CN-122017604-A - Method and system for estimating state of health based on fuel cell

CN122017604ACN 122017604 ACN122017604 ACN 122017604ACN-122017604-A

Abstract

The invention relates to the technical field of fuel cell state evaluation, in particular to a fuel cell state-of-health estimation method and system. Carrying out ohm, activation and concentration polarization component decomposition on the polarization curve, extracting a corresponding voltage loss component, carrying out correlation matching calculation on the voltage loss component and the impedance spectrum, and constructing a polarization loss-impedance characteristic mapping relation. Extracting a pressure fluctuation mode synchronous with the time sequence of the polarization curve, and identifying an additional concentration loss component caused by gas diffusion disorder by correlating concentration polarization loss voltage. And (5) merging the components, the mapping relation and the reference aging model to calculate the comprehensive health state index. The invention can accurately reflect the multiple aging superposition states of the membrane electrode, the proton exchange membrane and the gas diffusion layer, refine aging cause distinction and improve the accuracy of the representation of the health state.

Inventors

WANG ZHENKUN
TANG XINGWANG
YU FENG
QIAO HAN
LI XINYU

Assignees

吉林大学

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. A fuel cell state of health estimation-based method, the method comprising: Continuously acquiring a dynamic electrochemical data set of a target fuel cell in a plurality of complete operation cycles, wherein the dynamic electrochemical data set comprises a cell polarization curve, an AC impedance spectrum in a galvanic pile and a gas flow channel pressure pulsation time sequence; Carrying out component decomposition treatment of ohmic polarization, activation polarization and concentration polarization on the battery monomer polarization curve, and extracting a voltage loss component set representing different aging mechanisms, wherein the voltage loss component set comprises ohmic polarization loss voltage, activation polarization loss voltage and concentration polarization loss voltage; Performing correlation matching calculation by using the voltage loss component set and the AC impedance spectrum in the pile, and establishing a polarization loss-impedance characteristic mapping relation which is used for calculating a corresponding polarization loss component from the impedance spectrum; extracting a pressure fluctuation mode synchronous with the acquisition time of the cell polarization curve from the gas flow channel pressure fluctuation time sequence, performing correlation analysis on the pressure fluctuation mode and the concentration polarization loss voltage, and identifying an additional concentration loss component caused by gas diffusion barriers; And (3) integrating the additional concentration difference loss component, the polarization loss-impedance characteristic mapping relation and a preset reference aging model, and calculating the comprehensive health state index of the target fuel cell, wherein the comprehensive health state index reflects the superposition effects of membrane electrode catalytic activity attenuation, proton exchange membrane dry-wet aging and gas diffusion layer blocking state.
2. The method according to claim 1, wherein the performing component decomposition processing of ohmic polarization, active polarization and concentration polarization on the cell polarization curve to extract a set of voltage loss components characterizing different aging mechanisms comprises: Obtaining a curve slope through linear regression fit in a high current density linear region of the battery cell polarization curve, and calculating voltage loss corresponding to the curve slope as the ohmic polarization loss voltage; In a low current density initial region of the battery monomer polarization curve, performing nonlinear fitting on an activation polarization process by utilizing a Tafel equation, and calculating overpotential obtained by fitting as the activation polarization loss voltage; Subtracting the ohmic polarization loss voltage and the activation polarization loss voltage in sequence from the total voltage loss of the cell polarization curve, and calculating the remaining voltage loss as the concentration polarization loss voltage; and recording the ohmic polarization loss voltage, the active polarization loss voltage and the concentration polarization loss voltage corresponding to different operation moments, and arranging according to a time sequence to form the voltage loss component set.
3. The fuel cell health estimation method according to claim 2, wherein said performing correlation matching calculation with the ac impedance spectrum inside the stack using the voltage loss component set, establishes a polarization loss-impedance characteristic mapping relationship, comprises: Performing relaxation time distribution analysis on the AC impedance spectrum in the electric pile to separate out characteristic impedance arcs corresponding to a charge transfer process, a proton conduction process and a mass transfer process; Extracting characteristic frequency and characteristic resistance value from the characteristic impedance arc, and carrying out linear regression analysis on the characteristic resistance value of the high-frequency region corresponding to the charge transfer process and the activated polarization loss voltage to obtain a first mapping coefficient; Carrying out linear regression analysis on the characteristic resistance value of the intermediate frequency region corresponding to the proton conduction process and the ohmic polarization loss voltage to obtain a second mapping coefficient; carrying out trend correlation analysis on the characteristic impedance characteristic of the low-frequency region corresponding to the quality transmission process and the concentration polarization loss voltage to obtain a third mapping coefficient; And integrating the first mapping coefficient, the second mapping coefficient and the third mapping coefficient to construct the polarization loss-impedance characteristic mapping relation.
4. A fuel cell state of health estimation method according to claim 3, wherein said extracting a pressure fluctuation pattern synchronized with the cell polarization curve acquisition time from the gas flow channel pressure fluctuation time sequence, and performing a correlation analysis of the pressure fluctuation pattern and the concentration polarization loss voltage, and identifying an additional concentration loss component due to gas diffusion barrier, comprises: Locking a time segment which is completely overlapped with the battery cell polarization curve data acquisition window in the gas flow channel pressure pulsation time sequence; performing fast Fourier transform on the pressure pulsation data of the time segment, and calculating the energy duty ratio of the pressure pulsation data in a low frequency band; Comparing the low-frequency energy duty ratio with a reference pressure pulsation spectrum, and judging that a pressure pulsation mode related to gas diffusion barrier exists in the time segment when the low-frequency energy duty ratio exceeds a set threshold; Subtracting the concentration polarization loss voltage of the time segment from the concentration polarization loss voltage in the reference state without diffusion barrier, and quantifying the difference value as the additional concentration loss component.
5. The fuel cell health estimation method of claim 4, wherein said performing a fast fourier transform on said time slice pressure pulsation data to calculate the energy duty cycle in the low frequency band comprises: Performing zero-mean preprocessing on the pressure pulsation data of the time slices to obtain a centralized pressure pulsation signal; Applying a hanning window function to the centralized pressure pulsation signal, and then performing a fast fourier transform operation to obtain a pressure pulsation frequency spectrum; Defining a frequency band lower than a set cut-off frequency as a low frequency band in the pressure pulsation frequency spectrum, and calculating the sum of the squares of all frequency component amplitudes in the low frequency band to obtain low frequency band energy; Calculating the sum of the squares of all frequency component amplitudes of the pressure pulsation spectrum in the full frequency band to obtain full frequency band energy; Dividing the low-frequency band energy by the full-frequency band energy to obtain the low-frequency band energy duty ratio.
6. The fuel cell state of health estimation method according to claim 5, wherein said integrating the additional concentration loss component, the polarization loss-impedance feature mapping relationship, and a preset reference aging model, calculates a comprehensive state of health index of the target fuel cell, comprising: Reading the reference polarization loss component and the reference impedance characteristic of the fuel cell under the standard working condition and the brand new state from the reference aging model; inputting the current acquired AC impedance spectrum in the pile into the polarization loss-impedance characteristic mapping relation, and calculating to obtain a mapping polarization loss component in the current state; subtracting the corresponding value of the mapped polarization loss component from the current actual voltage loss component set to obtain a polarization loss residual error which cannot be mapped directly by impedance; adding the extra concentration loss component to a part belonging to concentration polarization in the polarization loss residual to obtain a total aging loss component related to gas diffusion; And comparing the mapped polarization loss component, the total aging loss component and the corresponding components in the reference aging model, and calculating the comprehensive health state index through a weighted fusion algorithm.
7. The fuel cell health estimation method according to claim 6, wherein subtracting the values corresponding to the mapped polarization loss components from the current actual voltage loss component set, respectively, to obtain polarization loss residuals that cannot be directly mapped by impedance, comprises: Reading the actual value of the ohmic polarization loss voltage, the actual value of the active polarization loss voltage and the actual value of the concentration polarization loss voltage from the current actual voltage loss component set; reading a mapped ohmic polarization loss voltage, a mapped activation polarization loss voltage, and a mapped concentration polarization loss voltage from the mapped polarization loss component obtained by impedance spectroscopy mapping; Subtracting the mapped ohmic polarization loss voltage from the actual value of the ohmic polarization loss voltage to obtain an ohmic polarization loss residual error; subtracting the mapped active polarization loss voltage from the actual value of the active polarization loss voltage to obtain an active polarization loss residual error; subtracting the mapped concentration polarization loss voltage from the actual value of the concentration polarization loss voltage to obtain a concentration polarization loss residual error; and combining the ohmic polarization loss residual error, the active polarization loss residual error and the concentration polarization loss residual error to form the polarization loss residual error.
8. The fuel cell state of health estimation method of claim 7, further comprising the step of extrapolating a predicted trend of aging based on said integrated state of health index and historical operating data: Acquiring the comprehensive health state indexes of the target fuel cell at a plurality of historical time points to form a health state time sequence; Smoothing filtering is carried out on the health state time sequence, short-term fluctuation noise is eliminated, and a smooth health state sequence reflecting long-term change trend is obtained; Performing curve fitting on the smooth health state sequence to obtain an aging trend function describing the time evolution of the health state; extrapolating and predicting a predicted value of the state of health of the target fuel cell at a specified future time point by using the aging trend function; And comparing the health state predicted value with a preset failure threshold value, and judging that the target fuel cell possibly reaches the service life end at the future appointed time point when the health state predicted value is lower than the failure threshold value.
9. The fuel cell state of health estimation method according to claim 8, wherein said smoothing the state of health time series to eliminate short-term fluctuation noise and obtain a smoothed state of health series reflecting a long-term trend comprises: processing the health state time sequence by adopting a sliding average algorithm, and setting the length of a sliding window, wherein the length of the sliding window is larger than the typical period of short-term fluctuation of the health state; Sequentially calculating arithmetic average values of data points covered by the sliding window from the starting point of the health state time sequence, and taking the arithmetic average values as smoothed health state values corresponding to window center points; moving the sliding window backwards by one data point, and repeating the step of calculating the arithmetic average value until the whole health state time sequence is processed; and arranging the smoothed health state values corresponding to all window center points according to the original time sequence to form the smoothed health state sequence.
10. A fuel cell state of health estimation system comprising a master processor and a co-data memory electrically coupled to the master processor and configured to store program instructions and a reference database, and wherein the master processor is configured to read and execute the program instructions in the co-data memory to implement the overall process of a fuel cell state of health estimation method according to any one of claims 1 to 9.

Description

Method and system for estimating state of health based on fuel cell Technical Field The present invention relates to the field of fuel cell state estimation technologies, and in particular, to a method and a system for estimating a state of health of a fuel cell. Background The current fuel cell health state estimation related technologies mostly rely on single polarization curve data or alternating current impedance spectrum data for analysis, part of the technologies can simply divide polarization types of the polarization curve, only basic polarization loss values are obtained, and pressure pulsation time sequence data of a gas flow channel are mostly used for monitoring gas transmission states independently and are not combined with electrochemical data for application. When the aging state of the fuel cell is analyzed by the conventional technology, only the integral polarization loss result can be obtained, different polarization loss components cannot be correspondingly associated with impedance spectrum characteristics, concentration polarization loss is not subdivided, pressure data and the acquisition time sequence of a polarization curve are not synchronously matched, and different generation causes of concentration polarization loss cannot be distinguished. In the operation process of the fuel cell, ohmic polarization, activation polarization and concentration polarization respectively correspond to different aging mechanisms, a single data dimension cannot accurately correspond to various aging characteristics, impedance spectrum data cannot be directly converted into polarization loss components, and analysis deviation of the aging mechanisms can be caused. The gas diffusion layer is blocked, the gas flow passage is not smooth to transmit and can generate additional concentration polarization loss, the conventional technology cannot identify the special loss component, the health state evaluation cannot be combined with the multiple aging effects of membrane electrode catalytic activity attenuation, proton exchange membrane aging and gas diffusion layer blocking, the corresponding correlation between the polarization loss component and the impedance spectrum is required to be established, the pressure fluctuation mode of the synchronous time sequence is extracted, the concentration polarization loss is correlated, and the additional concentration loss component caused by gas diffusion disorder is screened. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides a fuel cell health state estimation method and system. In order to achieve the above purpose, the invention adopts the following technical scheme that the fuel cell based state of health estimation method comprises the following steps: Continuously acquiring a dynamic electrochemical data set of a target fuel cell in a plurality of complete operation cycles, wherein the dynamic electrochemical data set comprises a cell polarization curve, an AC impedance spectrum in a galvanic pile and a gas flow channel pressure pulsation time sequence; Carrying out component decomposition treatment of ohmic polarization, activation polarization and concentration polarization on the battery monomer polarization curve, and extracting a voltage loss component set representing different aging mechanisms, wherein the voltage loss component set comprises ohmic polarization loss voltage, activation polarization loss voltage and concentration polarization loss voltage; Performing correlation matching calculation by using the voltage loss component set and the AC impedance spectrum in the pile, and establishing a polarization loss-impedance characteristic mapping relation which is used for calculating a corresponding polarization loss component from the impedance spectrum; extracting a pressure fluctuation mode synchronous with the acquisition time of the cell polarization curve from the gas flow channel pressure fluctuation time sequence, performing correlation analysis on the pressure fluctuation mode and the concentration polarization loss voltage, and identifying an additional concentration loss component caused by gas diffusion barriers; And (3) integrating the additional concentration difference loss component, the polarization loss-impedance characteristic mapping relation and a preset reference aging model, and calculating the comprehensive health state index of the target fuel cell, wherein the comprehensive health state index reflects the superposition effects of membrane electrode catalytic activity attenuation, proton exchange membrane dry-wet aging and gas diffusion layer blocking state. As a further aspect of the present invention, the performing component decomposition processing of ohmic polarization, active polarization and concentration polarization on the polarization curve of the battery cell to extract a set of voltage loss components characterizing different aging mechanisms includes: Obtaining a cur