Aero-engine assembly process optimization method based on assembly data

Abstract

The invention discloses an aero-engine assembly process optimization method based on assembly data, and the method comprises the steps: collecting the assembly data of all unit bodies in an aero-engine when an aero-engine test passes, and carrying out the screening of assembly indexes of the assembly data of the unit bodies; calculating correlation among the assembly indexes after unit screening, and constructing a unit assembly network model according to a complex network theory; introducing a network dynamics theory, and performing neural network dynamics model fitting on the unit assembly network model to obtain a unit dynamics model; on the basis of calculation of the unit dynamical model, the fitting error of each assembly index is determined; on the basis of fitting errors of the assembly indexes, a unit evaluation interval is constructed by means of the average field theory; for a to-be-evaluated unit body, determining an evaluation index of the unit body by using the assembly index of the unit body; and judging whether the to-be-evaluated unit body is qualified or not based on the relationship between the evaluation index and the unit body evaluation interval.

Description

Technical Field

[0001] The invention relates to the field of aircraft engine assembly process optimization, and in particular to an aircraft engine assembly process optimization method based on assembly data. Background Art

[0002] Aircraft engines have the characteristics of high technology concentration and content, complex system sets, high added value, and great driving force. As a complex and high-precision thermal mechanical system, aircraft engines have many components and complex assembly processes. The assembly process of aircraft engines can directly affect its performance, reliability and service life. The assembly efficiency and quality can be improved through assembly-testing linkage, and resource allocation can be further optimized.

[0003] The assembly process and test run process of aircraft engines are two crucial processes in the manufacturing process. The assembly of aircraft engines directly affects the performance, reliability and service life of aircraft engines. After assembly, aircraft engines must undergo rigorous test runs before they can be put into actual use. However, in the manufacturing process, there are often some aircraft engines whose parts and components meet the design requirements in their manufacturing dimensions, and whose corresponding assembly work is strictly carried out in accordance with regulations, but fail to pass the test run procedures. Therefore, the entire machine has to be disassembled, reassembled, and retested until it meets the test run requirements, which ultimately increases the production costs of the enterprise.

[0004] In recent years, aero-engine manufacturers have accumulated a large amount of aero-engine assembly data and test results. In theory, the assembly and test of aero-engines should be complementary processes; however, in actual engineering applications, assembly data and test results are still in an artificial "separation" state, resulting in a certain degree of blindness in the current optimization of aero-engine assembly processes. Secondly, assembly data includes assembly process parameters of each component, pre- and post-assembly inspection data, error records during the assembly process, and various correction operations. The data is highly dimensional and heterogeneous, and the correlation between indicators is strong, making it difficult to accurately guide assembly process optimization and improve assembly efficiency and quality. Finally, in actual engineering, aero-engines mostly adopt a unit structure design to achieve the interchangeability of units and improve the maintainability of aero-engines. However, the existing technology cannot effectively evaluate at the unit level. It is necessary to wait until the whole machine is assembled and then test it. Then, according to the test results, it is determined whether the whole machine needs to be disassembled or delivered. This greatly limits the assembly efficiency and quality of aero-engines. Therefore, in view of the above defects, it is an urgent problem to be solved in the field of aero-engine manufacturing to study how to use assembly data and test data in actual engineering to improve the assembly efficiency and test pass rate of enterprises.

[0005] The prior art discloses a digital twin model of aircraft engine performance that integrates assembly data and a method for establishing the model (a patent application with publication number CN118378141A): first, collect assembly test data of different test procedures of multiple aircraft engines of the same model, and determine the sensor monitoring data and assembly data related to the target performance parameters; second, construct a performance digital twin model that integrates assembly data, and train the performance digital twin model using sensor monitoring data and assembly data; finally, calculate the prediction accuracy of the performance digital twin model after training, and if the prediction accuracy meets the requirements, then end; if the prediction accuracy does not meet the requirements, then retrain the performance digital twin model. However, the shortcomings of this method are: this method uses a multi-head attention mechanism to learn component weight coefficients and establishes a performance digital twin model through a recurrent neural network, which means that the result generation mechanism of the model cannot be clearly defined; in addition, this method can only predict the performance parameters of aircraft engine tests, and cannot provide an effective guidance scheme for the actual assembly of aircraft engines. Summary of the invention

[0006] The purpose of the present invention is to provide an aircraft engine assembly process optimization method based on assembly data, to fit a unit body dynamics model based on a neural network dynamics model to evaluate the unit body state so as to provide a practical solution for aircraft engine assembly and improve aircraft engine assembly efficiency and test pass rate.

[0007] In order to achieve the above tasks, the present invention adopts the following technical solutions:

[0008] The aircraft engine assembly process optimization method based on assembly data includes:

[0009] When the aircraft engine passes the test run, the assembly data of each unit in the aircraft engine is collected, and the assembly index of the unit assembly data is screened;

[0010] Calculate the correlation between the assembly indicators after unit screening, and build a unit assembly network model based on complex network theory;

[0011] Introducing the network dynamics theory, the neural network dynamics model is fitted to the unit body assembly network model to obtain the unit body dynamics model; based on the solution of the unit body dynamics model, the fitting error of each assembly index is determined;

[0012] Based on the fitting error of the assembly index, the unit body evaluation interval is constructed using the mean field theory; for the unit body to be evaluated, the evaluation index of the unit body is determined using the assembly index of the unit body; based on the relationship between the evaluation index and the unit body evaluation interval, it is determined whether the unit body to be evaluated is qualified.

[0013] Furthermore, the method further comprises:

[0014] The evaluation index of the unit body is calculated by using the assembly index of the unit body after screening when the aircraft engine test passes or fails, and the evaluation index of the unit body is divided into state intervals. A Bayesian network model is constructed with the test index of the aircraft engine as the target variable and the state interval as the attribute variable. The posterior probability that the test index of the aircraft engine to be evaluated is qualified is predicted based on the Bayesian network model.

[0015] Furthermore, the assembly data of each unit body is composed of multiple assembly indicators; variance analysis is performed on all assembly indicators in the assembly data of each unit body, and assembly indicators with zero variance are eliminated based on statistical principles, thereby obtaining the screened assembly indicators of each unit body.

[0016] Furthermore, the correlation between the assembly indices after unit screening is calculated, and a unit assembly network model is constructed based on complex network theory, including:

[0017] Network modeling is performed based on the assembly data of the unit body, the assembly indicators screened in the unit body assembly data are abstracted into nodes in the network, the mutual relationship between the assembly indicators is determined based on the correlation between the assembly indicators, and the mutual relationship is abstracted into the edge in the network to construct a unit body assembly network model; the adjacency matrix is ​​constructed using the mutual relationship, and the positive adjacency matrix and the negative adjacency matrix are determined according to the size of the correlation; the correlation is calculated using the Pearson correlation coefficient formula.

[0018] Furthermore, the unit cell dynamics model is expressed as:

[0019] ;

[0020] Where t is the time parameter, , represents the i-th and j-th assembly indexes, Assembly index The initial state of Assembly index The inverse of its own rate of change, is the number of assembly indicators in the unit body assembly network model, a is the trigger threshold, n is the slope of the activation function, and The positive adjacency matrix and the negative adjacency matrix The element in row i and column j in and Assembly index of the excitation and inhibition strengths.

[0021] Furthermore, the specific formula for determining the excitation intensity and the suppression intensity is:

[0022] , ;

[0023] in, and They are the positive correlation coefficient matrices of the unit cell assembly network model And the negative correlation coefficient matrix The element in the i-th row and j-th column of .

[0024] Furthermore, based on the solution of the unit body dynamics model, the fitting errors of various assembly indicators are determined, including:

[0025] The value of the unit body dynamics model is set to zero; the solution value of the assembly index is obtained by solving the unit body dynamics model as the simulation value; and the difference between the simulation value and the actual value of the assembly index is calculated as the fitting error of the assembly index.

[0026] Furthermore, based on the fitting error of the assembly index, the unit cell evaluation interval is constructed using the mean field theory, including:

[0027] Evaluation indicators of unit body By calculating the expectation of all assembly indicators in the unit body assembly network model, we can get:

[0028] ;

[0029] in, is the number of assembly indices in the unit cell assembly network model, Assemble the network model for the unit assembly index; Simulated values ​​of assembly indices As Substitute the value of into the above formula, and based on the The fitting error of the assembly index , the lower and upper bounds of the unit cell evaluation interval are constructed by the following formula:

[0030] ;

[0031] in, is the middle value of the unit cell evaluation interval, and are the lower and upper bounds of the unit evaluation interval, respectively. Then the unit evaluation interval is .

[0032] Furthermore, the actual values ​​of the assembly indexes after screening when the aircraft engine passes or fails the test are used to calculate the evaluation indexes of the unit body when the test passes or fails, and the ranges of all the evaluation indexes are counted; the elbow method is used to determine the optimal clustering number of the evaluation index, the K-Means clustering algorithm is used to cluster the evaluation indexes, and the range of the evaluation indexes is divided into state intervals; a unique code is assigned to each divided state interval, so as to obtain the state interval of the evaluation index and its corresponding code;

[0033] The Bayesian network model is used to calculate the combination of attribute variables when the target variable is qualified, from which the combination of attribute variables greater than the probability threshold set by the enterprise is screened out, and the encoding of each attribute variable in the combination is used to construct an assembly suggestion table to provide assembly suggestions for the enterprise for the assembly of aircraft engines.

[0034] An aircraft engine assembly process optimization device comprises a processor, a memory and a computer program stored in the memory; when the processor is executed by a computer, the aircraft engine assembly process optimization method based on assembly data is implemented.

[0035] A computer-readable storage medium stores a computer program; when the computer program is executed by a processor, the assembly data-based aircraft engine assembly process optimization method is implemented.

[0036] Compared with the prior art, the present invention has the following technical features:

[0037] 1. The aircraft engine assembly process optimization method based on assembly data constructed by the present invention can effectively break the "barrier" between the assembly process and the test process in engineering applications, predict the qualified probability of aircraft engine test indicators and provide a practical assembly suggestion table for the assembly process;

[0038] 2. The evaluation indexes and evaluation intervals of the fan unit, compressor unit and turbine unit constructed by the present invention can effectively evaluate the unit, thereby effectively improving the assembly quality and efficiency, and greatly reducing the enterprise's test run costs;

[0039] 3. The assembly network model and dynamic model of each unit body of the aircraft engine proposed in the present invention can accurately describe the relationship between each assembly index and explain the relationship between the assembly index and the system state change, and is interpretable. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the process of the present invention;

[0041] Figure 2 It is a schematic diagram of the structure of each unit of the aircraft engine;

[0042] Figure 3a A fan assembly network model in an embodiment of the present invention;

[0043] Figure 3b A compressor assembly network model according to an embodiment of the present invention;

[0044] Figure 3c The turbine assembly network model in the embodiment of the present invention;

[0045] Figure 4a is a dynamic fitting error diagram of a fan in an embodiment of the present invention;

[0046] Figure 4b is a dynamic fitting error diagram of a compressor in an embodiment of the present invention;

[0047] Figure 4c is a dynamic fitting error diagram of a turbine in an embodiment of the present invention;

[0048] Figure 5a Schematic diagram of the evaluation index and evaluation interval verification of the fan in an embodiment of the present invention;

[0049] Figure 5b Schematic diagram of evaluation index and evaluation interval verification of a compressor in an embodiment of the present invention;

[0050] Figure 5c It is a schematic diagram of the evaluation index and evaluation interval verification of the turbine in an embodiment of the present invention;

[0051] Figure 6a is a clustering performance diagram of evaluation indicators of a fan in an embodiment of the present invention;

[0052] Figure 6b is a clustering performance diagram of evaluation indicators of a compressor in an embodiment of the present invention;

[0053] Figure 6c is a clustering performance diagram of evaluation indicators of a turbine in an embodiment of the present invention;

[0054] Figure 7 Schematic diagram of the Bayesian network model;

[0055] Figure 8 It is a ROC curve diagram used to characterize the accuracy of the Bayesian network model. DETAILED DESCRIPTION

[0056] In order to illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description in conjunction with embodiments.

[0057] Aiming at the assembly data and test data of aircraft engines, the present invention combines the advantages of complex networks and Bayesian network theory to propose an aircraft engine assembly process optimization method based on assembly data, which can establish a bridge between aircraft engine assembly indicators and test indicators, construct evaluation standards for qualified diagnosis of each unit of the aircraft engine, give the qualified probability of the predicted test indicators of the aircraft engine, and derive an aircraft engine assembly recommendation table that can guide the actual assembly process.

[0058] See attached Figure 1 The present invention provides an aircraft engine assembly process optimization method based on assembly data, comprising the following steps:

[0059] Step 1, when the aircraft engine passes the test run, the assembly data of each unit in the aircraft engine is collected, and the assembly index of the assembly data of the unit is screened; the unit includes a fan, a compressor and a turbine.

[0060] In addition to the fan, compressor and turbine mentioned above, the unit body may also include other parts of the aircraft engine in actual use.

[0061] The assembly data of each unit body is composed of multiple assembly indicators. In this scheme, variance analysis is performed on all assembly indicators in the assembly data of each unit body, and assembly indicators with zero variance are eliminated based on statistical principles, so as to obtain the screened assembly indicators of each unit body, so as to achieve scientific screening of assembly indicators.

[0062] Step 2: Calculate the correlation between the assembly indices after unit screening, and construct a unit assembly network model based on complex network theory.

[0063] According to complex network theory, network modeling is performed based on the assembly data of unit bodies. The assembly indicators screened in the unit body assembly data are abstracted into nodes in the network. The mutual relationships between the assembly indicators are determined based on the correlation between the assembly indicators, and the mutual relationships are abstracted into edges in the network to construct a unit body assembly network model. The adjacency matrix is ​​constructed using the mutual relationships, and the positive adjacency matrix and the negative adjacency matrix are determined according to the size of the correlation.

[0064] (2-1) Determine the relationship between assembly indicators and the adjacency matrix.

[0065] Based on statistical principles, the Pearson correlation coefficient formula is used to calculate the first Item Assembly Index The correlation between :

[0066] (1);

[0067] in, Is the assembly indicator The corresponding covariance, and Is the assembly indicator The corresponding standard deviation.

[0068] The 0-1 adjacency matrix is ​​used to characterize the relationship between nodes in the unit cell assembly network model, which is expressed as follows:

[0069] (2);

[0070] in, is a 0-1 adjacency matrix The element in row i and column j in is represented as the assembly index The relationship between them; when the assembly index Correlation coefficient between The absolute value of the correlation coefficient is greater than the preset threshold value. When There are edges between the nodes in the corresponding assembly network model; otherwise, there are no edges; in this way, the 0-1 adjacency matrix corresponding to the unit body is constructed ; In this embodiment, the correlation coefficient threshold is set to 0.25.

[0071] (2-2) Determine the positive adjacency matrix and the negative adjacency matrix.

[0072] Based on the correlation between assembly indicators Constructing the positive adjacency matrix of the unit cell assembly network model and the negative adjacency matrix ;in .

[0073] when Greater than or equal to Assembly index Positive correlation, then the positive adjacency matrix The element in row i and column j in is 1; when Less than or equal to - Assembly index Negative correlation, then the negative adjacency matrix The element in row i and column j in is 1; the specific representation is as follows:

[0074] (3);

[0075] (4);

[0076] exist Figure 3a to Figure 3cThe given example shows that when the unit bodies are fans, compressors and turbines, the corresponding fan assembly network model, compressor assembly network model and turbine assembly network model are constructed.

[0077] Step 3, introduce network dynamics theory, fit the neural network dynamics model to the unit body assembly network model to obtain the unit body dynamics model; based on the solution of the unit body dynamics model, determine the fitting error of each assembly index.

[0078] (3-1) Construct a unit body dynamics model.

[0079] In the constructed unit assembly network model, positive and negative correlations between assembly indicators are common; positive correlation indicates that the increase of one assembly indicator will lead to the synchronous growth of another assembly indicator; while negative correlation means that the increase of one assembly indicator will lead to the decrease of another assembly indicator.

[0080] In order to quantitatively describe these dynamic interactions, differential equations can be used to model the changing rules between assembly indicators, thereby simulating their interactions. Neural network dynamics, as a mature modeling method, can simulate the information transfer process between neurons based on the interconnection structure and connection strength of neurons, and effectively characterize their positive and negative correlation interaction mechanisms. The original equation form of this dynamic model is as follows:

[0081] (5);

[0082] Where t is the time parameter, is the trigger threshold, is the slope of the activation function, I is the basal activity, R is the inverse of the neuron death rate, is the number of neurons in the neural network, and They represent the and The activity value of a neuron, and represent the excitation and inhibition strengths, respectively. and are the stimulation adjacency matrix and inhibition adjacency matrix of the neural network, and They are and The element in the i-th row and j-th column of .

[0083] Based on neural network dynamics, the assembly index and node number in the unit assembly network model are , positive adjacency matrix and the negative adjacency matrix Introduce and use the correlation between assembly indicators to characterize and , in order to construct a unit body dynamics model, which is used to quantitatively describe the interaction mechanism between various assembly indicators in the unit body assembly network model:

[0084] (6);

[0085] Where t is the time parameter, , represents the i-th and j-th assembly indexes, Assembly index The initial state, i.e. the standard value of the assembly index; Assembly index The inverse of its own rate of change, is the number of assembly indicators in the unit body assembly network model, a is the trigger threshold, which is generally set to -1; n is the slope of the activation function, which is generally set to 2; and The positive adjacency matrix and the negative adjacency matrix The element in row i and column j in and Assembly index The excitation intensity and inhibition intensity can be calculated by the following formula:

[0086] (7);

[0087] (8);

[0088] in, and They are the positive correlation coefficient matrices of the unit cell assembly network model And the negative correlation coefficient matrix The element in row i and column j in can be calculated as follows:

[0089] (9);

[0090] (10);

[0091] (3-2) Determine the fitting error of each assembly index.

[0092] Let the value of the unit body dynamics model be zero (that is, when the system is in steady state), that is ; The assembly index is obtained by solving equation (6) of the unit body dynamics model The solution value is used as the simulation value ; Calculate the simulation value With assembly indicators The difference between the actual values ​​of The fitting error.

[0093] In one example of the present invention, the construction of the fan network dynamics model, the compressor network dynamics model and the turbine network dynamics model and the calculation of the fitting errors of the corresponding assembly indicators are completed according to the above method, such as Figures 4a to 4c shown; from Figures 4a to 4c It can be seen that the actual value of each assembly index is very close to the simulated value calculated by the network dynamics model, indicating that the unit body dynamics model constructed in this scheme can accurately characterize each assembly index and the assembly state of the unit body.

[0094] Step 4, based on the fitting error of the assembly index, construct a unit body evaluation interval using the mean field theory; for the unit body to be evaluated, determine the evaluation index of the unit body using the assembly index of the unit body; based on the relationship between the evaluation index and the unit body evaluation interval, determine whether the unit body to be evaluated is qualified.

[0095] According to the mean field theory, the system average behavior value can characterize the average state of the system. Therefore, the system average behavior value is used as the evaluation index of the unit body to evaluate its assembly state. Specifically, the system average behavior value of the unit body can be obtained by calculating the expectation of all nodes (i.e., assembly index) in the unit body assembly network model:

[0096] (11);

[0097] in, is the evaluation index of the unit body, is the number of assembly indices in the unit cell assembly network model, Assemble the network model for the unit Assembly indicators.

[0098] Based on the above principle, the simulation value of each assembly index As The value of is substituted into formula (11), and the fitting error based on each assembly index is , the lower and upper bounds of the unit cell evaluation interval are constructed by the following formula:

[0099] (12);

[0100] in, is the middle value of the unit cell evaluation interval, and are the lower and upper bounds of the unit evaluation interval, respectively. Then the unit evaluation interval is .

[0101] For the unit body to be evaluated, the assembly data of the unit body is obtained and the assembly index is screened. The actual value of the assembly index after screening is used to calculate the evaluation index of the unit body through formula (11): ,like Falling in the evaluation range If the value is between 0 and 1, the unit is considered qualified, otherwise it is considered unqualified.

[0102] The above evaluation interval can be used to determine whether the unit to be evaluated is qualified:

[0103] The units to be evaluated include units that have not yet been assembled into aircraft engines, and units included in aircraft engines that have completed overall assembly but have not yet been tested. When a unit to be evaluated is determined to be unqualified by the above method, it is necessary to replace it before assembling it into an aircraft engine. If the unit to be evaluated has been assembled into an aircraft engine but is found to be unqualified after evaluation, there is no need to conduct subsequent aircraft engine tests, and the test can be carried out only after replacing it with other qualified units.

[0104] Through the above method, a fan evaluation interval, a compressor evaluation interval and a turbine evaluation interval can be constructed, which are used to perform qualification evaluation on the fan, compressor and turbine to be evaluated respectively.

[0105] Based on the above scheme, the process optimization method may further include the following steps:

[0106] Step 5, after obtaining the assembly data of each unit body when the aircraft engine fails the test and screening the assembly indicators, the evaluation indicators of the unit body are calculated using the screened assembly indicators of the unit body when the aircraft engine passes the test and fails the test, and the evaluation indicators of the unit body are divided into state intervals, and a Bayesian network model is constructed with the test indicators of the aircraft engine as the target variable and the state interval as the attribute variable; based on the Bayesian network model, the posterior probability when the test indicators of the aircraft engine to be evaluated are qualified is predicted, thereby providing guidance for the production and assembly of the aircraft engine.

[0107] When the aircraft engine fails the test run, the assembly data of each unit body is collected and the assembly index is screened; the actual values ​​of the assembly indexes after screening when the aircraft engine passes the test run and fails the test run are used to calculate the evaluation index of the unit body when the test run passes and fails based on formula (11): , and count all evaluation indicators The range of; The elbow method is used to determine the evaluation index The optimal number of clusters is obtained, and the K-Means clustering algorithm is used to evaluate the evaluation index Clustering and evaluation indicators The state interval is divided into two parts according to the range of the state interval, and a unique code is given to each divided state interval to obtain the evaluation index. The state interval and its corresponding code.

[0108] Based on Bayesian theory, the evaluation index The state interval of the engine is taken as the attribute variable, and the test index of the aircraft engine is taken as the target variable to construct a Bayesian network model. The Bayesian network model is used to calculate the combination of attribute variables when the target variable is qualified, and the combination of attribute variables greater than the probability threshold set by the enterprise is screened out. The encoding of each attribute variable in the combination is used to construct an assembly suggestion table, which is used to provide assembly suggestions for the enterprise to assemble aircraft engines.

[0109] In an example of the present invention, the unit body is set to be a fan, a compressor, and a turbine, and the process of constructing the Bayesian network model is:

[0110] First calculate the joint probability distribution:

[0111] (13);

[0112] Where G is the target variable of a given category, i.e., the test index; wherein the given category is qualified and unqualified; is the prior probability distribution of the target variable G, indicating the probability of G taking different values ​​in the absence of any other information; , , are attribute variables, corresponding to the evaluation indicators of fans, compressors, and turbines respectively The state interval of It means that when G is determined, The conditional probability of taking different values, i.e. the fan evaluation index Take the conditional probability of different state intervals; similarly, , The meanings are the same, corresponding to the compressor and turbine respectively; is a joint probability distribution describing the target variable G and the attribute variable , , At the same time, the probability of taking a specific state interval is taken.

[0113] Secondly, for formula (13), a Bayesian network model is constructed to calculate the posterior probability of the target variable G:

[0114] (14);

[0115] in, Indicates that in known attribute variables , , The posterior probability of the target variable G under the corresponding state interval is taken, that is, the probability of the target variable taking the corresponding pass and fail; Represents attribute variables , , Take the total probability of the corresponding state interval.

[0116] According to the above method, the posterior probability of the target variable G being qualified under the state interval combination corresponding to the different encodings of each attribute variable is calculated, thereby obtaining a posterior probability table; in this posterior probability table, the first three columns are the attribute variables , , The fourth column is the posterior probability that the target variable G is qualified under the coding combination of the first three columns, that is, the posterior probability that the aircraft engine test performance index is qualified.

[0117] For an aircraft engine to be evaluated, the assembly index of the aircraft engine assembly data is screened and the evaluation index of each unit body is calculated by formula (11); the state interval in which the evaluation index falls is determined according to the value of the evaluation index, so as to obtain the coding combination of the state interval in which the evaluation index of each unit body falls; based on the coding combination, the posterior probability of the target variable G passing the test is queried from the posterior probability table. If the posterior probability is greater than the probability threshold, the aircraft engine is allowed to be tested; otherwise, it is directly disassembled and reassembled to reduce the test cost of the enterprise.

[0118] On the basis of the above scheme, the following processing can also be performed:

[0119] Based on the posterior probability table, the state interval combinations of each unit body whose target variable is qualified and whose posterior probability is greater than the probability threshold are selected, and the codes of each state interval in the combination are used as the assembly suggestion table; for the unit bodies that are determined to be qualified in step 4, they can be assembled into an aircraft engine in accordance with the combination method of the assembly suggestion table to ensure that they have a high probability of passing the test run. In this way, the assembly process is optimized and guidance is provided for the aircraft engine assembly process.

[0120] The following is an application description of an embodiment of the present invention.

[0121] This embodiment takes a certain aircraft engine manufacturer as the research object, and collects the assembly data and test data of 49 aircraft engines of a certain model, including 37 aircraft engines that passed the test and 12 aircraft engines that did not pass the test; the aircraft engine assembly data records in detail the assembly indexes of each unit body during the assembly process. The data has the characteristics of high-dimensional heterogeneity, multivariate and multimodal, and the correlation between the assembly indexes is strong; the structure of each unit body of the aircraft engine is shown in Figure 2The comprehensive vibration value of an aircraft engine is one of the most critical test indicators in the test data, and can comprehensively reflect the overall performance of the aircraft engine to a certain extent. In actual engineering, the upper limit of the comprehensive vibration value set by the aircraft engine manufacturer is 30, that is, when the comprehensive vibration value exceeds 30, the engine is judged to have failed the test.

[0122] This embodiment aims to construct evaluation indicators of each unit body of the aircraft engine and calculate its qualified interval based on the above-mentioned real assembly data and test data for fault diagnosis; through the state interval combination of the unit body evaluation indicators, predict the probability of the unknown aircraft engine test indicator being qualified, that is, the probability that the comprehensive vibration value is less than 30, and provide an assembly suggestion table to achieve assembly process optimization; the specific steps are:

[0123] By screening the assembly indexes of the aircraft engine assembly data, 23 assembly indexes of the fan are extracted (corresponding to Figure 3a 1 to 23), the compressor extracts 46 assembly indicators (corresponding to Figure 3b 1 to 46), turbo extractor 59 assembly indicators (corresponding to Figure 3c 1 to 59); the assembly indexes of the fan include the initial quantity after adjustment, angular direction, coaxiality, etc.; the assembly indexes of the compressor include the outer circle runout of the drum shaft end stop, coaxiality, assembly clearance at different positions, etc.; the assembly indexes of the turbine include runout value, angular direction, unbalance, etc.; they can be set according to actual needs in the application.

[0124] In this embodiment, for the assembly data of 37 aircraft engines that have passed the test after screening, the fan assembly network model, the compressor assembly network model, and the turbine assembly network model of the qualified aircraft engines are respectively constructed according to the unit body.

[0125] The network dynamics theory is introduced to perform network dynamics fitting on the assembly network model of each unit and calculate the fitting error; the fitting error of the assembly index of the fan, compressor and turbine is The mean values ​​are 0.0412, 0.1206 and 0.0457 respectively, and the kinetic fitting error diagram is drawn, as shown in Figure 4a , Figure 4b and Figure 4c As shown in the figure, there is little difference between the actual value of each unit assembly index and the simulated value calculated by the dynamic model, indicating that the dynamic model of each unit of the aircraft engine that has passed the test run can accurately characterize each assembly index and unit assembly status.

[0126] In this embodiment, the evaluation interval of each unit body can be constructed by formula (12), specifically: First, the simulated value of the assembly index of each unit body in steady state is Substitute to obtain the middle value of the evaluation interval , respectively 0.5933, 0.8246 and 0.7091; then, the fitting error of each unit assembly index Substitute the mean of into the upper bound of the unit evaluation interval and the lower bound , the evaluation intervals of the fan, compressor and turbine are [0.5521, 0.6345], [0.704, 0.9452] and [0.6634, 0.7548] respectively.

[0127] For the unit body to be evaluated, the value of the evaluation index of the unit body is calculated using formula (11). When the value falls within the evaluation interval of the unit body, the unit body can be judged to be qualified. Figure 5a , Figure 5b and Figure 5c As shown; in the figure, green represents the unit body that falls within the unit body evaluation interval, and orange represents the unit body that falls outside the unit body evaluation interval. It can be found that this method can accurately distinguish between qualified and unqualified units.

[0128] In this embodiment, the evaluation indexes of each unit body of the aircraft engine that passed and failed the test run are Take the value, calculate the sum of squares of distances under each number of clusters, and use the elbow rule to select the number of clusters corresponding to the inflection point as the optimal number of clusters for the evaluation index of each unit body; Figure 6a , Figure 6b and Figure 6c As shown in the figure, the optimal clustering number of the fan, compressor and turbine evaluation indicators is 4; on this basis, K-Means is used to cluster the evaluation indicator data of each unit and divide the state interval, and a unique code is assigned to each divided state interval, so as to obtain the state interval of each unit evaluation indicator and its corresponding code, as follows:

[0129] Table 1: Division of status intervals of evaluation indicators of each unit and their corresponding codes.

[0130]

[0131] According to Bayesian theory, the evaluation index of each unit As attribute variables, the test index of the aircraft engine is used as the target variable, that is, the comprehensive vibration value is used as the target variable to construct a Bayesian network model. For the specific structure, see Figure 7 shown.

[0132] In terms of model evaluation, the prediction performance of the model is measured by drawing the receiver operating characteristic curve (ROC curve); the area under the ROC curve is an important indicator for evaluating the prediction accuracy of the model. The larger the value, the stronger the prediction ability of the model. The area under the ROC curve of the Bayesian network model is 71.61%. Figure 8As shown, this indicates that the model has a high prediction accuracy.

[0133] Based on the Bayesian network model, the posterior probability of the target variable G being qualified under the combination of state intervals of different attribute variables is calculated by formula (14), that is, the probability that the comprehensive vibration value of the aircraft engine meets the requirements under the combination of evaluation index values ​​of different unit bodies, thereby obtaining the posterior probability table.

[0134] Specifically, all possible combinations of state intervals of unit evaluation indicators are listed, and the posterior probability of the target variable G being qualified under each combination is calculated to obtain the posterior probability table of the target variable G being qualified, as shown in Table 2. In the posterior probability table, the first three columns are the state intervals corresponding to the evaluation indicator values ​​of the fan, compressor, and turbine, and the fourth column is the posterior probability of the target variable G being qualified under this combination, that is, the probability that the comprehensive vibration value of the aircraft engine test meets the requirements.

[0135] Table 2: Posterior probability table.

[0136]

[0137] Based on the posterior probability table, a probability threshold can be set in combination with the actual production requirements of the enterprise. When the pass probability of the target variable G is greater than the probability threshold, it is considered that the aircraft engine has a high probability of passing the test run. In this embodiment, the probability threshold is set to 90%.

[0138] On this basis, if there is an aircraft engine to be evaluated, the evaluation index of each unit body can be calculated by using the assembly data through formula (11), and the code corresponding to the state interval in which the evaluation index of each unit body falls can be queried by comparing Table 1, and the qualified probability of the target variable G under this coding combination can be substituted into Table 2. If it is greater than 90%, it is allowed to be tested, otherwise it is directly disassembled and reassembled to reduce the test cost of the enterprise.

[0139] Based on the posterior probability table, the combination of the state intervals of the unit body when the posterior probability is greater than 90% is selected as the assembly recommendation table, as shown in Table 3. In the actual assembly process of the enterprise, the evaluation index can be calculated for each assembled unit body using formula (11). If the value falls within the unit body evaluation interval, the unit body is considered qualified and can be used for the assembly of aircraft engines. For qualified units, they can be assembled into aircraft engines according to the combination of the assembly recommendation table to ensure that they have a high probability of passing the test run. In this way, the assembly process is optimized and guidance is provided for the assembly process of aircraft engines.

[0140] Table 3: Assembly suggestions.

[0141]

[0142] The embodiments show that the aircraft engine assembly process optimization method based on assembly data constructed by the present invention can effectively break the "barriers" existing in the assembly process and the test process in engineering applications, can predict the pass probability of aircraft engine test indicators, and provide a practical assembly suggestion table for the assembly process; the evaluation indicators and evaluation intervals of the fan, compressor and turbine constructed by the present invention can effectively evaluate the unit body, thereby effectively improving the assembly quality and efficiency, and can greatly reduce the enterprise test cost; the assembly network model and dynamic model of each unit body of the aircraft engine proposed by the present invention can accurately characterize the relationship between each assembly indicator, and explain the relationship between the assembly indicator and the system state change, and are interpretable.

[0143] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An aircraft engine assembly process optimization method based on assembly data, characterized in that: include: When the aircraft engine passes the test run, the assembly data of each unit in the aircraft engine is collected, and the assembly index of the unit assembly data is screened; Calculate the correlation between the assembly indicators after unit screening, and build a unit assembly network model based on complex network theory; Introducing the network dynamics theory, the neural network dynamics model is fitted to the unit body assembly network model to obtain the unit body dynamics model; based on the solution of the unit body dynamics model, the fitting error of each assembly index is determined; Based on the fitting error of the assembly index, the unit body evaluation interval is constructed using the mean field theory; for the unit body to be evaluated, the evaluation index of the unit body is determined using the assembly index of the unit body; based on the relationship between the evaluation index and the unit body evaluation interval, it is determined whether the unit body to be evaluated is qualified.

2. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 1, characterized in that: The method further comprises: The evaluation index of the unit body is calculated by using the assembly index of the unit body after screening when the aircraft engine test passes or fails, and the evaluation index of the unit body is divided into state intervals. A Bayesian network model is constructed with the test index of the aircraft engine as the target variable and the state interval as the attribute variable. The posterior probability that the test index of the aircraft engine to be evaluated is qualified is predicted based on the Bayesian network model.

3. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 1, characterized in that: The assembly data of each unit body is composed of multiple assembly indicators. A variance analysis is performed on all the assembly indicators in the assembly data of each unit body, and the assembly indicators with zero variance are eliminated based on statistical principles, so as to obtain the screened assembly indicators of each unit body.

4. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 1, characterized in that: Calculate the correlation between the assembly indicators after unit screening, and build a unit assembly network model based on complex network theory, including: Network modeling is performed based on the assembly data of the unit body, the assembly indicators screened in the unit body assembly data are abstracted into nodes in the network, the mutual relationship between the assembly indicators is determined based on the correlation between the assembly indicators, and the mutual relationship is abstracted into the edge in the network to construct a unit body assembly network model; the adjacency matrix is ​​constructed using the mutual relationship, and the positive adjacency matrix and the negative adjacency matrix are determined according to the size of the correlation; the correlation is calculated using the Pearson correlation coefficient formula.

5. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 4, characterized in that: The unit cell dynamics model is expressed as: ； Where t is the time parameter, , represents the i-th and j-th assembly indexes, Assembly index The initial state of Assembly index The inverse of its own rate of change, is the number of assembly indicators in the unit body assembly network model, a is the trigger threshold, n is the slope of the activation function, and The positive adjacency matrix and the negative adjacency matrix The element in row i and column j in and Assembly index of the excitation and inhibition strengths.

6. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 5, characterized in that: The specific formula for determining the excitation intensity and the inhibition intensity is: ， ； in, and They are the positive correlation coefficient matrices of the unit cell assembly network model And the negative correlation coefficient matrix The element in the i-th row and j-th column of .

7. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 1, characterized in that: Based on the solution of the unit body dynamics model, the fitting error of each assembly index is determined, including: The value of the unit body dynamics model is set to zero; the solution value of the assembly index is obtained by solving the unit body dynamics model as the simulation value; and the difference between the simulation value and the actual value of the assembly index is calculated as the fitting error of the assembly index.

8. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 1, characterized in that: Based on the fitting error of the assembly index, the unit cell evaluation interval is constructed using the mean field theory, including: Evaluation indicators of unit body By calculating the expectation of all assembly indicators in the unit body assembly network model, we can get: ； in, is the number of assembly indices in the unit cell assembly network model, Assemble the network model for the unit assembly index; Simulated values ​​of assembly indices As Substitute the value of into the above formula, and based on the The fitting error of the assembly index , the lower and upper bounds of the unit cell evaluation interval are constructed by the following formula: ； in, is the middle value of the unit cell evaluation interval, and are the lower and upper bounds of the unit evaluation interval, respectively. Then the unit evaluation interval is .

9. The method for optimizing the assembly process of an aircraft engine based on assembly data according to claim 2, characterized in that: The actual values ​​of the assembly indexes after screening when the aircraft engine passes or fails the test are used to calculate the evaluation indexes of the unit body when the test passes or fails, and the range of all evaluation indexes is counted; the elbow method is used to determine the optimal clustering number of the evaluation index, the K-Means clustering algorithm is used to cluster the evaluation index, and the range of the evaluation index is divided into state intervals; each divided state interval is given a unique code, so as to obtain the state interval of the evaluation index and its corresponding code; The Bayesian network model is used to calculate the combination of attribute variables when the target variable is qualified, from which the combination of attribute variables greater than the probability threshold set by the enterprise is screened out, and the encoding of each attribute variable in the combination is used to construct an assembly suggestion table to provide assembly suggestions for the enterprise for the assembly of aircraft engines.

10. An aircraft engine assembly process optimization device, comprising a processor, a memory, and a computer program stored in the memory; characterized in that: When the processor is executed by a computer, the method for optimizing the assembly process of an aircraft engine based on assembly data according to any one of claims 1 to 9 is implemented.

Images