the first abnormal sensor data set and the second abnormal sensor data set. Specifically, the first abnormal sensor data set is input into the corresponding fault prediction model, and the fault prediction model automatically extracts key features, calculates the similarity between the abnormal data and the historical fault samples, and gives the probability prediction value of the fault occurrence based on the decision logic (such as threshold comparison, probability calculation, etc.), that is, the first fault probability and the second fault probability, which quantifies the possibility of the current folding mechanism and the propeller jamming. For example, if the first fault probability is as high as 90%, and the second fault probability is only 10%, it means that the current folding mechanism is very likely to jam, which requires special attention and processing; while the failure risk of the propeller is relatively low, and the maintenance frequency can be appropriately reduced.

[0073] Furthermore, the embodiment of the present application also includes:

[0074] The folding mechanism comprises a folding rod, and the folding rod comprises a first rod and a second rod, and the first rod and the second rod are a nested structure to achieve folding of the folding fan.

[0075] In one embodiment, the folding mechanism adopts a structural design including a folding rod to realize the folding function of the folding fan. The folding rod is composed of two parts, a first rod and a second rod. A nested connection structure is adopted between the first rod and the second rod, that is, one end of the first rod is inserted into the inner cavity of the second rod, and the two realize telescopic movement by sliding cooperation. In the extended state of the folding rod, the first rod and the second rod are coaxially aligned, and the axial direction thereof coincides with the rotation axis of the fan blade; in the folded state, the first rod will slide into the inner cavity of the second rod along the axial direction, so that the length of the entire folding rod is greatly reduced, thereby driving the fan blades connected thereto to fold and reduce the volume of the fan.

[0076] By adopting a folding rod composed of a first rod and a second rod nested together, the folding fan can be folded with high reliability. Through the coaxial nesting design, the axial size of the fan can be greatly reduced in the folded state, and in the working state, the fan blades can be supported in place through telescopic movement to ensure the normal operation of the fan.

[0077] In summary, the folding propulsion fan fault prediction method based on multi-source monitoring data processing provided by the embodiment of the present application has the following technical effects:

[0078] A first sensor monitoring module is provided, and the first sensor monitoring module is provided on the folding mechanism of the folding fan. The first sensor monitoring module includes a stress sensor, a vibration sensor and a displacement sensor. By arranging various types of sensors on the folding mechanism, the stress, vibration and displacement signals in the folding process are fully collected, and the data of the operating state of the folding mechanism is obtained, so as to provide targeted data support for subsequent fault prediction. A second sensor monitoring module is provided, and the second sensor monitoring module is provided on the propeller. The propeller is connected to the folding fan and is used to propel the folding fan to perform a folding operation. The second sensor monitoring module includes a speed sensor, a vibration sensor and an operation sensor. By arranging various types of sensors on the propeller, the speed, vibration and operation signals in the propulsion process are fully collected, and the data of the working state of the propeller is obtained, so as to provide targeted data support for subsequent fault prediction. According to the first sensor monitoring module and the second sensor monitoring module, a first multi-source sensor monitoring data set and a second multi-source sensor monitoring data set are obtained, and monitoring data are obtained from the multi-source sensor monitoring modules of the folding mechanism and the propeller, respectively, to form the multi-source monitoring data sets of the folding mechanism and the propeller, respectively, to provide comprehensive and targeted data for the jamming fault prediction model. The first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set are input into the jamming fault prediction model, and the fault probability prediction is performed respectively according to the jamming fault prediction model, and the corresponding first fault probability and second fault probability are output to provide a quantitative basis for subsequent fault location and early warning. According to the first fault probability and the second fault probability, a jamming warning signal is generated, and the jamming warning signal is used to control the folding mechanism and / or the propeller to handle the jamming fault. The source of the jamming fault is determined by the failure probability of the folding mechanism and the propeller, and a corresponding jamming warning signal is generated to guide the accurate location and targeted handling of the fault, thereby improving the timeliness and effectiveness of fault prevention and handling.

[0079] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0080] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to multiple elements that can be selected individually or in full. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these changes and variations.

Claims

1. A folding propulsion fan fault prediction method based on multi-source monitoring data processing, characterized in that: The method comprises: A first sensor monitoring module is provided, wherein the first sensor monitoring module is provided on the folding mechanism of the folding fan, and the first sensor monitoring module includes a stress sensor, a vibration sensor and a displacement sensor; A second sensor monitoring module is provided, wherein the second sensor monitoring module is provided on a screw propeller, the screw propeller is connected to the folding fan, and is used to propel the folding fan to perform a folding operation, and the second sensor monitoring module includes a speed sensor, a vibration sensor, and an operation sensor; Acquire a first multi-source sensor monitoring data set and a second multi-source sensor monitoring data set according to the first sensor monitoring module and the second sensor monitoring module; Inputting the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set into a jamming fault prediction model, performing fault probability predictions according to the jamming fault prediction model, and outputting corresponding first fault probabilities and second fault probabilities; A jam warning signal is generated according to the first failure probability and the second failure probability, and the jam warning signal is used to control the folding mechanism and / or the screw propeller to handle the jam failure.

2. The method according to claim 1, characterized in that Generating a jam warning signal according to the first failure probability and the second failure probability, the method comprising: Determining whether the first failure probability and the second failure probability are greater than a preset failure probability; If both the first fault probability and the second fault probability are greater than the preset fault probability, performing a concomitant fault detection on the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set using a concomitant detection module to obtain a concomitant detection result; The folding mechanism and the screw propeller are identified according to the accompanying detection results, and a jam warning signal is generated.

3. The method according to claim 2, characterized in that The device attributes of the folding mechanism and the screw propeller are identified according to the accompanying detection results, wherein the identification information includes main fault identification information and accompanying fault identification information of the accompanying fault caused by the main fault device.

4. The method according to claim 3, characterized in that Generating a jam warning signal according to the device identifying the main fault identification information, the method further includes: Determine whether the folding mechanism and the screw propeller are both identified as main fault identification information, and if the folding mechanism and the screw propeller are both identified as main fault identification information, obtain a first jam warning signal and a second jam warning signal; Integrate the first jam warning signal and the second jam warning signal into a jam warning signal to perform jam fault processing on the folding mechanism and the propeller; If the folding mechanism and the screw propeller are identified as main fault identification information and accompanying fault identification information, a jam warning signal is generated according to the device that identifies the main fault identification information.

5. The method according to claim 3, characterized in that Performing accompanying fault detection on the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set using an accompanying detection module, the method comprising: Using a companion detection module to perform abnormal data identification on the first multi-source sensor monitoring data set and the second multi-source sensor monitoring data set to obtain a first abnormal sensor data set and a second abnormal sensor data set; Identifying abnormal time series concomitancy, abnormal data concomitancy, and abnormal pattern concomitancy of the first abnormal sensor data set and the second abnormal sensor data set; Based on the abnormal time series accompaniment, the abnormal data accompaniment and the abnormal pattern accompaniment, an accompaniment detection result is generated.

6. The method according to claim 5, characterized in that According to the abnormal sequence concomitancy, the device that fails first in the abnormal sequence window is pre-marked as the main fault device, and the device that fails later is pre-marked as the accompanying fault device; Identify according to the abnormal data accompanyingness and the abnormal pattern accompanyingness, and generate a dual constraint condition, wherein the dual constraint condition includes a preset proportion and a preset similarity; If any one of the dual constraints is not satisfied, the pre-identified result is optimized.

7. The method according to claim 6, characterized in that in, Identify the abnormal proportion of the abnormal data, including identifying the proportion of the amount of accompanying data to the total amount of abnormal data, and obtaining the abnormal proportion; Performing accompanying pattern similarity recognition on the abnormal pattern accompanyingness to obtain pattern similarity; The abnormal proportion and the pattern similarity are used as input information of the dual constraint conditions, and compared with the preset proportion and the preset similarity. If the abnormal proportion is greater than the preset proportion and the pattern similarity is greater than the preset similarity, the pre-identification result is output.

8. The method according to claim 5, characterized in that According to the jamming fault prediction model, fault probability prediction is performed respectively, and a corresponding first fault probability and a second fault probability are outputted, and the method further includes: Acquire a first abnormal sensing data set and a second abnormal sensing data set; Establishing a first fault prediction model and a second fault prediction model, wherein the first fault prediction model includes a preset jamming fault sample corresponding to the folding mechanism, and the second fault prediction model includes a preset jamming fault sample corresponding to the propeller; The first abnormal sensor data set and the second abnormal sensor data set are predicted respectively according to the first fault prediction model and the second fault prediction model, and a first fault probability and a second fault probability are output.

9. The method according to claim 1, characterized in that The folding mechanism comprises a folding rod, and the folding rod comprises a first rod and a second rod, and the first rod and the second rod are a nested structure to achieve folding of the folding fan.

Images