Engine test bed fault prediction method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a test bed fault prediction method and device of an engine, electronic equipment and a storage medium, and relates to the technical field of computers. The method comprises the steps of obtaining a first data set; analyzing the first data set based on the value range of each parameter in the first data set to obtain an analysis result of the first data set; and if the analysis result represents that the first data set has the abnormal parameter value, inputting the first data set into the fault prediction model, and outputting predicted fault information. The parameter values collected on the sampling points are combined into a first data set, the first data set is analyzed, whether abnormal data exist in the first data set or not is judged, if yes, the first data set is input into the trained fault prediction model, and a fault prediction result of the test bed is output. According to the scheme, the fault of the test bed does not need to be detected manually, manpower can be saved, meanwhile, the analysis standard of abnormal data is unified, and high precision is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular, the present application relates to an engine fault prediction method and device, electronic equipment and storage medium. BACKGROUND

[0002] In the engine test, accurate analysis of test data is to determine whether the test bench system has a fault, which is a key means to ensure the safety of the test run. At present, the processing of engine test data mainly relies on manual operation.

[0003] However, due to the large amount of data, it is difficult to analyze the data of a large number of measuring points on the screen manually during each engine test, and at present, only the upper and lower limits can be set to highlight the display to alarm or the fault can be judged by the personal experience of the staff. However, different experience levels of staff may draw different fault judgment results for the same abnormal data, and it is impossible to predict the occurrence of fault through the trend of data or the correlation change of data. SUMMARY

[0004] The purpose of the present application is to at least solve one of the above technical defects, and the technical scheme provided by the embodiments of the present application is as follows: In a first aspect, the embodiments of the present application provide an engine test bench fault prediction method, comprising: obtaining a first data set, the first data set comprising a plurality of first data subsets, each first data subset corresponding to a unique parameter, the first data subset comprising parameter values of the corresponding parameter at a plurality of first time points after the current start of the engine, the parameter value of each parameter being obtained by collecting the corresponding sampling point located on the engine or the test bench where the engine is located, and each parameter corresponding to a unique sampling point; analyzing the first data set based on the value range of each parameter in the first data set to obtain an analysis result of the first data set, the analysis result representing whether there is abnormal real-time data in the first data set; If the analysis result represents that there is abnormal parameter value in the first data set, input the first data set into a fault prediction model to output the predicted fault information of the test bench.

[0005] In a second aspect, the embodiments of the present application provide an engine test bench fault prediction device, comprising: a data set acquisition module for acquiring a first data set, the first data set comprising a plurality of first data subsets, each first data subset corresponding to a unique parameter, the first data subset comprising parameter values of the corresponding parameter at a plurality of first time points after the current start of the engine, the parameter value of each parameter being obtained by collecting the corresponding sampling point located on the engine or the test bench where the engine is located, and each parameter corresponding to a unique sampling point; a data set analysis module, configured to analyze the first data set based on the value range of each parameter in the first data set, and obtain an analysis result of the first data set, wherein the analysis result represents whether there is real-time data of abnormal parameters in the first data set; a fault prediction module, configured to input the first data set into a fault prediction model if the analysis result represents that there is an abnormal parameter value in the first data set, and output predicted fault information of the test bench.

[0006] In a third aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory; The processor executes the computer program to implement the method provided in the first aspect or any optional embodiment of the first aspect.

[0007] In a fourth aspect, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. When the computer program is executed by a processor, the method provided in the first aspect or any optional embodiment of the first aspect is implemented.

[0008] The technical scheme provided in the embodiments of the present application has the following beneficial effects: The scheme provided in the embodiments of the present application combines the parameter values collected at multiple sampling points into a first data set, analyzes the first data set, determines whether there is abnormal data in the first data set, inputs the first data set into a trained fault prediction model if there is abnormal data, and outputs the fault prediction result of the test bench. The above scheme does not need manual detection of the test bench fault, can save manpower, and has a relatively unified analysis standard for abnormal data and a relatively high precision. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical scheme in the embodiments of the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced.

[0010] Figure 1 A flowchart of a test bench fault prediction method of an engine provided in the embodiments of the present application; Figure 2 A flowchart of a test bench fault prediction method of an engine in an example of the embodiments of the present application; Figure 3 A flowchart of a training method of a fault prediction model in an example of the embodiments of the present application; Figure 4 A flowchart of a data analysis tool training method in an example of the embodiments of the present application; Figure 5 A block diagram of a test bench fault prediction device of an engine provided in the embodiments of the present application; Figure 6 Fig. 1 is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0011] Embodiments of the present application will be described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions of the technical solutions of the embodiments of the present application, and do not limit the technical solutions of the embodiments of the present application.

[0012] Those skilled in the art can understand that the singular forms "a", "an" and "the" used herein include plural forms unless specifically stated otherwise. It should be further understood that the terms "comprise" and "include" used in the embodiments of the present application mean that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements and / or components, but do not exclude other features, information, data, steps, operations, elements, components and / or combinations thereof supported by the present technology. It should be understood that when we say that an element is "connected" or "coupled" to another element, the element can be directly connected or coupled to the other element, or it can mean that the element and the other element are connected through an intermediate element. In addition, "connected" or "coupled" used herein can include wireless connection or wireless coupling. The term "and / or" used herein means that at least one of the items defined by the term, for example, "A and / or B" can be implemented as "A", or as "B", or as "A and B".

[0013] To make the purposes, technical solutions and advantages of the present application clearer, the embodiments of the present application will be described in further detail below with reference to the accompanying drawings.

[0014] The technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application will be described below through the description of several exemplary embodiments. It should be pointed out that the following embodiments can be mutually referenced, borrowed or combined, and the same terms, similar features and similar implementation steps in different embodiments will not be described repeatedly.

[0015] Figure 1 Fig. 1 is a structural schematic diagram of an electronic device according to an embodiment of the present application. Figure 1 As shown in Fig. 1, the method can include: In step S101, a first data set is acquired, the first data set comprising a plurality of first data subsets, each first data subset corresponding to a unique parameter, the first data subset comprising parameter values of the corresponding parameter at a plurality of first time points after the current start of the engine, the parameter value of each parameter being obtained by collecting the corresponding sampling point on the engine or the test bench on which the engine is located, and each parameter corresponding to a unique sampling point.

[0016] In the embodiments of the present application, the engine can be a carrier that can be used for spacecraft propulsion flight, and the spacecraft has the ability to fly by igniting the propellant in the engine to generate thrust. The test bench can be a platform for testing the performance of the engine. Each engine needs to be tested a lot before it is put into use. Only the engine that passes the test can be formally put into use. The sampling point can be on the test bench or the engine, such as the tail of the engine, the connection between the test bench and the engine, the test bench pipeline, the engine pipeline, the control gas pipeline, and the like, which are not limited in the embodiments of the present application. The parameter can be an index that needs to be sampled at the sampling point, such as engine liquid oxygen inlet pressure, fuel inlet pressure, engine liquid oxygen inlet temperature before starting, propellant flow, control gas pressure, and blow-off gas pressure. Different sampling points can collect parameter values of different parameters through different types of sensors. Optionally, one or more sensors can be arranged at one sampling point to collect parameter values of different parameters, which are not limited in the embodiments of the present application. The parameter value is the specific data value collected by the sensor. One or more different sensors can be installed at each sampling point to simultaneously collect parameter values of one or more parameters.

[0017] Specifically, since the engine needs to ensure the accuracy of the test bench during the test process, it is necessary to determine whether the test bench has a fault in real time. In order to determine the fault of the test bench, the parameter values of each parameter during the test are acquired, and whether the test bench has a fault is determined by judging whether the parameter values are normal. Therefore, in the embodiments of the present application, sensors are arranged at each meaningful sampling point on the test bench and the engine, and each first time point that needs to be sampled is set, and the parameter values of each parameter are continuously collected in real time from the start of the engine in each test. After the parameter values of each parameter are collected, the parameter values are taken as the first data set for analysis.

[0018] It should be noted that different sampling frequencies can be set for different sampling points in the embodiments of the present application. For example, for a parameter such as temperature, the change in a short time is not particularly obvious, and therefore the sampling frequency of the parameter can be set slightly lower than that of other parameters to save more resources. For example, pressure values, flow values, temperature values, etc. are sampled at a sampling rate of 1 kHz, and vibration, pulsating pressure, etc. are sampled at a sampling rate of 25 kHz. The embodiments of the present application are not limited herein.

[0019] It can be understood that, since the test bench may fail due to partial reasons during the test process, the first data set can be automatically generated and analyzed every preset period in the embodiments of the present application, so that the fault state of the test bench can be analyzed in real time during the test process, and a large amount of loss caused by the failure of the test bench can be avoided.

[0020] In step S102, the first data set is analyzed based on the value range of each parameter in the first data set, and an analysis result of the first data set is obtained. The analysis result represents whether there is abnormal real-time data in the first data set.

[0021] In the embodiments of the present application, the real-time data is the parameter value of each parameter collected in the test.

[0022] Specifically, the fault analysis in the embodiments of the present application can be completed by a pre-prepared AI (Artificial Intelligence) tool. The AI tool can include a data analysis tool and a fault prediction model. As shown in Figure 2 When the parameter value of the parameter is collected and the first data set is generated, the first data set can be preliminarily analyzed by the data analysis tool first, and it is determined whether there is abnormal real-time data in the first data set. If the analysis result represents that there is abnormal real-time data, the first data set is input into the fault prediction model for further fault analysis and prediction. If the analysis result represents that there is no abnormal real-time data, it can be considered that the test bench is not currently in failure, and the analysis process can be directly ended.

[0023] It should be noted that in some cases, the abnormal real-time data in the first data set is not caused by the failure of the test bench itself, but is a result of the normal test process. For example, through analysis of a large amount of normal running data, it is found that the small engine liquid oxygen inlet pressure is likely to be caused by large propellant flow, resulting in large flow resistance. Therefore, the abnormal data needs to be further analyzed and judged by the fault prediction model.

[0024] In step S103, if the analysis result represents that the first data set has abnormal parameter values, the first data set is input into the fault prediction model, and the predicted fault information of the test bench is output.

[0025] In the embodiments of the present application, the fault prediction model can be a multilayer perceptron neural network model (MLP) containing an input layer, a hidden layer and an output layer, wherein the input layer includes a plurality of input nodes, the number of input nodes is the same as the number of parameters, and each input node corresponds to a parameter; the output layer includes a plurality of output nodes, the number of output nodes is the same as the number of predicted fault information, and each output node corresponds to a predicted fault information. The number of layers and the number of nodes of the hidden layer are determined through multiple tests and optimizations, for example, can be obtained according to an empirical formula:

[0026] wherein n h is the number of hidden layer nodes, n i is the number of input layer nodes (the number of measuring points), n o is the number of output layer nodes (the number of fault modes), and k is an adjustment constant of 1-10.

[0027] In the embodiments of the present application, a two-layer structure can be designed, the first layer processes a single parameter (time domain analysis of a single parameter), and the second layer processes combined parameter features (a plurality of abnormal data or a plurality of parameters collectively changing correspond to a certain fault mode). The predicted fault information can be a specific fault type, which is obtained by the fault prediction model.

[0028] Specifically, as shown in Figure 2 When it is found through data analysis that there is abnormal real-time data in the first data set, the first data set is input into the fault prediction model, the fault prediction model can further judge whether the abnormal real-time data is caused by the test bench fault according to the parameter values in the first data set, and finally output the predicted fault information. It can be understood that the predicted fault information can also include "no fault".

[0029] The scheme provided in the embodiments of the present application combines the parameter values collected at a plurality of sampling points into a first data set, analyzes the first data set, judges whether there is abnormal data in the first data set, if there is abnormal data, inputs the first data set into the trained fault prediction model, and outputs the fault prediction result of the test bench. The above scheme does not need manual detection of test bench faults, can save manpower, and has a relatively unified analysis standard for abnormal data and high precision.

[0030] On the basis of each of the above embodiments, as an optional embodiment, the first data set is analyzed based on the value range of each parameter in the first data set to obtain an analysis result of the first data set, specifically including: for each parameter value of each parameter, determining whether the parameter value is within the value range of the parameter; if the parameter values of all parameters are within the corresponding value ranges, determining that the analysis result of the first data set represents that the first data set does not have real-time data with abnormality; if the parameter value of at least one parameter is not within the corresponding value range, obtaining data relationship information of the at least one parameter, the data relationship information of the parameter representing an association relationship between the parameter value of the parameter and parameter values of other parameters; analyzing the first data set based on the data relationship information of the at least one parameter to obtain an analysis result.

[0031] In the embodiments of the present application, the value range of each parameter can represent the numerical value that the parameter should have in the normal state of the test bench. The data relationship information can represent the relationship between different parameters, which can be a functional relationship or a change trend (for example, when parameter A increases, parameter B should correspondingly decrease, etc.).

[0032] Specifically, when analyzing each parameter value, the data analysis tool of the embodiments of the present application first determines whether the parameter value is within the value range of the corresponding parameter. If it is, it means that the sampling point corresponding to the parameter value does not have abnormality. When the parameter values corresponding to all sampling points are within the corresponding value ranges, it can be considered that the test bench currently does not have a fault. If there is any parameter value that is not within the corresponding value range, it is considered that there is abnormal data in the first data set, and it can be considered that the test bench currently may have a fault. The abnormal data may be caused by a fault of the test bench, or it may be abnormal data that appears temporarily because of a specific test. At this time, the data relationship information between the parameter corresponding to the abnormal parameter value and other parameters needs to be obtained in real time, and the parameter value is analyzed based on the association relationship to further determine the analysis result.

[0033] It should be noted that the data relationship information in the embodiments of the present application can be pre-set by the user according to experience, or can be learned by the data analysis tool itself after the user inputs the abnormal data and the result of “no fault” into the data analysis tool in the case of abnormal data but no fault of the test bench. The embodiments of the present application do not limit this.

[0034] Based on the above various embodiments, as an optional embodiment, the first data set is analyzed based on the data relationship information of the at least one parameter to obtain an analysis result, specifically including: for each parameter value of each parameter, obtaining at least one first parameter having an association relationship with the parameter based on the data relationship information, and obtaining parameter values of each first parameter at the same time as first parameter values; For any parameter value of the parameter, if there is at least one first parameter value which does not satisfy the association relationship between the parameter value and the first parameter value, it is determined that the analysis result of the first data set represents that the first data set has abnormal real-time data; For each parameter value of the parameter, if each first parameter value satisfies the association relationship between the parameter value and each first parameter value, it is determined that the analysis result of the first data set represents that the first data set does not have abnormal real-time data.

[0035] In the embodiments of the present application, the first parameter can be another parameter which has an association relationship with the parameter that needs to be analyzed for abnormal data. For example, since the condition that "the engine liquid oxygen inlet pressure is small because the propellant flow is large, resulting in large flow resistance", "flow resistance" is the parameter of abnormal data, the first parameter which has an association relationship with it can be "liquid oxygen inlet pressure".

[0036] Specifically, when a certain parameter value exceeds the value range of the corresponding parameter, the other first parameters which have an association relationship with the parameter need to be obtained through the data relationship information, and the real-time data (i.e. the parameter value of the first parameter) corresponding to the first parameter at the same time is obtained, and then it is judged whether the real-time data of the first parameter and the parameter value of the parameter satisfy the association relationship between the two parameters. If yes, it is indicated that the abnormal data is not caused by the test bench failure. When all parameter values satisfy the association relationship between the parameter value of each first parameter, it is determined that the test bench does not have a failure, and the analysis result that the first data set does not have abnormal real-time data can be obtained. On the contrary, if there is only one parameter value which does not satisfy the association relationship between the parameter value of any one of the first parameters, it is considered that the abnormal real-time data is caused by the test bench failure, and the analysis result that the first data set has abnormal real-time data can be obtained.

[0037] On the basis of each of the above embodiments, as an optional embodiment, the method further comprises the step of predicting data at multiple future time points: obtaining a second data set, the second data set comprising multiple second data subsets, each second data subset corresponding to a unique parameter, and each second data subset comprising parameter values of the corresponding parameter at multiple second time points after a historical start of the engine; for each parameter, determining a first function image describing a rule of change of the parameter value of the parameter over time after the historical start based on the parameter values at the multiple second time points, and determining a second function image describing a rule of change of the parameter value of the parameter over time after the current start of the engine based on the parameter values at the multiple first time points; for each parameter, determining a first sub-image identical to the second function image from the first function image, and obtaining a second sub-image from the first function image which is in time sequence after the first sub-image. For each parameter, the parameter values of multiple future time instants are predicted according to the rule described by the second sub-image.

[0038] In the embodiments of the present application, each second time instant can be a historical time instant. After each engine test is completed, the terminal records the parameter values collected from each sampling point during the test and the corresponding time instants, which are the second time instants. The function images can be obtained by time series analysis algorithms. The first function image can be obtained by analyzing the parameter values in the second data set and the corresponding second time instants, and the second function image can be obtained by analyzing the parameter values in the first data set and the corresponding first time instants.

[0039] Specifically, although the real-time data collection at each sampling point can effectively and timely detect possible faults of the test bench, the above-mentioned method is based on the premise that the test bench has already failed, which still causes certain losses. Therefore, in order to minimize the losses, the embodiments of the present application also provide a method for predicting the data of each sampling point at future time instants. Specifically, the parameter value changes of each sampling point during each engine test can be sorted out by including the parameter values of each historical second time instant, and a first function image describing the rule of the parameter value change of each parameter over time can be generated. Each first function image can be regarded as the rule of the parameter value change of the corresponding parameter over time during a test. Since the test process is similar each time, the first function images of each parameter can be used as a reference for the change rule of each parameter during the test.

[0040] Then, from the start of the engine test, the parameter values of each parameter until the current time instant are obtained, and the second function image of each parameter is determined in the same way. It can be understood that since the first function image contains at least one complete engine test process, the second function image will have a corresponding same part in the first function image. This same part can be determined as the first sub-image, and the last second time instant in the first sub-image corresponds to the current time instant. Therefore, from the second sub-image after the second time instant, the rule of the parameter value change of each parameter at future time instants after the current time instant is described. Therefore, the second sub-image of each parameter can be used as a reference to predict the parameter values of multiple future time instants during the test. While obtaining the predicted parameter values of each parameter, the analysis method described above can also be used to predict the possible fault information of the test bench at future time instants. If possible fault information is predicted, corresponding measures can be taken in advance or corresponding warning information can be issued to avoid the occurrence of the fault, thereby reducing unnecessary losses.

[0041] It should be noted that there is a certain error between the parameter values of each parameter between each engine test, and there may be no part in the first function image that is completely the same as the second function image, so the part with the highest coincidence degree with the second function image can also be determined as the first sub-image.

[0042] On the basis of the above various embodiments, as an optional embodiment, the predicted fault information further includes a dangerous degree of the predicted fault information. If the dangerous degree is not less than a preset threshold, after outputting the predicted fault information of the test bench, the method further includes: Based on the predicted fault information, a corresponding countermeasure is matched from a preset list, and the countermeasure is executed; the preset list includes a corresponding relationship between the fault information and the countermeasure; If the dangerous degree is less than the preset threshold, after outputting the predicted fault information of the test bench, the method further includes: Based on the predicted fault information, a corresponding countermeasure is matched from a preset list, and the countermeasure is executed; the preset list includes a corresponding relationship between the fault information and the countermeasure;

[0043] In the embodiments of the present application, the dangerous degree can be represented by a series, such as "first level", "second level", "third level", etc., which is not limited in the embodiments of the present application. The preset list can be designed by the user in advance and stored in the terminal, which is used to instruct the terminal to take corresponding countermeasures for different fault information. The preset terminal is not the terminal described above, which is used to receive the alarm information sent by the terminal described above and display it to the user.

[0044] Specifically, the predicted fault information output by the fault prediction model in the embodiments of the present application further includes a dangerous degree corresponding to the fault. When the dangerous degree is low, the terminal can only match the corresponding countermeasure from the preset list according to the fault information output by the model, and then send the fault information and the countermeasure to the preset terminal in the form of alarm information, so that the preset terminal displays the alarm information. After the user sees the alarm information through the preset terminal, the user can choose to execute the countermeasure matched by the terminal from the preset list, or take other countermeasures, etc., which is not limited in the embodiments of the present application. When the dangerous degree is high, it can be considered that the fault is urgent, and further judgment by the user cannot be waited for, so the terminal can directly execute the countermeasure matched from the preset list (for example, real-time detection of four key parameters, as long as two of them are simultaneously out of limit (including upper limit and lower limit), emergency shutdown is directly executed), so as to solve the fault as soon as possible or avoid greater loss caused by the fault.

[0045] On the basis of each of the above embodiments, as an optional embodiment, the fault prediction model is trained by a plurality of training samples and a training label corresponding to each training sample; Each training sample includes a third data set, the third data set includes a plurality of third data subsets, each third data subset corresponds to a unique parameter, the third data subset includes parameter values of the corresponding parameter at a plurality of third time points, and the training label is real fault information of the test bench occurring within a preset time period. Each third time point is within the preset time period.

[0046] In the embodiments of the present application, the fault prediction model can be constructed based on a multilayer perceptron neural network model (MLP) in deep learning. The third time point can be a plurality of historical time points, and the training sample in the embodiment of the present application can be collected according to the test on the engine in the history. The real fault information in the training label can be the actual fault occurring in each corresponding test.

[0047] Specifically, an initial model is first constructed based on a multilayer perceptron neural network model in deep learning in the embodiment of the present application. Since the initial model has a large error in predicting fault information, it needs to be trained. In order to enable the model to have accurate fault prediction capability, training samples and training labels for training can be constructed according to historical test data, which can specifically include the following steps: For each training sample, the training sample is input into the initial model, and a fault prediction result corresponding to the training sample is output; Based on the real fault information corresponding to each training sample, a training loss of the initial model is determined, and each model parameter of the initial model is adjusted based on the training loss until a preset training termination condition is met, and a fault prediction model is obtained.

[0048] As shown in Figure 3 First, the initial model is first trained according to the third data set X to obtain a fault prediction model 1, and then the fault prediction model 1 is further trained by the third data set Y. The result output by the fault prediction model 1 is compared with the real fault information corresponding to the third data set Y, and the model parameters of the fault prediction model 1 are adjusted according to the comparison result to obtain a fault prediction model 2. Then the above process is repeatedly performed until the training termination condition is met, and a trained fault prediction model (i.e. fault prediction model N) is obtained.

[0049] The preset termination training condition in the embodiment of the present application can be that a certain number of training times (such as 50 times) are met, or the loss value between the output of the model and the training label in the training process is less than a threshold, and the like. The embodiment of the present application is not limited herein.

[0050] Optionally, as shown in Figure 4As shown, the data analysis tool in the embodiment of the present application can also determine the value range of each parameter and the correlation with other parameters in advance based on the data envelopment analysis algorithm. For example, the data unrelated to the fault in each test in the third training set is input into the data analysis tool, and the analysis result of whether there is abnormal data in the third training set is input into the data analysis tool. The data analysis tool can explore the correlation information between parameters according to the parameter values in the third training set and the analysis result, and determine the value range corresponding to each parameter.

[0051] On the basis of each of the above embodiments, as an optional embodiment, after outputting the preset fault result of the test bench, the method further specifically comprises: obtaining real fault information corresponding to the first data set; taking the first data set as a training sample and taking the real fault information as a corresponding training label; adjusting parameters of the fault prediction model based on the training sample and the training label.

[0052] Specifically, after each engine test is completed, the user can determine the real fault information occurring in the test process by means of artificial detection, and then sends the first data set as a training sample and the real fault information of the engine test to the current fault prediction model as a training label, so that the fault prediction model further adjusts the model parameters according to the test data, ensures that the model can continuously adapt to new test data and fault characteristics, and continuously improves the ability of fault analysis and judgment.

[0053] Figure 5 A structural block diagram of an engine test bench fault prediction device provided by the embodiment of the present application is shown in Figure 5 As shown, the engine test bench fault prediction device 500 can include a data set acquisition module 501, a data set analysis module 502, and a fault prediction module 503, wherein, The data set acquisition module 501 is configured to acquire a first data set, the first data set including a plurality of first data subsets, each first data subset corresponding to a unique parameter, the first data subset including parameter values of the corresponding parameter at a plurality of first time points after the engine is started this time, the parameter values of each parameter being obtained by collecting the corresponding sampling point located on the engine or the test bench where the engine is located, and each parameter corresponding to a unique sampling point. The data set analysis module 502 is configured to analyze the first data set based on the value range of each parameter in the first data set to obtain an analysis result of the first data set, the analysis result indicating whether there is abnormal real-time data in the first data set. The fault prediction module 503 is configured to input the first data set into the fault prediction model to output the predicted fault information of the test bench if the analysis result indicates that the first data set has abnormal parameter values.

[0054] The scheme provided by the embodiments of the present application combines the parameter values collected at multiple sampling points into a first data set, analyzes the first data set, determines whether there is abnormal data in the first data set, inputs the first data set into the trained fault prediction model if there is abnormal data, and outputs the fault prediction result of the test bench. The above scheme does not need manual detection of the test bench fault, can save manpower, and has a relatively unified analysis standard for abnormal data and high precision.

[0055] Based on the above embodiments, as an optional embodiment, the data set analysis module is specifically configured to: For the parameter value of each parameter, determine whether the parameter value is within the value range of the parameter; If the parameter values of the parameters are within the corresponding value ranges, it is determined that the analysis result of the first data set indicates that the first data set does not have abnormal real-time data; If the parameter value of at least one parameter is not within the corresponding value range, obtain data relationship information of the at least one parameter, and the data relationship information of the parameter indicates the association relationship between the parameter value of the parameter and the parameter values of other parameters; Based on the data relationship information of the at least one parameter, analyze the first data set to obtain an analysis result.

[0056] Based on the above embodiments, as an optional embodiment, the data set analysis module is further configured to: For the parameter value of each parameter, obtain at least one first parameter having an association relationship with the parameter based on the data relationship information, and obtain the parameter value of each first parameter at the same time as a first parameter value; For the parameter value of any one parameter, if at least one first parameter value does not satisfy the association relationship between the parameter value and the first parameter value, it is determined that the analysis result of the first data set indicates that the first data set has abnormal real-time data; For the parameter value of each parameter, if each first parameter value satisfies the association relationship between the parameter value and each first parameter value, it is determined that the analysis result of the first data set indicates that the first data set does not have abnormal real-time data.

[0057] Based on the above embodiments, as an optional embodiment, the device further includes a data prediction module, which is specifically configured to: obtaining a second data set, the second data set comprising a plurality of second data subsets, each second data subset corresponding to a unique parameter, the second data subset comprising parameter values of the corresponding parameter at a plurality of second time points after a historical start of the engine; For each parameter, based on the parameter values at the plurality of second time points, determining a first function image describing a rule of variation of the parameter values of the parameter over time after the historical start, and based on the parameter values at the plurality of first time points, determining a second function image describing a rule of variation of the parameter values of the parameter over time after the current start of the engine; For each parameter, determining a first sub-image identical to the second function image from the first function image, and obtaining a second sub-image from the first function image, the second sub-image being chronologically after the first sub-image; For each parameter, predicting parameter values at a plurality of future time points according to the rule described by the second sub-image.

[0058] On the basis of each of the above embodiments, as an optional embodiment, the predicted fault information further comprises a risk degree of the predicted fault information; The device further comprises a fault handling module, specifically configured to: Based on the predicted fault information, matching a corresponding countermeasure from a preset list, and executing the countermeasure; the preset list comprising a corresponding relationship between the fault information and the countermeasure; If the risk degree is less than a preset threshold, after outputting the predicted fault information of the test bench, the method further comprises: Based on the predicted fault information, matching a corresponding countermeasure from a preset list, and sending an alarm information to a preset terminal, the alarm information comprising the predicted fault information and the countermeasure corresponding to the predicted fault information.

[0059] On the basis of each of the above embodiments, as an optional embodiment, the fault prediction model is trained by a plurality of training samples and a training label corresponding to each training sample; Each training sample comprises a third data set, the third data set comprising a plurality of third data subsets, each third data subset corresponding to a unique parameter, the third data subset comprising parameter values of the corresponding parameter at a plurality of third time points, and the training label being real fault information of the test bench occurring within a preset period; each third time point being within the preset period.

[0060] On the basis of each of the above embodiments, as an optional embodiment, the device further comprises a model training module, specifically configured to: Obtaining real fault information corresponding to the first data set; Taking the first data set as a training sample, and taking the real fault information as a corresponding training label; Adjusting parameters of the fault prediction model based on the training sample and the training label.

[0061] Further, compared with related art, the beneficial effects of the embodiments of the present application are: 1. Improve fault judgment accuracy: Traditional manual analysis methods are affected by the difference in the experience level of staff, and different fault judgment results may be obtained for the same abnormal data. However, the AI tool of the present application can scientifically judge the fault according to the data trend and correlation change by using the powerful data analysis capability of the AI tool and constructing a precise abnormal data and fault mode correlation model through learning a large number of normal and abnormal data samples.

[0062] 2. Realize scientific fault prediction: The existing method relying on personnel experience cannot scientifically predict the occurrence of faults through data trends or correlation changes, while the AI tool of the present application can analyze the historical trend of real-time data with the help of time series analysis algorithms (such as ARIMA). When the engine liquid oxygen inlet pressure appears a sustained rise and deviates from the normal envelope in relation to the propellant flow, the AI tool can predict whether it will further deteriorate and the time node of possible fault occurrence. This enables test personnel to make preparations in advance, take preventive measures, avoid the occurrence of faults or reduce the loss caused by faults, and changes the passive situation of being able to deal with only after the fault occurs.

[0063] 3. Reduce cost and speed up development progress: Faults often lead to engine and test bench damage, and the repair process requires a large amount of material and labor costs, and also slows down the engine development progress. The present application can effectively avoid or reduce the occurrence of faults through accurate fault judgment and scientific fault prediction, thereby reducing the probability of engine and test bench damage.

[0064] 4. Continuous optimization and adaptability: The AI tool of the present application has the ability of relearning and model updating. After each test, the test data and fault mode are input to correct the intelligent fault analysis model. With the increase of test times and the appearance of new fault modes, the model can continuously adjust parameters, optimize structure, and add new output nodes to adapt to new fault conditions. This enables the fault analysis and judgment capability of the AI tool to continuously improve, always maintaining the advanced nature and effectiveness of the fault analysis of the engine test bench, and better coping with increasingly complex engine test scenarios.

[0065] 5. Improve test efficiency and safety: During the test, the AI tool can process massive test data in real time and quickly give fault analysis and prediction results to provide decision support for test personnel. Test personnel can make timely and accurate decisions based on these results, such as timely emergency shutdown, to avoid more serious consequences due to untimely judgment and decision making when a fault occurs, thereby ensuring the safety of the test process.

[0066] Reference will now be made to Figure 6which shows a structural diagram of an electronic device (e.g., a terminal device or a server) 600 adapted to implement the methods of the present application. The electronic device in the present application can include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (Personal Digital Assistant), a PAD (Tablet Personal Computer), a PMP (Portable Multimedia Player), a car terminal (e.g., a car navigation terminal), a wearable device, and the like, as well as a stationary terminal such as a digital TV, a desktop computer, and the like. Figure 1 The electronic device shown is merely an example and should not impose any limitation on the functions and use range of the embodiments of the present application. Figure 6 The electronic device shown is merely an example and should not impose any limitation on the functions and use range of the embodiments of the present application.

[0067] The electronic device includes a memory for storing a program for executing the methods described in the above various method embodiments and a processor configured to execute the program stored in the memory. Here, the processor can be referred to as a processing means 601 described below, and the memory can include at least one of a read-only memory (ROM) 602, a random access memory (RAM) 603, and a storage device 608, as shown below: As shown in the figure, the electronic device 600 can include a processing means (e.g., a central processing unit, a graphic processing unit, etc.) 601 that can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing means 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604. Figure 6 As shown in the figure, the electronic device 600 can include a processing means (e.g., a central processing unit, a graphic processing unit, etc.) 601 that can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing means 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0068] Generally, the following devices can be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, and the like; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, and the like; a storage device 608 including, for example, a magnetic tape, a hard disk, and the like; and a communication device 609. The communication device 609 can allow the electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 The electronic device is shown with various devices, but it is understood that all of the shown devices are not required to be implemented or provided. More or less devices can alternatively be implemented or provided.

[0069] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program in accordance with embodiments of the present application. For example, embodiments of the present application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program comprising program code for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-described functions defined in the methods of embodiments of the present application are performed.

[0070] It should be noted that the computer-readable storage medium described above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus. In the present application, the computer-readable signal medium can include a data signal carried in a baseband or as a part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take many forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or apparatus. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to a wire, an optical fiber, an RF (radio frequency) or the like, or any suitable combination of the above.

[0071] In some embodiments, the client, server, or other computing machines can communicate using any known or later developed form of computer-readable data signals over any type of transmission medium when communicating data. Such computer-readable data signals can be of the type generally known to those skilled in the art, that can be generated and transmitted, for example, by modulating a carrier wave or other signal, over a communication channel. The communication channel can be for example a communication network, such as the Internet, Intranet, Extranet, or the like. The network can be wired or wireless. Examples of wireless networks include IEEE 802.11, 802.16, 802.20, cellular networks, Bluetooth, and the like. Examples of wired networks include Ethernet, Proximity network, and the like. The network can be any combination of wired and / or wireless networks. The network can also be protected networks obtained through firewalls or other security measures.

[0072] The computer-readable medium described above can be included in the electronic device described above; alternatively, the computer-readable medium can exist as a standalone entity.

[0073] The computer-readable medium described above can be included in the electronic device described above; alternatively, the computer-readable medium can exist as a standalone entity. The computer-readable medium described above can be included in the electronic device described above; alternatively, the computer-readable medium can exist as a standalone entity.

[0074] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network ("LAN") or a wide area network ("WAN"), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, application specific circuitry, or field programmable gate array ("FPGA") circuitry can execute the computer program code.

[0075] The computer program product of the first aspect of the present application can include a computer readable storage medium. The computer readable storage medium can be a tangible medium that can contain or store the program for use by or in connection with the instruction execution system, apparatus, or device. The computer readable storage medium can be a machine-readable signal medium or a machine-readable storage medium.

[0076] The modules or units described in the embodiments of the present application can be implemented by software, or by hardware, or by a combination of software and hardware. In some cases, the names of the modules or units do not limit the functions of the modules or units themselves. For example, the first constraint obtaining module can also be referred to as a module that obtains a first constraint.

[0077] The functions described in the above description can be performed by one or more hardware logic components. For example, non-limiting examples of hardware logic components that can be used include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SOCs), complex programmable logic devices (CPLDs), etc.

[0078] In the context of the present application, a machine-readable medium can be a tangible medium that can contain or store the program for use by or in connection with the instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the above. More specific examples of the machine-readable storage medium will include one or more of an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0079] It should be understood that although the steps in the flowcharts of the drawings are shown in sequence according to the indication of the arrows, the steps are not necessarily executed in sequence according to the indication of the arrows. Unless explicitly stated herein, the execution of the steps is not strictly limited in sequence, and the steps can be executed in other sequences. Moreover, at least part of the steps in the flowcharts of the drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0080] The above only describes some embodiments of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method for predicting engine test bench failures, characterized in that, include: Obtain a first dataset, which includes multiple first data subsets, each first data subset corresponding to a unique parameter. The first data subset includes the parameter values ​​of the corresponding parameter at multiple first moments after the engine starts this time. The parameter value of each parameter is obtained by collecting data from the corresponding sampling point located on the engine or the test bench on which the engine is located. Each parameter corresponds to a unique sampling point. The first dataset is analyzed based on the value range of each parameter in the first dataset to obtain the analysis results of the first dataset; The analysis results indicate whether there is any abnormal real-time data in the first dataset; If the analysis results indicate that there are abnormal parameter values ​​in the first dataset, then the first dataset is input into the fault prediction model, and the predicted fault information of the test bench is output.

2. The method according to claim 1, characterized in that, The analysis of the first dataset based on the value range of each parameter in the first dataset to obtain the analysis results of the first dataset includes: For each parameter value, determine whether the parameter value is within the range of the parameter's value; If the values ​​of each parameter are all within their corresponding ranges, then the analysis results of the first dataset indicate that there is no abnormal real-time data in the first dataset. If at least one parameter's value is not within its corresponding range, then the data relationship information of the at least one parameter is obtained. The data relationship information of the parameter represents the association between the parameter's value and the parameter values ​​of other parameters. The first dataset is analyzed based on the data relationship information of at least one parameter to obtain the analysis results.

3. The method according to claim 2, characterized in that, The analysis of the first dataset based on the data relationship information of at least one parameter to obtain the analysis results includes: For each parameter value, at least one first parameter that is related to the parameter is obtained based on the data relationship information, and the parameter value of each first parameter at the same time is obtained as the first parameter value. For any parameter value, if there exists at least one first parameter value that does not satisfy the correlation between the parameter value and the first parameter value, then the analysis result of the first dataset indicates that there is abnormal real-time data in the first dataset. For each parameter value, if each first parameter value satisfies the correlation relationship between the parameter value and each first parameter value, then the analysis result of the first dataset indicates that there is no abnormal real-time data in the first dataset.

4. The method according to claim 2, characterized in that, The method also includes the step of predicting data from multiple future times: Obtain a second dataset, which includes multiple second data subsets, each of which corresponds to a unique parameter. The second data subset includes the parameter values ​​of the corresponding parameter at multiple second moments after the engine's historical start-up. For each parameter, based on the parameter values ​​at the plurality of second moments, a first function graph describing the pattern of the parameter value changing over time after a historical start is determined, and based on the parameter values ​​at the plurality of first moments, a second function graph describing the pattern of the parameter value changing over time after the current engine start is determined; For each parameter, a first sub-image that is the same as the second function image is determined from the first function image, and a second sub-image that is temporally following the first sub-image is obtained from the first function image; For each parameter, the parameter value at the plurality of future moments is predicted based on the pattern described by the second sub-image.

5. The method according to claim 1, characterized in that, The predicted fault information also includes the degree of danger of the predicted fault information; If the degree of danger is not less than a preset threshold, after outputting the predicted fault information of the test bench, the method further includes: Based on the predicted fault information, corresponding countermeasures are matched from a preset list and executed; the preset list includes the correspondence between fault information and countermeasures. If the degree of danger is less than a preset threshold, after outputting the predicted fault information of the test bench, the method further includes: Based on the predicted fault information, corresponding countermeasures are matched from the preset list, and alarm information is sent to the preset terminal. The alarm information includes the predicted fault information and the corresponding countermeasures.

6. The method according to claim 1, characterized in that, The fault prediction model is trained using multiple training samples and the training labels corresponding to each training sample. Each training sample includes a third dataset, which includes multiple third data subsets. Each third data subset corresponds to a unique parameter. The third data subset includes the parameter values ​​of the corresponding parameter at multiple third time points. The training label is the real fault information that occurred on the test bench within a preset time period. Each third time point is within the preset time period.

7. The method according to claim 1, characterized in that, After outputting the preset fault result of the test bench, the method further includes: Obtain the actual fault information corresponding to the first dataset; The first dataset is used as the training sample, and the real fault information is used as the corresponding training label. The parameters of the fault prediction model are adjusted based on the training samples and the training labels.

8. A test bench fault prediction device for an engine, characterized in that, include: The dataset acquisition module is used to acquire a first dataset, which includes multiple first data subsets. Each first data subset corresponds to a unique parameter. The first data subset includes the parameter values ​​of the corresponding parameter at multiple first moments after the engine starts this time. The parameter value of each parameter is obtained by collecting data from the corresponding sampling point located on the engine or the test bench on which the engine is located. Each parameter corresponds to a unique sampling point. The dataset analysis module is used to analyze the first dataset based on the value range of each parameter in the first dataset, and obtain the analysis results of the first dataset. The analysis results indicate whether there is any abnormal real-time data in the first dataset; The fault prediction module is used to input the first dataset into the fault prediction model and output the predicted fault information of the test bench if the analysis results indicate that there are abnormal parameter values ​​in the first dataset.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the method of any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1-7.

Images