A Fault Diagnosis Method for Anti-recoil Devices Based on Multi-parameter Fusion

Abstract

This invention discloses a fault diagnosis method for an anti-recoil device based on multi-parameter fusion, comprising: establishing the differential equations of recoil motion and recoil motion of the anti-recoil device according to Newton's second law; substituting the basic parameters of a fault-free anti-recoil device into the differential equations of recoil motion and recoil motion of the anti-recoil device to obtain a fault-free sample dataset under zero angle of attack; establishing a fault analysis and diagnosis model for the anti-recoil device; obtaining a mapping relationship dataset to form a fault test and training set; establishing a multi-parameter fusion artificial neural network diagnosis model based on the fault analysis and diagnosis model of the anti-recoil device, and training and optimizing it using the fault test and training set to obtain an intelligent model for fault diagnosis of the anti-recoil device; using the intelligent model for fault diagnosis of the anti-recoil device, inputting the working status data of the anti-recoil device, performing fault discrimination and diagnosis, and obtaining the diagnosis result.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology for special mechanical devices, and in particular to a fault diagnosis method for special mechanical devices based on multi-parameter fusion. Background Technology

[0002] The recoil countermeasure is a special mechanical device that can rapidly buffer and attenuate vibration and impact energy during the equipment's recoil stroke, and accurately and quickly reset under the action of an energy storage buffer. Its comprehensive support capability is a crucial factor restricting the effective operation of the equipment. For a long time, the comprehensive support capability of recoil countermeasures has mostly remained at the level of post-failure maintenance, with insufficient research into the failure mechanisms and overly simplistic diagnostic methods, largely relying on single-parameter fault diagnosis. Therefore, based on research into the fault diagnosis mechanism of recoil countermeasures, this paper proposes a multi-parameter fusion fault diagnosis method. This method can obtain important information about the recoil countermeasure in a timely manner, enabling targeted preventative measures to ensure safer and better mission completion.

[0003] The technical problem to be solved by the present invention is to provide a high-efficiency, multi-parameter fusion fault diagnosis method for equipment, which can effectively diagnose various faults of the anti-recoil device. Summary of the Invention

[0004] This invention aims to solve the problem of efficient, multi-parameter fusion-based fault diagnosis methods for equipment. According to the method of this invention, a fault diagnosis method for an anti-recoil device based on multi-parameter fusion is provided, comprising: establishing the recoil motion differential equation and the recoil motion differential equation of the anti-recoil device according to Newton's second law; substituting the basic parameters of the fault-free anti-recoil device into the recoil motion differential equation and the recoil motion differential equation of the anti-recoil device, and taking the angle of attack as zero, solving the recoil motion differential equation and the recoil motion differential equation using the ode45 function of the fourth-order fifth-level Runge-Kutta method in MATLAB to obtain a fault-free sample dataset under zero angle of attack; establishing a fault analysis and diagnosis model of the anti-recoil device based on the recoil motion differential equation and the recoil motion differential equation of the anti-recoil device; and using MATLAB or FORTRAN to program the fault analysis and diagnosis... The simulation program of the model obtains the mapping relationship between typical faults of the anti-recoil device and their fault characteristic parameters, forming a mapping relationship dataset. Based on the mapping relationship dataset, the trend curve of the relationship between the extreme values ​​of the fault characteristic parameters and the degree of occurrence of typical faults is obtained, and the trend curve is used to form a fault test and training set. Based on the fault analysis and diagnosis model of the anti-recoil device, a multi-parameter fusion artificial neural network diagnostic model is established, and the artificial neural network diagnostic model is trained and optimized using a fault-free sample dataset and the fault test and training set to obtain an intelligent model for fault diagnosis of the anti-recoil device. Using the intelligent model for fault diagnosis of the anti-recoil device, the working status data of the anti-recoil device is input to perform fault discrimination and diagnosis, and obtain the diagnostic results.

[0005] Preferably, the differential equation of the recoil motion is:

[0006]

[0007] in:

[0008] m h For the mass of the rear seat portion;

[0009] x represents the recoil displacement;

[0010] v is the recoil speed;

[0011] t is the recoil time;

[0012] F R For recoil resistance, and

[0013]

[0014] in:

[0015] For the angle of attack; F ΦhIt is the hydraulic resistance of the retractor, and its expression is...

[0016] F Φh =f(a x V 2 =f(a x0 +Δa x V 2

[0017] in:

[0018] a x The annular leak area at the control ring (where a is the annular leak area when the control ring is not worn) is the area of ​​the annular leak. x0 With the area Δa worn away x (the sum of the two), where v is the recoil velocity;

[0019] f(a x The expression for ) is

[0020]

[0021] in:

[0022] A1 is the minimum tributary cross-sectional area;

[0023] K1 and K2 are hydraulic resistance coefficients, K1 = 1.2 - 1.6, K2 = (2 - 4)K1;

[0024] A p To control the area of ​​the annular hole.

[0025] A fj The working area of ​​the controller. d1 is the inner diameter of the brake lever;

[0026] a x The area of ​​the fluid flow orifice. d x To control the outer diameter of the rod;

[0027] ρ is the density of the desiccant.

[0028] Preferably, the F f For the reciprocating force, its expression is:

[0029]

[0030] in:

[0031] A f This refers to the working area of ​​the piston in the reciprocating machine.

[0032] p f0 This is the initial gas pressure in the gas storage chamber of the recoil engine;

[0033] V0 is the initial volume of gas in the gas storage chamber of the recoil engine;

[0034] x represents the recoil displacement.

[0035] F represents the frictional force in the sealing element of the recoil mechanism.

[0036] F=νm h g

[0037] In the formula: ν is the equivalent friction coefficient of the sealing link, which takes a value of 0.3-0.5;

[0038] The F T For the friction of the cradle guide rail,

[0039]

[0040] In the formula: f is the friction coefficient of the cradle guide rail, which takes a value of 0.16-0.20.

[0041] Preferably, the differential equation of the recursive motion is as follows:

[0042]

[0043] in:

[0044] m h For the mass of the rear seat portion;

[0045] x is the recoil displacement;

[0046] v is the reciprocating velocity;

[0047] t is the re-entry time;

[0048] F f For recoil mechanism;

[0049] The angle of attack;

[0050] in:

[0051] F Φfy It is the reciprocating hydraulic resistance, and its expression is:

[0052] F Φfy =F Φf +F Φfj +F Φkh in:

[0053] F Φf It is the hydraulic resistance provided by the recoil mechanism during the return-forward movement.

[0054]

[0055] in:

[0056] K 1f The hydraulic resistance coefficient of the retraction mechanism during the return stroke is typically 1.4-1.6.

[0057] A 0f The working area of ​​the piston in the retraction mechanism during the return stroke:

[0058]

[0059] a x To control the area of ​​the annular leak at the control ring;

[0060] v is the recoil speed.

[0061] Preferably, the F Φfj The hydraulic resistance is provided by the reciprocating controller:

[0062]

[0063] in:

[0064] K 2f The hydraulic resistance coefficient of the reciprocating controller is generally K. 2f = (3-4)K 1f ;

[0065] A fj Working area of ​​the reciprocating regulator piston:

[0066]

[0067] a f This represents the area of ​​the groove through which the liquid passes by the inner cavity of the brake rod.

[0068]

[0069] v is the recoil speed;

[0070] The F Φkh It is the hydraulic resistance of the reciprocating control valve:

[0071]

[0072] in:

[0073] K v The hydraulic resistance coefficient of the reciprocating control valve flow hole;

[0074] a v To control the working area of ​​the valve flow orifice:

[0075] A f This represents the working area of ​​the piston in the reciprocating machine.

[0076] Preferably, the fault-free sample dataset includes: the recoil recovery displacement-time relationship curve of the recoil mechanism under fault-free conditions, the recoil recovery velocity-time relationship curve of the recoil mechanism under fault-free conditions, and the recoil recovery acceleration-time relationship curve of the recoil mechanism under fault-free conditions.

[0077] Preferably, the fault analysis and diagnosis model of the recoil device reflects the nonlinear relationship between fault characteristic parameters and the main fault types of the recoil device, and reflects the mapping relationship between the main fault types of the recoil device and its fault characteristic parameters. The fault characteristic parameters include the maximum recoil displacement, the maximum recoil speed, the maximum recoil force, and the maximum retraction force. The main fault types include: the degree of wear of the control ring, the degree of air leakage of the recoil mechanism, and the degree of liquid leakage of the retraction mechanism.

[0078] Preferably, the fault test and training set includes: the influence curve of the wear fault of the inner diameter of the control ring on the maximum value of the recoil return stroke, the influence curve of the wear fault of the inner diameter of the control ring on the maximum value of the recoil reaction force FR; the influence curve of the external oil leakage fault of the retractor piston on the maximum value of the recoil return stroke; the influence curve of the external oil leakage fault of the retractor piston on the maximum value of the recoil return speed, and the influence curve of the external oil leakage fault of the retractor piston on the minimum value of the recoil return acceleration.

[0079] Preferably, the artificial neural network diagnostic model includes an input layer, an output layer, and a hidden layer. The input layer has four nodes, namely the maximum recoil displacement, the maximum recoil velocity, the maximum recoil force, and the maximum retraction force. The output layer has three nodes, namely the wear degree of the control ring, the air leakage degree of the recoil mechanism, and the liquid leakage degree of the retraction mechanism. The hidden layer is adjusted to have seven nodes.

[0080] Preferably, the working status data is acquired using vibration sensors and pressure sensors. The acquired working status data is saved and preprocessed before being sent to the fault diagnosis intelligent model of the anti-recoil device for fault diagnosis. The steps for training and optimizing the artificial neural network diagnostic model include: inputting the training set into the initial neural network, learning using the gradient descent algorithm, and adjusting the neural network model coefficients; inputting the test set into the fault diagnosis model of the anti-recoil device, evaluating the prediction performance based on the results, and choosing to retrain after adjustment or accept the current training to determine the fault diagnosis intelligent model.

[0081] According to the method of the present invention, compared with the prior art, it is possible to correctly diagnose more than three types of malfunctions of the anti-recoil device. Attached Figure Description

[0082] Various embodiments or examples (“Examples”) of this disclosure are disclosed in the following detailed description and accompanying drawings. The drawings are not necessarily drawn to scale. Generally, unless otherwise specified in the claims, the products or methods disclosed in this invention can be performed in any order. In the drawings:

[0083] Figure 1 The present invention provides a fault-free recoil recovery displacement-time relationship curve for the anti-recoil device under the following conditions;

[0084] Figure 2 The recoil recovery speed-time curve of the anti-recoil device under fault-free conditions;

[0085] Figure 3 The curve showing the recoil recovery acceleration-time relationship of the anti-recoil device under fault-free conditions;

[0086] Figure 4 The curves showing the impact of ring wear failure on recoil and recoil strokes;

[0087] Figure 5 The curve showing the impact of oil leakage fault inside the recoil barrel on the dynamic stroke of the recoil mechanism;

[0088] Figure 6 The curve showing the impact of oil leakage fault inside the recoil barrel on the dynamic stroke of the recoil mechanism;

[0089] Figure 7 The curve showing the impact of external oil leakage fault of the retraction mechanism on the recoil recovery stroke of the recoil device;

[0090] Figure 8 The curve showing the influence of air leakage fault in the recoiler on the recoiler force Ff;

[0091] Figure 9 The curve showing the influence of wear on the inner diameter of the control ring on the recoil and precession stroke;

[0092] Figure 10 The curve showing the influence of ring inner diameter wear fault on recoil reaction force FR;

[0093] Figure 11 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the recoil and advance stroke (%L is the amount of oil leakage);

[0094] Figure 12 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the recoil recovery speed (%L is the amount of oil leakage);

[0095] Figure 13 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the minimum value of recoil recovery acceleration (%L is the amount of oil leakage);

[0096] Figure 14 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the recoil recovery time (%L is the amount of oil leakage);

[0097] Figure 15 The curve showing the influence of internal oil leakage fault in the piston of the recoil mechanism on the recoil and advance stroke (%L represents the amount of oil leakage);

[0098] Figure 16 The curve showing the effect of internal oil leakage fault in the piston of the recoil mechanism on the recoil speed (%L is the amount of oil leakage);

[0099] Figure 17 The curve showing the influence of internal oil leakage fault in the piston of the recoil mechanism on the recoil recovery time (%L is the amount of oil leakage);

[0100] Figure 18 The curve showing the effect of air leakage fault in the recoil chamber on the recoil recovery time (%V is the leakage amount);

[0101] Figure 19 The curve showing the effect of air leakage fault in the recoil chamber on the recoil force (%V is the leakage amount);

[0102] Figure 20 The curve showing the effect of external leakage fault in the hydraulic chamber of the recoil mechanism on the recoil force (%L represents the leakage amount).

[0103] Figure 21 This is an artificial neural network diagnostic model for the anti-recoil device according to the present invention. Detailed Implementation

[0104] Before explaining one or more embodiments of this disclosure in detail, it should be understood that the embodiments are not limited to the construction details in their specific applications, and the steps or methods presented in the following embodiments or drawings. The systems and methods of the present invention will now be described in detail with reference to the accompanying drawings.

[0105] A fault diagnosis method for anti-recoil devices based on multi-parameter fusion, the specific steps of which include:

[0106] Step 1: Based on Newton's second law, establish the differential equations for the recoil motion and the recoiling motion of the anti-recoil mechanism;

[0107] 1) The differential equation of the recoil motion is as follows:

[0108]

[0109] in:

[0110] m h For the mass of the rear seat portion;

[0111] x represents the recoil displacement;

[0112] v is the recoil speed;

[0113] t represents the recoil time.

[0114] F R For recoil resistance, and

[0115]

[0116] In formula (2):

[0117] For the angle of attack; F Φh It is the hydraulic resistance of the retractor, and its expression is:

[0118] F Φh =f(a x V 2 =f(a x0 +Δa x V 2 (3)

[0119] In formula (3):

[0120] a x The annular leak area at the control ring (where a is the annular leak area when the control ring is not worn) is the area of ​​the annular leak. x0 With the area Δa worn away x (The sum of the two values), where v is the recoil speed.

[0121] f(a x The expression for ) is

[0122]

[0123] In formula (4):

[0124] A1 is the minimum tributary cross-sectional area;

[0125] K1 and K2 are hydraulic resistance coefficients, K1 = 1.2 - 1.6, K2 = (2 - 4)K1;

[0126] A p To control the area of ​​the annular hole.

[0127] A fj The working area of ​​the controller. d1 is the inner diameter of the brake lever;

[0128] a x The area of ​​the fluid flow orifice. d x To control the outer diameter of the rod;

[0129] ρ is the density of the desiccant.

[0130] In formula (2):

[0131] F f For the reciprocating force, its expression is:

[0132]

[0133] In formula (5):

[0134] A f This refers to the working area of ​​the piston in the reciprocating machine.

[0135] p f0 This is the initial gas pressure in the gas storage chamber of the recoil engine;

[0136] V0 is the initial volume of gas in the gas storage chamber of the recoil engine;

[0137] x represents the recoil displacement.

[0138] In formula (2):

[0139] F represents the frictional force in the sealing element of the recoil mechanism.

[0140] F=νm h g (6)

[0141] In the formula: ν is the equivalent friction coefficient of the sealing link, which takes a value of 0.3-0.5.

[0142] In formula (2):

[0143] F T For the friction of the cradle guide rail,

[0144]

[0145] In the formula: f is the friction coefficient of the cradle guide rail, which takes a value of 0.16-0.20.

[0146] 2) The differential equation of the recursive motion is as follows:

[0147]

[0148] in:

[0149] m h For the mass of the rear seat portion;

[0150] x is the recoil displacement;

[0151] v is the reciprocating velocity;

[0152] t is the re-entry time;

[0153] F f For recoil mechanism;

[0154] The angle of attack;

[0155] In formula (8):

[0156] F Φfy It is the reciprocating hydraulic resistance, and its expression is:

[0157] F Φfy =F Φf +F Φfj +F Φkh (9)

[0158] in:

[0159] F Φf It is the hydraulic resistance provided by the recoil mechanism during the return-forward movement.

[0160]

[0161] In formula (10):

[0162] K 1f The hydraulic resistance coefficient of the retraction mechanism during the return stroke is typically 1.4-1.6.

[0163] A 0f The working area of ​​the piston in the retraction mechanism during the return stroke:

[0164]

[0165] a x To control the area of ​​the annular leak at the control ring;

[0166] v is the recoil speed.

[0167] In formula (9):

[0168] F Φfj The hydraulic resistance is provided by the reciprocating controller.

[0169]

[0170] In formula (12):

[0171] K 2f The hydraulic resistance coefficient of the reciprocating controller is generally K. 2f = (3-4)K 1f ;

[0172] A fj Working area of ​​the reciprocating regulator piston:

[0173]

[0174] a f This represents the area of ​​the groove through which the liquid passes by the inner cavity of the brake rod.

[0175]

[0176] v is the recoil speed.

[0177] In formula (9):

[0178] F Φkh It is the hydraulic resistance of the reciprocating control valve.

[0179]

[0180] In the formula:

[0181] K v The hydraulic resistance coefficient of the reciprocating control valve flow hole;

[0182] a v To control the working area of ​​the valve flow orifice:

[0183] A f This represents the working area of ​​the piston in the reciprocating machine.

[0184] Step 2: Substitute the basic parameters of the fault-free anti-recoil device into the recoil motion differential equation and the recoil motion differential equation of the recoil device, and set the angle of attack to zero. Solve the recoil motion differential equation and the recoil motion differential equation using the ode45 function of the fourth-order fifth-order Runge-Kutta method in MATLAB to obtain the fault-free sample dataset at zero angle of attack. The characteristic parameter curves formed by some of the fault-free sample datasets are shown below. Figures 1 to 3 As shown, where Figure 1 This is the recoil recovery displacement-time relationship curve of the anti-recoil device under fault-free conditions according to an embodiment of the present invention. Figure 2 The recoil recovery speed-time curve of the anti-recoil device under fault-free conditions; Figure 3 The curve showing the recoil recovery acceleration-time relationship of the anti-recoil device under fault-free conditions;

[0185] Step 3: Based on the differential equations of recoil motion and recoil motion of the anti-recoil device, establish a fault analysis and diagnosis model for the anti-recoil device. The fault analysis and diagnosis model reflects the nonlinear relationship between fault characteristic parameters (e.g., recoil and recoil displacement, velocity, acceleration, etc.) and the main fault types of the anti-recoil device, and reflects the mapping relationship between the main fault types of the anti-recoil device and its fault characteristic parameters.

[0186] Step 4: Develop a simulation program using MATLAB or FORTRAN to obtain the mapping relationship between typical faults of the anti-recoil device and their fault characteristic parameters, and form a mapping relationship dataset. Typical faults shown in this invention include control ring wear, oil leakage inside the recoil mechanism barrel, external oil leakage from the recoil mechanism, and air leakage from the return mechanism. Figures 4 to 8The simulation results for step 4 are presented. Among them, Figure 4 The curves showing the impact of ring wear failure on recoil and recoil strokes; Figure 5 The curve showing the impact of oil leakage fault inside the recoil barrel on the dynamic stroke of the recoil mechanism; Figure 6 The curve showing the impact of oil leakage fault inside the recoil barrel on the dynamic stroke of the recoil mechanism; Figure 7 The curve showing the impact of external oil leakage fault of the retraction mechanism on the recoil recovery stroke of the recoil device; Figure 8 The curve showing the influence of air leakage in the recoil mechanism on the recoil force Ff is used to define the fault characteristic parameters and fault types of the recoil device. The fault characteristic parameters preferably include: maximum recoil displacement, maximum recoil speed, maximum recoil force, and maximum retraction force; the fault types include, for example, the degree of wear of the control ring, the degree of air leakage in the recoil mechanism, and the degree of liquid leakage in the retraction mechanism.

[0187] Step 5: Based on the mapping relationship dataset between the typical faults and their fault characteristic parameters, obtain the trend curves showing the relationship between the extreme values ​​of the fault characteristic parameters and the severity of typical faults. Use these trend curves to form a fault test and training set. The fault characteristic parameters with extreme values ​​include recoil recovery stroke, recoil recovery velocity, recoil recovery acceleration, recoil reaction force FR, and recoil recovery time. The trend curves showing the relationship between their extreme values ​​and the severity of various faults are as follows: Figure 9 As shown in Figure 20. Among them, Figure 9 The curve showing the influence of wear on the inner diameter of the control ring on the maximum value of the recoil and precession stroke; Figure 10 The curve showing the influence of wear failure of the inner diameter ring on the maximum value of the recoil reaction force FR; Figure 11 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the maximum value of the recoil and advance stroke (%L is the amount of oil leakage); Figure 12 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the maximum recoil speed (%L is the amount of oil leakage); Figure 13 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the minimum value of recoil recovery acceleration (%L is the amount of oil leakage); Figure 14 The curve showing the influence of external oil leakage fault of the piston of the recoil mechanism on the maximum recoil recovery time (%L is the amount of oil leakage); Figure 15 The curve showing the influence of internal oil leakage fault in the piston of the recoil mechanism on the maximum value of the recoil and advance stroke (%L is the amount of oil leakage); Figure 16 The curve showing the influence of internal oil leakage fault in the piston of the recoil mechanism on the maximum recoil speed (%L represents the amount of oil leakage); Figure 17 The curve showing the influence of internal oil leakage fault in the piston of the recoil mechanism on the maximum recoil recovery time (%L is the amount of oil leakage); Figure 18 The curve showing the influence of air leakage fault in the recoil chamber on the maximum recoil recovery time (%V represents the leakage amount); Figure 19 The curve showing the influence of air leakage fault in the recoil chamber on the maximum recoil force (%V represents the leakage amount); Figure 20 The curve showing the influence of external leakage fault in the hydraulic chamber of the recoil mechanism on the maximum recoil force (%L represents the leakage amount).

[0188] Step 6: Based on the fault analysis and diagnosis model of the anti-recoil device, establish a multi-parameter fusion artificial neural network diagnosis model, and use the fault-free sample dataset and fault test and training set to train and optimize the artificial neural network diagnosis model to obtain the intelligent model for fault diagnosis of the anti-recoil device.

[0189] The artificial neural network diagnostic model of the anti-recoil device fault diagnosis method according to the present invention is as follows: Figure 21 As shown, this artificial neural network is as follows Figure 21 As shown, the diagnostic model comprises an input layer, an output layer, and a hidden layer. The input layer has four nodes, which represent the multiple parameters of this invention: maximum recoil displacement, maximum recoil speed, maximum recoil force, and maximum retraction force. The output layer contains three nodes, which represent the typical faults of this invention: the wear degree of the control ring, the air leakage degree of the recoil mechanism, and the liquid leakage degree of the retraction mechanism. The hidden layer has been adjusted to have seven nodes.

[0190] Step 7: Using the intelligent model for fault diagnosis of the anti-recoil device, input the working status data of the anti-recoil device, perform fault identification and diagnosis, and obtain the diagnosis results.

[0191] The working status data is acquired using vibration sensors, pressure sensors, etc., and the acquired working status data is saved for subsequent processing.

[0192] In step six, establishing a multi-parameter fusion artificial neural network diagnostic model includes: determining the number of nodes in each layer of the neural network by utilizing the mapping relationship between the main fault types of the anti-backseat device and their fault characteristic parameters; initializing the basic data of the artificial neural network, such as setting the iteration rounds, setting the loss target, and initializing the model coefficients; and determining the convergence algorithm and algorithm parameters, such as setting the initial step size and small parameters.

[0193] In step six, the steps for training and optimizing the artificial neural network diagnostic model include: inputting the training set into the initial neural network, learning with the help of the gradient descent algorithm, and finally obtaining the intelligent model for diagnosing the anti-recoil device fault; inputting the test set into the anti-recoil device fault diagnosis model; evaluating the prediction performance based on the results; and choosing to retrain after adjustment or accept the current training.

[0194] Although the invention has been described with reference to embodiments shown in the accompanying drawings, equivalent or alternative means may be used without departing from the scope of the claims. The components described and illustrated in this invention are merely examples of systems / apparatus and methods that can be used to implement embodiments of this disclosure, and may be replaced with other devices and components without departing from the scope of the claims.

Claims

1. A fault diagnosis method for anti-recoil devices based on multi-parameter fusion, comprising: Based on Newton's second law, establish the differential equations for the recoil motion and the recoil motion of the anti-recoil device; Substitute the basic parameters of the fault-free anti-recoil device into the recoil motion differential equation and the recoil motion differential equation of the recoil device, and take the angle of attack as zero. Solve the recoil motion differential equation and the recoil motion differential equation using the ode45 function of the fourth-level fifth-order Runge-Kutta method in MATLAB to obtain the fault-free sample dataset under zero angle of attack. Based on the differential equations of recoil motion and recoil motion of the anti-recoil device, a fault analysis and diagnosis model for the anti-recoil device is established. A simulation program for the fault analysis and diagnosis model is developed using MATLAB or FORTRAN. Through the simulation program, the mapping relationship between typical faults of the anti-recoil device and their fault characteristic parameters is obtained, and a mapping relationship dataset is formed. Based on the mapping relationship dataset between the typical faults and their fault characteristic parameters, the relationship trend curve between the extreme values ​​of the fault characteristic parameters and the degree of occurrence of typical faults is obtained, and the relationship trend curve is used to form a fault test and training set. Based on the fault analysis and diagnosis model of the anti-recoil device, a multi-parameter fusion artificial neural network diagnosis model is established, and the artificial neural network diagnosis model is trained and optimized using a fault-free sample dataset and a fault test and training set to obtain an intelligent model for fault diagnosis of the anti-recoil device. Using the intelligent model for fault diagnosis of the anti-recoil device, the working status data of the anti-recoil device is input to perform fault identification and diagnosis, and obtain the diagnosis results.

2. The method for fault diagnosis of anti-recoil devices based on multi-parameter fusion as described in claim 1, characterized in that, The differential equation for recoil motion is: in: m h For the mass of the rear seat portion; x represents the recoil displacement; v is the recoil speed; t is the recoil time; F R For recoil resistance, and in: F f F is the recoil force; F is the frictional force of the sealing element in the recoil mechanism; F T The friction force of the cradle guide rail; For the angle of attack; F Φh It is the hydraulic resistance of the retractor, and its expression is... F Φh =f(a x )IN 2 =f(a x0 +Δa x )IN 2 in: a x Let be the area of ​​the annular leak at the control ring, and let a be the area of ​​the annular leak when the control ring is not worn. x0 With the area Δa worn away x The sum of these, where v is the recoil velocity; f(a x The expression for ) is in: A1 is the minimum tributary cross-sectional area; K1 and K2 are hydraulic resistance coefficients. The value of K1 is between 1.2 and 1.6, and the value of K2 is 2 to 4 times that of K1. A p To control the area of ​​the annular hole. A fj The working area of ​​the controller. d1 is the inner diameter of the brake lever; a x The area of ​​the fluid flow orifice. d x To control the outer diameter of the rod; ρ is the density of the desiccant.

3. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 2, characterized in that, The F f For the reciprocating force, its expression is: in: A f This refers to the working area of ​​the piston in the reciprocating machine. p f0 This is the initial pressure of the gas in the gas storage chamber of the recoil engine; V0 is the initial volume of gas in the gas storage chamber of the recoil engine; x represents the recoil displacement; F represents the frictional force in the sealing element of the recoil mechanism. F=νm h g In the formula: ν is the equivalent friction coefficient of the sealing link, which takes a value of 0.3-0.5; The F T For the friction of the cradle guide rail, In the formula: f is the friction coefficient of the cradle guide rail, which takes a value of 0.16-0.

20.

4. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The differential equation of the recursive motion is as follows: in: m h For the mass of the rear seat portion; x f For recurrent displacement; v f For the reciprocating velocity; t f For the time to resume; F f F is the recoil force; F is the frictional force of the sealing element in the recoil mechanism; F T The friction force of the cradle guide rail; The angle of attack; in: It is the reciprocating hydraulic resistance, and its expression is: F Φfy =F Φf +F Φfj +F Φkh in: It is the hydraulic resistance provided by the reciprocating controller; It is the hydraulic resistance of the reciprocating control valve; It is the hydraulic resistance provided by the recoil mechanism during the return-forward movement. in: ρ is the density of the desiccant; K 1f The hydraulic resistance coefficient of the retraction mechanism during the return stroke ranges from 1.4 to 1.

6. A 0f The working area of ​​the piston in the retraction mechanism during the return stroke: a x To control the area of ​​the annular leak at the control ring; v f This refers to the reciprocating speed.

5. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 4, characterized in that, The The hydraulic resistance is provided by the reciprocating controller: in: K 2f K is the hydraulic resistance coefficient of the reciprocating controller. 2f The value is K 1f 3-4 times; A fj Working area of ​​the reciprocating regulator piston: a f This represents the area of ​​the groove through which the liquid passes by the inner cavity of the brake rod. V f For the reciprocating velocity; The It is the hydraulic resistance of the reciprocating control valve: in: K v The hydraulic resistance coefficient of the reciprocating control valve flow hole; a v To control the working area of ​​the valve flow orifice; A f This represents the working area of ​​the piston in the reciprocating machine.

6. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The fault-free sample dataset includes: the recoil recovery displacement-time curve of the recoil mechanism under fault-free conditions, the recoil recovery velocity-time curve of the recoil mechanism under fault-free conditions, and the recoil recovery acceleration-time curve of the recoil mechanism under fault-free conditions.

7. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The fault analysis and diagnosis model of the recoil mechanism reflects the nonlinear relationship between fault characteristic parameters and the main fault types of the recoil mechanism, and reflects the mapping relationship between the main fault types of the recoil mechanism and its fault characteristic parameters. The fault characteristic parameters include the maximum recoil displacement, the maximum recoil speed, the maximum recoil force, and the maximum retraction force. The main fault types include: the degree of wear of the control ring, the degree of air leakage of the recoil mechanism, and the degree of liquid leakage of the retraction mechanism.

8. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The fault test and training set includes: the influence curve of the wear fault of the inner diameter of the control ring on the maximum value of the recoil return stroke; the influence curve of the wear fault of the inner diameter of the control ring on the maximum value of the recoil reaction force FR; the influence curve of the external oil leakage fault of the retractor piston on the maximum value of the recoil return stroke; the influence curve of the external oil leakage fault of the retractor piston on the maximum value of the recoil return speed; and the influence curve of the external oil leakage fault of the retractor piston on the minimum value of the recoil return acceleration.

9. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The artificial neural network diagnostic model comprises an input layer, an output layer, and a hidden layer. The input layer has four nodes: maximum recoil displacement, maximum recoil velocity, maximum recoil force, and maximum retraction force. The output layer has three nodes: control ring wear degree, recoil air leakage degree, and retraction liquid leakage degree. The hidden layer has been adjusted to have seven nodes.

10. The method for fault diagnosis of anti-recoil device based on multi-parameter fusion as described in claim 1, characterized in that, The working status data is acquired using vibration sensors and pressure sensors. The acquired working status data is saved and preprocessed, and then sent to the fault diagnosis intelligent model of the anti-recoil device for fault diagnosis. The steps for training and optimizing the artificial neural network diagnostic model include: inputting the training set into the initial neural network, learning with the help of the gradient descent algorithm, and adjusting the neural network model coefficients; inputting the test set into the anti-recoil device fault diagnosis model, evaluating the prediction performance based on the results, and choosing to train again after adjustment or accept this training to determine the fault diagnosis intelligent model.

Images