Impedance optimization method for porous gradient anode supported solid fuel cells

Abstract

The present invention relates to the field of solid fuel cell technology, and discloses an impedance optimization method for a porous gradient anode-supported solid fuel cell, which is applied to a system comprising a plurality of impedance sensors and an optimization controller for a porous gradient anode. The method comprises: the sensor collects voltage and current data to generate an impedance state data set; the optimization controller constructs a regional impedance state matrix, extracts features after processing, and predicts the impedance distribution in combination with the potential gradient, etc., outputs the spatiotemporal distribution features through an impedance optimization model and updates the data set; and dynamically optimizes regional electrode parameters based on the distribution features, including operations such as voltage adjustment, current aggregation, and resistance distribution. This method achieves accurate prediction of the spatiotemporal distribution of electrode impedance and dynamic parameter control through multi-sensor distributed acquisition, multi-dimensional feature analysis, and dynamic modeling, thereby improving the performance and stability of solid fuel cells and being applicable to the field of solid fuel cell management.

Description

Technical Field

[0001] The present invention relates to the technical field of solid fuel cells, and in particular to an impedance optimization method for porous gradient anode supported solid fuel cells. Background Art

[0002] Solid fuel cells, due to their high efficiency, environmental friendliness, and strong fuel adaptability, have shown broad application prospects in fields such as distributed power generation and transportation. As an important type of SOFC, porous gradient anode-supported solid fuel cells (SOFCs) utilize a porous gradient structure in the anode layer to effectively improve fuel diffusion efficiency and electrode reaction activity. However, in actual operation, the uneven distribution of electrode impedance significantly affects battery performance and lifespan.

[0003] Existing solid fuel cell management systems face numerous challenges in impedance optimization. For one thing, traditional single-point measurement methods struggle to capture the spatiotemporal distribution of electrode impedance and cannot accurately reflect the complex impedance variations within porous gradient anodes. The porous gradient structure of the anode leads to differences in porosity, pore size distribution, and material composition across different regions, resulting in uneven distribution of fuel gas diffusion paths and electrochemical reaction sites. This, in turn, causes a spatially non-uniform impedance distribution. Single-point measurement only captures impedance information for a localized region and cannot fully reflect the impedance state of the entire anode. Furthermore, existing optimization algorithms are mostly based on static models and lack the ability to track impedance evolution in real time under dynamic operating conditions. Under varying load, temperature, and fuel composition, parameters such as electrode reaction rate and gas diffusion coefficient undergo dynamic changes, causing impedance characteristics to evolve over time. Traditional static models are unable to adapt to these changes, making it difficult to dynamically optimize electrode parameters. Furthermore, issues such as mutual interference between electrodes, transmission delays, and impedance blind spots further exacerbate the complexity of impedance optimization. In porous gradient anodes, interactions such as electromagnetic coupling exist between electrodes in different regions, which can interfere with impedance measurement results. Furthermore, the transmission of signals in the electrodes requires a certain amount of time, and there is a transmission delay, which may result in a time deviation between the measured data and the actual impedance state. Furthermore, the existence of impedance blind spots (such as areas where it is difficult to accurately obtain impedance data due to structural or measurement limitations) makes it impossible to fully grasp the impedance distribution of the entire anode, affecting the formulation of optimization strategies.

[0004] In existing technologies, some studies have attempted to optimize impedance management by increasing the number of sensors or improving algorithms, but the following deficiencies remain: Sensor network deployment lacks systematicity and fails to fully consider the structural characteristics of porous gradient anodes, resulting in insufficient spatial coverage and accuracy of data acquisition; at the algorithmic level, the ability to fuse and process multi-source data is limited, making it impossible to effectively extract key information such as potential intensity distribution and transmission delay characteristics. Furthermore, the inherent relationship between electrode structure and impedance distribution is not fully considered during the modeling process, resulting in poor accuracy and generalization of the optimization model. Therefore, how to construct a high-precision, real-time impedance optimization method based on the structural characteristics of porous gradient anodes to accurately predict and dynamically control the spatiotemporal distribution of electrode impedance has become a key technical issue that needs to be urgently addressed in the field of solid fuel cells. Summary of the Invention

[0005] The object of the present invention is to provide a method for optimizing the impedance of a porous gradient anode-supported solid fuel cell to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the impedance of a porous gradient anode-supported solid fuel cell, applied to a solid fuel cell management system, the system comprising a plurality of impedance sensors deployed on a porous gradient anode and an optimization controller connected to the impedance sensors, with two adjacent impedance sensors being separated by a set distance. The method comprises:

[0007] The impedance sensor collects voltage data and current data of electrodes in the region to generate an impedance state data set;

[0008] The optimization controller receives real-time impedance state data of multiple impedance sensors and constructs a regional impedance state matrix;

[0009] Determining distribution characteristics of electrode impedance based on the impedance state data set and real-time data from each impedance sensor, wherein determining the distribution characteristics of electrode impedance includes: processing an impedance state matrix, extracting impedance characteristics based on the impedance state data set, predicting impedance distribution based on potential gradient and electrolyte path information, outputting spatiotemporal distribution characteristics of electrode impedance through an impedance optimization model, and updating the impedance state data set based on the spatiotemporal distribution characteristics;

[0010] According to the distribution characteristics, regional electrode parameters are dynamically optimized.

[0011] Preferably, the determining the distribution characteristics of the electrode impedance includes:

[0012] Processing the impedance state matrix to extract potential intensity distribution, transmission delay characteristics and impedance change trends;

[0013] Performing electric potential intensity modeling on the impedance state matrix according to the electric potential intensity distribution and transmission delay characteristics, dividing the regional electrode into a plurality of sub-regions and marking the regional identifiers, performing correlation matching with the impedance state data set according to the electric potential intensity of the sub-regions, and marking the regional identifiers in the impedance state data set;

[0014] Calculating the potential intensity gradient according to the position of the impedance sensor, predicting the distribution of the electrode impedance according to the potential intensity gradient and the impedance change trend, and calculating the impedance prediction information of each sub-region;

[0015] Constructing an impedance optimization model, using the impedance prediction information as an input parameter of the impedance optimization model, performing spatial correlation modeling on the impedance prediction information through the impedance optimization model, and outputting spatiotemporal distribution characteristics of electrode impedance;

[0016] The impedance state data set is updated according to the spatiotemporal distribution characteristics to obtain the distribution characteristics of the electrode impedance.

[0017] Preferably, the processing of the impedance state matrix includes:

[0018] The impedance state matrix is ​​normalized, the impedance hotspot area in the matrix is ​​intercepted by a sliding window, the hotspot area is noise filtered, and the impedance change trend is calculated by an eigendecomposition algorithm;

[0019] Calculating the spatial correlation characteristics of the impedance state matrix, calculating the interference intensity, link stability coefficient and impedance blind zone index between regions based on the spatial correlation characteristics, constructing a feature fusion network, and calculating the transmission delay characteristics through the feature fusion network;

[0020] The time domain features and frequency domain features collected by each impedance sensor are extracted, and the impedance characteristic vector of the sensor is calculated based on the phase difference between the time domain features and the frequency domain features. The impedance sensors at different positions are feature matched based on the impedance characteristic vector, and the impedance change trend is calculated.

[0021] Preferably, the performing potential intensity modeling on the impedance state matrix includes:

[0022] Extracting electric potential intensity sampling points from each frame of data based on the electric potential intensity distribution, performing correlation mapping between the sampling points and transmission delay characteristics to generate an electric potential intensity map, spatially aligning the electric potential intensity maps collected by multiple sensors, and calculating the electric potential intensity distribution of the region;

[0023] Set the interference threshold, locate the interference source based on the potential strength value of the multi-frame impedance state matrix, and calculate the interference strength difference. If the interference strength difference is greater than or equal to the interference threshold, it indicates that there is an impedance blind spot in the area. Perform propagation model constraint compensation on the current area, iteratively correct the potential strength distribution of the current area based on the path loss model corresponding to the current area, and calculate the blind spot potential strength compensation value based on the correction result.

[0024] The potential intensity modeling is performed on the impedance state matrix according to the potential intensity distribution, and the time delay is marked on the potential model of the region through the transmission delay characteristics.

[0025] Preferably, the calculating of the electric potential intensity gradient according to the position of the impedance sensor comprises:

[0026] Based on multiple sets of impedance coverage data, the potential intensity change points are extracted and mapped to a unified electrode coordinate system based on the deployment location of the sensor. The change points are fitted using a spatial interpolation algorithm to generate a regional potential field model.

[0027] Performing equal-interval sampling along the transmission path of the electric potential field model, calculating the voltage attenuation rate, interference fluctuation index, and potential change slope of the path based on the sampling results, and calculating the potential change parameter based on the voltage attenuation rate, interference fluctuation index, and potential change slope;

[0028] Based on the deployment parameters and acquisition accuracy of the impedance sensors, the distribution characteristics of the impedance change in each frame of data are projected onto the electric potential field model. The electric potential field model is partitioned along the transmission direction according to the number of sensors. The variation pattern of the impedance change within the partition is analyzed, and the impedance distribution characteristics are calculated based on the variation pattern.

[0029] The electric potential intensity gradient is calculated based on the electric potential change parameter and the impedance distribution characteristic. The calculation process of the electric potential intensity gradient includes: based on the electrode position range from the first impedance sensor to the last impedance sensor, selecting spatial coordinate points in the sensor deployment direction, cumulatively calculating the product of the electric potential field strength characteristic weight value and the impedance distribution characteristic weight value within the spatial resolution range, and superimposing the influence value of the sensor acquisition frequency on the potential intensity change rate.

[0030] Preferably, the calculation of the impedance prediction information of each sub-region includes:

[0031] Taking the main transmission path of the electric potential field model as the baseline and the peak position of the impedance change in each frame of data as the reference point, the impedance offset is calculated and the impedance distribution curve is drawn according to the electrode coordinates;

[0032] Correcting the rate and direction of change in the impedance change trend according to the potential intensity gradient;

[0033] Starting from the nearest impedance distribution point, the distribution curve is continuously drawn according to the correction results of the change rate and direction to generate the impedance distribution points of the next period until the distribution points cover the entire target area and generate impedance prediction information.

[0034] Preferably, the constructing of the impedance optimization model includes:

[0035] The input layer is used to organize the impedance prediction information into spatial distribution data and perform normalization processing;

[0036] The feature fusion layer is used to extract the regional correlation features of impedance by processing spatial distribution data and construct the dependency relationship between electrode units;

[0037] The parameter optimization layer is used to integrate the correlation between electrode impedances in spatial units and generate electrode parameter optimization strategies.

[0038] Preferably, obtaining the distribution characteristics of the electrode impedance includes:

[0039] According to the spatiotemporal distribution characteristics of the electrode impedance output by the impedance optimization model, the sub-region identifiers are matched with the spatiotemporal distribution characteristics;

[0040] The regional data in the impedance state dataset are reorganized according to the spatiotemporal characteristics to generate a regional distribution map sorted by impedance intensity;

[0041] According to the reorganized regional distribution map, the optimized electrode impedance distribution characteristics are output.

[0042] Preferably, the dynamic optimization of regional electrode parameters includes:

[0043] Mapping the regional identifiers to the regions of the distribution characteristics of the electrode impedance one by one;

[0044] According to the spatiotemporal distribution characteristics of electrode impedance, the execution actions of electrode parameters are controlled, including voltage adjustment, current aggregation and resistance distribution operations;

[0045] Based on the spatial distribution of electrode impedance and the preset parameter optimization strategy, the electrode parameters are dynamically allocated to the corresponding electrode areas.

[0046] Preferably, the execution action of controlling the electrode parameters includes:

[0047] When the electrode impedance reaches a preset intensity threshold in the target area, a voltage increase instruction is triggered for the adjacent electrodes;

[0048] Dynamically combine available circuit resources according to current aggregation strategy to generate resistance allocation vectors;

[0049] Electrode parameters of transmission nodes in the target area are adjusted based on the resistance allocation vector.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] By deploying multiple impedance sensors on the porous gradient anode, the voltage and current data of the electrodes within the region are collected at set intervals to generate an impedance state dataset. Compared to traditional single-point measurement, this distributed sensor network coverage model can comprehensively capture the impedance information of different regions of the electrode, effectively solving the problem of local measurement deviation caused by uneven impedance distribution in the porous gradient structure, significantly improving the spatial resolution and integrity of data collection, and laying a solid data foundation for subsequent accurate analysis of impedance distribution characteristics.

[0052] The optimization controller receives real-time impedance state data from multiple impedance sensors and constructs a regional impedance state matrix. Through matrix standardization, noise filtering, and feature decomposition, it extracts key information such as potential intensity distribution, transmission delay characteristics, and impedance variation trends. The calculation of spatial correlation features and the construction of a feature fusion network enable in-depth analysis of inter-regional interference intensity, link stability coefficients, and impedance blind zone indexes. This effectively overcomes the impact of complex factors such as mutual interference between electrodes, transmission delays, and impedance blind zones on impedance analysis, ensuring the accuracy and reliability of the extracted features.

[0053] In the process of determining the electrode impedance distribution characteristics, the regional electrodes are divided into multiple sub-regions through potential intensity modeling and identified and associated. Combined with the calculation of potential intensity gradient and the generation of impedance prediction information, the dynamic prediction of the spatiotemporal distribution of electrode impedance is achieved. Spatial correlation modeling based on the impedance optimization model can accurately capture the dependencies between electrode units and output distribution characteristics that reflect the spatiotemporal evolution of electrode impedance, providing a scientific basis for the dynamic optimization of electrode parameters. This complete process from data acquisition, feature extraction to modeling and prediction forms a closed-loop analysis system for impedance distribution characteristics, significantly improving the depth of understanding and prediction accuracy of the impedance distribution of porous gradient anodes.

[0054] The regional electrode parameters are dynamically optimized according to the impedance distribution characteristics, and the spatial precision of the electrode parameter adjustment is achieved by mapping the regional identification with the impedance distribution characteristics. The electrode parameters are controlled to perform operations such as voltage adjustment, current aggregation and resistance distribution, and the working parameters can be adjusted in real time according to the impedance state of different regions. For example, when the electrode impedance in the target area reaches the preset intensity threshold, the voltage increase instruction of the adjacent electrode is triggered, and the resistance distribution vector is generated by dynamically combining the available circuit resources and adjusting the transmission node parameters, thereby achieving rapid response and local regulation of the impedance abnormality area. This dynamic optimization strategy based on spatiotemporal distribution characteristics breaks the limitations of traditional static parameter adjustment, enables dynamic matching of the electrode working state and the impedance distribution, and effectively improves the overall performance and operational stability of the solid fuel cell.

[0055] By updating the impedance state dataset in real time, a closed-loop optimization mechanism of "data acquisition - modeling analysis - parameter optimization - data feedback" has been formed. This mechanism continuously tracks the dynamic evolution of electrode impedance, continuously optimizes model parameters and adjusts electrode operating conditions, ensuring the system is always in optimal operating condition. This significantly extends the service life of solid fuel cells, reduces maintenance costs, and provides a strong guarantee for their long-term stable operation in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a working principle diagram of the impedance optimization method for porous gradient anode supported solid fuel cells according to the present invention;

[0057] Figure 2 Schematic diagram of the working principle for determining the electrode impedance distribution characteristics;

[0058] Figure 3 This is a diagram showing the working principle of the electric potential intensity gradient calculation. DETAILED DESCRIPTION

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0060] See also Figure 1-Figure 3 The present invention relates to an impedance optimization method for a porous gradient anode-supported solid fuel cell, which is applied to a solid fuel cell management system. The system includes multiple impedance sensors deployed on the porous gradient anode and an optimization controller connected to the impedance sensors, with adjacent impedance sensors spaced a set distance apart. The specific implementation scheme is as follows:

[0061] Step 1: Multiple impedance sensors deployed on the porous gradient anode collect real-time voltage and current data from the electrodes in each area. Each impedance sensor periodically samples the electrical parameters of the electrode at its location at a preset sampling frequency. The collected voltage and current data are timestamped and packaged according to a preset data format to generate an impedance state dataset containing information such as the sensor location, sampling time, voltage, and current values. This dataset serves as the basis for subsequent analysis and processing, reflecting the impedance state of the electrode at different times and locations.

[0062] Step 2: The optimization controller receives real-time impedance state data from multiple impedance sensors via a communication link. Based on the deployment location and number of each impedance sensor, the received data is arranged in spatial order to construct a regional impedance state matrix. The rows and columns of this matrix correspond to the spatial coordinate directions of the electrode region, and each element in the matrix represents the impedance state data collected by the sensor at the corresponding location. This intuitive matrix presents the spatial distribution of the electrode impedance, providing a structured data foundation for subsequent analysis of the impedance distribution characteristics.

[0063] Step 3: Process the impedance state matrix, extract the impedance features based on the impedance state data set, and predict the impedance distribution based on the potential gradient and electrolyte path information. Output the spatiotemporal distribution characteristics of the electrode impedance through the impedance optimization model, and update the impedance state data set based on the spatiotemporal distribution characteristics. The specific process is as follows:

[0064] The impedance state matrix is ​​processed to extract the potential intensity distribution, transmission delay characteristics and impedance change trend; the impedance state matrix is ​​modeled for potential intensity based on the potential intensity distribution and transmission delay characteristics, the regional electrode is divided into multiple sub-regions and the regional identifiers are marked, the potential intensity of the sub-region is correlated and matched with the impedance state dataset, and the regional identifiers are marked in the impedance state dataset; the potential intensity gradient is calculated according to the position of the impedance sensor, the distribution of the electrode impedance is predicted based on the potential intensity gradient and the impedance change trend, and the impedance prediction information of each sub-region is calculated; an impedance optimization model is constructed, the impedance prediction information is used as an input parameter, the impedance prediction information is spatially correlated modeled through the impedance optimization model, and the spatiotemporal distribution characteristics of the electrode impedance are output; the impedance state dataset is updated according to the spatiotemporal distribution characteristics to obtain the distribution characteristics of the electrode impedance.

[0065] Step 4: Dynamically adjust regional electrode parameters based on the electrode impedance distribution characteristics determined in step 3. Specifically, the distribution characteristics of the electrode impedance are associated with the physical location of the electrode region, and the causes of impedance anomalies in different regions are analyzed. Based on the impedance conditions of each region, the optimization controller generates corresponding parameter adjustment instructions, and dynamically optimizes the electrode voltage, current, resistance and other parameters to reduce the unevenness of the electrode impedance and improve the overall performance and stability of the solid fuel cell.

[0066] The present invention will be further described below in conjunction with Examples 1 to 5:

[0067] Example 1:

[0068] In this embodiment, the process of determining the electrode impedance distribution characteristics requires detailed operations from the impedance state matrix processing, potential intensity modeling, and related feature extraction. First, when processing the impedance state matrix, a normalization step must be performed. That is, a data normalization algorithm is used to convert the voltage and current data of different dimensions in the matrix into values ​​of a unified dimension, making the data at each location comparable. For example, a minimum-maximum normalization method can be used to map the value of each data point to the interval [0,1], eliminating the impact of differences in sensor acquisition accuracy and different physical quantity units on subsequent analysis.

[0069] After standardization, a sliding window technique is used to identify impedance hotspots in the matrix. The size and step size of the sliding window can be set based on the sensor deployment density and electrode area characteristics. For example, the window can be set to cover a 5×5 sensor area with a step size of 1 sensor spacing. The matrix is ​​scanned row by row and column by column to identify hotspots where the impedance value is significantly higher or lower than the surrounding area. For these hotspots, median or Gaussian filtering algorithms are used to filter out noise, removing anomalous data points caused by factors such as electromagnetic interference and transient sensor failures, while retaining valid data that reflects the true impedance state.

[0070] Eigendecomposition algorithms (such as principal component analysis (PCA)) are used to process the filtered matrix data and calculate impedance trends. Eigendecomposition maps high-dimensional impedance state data into a low-dimensional feature space, extracts the primary eigenvectors, and analyzes their temporal variations to determine whether the impedance is rising, falling, or fluctuating within the overall region. For example, by observing the principal component scores over time, the rate and periodicity of impedance change can be determined.

[0071] When calculating the spatial correlation characteristics of the impedance state matrix, spatial autocorrelation analysis methods (such as the Moran index) can be used to measure the correlation between sensor data at different locations. Based on the calculation results, the interference intensity between regions is determined, that is, the degree to which impedance changes in adjacent regions influence each other. The link stability coefficient is calculated to assess the reliability of the data transmission link between sensors. The impedance blind zone index is calculated to identify areas where data may be missing or where impedance is abnormally stable. When constructing a feature fusion network, a fully connected neural network structure can be used. Parameters such as interference intensity, link stability coefficient, and impedance blind zone index are used as inputs. Through nonlinear transformations of multiple layers of neurons, the transmission delay characteristics are output, that is, the time delay value of signal transmission between different regions.

[0072] When extracting the time and frequency domain features of the data collected by each impedance sensor, the time domain features can include voltage mean, current variance, and rise time. Frequency domain features can be obtained through a fast Fourier transform (FFT) to obtain the amplitude and phase of each frequency component. Based on the phase difference between the time and frequency domain features, the sensor's impedance characteristic vector is calculated. This vector contains the amplitude and phase information of the impedance at the sensor's location. Using algorithms such as cosine similarity, the impedance characteristic vectors of sensors at different locations are matched to analyze the consistency of impedance change trends. For example, it can be determined whether the impedance changes of adjacent sensors are synchronized or there is a phase difference.

[0073] When modeling the potential intensity of the impedance state matrix, the potential intensity sampling points are first extracted from each frame of data based on the potential intensity distribution. The sampling points can be selected based on the sensor location or evenly distributed within the electrode area through interpolation. The sampling points are associated with the transmission delay characteristics and mapped. That is, for each sampling point, its potential intensity value and the corresponding signal transmission delay time are recorded to generate a potential intensity map. The potential intensity maps collected by multiple sensors are spatially aligned through spatial transformation algorithms (such as translation, rotation, and scaling) so that the maps of different sensors overlap in a unified electrode coordinate system, thereby accurately calculating the potential intensity distribution of the region, for example, determining the location and range of high potential and low potential areas.

[0074] Set the interference threshold value (such as 2 times the average interference intensity based on historical data statistics) and analyze the potential intensity value of the multi-frame impedance state matrix. Locate the possible interference source by comparing the potential intensity difference between adjacent frames or adjacent areas. Calculate the interference intensity difference. If the difference is greater than or equal to the interference threshold value, it indicates that there may be an impedance blind area in the area, that is, an area where impedance data cannot be accurately obtained due to interference. At this time, perform propagation model constraint compensation on the current area, and iteratively correct the potential intensity distribution according to the path loss model corresponding to the area (such as the free space propagation model, the logarithmic distance attenuation model). For example, assuming the path loss model is (in is the reference distance loss, is the path loss exponent, For distance, is Gaussian noise), and the model parameters are iteratively adjusted to make the corrected potential intensity value consistent with the actual measured value, and then the intensity compensation value of the blind area potential is calculated to correct the potential data deviation caused by the blind area.

[0075] Finally, the impedance state matrix is ​​modeled based on the corrected potential intensity distribution, and the corresponding potential intensity value is annotated at each element in the matrix. At the same time, the regional potential model is annotated with time delays using the transmission delay feature. This means that the time delay information of each regional potential change is recorded in the potential intensity map or matrix, such as the time difference between the potential change in a certain region and the reference point, so that the influence of time factors can be considered in the subsequent analysis of the impedance distribution.

[0076] When processing the impedance state matrix, it's crucial to ensure logical coherence across all steps. For example, standardization provides a unified data foundation for subsequent feature extraction and modeling. Noise filtering prevents abnormal data from interfering with analysis results. Feature decomposition and spatial correlation analysis reveal patterns in impedance variation from different dimensions. Potential intensity modeling, through sampling point selection, map alignment, and interference compensation, accurately depicts the potential distribution across the electrode region, providing a reliable basis for subsequent calculations of potential intensity gradients and impedance predictions.

[0077] Example 2:

[0078] In this embodiment, calculating the potential intensity gradient is a key step in determining the characteristics of the electrode impedance distribution. It is necessary to start from multiple sets of impedance coverage data and implement it through a series of operations such as data processing, model construction, and parameter calculation. First, the multiple sets of impedance coverage data are preprocessed to identify and extract the potential intensity change points. These change points are locations where the potential intensity in the electrode area changes significantly, and can be determined by comparing the potential difference values ​​of adjacent data points. For example, a potential change threshold is set. When the potential difference value of two adjacent points exceeds the threshold, the position between the two points is marked as a potential intensity change point.

[0079] Based on the sensor's deployment location, the extracted potential intensity change points are mapped to a unified electrode coordinate system. This coordinate system, with a fixed point on the electrode as its origin, defines clear coordinate axis directions and unit lengths, ensuring that change points at different locations have a unified coordinate representation. This mapping allows the scattered change points to be integrated into a complete spatial framework, facilitating subsequent analysis and processing.

[0080] After the coordinates of the change points are mapped, a spatial interpolation algorithm is used to fit these points and generate an electric potential field model. This algorithm can use methods such as polynomial interpolation or spline interpolation to estimate the electric potential intensity at any location within the entire electrode region based on the known change point locations and potential intensity values. For example, polynomial interpolation fits a polynomial function so that its value at the known change points is equal to the actual potential intensity value, thereby obtaining the electric potential distribution for the entire region. The resulting electric potential field model represents the electric potential distribution within the electrode region as a continuous function.

[0081] Sampling is performed at equal intervals along the transmission path of the electric potential field model, acquiring data such as voltage, interference, and potential at each sampling point along the path. The sampling interval can be set based on the length of the transmission path and the required accuracy to ensure that the potential variation characteristics along the path are fully captured. For each sampling point, its position coordinates in the electrode coordinate system and the corresponding voltage, interference, and potential values ​​are recorded to form a set of sampled data.

[0082] Based on the sampling results, the voltage decay rate, interference fluctuation index, and potential change slope of the path are calculated. The voltage decay rate reflects the change in voltage with distance along the transmission path and is calculated by comparing the voltage differences between adjacent sampling points and incorporating distance information. The interference fluctuation index measures the degree of fluctuation of the interference signal along the path and can be determined by calculating statistics such as the standard deviation or variance of the interference values. The potential change slope indicates the rate of change of potential with distance and is calculated by taking the difference between the potential values ​​of adjacent sampling points. These three parameters are combined to calculate the potential change parameter, which describes the overall variation characteristics of the potential along the transmission path and provides an important basis for subsequent analysis of the potential distribution.

[0083] Based on the impedance sensor deployment parameters and acquisition accuracy, the distribution characteristics of the impedance change in each frame of data are projected onto the electric potential field model. Sensor deployment parameters include sensor spacing and location coordinates, while acquisition accuracy determines the resolution and accuracy of the data. By spatially aligning the distribution characteristics of the impedance change with the electric potential field model, the relationship between the impedance change and the electric potential distribution can be analyzed. For example, by observing the correspondence between high-impedance areas and the electric potential distribution, it can be determined whether areas with large potential gradients correspond to high-impedance areas.

[0084] Based on the number of sensors, the electrode area is divided into multiple zones along the transmission direction of the electric potential field model. The size of each zone can be adjusted based on the density of sensor distribution and the characteristics of the electrode area to ensure that the impedance changes within each zone have a certain degree of similarity and regularity. Analyze the impedance variation patterns within each zone. For example, calculate the distribution range of impedance values ​​and determine their maximum, minimum, and average values. Observe the impedance change trend over time to determine whether it is increasing, decreasing, or fluctuating.

[0085] Based on the impedance variation patterns within the subregion, the impedance distribution characteristics are calculated. These characteristics include statistical quantities such as the mean, variance, maximum, and minimum values ​​of the impedance, as well as characteristic parameters such as the skewness and peak state of the impedance distribution. Analysis of these characteristics provides a comprehensive understanding of the impedance distribution within the electrode region, providing data support for subsequent calculations of the potential intensity gradient.

[0086] When calculating the potential intensity gradient, multiple spatial coordinate points are selected along the sensor deployment direction based on the electrode position range from the first impedance sensor to the last impedance sensor. These coordinate points should cover the entire electrode area and be distributed as evenly as possible to ensure that the spatial variation of the potential intensity is accurately reflected.

[0087] For each selected spatial coordinate point, the product of the potential field strength characteristic weight and the impedance distribution characteristic weight within the spatial resolution range of that point is cumulatively calculated. The spatial resolution range can be set based on the required computational accuracy and actual conditions, for example, a small area centered on the point. The potential field strength characteristic weight reflects the influence of the potential field strength around the point on the overall gradient, while the impedance distribution characteristic weight indicates the contribution of the impedance distribution around the point to the gradient. The specific calculation of these two weights can be determined based on the analysis results of the potential field model and the impedance distribution characteristics, combined with the actual application scenario.

[0088] The effect of the sensor acquisition frequency on the rate of change of the electric potential intensity is also considered. A higher sensor acquisition frequency yields denser data, allowing for more accurate capture of rapid changes in the electric potential intensity. Conversely, a lower acquisition frequency may miss some details of rapidly changing events. Therefore, when calculating the electric potential intensity gradient, the effect of the sensor acquisition frequency on the rate of change of the electric potential intensity needs to be added to the previous calculation results. By comprehensively considering factors such as the electric potential field strength characteristics, the impedance distribution characteristics, and the sensor acquisition frequency, a potential intensity gradient that accurately reflects the spatial rate of change of the electric potential intensity is ultimately derived.

[0089] Throughout the computational process, each data processing step is closely linked, with the results of the previous step providing the foundation for the subsequent calculations. For example, the construction of the electric potential field model relies on the extraction and interpolation of potential intensity change points, while the calculation of impedance distribution characteristics is based on the electric potential field model and partition analysis. Sensor deployment parameters and acquisition accuracy also play a significant role in the entire process, affecting data quality and the accuracy of the calculation results.

[0090] Example 3:

[0091] In this embodiment, the impedance prediction information for each sub-region is calculated by combining multiple aspects of information, including the potential field model, potential intensity gradient, and impedance change trend. This is achieved through steps such as baseline setting, trend correction, and curve drawing. First, the main transmission path of the potential field model is used as the baseline. This main path is generally the main direction of potential transmission within the electrode region and can be determined by analyzing the distribution pattern of potential intensity in the potential field model. For example, the direction with the most significant potential intensity change or the longest transmission path is selected as the main path. In each frame of data, the peak position of the impedance change is identified, that is, the position where the impedance value reaches a local maximum or minimum, and it is used as the reference point. The impedance offset of the reference point relative to the baseline is calculated. This offset is determined by the difference between the horizontal coordinate of the reference point in the electrode coordinate system and the baseline coordinate, and is used to measure the degree of deviation of the impedance distribution from the main transmission path. For example, if the baseline is the x-axis, the horizontal coordinate of the reference point is x1, and the baseline coordinate is x0, then the impedance offset is |x1-x0|.

[0092] The impedance distribution curve is drawn by mapping the impedance value of each reference point to its positional information according to the coordinate order of the electrode coordinate system. The horizontal axis represents the spatial position of the electrode region, and the vertical axis represents the impedance value. A smooth curve connects the points to visually display the spatial distribution of the impedance, such as whether there is a single peak, multiple peaks, or a uniform distribution.

[0093] Based on the calculated potential intensity gradient, the rate of change and direction of the impedance change trend are corrected. The potential intensity gradient reflects the rate and direction of change of the potential in space. When the potential intensity gradient in a certain area is large, it means that the potential in the area changes dramatically, which may have a significant impact on the transmission of ions or electrons, thereby causing the impedance change rate to accelerate or change in direction. For example, in an area with a positive potential intensity gradient (the potential increases with increasing spatial position), if the original impedance change trend is rising, the rate of change may increase due to the potential driving effect; if the trend is falling, the rate of decline may slow down due to the potential hindering effect. During correction, the adjustment amount for the rate of change and direction is obtained by multiplying the potential intensity gradient value with the preset correction coefficient, and then the adjustment amount is superimposed on the original change trend parameter to form the corrected rate of change and direction parameters.

[0094] Starting from the nearest impedance distribution point, that is, selecting the impedance distribution point at the current moment or the most recent sampling as the starting point, and gradually predicting the impedance distribution point of the next period according to the corrected rate of change and direction. When predicting, according to the preset time interval (such as ) or spatial intervals (e.g. ), calculate the impedance value and position of the next point. For example, if the impedance value of the current point is , the location is The corrected rate of change is (Unit is or ), the direction is positive (indicating that the impedance increases with time or space), then the impedance value of the next point is: (or ), the location is: (or time is ).

[0095] When plotting the distribution curve, it is important to ensure that changes between adjacent points conform to the corrected trend to avoid sudden changes or unreasonable fluctuations. By predicting and plotting point by point, the distribution curve gradually covers the entire target area until all sub-areas are included. The resulting impedance prediction information contains the predicted impedance value and spatial location of each sub-area in the future time period. This information is stored in the form of a data table or graph, facilitating subsequent input into the impedance optimization model for processing.

[0096] In the entire calculation process, the accurate selection of the baseline is the key, and its rationality directly affects the calculation of the impedance offset and the shape of the distribution curve. The correction effect of the potential intensity gradient needs to be based on the theoretical analysis of the physical characteristics and transmission mechanism of the electrode to ensure that the correction logic conforms to the actual physical laws. The time interval or spatial interval setting in the prediction process needs to comprehensively consider the sensor acquisition frequency and the electrode reaction speed to avoid distortion of the prediction results due to excessive intervals or increased calculation complexity due to excessively small intervals. In addition, the drawing of the impedance distribution curve requires the use of a suitable smoothing algorithm, such as the moving average method or the spline smoothing method, to eliminate curve fluctuations caused by data noise and make the prediction results more credible.

[0097] Example 4:

[0098] In this example, constructing an impedance optimization model is the core step in analyzing the spatiotemporal distribution characteristics of electrode impedance and optimizing its parameters. This requires detailed explanation from the perspectives of model architecture design, functional implementation of each layer, and data processing flow. The model is divided into an input layer, a feature fusion layer, and a parameter optimization layer. Data transfer and algorithmic processing between these layers form a complete optimization logic chain.

[0099] The primary task of the input layer is to organize the calculated impedance prediction information into spatially distributed data. This information contains the impedance values ​​and corresponding spatial coordinates of each subregion in the future. The input layer must arrange these discrete point data in the order of the spatial coordinates of the electrode regions, forming a dataset with clear spatial location indexes. For example, the data can be stored in a two-dimensional matrix, where the rows and columns correspond to the horizontal and vertical coordinates of the electrode regions, respectively, and the matrix elements are the impedance prediction values ​​at the corresponding locations, so that the data presents a spatial distribution structure consistent with the physical layout of the electrodes.

[0100] The input layer performs normalization on the spatial distribution data. Since the impedance values ​​of different sub-regions may have different dimensions or numerical ranges, normalization can eliminate the impact of these differences on model training. In specific operations, a normalization method can be used to map each impedance value to a specific interval (such as [0,1] or [-1,1]). This is usually achieved by calculating the maximum, minimum, mean, and standard deviation of the data and using a linear transformation formula. For example, for a certain impedance value Z, the formula Calculate the normalized value ,in and are the maximum and minimum values ​​in the batch of data respectively. The normalized spatial distribution data is output as the input layer and passed to the feature fusion layer.

[0101] The core function of the feature fusion layer is to process spatially distributed data through algorithms, extract regional impedance correlation features, and construct dependencies between electrode units. This layer can employ a variety of neural network structures or signal processing algorithms, such as convolutional neural networks (CNNs), graph neural networks (GNNs), or self-attention mechanisms. Taking convolutional neural networks as an example, by designing convolution kernels of different sizes, convolution operations are performed on spatially distributed data to capture local regional correlation features at different scales. For example, a 3×3 convolution kernel can extract direct correlations between adjacent subregions, while a 5×5 convolution kernel can capture indirect influences over a larger area.

[0102] During feature extraction, the normalized spatial distribution data is first input to the convolutional layer. A sliding window operation is performed on the convolution kernel to calculate the characteristic response value at each location. This characteristic response value reflects the impedance distribution characteristics of that location and its neighborhood. Activation functions (such as ReLU) introduce nonlinear transformations to enhance the model's ability to represent complex correlated features. Through alternating layers of convolutional and pooling layers, features are extracted from low-level to high-level levels, such as the impedance value of a single subregion, gradient features of adjacent regions, and long-range dependencies across regions.

[0103] In addition to convolutional neural networks, graph neural networks are also an effective feature fusion method. Each electrode sub-region is regarded as a node in a graph structure, the node attribute is the impedance prediction value of the region, and the edges between nodes represent the spatial adjacency or dependency strength between sub-regions. Through graph convolution operations (GCN) or graph attention operations (GAT), nodes can aggregate the feature information of neighboring nodes to capture the mutual influence between regions. For example, the graph attention mechanism can dynamically adjust the contribution of different neighboring nodes to the current node feature update by calculating the attention weights between nodes, thereby constructing a more flexible electrode unit dependency model.

[0104] The output of the feature fusion layer is a high-dimensional feature vector containing regional correlation features. These feature vectors not only preserve the impedance information of each subregion but also encode the spatial dependencies between subregions, such as the synchronization of impedance changes in adjacent regions and the coupling effects of distant regions. These features provide in-depth spatial correlation information to the parameter optimization layer, enabling the model to analyze the impedance distribution problem from a global perspective.

[0105] The main function of the parameter optimization layer is to integrate the correlation between electrode impedances in spatial units and generate an electrode parameter optimization strategy. This layer is usually composed of models such as a fully connected neural network or a decision tree. It receives the high-dimensional feature vector output by the feature fusion layer and performs feature mapping through multiple layers of nonlinear transformations. First, the fully connected layer compresses the high-dimensional feature vector to a lower dimension and extracts the most critical feature components for parameter optimization. For example, by setting up multiple fully connected layers, the feature dimension can be gradually reduced from hundreds of dimensions to tens of dimensions. At the same time, through the learning of the weight matrix, the features related to parameter optimization are strengthened and redundant information is suppressed.

[0106] When integrating spatial correlations, the parameter optimization layer must consider the physical constraints and optimization objectives for adjusting electrode parameters. Physical constraints include voltage adjustment range, current carrying capacity, and resistance adjustment accuracy. The optimization objectives are typically to minimize overall impedance, balance impedance distribution, or increase fuel cell output power. For example, if the optimization objective is to minimize overall impedance, the model can measure the difference between the predicted impedance and the target impedance using a loss function (such as mean squared error). The model parameters are then updated using a backpropagation algorithm, allowing the resulting optimization strategy to effectively reduce impedance.

[0107] When generating an electrode parameter optimization strategy, the parameter optimization layer outputs parameter adjustment values ​​corresponding to each subregion, such as voltage adjustment, current aggregation coefficient, and resistance distribution ratio. These adjustment values ​​must correspond to the spatial location of the subregion to ensure that the optimization controller can accurately send parameter instructions to the target region. For example, for a high-impedance subregion, the model may generate a strategy that increases the voltage in adjacent regions and reduces the resistance of the region, thereby reducing its impedance through potential gradient adjustment and current path optimization.

[0108] During model training, historical impedance data and the corresponding parameter adjustment results are used as training samples. Through continuous iterative training, the model learns the optimal parameter adjustment strategy for different impedance distribution patterns. Care should be taken to avoid overfitting during training. Regularization techniques (such as L2 regularization), dropout layers, and cross-validation can be used to improve the model's generalization capabilities.

[0109] The entire impedance optimization model construction process must be closely integrated with the electrode physical properties and fuel cell operating mechanisms. For example, the regional correlation features extracted by the feature fusion layer must conform to the physical laws of ion transport, and the strategy generated by the parameter optimization layer must meet the electrical performance constraints of the electrode material. The model's input and output design should form a closed loop with the front-end data acquisition and back-end execution mechanisms to ensure a smooth transition from impedance data acquisition, feature analysis, to parameter optimization.

[0110] Through data standardization in the input layer, spatial correlation feature extraction in the feature fusion layer, and strategy generation in the parameter optimization layer, the impedance optimization model can efficiently process complex impedance distribution data and output optimization solutions that are physically meaningful and practically feasible.

[0111] Example 5:

[0112] In this embodiment, dynamic optimization of regional electrode parameters is achieved through regional identification mapping, action control, and parameter allocation. Each step is closely centered around the spatiotemporal distribution characteristics of electrode impedance, ensuring the targeted and effective optimization strategy. First, regional mapping is performed between regional identification and electrode impedance distribution characteristics. After determining the spatiotemporal distribution characteristics of electrode impedance, each subregion is assigned a unique regional identification (e.g., a number or coordinate interval) through the potential intensity modeling process. At this point, a one-to-one mapping relationship is established between these regional identifications and corresponding impedance distribution characteristics (e.g., impedance value range, variation trend, and spatial location), forming a regional-impedance characteristic comparison table. For example, subregion A1 corresponds to a high-impedance region, with impedance values ​​exceeding a preset threshold and showing an upward trend, located at the left edge of the electrode. Subregion B3 corresponds to a low-impedance region, with impedance values ​​stable within a normal range, located in the center of the electrode. This comparison table serves as the basis for subsequent parameter adjustments, enabling the optimization controller to quickly locate the regions requiring adjustment.

[0113] According to the spatiotemporal distribution characteristics of the electrode impedance, the execution actions of the electrode parameters are controlled, including voltage adjustment, current aggregation, and resistance distribution operations. When the electrode impedance of a target area is detected to reach a preset intensity threshold (such as higher than the upper limit of the normal operating range), the voltage increase instruction of the adjacent electrode is triggered. Specifically, by looking up the region-impedance characteristic comparison table, the position of the target area and the adjacent area identification are determined, and a voltage increase signal is sent to the electrode in the adjacent area. By increasing the potential of the adjacent area, a potential gradient difference is formed, and ions or electrons are guided to flow to the target area, thereby reducing the impedance of the target area. For example, if the target area is area A1 on the left edge and its adjacent right area is A2, the voltage of area A2 can be increased by 5% to form a potential gradient from A2 to A1, thereby promoting charge transfer.

[0114] Current aggregation dynamically combines available circuit resources based on a pre-defined current aggregation strategy. These resources include current channels, power modules, and load cells within each subregion. By analyzing the impedance state of each subregion, the current channels in each region can be activated or deactivated to achieve optimal current distribution. For example, in the low-impedance central region B3, due to its high charge transfer efficiency, more current channels can be activated to aggregate more current. In the high-impedance region A1, the current load can be temporarily reduced to avoid energy loss. When generating a resistance allocation vector, each subregion is assigned a corresponding resistance adjustment value. This vector is stored in a matrix or list format, with each element corresponding to the direction (increase or decrease) and magnitude (e.g., percentage) of the resistance adjustment for a subregion. For example, a resistance adjustment value of -10% for region A1 in the resistance allocation vector indicates that the resistance of that region needs to be reduced by 10% to lower the impedance; a resistance adjustment value of 0% for region B3 indicates that the current resistance state should be maintained.

[0115] Based on the resistance allocation vector, the electrode parameters of the transmission nodes in the target region are adjusted. Transmission nodes are key locations within the electrode region responsible for current transmission, and each node corresponds to one or more subregions. Resistance adjustment instructions are sent via a communication link to the actuator (such as an adjustable resistor) of the target transmission node. The actuator adjusts the node's resistance value according to the instruction, thereby changing the current transmission characteristics of that region. The adjustment process must follow a sequential order of low-priority regions followed by high-priority regions to avoid adjusting multiple critical regions simultaneously, which could cause system fluctuations.

[0116] Based on the spatial distribution of electrode impedance and the preset parameter optimization strategy, the electrode parameters are dynamically allocated to the corresponding electrode area. The parameter optimization strategy predefines the adjustment rules for different impedance distribution scenarios. For example, when a local high impedance area appears, the voltage and resistance of the area and its adjacent areas are adjusted first; when the overall impedance distribution is uniform but high, a global current aggregation strategy is adopted. According to the spatial distribution pattern of the current impedance, the corresponding optimization strategy is matched to generate a specific parameter allocation scheme. For example, if a strip-shaped high impedance area is detected on the right side of the electrode, the "local area voltage-resistance coordinated adjustment strategy" is matched to adjust the voltage and resistance of the strip area and its upper and lower adjacent areas in a coordinated manner. The voltage increase is 3%-5%, and the resistance reduction is 8%-12%. The specific values ​​are determined according to the degree of regional impedance deviation.

[0117] When dynamically allocating electrode parameters, the timing and coordination of parameter adjustments must be considered. For example, voltage and resistance adjustments must be performed sequentially, typically adjusting the voltage first to establish the potential gradient, followed by adjusting the resistance to optimize the current path, to avoid disrupting charge transfer caused by simultaneous adjustments. For cross-region parameter adjustments (e.g., simultaneous adjustments to multiple adjacent regions), synchronization signals are required to ensure that the execution actions of each region are consistent in time to prevent new impedance imbalances caused by differences in adjustment timing.

[0118] The entire dynamic optimization process requires real-time monitoring of impedance feedback data after electrode parameter adjustments. Impedance sensors collect real-time voltage and current data from each region, calculate the adjusted impedance value, and compare it with the optimization target. If the adjusted impedance value does not achieve the desired effect (for example, it remains above the threshold or the distribution remains uneven), a secondary optimization process is triggered, reanalyzing the impedance distribution characteristics and adjusting the optimization strategy and parameter allocation scheme until the optimization target is met.

[0119] During implementation, strict adherence to the physical properties of the electrodes and fuel cell safety operating specifications is crucial. For example, voltage adjustments must not exceed the tolerance limits of the electrode material, and resistance adjustments must be within the actuator's adjustable range. Furthermore, frequent parameter adjustments that can lead to fatigue and wear of the electrode assembly must be avoided, and reasonable adjustment intervals must be established (e.g., waiting 5-10 minutes after each adjustment to allow the system to stabilize before conducting the next monitoring and adjustment).

[0120] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0121] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing the impedance of a porous gradient anode-supported solid fuel cell, applied to a solid fuel cell management system, wherein the system comprises a plurality of impedance sensors deployed on a porous gradient anode and an optimization controller connected to the impedance sensors, wherein two adjacent impedance sensors are separated by a set distance, and wherein: The method comprises: The impedance sensor collects voltage data and current data of electrodes in the region to generate an impedance state data set; The optimization controller receives real-time impedance state data of multiple impedance sensors and constructs a regional impedance state matrix; Determining distribution characteristics of electrode impedance based on the impedance state data set and real-time data from each impedance sensor, wherein determining the distribution characteristics of electrode impedance includes: processing an impedance state matrix, extracting impedance characteristics based on the impedance state data set, predicting impedance distribution based on potential gradient and electrolyte path information, outputting spatiotemporal distribution characteristics of electrode impedance through an impedance optimization model, and updating the impedance state data set based on the spatiotemporal distribution characteristics; According to the distribution characteristics, regional electrode parameters are dynamically optimized.

2. The impedance optimization method of a porous gradient anode supported solid fuel cell according to claim 1, characterized in that: The determining of the distribution characteristics of the electrode impedance includes: Processing the impedance state matrix to extract potential intensity distribution, transmission delay characteristics and impedance change trends; Performing electric potential intensity modeling on the impedance state matrix according to the electric potential intensity distribution and transmission delay characteristics, dividing the regional electrode into a plurality of sub-regions and marking the regional identifiers, performing correlation matching with the impedance state data set according to the electric potential intensity of the sub-regions, and marking the regional identifiers in the impedance state data set; Calculating the potential intensity gradient according to the position of the impedance sensor, predicting the distribution of the electrode impedance according to the potential intensity gradient and the impedance change trend, and calculating the impedance prediction information of each sub-region; Constructing an impedance optimization model, using the impedance prediction information as an input parameter of the impedance optimization model, performing spatial correlation modeling on the impedance prediction information through the impedance optimization model, and outputting spatiotemporal distribution characteristics of electrode impedance; The impedance state data set is updated according to the spatiotemporal distribution characteristics to obtain the distribution characteristics of the electrode impedance.

3. The impedance optimization method of a porous gradient anode supported solid fuel cell according to claim 2, characterized in that: The processing of the impedance state matrix includes: The impedance state matrix is ​​normalized, the impedance hotspot area in the matrix is ​​intercepted by a sliding window, the hotspot area is noise filtered, and the impedance change trend is calculated by an eigendecomposition algorithm; Calculating the spatial correlation characteristics of the impedance state matrix, calculating the interference intensity, link stability coefficient and impedance blind zone index between regions based on the spatial correlation characteristics, constructing a feature fusion network, and calculating the transmission delay characteristics through the feature fusion network; The time domain features and frequency domain features collected by each impedance sensor are extracted, and the impedance characteristic vector of the sensor is calculated based on the phase difference between the time domain features and the frequency domain features. The impedance sensors at different positions are feature matched based on the impedance characteristic vector, and the impedance change trend is calculated.

4. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 3, characterized in that: The performing electric potential intensity modeling on the impedance state matrix includes: Extracting electric potential intensity sampling points from each frame of data based on the electric potential intensity distribution, performing correlation mapping between the sampling points and transmission delay characteristics to generate an electric potential intensity map, spatially aligning the electric potential intensity maps collected by multiple sensors, and calculating the electric potential intensity distribution of the region; Set the interference threshold, locate the interference source based on the potential strength value of the multi-frame impedance state matrix, and calculate the interference strength difference. If the interference strength difference is greater than or equal to the interference threshold, it indicates that there is an impedance blind spot in the area. Perform propagation model constraint compensation on the current area, iteratively correct the potential strength distribution of the current area based on the path loss model corresponding to the current area, and calculate the blind spot potential strength compensation value based on the correction result. The potential intensity modeling is performed on the impedance state matrix according to the potential intensity distribution, and the time delay is marked on the potential model of the region through the transmission delay characteristics.

5. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 4, characterized in that: Calculating the electric potential intensity gradient according to the position of the impedance sensor includes: Based on multiple sets of impedance coverage data, the potential intensity change points are extracted and mapped to a unified electrode coordinate system based on the deployment location of the sensor. The change points are fitted using a spatial interpolation algorithm to generate a regional potential field model. Performing equal-interval sampling along the transmission path of the electric potential field model, calculating the voltage attenuation rate, interference fluctuation index, and potential change slope of the path based on the sampling results, and calculating the potential change parameter based on the voltage attenuation rate, interference fluctuation index, and potential change slope; Based on the deployment parameters and acquisition accuracy of the impedance sensors, the distribution characteristics of the impedance change in each frame of data are projected onto the electric potential field model. The electric potential field model is partitioned along the transmission direction according to the number of sensors. The variation pattern of the impedance change within the partition is analyzed, and the impedance distribution characteristics are calculated based on the variation pattern. The electric potential intensity gradient is calculated based on the electric potential change parameter and the impedance distribution characteristic. The calculation process of the electric potential intensity gradient includes: based on the electrode position range from the first impedance sensor to the last impedance sensor, selecting spatial coordinate points in the sensor deployment direction, cumulatively calculating the product of the electric potential field strength characteristic weight value and the impedance distribution characteristic weight value within the spatial resolution range, and superimposing the influence value of the sensor acquisition frequency on the potential intensity change rate.

6. The impedance optimization method for a porous gradient anode-supported solid fuel cell according to claim 5, characterized in that: The calculating of impedance prediction information for each sub-region includes: Taking the main transmission path of the electric potential field model as the baseline and the peak position of the impedance change in each frame of data as the reference point, the impedance offset is calculated and the impedance distribution curve is drawn according to the electrode coordinates; Correcting the rate and direction of change in the impedance change trend according to the potential intensity gradient; Starting from the nearest impedance distribution point, the distribution curve is continuously drawn according to the correction results of the change rate and direction to generate the impedance distribution points of the next period until the distribution points cover the entire target area and generate impedance prediction information.

7. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 2, characterized in that: The constructing of the impedance optimization model includes: The input layer is used to organize the impedance prediction information into spatial distribution data and perform normalization processing; The feature fusion layer is used to extract the regional correlation features of impedance by processing spatial distribution data and construct the dependency relationship between electrode units; The parameter optimization layer is used to integrate the correlation between electrode impedances in spatial units and generate electrode parameter optimization strategies.

8. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 2, characterized in that: The obtaining of distribution characteristics of electrode impedance includes: According to the spatiotemporal distribution characteristics of the electrode impedance output by the impedance optimization model, the sub-region identifiers are matched with the spatiotemporal distribution characteristics; The regional data in the impedance state dataset are reorganized according to the spatiotemporal characteristics to generate a regional distribution map sorted by impedance intensity; According to the reorganized regional distribution map, the optimized electrode impedance distribution characteristics are output.

9. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 1, characterized in that: The dynamic optimization of regional electrode parameters includes: Mapping the regional identifiers to the regions of the distribution characteristics of the electrode impedance one by one; According to the spatiotemporal distribution characteristics of electrode impedance, the execution actions of electrode parameters are controlled, including voltage adjustment, current aggregation and resistance distribution operations; Based on the spatial distribution of electrode impedance and the preset parameter optimization strategy, the electrode parameters are dynamically allocated to the corresponding electrode areas.

10. The impedance optimization method for a porous gradient anode supported solid fuel cell according to claim 9, characterized in that: The execution action of controlling the electrode parameters includes: When the electrode impedance reaches a preset intensity threshold in the target area, a voltage increase instruction is triggered for the adjacent electrodes; Dynamically combine available circuit resources according to current aggregation strategy to generate resistance allocation vectors; Electrode parameters of transmission nodes in the target area are adjusted based on the resistance allocation vector.

Images