A method for predicting the remaining life of aerospace BDR modules based on particle swarm optimization

Through the method based on particle swarm algorithm and combined with the degradation mechanism model, the accuracy of the life prediction of the aerospace BDR module is solved, and the high-accurate life prediction of aerospace electronic products is achieved, reducing system complexity and cost.

CN114239373BActive Publication Date: 2024-11-12SHANGHAI INST OF SPACE POWER SOURCES
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
CN202111551123.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-17
Publication Date
2024-11-12
Estimated Expiration
2041-12-17

AI Technical Summary

Technical Problem

The prior art cannot accurately predict the remaining life of aerospace BDR module, resulting in the dependence of backup modules to increase system complexity, volume, weight and cost, and traditional statistical methods are not sufficient to meet the life prediction needs of aerospace electronic products.

Method used

Using a method based on particle swarm algorithm, the degradation mechanism model is constructed through degradation mechanism analysis, the particle swarm algorithm is used to achieve self-excitation of model parameters, and an accurate residual life prediction model of aerospace BDR module is established, and the degradation mechanism model and particle swarm algorithm are integrated to improve prediction. Accuracy.

Benefits of technology

Accurate residual life prediction of aerospace BDR module is achieved, the modeling difficulties and parameter determination difficulties in traditional methods are overcome, the accuracy of life prediction of aerospace electronic products is improved, and prediction support is provided for other aerospace electronic products.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114239373B_ABST
    Figure CN114239373B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm, including: S1, determining key components; S2, carrying out degradation mechanism analysis for key components; S3, constructing a degradation mechanism model of an aerospace BDR module; S4, calculating the root mean square error of the module remaining life prediction, and using it as the fitness function of the particle swarm algorithm; S5, calculating the particle fitness value according to the fitness function of the particle swarm; S6, recording the optimal position of individuals and populations, updating the particle speed and position, and determining the global optimal position of individuals and populations; S7, judging whether the root mean square error reaches the minimum, if it does not reach the minimum value, returning to step S5, continuing to iterate, if it reaches the minimum value or the number of iterations is equal to the maximum number of iterations, ending the iteration. The present invention can predict the remaining life of an aerospace BDR module, and the remaining life prediction method obtained also provides a method support for guiding the remaining life prediction of other aerospace electronic products.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of remaining life prediction of aerospace electronic products, and in particular to a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm. Background Art

[0002] The aerospace BDR module is a key module in the power controller and is a type of aerospace electronic product. Its remaining service life also determines the remaining service life of the spacecraft. At present, there is little research on the remaining service life prediction of aerospace BDR modules in China. Traditional statistical methods cannot accurately predict the remaining service life of aerospace BDR modules. Many spacecraft still use backup aerospace BDR modules to improve system stability. This method that relies on product over-design not only increases the volume, weight and production costs, but also increases the complexity of system control. Summary of the invention

[0003] In order to improve the research level of the remaining life prediction of aerospace electronic products in my country, the present invention provides a method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm. The degradation mechanism model is constructed through degradation mechanism analysis, and the particle swarm algorithm is used to realize self-optimization of model parameters, and an accurate remaining life prediction model of aerospace BDR modules is established. It can support relevant aerospace power design units and users to implement effective life prediction for aerospace BDR modules, and further improve the level of remaining life prediction of aerospace electronic products in my country.

[0004] The present invention provides a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm, which is characterized by comprising:

[0005] S1. Analyze the historical data of the aerospace BDR module and determine the key components that affect the module life;

[0006] S2. Carry out degradation mechanism analysis for key components;

[0007] S3. Based on the degradation mechanism analysis, the degradation mechanism model of aerospace BDR module is constructed;

[0008] S4. Calculate the root mean square error RMSE of the remaining life prediction of the aerospace BDR module and use it as the fitness function of the particle swarm algorithm.

[0009] S5. Calculate the particle fitness value according to the particle swarm fitness function;

[0010] S6, record the optimal position of individuals and populations, update particle speed and position, and determine the global optimal position of individuals and populations;

[0011] S7. Determine whether the root mean square error RMSE reaches the minimum. If not, return to step S5 and continue iterating. If it reaches the minimum or the number of iterations is equal to the maximum number of iterations, end the iteration and complete the remaining life modeling of the aerospace BDR module.

[0012] Optionally, the step S1 specifically includes:

[0013] S11. Collect historical data of the aerospace BDR module and conduct data analysis on it;

[0014] S12. Compare possible failure mechanisms and ignore failure factors with less impact, so as to determine the key components that affect the module life.

[0015] Optionally, the step S2 specifically includes:

[0016] S21. Considering the working time and working temperature, the Arrhenius model is used to analyze the degradation mechanism of the resistor. The expression is:

[0017]

[0018] Among them, ΔR and R are the resistance difference and resistance value at time t respectively, T is the temperature, and B, C, D and F are all constants;

[0019] S22. Considering the operating temperature, the degradation mechanism of the capacitor is analyzed using the Arrhenius model, and the expression is:

[0020] c=A·exp(-E / kT)

[0021] Where c is capacitance, T is temperature, E is the voltage across the capacitor, and A and k are constants;

[0022] S23. Taking the inductance value as the fault characteristic parameter of the inductor, the degradation mechanism analysis of the inductor is carried out, and the expression is:

[0023] L(t)=L 0 -λt

[0024] Where, L(t) is the inductance value at time t, L 0 is the nominal value of the inductance at the initial moment, λ is the degradation model parameter;

[0025] S24. The on-resistance is indirectly used as a fault characteristic parameter reflecting the degree of degradation of the power diode, and the degradation mechanism analysis of the power diode is carried out. The expression is:

[0026] ΔR D =a·[e b·t -1]

[0027] Among them, ΔR DThe on-resistance at time t, a and b are constants;

[0028] S25. Without considering the failure of the package structure and only considering the failure of the chip structure, the degradation mechanism analysis of MOSFET is carried out, and the expression is:

[0029]

[0030] Where, ΔV th is the gate-source voltage when the device is just turned on, C ox is the oxide layer capacitance, ΔQ ot and ΔQ it They represent the changes in the number of oxide layer trapped charges and interface state trapped charges respectively.

[0031] Optionally, the step S3 specifically includes:

[0032] S31. Establish a degradation model based on moisture and heat, the expression is:

[0033]

[0034] Among them, L(H,T) represents the wet heat aging life, d and c are both unknown model parameters, A is a constant, H is the relative humidity, and T is the absolute temperature;

[0035] S32, based on the degradation model, using the least square method to implement parameter identification to determine the parameters of the degradation model;

[0036] S33, considering the data errors, the RANSAC algorithm is used to remove outliers;

[0037] S34, fitting the data after removing the outliers again, and extrapolating the degradation model of the fitting result to the failure threshold, to obtain the failure time of the key components, thereby obtaining the remaining service life prediction value of the key components.

[0038] Optionally, the step S4 specifically includes:

[0039] S41, calculating the root mean square error RMSE of the remaining service life prediction of the aerospace BDR module according to the predicted value of the remaining service life obtained in step S3;

[0040] The root mean square error RMSE formula is as follows:

[0041]

[0042] Where n represents the number of predictions, x(i) represents the actual life of the aerospace BDR module, Represents the predicted life of the aerospace BDR module;

[0043] S42. Use the root mean square error as the fitness function of the particle swarm algorithm.

[0044] Optionally, the step S5 specifically includes:

[0045] S51, determining the speed and position of the current particle;

[0046] S52. Determine the fitness value of the particle according to the speed and position of the particle.

[0047] Optionally, the step S6 specifically includes:

[0048] S61. During each search, record the optimal position of each particle:

[0049]

[0050] Among them, p i is the local optimal position;

[0051] S62, particle speed v i and position x i The update method is as follows:

[0052]

[0053] Where, d = 1, 2, 3...N, w is the inertia factor, c 1 and c 2 is the acceleration constant, α is the constraint factor, and p g is the global optimal position;

[0054] S63. The optimal solution of the optimal position of all particles can be regarded as the optimal position of the entire particle swarm:

[0055]

[0056] Among them, g i and G i are the optimal positions of individuals and populations, respectively.

[0057] Optionally, the step S7 specifically includes:

[0058] S71, recording the root mean square error RMSE value in each iteration process;

[0059] S72, if RMSE does not reach the minimum value, return to step S5, continue to iterate, update the particle speed and position, and continue to find the global optimal position of the particle swarm;

[0060] S73. If RMSE reaches the minimum value or the maximum number of iterations, the particle swarm optimization is completed and the remaining life modeling of the aerospace BDR module is completed.

[0061] The technical solution provided by the present invention may include the following beneficial effects:

[0062] The present invention provides a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm, which has obvious advantages. It integrates the degradation mechanism model and the particle swarm algorithm, realizes the self-optimization of the key parameters of the degradation model, and can accurately predict the remaining life of the aerospace BDR module. In addition, the present invention further overcomes the problems of difficulty in modeling and parameter determination based on traditional mechanism degradation methods by integrating the degradation mechanism model and the particle swarm algorithm, improves the accuracy of the remaining life prediction of the aerospace BDR module, and the remaining life prediction method obtained also provides a method support for guiding the remaining life prediction of other aerospace electronic products.

[0063] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for describing the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0065] Figure 1 The present invention is a flowchart of a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm. DETAILED DESCRIPTION

[0066] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0067] The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the" and "the" used in the present invention and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0068] It should be understood that although the terms "first", "second", "third", etc. may be used to describe various information in the present invention, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present invention, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0069] The embodiment of the present invention provides a method for predicting the remaining life of an aerospace BDR module based on a particle swarm algorithm, which integrates the degradation mechanism model and the particle swarm algorithm, realizes self-optimization of key parameters of the degradation model, and can accurately predict the remaining life of the aerospace BDR module.

[0070] The technical solution of the embodiment of the present invention will be described in detail below with reference to the accompanying drawings.

[0071] See also Figure 1 This embodiment provides a method for predicting the remaining life of an aerospace BDR module using a particle swarm algorithm. Figure 1 As shown, it includes:

[0072] S1. Analyze historical data of aerospace BDR modules and identify key components that affect module life;

[0073] S11. Collect historical data of the Aerospace BDR module, namely, the voltage, current, humidity and temperature data of components such as resistors, capacitors, inductors, power diodes and MOSFETs, and conduct data analysis on them;

[0074] S12. By analyzing historical data, the impact of the failure mechanism is ranked, the possible failure mechanisms are compared, and the failure factors with smaller impact are ignored, so as to determine the key components that affect the module life: resistors, capacitors, inductors, power diodes and MOSFETs;

[0075] S2. Carry out degradation mechanism analysis for key components;

[0076] S21. Considering the working time and working temperature, the Arrhenius model is used to analyze the degradation mechanism of the resistor. The common expression is:

[0077]

[0078] Among them, ΔR and R are the resistance difference and resistance value at time t respectively, T is the temperature, and B, C, D and F are all constants;

[0079] S22. Considering the operating temperature, the degradation mechanism of the capacitor is analyzed using the Arrhenius model. The commonly used expression is:

[0080] c=A·exp(-E / kT)

[0081] Where c is capacitance, T is temperature, E is the voltage across the capacitor, and A and k are constants;

[0082] S23. Taking the inductance value as the fault characteristic parameter of the inductor, the degradation mechanism analysis of the inductor is carried out. The commonly used expression is:

[0083] L(t)=L 0 -λt

[0084] Where, L(t) is the inductance value at time t, L 0 is the nominal value of the inductance at the initial moment, λ is the degradation model parameter;

[0085] S24. The on-resistance is indirectly used as a fault characteristic parameter to reflect the degree of degradation of the power diode. The degradation mechanism analysis of the power diode is carried out. The commonly used expression is:

[0086] ΔR D =a·[e b·t -1]

[0087] Among them, ΔR D The on-resistance at time t, a and b are constants;

[0088] S25. Without considering the failure of the package structure, only considering the failure of the chip structure, the degradation mechanism analysis of MOSFET is usually expressed as:

[0089]

[0090] Where, ΔV th is the gate-source voltage (threshold voltage) when the device is just turning on, C ox is the oxide layer capacitance, ΔQ ot and ΔQ it They represent the changes in the number of oxide layer trap charges and interface state trap charges respectively;

[0091] S3. Based on the degradation mechanism analysis, the degradation mechanism model of aerospace BDR module is constructed;

[0092] S31. Establishment of a degradation model based on moisture and heat:

[0093]

[0094] Among them, L(H,T) represents the wet heat aging life, d and c are both unknown model parameters, A is a constant, H is the relative humidity, and T is the absolute temperature;

[0095] S32, based on the degradation model, using the least square method to implement parameter identification to determine the parameters of the degradation model;

[0096] S33, considering the data errors, the RANSAC algorithm is used to remove outliers;

[0097] The RANSAC algorithm is a fast algorithm for eliminating outliers in estimating camera poses. It uses a minimum subsample of the observed data to estimate model parameters. This method is particularly effective when the data contains a large number of outliers. In order to obtain good performance, normal values ​​should account for at least 50% of the data set.

[0098] S34, fitting the data after removing the outliers again, and extrapolating the degradation model of the fitting result to the failure threshold, to obtain the failure time of the key components, thereby obtaining the remaining service life prediction value of the key components;

[0099] S4. Calculate the root mean square error RMSE of the remaining life prediction of the aerospace BDR module and use it as the fitness function of the particle swarm algorithm.

[0100] S41, calculating the root mean square error RMSE of the remaining service life prediction of the aerospace BDR module according to the predicted value of the remaining service life obtained in step S3;

[0101] The root mean square error RMSE formula is as follows:

[0102]

[0103] Where n represents the number of predictions, x(i) represents the actual life of the aerospace BDR module, Represents the predicted life of the aerospace BDR module;

[0104] S42, taking the root mean square error as the fitness function of the particle swarm algorithm;

[0105] S5. Calculate the particle fitness value according to the particle swarm fitness function;

[0106] S51, determining the speed and position of the current particle;

[0107] S52, determining the particle fitness value according to the particle speed and position;

[0108] S6, record the optimal position of individuals and populations, update particle speed and position, and determine the global optimal position of individuals and populations;

[0109] S61. During each search, record the optimal position of each particle:

[0110]

[0111] Among them, p i is the local optimal position.

[0112] S62, particle speed v i and position x i The update method is as follows:

[0113]

[0114] Where, d = 1, 2, 3...N, w is the inertia factor, c 1 and c 2 is the acceleration constant, α is the constraint factor, and p g is the global optimal position.

[0115] S63. The optimal solution of the optimal position of all particles can be regarded as the optimal position of the entire particle swarm:

[0116]

[0117] Among them, g i and G i are the optimal positions of individuals and populations respectively;

[0118] S7. Determine whether the root mean square error RMSE reaches the minimum. If not, return to step S5 and continue iterating. If it reaches the minimum or the number of iterations is equal to the maximum number of iterations, end the iteration and complete the remaining life modeling of the aerospace BDR module.

[0119] S71, recording the root mean square error RMSE value in each iteration process;

[0120] S72, if the root mean square error RMSE does not reach the minimum value, return to step S5, continue to iterate, update the particle speed and position at the same time, and continue to find the global optimal position of the particle swarm;

[0121] S73. If the root mean square error RMSE reaches the minimum value or the maximum number of iterations, the particle swarm optimization is completed and the remaining life modeling of the aerospace BDR module is completed.

[0122] The above description is only an embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement and improvement made within the spirit and scope of the present invention are included in the protection scope of the present invention.

Claims

1. A method for predicting the remaining life of aerospace BDR modules based on particle swarm optimization, characterized in that: include: S1. Analyze the historical data of the aerospace BDR module and determine the key components that affect the module life; S2. Carry out degradation mechanism analysis for key components: S21. Considering the working time and working temperature, the Arrhenius model is used to analyze the degradation mechanism of the resistor. The expression is: Among them, ΔR and R are the resistance difference and resistance value at time t respectively, T is the temperature, and B, C, D and F are all constants; S22. Considering the operating temperature, the degradation mechanism of the capacitor is analyzed using the Arrhenius model, and the expression is: c=A·exp(-E / kT) Where c is capacitance, T is temperature, E is the voltage across the capacitor, and A and k are constants; S23. Taking the inductance value as the fault characteristic parameter of the inductor, the degradation mechanism analysis of the inductor is carried out, and the expression is: L(t)=L0-λt Where, L(t) is the inductance value at time t, L0 is the nominal value of the inductance at the initial time, and λ is the degradation model parameter; S24. The on-resistance is indirectly used as a fault characteristic parameter reflecting the degree of degradation of the power diode, and the degradation mechanism analysis of the power diode is carried out. The expression is: ΔR D =a·[e b·t -1] Among them, ΔR D The on-resistance at time t, a and b are constants; S25. Without considering the failure of the package structure and only considering the failure of the chip structure, the degradation mechanism analysis of MOSFET is carried out, and the expression is: Where, ΔV th is the gate-source voltage when the device is just turned on, C ox is the oxide layer capacitance, ΔQ ot and ΔQ it They represent the changes in the number of oxide layer trap charges and interface state trap charges respectively; S3. Based on the degradation mechanism analysis, the degradation mechanism model of the aerospace BDR module is constructed: S31. Establish a degradation model based on moisture and heat, the expression is: Among them, L(H,T) represents the wet heat aging life, d and c are both unknown model parameters, A is a constant, H is the relative humidity, and T is the absolute temperature; S32, based on the degradation model, using the least square method to implement parameter identification to determine the parameters of the degradation model; S33, considering the data errors, the RANSAC algorithm is used to remove outliers; S34, fitting the data after removing the outliers again, and extrapolating the degradation model of the fitting result to the failure threshold, to obtain the failure time of the key components, thereby obtaining the remaining service life prediction value of the key components; S4. Calculate the root mean square error RMSE of the remaining life prediction of the aerospace BDR module and use it as the fitness function of the particle swarm algorithm. S5. Calculate the particle fitness value according to the particle swarm fitness function; S6, record the optimal position of individuals and populations, update particle speed and position, and determine the global optimal position of individuals and populations; S7. Determine whether the root mean square error RMSE reaches the minimum. If not, return to step S5 and continue iterating. If it reaches the minimum or the number of iterations is equal to the maximum number of iterations, end the iteration and complete the remaining life modeling of the aerospace BDR module.

2. The method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm according to claim 1, characterized in that: The step S1 specifically includes: S11. Collect historical data of the aerospace BDR module and conduct data analysis on it; S12. Compare possible failure mechanisms and ignore failure factors with less impact, so as to determine the key components that affect the module life.

3. The method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm according to claim 1, characterized in that: The step S4 specifically includes: S41, calculating the root mean square error RMSE of the remaining service life prediction of the aerospace BDR module according to the predicted value of the remaining service life obtained in step S3; The root mean square error RMSE formula is as follows: Where n represents the number of predictions, x(i) represents the actual life of the aerospace BDR module, Represents the predicted life of the aerospace BDR module; S42. Use the root mean square error as the fitness function of the particle swarm algorithm.

4. The method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm according to claim 1, characterized in that: The step S5 specifically includes: S51, determining the speed and position of the current particle; S52. Determine the particle fitness value according to the particle speed and position.

5. The method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm according to claim 1, characterized in that: The step S6 specifically includes: S61. During each search, record the optimal position of each particle: Among them, p i is the local optimal position; S62, particle speed v i and position x i The update method is as follows: Where d = 1, 2, 3...N, w is the inertia factor, c1 and c2 are acceleration constants, α is the constraint factor, p g is the global optimal position; S63. The optimal solution of the optimal position of all particles can be regarded as the optimal position of the entire particle swarm: Among them, g i and G i are the optimal positions of individuals and populations, respectively.

6. The method for predicting the remaining life of aerospace BDR modules based on particle swarm algorithm according to claim 1, characterized in that: The step S7 specifically includes: S71, recording the root mean square error RMSE value in each iteration process; S72, if RMSE does not reach the minimum value, return to step S5, continue to iterate, update the particle speed and position, and continue to find the global optimal position of the particle swarm; S73. If RMSE reaches the minimum value or the maximum number of iterations, the particle swarm optimization is completed and the remaining life modeling of the aerospace BDR module is completed.

Citation Information

Patent Citations

  • Particle filter residual life forecasting method based on accelerated degradation data

    CN102542155A

  • Method for analyzing reliability of explosive logic network based on evaluation of accelerated life

    CN107228926A

  • A residual life prediction method considering the degradation characteristics of lithium batteries

    CN109446661A

  • Lithium ion battery remaining life prediction method based on wolf pack optimization LSTM network

    CN109993270A

  • Lithium-ion battery remaining life prediction method

    CN110007240A