Method and device for constructing health baseline of multiple parameters of ultra-short wave power amplifier assembly

Abstract

The application discloses a kind of ultra-short wave power amplifier assembly multi-parameter fusion health baseline construction method and device, it is related to data processing technical field, the method, comprising: obtaining the operating data of multiple groups of power amplifier type equipment, wherein, each group of operating data record is generated by pre-set flight number, determine the operating principle and failure mechanism of each group of power amplifier type equipment based on each group of operating data, establish screening model, determine the input data label and output data label of screening model according to operating principle and failure mechanism, extract target data set from operating data based on input data label and output data label, align and merge target data set using timestamp, to generate the multi-parameter fusion health baseline of power amplifier device and construct data set, can consider operating characteristics, time sequence relationship and other factors, analyze the parameter representation influence of the coupling characteristics of data on abnormal mode, realize accurate detection to failure and accurate evaluation of equipment health state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a method and device for constructing a multi-parameter fusion health baseline of an ultrashort wave power amplifier assembly. BACKGROUND

[0002] An airborne power amplifier device is one of the key core components of an aircraft. Due to external factors such as electromagnetic interference, mechanical vibration, and the like, and due to the aging of the equipment itself, drift of device indicators, wear and tear, and the like, the airborne power amplifier device may malfunction, seriously affecting the normal execution of the aircraft mission.

[0003] In related technologies, the traditional maintenance mode mainly includes two kinds of post-maintenance and planned maintenance. Post-maintenance is maintenance and repair after the occurrence of an electronic device fault, which is a maintenance mode driven by fault events, and has high maintenance costs and cannot avoid the adverse consequences caused by device malfunctions. Planned maintenance is a regular maintenance and component replacement of electronic devices according to reliability theory, which is a maintenance mode driven by time, and frequent maintenance, disassembly, and replacement result in low maintenance efficiency and high maintenance costs.

[0004] In order to solve the problems of insufficient maintenance of post-maintenance and excessive maintenance of planned maintenance, a condition-based maintenance mode based on device state monitoring and with fault diagnosis algorithm as the core has gradually developed. Therefore, how to provide effective data support for performance monitoring, state evaluation, and fault prediction of the device is a problem to be solved. SUMMARY

[0005] In view of the above problems, the present application provides a method and device for constructing a multi-parameter fusion health baseline of an ultrashort wave power amplifier assembly to at least solve the problems in related technologies.

[0006] In a first aspect, an embodiment of the present application provides a method for constructing a multi-parameter fusion health baseline of an ultrashort wave power amplifier assembly, which comprises:

[0007] Obtaining running data of a plurality of groups of power amplifier devices, wherein each group of running data records is generated by a preset number of flights;

[0008] Determining the running principle and fault mechanism of the power amplifier devices of each group based on each group of running data;

[0009] Establishing a screening model, and determining the input data label and output data label of the screening model according to the running principle and fault mechanism;

[0010] Extracting a target data set from the running data based on the input data label and output data label;

[0011] aligning and merging the target data sets by timestamps to generate a multi-parameter fusion health baseline of the power amplifier device and build a data set.

[0012] In some embodiments, the aligning and merging the target data sets by timestamps comprises:

[0013] merging the function record data in the target data sets and reading PTT data, obtaining a first timestamp of any of the PTT data;

[0014] finding corresponding working modes and wave channel numbers based on the first timestamp and the PTT data;

[0015] data merging the PTT data, the working modes and the wave channel numbers.

[0016] In some embodiments, the aligning and merging the target data sets by timestamps further comprises:

[0017] determining monitoring data in the target data sets and obtaining a second timestamp of the monitoring data;

[0018] finding any PTT data in a preset time period before the second timestamp and selecting a target PTT data closest to the second timestamp;

[0019] data merging the monitoring data, the target PTT data and the function record data.

[0020] In some embodiments, the multi-parameter fusion health baseline construction method of the ultrashort wave power amplifier assembly further comprises:

[0021] dividing the data set into a test set and a training set, and dividing into a support set and a query set;

[0022] randomly selecting N categories from all categories in the data set;

[0023] randomly selecting K samples from the selected category samples to form a support set, and selecting X samples from the remaining samples of the selected category to form a query set, wherein the support set has samples, and the query set has samples.

[0024] In some embodiments, the multi-parameter fusion health baseline construction method of the ultrashort wave power amplifier assembly further comprises:

[0025] extracting features of each sample in the support set to obtain an encoding of each sample;

[0026] Summing and averaging the encoding of all samples under each classification in the support to serve as a prototype representation of each label classification;

[0027] When a data sample is input, an encoding is used on the data sample to generate a corresponding encoding representation;

[0028] Based on the Euclidean distance, the distance between the encoding representation of the new data sample and each classification prototype representation is calculated;

[0029] Based on the activation function, the distance is converted into probability to determine the probability of the data sample belonging to the corresponding output label, and the class with the highest output probability is calculated.

[0030] In some embodiments, the ultra-short wave power amplifier component multi-parameter fusion health baseline construction method further comprises:

[0031] Using a quantile-based statistical method, the normal range of power values of different working modes and wave channels in the data set is determined.

[0032] In some embodiments, the quantile-based statistical method is used to determine the normal range of power values of different working modes and wave channels in the data set, comprising:

[0033] Reading the relevant parameters in the data set;

[0034] Grouping the relevant parameters according to the working mode and wave channel number, and cyclically calculating the quantiles of the power values corresponding to each group of relevant parameters;

[0035] Determine the upper and lower bounds of the quantiles, and construct the power value health baseline of each group of data based on the relationship between the input sample and the output label to determine the normal range of power values of different working modes and wave channels.

[0036] In a second aspect, the embodiments of the present application provide an ultra-short wave power amplifier component multi-parameter fusion health baseline construction device, comprising:

[0037] An acquisition module is configured to acquire running data of multiple groups of power amplifier type devices, wherein each group of running data is generated by a preset number of flights;

[0038] A determination module is configured to determine the running principle and failure mechanism of the power amplifier type devices in each group based on each group of running data;

[0039] A screening module is configured to establish a screening model, and determine input data labels and output data labels of the screening model according to the running principle and failure mechanism;

[0040] An extraction module is configured to extract a target data set from the running data based on the input data labels and the output data labels;

[0041] A generating module is configured to align and merge the target data set by using the time stamp to generate the multi-parameter fusion health baseline of the power amplifier device and construct a data set.

[0042] The multi-parameter fusion health baseline construction method and device for the ultrashort wave power amplifier assembly provided by the embodiment of the present application can comprehensively consider factors such as operation characteristics and time sequence relationship, analyze the influence of the coupling characteristics of data on the parameter representation of abnormal patterns, and realize accurate detection of faults and accurate evaluation of the health state of the device.

[0043] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0044] The present application will be described in more detail below based on the embodiments and with reference to the accompanying drawings.

[0045] Figure 1 A flowchart of the multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly is shown in an embodiment of the present application;

[0046] Figure 2 An exemplary schematic diagram of the multi-parameter fusion health baseline construction is shown in an embodiment of the present application;

[0047] Figure 3 An exemplary training flowchart in the multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly is shown in an embodiment of the present application;

[0048] Figure 4 An exemplary support set and query set division schematic diagram is shown in an embodiment of the present application;

[0049] Figure 5 A structural block diagram of the multi-parameter fusion health baseline construction device for the ultrashort wave power amplifier assembly is shown in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the objects, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with embodiments and drawings, the schematic embodiments and their descriptions are only used to explain the present application, and do not limit the present application.

[0051] In order to solve the problems of insufficient maintenance of after-service and excessive maintenance of planned maintenance, the condition-based maintenance mode based on equipment state monitoring and taking fault diagnosis algorithm as the core gradually develops. The key problem of the condition-based maintenance mode is how to construct a health assessment model according to the equipment state measurement data, and the construction of the health baseline plays a crucial role in state assessment. The present application can provide a benchmark model based on historical data, equipment specifications and operating conditions for abnormal detection and health assessment through the construction of the multi-parameter fusion health baseline, and provides effective data support for subsequent performance monitoring, state assessment and fault prediction.

[0052] In view of the above problems, the applicant proposes a multi-parameter fusion health baseline construction method and device for an ultrashort wave power amplifier assembly. The multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly comprehensively considers the system operation characteristics, time sequence relationship and other factors related to the ultrashort wave power amplifier assembly, analyzes the parameter characterization influence of the coupling characteristics of the system operation monitoring data on the abnormal mode, and constructs an electronic equipment abnormal detection data set and a health assessment data set according to different types of parameters such as power amplifier equipment state, control parameter, temperature, power value and module monitoring result. Using the small sample learning method, the effective classification model can be learned sufficiently under the condition of a small amount of fault samples, and accurate detection of faults and accurate assessment of the health state of the equipment can be realized.

[0053] The multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly is described in detail in subsequent embodiments.

[0054] The application scenario of the multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly provided in the embodiments of the present application is introduced as follows:

[0055] Please refer to Figure 1 , Figure 1 The multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly provided in the embodiments of the present application is described in detail in subsequent embodiments. Figure 5 The multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly provided in the embodiments of the present application is described in detail in subsequent embodiments. Figure 1 The multi-parameter fusion health baseline construction method for the ultrashort wave power amplifier assembly provided in the embodiments of the present application is described in detail in subsequent embodiments.

[0056] S110: Obtain running data of multiple groups of power amplifier type devices, wherein each group of running data records is generated by a preset number of flights.

[0057] In the embodiments of the present application, multiple groups of power amplifier type device running data are collected, and each group of running data file is power amplifier type device running data generated by one flight.

[0058] S120: Determine the operation principle and failure mechanism of the power amplifier type device of each group based on each group of running data.

[0059] S130: Establish a screening model, and determine the input data label and output data label of the screening model according to the operation principle and failure mechanism.

[0060] In the embodiments of the present application, for each group of running data, the input is confirmed according to the operation principle and failure mechanism of the power amplifier type device, which can include: power amplifier temperature, power indication, power measured value, standing wave state, PTT valid bit and wave channel number as input. The output label can include "normal", "standing wave failure", "power indication failure", "input excitation failure" and "false alarm", etc.

[0061] S140: Extract a target data set from the running data based on the input data label and the output data label.

[0062] In the embodiments of the present application, the required data can be extracted from the original data of the airborne power amplifier type device, and then the data is cleaned to screen out incomplete data, as shown in Figure 2 The data can be extracted according to the demand analysis from the original data, and the data can be cleaned.

[0063] S150: Align and merge the target data set by using a time stamp to generate a multi-parameter fusion health baseline of the power amplifier type device and construct a data set.

[0064] In some embodiments, S150 includes S151 to S153.

[0065] S151: Merge the function record data in the target data set and read the PTT data to obtain a first time stamp of any one PTT data.

[0066] S152: Find the corresponding working mode and wave channel number based on the first time stamp and the PTT data.

[0067] S153: Merge the PTT data, the working mode and the wave channel number.

[0068] In the embodiments of the present application, according to the system operation principle, the time sequence relationship between the data of each part is followed, the timestamp is used to align and merge the module monitoring parameters and the function use parameters, and a complete data is combined. Specifically, first, the function record data is merged, the PTT data is read line by line, the timestamp of a piece of PTT data is obtained, and then the corresponding working mode and wave channel number are found. (The timestamp of the wave channel number working mode is before the PTT data and is closest to the timestamp of the PTT), and the PPT data is merged with the wave channel number and the working mode data.

[0069] In some embodiments, S150 further includes S154 to S156.

[0070] S154: determining the monitoring data in the target data set, and obtaining the second timestamp of the monitoring data.

[0071] S155: finding any PTT data in a preset time period before the second timestamp, and selecting the target PTT data closest to the second timestamp.

[0072] S156: merging the monitoring data, the target PTT data and the function record data.

[0073] In the embodiments of the present application, the module monitoring data is obtained line by line, the timestamp of the module monitoring data is obtained, then a PTT within 6 seconds before the timestamp is found (the effective time of PTT is 4-6 seconds), the PTT closest to the timestamp of the module monitoring data is selected, and then the module monitoring data and the function record data are merged. Repeat the operation until all data merging of this group is completed.

[0074] Repeat steps S151-S156 to complete the collection, processing and data merging of multiple groups of operation data, and form the health baseline construction data set of the multi-parameter fusion of the power distribution class equipment.

[0075] In some embodiments, the data set is divided.

[0076] Referring to Figure 1 and Figure 4 , the merged data set can be divided into a test set and a training set, and needs to be divided into a support set and a query set.

[0077] First, for all categories, randomly select categories.

[0078] In the selected category samples, randomly select samples to constitute a support set, wherein there are samples.

[0079] Then, in the remaining samples of the selected category, select samples, to form a query set, wherein there are samples.

[0080] In some embodiments, the construction of the prototype network is performed.

[0081] The multi-parameter fusion health baseline construction method of the ultra-short wave power amplifier assembly further includes:

[0082] Feature extraction is performed on each sample in the support set to obtain the encoding of each sample;

[0083] The encodings of all samples under each classification in the support set are summed and averaged to obtain the prototype representation of each label classification.

[0084] When a data sample is input, the encoding is used to generate a corresponding encoding representation for the data sample;

[0085] The distances between the encoding representation of the new data sample and each prototype representation of the classification are calculated based on the Euclidean distance;

[0086] The distances are converted into probabilities based on the activation function to determine the probability that the data sample belongs to the corresponding output label, and the class with the highest output probability is calculated.

[0087] In the specific embodiments:

[0088] First, for each support set containing n classes of data, the encoding Feature extraction is performed on the samples to obtain the encoding representation of each sample as follows:

[0089]

[0090] wherein, is a query sample, and the number of samples in each class can be different, for example, the first query sample can be represented as , is a sample label, and is a feature extraction network, which can obtain the feature encoding of the query sample .

[0091] After obtaining the encoding representation of each sample, the encodings of all samples under each classification are summed and averaged, and the result is taken as the prototype vector of each label classification.

[0092] ​​The distance between the encoding representation of the new sample and the representation of each classification prototype is calculated based on the Euclidean distance, and finally the distance is converted into a probability form by the Softmax activation function, representing the probability of the data sample belonging to the corresponding output label. The class with the highest output probability is expressed as:

[0093]

[0094] wherein, is the query sample the probability of belonging to the class label, is the prototype vector of the class label, is the feature extraction network, is the distance metric function, is the exponential function.

[0095] In some embodiments, the method for constructing a multi-parameter fusion health baseline of the ultra-short wave power amplifier assembly further comprises:

[0096] A statistical method based on quantiles is used to determine the normal range of power values of different working modes and wave channels in the data set, specifically including:

[0097] Reading the relevant parameters in the data set;

[0098] Grouping the relevant parameters according to the working mode and wave channel number, and cyclically calculating the quantiles of the power values corresponding to each group of relevant parameters;

[0099] Determine the upper and lower bounds of the quantiles, and construct the power value health baseline of each group of data based on the relationship between the input sample and the output label to determine the normal range of power values of different working modes and wave channels.

[0100] In the specific embodiment, referring to Figure 3 , by reading the mode, wave channel, power value measured value and power control value in the data set, then grouping the data according to the mode and wave channel, finally cyclically calculating the 2.5% quantile of the power value of each group of data as the lower bound and the 97.5% quantile as the upper bound, combining the relationship between the input sample and the output label generated in step 4, constructing the power value health baseline of each group of data, and finally forming the dynamic health baseline of the power amplifier device and saving.

[0101] In some embodiments, each packet of data can be sorted by timestamp, and data with incorrect timestamps due to unsynchronized system time when the system is powered on can be deleted.

[0102] In summary, the application considers the system operation characteristics, time sequence relationship and other factors, analyzes the influence of the coupling characteristics of the system operation monitoring data on the parameter representation of the abnormal mode, constructs the electronic equipment abnormal detection data set and health evaluation data set according to different types of parameters such as the state of the power amplifier type device, the control parameter, the temperature, the power value and the module monitoring result, can learn an effective classification model under the condition of a small amount of fault samples, realizes accurate detection of faults and accurate evaluation of the health state of the equipment, and provides effective data support for performance monitoring, state evaluation and fault prediction.

[0103] Please refer to Figure 5 , Figure 5 A structure block diagram of a multi-parameter fusion health baseline construction device of an ultrashort wave power amplifier assembly provided by the application, comprising: an acquisition module 310, a determination module 320, a screening module 330, an extraction module 340 and a generation module 350, wherein:

[0104] The acquisition module 310 is used for acquiring operation data of multiple groups of power amplifier type devices, wherein each group of operation data records is generated by a preset flight number.

[0105] The determination module 320 is used for determining the operation principle and the fault mechanism of the power amplifier type device of each group based on each group of operation data.

[0106] The screening module 330 is used for establishing a screening model, and determining the input data label and the output data label of the screening model according to the operation principle and the fault mechanism.

[0107] The extraction module 340 is used for extracting a target data set from the operation data based on the input data label and the output data label.

[0108] The generation module 350 is used for aligning and merging the target data set by using a time stamp, to generate a multi-parameter fusion health baseline of the power amplifier type device and construct a data set.

[0109] It should be noted that the device embodiments in the application correspond to the foregoing method embodiments, and the specific principles in the device embodiments can be referred to the contents in the foregoing method embodiments, which will not be described here.

[0110] In several embodiments provided in the embodiment, the coupling between the modules can be electrical, mechanical or other forms of coupling.

[0111] In addition, each functional module in each embodiment of the application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software functional module.

[0112] The application further provides an electronic device capable of executing the method for constructing a multi-parameter fusion health baseline of an ultrashort wave power amplifier assembly.

[0113] The electronic device further includes a processor and a memory. The memory stores a program capable of executing the content of the foregoing embodiments, and the processor can execute the program stored in the memory.

[0114] The application further provides a computer readable storage medium. The computer readable storage medium stores program codes, and the program codes can be invoked by a processor to execute the method described in the foregoing method embodiments.

[0115] The application further provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method for constructing a multi-parameter fusion health baseline of an ultrashort wave power amplifier assembly described in the various optional implementation manners.

[0116] The above embodiments are only used to illustrate the technical solutions of the present application, rather than limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for constructing a health baseline of a multi-parameter fusion of an ultrashort wave power amplifier assembly, characterized in that, The method comprises the following steps: Obtain operation data of multiple groups of power amplifier devices, wherein each group of operation data records are generated by a preset flight number; Determine the operation principle and failure mechanism of the power amplifier devices in each group based on each group of operation data; Establish a screening model, and determine the input data label and output data label of the screening model according to the operation principle and failure mechanism; Extract a target data set from the operation data based on the input data label and output data label; Align and merge the target data set using a timestamp, including: merging the function record data in the target data set and reading PTT data, obtaining a first timestamp of any PTT data; based on the first timestamp and PTT data, find the corresponding working mode and wave channel number; data merge the PTT data, working mode and wave channel number; determine the monitoring data in the target data set, and obtain a second timestamp of the monitoring data; find any PTT data in a preset time period before the second timestamp, and select the target PTT data closest to the second timestamp; data merge the monitoring data, target PTT data and function record data to generate a multi-parameter fusion health baseline of the power amplifier device and construct a data set.

2. The method of claim 1, wherein the method further comprises: The method for constructing a multi-parameter fusion health baseline of the ultrashort wave power amplifier assembly further comprises: Divide the data set into a test set and a training set, and divide it into a support set and a query set; Randomly select N categories from all categories in the data set; According to the selected category sample, K samples are randomly selected to form a support set, and X samples are selected from the remaining samples of the selected category to form a query set, wherein the support set has samples, and the query set has samples.

3. The method of claim 2, wherein the method further comprises: The method for constructing a multi-parameter fusion health baseline of the ultrashort wave power amplifier assembly further comprises: Extract features from each sample in the support set to obtain the encoding of each sample; Sum and average the encoding of all samples under each category in the support set to obtain the prototype representation of each label category; When a data sample is input, use the encoding of the data sample to generate a corresponding encoding representation; Calculate the distance between the encoding representation of the new data sample and each category prototype representation based on the Euclidean distance; Convert the distance into a probability based on an activation function to determine the probability that the data sample belongs to the corresponding output label, and calculate the category with the highest output probability.

4. The method of claim 1, wherein the method further comprises: The method for constructing a multi-parameter fusion health baseline of the ultrashort wave power amplifier assembly further comprises: Determine the normal range of power values of different working modes and wave channels in the data set using a quantile-based statistical method.

5. The method of claim 4, wherein the method further comprises: The method for determining the normal range of power values of different working modes and wave channels in the data set using a quantile-based statistical method comprises: Read the related parameters in the data set; Group the related parameters according to the working mode and wave channel number, and calculate the quantile of the power value corresponding to each group of related parameters; Determine the upper and lower bounds of the quantile, and construct the power value health baseline of each group of data based on the relationship between the input sample and the output label to determine the normal range of power values of different working modes and wave channels.

6. An ultrashort wave power amplifier assembly multi-parameter fusion health baseline construction device, characterized in that, The device comprises: An acquisition module for obtaining operation data of multiple groups of power amplifier devices, wherein each group of operation data records are generated by a preset flight number; The determining module is configured to determine the operation principle and the failure mechanism of the power amplifier type device of each group based on the operation data of each group. The screening module is configured to establish a screening model, and determine input data tags and output data tags of the screening model according to the operation principle and the failure mechanism. The extracting module is configured to extract a target data set from the operation data based on the input data tags and the output data tags. The generating module is configured to align and merge the target data set by using a time stamp, including: merging function record data in the target data set and reading PTT data, obtaining a first time stamp of any one of the PTT data; finding corresponding working modes and wave channel numbers based on the first time stamp and the PTT data; merging the PTT data, the working modes and the wave channel numbers; determining monitoring data in the target data set, and obtaining a second time stamp of the monitoring data; finding any PTT data in a preset time period before the second time stamp, and selecting target PTT data closest to the second time stamp; and merging the monitoring data, the target PTT data and the function record data to generate a multi-parameter fusion health baseline of the power amplifier type device and construct a data set.

7. An electronic device, comprising: The electronic device includes a memory and a processor, and the memory stores program code executable on the processor. When the program code is executed by the processor, the program code implements the multi-parameter fusion health baseline construction method of the ultrashort wave power amplifier assembly according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores program code, and the program code can be called and executed by one or more processors to implement the multi-parameter fusion health baseline construction method of the ultrashort wave power amplifier assembly according to any one of claims 1-5.

Images