Electronic equipment health assessment device

Abstract

The present invention relates to the technical field of aircraft health management, and discloses an electronic equipment health assessment device, including a component-level health assessment module and a system-level health assessment module. The component-level health assessment module includes a fuzzy severity level calculation module and a component health calculation module, and the system-level health assessment module includes a component weight adaptive construction module and a system health calculation module. The present invention realizes the health assessment of electronic equipment in the absence of status monitoring parameters, dynamically assesses the health status of components based on component failure risks and in-machine self-test information, and dynamically assesses the health status of the system based on the system reliability block diagram and the adaptive weights of the components, thereby improving the health management capability of the aircraft and contributing to the realization of autonomous health management of the aircraft.

Description

Technical Field

[0001] The present invention relates to the technical field of aircraft health management, and in particular to an electronic equipment health assessment device. Background Art

[0002] In the field of aircraft health assessment, the health status of an object is generally evaluated based on real-time performance parameters. With the digital development of aircraft, the application of electronic equipment in aircraft is increasing, and the importance of electronic equipment health assessment is becoming increasingly prominent. Currently, aircraft onboard electronic equipment relies on built-in self-test diagnostics and lacks real-time performance parameters for health assessment, making health assessment of aircraft electronic equipment difficult to implement. Furthermore, most aircraft health assessment methods do not distinguish between the logical structure of the system, and there is little difference between system-level and component-level health assessment methods. In short, system-level health assessment does not reflect the series and parallel logical structure of the system. To achieve the health assessment of electronic equipment, it is urgent to evaluate the health status of components in the absence of performance parameters. On this basis, considering the series and parallel logical structure of the system, the health assessment of electronic equipment at the system level can be performed, thereby improving the health assessment capabilities of electronic equipment. Summary of the Invention

[0003] The purpose of the present invention is to provide an electronic equipment health assessment device to solve the problem of electronic system health assessment based on failure risk, thereby supporting the health management of aircraft.

[0004] The object of the present invention is achieved through the following technical solutions:

[0005] An electronic equipment health assessment device includes a component-level health assessment module and a system-level health assessment module;

[0006] The component-level health assessment module includes a fuzzy severity level calculation module and a component health calculation module. The fuzzy severity level calculation module is used to defuzzify the fuzzy membership of the component severity level to obtain the fuzzy severity level of the component. The component health calculation module is used to analyze the reliability, fuzzy severity level and built-in self-test of the component, and evaluate the health of the component based on the failure risk of the component.

[0007] The system-level health assessment module includes a component weight adaptive construction module and a system health calculation module. The component weight adaptive construction module constructs a relative importance judgment matrix based on the fuzzy severity level of the component, and uses an adaptive adjustment algorithm to adjust the relative importance judgment matrix so that the relative importance judgment matrix and the component weight vector pass the consistency test; the system health calculation module first analyzes the reliability block diagram of the electronic equipment based on the component health and the weight vector of the components in the system that pass the consistency test, and obtains the logical results of each system based on this, and then selects the system health calculation method according to the series-parallel relationship to obtain the system health assessment.

[0008] Preferably, the fuzzy severity level calculation module uses the fault type impact and fatality analysis method to analyze the components and construct the fuzzy membership of the severity level of the components: after determining the severity level set V, an evaluation team is established to evaluate the severity levels of each component, obtain the evaluation of each severity level of each component, and count the number of supporters x for each severity level of each component. ij , based on which we get the membership degree b ij =x ij / x i , where x i represents the number of people who evaluated the i-th component; finally, the fuzzy membership vector B of the severity level of the i-th component is calculated;

[0009] After constructing the fuzzy membership vector B of the component severity level, the fuzzy severity level S of the component is calculated using the centroid method according to the severity level set V, and normalized, as shown in formula (1):

[0010] S=VB T / maxV (1).

[0011] Preferably, the component health calculation module analyzes and processes the failure rate h and initial reliability R0 of the component, calculates the failure probability OPR of the component based on the reliability analysis, and calculates the failure risk RNK of the component by combining the fuzzy severity level S of the component calculated by the fuzzy severity level calculation module and the adjustment coefficient K of S, as shown in formula (2):

[0012]

[0013] The health of the component is calculated based on the failure risk of the component, and the calculation method of the component health is selected according to the component's built-in self-test diagnostic report. Finally, the health of the component is calculated as shown in formula (3):

[0014]

[0015] Wherein, BIT 0 indicates that the component is not faulty, and BIT 1 indicates that the component is faulty.

[0016] Preferably, the component weight adaptive construction module includes a relative importance judgment matrix module, a component weight calculation module, a consistency index inspection module and a judgment matrix adaptive adjustment module;

[0017] The relative importance judgment matrix module calculates the relative importance judgment matrix R of each component according to the fuzzy severity level S of each component in the system, and adjusts the relative importance matrix according to the range of the nine-scale judgment scale. The adjustment coefficient is L. The relative importance judgment matrix of the system is shown in formula (4):

[0018] R←[S T S] L (4)

[0019] The component weight calculation module calculates the component weight vector based on the relative importance judgment matrix R: First, sum the relative importance judgment matrix R in the row and column directions to obtain the column sum vector CS R , and then normalize the relative importance judgment matrix in the column direction, the relative importance judgment matrix R and the column sum vector CS R Divide the number to get the normalized judgment matrix R C =R / CS R , and then normalize the judgment matrix R C Take the row summation process to get the row sum vector RS R , for the normalized judgment matrix R C Sum all elements and get the element and TS R , and then calculate the weight vector W of each component in the system, as shown in formula (5):

[0020] W=RS R / TS R (5)

[0021] The consistency index test module performs consistency test on the relative importance judgment matrix and component weight vector calculated by the relative importance judgment matrix module and the component weight calculation module: first, the dot product of the relative importance judgment matrix R and the weight vector W is calculated, and then the vector λ is divided by the weight vector W to obtain the vector λ. The average value of the vector λ is obtained to obtain the maximum characteristic root λ max Then, according to the number n of components in the system, the consistency index CI is calculated, and the consistency ratio CR is calculated according to the random consistency index RI corresponding to the number n of components. Finally, if CR < 0.1, the consistency test passes, and the weight vector of the component is output to the system health calculation module. Otherwise, the consistency test fails, and the judgment matrix adaptive adjustment module is enabled. The relevant calculation is shown in formula (6):

[0022]

[0023] The judgment matrix adaptive adjustment module performs adaptive adjustment on the relative importance judgment matrix. First, the adjustment coefficient d is initialized, and the elements in the upper right corner of the relative importance judgment matrix are adjusted one by one. The adjustment element r=r d , assign r = 1 / r to the elements symmetrical along the main diagonal of element r, and obtain the adjusted relative importance judgment matrix R←R based on the adjusted elements rThe relative importance judgment matrix module and the component weight calculation module are recalculated to obtain the adjusted weight vector W. If the consistency check passes, the component weight is output to the system health calculation module. Otherwise, if the consistency check fails, the judgment matrix adaptive adjustment module continues to adjust.

[0024] Preferably, the system health calculation module inputs the component health HI completed by the component health calculation module and the weight vector W of the system components completed by the component weight adaptive construction module, and calculates the health of the system hierarchically according to the reliability block diagram of the system and the logical structure of the reliability block diagram, as shown in formula (7):

[0025]

[0026] The beneficial effects of the present invention are:

[0027] In the absence of performance parameters, the health assessment of electronic equipment is realized, and a new component health assessment method is provided, that is, the component health assessment is realized based on the failure risk. At the same time, different from other system health assessment methods, the system health calculation method is selected based on the system reliability block diagram or logical structure, which avoids the weight dilution problem and realizes system-level health assessment. The health assessment based on failure risk adopts an analytical calculation method with fast calculation speed. This method can be applied not only to the health assessment of ground systems, but also to the health assessment of airborne systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A diagram of the device structure for system health assessment based on failure risk;

[0029] Figure 2 This is a reliability block diagram for an avionics system;

[0030] Figure 3 This is a framework diagram of the system health assessment algorithm based on failure risk;

[0031] Figure 4 It is a relationship diagram between component health, failure probability and fuzzy severity level;

[0032] Figure 5 This is a statistical chart of the health classification of an avionics system;

[0033] Figure 6 This is the health distribution diagram of an avionics equipment system. DETAILED DESCRIPTION

[0034] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0035] See also Figure 1As shown, an electronic device health assessment device shown in this embodiment includes a component-level health assessment module and a system-level health assessment module;

[0036] The component-level health assessment module includes a fuzzy severity level calculation module and a component health calculation module. The fuzzy severity level calculation module is used to defuzzify the fuzzy membership of the component severity level to obtain the fuzzy severity level of the component. The component health calculation module is used to analyze the component's reliability, fuzzy severity level, and built-in self-test information, and evaluate the component's health based on the component's failure risk.

[0037] The system-level health assessment module includes a component weight adaptive construction module and a system health calculation module. The component weight adaptive construction module constructs a relative importance judgment matrix based on the fuzzy severity level of the component, and uses an adaptive adjustment algorithm to adjust the relative importance judgment matrix so that the relative importance judgment matrix and the component weight vector pass the consistency test; the system health calculation module first analyzes the reliability block diagram of the electronic equipment according to the component health and the weight vector of the components that pass the consistency test in the system, and obtains the logical results of each system based on this, and then selects the calculation method of the system health according to the series-parallel relationship. In order to avoid the weight dilution problem, the component health is exponentially calculated to realize system health assessment based on the reliability block diagram.

[0038] This embodiment can implement component-level health assessment based on failure risk and system-level health assessment based on reliability block diagrams when electronic devices lack performance parameters, thereby improving the health assessment capabilities of electronic devices. An electronic device will be used as an example for illustration.

[0039] Abstract the system structure of an electronic device and perform reliability analysis on the system structure to obtain the reliability block diagram of the electronic device, such as Figure 2As shown in Figure 1, the entire electronic device consists of four subsystems connected in series and parallel, totaling 20 components. Tables 1, 2, and 3 show the severity level fuzzy membership, built-in self-test (BIST) diagnostic reports, failure rates, and initial reliability of some components. In these tables, PRODUCTION_NUMBER represents the component's unique identifier, LEVEL represents the severity level (1 to 5), RUNNING_TIME represents the component's operating time, BIT represents the component's current BIST result (0 and 1 correspond to no fault and fault, respectively), INITIAL_RELIABILITY represents the initial reliability, and FAILURE_RATE represents the failure rate (number of failures per unit time). The severity evaluation level, failure rate, and initial reliability of the remaining components (components 6 to 20) are the same as those of components 1 to 5. Assuming the equipment is a newly commissioned system, the operating time of components 6 to 20 is consistent with that of components 1 to 5, and no failures have occurred to date, meaning that the BIT results are all 0.

[0040] Table 1 Fuzzy membership of component severity levels

[0041]

[0042] Table 2 Built-in self-test

[0043]

[0044]

[0045] Table 3 Failure rate and initial reliability of components

[0046]

[0047] In the fuzzy severity level calculation module, the failure type effect and criticality analysis (FMECA) method is used to analyze the components and construct the fuzzy membership of the severity level of the components.

[0048] In this embodiment, the severity level is divided into 1 to 5 levels. The greater the fuzzy membership of the severity level, the greater the degree of membership of the severity level. After determining the severity level set V, an evaluation team is established to evaluate the severity levels of each component and obtain the evaluation of each severity level of each component, as shown in Table 1. The fuzzy severity level calculation module counts the number of supporters x for each severity level according to Table 1. ij (x ij = (number of people whose severity level of component i is level j), and the membership degree b is obtained accordingly. ij =x ij / x i (Membership degree b ijrepresents the degree to which the severity level of component i belongs to the jth level), where x i represents the number of people who evaluate the i-th component; finally, the fuzzy membership vector B of the severity level of the i-th component is calculated, where B = [b i1 ,b i2 ,b i3 ,b i4 ,b i5 ].

[0049] After constructing the severity membership vector B of the component, the fuzzy severity level S of the component is calculated using the centroid method according to the severity level set V, and normalized, as shown in formula (1):

[0050] S=VB T / maxV (1)

[0051] The component health calculation module analyzes and processes the failure rate h and initial reliability R0 of the components in Table 3, calculates the failure probability OPR of the components based on the reliability analysis, and calculates the failure risk RNK of the components by combining the fuzzy severity level S of the components calculated by the fuzzy severity level calculation module and the adjustment coefficient K=0.5 of S, as shown in formula (2):

[0052]

[0053] The health of the component is calculated based on the failure risk of the component, and the calculation method of the component health is selected according to the component built-in self-test (BIT) diagnostic report of Table 2 (BIT is 0 for no fault, and 1 for a fault). Finally, the health of the component is calculated as shown in formula (3):

[0054]

[0055] The component weight adaptive construction module includes a relative importance judgment matrix module, a component weight calculation module, a consistency index verification module and a judgment matrix adaptive adjustment module.

[0056] The relative importance judgment matrix module calculates the relative importance judgment matrix R of each component according to the fuzzy severity level S of each component in the system, and adjusts the relative importance matrix according to the range of the nine-scale judgment scale. The adjustment coefficient is L = 1.365. The relative importance judgment matrix of the system is shown in formula (4):

[0057] R←[S T S] L (4)

[0058] The component weight calculation module calculates the component weight vector according to the relative importance judgment matrix R. First, the relative importance judgment matrix R is summed in the row and column directions to obtain the column sum vector CS. R , and then normalize the relative importance judgment matrix in the column direction, the relative importance judgment matrix R and the column sum vector CS R Divide the number to get the normalized judgment matrix R C =R / CS R , and then normalize the judgment matrix R C Take the row summation process to get the row sum vector RS R , for the normalized judgment matrix R C Sum all elements and get the element and TS R , and then calculate the weight vector W of each component in the system, as shown in formula (5):

[0059] W=RS R / TS R (5)

[0060] The consistency index test module performs consistency test on the relative importance judgment matrix and component weight vector calculated by the relative importance judgment matrix module and the component weight calculation module. First, the dot product of the relative importance judgment matrix R and the weight vector W is calculated, and then the vector λ is divided by the weight vector W to obtain the vector λ. The average value of the vector λ is obtained to obtain the maximum characteristic root λ. max Then, according to the number n of components in the system, the consistency index CI is calculated, and the consistency ratio CR is calculated according to the random consistency index RI corresponding to the number n of components. Finally, if CR < 0.1, the consistency test passes, and the weight vector of the component is output to the system health calculation module. Otherwise, the consistency test fails, and the judgment matrix adaptive adjustment module is enabled. The relevant calculation is shown in formula (6):

[0061]

[0062] The judgment matrix adaptive adjustment module performs adaptive adjustment on the relative importance judgment matrix. First, the adjustment coefficient d=0.5 is initialized, and the elements in the upper right corner of the relative importance judgment matrix are adjusted one by one. The adjustment element r=r d , assign r = 1 / r to the elements symmetrical along the main diagonal of element r, and obtain the adjusted relative importance judgment matrix R←R based on the adjusted elements r The relative importance judgment matrix module and the component weight calculation module are recalculated to obtain the adjusted weight vector W. If the consistency check passes, the component weight is output to the system health calculation module. Otherwise, if the consistency check fails, the judgment matrix adaptive adjustment module continues to adjust.

[0063] The system health calculation module inputs the component health HI completed by the component health calculation module and the weight vector W of the system components completed by the component weight adaptive construction module. According to the system reliability block diagram, the system health is calculated hierarchically according to the logical structure of the reliability block diagram, as shown in formula (7):

[0064]

[0065] According to this embodiment, a device for evaluating the health of an electronic device is provided. Figure 3 As shown, the pseudo code for calculating component and system health assessment is as follows:

[0066] Step S1: Input the component severity level fuzzy membership file, built-in self-test file, component failure rate and initial reliability file, as shown in Tables 1 to 3;

[0067] Step S2: Input the component severity level fuzzy membership file and calculate the component's fuzzy severity level through the fuzzy severity level algorithm (SEVERITY) module. The SEVERITY algorithm is as follows:

[0068] Input: Fuzzy membership vector B of component severity level

[0069] Algorithm flow:

[0070] 1: Get the component severity level set V;

[0071] 2: Calculate the fuzzy severity level of the component based on the centroid method, then S←VB T ;

[0072] 3: Normalize the fuzzy severity level of the component, then S←S / maxV;

[0073] Output: Fuzzy severity level S of the component

[0074] Step S3: Input the built-in self-test file, the component failure rate and initial reliability, and the component fuzzy severity level, and calculate the component health through the component health algorithm (HEALTH) module. The HEALTH algorithm is as follows:

[0075] Input: component fuzzy severity level S, failure rate h, initial reliability R0, built-in self-test result BIT

[0076] Algorithm flow:

[0077] 1: Calculate the probability of component failure, OPR←1-R0exp[-ht];

[0078] 2: Get the severity level adjustment coefficient K←0.5;

[0079] 3: Calculate the component failure risk RNK←OPR 1-KS ;

[0080] 4: Calculate the health of the component HI←1-RNK;

[0081] 5: Adjust component health based on built-in self-test results, HI←HI(1-BIT);

[0082] Output: Component health HI

[0083] The relationship between component health HI, failure probability OPR, and fuzzy severity level S is as follows: Figure 4 , the larger the failure probability OPR and the fuzzy severity level S are, the smaller the health of the component is.

[0084] Step S4: Input the fuzzy severity level of the component and calculate the relative importance judgment matrix of each component in the system through the relative importance judgment matrix construction algorithm (JUDGEMENT) module. The JUDGEMENT algorithm is as follows:

[0085] Input: Fuzzy severity level vector S of each component in the system

[0086] Algorithm flow:

[0087] 1: Get the judgment matrix adjustment coefficient, L←1.365;

[0088] 2: Construct judgment matrix based on fuzzy severity level, R←S T S;

[0089] 3: Adjust the judgment matrix, R←R L ;

[0090] Output: Relative importance judgment matrix R of each component in the system

[0091] Step S5: Input the relative importance judgment matrix of the components, and calculate the weight of each component in the system through the weight calculation algorithm (WEIGHT) module. The WEIGHT algorithm is as follows:

[0092] Input: Relative importance judgment matrix R of each component in the system

[0093] Algorithm flow:

[0094] 1: Sum the relative importance judgment matrix R in the column direction to obtain the column sum vector CS R ;

[0095] 2: Normalize the relative importance judgment matrix in the column direction to obtain the normalized judgment matrix R C ←R / CS R ;

[0096] 3: Normalized judgment matrix R C Perform row summation to obtain the row sum vector RS R ;

[0097] 4: Normalized judgment matrix R C Sum all elements and get the element and TS R ;

[0098] 5: Calculate the weight vector of each component in the system, W←RS R / TS R ;

[0099] Output: Weight vector W of each component in the system

[0100] Step S6: Input the component weights and relative importance judgment matrix, perform consistency verification through the consistency verification algorithm (VALIDATION) module, and obtain the consistency verification result (pass / fail). The VALIDATION algorithm is as follows:

[0101] Input: Relative importance judgment matrix R and weight vector W of each component in the system

[0102] Algorithm flow:

[0103] 1: Calculate the dot product of the relative importance judgment matrix R and the weight vector W, and then divide it by the weight vector W to obtain the vector λ←RW / W;

[0104] 2: Calculate the average value of vector λ and get the maximum characteristic root λ max ;

[0105] 3: Get the number n of components in the system;

[0106] 4: Calculate the consistency index CI←(λ max -n) / (n-1);

[0107] 5: Get the random consistency index RI corresponding to the number of components n;

[0108] 6: Calculate the consistency ratio CR←CI / RI;

[0109] 7: If CR<0.1then

[0110] The consistency test result is V←True;

[0111] Else

[0112] V←False;

[0113] End if

[0114] Output: consistency test result V (True / False)

[0115] Step S7: If the result fails, the relative importance judgment matrix is fine-tuned through the judgment matrix adaptive adjustment algorithm (ADJUST) module, and steps S5 to S7 are repeated; if the result passes, step S8 is entered. The ADJUST algorithm is as follows:

[0116] Input: Relative importance judgment matrix R of each component in the system

[0117] Algorithm flow:

[0118]

[0119]

[0120] 4: Based on the updated relative importance judgment matrix R, call the ADJUST algorithm again

[0121] Output: Adaptively adjusted relative importance judgment matrix R

[0122] Step S8: Input the component health and component weight, and calculate the system health through the system health algorithm (SYSTEM_HEALTH) module. The SYSTEM_HEALTH algorithm is as follows:

[0123] Input: The relative importance judgment matrix R and weight vector W of each component in the system, as well as the health vector HI of each component

[0124] Algorithm flow:

[0125]

[0126]

[0127] Using the above calculation steps, first, the component-level health is calculated according to steps S1 to S2, and then the system-level health is calculated according to steps S3 to S8. The calculation results are shown in Table 4, where SYSTEM_NUMBER identifies the unique identifier of the system or subsystem, and HEALTH identifies the health of the component and system.

[0128] Table 4 Calculation results of the health of an avionics system

[0129]

[0130]

[0131] According to the health of components and systems, the health status of an avionics system is classified and counted, such as Figure 5 As shown in the classification statistics chart, it can be seen from the chart that a large number of components or systems are in abnormal conditions, and it is necessary to troubleshoot a certain avionics equipment system.

[0132] According to the health distribution diagram of components and systems, such as Figure 6 As shown in the figure, it can be seen that the health of the system is in normal state, but there are abnormalities in the level 3 subsystem and the level 1 subsystem. Then the health of the components is checked and it is found that the health status of the series components is in abnormal state. Therefore, it is necessary to take maintenance and repair measures for the components in abnormal state.

[0133] It is understandable that those skilled in the art can make equivalent substitutions or changes based on the technical solution and inventive concept of the present invention, and all these changes or substitutions should fall within the scope of protection of the claims attached to the present invention.

Claims

1. An electronic equipment health assessment device, comprising a component-level health assessment module and a system-level health assessment module, characterized in that: The component-level health assessment module includes a fuzzy severity level calculation module and a component health calculation module. The fuzzy severity level calculation module is used to defuzzify the fuzzy membership of the component severity level and obtain the fuzzy severity level of the component. The component health calculation module is used to analyze the reliability of the component and calculate the failure probability OPR of the component based on the reliability analysis. Combined with the fuzzy severity level S of the component calculated by the fuzzy severity level calculation module and the adjustment coefficient K of S, the failure risk RNK of the component is calculated, as shown in formula (2): Where h is the failure rate of the component, R0 is the initial reliability of the component; t is the operating time of the component The health of the component is evaluated based on the component failure risk, and the calculation method of the component health is selected according to the component's built-in self-test diagnostic report. Finally, the health of the component is calculated as shown in formula (3): Among them, BIT is 0, which means the component is not faulty, and 1, which means the component is faulty; The system-level health assessment module includes a component weight adaptive construction module and a system health calculation module. The component weight adaptive construction module constructs a relative importance judgment matrix based on the fuzzy severity level of the component, and uses an adaptive adjustment algorithm to adjust the relative importance judgment matrix so that the relative importance judgment matrix and the component weight vector pass the consistency test; the system health calculation module first analyzes the reliability block diagram of the electronic equipment based on the component health and the weight vector of the components in the system that pass the consistency test, and obtains the logical results of each system based on this, and then selects the system health calculation method according to the series-parallel relationship to obtain the system health assessment.

2. The electronic device health assessment device according to claim 1, characterized in that The fuzzy severity level calculation module uses the fault type impact and fatality analysis method to analyze the components and construct the fuzzy membership of the severity level of the components: after determining the severity level set V, an evaluation team is established to evaluate the severity levels of each component, obtain the evaluation of each severity level of each component, and count the number of supporters x for each severity level of each component. ij , based on which we get the membership degree b ij =x ij / x i , where x i represents the number of people who evaluated the i-th component, and j represents the severity level of the component; finally, the fuzzy membership vector B of the severity level of the i-th component is calculated; After constructing the fuzzy membership vector B of the component severity level, the fuzzy severity level S of the component is calculated using the centroid method according to the severity level set V, and normalized, as shown in formula (1): S=VB T / maxV (1)。 3. The electronic device health assessment device according to claim 1, characterized in that The component weight adaptive construction module includes a relative importance judgment matrix module, a component weight calculation module, a consistency index test module and a judgment matrix adaptive adjustment module; The relative importance judgment matrix module calculates the relative importance judgment matrix R of each component according to the fuzzy severity level S of each component in the system, and adjusts the relative importance matrix according to the range of the nine-scale judgment scale. The adjustment coefficient is L. The relative importance judgment matrix of the system is shown in formula (4): R←[S T S] L (4) The component weight calculation module calculates the component weight vector based on the relative importance judgment matrix R: First, sum the relative importance judgment matrix R in the row and column directions to obtain the column sum vector CS R , and then normalize the relative importance judgment matrix in the column direction, the relative importance judgment matrix R and the column sum vector CS R Divide the number to get the normalized judgment matrix R C =R / CS R , and then normalize the judgment matrix R C Take the row summation process to get the row sum vector RS R , for the normalized judgment matrix R C Sum all elements and get the element and TS R , and then calculate the weight vector W of each component in the system, as shown in formula (5): W=RS R / TS R (5) The consistency index test module performs consistency test on the relative importance judgment matrix and component weight vector calculated by the relative importance judgment matrix module and the component weight calculation module: first, the dot product of the relative importance judgment matrix R and the weight vector W is calculated, and then the vector λ is divided by the weight vector W to obtain the vector λ. The average value of the vector λ is obtained to obtain the maximum characteristic root λ max Then, according to the number n of components in the system, the consistency index CI is calculated, and the consistency ratio CR is calculated according to the random consistency index RI corresponding to the number n of components. Finally, if CR < 0.1, the consistency test passes, and the weight vector of the component is output to the system health calculation module. Otherwise, the consistency test fails, and the judgment matrix adaptive adjustment module is enabled. The relevant calculation is shown in formula (6): The judgment matrix adaptive adjustment module performs adaptive adjustment on the relative importance judgment matrix. First, the adjustment coefficient d is initialized, and the elements in the upper right corner of the relative importance judgment matrix are adjusted one by one. The adjustment element r=r d , assign r = 1 / r to the elements symmetrical along the main diagonal of element r, and obtain the adjusted relative importance judgment matrix R←R based on the adjusted elements r The relative importance judgment matrix module and the component weight calculation module are recalculated to obtain the adjusted weight vector W. If the consistency check passes, the component weight is output to the system health calculation module. Otherwise, if the consistency check fails, the judgment matrix adaptive adjustment module continues to adjust.

4. The electronic device health assessment device according to claim 1, characterized in that The system health calculation module inputs the component health HI completed by the component health calculation module and the weight vector w of the system component i completed by the component weight adaptive construction module. i According to the reliability block diagram of the system, the health of the system is calculated hierarchically according to the logical structure of the reliability block diagram, as shown in formula (7): Where n is the number of components.

Images