Artillery recoil brake state evaluation method based on cusum-rhd

Abstract

The application discloses a CUSUM-RHD-based artillery recoil mechanism state evaluation method, and belongs to the technical field of artillery recoil mechanism state evaluation. The method comprises the following steps: analyzing and extracting a fault characteristic data signal; performing normalization processing on the fault characteristic data, so as to eliminate the influence between dimensions; constructing a relative Hausdorff distance calculation model to calculate the Hausdorff distance between two adjacent shootings of the artillery; constructing a cumulative sum relative Hausdorff distance calculation model to calculate the cumulative sum relative Hausdorff distance, and defining the cumulative sum relative Hausdorff distance as a health factor for evaluating the state of the artillery recoil mechanism; and finally, performing state evaluation on the artillery recoil mechanism according to the health factor index. The state evaluation method has good evaluation accuracy, robustness and anti-interference ability. The application also has the advantage that the state of the artillery recoil mechanism can be evaluated without obtaining the full life cycle data.

Description

TECHNICAL FIELD

[0001] The application relates to a CUSUM-RHD-based state evaluation method for a recoil device of a gun. BACKGROUND

[0002] The gun is widely used in various military services and has a broad military application market due to its significant advantages in combat continuity, simple use and low cost. The gun has a complex structure and is composed of a recoil device, a bullet supply and a barrel. The recoil device is called the heart of the gun and is a key component for controlling the shooting and movement of the gun. The working condition of the recoil device is complex and changeable, and the recoil device is a high-frequency fault component among all gun components. If the fault of the recoil device is not monitored and disposed in time, the shooting accuracy of the bullet will be seriously affected, and the overall combat performance of the gun will be reduced. Therefore, the state evaluation of the recoil device of the gun has important military significance.

[0003] The state evaluation of equipment usually includes the steps of collecting equipment signals, extracting feature signal data, data preprocessing, constructing health factors and monitoring evaluation. Many scholars have carried out a large amount of state evaluation work on bearings in rotating machinery. The bearing data is easy to obtain in the whole life cycle. After extracting the bearing signal features, the maximum and minimum normalization method is usually used for data preprocessing, and then the traditional distance measurement indexes such as Euclidean distance and Mahalanobis distance are used to construct health factors. The state of the bearing is evaluated by the size of the health factor. However, due to the limitations of the real military environment, it is difficult to obtain the whole life cycle data of the recoil device. Therefore, the commonly used maximum and minimum normalization data processing method cannot be adapted to the data processing of the gun recoil device. In addition, the gun recoil device belongs to reciprocating machinery. Unlike rotating machinery, the recoil device needs to construct health factors based on multiple features. Therefore, the traditional Euclidean distance and Mahalanobis distance indexes cannot be used for the health factor construction of the gun recoil device. Moreover, there is little state evaluation work on the gun recoil device at present. SUMMARY

[0004] The application aims at providing a CUSUM-RHD-based artillery recoil device state evaluation method, which can solve the problems of difficult data preprocessing and difficult health factor construction of the characteristic signals of the artillery recoil device.

[0005] The application aims at providing a CUSUM-RHD-based artillery recoil device state evaluation method, which can solve the problems of difficult data preprocessing and difficult health factor construction of the characteristic signals of the artillery recoil device.

[0006] The CUSUM-RHD-based artillery recoil device state evaluation method disclosed by the application comprises the following steps.

[0007] Step 1: determining the fault characteristic signals of the artillery recoil device.

[0008] The speed-time curve and the displacement-time curve of the artillery recoil device under different fault states are obtained, and four characteristic values, i.e., the inflection point (the maximum recoil displacement Xmax) of the speed-time curve, the inflection point (the maximum recoil speed Vmax and the maximum return speed Umax) and the endpoint (the return-to-position speed Uend) of the displacement-time curve, are taken to represent the fault state of the artillery recoil device.

[0009] Step 2: the different fault characteristic values of the artillery recoil device selected in step 1 have different dimension ranges, and the initial value method is adopted to normalize the different fault characteristic value data, so as to eliminate the dimension influence of the four different fault characteristic values, facilitate the subsequent step 3 to uniformly solve the Hausdorff distance between the adjacent two shootings of the artillery, and further uniformly process the different fault characteristic value data by using the initial value method, compared with the traditional maximum and minimum normalization method which has no special limitation on the fault characteristic value data, and the artillery recoil device state can still be evaluated without obtaining the full life cycle data.

[0010] In step 2, the initial value method is adopted to uniformly process the different fault characteristic value data, and the normalization processing formula is as follows:

[0011]

[0012] where i = 1, 2, 3, 4, respectively represent the maximum setback displacement X max , the maximum setback velocity V max , the maximum ramming velocity U max and the ramming to position velocity Uend four fault characteristics; j = 1, 2, …, n, n represents the number of gun firing times; x i j represents the data value of the fault characteristic i of the gun recoil system after the jth firing.

[0013] Step 3: The four fault characteristic data of the gun after each firing, the maximum setback displacement X max , the maximum setback velocity V max , the maximum ramming velocity U max and the ramming to position velocity Uend are normalized by using the initialization method in step 2 to construct the Hausdorff space corresponding to each firing of the gun; the Hausdorff distance (HD) between the Hausdorff spaces corresponding to the adjacent two firings of the gun is solved based on the constructed Hausdorff space corresponding to each firing.

[0014] Step 3.1: Construction of the Hausdorff space.

[0015] The four fault characteristic data of the gun after each firing, the maximum setback displacement X max , the maximum setback velocity V max , the maximum ramming velocity U max and the ramming to position velocity Uend are normalized by using the initialization method in step 2 to construct the Hausdorff space corresponding to each firing of the gun; the Hausdorff distance (HD) between the Hausdorff spaces corresponding to the adjacent two firings of the gun is solved based on the constructed Hausdorff space corresponding to each firing. The Hausdorff space corresponding to the jth firing of the gun is constructed, expressed as:

[0016]

[0017] Step 3.2: Solution of the Hausdorff distance.

[0018] Based on step 3.1, the Hausdorff space corresponding to the j+1th firing of the gun is The Hausdorff distance (HD) between the Hausdorff spaces corresponding to the jth and j+1th firings of the gun is solved, expressed as:

[0019]

[0020] In formula (3), h(space j ,space j+1) represents space j to space j+1 , while HD(space j , space j+1 ) represents the Hausdorff distance between the jth and (j+1)th shots of the gun.

[0021] Step 4: On the basis of the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shots of the gun obtained in step 3, a relative Hausdorff distance calculation model (RHD) is constructed, the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shots of the gun is calculated in two directions through the relative Hausdorff distance calculation model, and the Hausdorff distances between the Hausdorff spaces corresponding to the adjacent two shots of the gun in two directions are averaged to obtain the relative Hausdorff distance containing data information of two directions.

[0022] HD(space j , space j+1 ) obtained in step 3.2 only considers a maximum one-way distance between space j and space j+1 , and does not consider the data information contained in the other one-way distance, so the directionality of space j and space j+1 is considered, and a relative Hausdorff distance (RHD) calculation model as shown in formula (4) is constructed:

[0023]

[0024] According to the relative Hausdorff distance calculation model, the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shots of the gun is calculated in two directions through the relative Hausdorff distance calculation model, and the Hausdorff distances between the Hausdorff spaces corresponding to the adjacent two shots of the gun in two directions are averaged to obtain the relative Hausdorff distance between the adjacent two shots (jth and (j+1)th) of the gun as RHD(space j , space j+1 ), which contains data information of two directions, can effectively mine the information between the adjacent two shots of the gun, and is beneficial to improve the reliability of the state evaluation result of the recoil device.

[0025] Step 5: Based on the relative Hausdorff distance calculated in step 4, a cumulative sum (CUSUM) method is applied to construct a cumulative sum relative Hausdorff distance (CUSUM-RHD) calculation model. The relative Hausdorff distance for evaluating the state of the recoil mechanism is accumulated through the cumulative sum relative Hausdorff distance calculation model. As the number of artillery firings increases, the accumulated relative Hausdorff distance for evaluating the state of the recoil mechanism realizes a differentiated amplification effect and information accumulation effect, further improving the accuracy, robustness, and anti-interference ability of the state evaluation of the recoil mechanism.

[0026] Based on the relative Hausdorff distance (RHD) calculated in step 4 and the cumulative sum (CUSUM) method, a cumulative sum relative Hausdorff distance (CUSUM-RHD) solution model is constructed. When the artillery is fired once, the cumulative sum relative Hausdorff distance is defined as 0. When the artillery is fired more than once, the cumulative sum relative Hausdorff distance solution model is:

[0027]

[0028] In formula (5), n represents the number of artillery firings, and n > 1.

[0029] Step 6: The cumulative sum relative Hausdorff distance calculated in step 5 is defined as a health index (HI) for evaluating the state of the recoil mechanism, which is represented as:

[0030] HI n = CUSUM-RHD n (6)

[0031] The smaller the health index, the healthier the recoil mechanism. By monitoring the health index, the state of the recoil mechanism is evaluated, further improving the accuracy, robustness, and anti-interference ability of the state evaluation of the recoil mechanism.

[0032] As a preferred embodiment, HI ∈ (0, 0.1] is the initial failure period of the recoil mechanism, HI ∈ (0.1, 0.3] is the general failure period of the recoil mechanism, and HI > 0.3 is the serious failure period of the recoil mechanism.

[0033] Advantages:

[0034] 1. The data normalization method involved in the existing state evaluation technology mostly needs the full life cycle data of the known equipment, however, it is difficult to obtain the full life cycle data for weapons such as artillery, and therefore the existing normalization data processing method (such as the maximum and minimum normalization method) is not suitable for the data processing of the artillery recoil device; the state evaluation method of the artillery recoil device based on CUSUM-RHD disclosed in the application is that, in view of the fact that different fault characteristic values of the artillery recoil device have different dimension ranges, the initial value method is used to normalize the different fault characteristic value data, so as to eliminate the dimension influence of the four different fault characteristic values, so as to facilitate the unified solution of the Hausdorff distance between the adjacent two shootings of the artillery, and in addition, the initial value method is used to homogenize the different fault characteristic value data, compared with the normalization method which has no special requirements for the fault characteristic value data, the state of the artillery recoil device can still be evaluated without obtaining the full life cycle data.

[0035] 2. The Euclidean distance and Mahalanobis distance based health factor used in the traditional state evaluation cannot represent the state of the artillery recoil device; the state evaluation method of the artillery recoil device based on CUSUM-RHD disclosed in the application is that, on the basis of realizing the beneficial effect 1, the Hausdorff space corresponding to each shooting of the artillery is constructed, and the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shootings of the artillery is solved; on the basis of obtaining the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shootings of the artillery, a relative Hausdorff distance solving model is constructed, the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shootings of the artillery is solved in two directions through the relative Hausdorff distance solving model, and the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two shootings of the artillery in the two directions is averaged to obtain the relative Hausdorff distance containing the data information of the two directions. The relative Hausdorff distance RHD contains the data information of the two directions, which can effectively mine the information between the adjacent two shootings of the artillery, and is beneficial to improving the reliability of the state evaluation result of the recoil device.

[0036] 3. The state evaluation method of the artillery recoil device based on CUSUM-RHD disclosed in the application is that, on the basis of obtaining the relative Hausdorff distance, a cumulative sum (CUSUM) method is used to construct a cumulative sum relative Hausdorff distance solving model, the relative Hausdorff distance used for the state evaluation of the artillery recoil device is accumulated through the cumulative sum relative Hausdorff distance model, with the increase of the shooting times of the artillery, the relative Hausdorff distance used for the state evaluation of the artillery recoil device accumulated realizes the differential amplification effect and the information accumulation effect, and further improves the state evaluation precision, robustness and anti-interference ability of the artillery recoil device.

[0037] 4. The CUSUM-RHD-based artillery recoil device state evaluation method disclosed in the application, based on the beneficial effect 3, takes the cumulative sum and the relative Hausdorff distance as health factors for evaluating the state of the artillery recoil device, divides the health factors into different failure periods of the recoil device, defines HI∈(0, 0.1] as the initial failure period of the recoil device, HI∈(0.1, 0.3] as the general failure period of the recoil device, and HI>0.3 as the serious failure period of the recoil device, and according to the corresponding failure type, corresponding failure treatment measures are taken to improve the efficiency of failure treatment and the safety of the artillery.

[0038] 5. The CUSUM-RHD-based artillery recoil device state evaluation method disclosed in the application, by analyzing the velocity-time curve and the displacement-time curve of multiple groups of artillery recoil devices under different failure states, takes four characteristic values, i.e., the inflection point (the maximum recoil displacement Xmax) of the velocity-time curve, the inflection point (the maximum recoil velocity Vmax and the maximum return velocity Umax) and the endpoint (the return-to-position velocity Uend) of the displacement-time curve, to represent the failure state of the artillery recoil device, the four characteristic values are easy to obtain and can improve the sensitivity of failure representation, and the relative Hausdorff distance model is constructed based on the four characteristic values for the first time, which can further improve the accuracy of the state evaluation result of the artillery recoil device. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a CUSUM-RHD-based artillery recoil device state evaluation method flowchart of the application;

[0040] Figure 2 is a velocity-time curve of the recoil device under different failure states;

[0041] Figure 3 is a velocity-displacement curve of the recoil device under different failure states;

[0042] Figure 4 is the change of HI with the number of shots. DETAILED DESCRIPTION

[0043] In order to better illustrate the purpose and advantages of the application, the content of the application is further described below in combination with the drawings and examples.

[0044] As Figure 1 shown; the CUSUM-RHD-based artillery recoil device state evaluation method disclosed in the embodiment is specifically implemented as follows:

[0045] Step 1: Determine the failure characteristic signal of the artillery recoil device:

[0046] Obtain the velocity-time curves and displacement-time curves of the artillery recoil mechanism under different fault conditions. Take the velocity-time curve, such as... Figure 2 , Figure 3 As shown, observe Figure 2 and Figure 3 It can be seen that there are significant differences in the data values ​​at the inflection points of the velocity-time curve, the inflection points of the displacement-time curve, and the endpoints. Therefore, four characteristic values ​​are taken to characterize the fault state of the recoil device of the artillery: the inflection point of the velocity-time curve (maximum recoil displacement Xmax), the inflection point of the displacement-time curve (maximum recoil velocity Vmax, maximum recoil velocity Umax), and the endpoint (recovery speed Uend). Then, the fault characteristic signal data of the recoil device under different fault states (three types of data correspond to three fault states) are given, as shown in Table 1.

[0047] Table 1 shows the three types of data obtained.

[0048]

[0049] Table 1 lists the data of the recoil mechanism from the normal state to the occurrence of three typical fault states (control ring wear, recoil mechanism air leakage, and brake rod piston wear). Each type of data records four fault signal data of the recoil mechanism from the normal state to the occurrence of a certain fault state.

[0050] Step 2: Normalization of fault characteristic data

[0051] Different fault characteristic values ​​have different dimensional ranges. Therefore, an initialization method is used to normalize the data of different fault characteristic values ​​to eliminate the influence of dimensions. The formula is as follows:

[0052]

[0053] In equation (1), i = 1, 2, 3, 4, representing the four fault characteristics: maximum recoil displacement Xmax, maximum recoil velocity Vmax, maximum recoil velocity Umax, and recoil end velocity Uend, respectively; j = 1, 2, ..., n, where n represents the number of artillery shots. The data value represents the fault characteristic i of the artillery recoil device after the j-th firing. Taking the I-th type of data as an example, some of the results obtained are shown in Table 2.

[0054] Table 2. Data of Type I after standardization.

[0055]

[0056] Step 3: Calculate the Hausdorff Distance (HD) between two consecutive shots from the artillery.

[0057] Step 3.1: Constructing Hausdorff space

[0058] The four fault feature data of maximum recoil displacement Xmax, maximum recoil velocity Vmax, maximum ramming velocity Umax and ramming to position velocity Uend after each shot of the gun are normalized by using the initialization method in step 2, and the normalized fault feature data after the jth shot is: The Hausdorff space corresponding to the jth shot of the gun is constructed, denoted as:

[0059]

[0060] Taking the type I data as an example, the constructed Hausdorff space is shown in Table 3.

[0061] Table 3 Hausdorff space corresponding to type I data

[0062]

[0063] Step 3.2: Solving Hausdorff distance

[0064] Based on step 3.1, the Hausdorff space corresponding to the j+1th shot of the gun is: The Hausdorff distance (HD) between the Hausdorff spaces corresponding to the jth and j+1th shots of the gun is solved, and the expression is:

[0065]

[0066] In formula (3), h(space j ,space j+1 ) represents the distance from space j to space j+1 , and HD(space j ,space j+1 ) represents the Hausdorff distance after the j+1th shot of the gun. The Hausdorff distance between the adjacent two shots is shown in Table 4.

[0067] Table 4 Hausdorff distance

[0068]

[0069] Step 4: Constructing relative Hausdorff distance (RHD) to obtain the model

[0070] The HD(space j ,space j+1) only considering space j With space j+1 , the data information contained in the other one-way distance is not considered, so all one-way distances between space j and space j+1 are considered to construct a relative Hausdorff distance (RHD) calculation model, and the expression is:

[0071]

[0072] According to the relative Hausdorff distance calculation model, the Hausdorff distance between the Hausdorff spaces corresponding to the adjacent two times of firing of the gun is calculated in two directions by the relative Hausdorff distance calculation model, and the Hausdorff distances between the Hausdorff spaces corresponding to the adjacent two times of firing of the gun in two directions are averaged to obtain the relative Hausdorff distance between the adjacent two times of firing (the jth and the j+1th) of the gun, which is RHD(space j ,space j+1 ), which contains two-way data information, can effectively mine the information between the adjacent two times of firing of the gun, and is beneficial to improve the reliability of the state evaluation result of the recoil device.

[0073] The relative Hausdorff distance between the adjacent two times of firing is shown in Table 5.

[0074] Table 5 Relative Hausdorff distance

[0075]

[0076] Step 5: Constructing a cumulative sum relative Hausdorff distance (CUSUM-RHD) calculation model

[0077] Based on the relative Hausdorff distance calculated in step 4, a cumulative sum (CUSUM) method is applied to construct a cumulative sum relative Hausdorff distance calculation model, which accumulates the relative Hausdorff distance used for gun recoil device state evaluation. With the increase of the number of gun firing, the accumulated relative Hausdorff distance used for gun recoil device state evaluation realizes the differential amplification effect and information accumulation effect, further improves the state evaluation precision, robustness and anti-interference ability of the gun recoil device.

[0078] Based on the relative Hausdorff distance (RHD) and the cumulative sum (CUSUM) method, a CUSUM-RHD model was constructed. The CUSUM-RHD was defined as 0 when the first shot was fired, and the model for calculating the CUSUM-RHD was as follows:

[0079]

[0080] The CUSUM-RHD is shown in Table 6.

[0081] Table 6 CUSUM-RHD

[0082]

[0083]

[0084] Step 6: The CUSUM-RHD calculated in step 5 was defined as a health index (HI) for evaluating the state of the recoil brake, and was expressed as:

[0085] HI n = CUSUM-RHD n (6)

[0086] The smaller the health index, the healthier the recoil brake. By monitoring the health index, the state of the recoil brake can be evaluated.

[0087] Table 7 Health index

[0088]

[0089] The evaluation of the health index mainly considered its monotonicity and robustness. After the gun was fired, the performance of the recoil brake would irreversibly degrade without human maintenance. Therefore, the constructed HI should have monotonicity. The formula for solving the monotonicity was as follows:

[0090]

[0091] In formula (7), f'(HI) represents the derivative value corresponding to the HI value; the monotonicity of the HI ranges from 0 to 1, and the closer the value of Mon is to 1, the better the monotonicity of the health index.

[0092] Due to the complexity of artillery operation, randomness and uncertainty may exist during data acquisition, interfering with the evaluation of HI (High Intensity Level). A good HI must be robust to disturbances during data acquisition. The formula for calculating robustness is:

[0093]

[0094] In equation (8), smoothed HI j The smoothed value of HI after the j-th shot of the artillery is represented by the moving average smoothing method, with a window size of 10. The robustness of HI ranges from Rob∈[0,1]. The closer the value of Rob is to 1, the better the robustness of the health factor.

[0095] To verify the effectiveness of the health factors constructed in this invention, the analysis results of health factors based on Hausdorff distance (HD) and relative Hausdorff distance (RHD) were compared, as shown in Table 8.

[0096] Table 8 Comparison results of different types of health factors

[0097]

[0098] As shown in Table 8, the health factor constructed in this invention has good monotonicity and robustness, and can be used for the status assessment of artillery recoil devices.

[0099] Based on the constructed health factors, the changes in health factors with the number of shots during artillery firing are solved. The health factor thresholds for different failure periods can usually be determined by expert experience, and the recoil device can be evaluated for its condition.

[0100] Preferably, HI∈(0,0.1] is given as the initial failure period of the recoil device, HI∈(0.1,0.3] is given as the general failure period of the recoil device, and HI>0.3 is given as the severe failure period of the recoil device.

[0101] The following is a graph showing the trend of health factors with the number of shots in the three types of data. Figure 4 .

[0102] From the appendix Figure 4It can be seen that the three types of data are basically consistent with the change trend of the corresponding health factors, and the health factors are gradually increasing, which represents that the performance state of the recoil brake device gradually declines after different faults occur. However, the growth rates of the three types of data are not consistent, and the size of the amplitude represents the speed of performance degradation of the recoil brake device. Among them, the growth rate of the II type data (normal data + recoil machine air leakage) is the smallest, indicating that the recoil machine air leakage fault has less impact on the performance of the recoil brake device, and the growth rate of the III type data (normal data + brake rod piston wear) is the largest, indicating that the brake rod piston wear has the greatest impact on the performance of the recoil brake device.

[0103] In addition, for the I type data, after the artillery shoots 200 times, the state of the recoil brake device enters the initial failure period, and at point A, after shooting 687 times, the recoil brake device enters the general failure period; for the II type data, after the artillery shoots 200 times, the state of the recoil brake device enters the initial failure period, and after subsequent shooting of 1000 times, the recoil brake device is still in the initial failure period; for the III type data, after the artillery shoots 200 times, the state of the recoil brake device enters the initial failure period, and at point B, after shooting 371 times, the state of the recoil brake device enters the general failure period, and at point B, after shooting 755 times, the recoil brake device enters the serious failure period. When the artillery is in the initial failure period, the combat performance degradation of the artillery is not serious, and no intervention is needed, but when the artillery enters the general failure period, the shooting accuracy and efficiency of the artillery will be affected, and the fault should be checked in time to prevent the artillery from entering the serious failure period and to avoid more serious military accidents.

[0104] The above specific description of the disclosure further illustrates the purpose, technical scheme and effective effect of the invention, but the embodiments of the invention are not limited thereto, and any modification made within the spirit and principles of the invention should be included in the protection scope of the invention.

Claims

1. A method for assessing the condition of artillery recoil devices based on CUSUM-RHD, characterized in that, Includes the following steps: Step 1: Identify the fault characteristic signals of the artillery recoil mechanism; Obtain velocity-time curves and displacement-time curves of the artillery recoil device under different fault states, and take four characteristic values, namely the inflection point of the velocity-time curve, the inflection point and the endpoint of the displacement-time curve, to characterize the fault state of the artillery recoil device. The inflection point of the velocity-time curve is the maximum recoil displacement Xmax. The inflection point of the displacement-time curve includes the maximum recoil velocity Vmax and the maximum recovery velocity Umax. The endpoint is the recovery velocity Uend. Determine the fault characteristic signals of the artillery recoil device; obtain the velocity-time curve and displacement-time curve of the artillery recoil device under different fault states, and take four characteristic values ​​to characterize the fault state of the artillery recoil device: the inflection point Xmax of the velocity-time curve, the inflection point Vmax of the displacement-time curve, the maximum recoil velocity Umax, and the endpoint Uend. Step 2: Since the different fault characteristic values ​​of the artillery recoil device selected in Step 1 have different dimension ranges, the initialization method is used to normalize the different fault characteristic value data to eliminate the influence of the four different fault characteristic values' dimensions. This is to facilitate the unified solution of the Hausdorff distance between two adjacent shots of the artillery in the subsequent Step 3. In addition, the initialization method is used to unify the different fault characteristic value data. Compared with the traditional maximum-minimum normalization method, it has no special restrictions on the fault characteristic value data and can still evaluate the state of the artillery recoil device without obtaining full life cycle data. Step 3: Normalize the four fault characteristic data after each artillery firing—maximum recoil displacement Xmax, maximum recoil velocity Vmax, maximum recoil velocity Umax, and recoil return velocity Uend—using the initialization method in Step 2 to construct the Hausdorff space corresponding to each artillery firing; and solve the Hausdorff distance between the Hausdorff spaces corresponding to two adjacent artillery firings based on the constructed Hausdorff space corresponding to each firing. Step 4: Based on the Hausdorff distance between two adjacent artillery shots obtained in Step 3, construct a relative Hausdorff distance calculation model. Calculate the Hausdorff distance between two adjacent artillery shots in both directions using the relative Hausdorff distance calculation model. Then, average the Hausdorff distances between two adjacent artillery shots in the two directions to obtain the relative Hausdorff distance containing data information from both directions. Step 5: Based on the relative Hausdorff distance obtained in Step 4, the cumulative sum CUSUM method is applied to construct a cumulative sum relative Hausdorff distance calculation model. The cumulative sum relative Hausdorff distance used for the condition assessment of the artillery recoil device is accumulated through the cumulative sum relative Hausdorff distance calculation model. As the number of artillery shots increases, the accumulated relative Hausdorff distance used for the condition assessment of the artillery recoil device achieves a differential amplification effect and information accumulation effect, further improving the accuracy, robustness and anti-interference ability of the artillery recoil device condition assessment. Step 6: Define the cumulative sum and relative Hausdorff distance obtained in Step 5 as the health factor HI used to evaluate the state of the artillery recoil device, and complete the state evaluation work.

2. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 1, characterized in that, The different fault characteristic values ​​in step 2 have different dimensional ranges. Therefore, an initialization method is used to normalize the data of different fault characteristic values. The formula is as follows: In equation (1), i = 1, 2, 3, 4, representing the four fault characteristics: maximum recoil displacement Xmax, maximum recoil velocity Vmax, maximum recoil velocity Umax, and recoil end velocity Uend, respectively; j = 1, 2, ..., n, where n represents the number of artillery shots. The data value represents the fault characteristic i of the artillery recoil device after the j-th shot.

3. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 2, characterized in that, Step 3 includes the following steps: Step 3.1: Construct Hausdorff space: The four fault characteristic data after each artillery firing—maximum recoil displacement Xmax, maximum recoil velocity Vmax, maximum recoil velocity Umax, and recoil return velocity Uend—are normalized using the initialization method in step 2. Therefore, the normalized fault characteristic data after the j-th firing is as follows: Construct the Hausdorff space corresponding to the j-th firing of the cannon, denoted as: Step 3.2: Solve for the Hausdorff distance: Based on step 3.1, the Hausdorff space corresponding to the (j+1)th shot of the cannon can be obtained as follows: The Hausdorff distance (HD) between the Hausdorff spaces corresponding to the j-th and j+1-th artillery shots is expressed as follows: In equation (3), h(space) j ,space j+1 ) represents space j to space j+1 The distance, and at the same time HD (sapce) j ,space j+1 ) represents the Hausdorff distance after the j-th and j+1-th shots of the artillery.

4. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 3, characterized in that, Step 4 includes the following steps: The HD (space) obtained in step 3.2 j ,space j+1 Only space is considered j With space j+1 The maximum one-way distance between them is considered, without taking into account the data information contained in the other one-way distance. Therefore, space is considered. j With space j+1 To determine the directionality, a relative Hausdorff distance (RHD) calculation model is constructed as shown in formula (4): Based on the relative Hausdorff distance calculation model, the Hausdorff distance between two adjacent shots of the artillery is calculated bidirectionally using this model. The average of these Hausdorff distances in both directions is then used to obtain the relative Hausdorff distance between two adjacent shots (the j-th shot and the (j+1)-th shot), which is RHD(sapce) j ,space j+1 By utilizing the relative Hausdorff distance (RHD) which contains two directional data information, it is possible to effectively extract information between two adjacent shots of the artillery, which is beneficial to improving the reliability of the recoil device status assessment results.

5. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 4, characterized in that, Step 5 includes the following steps: Based on the relative Hausdorff distance obtained in step 4, the cumulative sum CUSUM method is applied to construct a cumulative sum relative Hausdorff distance calculation model. The cumulative sum relative Hausdorff distance used for the condition assessment of the artillery recoil device is accumulated through the cumulative sum relative Hausdorff distance calculation model. As the number of artillery shots increases, the accumulated relative Hausdorff distance used for the condition assessment of the artillery recoil device achieves a differential amplification effect and information accumulation effect, further improving the accuracy, robustness and anti-interference ability of the condition assessment of the artillery recoil device. Based on the Relative Hausdorff Distance (RHD) obtained in step 4 and the Cumulative Summation (CUSUM) method, a Cumulative Summation Relative Hausdorff Distance (CUSUM-RHD) calculation model is constructed. For the first firing of the artillery, the Cumulative Summation Relative Hausdorff Distance is defined as 0. The calculation model for the Cumulative Summation Relative Hausdorff Distance after more than one firing is as follows: In equation (5), j represents the j-th shot, and j>1.

6. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 5, characterized in that, Step 6 includes the following steps: The cumulative sum and relative Hausdorff distance obtained in step 5 are defined as the Health Index (HI) for evaluating the state of the artillery recoil mechanism, and are expressed as: HI n =CUSUM-RHD n (6) The lower the health factor, the healthier the recoil mechanism. Monitoring the health factor allows for the assessment of the recoil mechanism's condition.

7. The method for assessing the condition of a gun recoil device based on CUSUM-RHD according to claim 6, characterized in that, Given HI∈(0,0.1] as the initial failure period of the recoil device, HI∈(0.1,0.3] as the general failure period of the recoil device, and HI>0.3 as the severe failure period of the recoil device, corresponding failure handling measures are taken according to the corresponding failure type to improve the efficiency of failure handling and the safety of the artillery.

Images