Hydraulic system temperature monitoring method based on improved multivariate estimation logic signal

Abstract

The invention relates to the technical field of aircraft hydraulic system fault diagnosis, in particular to a hydraulic system temperature monitoring method based on improved multivariate estimation logic signals, which comprises the following steps: acquiring a fault transfer diagram of a hydraulic system; acquiring system parameters influencing high-temperature faults of the oil tank of the hydraulic system; constructing an MSET model for monitoring the temperature of the hydraulic system; an RSA-MSET model is obtained; and determining a hydraulic system temperature monitoring result. Accurate monitoring of the temperature is achieved, so that the flight safety of the civil aircraft is guaranteed, and the operation safety of a domestic civil aircraft hydraulic system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of aircraft hydraulic system fault diagnosis, and particularly to a hydraulic system temperature monitoring method based on improved multi - element estimation logic signals. Background Art

[0002] The aircraft hydraulic system applies pressure to the hydraulic oil in the fuel tank through a hydraulic pump, transmits the pressure to the users of the hydraulic system, and controls the retraction and extension of the landing gear and the adjustment of the flight attitude during the take - off, landing, and level flight of the aircraft to ensure the safe and stable operation of the aircraft. During take - off, the flap movement is controlled by the hydraulic system to ensure the lift of the aircraft. During level flight, the flight attitude of the aircraft is adjusted by the hydraulic system. During landing, the engine reverse thrust angle, the retraction and extension of the landing gear, and the flap angle are controlled by the hydraulic system to ensure a smooth landing of the aircraft. During the actual operation of civil aircraft, the overheat fault of the civil aircraft hydraulic system has always been one of the typical research objects of hydraulic system faults due to its high fault risk and serious fault consequences.

[0003] The traditional method for monitoring the overheat of the hydraulic system fuel tank is the threshold method. This method relies on traditional design experience and is difficult to characterize the current degradation state of civil aircraft. On the other hand, due to the low failure rate of the fuel tank overheat fault and the small amount of fault data, it is difficult to extract fault characteristics from the actual operation data to achieve fault modeling. Moreover, in the traditional method, the similarity matrix may cause matrix singularity during the calculation process, which may cause a large jump in the performance monitoring parameters and affect the monitoring effect.

[0004] Therefore, to ensure the flight safety of civil aircraft and timely detect abnormal temperature changes in the hydraulic system, a hydraulic system temperature monitoring method based on improved multi - element estimation logic signals is needed to solve the above problems. Summary of the Invention

[0005] The present invention provides a hydraulic system temperature monitoring method based on improved multi - element estimation logic signals to solve the problems that the existing method for monitoring the overheat of the hydraulic system fuel tank is the threshold method, which relies on traditional design experience and is difficult to characterize the current degradation state of civil aircraft. On the other hand, due to the low failure rate of the fuel tank overheat fault and the small amount of fault data, it is difficult to extract fault characteristics from the actual operation data to achieve fault modeling. Moreover, in the traditional method, the similarity matrix may cause matrix singularity during the calculation process, which may cause a large jump in the performance monitoring parameters and affect the monitoring effect.

[0006] A hydraulic system temperature monitoring method based on improved multi - element estimation logic signals of the present invention adopts the following technical solutions, including: Conduct a system safety analysis on the hydraulic system to obtain the fault transfer diagram of the hydraulic system; Screen out the system parameters that affect the high temperature fault of the hydraulic system oil tank from the flight parameters of the aircraft based on the input parameters in the fault transfer diagram. The system parameters include: EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed; Construct a memory matrix with the historical system parameters when the hydraulic system is in a normal working state, construct an initial MSET model based on the memory matrix, take the system parameters as the input values of the initial MSET model to obtain the predicted values of the initial MSET model, obtain the residuals according to the input values and the predicted values, and obtain the target weight matrix when the residuals are the smallest. Construct the MSET model for hydraulic system temperature monitoring according to the target weight matrix and the memory matrix; Use the ridge regularization algorithm to optimize the MSET model to obtain the optimized target MSET model. Use the RSA method to optimize the regularization parameter and the Gaussian kernel operator bandwidth of the target MSET model until the difference between the input value and the output parameter of the target MSET model is the smallest, and obtain the RSA-MSET model; Use the sequential probability ratio test method to test the output results of the RSA-MSET model, and determine the hydraulic system temperature monitoring results according to the test results.

[0007] Preferably, the steps to obtain the fault transfer diagram of the hydraulic system are as follows: Use the FHA analysis method to analyze the fault modes of the high temperature fault of the oil tank; according to the fault modes of the high temperature fault of the oil tank, obtain the typical components that cause the high temperature fault of the hydraulic system oil tank; use the FMEA analysis method to construct the FMEA table corresponding to the typical components of the high temperature fault of the oil tank; use the FTA analysis method to analyze the influence of the component failure modes corresponding to the FMEA table of the typical components of the high temperature fault of the oil tank on the relevant parameters in the upstream system and downstream system of the hydraulic system, and determine the input parameters of the fault transfer diagram; take the component failure mode of the high temperature fault of the oil tank as the bottom event, and construct the fault transfer model of the target fault of the hydraulic system; based on the input parameters, take the logical transfer relationship between the component failure modes in the fault transfer model as the logical path of the fault logic diagram, and the high temperature fault of the oil tank as the output parameter of the fault logic diagram, and construct the fault logic diagram of the high temperature fault of the oil tank.

[0008] Preferably, the typical components of the high temperature fault of the oil tank include: oil filter assembly, unloading valve, hydraulic oil circuit, EDP, ACMP, temperature switch.

[0009] Preferably, the input parameters of the fault transfer diagram are EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed.

[0010] Preferably, constructing the memory matrix includes:

[0011] In the formula, represents the memory matrix; X （m） represents the historical system parameter vector of the hydraulic system in the normal working state at the m-th moment; xn （m） represents the n-th historical system parameter in the historical system parameter vector of the hydraulic system in the normal working state at the m-th moment.

[0012] Preferably, the initial MSET model is:

[0013] In the formula, represents the predicted value of the initial MSET model; represents the memory matrix; W represents the initial weight matrix of the initial MSET model; wm represents the m-th matrix parameter in the initial weight matrix of the initial MSET model; x （m） represents the historical system parameter vector of the hydraulic system in the normal working state for the m-th time.

[0014] Preferably, the expression of the target weight matrix is:

[0015] In the formula, represents the target weight matrix; represents the input value of the MSET model; represents the memory matrix.

[0016] Preferably, the expression of the MSET model for hydraulic system temperature monitoring is:

[0017] In the formula, represents the predicted value of the MSET model; represents the memory matrix; represents the similarity calculation symbol of the corresponding position vectors in matrix multiplication.

[0018] Preferably, the expression of the optimized target MSET model is:

[0019] In the formula, Represents the predicted value of the target MSET model; Represents the regularization parameter; Represents the identity matrix; Represents the similarity calculation symbol of the corresponding position vectors in matrix multiplication.

[0020] Preferably, the sequential probability ratio test method is used to test the output result of the RSA-MSET model, and the steps to determine the temperature monitoring result of the hydraulic system according to the test result are as follows: The expression of the sequential probability ratio is:

[0021] In the formula, is the sequential probability ratio; is the warning threshold; is the residual variance of the RSA-MSET model; is the residual mean of the RSA-MSET model; If , then the hydraulic system is in the normal temperature range; if , then continue to collect system parameters; if , then the hydraulic system is in an abnormal temperature state and an alarm is issued, where is the miss alarm rate, is the false alarm rate.

[0022] The beneficial effects of the present invention are: By establishing a fault logic diagram of the hydraulic system, the fault parameters of the high-temperature fault of the hydraulic system and the fault-related parameters of the upstream system and the downstream system are determined based on the output parameters of the fault logic diagram as system parameters, and the system parameters in the actual flight parameter data are used as the database for constructing the MSET model; for the establishment of the MSET model, on the basis of the traditional MSET model, the present invention introduces the ridge regularization optimization algorithm to improve the model, prevent the generation of a singular matrix in model calculation, and further uses the difference between the predicted value and the input value as the model accuracy index to determine the optimal solutions of the Gaussian kernel function bandwidth and the ridge height in the MSET algorithm, improve the prediction accuracy of the traditional MSET model, and finally use the sequential probability ratio test method to achieve accurate temperature monitoring, so as to ensure the flight safety of civil aircraft and improve the operation safety of the hydraulic system of domestic civil aircraft. Brief Description of the Drawings

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0024] Figure 1 It is a flowchart of a hydraulic system temperature monitoring method based on an improved multi - estimation logic signal of the present invention; Figure 2 It is a detailed flowchart of an embodiment of a hydraulic system temperature monitoring method based on an improved multi - estimation logic signal of the present invention; Figure 3 It is a fault transmission diagram of the hydraulic system of the present invention; Figure 4 It is a diagram showing the change of the health curve of the hydraulic system before and after replacing components in case of a high - temperature fault during the actual process in this embodiment. Detailed implementation manners

[0025] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0026] An embodiment of a hydraulic system temperature monitoring method based on an improved multi - estimation logic signal of the present invention, as Figure 1 shown, includes: S1. Obtain the fault transmission diagram of the hydraulic system; Step 11. Analysis preparation: Analyze the fault transfer model of the domestic civil aircraft hydraulic system based on the working principle of the aircraft and the manual information. Conduct a requirements analysis on the relevant parameters of the basic events of the fault transfer model, and then construct the fault logic diagram of the hydraulic system. Among them, the functional faults of the hydraulic system include pressure supply function faults, oil supply / return function faults, safety function faults, indication function faults, and logic control function faults. The pressure supply function faults include abnormal low pressure in the hydraulic system, loss of unloading function in the hydraulic system, failure of the unloading valve, and failure of the PTU; the oil supply / return function faults include leakage of the hydraulic oil circuit, high temperature of the hydraulic oil tank, and overheat of the hydraulic oil tank; the safety function fault is the failure of the fire cut-off valve in the hydraulic system; the indication function faults include the failure of the temperature sensor in the hydraulic system, the failure of the pressure sensor in the hydraulic system, and the failure of the oil quantity sensor in the hydraulic system; the logic control function faults include the failure of HCLE1 logic control and the failure of HCLE2 logic control. Use the FHA analysis method to analyze the influence level of each fault. Among them, abnormal low pressure in the hydraulic system, loss of unloading function in the hydraulic system, failure of the unloading valve, failure of the PTU, leakage of the hydraulic oil circuit, and high temperature of the hydraulic oil tank are of grade III danger level; the failure of the temperature sensor in the hydraulic system, the failure of the pressure sensor in the hydraulic system, and the failure of the oil quantity sensor in the hydraulic system are of grade IV danger level; overheat of the hydraulic oil tank is of grade II danger level. Since the frequency of high temperature of the hydraulic oil tank is the highest, the high temperature fault of the oil tank is selected as the research object. Based on the functional principle analysis and the FHA analysis results, screen the typical components that can cause the high temperature fault of the oil tank. The typical components of the high temperature fault of the oil tank include the oil filter assembly, unloading valve, hydraulic oil circuit, EDP, ACMP, and temperature switch. Use the FMEA analysis method to construct the FMEA table corresponding to the typical components of the high temperature fault of the oil tank, specifically including the FMEA table of the oil filter assembly, the FMEA table of the unloading valve, the FMEA table of the hydraulic oil circuit, the FMEA table of the EDP, the FMEA table of the ACMP, and the FMEA table of the temperature switch.

[0027] Step 12. Obtain the fault transfer diagram of the hydraulic system: Use the FHA analysis method to analyze the fault forms of the high temperature fault of the oil tank; according to the fault forms of the high temperature fault of the oil tank, obtain the typical components that cause the high temperature fault of the oil tank in the hydraulic system; use the FMEA analysis method to construct the FMEA table corresponding to the typical components of the high temperature fault of the oil tank; use the FTA analysis method to analyze the influence of the component failure forms corresponding to the FMEA table of the typical components of the high temperature fault of the oil tank on the relevant parameters in the upstream system and downstream system of the hydraulic system, and determine the input parameters of the fault transfer diagram; take the component failure forms of the high temperature fault of the oil tank as the basic events, and construct the fault transfer model of the target fault of the hydraulic system; based on the input parameters, take the logical transfer relationship between the component failure forms in the fault transfer model as the logical path of the fault logic diagram, and the high temperature fault of the oil tank as the output parameter of the fault logic diagram, and construct the fault logic diagram of the high temperature fault of the oil tank.

[0028] So far, the fault logic diagram of the fuel tank high-temperature fault and the input parameters of the fault logic diagram have been obtained.

[0029] S2. Obtain the system parameters that affect the high-temperature fault of the hydraulic system fuel tank; Specifically, based on the input parameters in the fault transfer diagram, screen out the system parameters that affect the high-temperature fault of the hydraulic system fuel tank from the flight parameters of the aircraft. Among them, the system parameters include: EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed.

[0030] Since the input parameters in the fault transfer diagram are EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed, first obtain the flight parameter data during the level flight of the aircraft. Then, based on the input parameters in the fault transfer diagram, screen out the EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed that affect the high-temperature fault of the hydraulic system fuel tank from the flight parameters of the aircraft. The EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature, and engine high-pressure rotor speed are used as system parameters.

[0031] S3. Construct the MSET model for monitoring the hydraulic system temperature; Specifically, construct a memory matrix with the historical system parameters when the hydraulic system is in a normal working state. Based on the memory matrix, construct an initial MSET model. Use the system parameters as the input values of the initial MSET model to obtain the predicted values of the initial MSET model. Obtain the residuals according to the input values and the predicted values, and obtain the target weight matrix when the residuals are the smallest. Construct the MSET model for monitoring the hydraulic system temperature according to the target weight matrix and the memory matrix.

[0032] Step 31. Construct the memory matrix: Let the historical system parameters when the hydraulic system is in a normal working state be , where x1 is the EDP output pressure, x2 is the ACMP output pressure, x3 is the hydraulic system pressure, x4 is the ambient temperature, x5 is the engine high-pressure rotor speed, and x6 is the hydraulic system temperature. Among them, the hydraulic system temperature is used as the main monitoring parameter, and the EDP output pressure, ACMP output pressure, hydraulic system pressure, ambient temperature, and engine high-pressure rotor speed are used as auxiliary monitoring parameters. That is, record the system parameters collected at the t-th moment as , collect the historical system parameters during the level flight period of multiple aircraft. Take 1000 consecutive historical system parameters during the level flight stage of each of the 20 flights as the database of the memory matrix. On this basis, extract a set of historical system parameters every 10 seconds to form a historical system parameter vector, and construct a memory matrix based on the extracted historical system parameter vector. Specifically, the expression of the memory matrix is: (1) In the formula, represents the memory matrix; X （m） represents the historical system parameter vector when the hydraulic system is in the normal working state for the mth time; Xn （m） represents the nth historical system parameter in the historical system parameter vector when the hydraulic system is in the normal working state for the mth time. The subspace (denoted by D) spanned by the m historical system parameter vectors in the memory matrix constructed by this process can represent the entire dynamic process of the normal operation of the process or equipment.

[0033] Step 32: Establish an initial MEST model; The input of the initial MEST model is the current system parameter (i.e., the input value) Xobs of the process or equipment at the current moment, and the output of the initial MEST model is the predicted value Xest corresponding to the current moment. For any input value Xobs, the initial MEST model will output an m-dimensional initial weight matrix, denoted as , and establish an initial MEST model based on the memory matrix composed of the preset initial weight matrix and the current input value Xobs, that is, the initial MEST model is: (2) In the formula, represents the predicted value of the initial MSET model; represents the memory matrix; W represents the initial weight matrix of the initial MSET model; wm represents the mth matrix parameter in the initial weight matrix of the initial MSET model; x （m） represents the historical system parameter vector when the hydraulic system is in the normal working state at the mth moment.

[0034] Step 33: Obtain the target weight matrix: The residual between the input value and the output value in the initial MEST model is: (3) In the formula, represents the residual between the input value and the output value in the initial MEST model.

[0035] Select the initial weight matrix W to minimize the sum of the squares of the residuals. The sum of the squares of the residuals is: (4) In the formula, represents the sum of squares of residuals; represents the residual between the input value and the output value in the initial MEST model; represents the memory matrix; W represents the initial weight matrix of the initial MSET model.

[0036] The transformed sum of squares is: (5) Take the partial derivatives of S(w) with respect to w 1 , w 2 , ⋯, w m and set them equal to 0: (6) Equation (6) can be transformed into: (7) In the formula, represents the parameter in the j-th row and k-th column of the memory matrix; represents the parameter in the i-th row and j-th column of the memory matrix; represents the j-th initial weight parameter of the initial weight matrix of the initial MSET model; represents the i-th historical system parameter in the vector of historical system parameters input to the initial MSET model.

[0037] Transforming Equation (7) into matrix form, we get: (8) From this, the target weight matrix is obtained: (9) Replacing the initial weight matrix in Equation (2) with the target weight matrix in Equation (9), we can obtain: (10) Since the data type of the system parameters is non-linear data, to ensure the accuracy of the algorithm and avoid matrix singularity, therefore, a similarity calculation method is used to represent the similarity of each pair of elements in the matrix instead of the multiplication symbol in Equation (10). Then, the expression of the MSET model for hydraulic system temperature monitoring after the change is: (11) In the formula, represents the predicted value of the MSET model; represents the memory matrix; represents the similarity calculation symbol for the corresponding position vectors in matrix multiplication. The expression for similarity calculation is: (12) In the formula, is the similarity with ; is the bandwidth of the Gaussian kernel operator; is the row vector at the corresponding position of the pre-matrix; is the column vector at the corresponding position of the post-matrix.

[0038] S4. Obtain the RSA-MSET model; Specifically, use the ridge regularization algorithm to optimize the MSET model to obtain the optimized target MSET model, and use the RSA method to optimize the regularization parameter and the Gaussian kernel operator bandwidth of the target MSET model until the difference between the input value and the output parameter of the target MSET model is minimized, and the RSA-MSET model is obtained.

[0039] Step 41. Obtain the optimized target MSET model: Since may be a singular matrix, which may lead to a reduction in the accuracy of the calculation result. An optimization method is used to avoid matrix singularity to improve the algorithm accuracy. To improve the prediction accuracy of the MSET model, the ridge regularization algorithm is used to optimize the MSET model. The optimized target MSET model is:

[0040] In the formula, represents the predicted value of the target MSET model; represents the regularization parameter; represents the identity matrix; represents the similarity calculation symbol of the corresponding position vectors in matrix multiplication.

[0041] Step 42. Obtain the RSA-MSET model: Use the RSA method to optimize the regularization parameter of the target MSET model and the Gaussian kernel operator bandwidth to obtain the optimal regularization parameter and the optimal Gaussian kernel operator bandwidth when the difference between the input value and the predicted value is minimized. The specific solution process is as follows: For the two optimization parameters of the regularization parameter and the Gaussian kernel operator bandwidth , randomly generate a set of candidate solutions, and based on this, perform multiple iterations. The optimal solution of each iteration is regarded as approaching the best value: (14) In the formula, Let \(N\) be the number of candidate solutions, and \(M\) represent the candidate solution matrix. Among them, the generation process of candidate solutions is as follows: (15) In the formula, represents the \(j\)-th position of the \(i\)-th solution, \(N\) represents the number of candidate solutions, and \(n\) represents the dimension of the given problem; is a random number between 0 and 1; represents the upper bound of the given problem; represents the lower bound of the given problem, and the interval between \(LB\) and \(UB\) is the permitted interval.

[0042] Further, a surrounding exploration is carried out for the optimal solution, and the position update equation of the optimal solution is: (16) In the formula, is the \(j\)-th position in the best solution obtained by the cut-off time \(t\), represents the hunting operator of the \(j\)-th position in the \(i\)-th solution, is a sensitive parameter that controls the exploration accuracy of walking off the ground during the iteration process, and the value of the sensitive parameter is 0.1; the reduction function is the value used to reduce the search area, represents a random number between 0 and 1; refers to the point randomly selected from the candidate solutions, is a random number between [1, \(N\)], and the evolutionary meaning is the probability ratio of randomly taking a decreasing value between 2 and -2 within the entire number of iterations; is a minimum value used to prevent the denominator from being 0, is the percentage difference between the \(i\)-th position of the best solution obtained and the \(j\)-th position of the current solution; \(T\) represents the maximum number of iterations; Among them, the hunting operator The expression is: (17) Among them, The expression is: (18) In the formula, is a sensitive parameter with a value of 0.1; the main parameter of the residual in formula (3) is used as the fitness, and when the fitness iterates into the permitted interval, the alligator position with the highest satiety (the lowest fitness value) is returned, and the parameter returned by this optimal position is the optimal regularization parameter and the optimal Gaussian kernel operator bandwidth ; thus, the regularization parameter of the target MSET model using the RSA method is completed Optimize the bandwidth of the Gaussian kernel operator to obtain the optimized RSA-MSET model.

[0043] S5. Determine the temperature monitoring result of the hydraulic system; Specifically, the sequential probability ratio test method is used to test the output result of the RSA-MSET model, and the temperature monitoring result of the hydraulic system is determined according to the test result.

[0044] Step 51. The expression for testing the output result of the RSA-MSET model using the sequential probability ratio test method is: (19) In the formula, is the sequential probability ratio; is the alarm threshold; is the residual variance of the RSA-MSET model; is the residual mean of the RSA-MSET model.

[0045] If , the hydraulic system is in the normal temperature range; If , continue to collect system parameters; If , the hydraulic system is in an abnormal temperature state and an alarm is issued, where is the miss alarm rate, is the false alarm rate.

[0046] Thus, the temperature monitoring of the civil aircraft hydraulic system is completed.

[0047] The following specifically describes the embodiments of the present invention in conjunction with the attached Figure 2 : Step 1. Analysis preparation and construction of the logic fault diagram: Based on the working principle analysis of the domestic civil aircraft hydraulic system and FHA analysis, FMEA analysis, and FTA analysis methods, establish a fault transfer model of the domestic civil aircraft hydraulic system. Based on the fault transfer model, conduct a bottom event analysis, and then construct a fault logic diagram for the high temperature fault of the fuel tank of the hydraulic system. Specifically, the fault logic diagram constructed in this embodiment is as Figure 3 shown.

[0048] Step 2. Obtaining of model input parameters: Construct a hydraulic system fault logic diagram based on aircraft manual information, and determine the input parameters related to the temperature of the hydraulic system. Specifically, the input parameters are shown in Table 1 below. The normal operating data sample of the hydraulic system is the parameter sample during the level flight phase. When the aircraft is in the level flight phase, the external temperature of the hydraulic system is relatively fixed, and the hydraulic system users rarely generate actions. Therefore, the operating conditions of the heat generating components and heat dissipating components are relatively stable. By obtaining the flight parameter data in the aircraft's quick access recorder during the level flight phase of the aircraft, it is used as the system parameters under the normal operating state of the hydraulic system. The system parameters are used as input parameters to construct a data sample, and this data sample is used as the basis for establishing the baseline model.

[0049] Table 1

[0050] Step 3: Construct the RSA-MSET model: During the level flight phase of the aircraft, obtain the flight parameter data of different aircraft numbers and different routes, and obtain 2000 flight parameter data. Based on the historical system parameters in the flight parameter data, construct a memory matrix. The memory matrix is: (20) In the formula, represents the memory matrix, HYD_1PRESS represents the pressure of the No. 1 hydraulic system, EDP1_PRESS represents the output pressure of the No. 1 EDP, ACMP1_PRESS represents the output pressure of the No. 1 ACMP, and ENG_N2 represents the high-pressure rotor speed of the left engine; Perform a normalization operation on the constructed memory matrix to facilitate the execution of calculations.

[0051] The memory matrix and the input value are used to predict the system parameters, and the initial MSET model for the hydraulic temperature can be obtained. Then, the difference between the predicted value and the input value of the initial MSET model, that is, the residual, is used as the characterization of the health state of the hydraulic system. To illustrate the effectiveness of the RSA-MSET model in monitoring the temperature of the civil aircraft hydraulic system, this embodiment conducts an effectiveness analysis by combining the changes in the system health curves before and after replacing components due to high-temperature faults in the actual process. The changes in the health curves are as Figure 4 shown, Figure 4 The curve in it is the RSA-MSET health monitoring curve (blue broken line). Since the health state of the aircraft hydraulic system was good in the early stage, the operating data was relatively close to the recorded parameters in the memory matrix. The RSA-MSET model can predict the hydraulic system temperature data with extremely high accuracy. At this time, the residual fluctuates around 0. As the aircraft continues to operate, the state of the hydraulic system gradually shifts from a healthy state to a faulty state (the critical value is the red line). At this time, the fluctuation of the residual curve increases, and it reaches the maximum value of the residual before the component is replaced. The effectiveness of the prediction method is verified by comparing the changes in the residual before and after the component is replaced.

[0052] Step 4. Temperature Monitoring of the Hydraulic System: The expression for testing the output result of the RSA-MSET model using the sequential probability ratio test method is:

[0053] In the formula, is the sequential probability ratio; is the alarm threshold; is the residual variance of the RSA-MSET model; is the mean residual of the RSA-MSET model. If , the hydraulic system is in the normal temperature range; if , continue to collect system parameters; if , the hydraulic system is in an abnormal temperature state and an alarm is issued.

[0054] From Figure 4 the results, it can be seen that a replacement occurred at the 1534 time point. At this time, the health parameters mutated, and the system state changed from a faulty state to a healthy state. In addition, based on the fault alarm threshold obtained by the sequential probability ratio test method, the aircraft should replace the hydraulic system components at the 1480 flight time point, realizing the forward shift of fault alarm. Based on the RSA-MSET model and the MSET model, a comparative analysis of the accuracy of the hydraulic system temperature monitoring model was carried out. The comparative analysis results are shown in Table 2. From Table 2, it can be concluded that in terms of analysis accuracy, the RSA-MSET model is accurate and feasible.

[0055] Table 2

[0056] As can be seen from Table 2, the monitoring accuracy of the RSA-MSET model is higher than that of the traditional MSET model. When the hydraulic system data is in a healthy condition, 1000 data in the healthy state are taken for verification. The verification results show that the monitoring accuracy of the RSA-MSET model is higher than that of the traditional MSET model.

[0057] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A hydraulic system temperature monitoring method based on improved multivariate estimation logic signal, characterized in that: include: Conduct system safety analysis on the hydraulic system and obtain the fault transmission diagram of the hydraulic system; Based on the input parameters in the fault transfer diagram, the system parameters that affect the high temperature fault of the hydraulic system oil tank are screened from the flight parameters of the aircraft, wherein the system parameters include: EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature and engine high pressure rotor speed; The historical system parameters of the hydraulic system in normal working state are used to construct a memory matrix, an initial MSET model is constructed based on the memory matrix, the system parameters are used as the input values ​​of the initial MSET model to obtain the predicted values ​​of the initial MSET model, the residuals are obtained according to the input values ​​and the predicted values, and the target weight matrix when the residuals are minimum is obtained, and the MSET model for hydraulic system temperature monitoring is constructed according to the target weight matrix and the memory matrix; The MSET model is optimized by using the ridge regularization algorithm to obtain the optimized target MSET model. The regularization parameter and Gaussian kernel operator bandwidth of the target MSET model are optimized by using the RSA method until the difference between the input value and the output parameter of the target MSET model is minimized, thus obtaining the RSA-MSET model. The sequential probability ratio test method is used to test the output results of the RSA-MSET model, and the hydraulic system temperature monitoring results are determined based on the test results.

2. A hydraulic system temperature monitoring method based on improved multivariate estimation logic signal according to claim 1, characterized in that: The steps to obtain the fault transfer diagram of the hydraulic system are: The FHA analysis method is used to analyze the failure mode of the oil tank high temperature failure; According to the fault form of oil tank high temperature fault, obtain the typical components that cause oil tank high temperature fault in the hydraulic system; Use FMEA analysis method to construct FMEA tables corresponding to typical components of fuel tank high temperature failure; Use the FTA analysis method to analyze the influence of the failure form of the components corresponding to the high temperature failure of the oil tank in the hydraulic system on the relevant parameters of the upstream and downstream systems of the hydraulic system in the FMEA table of the typical components of the oil tank high temperature failure, and determine the input parameters of the fault transfer diagram; Taking the component failure form of oil tank high temperature fault as the bottom event, the fault transmission model of the target fault of the hydraulic system is constructed; Based on the input parameters, the logical transfer relationship between the component failure forms in the fault transfer model is used as the logical path of the fault logic diagram, and the fuel tank high temperature fault is used as the output parameter of the fault logic diagram to construct the fault logic diagram of the fuel tank high temperature fault.

3. A hydraulic system temperature monitoring method based on improved multivariate estimation logic signal according to claim 2, characterized in that: Typical components of oil tank high temperature failure include: oil filter assembly, unloading valve, hydraulic oil circuit, EDP, ACMP, and temperature switch.

4. The method for monitoring the temperature of a hydraulic system based on an improved multivariate estimation logic signal according to claim 2, characterized in that: The input parameters of the fault transfer diagram are EDP output pressure, ACMP output pressure, hydraulic system pressure, hydraulic system temperature, ambient temperature and engine high-pressure rotor speed.

5. The hydraulic system temperature monitoring method based on improved multivariate estimation logic signal according to claim 1 is characterized in that: Building a memory matrix includes: In the formula, represents the memory matrix; X （m） Represents the historical system parameter vector of the hydraulic system in normal working condition at the mth moment; xn （m） Represents the nth historical system parameter in the historical system parameter vector when the hydraulic system is in normal working condition at the mth moment.

6. The method for monitoring temperature of a hydraulic system based on an improved multivariate estimation logic signal according to claim 1, characterized in that: The initial MSET model is: In the formula, represents the predicted value of the initial MSET model; represents the memory matrix; W represents the initial weight matrix of the initial MSET model; wm represents the mth matrix parameter in the initial weight matrix of the initial MSET model; x （m） Represents the historical system parameter vector of the hydraulic system under normal working condition for the mth time.

7. The method for monitoring temperature of a hydraulic system based on an improved multivariate estimation logic signal according to claim 1, characterized in that: The expression of the target weight matrix is: In the formula, represents the target weight matrix; Represents the input value of the MSET model; Represents the memory matrix.

8. The method for monitoring temperature of a hydraulic system based on an improved multivariate estimation logic signal according to claim 1, characterized in that: The expression of the MSET model for hydraulic system temperature monitoring is: In the formula, represents the predicted value of the MSET model; represents the memory matrix; Represents the symbol for calculating the similarity of corresponding position vectors in matrix multiplication.

9. The method for monitoring temperature of a hydraulic system based on improved multivariate estimation logic signals according to claim 8, characterized in that: The expression of the optimized target MSET model is: In the formula, represents the predicted value of the target MSET model; represents the regularization parameter; represents the identity matrix; Represents the symbol for calculating the similarity of corresponding position vectors in matrix multiplication.

10. The method for monitoring temperature of a hydraulic system based on an improved multivariate estimation logic signal according to claim 1, characterized in that: The sequential probability ratio test method is used to test the output results of the RSA-MSET model, and the steps for determining the hydraulic system temperature monitoring results based on the test results are as follows: The expression of sequential probability ratio is: In the formula, is the sequential probability ratio; is the alarm threshold; is the residual variance of the RSA-MSET model; is the residual mean of the RSA-MSET model; like , the hydraulic system is in the normal temperature range; if , then continue to collect system parameters; if , the hydraulic system is in an abnormal temperature state and an alarm is issued, among which, is the missed alarm rate, is the false alarm rate.

Images