An Improved FMEA Method Based on the Analytic Hierarchy Process (AHP)

Abstract

An improved FMEA method for CubeSats based on the Analytic Hierarchy Process (AHP) is proposed. This method involves assembling an FMEA expert team based on the overall mission and subsystem characteristics of the CubeSat; performing subsystem-level FMEA analysis; and using the AHP to calculate the subjective risk factors (SOD) for each device failure mode, combining expert experience information to calculate expert subjective weights ω = [ω1 ω2 ω3]. The method then corrects the judgment matrix provided by multiple experts and the calculated subjective weight values ​​using a tangent space objective average weighting, resulting in a comprehensive risk factor for each device failure mode, a statistical functional failure mode risk factor, and a subsystem-level risk factor. Finally, the FMEA failure modes for the entire CubeSat are calculated. This invention, considering the characteristics of CubeSats, employs a quantitative and objective expert database and the AHP to comprehensively process various component-level and functional failure modes. This approach enables rapid identification of potentially risky components and analysis of their consequences, while simultaneously providing scientific FMEA analysis results for both subsystems and the entire CubeSat.

Description

Technical Field

[0001] This invention relates to a quantitative analysis method in the field of aerospace quality control technology, specifically an improved FMEA method for CubeSats based on an expert database and the analytic hierarchy process. Background Technology

[0002] In the field of CubeSat development and even aerospace quality control, Failure Mode and Effects Analysis (FMEA) is widely used as an effective method. The traditional analysis process involves identifying and summarizing various failure modes at the individual board and unit levels based on the designer's experience. The drawback of this approach is that it can only qualitatively analyze some potential failure modes at the unit or board level, failing to quantitatively identify specific risky components. Furthermore, it lacks an intuitive method for identifying failure modes at the subsystem level and for the entire satellite. Summary of the Invention

[0003] To address the aforementioned shortcomings of existing technologies, this invention proposes an improved FMEA method for CubeSats based on the Analytic Hierarchy Process (AHP). Combining the characteristics of CubeSats, this method employs a quantitative and objective expert database and AHP to comprehensively process various component-level and functional failure modes. On the one hand, it can quickly locate potentially risky components and analyze their consequences; on the other hand, it can scientifically obtain FMEA analysis results for subsystems and the entire CubeSat.

[0004] This invention is achieved through the following technical solution:

[0005] This invention relates to an improved FMEA method based on an expert database and the analytic hierarchy process (AHP), comprising:

[0006] Step 1: Based on the overall mission and subsystem characteristics of the CubeSat, establish an FMEA expert team, consisting of Chief Satellite Designer Z and n members. Z Person, Subsystem Chief Designer F (total n) F Person, single-machine supervisor D total n D People and single-board designers B together n B Composed of people.

[0007] Step 2: Subsystem-level FMEA analysis.

[0008] Step 3: Using the analytic hierarchy process (AHP), for each device failure mode, the SOD risk factor is calculated by combining expert experience information to determine the expert subjective weight ω = [ω1 ω2 ω3].

[0009] Step 4: Correct the judgment matrix given by multiple experts and the calculated subjective weight values ​​by performing a tangent space objective average weighting to obtain the comprehensive risk factor of each device failure mode, the statistical functional failure mode risk factor, and the subsystem-level risk factor.

[0010] Step 5: Calculate the whole satellite FMEA fault modes.

[0011] Technical effect

[0012] This invention employs a quantitative and objective expert database and the analytic hierarchy process (AHP) to comprehensively process various component-level and functional failure modes. It corrects the subjective expert scores with a cross-space objective average weighting, fully combining the rich experience and objective evaluation of experts to provide risk assessment results. Compared to existing technologies, this invention can quickly locate potentially risky components and analyze their associated risk consequences, while simultaneously obtaining scientific FMEA analysis results for subsystems and the entire satellite. Attached Figure Description

[0013] Figure 1 This is a flowchart of the present invention;

[0014] Figure 2 This is a diagram illustrating the key steps;

[0015] Figure 3 The following is a diagram illustrating the combined risk factors and device failure mode ranking effect of an example. Detailed Implementation

[0016] like Figure 1 and Figure 2 As shown, this embodiment relates to an improved FMEA method based on an expert database and the analytic hierarchy process (AHP), including:

[0017] Step 1: Based on the overall mission and subsystem characteristics of the CubeSat, establish an FMEA expert team, consisting of Chief Satellite Designer Z and n members. Z Person, Subsystem Chief Designer F (total n) F Person, single-machine supervisor D total n D People and single-board designers B together n B Composed of people.

[0018] Step 2: Subsystem-level FMEA analysis.

[0019] Step 2.1: FMEA analysis begins at the subsystem level. In this embodiment, the important measurement and control subsystem is used as an example. After joint consultation by the chief designer F of the measurement and control and related subsystems, the single-machine supervisor D, and the single-board designer B, n potential failure modes of the measurement and control subsystem are determined: FM1, FM2, ..., FMn.

[0020] Step 2.2: Since the CubeSat adopts the special architecture of PC104, the specific implementation of its subsystem level is basically completed at the single machine (single board) level. Therefore, in this embodiment, taking a certain CubeSat equipped with an actual measurement and control single machine as an example, the various fault modes and labels are shown in Table 1.

[0021] Table 1. FMEA Table and Fault Mode Numbers for CubeSat Control Subsystem

[0022]

[0023]

[0024]

[0025]

[0026] Step 2.3: Based on the fault mode numbers in Table 1, the FMEA expert team scores each type of fault according to three dimensions: severity (s), occurrence (o), and detectability (d). The scoring criteria are shown in Table 2.

[0027] Table 2 CubeStar FMEA Scoring Criteria

[0028]

[0029]

[0030] For the FMEA expert team composed of F (chief designer of measurement and control and related subsystems), D (single-machine supervisor), and B (single-board designer), the component-level failure mode analysis mainly refers to the opinions of the experienced single-machine supervisor D and single-board designer B. The three types of experts can be set as weights (a). F =0.2,a D =0.4,a B =0.4), so that the average value can be calculated based on the SOD scores of the three types of experts.

[0031] In this embodiment, the severity s is taken as an example, and its calculation method is as follows: First, consult the FM1 information in Table 1 to determine that FM1 indicates a short-circuit fault in the phase compensation capacitor C56. The potential causes of this fault are electrical / thermal / stress effects or component quality issues. For this level of product, it will result in a zero amplification factor and reduced servo gain. For higher-level products, it will reduce the overall loop gain and decrease stability. The expert scoring result is: 1 subsystem chief designer, score s. F =5.10; 1 single-machine supervisor, score s D =4.90; 3 single-board designers, score s D1 =4.95, s D2 =5.30, s D3 =5.12. Calculations show that the short-circuit fault severity s of the FM1 compensation capacitor C56 is 5.03. Other calculations follow the same pattern to obtain the scores for all fault modes.

[0032] Table 3. FMEA Scoring Table for CubeSat Tracking and Control Subsystem

[0033]

[0034] Step 3: Calculate the expert subjective weights ω = [ω1 ω2 ω3] of the SOD risk factors for each device failure mode using the analytic hierarchy process (AHP).

[0035] Step 3.1: Obtain advice from experts, compare the importance of the three risk factors (SOD), and construct a judgment matrix E = [e...]. ii' ]| 3×3 The scaling criteria for the judgment matrix are shown in Table 4.

[0036] Table 4. Criteria for Determining Matrix Scale

[0037]

[0038] Step 3.2: Calculate the consistency ratio C. R Determine the consistency of pairwise comparison matrices

[0039] When C R If the result is less than 0.1, then the judgment matrix is ​​considered to meet the consistency requirement. Where: C I =(λ max -n) / (n-1) is the consistency index, λ max Let n be the largest eigenvalue of matrix E, and let n be its order. R I The random consistency index is obtained from Table 5.

[0040] Table 5 Consistency Indicators

[0041]

[0042] Step 3.3, Weight Calculation: Calculate the relative weight EW = λ of each indicator relative to the elements of the previous layer from the judgment matrix using the eigenvalue method. max W, where W is the feature vector, and the corresponding index weights are obtained after normalization.

[0043] In this embodiment, it is assumed that the expert constructs a judgment matrix based on scaling table 4. The largest eigenvalue λ is calculated. max =3.0055, thus obtaining C I =0.0055, C R =0.0028 < 0.1, therefore the judgment matrix given by the experts meets the consistency requirement.

[0044] Further, the relative weight eigenvector W is obtained through the eigenvalue method, and after normalization, the final result is obtained.

[0045] ω=[0.8902 0.4132 0.1918].

[0046] Step 4: Correct the judgment matrix given by multiple experts and the calculated subjective weight values ​​by performing a tangent space objective average weighting to obtain the comprehensive risk factor of each device failure mode, the statistical functional failure mode risk factor, and the subsystem-level risk factor.

[0047] Step 4.1: First, based on Step 3, record E from the judgment matrix obtained from any expert's subjective scoring. i ,i=1,...,n, where n is the number of experts.

[0048] To avoid bias in the judgment matrix caused by subjective scoring from multiple experts, which could affect the objectivity of the comprehensive risk factors, this paper adopts the tangent space average matrix calculation method to seek the average judgment matrix. Where SO(κ) denotes the tangent space of a κ-dimensional matrix, d(E,E) i () represents two judgment matrices E, E i The tangent space distance, that is, under this definition, minimizes the sum of squared distances from the average judgment matrix to the subjective scoring judgment matrices given by all experts.

[0049] In this embodiment, the tangent space distance between the two judgment matrices is defined as d(E,E). i )=||log(E)-log(E i This allows us to write the calculation of the tangent space average matrix as follows: Where exp and log are matrix exponentiation and matrix logarithm operations, respectively.

[0050] After calculation, the average judgment matrix can be obtained. This updates the expert subjective weights ω = [ω1 ω2 ω3] to the objective average corrected weights of the tangent space. The corrected weights of the SOD risk factors for each device failure mode were calculated. And record it.

[0051] The comprehensive risk factor for each device failure mode is calculated as follows: The final comprehensive risk factors and ranking of device failure modes are shown in Table 6.

[0052] Table 6. Comprehensive Risk Factors and Ranking of Device Failure Modes

[0053]

[0054]

[0055] As shown in the table above, for single-unit measurement and control products, the contribution of each component to the risk varies. Among them, the FM18 feedback resistor has the greatest impact, ranking first. This indicates that in the actual space environment, due to the influence of electricity / heat / stress or the quality of the components, the resistance output of this product may deviate from the positive value, causing the upstream product to malfunction. This risk should be given special attention.

[0056] Step 4.2: According to Table 1, after obtaining the comprehensive risk factor of component failure modes, calculate the functional failure mode risk factor according to reasonable criteria in order to simplify the FMEA analysis at the subsystem level.

[0057] According to Table 1, component failure categories can be broadly classified into four types: open circuit, short circuit, parameter drift, and others. Based on the component grade information provided by the component suppliers, the probability distribution of these four failure modes can be determined, which is assumed to be f. k ,f d ,f p ,f q The four types of failure modes cover all failure causes for this function item; therefore, the functional failure modes can be calculated as follows: Where i = 1, 2, 3, ..., 18, representing 18 functional items.

[0058] Step 4.3: Calculate the fault risk factor value at the measurement and control subsystem level. First, based on expert opinions, obtain the pairwise comparison judgment matrix E of the importance of 18 functional items. f =[e jj' ]| 18×18 Then, a consistency judgment is performed using the eigenvalue method E. f W = λ max W obtains the eigenvectors of the judgment matrix, and after normalization, the corresponding risk factor index weights ω are obtained. j ,j=1,2,3,...,18;Calculate the fault risk factor at the measurement and control subsystem level.

[0059] Step 5: Calculate the whole satellite FMEA fault modes.

[0060] Step 5.1: First, the entire satellite subsystem is analyzed. In this embodiment, the CubeSat subsystem includes: structural subsystem, satellite operations subsystem, telemetry and control subsystem, power supply subsystem, control subsystem, payload subsystem, and thermal control subsystem. After analyzing and summarizing each subsystem, the overall satellite failure modes are summarized in Table 7.

[0061] Table 7: Full Star FMEA Table

[0062]

[0063]

[0064] Step 5.2: Calculate the overall satellite failure risk factor value. Based on expert opinions, obtain the pairwise comparison judgment matrix E for the importance of the seven subsystems. t =[e kk' ]| 7×7 Then, a consistency judgment is performed using the eigenvalue method E. t W = λ max W obtains the eigenvectors of the judgment matrix, and after normalization, the corresponding risk factor index weights ω are obtained. k k = 1, 2, 3, ..., 7; Calculate the overall satellite failure risk factor.

[0065] To verify the effectiveness of this method, a practical experiment was conducted, repeating step 2 and changing the expert composition. Assuming the FMEA expert team consists of the chief designer F, the single-machine supervisor D, and the single-board designer B, the three types of experts can be set with weights (a...). F =0.4,a D =0.2,a B =0.4), specifically, one chief designer of the subsystem, two single-machine supervisors, and two single-board designers, under the same conditions, give the scoring results of all failure modes, as shown in Table 8.

[0066] Table 8. FMEA Scoring Table for CubeSat Measurement and Control Subsystem after Expert Replacement

[0067]

[0068] Based on the normalization process in step 3 and the objective average weight correction in step 4, the comprehensive risk factor for each device failure mode is calculated as follows: The order is shown in Table 9.

[0069] Table 9. Comprehensive Risk Factors and Ranking of Component Failure Modes After Expert Replacement (Table 9: Comprehensive Risk Factors and Ranking of Component Failure Modes after Expert Replacement)

[0070]

[0071] It can be seen that after changing the experts, the risk ranking for single-unit measurement and control products remains basically unchanged compared to Table 6, such as... Figure 3 The FM18 feedback resistor was accurately identified as having the greatest impact, ranking it as number one. Similarly, the risks of other component failure modes were also accurately identified, indicating that in the actual space environment, due to electrical / thermal / stress effects or component quality issues, the output of this level of product may deviate from the positive value, causing the upstream product to malfunction. This risk should be given special attention.

[0072] Compared with existing technologies, this method combines the characteristics of CubeSats, adopts a quantitative and objective expert database and the analytic hierarchy process to comprehensively process various component-level failure modes and functional failure modes. It corrects the subjective scoring results of experts by using a tangent space objective average weighting, and fully combines the rich experience and objective evaluation of experts. On the one hand, it can quickly locate potential risky components and analyze the consequences, and on the other hand, it can scientifically obtain the FMEA analysis results of subsystems and the entire satellite.

[0073] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A cubestar-based improved FMEA method based on an expert database and the analytic hierarchy process, characterized in that, include: Step 1: Based on the overall mission and subsystem characteristics of the CubeSat, establish an FMEA expert team, consisting of Chief Satellite Designer Z and n members. Z Person, Subsystem Chief Designer F (total n) F Person, single-machine supervisor D total n D People and single-board designers B together n B Composed of people; Step 2: Subsystem-level FMEA analysis; Step 3: Using the analytic hierarchy process (AHP), for each device failure mode's SOD risk factor, and combining expert experience information, subjective weighting is performed by experts. calculate; Step 4: Based on the judgment matrix provided by multiple experts and the calculated subjective weight values, perform tangent space objective average weight correction to obtain the comprehensive risk factor for each device failure mode, the statistical functional failure mode risk factor, and the subsystem-level risk factor, specifically including: Step 4.1: First, based on Step 3, record the judgment matrix obtained from any expert's subjective scoring. , The number of experts is used; the average judgment matrix is ​​calculated using the tangent space average matrix method. ,in: express dimensional matrix tangent space, Represents two judgment matrices The tangent space distance, defined under this condition, is the distance between the average judgment matrix and the subjective scoring judgment matrices given by all experts, where the sum of the squared distances is minimized. The calculation of the tangent space average matrix is ​​written as Where exp and log are the matrix exponent and logarithm operations, respectively; thus, the average judgment matrix is ​​obtained. This updates the expert subjective weights. To adjust the weights by objective average of the tangent space The corrected weights of the SOD risk factors for each device failure mode were calculated. And record it; Step 4.2: After obtaining the comprehensive risk factor of each device's failure mode, calculate the functional failure mode risk factor according to reasonable criteria in order to simplify the subsystem-level FMEA analysis. Step 4.3: Calculate the fault risk factor value at the measurement and control subsystem level: First, based on expert opinions, obtain a pairwise comparison judgment matrix of the importance of 18 functional items. Then, a consistency judgment is performed using the eigenvalue method. The eigenvectors of the judgment matrix are obtained, and the corresponding risk factor index weights are derived after normalization. ; Calculate the fault risk factors at the measurement and control subsystem level ; Step 5: Calculate the overall satellite FMEA failure modes.

2. The CubeStar Improved FMEA Method based on Expert Database and Analytic Hierarchy Process (AHP) as described in claim 1, characterized in that, Step 2 specifically includes: Step 2.1: FMEA analysis begins at the subsystem level. In this embodiment, the important measurement and control subsystem is used as an example. After joint consultation by the chief designer F of the measurement and control and related subsystems, the single-machine supervisor D, and the single-board designer B, n potential failure modes of the measurement and control subsystem are determined: FM1, FM2, …, FMn. Step 2.2: Because CubeSat adopts the special architecture of PC104, its subsystem-level implementation is basically completed at the single-machine level; Step 2.3: Based on the fault mode number, the FMEA expert team scores each type of fault according to three dimensions: severity, occurrence, and detectability.

3. The CubeStar Improved FMEA Method based on Expert Database and Analytic Hierarchy Process (AHP) as described in claim 1, characterized in that, Step 3 specifically includes: Step 3.1: Obtain advice from experts, compare the importance of the three risk factors of SOD, and construct a judgment matrix. ; Step 3.2: Calculate the consistency ratio Determine the consistency of pairwise comparison matrices ;when If the judgment matrix satisfies the consistency requirement, then: As a consistency indicator, For matrix The largest eigenvalue, Its order, It is a random consistency indicator; Step 3.3, Weight Calculation: Calculate the relative weight of each indicator relative to the elements of the previous layer from the judgment matrix using the eigenvalue method. ,in: It is a feature vector, and the corresponding index weights are obtained after normalization.

4. The CubeStar Improved FMEA Method based on Expert Database and Analytic Hierarchy Process (AHP) as described in claim 1, characterized in that, Step 5 specifically includes: Step 5.1: First, the entire satellite subsystem is sorted out. In this embodiment, the CubeSat subsystem includes: structural subsystem, satellite service subsystem, telemetry and control subsystem, power supply subsystem, control subsystem, payload subsystem, and thermal control subsystem. After analyzing and summarizing each subsystem; Step 5.2: Calculate the overall satellite failure risk factor value, and based on expert opinions, obtain a pairwise comparison judgment matrix of the importance of the seven subsystems. Then, a consistency judgment is performed using the eigenvalue method. The eigenvectors of the judgment matrix are obtained, and the corresponding risk factor index weights are derived after normalization. ; Calculate the overall satellite failure risk factor .

Images