Reuse-based maintainability modeling, allocation and prediction method

Abstract

The invention discloses a maintainability modeling, distribution and prediction method based on reuse. The method comprises the following steps: establishing a hierarchical structure of a reuse carrier rocket; constructing a judgment matrix of the scheme layer; weight vectors of the criterion layer and the scheme layer are calculated, and consistency check of the judgment matrix is carried out; calculating a combined weight vector of each scheme in the scheme layer according to the weight vectors of the criterion layer and the scheme layer; obtaining the maintenance degree of each scheme in the scheme layer according to the combined weight vector; determining maintainability prediction parameters and a reference unit; calculating the average repair time of the subsystems; calculating the average restoration time interval of the whole rocket and the uncertainty of the average restoration time of the whole rocket; and predicting a maintenance degree index of the target layer. The method solves the problems of lack of distribution data, difficult prediction process, difficult prediction result quantification and uncertainty and the like of the reusing carrier rocket maintainability index under the background of minimum subsamples.

Description

Technical Field

[0001] The invention relates to a maintainability modeling, allocation and prediction method based on reuse, and belongs to the technical field of maintainability of reusable launch vehicles. Background Art

[0002] 1. Multi-level comprehensive weighted allocation method for the maintainability of reusable launch vehicles

[0003] Maintainability is the ability of a product to maintain or restore its specified state when it is maintained according to the specified procedures and methods under specified conditions and within a specified time. Maintainability engineering runs through all stages of the product's life cycle, including demonstration, verification, production, use, and retirement. Maintainability allocation, prediction, and evaluation are important work items specified in GJB368B, among which maintainability allocation is the source of equipment maintainability design analysis, prediction, and evaluation. Whether the maintainability allocation work can be effectively implemented is related to whether the equipment maintainability design analysis meets the maintainability requirements of development and whether the maintainability requirements can be converted into design measures. Therefore, it is very important to do a good job of maintainability allocation in the early stages of equipment development.

[0004] When designing maintainability, designers allocate maintainability according to the maintainability indicators of the product and the requirements of each level and function. The purpose of maintainability allocation is: ① The developers carry out repairs according to the repair performance indicators to meet the requirements of each component of the equipment; ② Through the issuance of maintenance at various levels, maintenance indicators are provided to the insurer to facilitate the management of the insurer. Maintainability allocation begins in the product demonstration stage and gradually deepens with the development of the product.

[0005] Maintainability allocation is an important task in the maintenance design stage and is directly related to the maintainability of the product. Maintainability allocation started earlier abroad. The United States conducted research on maintainability allocation in the 1960s and issued MIL-STD-470 "Maintainability Outline", which standardized maintainability allocation to a certain extent. China started a little later in comparison, and gave corresponding military standards with reference to American standards. The National Military Standard "Maintainability Allocation and Prediction Manual" has listed five commonly used methods: equal value allocation method, similar product allocation method, allocation method based on failure rate, comprehensive weighted allocation method based on failure rate and design characteristics, and weighted allocation method that ensures availability and considers the complexity differences of each unit. These methods are used in different stages of equipment design, and these commonly used maintainability allocation methods only consider the allocation of maintenance time.

[0006] At present, the distribution of equipment maintainability index mainly adopts the comprehensive weighted distribution method of failure rate and design characteristics in GJB / Z57 "Maintainability Distribution and Prediction Manual". The comprehensive weighted distribution method based on failure rate and design characteristics is applicable to the scheme stage and engineering development stage. It converts the factors considered in the distribution (complexity, testability, accessibility, replaceability, adjustability, maintenance environment, etc.) into weighted factors, and distributes maintainability indicators (such as mean repair time) according to the comprehensive weighting of failure rate and design characteristics. Because this method considers both failure rate and design characteristics, the distribution result is more accurate than the method based only on failure rate. When the complexity, testability, accessibility, replaceability, adjustability, maintenance environment, etc. of the equipment have been determined and the reliability data has been obtained, this method can improve the distribution accuracy. Therefore, this method is widely used in the maintainability distribution of complex aerospace equipment.

[0007] However, the comprehensive weighted allocation method based on failure rate and design characteristics also has some problems in its application: due to the different product information obtained during the system-level and equipment-level design phases, it is impossible to accurately control the time required for each maintenance activity, and the allocation value of MTTR (Mean Time to Repair) is not coordinated; in addition, when the failure rates of various systems vary greatly, there will be a problem of serious deviation between the allocation results and the actual situation. In addition, due to the lack of similar product data for reusable launch vehicles and incomplete failure rate information, if traditional methods are used to achieve the allocation of maintainability indicators, a large amount of failure rate data for the entire system and each subsystem needs to be collected, and the test cost is high, which makes it difficult to apply to very expensive, high-reliability, small-sample complex systems such as reusable launch vehicles.

[0008] 2. Grey prediction method for maintainability of reusable launch vehicle under small sample background

[0009] A reusable launch vehicle is a rocket that can return in whole or in part and land safely after completing the scheduled launch mission, and can perform the launch mission again after maintenance and refueling. This design aims to reduce the waste of expensive rocket bodies, engines and electrical equipment caused by one-time use, and reduce the production and launch costs of launch vehicles by sharing the costs through multiple uses. Reusable launch vehicles are an important means to reduce space transportation costs and an effective way to improve space rapid response capabilities. Maintainability is the ability of a product to maintain or restore its specified state when it is repaired under specified conditions and within a specified time according to specified procedures and methods. Maintainability, as an important parameter of general quality characteristics, directly affects the integrity of use and mission success. Maintainability engineering runs through all stages of the product's entire life cycle, including demonstration, verification, production, use, and retirement. Maintainability allocation, prediction and evaluation are important work items specified in GJB368B. Whether the maintainability prediction work can be effectively implemented is related to whether the equipment maintainability design analysis meets the maintainability requirements of development and whether the maintainability requirements can be converted into design measures. Therefore, it is very important to do the best possible maintenance prediction work in the early stages of reusable launch vehicle development.

[0010] Maintainability prediction is an important task in the maintenance design stage and is directly related to the maintainability of the product. At present, there are many research results on maintainability prediction at home and abroad. GJB / Z 57 provides about 6 maintainability prediction methods, and the scope of application and maintainability parameters used for estimation of each maintainability prediction method are different. First, the probability simulation prediction method, which is characterized by using probability simulation to obtain the distribution of system maintenance time based on the known time distribution of each maintenance operation of each component unit of the system, so as to make an estimate of the system maintenance time. It is suitable for products with complex systems and high test costs, but this method requires a lot of complex calculations and is not suitable for early design. Second, the functional level prediction method provides operation time in hours for each maintenance operation at different repair levels of the equipment, which may not be consistent with the current maintenance methods and characteristics, and is not sensitive to the failure mode of the grassroots replaceable unit. Third, the operation function prediction method conducts a unified analysis and study on the preventive maintenance and corrective maintenance time, and divides the task process into several operation functions. However, this method has more subjective components and is based on environmental judgment, not design. Fourth, the sampling scoring prediction method uses the principle of random sampling and a special checklist based on empirical data to score the maintenance work, and then uses an empirical formula to estimate the maintenance time. However, this method requires that the equipment has a sufficient sample size, and the checklist is mainly determined based on experience. Fifth, the time accumulation prediction method decomposes the equipment into several replaceable units, first estimates the time required for maintenance after each unit fails, and then calculates the average maintenance time of the equipment. However, the maintenance operation time standards given are mainly for electronic equipment. Sixth, the unit comparison prediction method requires finding a unit with a known maintenance time from the equipment, using it as a benchmark, and estimating the maintenance time of each unit by comparing it with the benchmark unit, and then determining the maintenance time of the equipment. However, at present, the maintenance time estimates of each unit in this method are all accurate real numbers, and the predicted results may not be consistent with the actual results, and the uncertainty of the predicted results cannot be quantified.

[0011] At present, the maintenance prediction mostly adopts the "functional level prediction method" and "unit comparison prediction method" to predict the repair time of the whole rocket of a disposable launch vehicle. However, the research on reusable launch vehicle technology started late, and the historical data, test information, sample information, etc. of the relevant maintainability indicators are extremely scarce. In engineering practice, the specific data of maintainability indicators is difficult to be accurate. If the traditional method is directly used to predict the maintainability indicators, it is difficult to obtain accurate maintainability results through the existing information and data calculation. Summary of the invention

[0012] The technical problem solved by the present invention is: to overcome the deficiencies of the prior art, to provide a maintainability modeling, allocation and prediction method based on reuse, and to solve the problem of lack of data on maintainability index allocation for reusable launch vehicles under extremely small sample background.

[0013] The technical solution of the present invention is:

[0014] The present invention discloses a maintainability modeling, allocation and prediction method based on reuse, comprising the following steps:

[0015] S1. Establish a hierarchical structure for reusable launch vehicles, including the target layer, the criterion layer, and the program layer;

[0016] S2, construct the judgment matrix of the scheme layer;

[0017] S3, according to the judgment matrix, calculate the weight vectors of the criterion layer and the solution layer, and perform a consistency check on the judgment matrix; if the consistency check passes, proceed to step S4, otherwise, return to step S2;

[0018] S4, calculating the combined weight vector of each scheme in the scheme layer according to the weight vectors of the criterion layer and the scheme layer;

[0019] S5. Obtain the maintainability of each solution in the solution layer according to the combined weight vector;

[0020] S6. Clarify the maintainability prediction parameters and reference units;

[0021] S7. Based on the benchmark unit, calculate the average repair time of the subsystem according to the relative failure rate coefficient and the repair time;

[0022] S8. Calculate the average repair time interval of the entire arrow and the uncertainty of the average repair time of the entire arrow according to the average repair time of the subsystem;

[0023] S9. According to the maintainability of each solution in the solution layer, the average repair time of the whole arrow and the uncertainty of the average repair time of the whole arrow, the maintainability index of the target layer is estimated.

[0024] Furthermore, in the above method, the judgment matrix is ​​specifically:

[0025]

[0026] a ij >0;a ij =1 / a ji ; a ii =1; i,j=1,2,…,n

[0027] Among them, a ij Take any two solutions u in the solution layer i and u j The importance of the criterion layer, n is the maximum number in the solution layer.

[0028] Furthermore, in the above method, the weight vectors of the criterion layer and the solution layer are determined, and the consistency check of the judgment matrix is ​​performed. The specific method is:

[0029] The weight vector of the criterion layer: k is the criterion number;

[0030] The weight vector of the solution layer:

[0031] According to the weight vector of the solution layer, calculate the maximum eigenvalue of the judgment matrix: ω is the weight vector of the judgment matrix, a i is the element in the judgment matrix;

[0032] According to the maximum eigenvalue λ of the matrix max , calculate the consistency index CI of the judgment matrix, Among them, n is the maximum number in the solution layer;

[0033] Look up the table to determine the random consistency index RI;

[0034] Calculate the judgment matrix consistency ratio CR,

[0035] If the consistency ratio of the judgment matrix is ​​greater than or equal to 1, the consistency test fails;

[0036] If the consistency ratio of the judgment matrix is ​​less than 1, the consistency test passes.

[0037] Furthermore, in the above method, the combined weight vector of each scheme in the calculation scheme layer is specifically:

[0038]

[0039]

[0040] Among them, ω i is the combined weight vector for each solution; is the relative weight of the target to the k criteria; is the relative weight of k criteria to n solutions.

[0041] Furthermore, in the above method, the maintenance degree is Among them, M 总 The design target of the maintainability of reusable launch vehicles is ω i is the combined weight vector of each solution in the solution layer.

[0042] Furthermore, in the above method, the average repair time of the subsystem is specifically:

[0043] Tcti =M Tcti +(2α-1)×D Tcti , 0≤α≤1, 1≤i≤n

[0044]

[0045] in, is the relative repair time coefficient of the jth unit, T ct0 is the average repair time of the benchmark unit, m is the number of components of the i-th subsystem; k j is the relative failure rate coefficient of the jth unit; is the repair time required to repair the i-th subsystem.

[0046] Furthermore, in the above method, the relative failure rate coefficient is specifically:

[0047] k j =λ ij / λ i0

[0048] Among them, λ i0 is the relative failure rate of the i-th system reference unit; ij is the relative failure rate of the jth unit in the ith system.

[0049] Furthermore, in the above method, the repair time required for the subsystem maintenance is specifically:

[0050]

[0051] in, is the kth maintenance operation time of the jth unit, T 0k is the corresponding maintenance operation time of the benchmark unit, h 0k It is the empirical value of the relative repair time of the benchmark unit.

[0052] Furthermore, in the above method, the average repair time interval of the whole arrow is specifically:

[0053]

[0054] in, is the average repair time interval of the subsystem, λ i is the failure rate of the ith subsystem, and n is the number of units that make up the subsystem.

[0055] Furthermore, in the above method, the uncertainty of the average repair time of the whole arrow is specifically:

[0056]

[0057] in, The average repair time interval for the entire arrow; They are the minimum and maximum average repair time of the whole arrow respectively.

[0058] The beneficial effects of the present invention and the prior art are:

[0059] (1) The multi-level comprehensive weighted distribution method of the maintainability of a reusable launch vehicle of the present invention solves the problems that when the failure rates of various systems of the maintainability distribution method of a reusable launch vehicle differ greatly, the distribution result deviates seriously from the actual situation, the maintenance rate is difficult to be accurate in engineering practice, and it is difficult to obtain an accurate maintainability by calculating the maintenance rate and maintenance time.

[0060] (2) The grey prediction method for maintainability of reusable launch vehicles under extremely small sample background of the present invention solves the problems in the prior art of lack of maintainability index data of reusable launch vehicles under extremely small sample background, difficulty in prediction process, and difficulty in quantifying the uncertainty of prediction results.

[0061] (3) The multi-level comprehensive weighted allocation method for the maintainability of a reusable launch vehicle of the present invention innovatively proposes to use maintainability as an indicator for maintainability allocation, and uses the analytic hierarchy process to calculate the weights of various factors affecting maintainability and to allocate maintainability, thereby solving the problems of the currently widely used "comprehensive weighted allocation method based on failure rate and design characteristics", such as the problem that the allocation result deviates seriously from the actual situation when the failure rates of various systems vary greatly.

[0062] (4) Due to the lack of similar product data for repeatable launch vehicles and incomplete failure rate information, it is impossible to use traditional methods to achieve the allocation of maintainability indicators. Compared with the existing traditional technology, the present invention does not require a large amount of failure rate data, reduces the number of related tests for repeatable launch vehicles, and reduces the launch cost while improving the reliability and maintainability of my country's launch vehicles.

[0063] (5) In engineering practice, compared with maintenance time, we are more concerned about whether the maintenance degree within a certain period of time can meet the target requirements. The traditional maintainability allocation method is to allocate maintenance time. Since the actual maintenance rate is difficult to accurately give, it is difficult to obtain the more concerned accurate maintainability through maintenance rate and maintenance time calculation. The multi-level comprehensive weighted allocation method for the maintainability of reusable launch vehicles proposed in the present invention directly uses the maintainability as an indicator for allocation. The allocation result is the maintainability indicator that the engineering is more concerned about, which better meets the needs of engineering practice.

[0064] (6) The present invention adopts a grey estimation method for the maintainability of reusable launch vehicles under the background of extremely small samples, integrates the grey system theory with the classical estimation method for maintainability indicators, uses interval grey numbers to calculate the maintainability indicator values ​​of each subsystem, and uses the grey unit estimation method to estimate the maintainability indicator of the entire rocket. This effectively solves the current problem of lack of maintainability indicator data and difficulty in the estimation process for reusable launch vehicles with extremely small sample characteristics.

[0065] (7) According to the analysis results of the example, the present invention increases the characterization and quantification of the uncertainty of the maintainability prediction results, and further improves the maintainability prediction technology of reusable launch vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a simplified diagram of the maintainability allocation hierarchy model of the present invention;

[0067] Figure 2 It is a hierarchical structure diagram of the reusable launch vehicle of the present invention. DETAILED DESCRIPTION

[0068] The present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0069] In view of the above problems, the present invention proposes a grey maintainability prediction method for reusable launch vehicles with extremely small sample background. This method innovatively integrates the grey system theory with the classical maintainability index prediction method, uses interval grey numbers to calculate the maintainability index values ​​of each subsystem, and uses the grey unit prediction method to predict the maintainability index. It solves the current problems of lack of data on maintainability indexes of reusable launch vehicles with extremely small sample background, difficulty in prediction process, and difficulty in quantifying uncertainty in prediction results.

[0070] The present invention discloses a maintainability modeling, allocation and prediction method based on reuse, comprising the following steps:

[0071] S1. Establish a hierarchical structure for reusable launch vehicles, including the target layer, the criterion layer, and the program layer;

[0072] S2, construct the judgment matrix of the scheme layer;

[0073] S3, according to the judgment matrix, calculate the weight vectors of the criterion layer and the solution layer, and perform a consistency check on the judgment matrix; if the consistency check passes, proceed to step S4, otherwise, return to step S2;

[0074] S4, calculating the combined weight vector of each scheme in the scheme layer according to the weight vectors of the criterion layer and the scheme layer;

[0075] S5. Obtain the maintainability of each solution in the solution layer according to the combined weight vector;

[0076] S6. Clarify the maintainability prediction parameters and reference units;

[0077] S7. Based on the benchmark unit, calculate the average repair time of the subsystem according to the relative failure rate coefficient and the repair time;

[0078] S8. Calculate the average repair time interval of the entire arrow and the uncertainty of the average repair time of the entire arrow according to the average repair time of the subsystem;

[0079] S9. According to the maintainability of each solution in the solution layer, the average repair time of the whole arrow and the uncertainty of the average repair time of the whole arrow, the maintainability index of the target layer is estimated.

[0080] Preferably, the judgment matrix is ​​specifically:

[0081]

[0082] a ij >0;a ij =1 / a ji ; a ii =1; i,j=1,2,…,n

[0083] Among them, a ij Take any two solutions u in the solution layer i and u j The importance of the criterion layer, n is the maximum number in the solution layer.

[0084] Preferably, the weight vectors of the criterion layer and the solution layer are determined, and the consistency check of the judgment matrix is ​​performed. The specific method is:

[0085] The weight vector of the criterion layer: k is the criterion number;

[0086] The weight vector of the solution layer:

[0087] According to the weight vector of the solution layer, calculate the maximum eigenvalue of the judgment matrix: ω is the weight vector of the judgment matrix, a i is the element in the judgment matrix;

[0088] According to the maximum eigenvalue λ of the matrix max , calculate the consistency index CI of the judgment matrix, Among them, n is the maximum number in the solution layer;

[0089] Look up the table to determine the random consistency index RI;

[0090] Calculate the judgment matrix consistency ratio CR,

[0091] If the consistency ratio of the judgment matrix is ​​greater than or equal to 1, the consistency test fails;

[0092] If the consistency ratio of the judgment matrix is ​​less than 1, the consistency test passes.

[0093] Preferably, the combined weight vector of each scheme in the calculation scheme layer is specifically:

[0094]

[0095] Among them, ω i is the combined weight vector for each solution; is the relative weight of the target to the k criteria; is the relative weight of k criteria to n solutions.

[0096] Preferably, the maintenance degree is Among them, M 总 The design target of the maintainability of reusable launch vehicles is ω i is the combined weight vector of each solution in the solution layer.

[0097] Preferably, the mean repair time of the subsystem is specifically:

[0098] T cti =M Tcti +(2α-1)×D Tcti , 0≤α≤1, 1≤i≤n

[0099]

[0100] in, is the relative repair time coefficient of the jth unit, T ct0 is the average repair time of the benchmark unit, m is the number of components of the i-th subsystem; k j is the relative failure rate coefficient of the jth unit; is the repair time required to repair the i-th subsystem.

[0101] Preferably, the relative failure rate coefficient is specifically:

[0102] k j =λ ij / λ i0

[0103] Among them, λ i0 is the relative failure rate of the i-th system reference unit; ij is the relative failure rate of the jth unit in the ith system.

[0104] Preferably, the repair time required for the subsystem maintenance is specifically:

[0105]

[0106] in, is the kth maintenance operation time of the jth unit, T 0k is the corresponding maintenance operation time of the benchmark unit, h 0k It is the empirical value of the relative repair time of the benchmark unit.

[0107] Preferably, the average repair time interval of the whole arrow is specifically:

[0108]

[0109] in, is the average repair time interval of the subsystem, λ i is the failure rate of the ith subsystem, and n is the number of units that make up the subsystem.

[0110] Preferably, the uncertainty of the average repair time of the whole arrow is specifically:

[0111]

[0112] in, The average repair time interval for the entire arrow; They are the minimum and maximum average repair time of the whole arrow respectively.

[0113] Example

[0114] like Figure 1 As shown, this embodiment provides a maintainability modeling, allocation and prediction method based on a reusable launch vehicle, comprising the following steps:

[0115] (1) Multi-level comprehensive weighted allocation method for the maintainability of reusable launch vehicles

[0116] At present, the allocation method using mean repair time as the maintainability index has the problem that the allocation result deviates seriously from the actual situation when the failure rates of various systems are quite different or the failure rate information is incomplete. Moreover, compared with the maintenance time, we are more concerned about whether the maintenance degree within a certain period of time can meet the target requirements. In engineering practice, the maintenance rate is difficult to be accurate and can be regarded as a gray number. If the maintenance time is allocated according to the traditional method, it is difficult to obtain the required accurate maintainability through the maintenance rate and maintenance time calculation. In view of the above problems, the present invention proposes a multi-level comprehensive weighted allocation method for the maintainability of reusable launch vehicles based on the hierarchical analysis method. The method uses the maintainability as the maintainability allocation index to allocate the maintainability, solves the problems existing in the current maintainability allocation method, improves the reliability and maintainability of reusable launch vehicles, and reduces the launch cost.

[0117] Specific method description:

[0118] Step 1. Establish a hierarchical model

[0119] According to the complexity of the problem of reusable launch vehicles, the hierarchical structure of reusable launch vehicles is divided into target layer, criterion layer and solution layer. The target layer A is the maintainability index; the criterion layer C is the various factors affecting maintainability, which can be further divided into levels according to the complexity of the problem, and the more indicators in the criterion layer, the more accurate the maintainability of the target layer; the solution layer P is the objects and indicators to be allocated, such as Figure 1 shown.

[0120] Step 2. Select the scale and construct the judgment matrix

[0121] After establishing the hierarchical model, it is necessary to construct a pairwise judgment matrix. In order to quantify the relative importance of each factor, it is necessary to select a quantitative judgment scale. The 1 to 9 scale is the most commonly used representation method, and the specific meaning is shown in Table 1.

[0122] Table 1 Judgment scale table

[0123]

[0124]

[0125] Construct the judgment matrix. The specific method of constructing the judgment matrix of the solution layer relative to the criterion layer is: first, take any two elements u from the solution layer i and u j , compare the importance of the two to the criterion layer, denoted as a ij , and so on, until all elements of the solution layer are compared pairwise and assigned corresponding scale values. Thus, the judgment matrix A is obtained:

[0126]

[0127] According to the construction principle of the judgment matrix, A has the following properties:

[0128] a ij >0;a ij =1 / a ji ; a ii =1; i,j=1,2,…,n

[0129] Step 3. Calculate the weight vectors of the criterion layer and the scheme layer (see Step 4) and perform consistency check

[0130] ① Calculate the maximum eigenvalue λ of the judgment matrix max ,Right now ω is the weight vector of the judgment matrix, that is, the weight vector of the criterion layer and the solution layer, a iis the element in the judgment matrix.

[0131] ② Calculate the consistency index CI of the judgment matrix

[0132]

[0133] ③ Look up the table to determine the random consistency index RI

[0134] ④ Calculate the consistency ratio CR of the judgment matrix

[0135]

[0136] Step 4. Calculate the combined weights of each unit to obtain the combined weight vector

[0137] The relative weight of the target layer to the criterion layer is recorded as The relative weight of a criterion to n objects in the solution layer is recorded as At this time, the relative weight ω of object i in the solution layer i It can be written as:

[0138]

[0139] Step 5. Assign maintainability indicators

[0140] The maintenance degree M of the i-th object in the component unit composed of k objects in the solution layer is i for:

[0141]

[0142] Among them, M 总 Overall maintainability design objectives for reusable launch vehicles.

[0143] Case analysis:

[0144] Step 1. Establish a hierarchical model

[0145] The maintenance degree of reusable launch vehicles is allocated based on the hierarchical analysis method. The reusable launch vehicle consists of 7 systems, and its connection mode is series connection. The present invention selects 5 main systems for analysis, namely: rocket body structure system, booster delivery system, engine, integrated electrical system, and ground test and launch control system. The minor maintenance degree is M 小修 =0.96, and now we use the analytic hierarchy process to distribute its maintainability. The hierarchical model selects testability, accessibility, replaceability, and adjustability as indicators that affect maintainability, such as Figure 2 As shown, the criterion layer is still composed of the 6 criteria shown in the figure.

[0146] Step 2. Construct a judgment matrix

[0147] The judgment matrix is ​​constructed based on expert experience, and the judgment matrix of the criterion layer (relative to the target layer) is:

[0148]

[0149] The judgment matrix is ​​constructed based on expert experience, and the judgment matrix of the solution layer (relative to the target layer) is:

[0150]

[0151]

[0152] Among them, A1 is the testability index judgment matrix; A2 is the accessibility index judgment matrix; A3 is the replaceability index judgment matrix; A4 is the adjustability index judgment matrix.

[0153] Step 3. Calculate the weight vector of each matrix and perform consistency check

[0154] The weight vector is calculated using the more general sum method.

[0155] ① Normalize the column vectors of each matrix, then sum them by row, and finally normalize the row vectors to obtain the weight vector:

[0156] ω=[0.25160.55490.09670.0967] T

[0157] ω1=[0.34240.12980.34240.12980.0557] T

[0158] ω2=[0.46240.19520.07370.19520.0737] T

[0159] ω3=[0.14290.14290.14290.14290.4286] T

[0160] ω4=[0.05570.12980.34240.12980.3424] T

[0161] because

[0162]

[0163]

[0164] From this we can determine that the maximum characteristic root of the matrix is:

[0165]

[0166]

[0167] ②Consistency index:

[0168]

[0169] ③ Check the random consistency index RI, as shown in Table 2:

[0170] Table 1 Random consistency index RI

[0171] m 1 2 3 4 5 6 7 8 9 10 RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49

[0172] When m = 4, the random consistency index RI = 0.9, so the consistency ratio is:

[0173] When m = 5, the random consistency index RI = 1.12, so the consistency ratio is:

[0174] All are less than 1, indicating that the inconsistency of matrices A, A1, A2, A3, and A4 is acceptable.

[0175] Step 4. Calculate the combined weight of each unit

[0176] According to the above weight vector, the combined weights of each scheme are calculated to obtain the combined weight vector:

[0177] V=[ω1,ω2,ω3,ω4,ω5][ω]

[0178] =[0.36200.16730.17390.16730.1294] T

[0179] Step 5. Allocate maintenance degree

[0180] Therefore, the maintenance degree allocated to each scheme is calculated as:

[0181]

[0182] Obviously, the maintenance degree allocation index of each component unit meets the allocation requirements.

[0183] (2) Grey prediction method for maintainability of reusable launch vehicle under extremely small sample background

[0184] In the maintainability design of reusable launch vehicles, maintainability prediction is one of the basic tasks to improve maintainability and is an important link in the maintainability design process. At present, the research on the reuse technology of launch vehicles in my country started late, and the historical data, test information, sample information, etc. of relevant maintainability indicators are extremely scarce. In engineering practice, the specific data of maintainability indicators are difficult to be accurate. If the maintainability indicators are directly predicted by traditional methods, it is difficult to obtain accurate maintainability results through existing information and data calculations. In view of this, the present invention proposes a gray prediction method for the maintainability of reusable launch vehicles with a small sample background. This method innovatively combines the gray system theory with the prediction method of the classical maintainability index, uses the interval gray number to calculate the maintainability index value of each subsystem, and uses the gray unit prediction method to predict the maintainability index of the whole rocket, which provides a reference for further improving the reliability of reusable launch vehicles and the research level of maintainability technology.

[0185] The specific steps of the method are as follows, and the flowchart of the expected method is shown in Figure 1 .

[0186] Step 1. Clarify the maintainability prediction parameters

[0187] Maintainability prediction parameters usually include the estimated mean time to repair (T CT ), average preventive maintenance time (T PT ), mean maintenance time (T M ), relative maintenance time coefficient (h i )wait.

[0188] Step 2. Determine the replaceable units of the product

[0189] Based on the specified maintenance level (for example, on-site maintenance), the various replaceable units of the product are divided and determined according to the product design plan and implementation possibilities.

[0190] Step 3. Select the base unit

[0191] The selection principle of the benchmark unit is, first, to be able to estimate its average repair time; second, to make it clearly comparable with other units in terms of complexity, maintainability, etc., so as to facilitate the determination of various coefficients. The benchmark unit for corrective maintenance and preventive maintenance can be the same benchmark unit, or different benchmark units can be selected according to needs.

[0192] Step 4. Determine the coefficients

[0193] 1) Relative failure rate coefficient. The relative failure rate coefficient of the jth replaceable unit is

[0194] k j =λ ij / λ i0

[0195] Where λ i0 —Relative failure rate of the i-th system reference unit; λ ij —Relative failure rate of the jth replaceable unit in the ith system. In actual prediction, k j Not necessarily through λ ij and λ i0 The value calculation can be directly estimated based on the product design characteristics.

[0196] 2) Relative repair time coefficient. The maintenance activities are divided into four basic maintenance operations: positioning and isolation, disassembly and assembly, installation and replacement, and adjustment and inspection. It can also be adjusted accordingly according to the situation, such as dividing it into seven basic maintenance operations: positioning, isolation, disassembly, replacement, combination, adjustment, and inspection. The relative repair time coefficient calculation formula is as follows:

[0197]

[0198] or

[0199]

[0200] In the formula, The kth maintenance operation time of the jth replaceable unit The maintenance operation time T corresponding to the benchmark unit 0k The ratio is determined, that is

[0201]

[0202] Since the technician cannot give an accurate value for the kth maintenance operation time of the jth replaceable unit, h jk is the interval gray number. For the reference unit, ∑h 0k =1,h 0k The value of is determined according to the ratio of the time taken by each basic maintenance operation. The other replaceable units are determined according to the multiple ratio of the time required relative to the benchmark. This value is generally given by experts in related fields or maintenance personnel through experience in the form of Table 3.

[0203] Table 3 Relative coefficients of each replaceable unit

[0204]

[0205] Therefore, it can be concluded that the repair time required for the jth replaceable unit is

[0206]

[0207] Step 5. Calculate the average repair time of the subsystem

[0208]

[0209] Step 6. Defuzzification of the average repair time interval gray number

[0210] The mean repair time for the subsystems obtained here is is an interval number, so it is necessary to defuzzify the interval number to calculate the average repair time of the subsystem. The specific calculation process is as follows:

[0211] make but

[0212] T cti =M Tcti +(2α-1)×D Tcti

[0213] In the formula, 0≤α≤1, 1≤i≤n, and n is the total number of component subsystems.

[0214] When α=0, At this time, the decision maker has a radical mentality; when α = 1 / 2, At this time, the decision maker holds a moderate mentality; when α=1, Decision makers are cautious at this time.

[0215] Step 7. Calculate the average repair time of the whole arrow

[0216] Finally, substitute the average repair time of each subsystem into the following formula to calculate the average repair time of the entire arrow:

[0217] Step 8. Calculate the uncertainty of the estimated average repair time for the entire arrow

[0218]

[0219] Case analysis:

[0220] Step 1. Determine the replaceable units of the product

[0221] The gray number unit comparison prediction method is used to predict the maintainability of the launch vehicle engine system. The failure rate of the engine system is λ3=5.13×10 -4 . It is mainly divided into two replaceable units: propellant supply unit U 31 、Engine control unit U 32 According to the FMEA table, the main replaceable units and their failure rates of this type of engine system are shown in Table 2.

[0222] Step 2. Select the base unit

[0223] Because the "propellant supply device U 31"This replaceable unit has the lowest failure rate, and the time required for each maintenance operation of this unit is relatively easy to determine, so "Propellant Supply Unit U 31 " is the benchmark unit.

[0224] Step 3. Determine the coefficients

[0225] "Propellant supply device 31 " is the reference unit, and its failure rate coefficient k0 = 1. The average time for detection and isolation of this reference unit is 16.56h, disassembly and assembly is 11.96h, installation and replacement is 9.2h, and adjustment and detection is 8.28h. The cumulative average repair time is 46h. Accordingly, it can be obtained

[0226] "Engine Control Unit U 32 "This replaceable unit has the highest failure rate, and its relative failure rate coefficient k2 = λ 32 / λ 31 =3.27×10 -4 / 1.86×10 -4 ≈1.758. The fault phenomenon corresponding to this fault mode is generally more obvious, and the final manifestation is that the deviation between the reading and the actual measured value exceeds the allowable range. Therefore, the required detection and isolation time is relatively short. Through comprehensive evaluation by experts, Since the external shielding needs to be removed during repair, it takes more time than the reference mode. The corresponding installation and replacement is easier than the reference mode. After installation, it is necessary to check whether the newly installed components can work normally and whether there is any leakage. The corresponding adjustment and detection time is longer than the reference mode.

[0227] Table 4 Relative coefficient table of engine system repair time

[0228]

[0229] Step 4. Calculate the mean repair time of the engine system

[0230] Calculate the sum of the coefficients as listed in Table 4. Then calculate the estimated maintainability value of the engine system according to the formula. The repair time of the benchmark unit should be the repair time of the "propellant supply device", which is 46 hours.

[0231]

[0232] Step 5. Defuzzification of interval gray numbers

[0233] The mean repair time for the subsystems obtained here is is an interval number, so it is necessary to defuzzify the interval number to calculate the average repair time of the subsystem. The specific calculation process is as follows:

[0234] make but

[0235] T cti =M Tcti +(2α-1)×D Tcti

[0236] In the formula, 0≤α≤1, 1≤i≤n, and n is the total number of component subsystems.

[0237] When α=0, At this time, the decision maker has a radical mentality; when α = 1 / 2, At this time, the decision maker holds a moderate mentality; when α=1, Decision makers are cautious at this time.

[0238] Here we take α = 0.5 and substitute it into the above formula to get

[0239]

[0240] By comparing the results of this paper with actual engineering experience, it can be concluded that the mean repair time of faults in this subsystem is basically in line with the actual situation and meets the index requirements.

[0241] Step 6. Calculate the average repair time of the whole arrow

[0242] According to the above method, the maintainability levels of all subsystems are estimated in turn and shown in Table 5.

[0243] Table 5 Average repair time of each subsystem of reusable launch vehicle

[0244]

[0245]

[0246] Finally, the average repair time interval of the entire arrow is estimated by the formula as follows:

[0247]

[0248] The average estimated repair time for the entire arrow is calculated as follows:

[0249]

[0250] The uncertainty of the mean repair time estimate is calculated as follows:

[0251]

[0252] According to the calculation results, the fastest repair time for the entire arrow is expected to be about 86.89 hours, the slowest repair time is about 94.74 hours, the average repair time is about 90.82 hours, and the uncertainty of the estimated result is about 3.93 hours. Maintenance technicians can reasonably plan the maintenance time based on the estimated results and their uncertainties, providing an important reference for the formulation of the overall maintenance strategy.

[0253] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

[0254] The contents not described in detail in the specification of the present invention belong to the common knowledge of the professionals in this field.

Claims

1. A method for modeling, allocating and predicting maintainability based on reuse, characterized in that: The following steps are involved: S1. Establish a hierarchical structure for reusable launch vehicles, including the target layer, the criterion layer, and the program layer; S2, construct the judgment matrix of the scheme layer; S3. Calculate the weight vectors of the criterion layer and the solution layer according to the judgment matrix, and perform consistency check of the judgment matrix; If the consistency check passes, proceed to step S4, otherwise, return to step S2; S4, calculating the combined weight vector of each scheme in the scheme layer according to the weight vectors of the criterion layer and the scheme layer; S5. Obtain the maintainability of each solution in the solution layer according to the combined weight vector; S6. Clarify the maintainability prediction parameters and reference units; S7. Based on the benchmark unit, calculate the average repair time of the subsystem according to the relative failure rate coefficient and the repair time; S8. Calculate the average repair time interval of the entire arrow and the uncertainty of the average repair time of the entire arrow according to the average repair time of the subsystem; S9. According to the maintainability of each solution in the solution layer, the average repair time of the whole arrow and the uncertainty of the average repair time of the whole arrow, the maintainability index of the target layer is estimated.

2. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The judgment matrix is ​​specifically: a ij >0;a ij =1 / a ji ;a ii =1;i,j=1,2,…,n Among them, a ij Take any two solutions u in the solution layer i and u j The importance of the criterion layer, n is the maximum number in the solution layer.

3. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: Determine the weight vectors of the criterion layer and the solution layer, and perform consistency check of the judgment matrix. The specific method is as follows: The weight vector of the criterion layer: k is the criterion number; The weight vector of the solution layer: According to the weight vector of the solution layer, calculate the maximum eigenvalue of the judgment matrix: ω is the weight vector of the judgment matrix, a i is the element in the judgment matrix; According to the maximum eigenvalue λ of the matrix max , calculate the consistency index CI of the judgment matrix, Among them, n is the maximum number in the solution layer; Look up the table to determine the random consistency index RI; Calculate the judgment matrix consistency ratio CR, If the consistency ratio of the judgment matrix is ​​greater than or equal to 1, the consistency test fails; If the consistency ratio of the judgment matrix is ​​less than 1, the consistency test passes.

4. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The combined weight vector of each solution in the calculation solution layer is specifically: Among them, ω i is the combined weight vector for each solution; is the relative weight of the target to the k criteria; is the relative weight of k criteria to n solutions.

5. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The maintenance degree is Among them, M 总 The design target of the maintainability of reusable launch vehicles is ω i is the combined weight vector of each solution in the solution layer.

6. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The mean repair time of the subsystem is specifically: T cti =M Tcti +(2α-1)×D Tcti ,0≤α≤1,1≤i≤n in, is the relative repair time coefficient of the jth unit, T ct0 is the average repair time of the benchmark unit, m is the number of components of the i-th subsystem; k j is the relative failure rate coefficient of the jth unit; is the repair time required to repair the i-th subsystem.

7. The method for modeling, allocating and predicting maintainability based on reuse according to claim 6, characterized in that: The relative failure rate coefficient is specifically: k j =λ ij / l i0 Among them, λ i0 is the relative failure rate of the i-th system reference unit; ij is the relative failure rate of the jth unit in the ith system.

8. The method for modeling, allocating and predicting maintainability based on reuse according to claim 6, characterized in that: The repair time required for the subsystem maintenance is specifically: in, is the kth maintenance operation time of the jth unit, T 0k is the corresponding maintenance operation time of the benchmark unit, h 0k It is the empirical value of the relative repair time of the benchmark unit.

9. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The average repair time interval of the whole arrow is specifically: in, is the average repair time interval of the subsystem, λ i is the failure rate of the ith subsystem, and n is the number of units that make up the subsystem.

10. The method for modeling, allocating and predicting maintainability based on reuse according to claim 1, characterized in that: The uncertainty of the average repair time of the whole arrow is specifically: in, The average repair time interval for the entire arrow; They are the minimum and maximum average repair time of the whole arrow respectively.

Images