A method for balancing the lifespan and performance of a fuel cell

Abstract

The present invention relates to the technical field of battery energy storage system evaluation, and more particularly to a method for balancing the life and performance of fuel cells, comprising the following steps: Step 1: Collect multi-source data of the fuel cell, apply a preset series of pulsed current load sequences to the fuel cell, and measure the response parameters output by the fuel cell in real time; Step 2: Perform Fourier transform processing on the response parameters to obtain the frequency-domain signal of the response parameters. By constructing a performance and life balance model based on the fusion feature vector including the dynamic performance, life characteristics and operating state of the fuel cell, the present invention can dynamically adjust the load distribution weight, balance the performance and life of a single fuel cell, thereby extending the overall life of the fuel cell stack and avoiding the impact of a single fuel cell on the performance of the entire system due to overload or excessive loss.

Description

Technical Field

[0001] The present invention relates to the technical field of battery energy storage, and particularly relates to a method for balancing the life and performance of a fuel cell. Background Art

[0002] A fuel cell is an efficient and clean electrochemical energy conversion device and is regarded as one of the important choices for future energy. Compared with traditional internal combustion engines, fuel cells have the advantages of zero emissions, higher energy conversion efficiency, and low noise. Therefore, fuel cells have been widely used in fields such as transportation, stationary power supplies, and portable power supplies. However, a major challenge faced by fuel cells in practical applications is how to extend their operating life while maintaining high-performance output.

[0003] After retrieval, Chinese Patent No. CN202310167828.5 discloses a method for balancing battery endurance and battery life, including: when the battery operates at a first voltage, regularly monitor the current temperature of the battery; if the current temperature of the battery exceeds a preset warning temperature, record a high-temperature data once; determine whether the continuous recording times of the high-temperature data are greater than a preset number of times; if the continuous recording times of the high-temperature data are greater than the preset number of times, control the battery to operate with a second voltage as the full charge voltage of the battery; the second voltage is less than the first voltage; if the continuous recording times of the high-temperature data are not greater than the preset number of times, continue to use the first voltage as the full charge voltage of the battery. The above solution can ensure battery life and safety by reducing the battery voltage at high temperatures, can increase the battery usage time by raising the charging voltage at low temperatures, and detects that the recording times of high-temperature data are greater than the preset number of times to prevent misjudgment. If the battery needs to be stepped down only by recording high-temperature data once, misjudgment is likely to occur. After normally switching to a low voltage as the full charge voltage, set a period of time to re-record the high-temperature times. If there is no continuous high temperature exceeding the preset number of times in the middle, set the full charge voltage of the battery to a higher standard voltage for operation.

[0004] However, in the existing battery life balancing methods, although the endurance of the battery is dynamically balanced through the temperature of the battery to alleviate the influence of high-temperature and high-voltage environments on the battery life, the voltage of the battery is changed during the balancing process, thereby affecting the performance efficiency of the battery; but during the change of the battery temperature and voltage, different parameter responses will also be caused, thereby resulting in possible non-linear changes in the performance of the battery. The existing balancing methods lack an effective balancing mechanism for fusing multi-source information, resulting in insufficient robustness and accuracy of the evaluation results; in addition, in a battery pack, the performance and life balance of individual batteries may be uneven. If the performance and life balance strategy based on the overall battery pack fails to fully consider the differences of individual batteries, it may lead to instability of the overall balance value of the battery pack, further exacerbating the consistency problem between individual batteries. Summary of the Invention

[0005] In view of the above-mentioned drawbacks of the prior art, the present invention provides a method for balancing the life and performance of a fuel cell, which can effectively solve the problem in the prior art that the balance of battery performance is ignored when adjusting the battery life based on temperature.

[0006] To achieve the above object, the present invention is realized through the following technical solutions:

[0007] The present invention provides a method for balancing the life and performance of a fuel cell. The technical solution adopted by the present invention is as follows: including the following steps:

[0008] Step 1: Collect multi-source data of the fuel cell, apply a preset series of pulsed current load sequences to the fuel cell, and measure the response parameters output by the fuel cell in real time;

[0009] Step 2: Perform Fourier transform processing on the response parameters to obtain the frequency-domain signal of the response parameters;

[0010] Step 3: Extract frequency-domain characteristics from the frequency-domain signal, perform vector-level fusion on the frequency-domain characteristics and the multi-source data to obtain a fused feature vector;

[0011] Step 4: Construct a performance and life balance model based on the fused feature vector, and predict the balance index score of the performance and life of the fuel cell ;

[0012] Step 5: Construct a fuel cell stack regulation model based on the balance index score of the single fuel cell to calculate the load distribution weight of the single fuel cell stack , and then calculate the actual load distribution based on the fuel cell load distribution weight , use the actual load distribution to update its response parameters , and obtain the consistency regulation parameter; return the consistency regulation parameter to the performance and life balance model, calculate the updated balance index score , repeat the above steps, calculate the distribution weight difference of the single fuel cell , where is the updated load distribution weight; if the distribution weight difference , calculate the excess load distribution , is the total output power of the battery stack, and distribute the excess load distribution to other fuel cells.

[0013] Among them, the multi-source data includes:

[0014] Electrochemical performance parameters: the discharge capacity of the fuel cell Open-circuit voltage and resistance ;

[0015] Ambient parameters: temperature and humidity ;

[0016] Gas supply parameters: oxygen flow rate and hydrogen flow rate ;

[0017] Core component status: catalyst active area and proton exchange membrane conductivity , where the proton exchange membrane conductivity is calculated by the formula: ; In the formula, is the optimal conductivity; is the humidity deviation, is the temperature deviation, is the humidity attenuation factor, is the temperature attenuation factor, is the exponential function.

[0018] Among them, the preset method of the series of pulsed current load sequences is:

[0019] Define the type of pulse waveform as a rectangular wave, and define the pulse waveform parameters of the rectangular wave, including: maximum current density ; minimum current density ; duty cycle , the calculation formula is: , where is the duration of the high-current state, is the pulse period time; pulse frequency , the calculation formula is: ;

[0020] Based on the maximum current density , minimum current density and active area calculate the peak current and valley current , the calculation formula is: and ;

[0021] Based on the above parameters, calculate the average current density and average power density , the calculation formula is:

[0022] ;

[0023] ​​​​​ ; is the average voltage actually output by the fuel cell within the pulse period, and the calculation formula is ; represents the output voltage of the fuel cell at any time scale t, where t is the time scale;

[0024] Combining the time scale t and the pulse waveform parameters to construct a pulse current load sequence ;

[0025] The method for measuring the response parameters of the fuel cell output includes:

[0026] Applying a dynamic current to the fuel cell according to a preset series of pulse current load sequences and collecting the response parameters ;

[0027] In the formula, is the dynamic output voltage of the fuel cell; is the dynamic output current of the fuel cell; is the dynamic change internal resistance of the fuel cell; is the dynamic operating temperature of the fuel cell; is the dynamic change humidity in the fuel of the fuel cell; is the dynamic hydrogen supply amount at the anode; is the dynamic oxygen supply amount at the cathode; is the dynamic conductivity of the membrane; is the dynamic discharge capacitance of the fuel cell; is the active area of the catalyst of the fuel cell.

[0028] Among them, the method for performing Fourier transform processing on the response parameters includes:

[0029] Extracting time series data from the response parameters , including the dynamic output voltage and the dynamic output current , defining the sampling frequency and the number of sampling points ;

[0030] Removing the DC offset in the time series data, and the removal formula is: ; In the formula, is the original time series data; n is the discrete index of the time series data; represents the value of the time series data corresponding to the nth time scale t; represents the DC component of the time series data, and the calculation formula is: , is the total length of the time series data; Represents the time series data after removing the DC offset;

[0031] For the time series data after removing the DC offset Perform windowing processing, and the processing formula is: ; In the formula, is the window function; is the signal after being processed by the window function, that is, the weighted time series data;

[0032] Perform Fourier transform on the weighted time series data, and the transform formula is:

[0033] :

[0034] In the formula, is the frequency domain signal output by the fast Fourier transform processing; is the discrete frequency, and the calculation formula is: , where k is the frequency index; is the Fourier basis function;

[0035] Among them, the method of vector-level fusion of the frequency domain characteristics and multi-source data includes:

[0036] Calculate the key characteristics of the battery based on the frequency domain characteristics and multi-source data, including:

[0037] Voltage response characteristics ;

[0038] Discharge capacity retention rate ; In the formula, is the initial discharge capacity;

[0039] Energy efficiency ; In the formula, is the time interval; is the calorific value of hydrogen combustion;

[0040] Catalyst activity ; In the formula, is the initial catalyst active area;

[0041] Proton exchange membrane conduction performance ;

[0042] Performance decay rate ; In the formula, is the time derivative of the voltage drop; Standardize the above parameters with unified dimensions and construct a fusion feature vector .

[0043] Among them, the method of constructing the performance and life balance model includes:

[0044] Define the balance index for fuel cell performance and lifespan , and its calculation formula is ;

[0045] In the formula, is the short-term performance efficiency, and its calculation formula is: ; is the lifespan loss rate; is the adjustment cost of operating conditions, and the calculation formula is: , where , and are the adjustment cost weights of the corresponding parameters, usually determined by providing experimental data; , and are the weight coefficients of the corresponding parameters, .

[0046] Among them, the acquisition method of the lifespan loss rate is:

[0047] Input the fused feature vector into the pre-trained lifespan loss model to predict the lifespan loss rate ; The training method of the lifespan loss model includes:

[0048] Define a recurrent neural network as the basic structure of the lifespan loss model. The basic structure includes an input layer, a hidden layer, and an output layer;

[0049] Collect the frequency domain characteristics and multi-source data of multiple fuel cells in the past fixed time, and correspondingly construct them into a fused feature vector, denoted as the historical fused feature vector , where represents the vector time step; Expand the historical fused feature vector by time step to form sequence data as the training input of the lifespan loss model, and correspondingly label the sequence data according to the vector time step; The label is the lifespan status index of the fuel cell; The annotation range of the lifespan status index is from 0 to 100%;

[0050] Initialize the network parameters of the lifespan loss model; The network parameters include the weight matrix from the input layer to the hidden layer , the recurrent weight matrix of the hidden layer , the bias vector of the hidden layer , the weight matrix from the hidden layer to the output layer and the bias vector of the output layer ;

[0051] Define the loss function of the lifespan loss model ; Among them, is the length of the sequence data; is the loss weight function weighted according to the life state; is the life state index of the th vector time step; is the life state index labeled at the th vector time step; is the predicted value of the life loss model at the th vector time step;

[0052] ; where, is the indicator function;

[0053] is the threshold interval of the preset life state, and the upper limit of the threshold interval is ; the lower limit of the threshold interval is ; if is less than , then is ; if is within the threshold interval, then is 1; if is greater than , then is ; where, and are preset weight coefficients, and is greater than 1, is less than 1;

[0054] The sequence data is passed to the input layer, then passes through the hidden layer and the output layer in sequence to output the predicted value; and calculate the value of the corresponding loss function, and backpropagate the error gradient along the output layer to the input layer; according to the error gradient, use the optimization algorithm to update the network parameters; repeat training on the sequence data until the life loss model converges or reaches the preset number of iterations in advance, that is, the training of the life loss model is completed.

[0055] Among them, the input layer is used to receive the fused feature vector as the input; the hidden layer consists of neurons;

[0056] The calculation formula of each neuron is: ; where, is the hidden state vector of the current vector time step , is the hidden state vector of the previous vector time step; is the activation function; is the weight matrix from the input layer to the hidden layer; is the recurrent weight matrix of the hidden layer; is the bias vector of the hidden layer;

[0057] The output layer calculates the predicted value based on the output of the hidden layer ; where, is the activation function of the output layer; is the weight matrix from the hidden layer to the output layer; is the bias vector of the output layer.

[0058] Among them, the method for constructing the fuel cell stack regulation model is:

[0059] Define the load distribution weight The calculation formula is: ; represents the load weight ratio that the th single fuel cell should be allocated; is the balance index score of the th fuel cell monomer; represents the sum of the balance index scores of all single fuel cells in the fuel cell stack, is the total number of single fuel cells, is the index of the fuel cell;

[0060] Define the actual load distribution The calculation formula is: ; In the formula is the total output probability of the battery stack, and the calculation formula is: , where and are the total output voltage and total output current of the fuel cell stack respectively;

[0061] Define the update formula of the consistency regulation parameter including:

[0062] ;

[0063] In the formula, is the actual current of the fuel cell under the actual load distribution ; is the operating voltage of the th fuel cell; is the internal resistance of the th fuel cell; is the voltage drop in the ohmic region; is the power loss caused by the internal resistance of the fuel cell; is the actual energy utilization efficiency of the th fuel cell; is the input energy of a single fuel cell, and its calculation formula is: , where is the Gibbs free energy of the reaction;

[0064] Calculate the distribution weight difference of a single fuel cell , and the calculation formula is: .

[0065] Among them, the constraint condition of the updated load distribution weight is: ; Excess load distribution 's constraint condition is: .

[0066] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are:

[0067] 1. In the present invention, collecting multi-source data can comprehensively reflect the operating state of the battery, provide a high-quality basis for subsequent data processing, and ensure the accuracy of the analysis results.

[0068] 2. In the present invention, by applying a pulsed current load sequence to the fuel cell, the dynamic response characteristics of the fuel cell under different load conditions can be captured, revealing its performance potential and limitations; and the response parameters of the fuel cell under different loads are collected. By performing Fourier transform processing on the response parameters, the non-linear behavior and dynamic response ability of the battery can be revealed, quantifying the response ability of the fuel cell to different frequency loads, and providing a basis for performance evaluation.

[0069] 3. In the present invention, by performing vector-level fusion of the frequency domain characteristics and multi-source data, the obtained fusion feature vector contains the dynamic performance, life characteristics, and operating state of the fuel cell, which can comprehensively reflect the health status of the battery. Moreover, the unified form of the feature vector is convenient for input into the machine learning model, improving the prediction ability of the model and ensuring that the model can reflect the latest state of the battery; then, based on the fusion feature vector, a performance and life balance model is constructed, which can dynamically adjust the load distribution weight, keeping the performance and life of a single fuel cell in balance, thereby extending the overall life of the fuel cell stack and avoiding the influence of a single fuel cell due to overloading or excessive loss on the performance of the entire system.

[0070] 4. In the present invention, by calculating the load distribution difference of a single fuel cell, dynamically adjusting the load distribution, protecting low-score batteries, and preferentially using high-score batteries, the operating consistency and overall health status of the battery stack can be improved, and the load distribution and operating conditions can be optimized, enhancing the overall efficiency and reliability of the battery stack. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 is a schematic diagram of the method flow in the embodiment of the present invention. Specific Embodiments

[0072] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0073] Embodiment: Referring to Figure 1 , a method for balancing the life and performance of a fuel cell is proposed in this case, including the following steps:

[0074] Step 1: Collect multi-source data of the fuel cell, apply a preset series of pulsed current load sequences to the fuel cell, and measure the response parameters output by the fuel cell in real time;

[0075] Step 2: Perform Fourier transform processing on the response parameters to obtain the frequency-domain signal of the response parameters;

[0076] Step 3: Extract frequency-domain characteristics from the frequency-domain signal, perform vector-level fusion on the frequency-domain characteristics and the multi-source data to obtain a fusion feature vector;

[0077] Step 4: Construct a performance and life balance model based on the fusion feature vector, and predict the balance index score of the fuel cell performance and life ;

[0078] Step 5: Construct a fuel cell stack regulation model based on the balance index score of a single fuel cell, calculate the load distribution weight of the single fuel cell stack , and then calculate the actual load distribution based on the fuel cell load distribution weight , use the actual load distribution to update its response parameters , and obtain a consistency regulation parameter; return the consistency regulation parameter to the performance and life balance model, calculate the updated balance index score , repeat the above steps, calculate the distribution weight difference of the single fuel cell , where is the updated load distribution weight; if the distribution weight difference , calculate the excess load distribution , and distribute the excess load distribution to other fuel cells.

[0079] Specifically, in this case, the multi-source data of the fuel cell includes:

[0080] Electrochemical performance parameters: the discharge capacity of the fuel cell , open circuit voltage and resistance ; Discharge capacity is a direct measure of the remaining available life of the fuel cell; Open circuit voltage characterizes the internal electrochemical energy of the fuel cell under no-current conditions, is used for reference voltage, and evaluates whether there is significant degradation;

[0081] Environmental parameters: Temperature and humidity ; Temperature affects the reaction rate on the catalyst surface, thus affecting the output performance of the fuel cell. It also corrects the conductivity of the proton exchange membrane and is a key regulator of the conductivity of the proton exchange membrane. Temperature deviation from the optimal range may lead to dehydration or over-hydration of the proton exchange membrane, thus affecting the conductivity of the proton exchange membrane; Humidity is used for hydration state evaluation and gas management optimization. Humidity is crucial for the ionic conductivity of the proton exchange membrane. Humidity too low will cause membrane dehydration, and humidity too high will cause difficulties in gas transport and flooding, and is also an important environmental condition for gas diffusion and reaction rate in the fuel cell;

[0082] Gas supply parameters: Oxygen flow rate and hydrogen flow rate ; Used for gas utilization analysis and mass transfer impedance analysis. The dynamic change of gas flow rate reflects the utilization efficiency of reactants. Whether the supply of hydrogen and oxygen is sufficient directly affects the working performance of the fuel cell; Insufficient gas supply will lead to an increase in concentration polarization in the oxygen or hydrogen region, thus increasing the internal resistance and affecting the performance;

[0083] Core component status: Catalyst active area and proton exchange membrane conductivity , The active area affects the range of the total current and is used to convert the current density into the actual applied current value, usually obtained by measuring the actual anode or cathode surface area of the fuel cell; Proton exchange membrane conductivity represents the dynamic change of the proton exchange membrane's conductivity per unit time and unit thickness, and is the core index of the proton transport ability in the fuel cell. It is significantly affected by temperature and humidity , and its calculation formula is: ; Among them, is the optimal conductivity of the proton exchange membrane under ideal humidity and temperature conditions; is the humidity deviation, which is the deviation between the current humidity and the ideal humidity , and the calculation formula is: ; is the temperature deviation, which is the current temperature minus the ideal temperature and the calculation formula is: ; is the humidity attenuation factor, which is a coefficient describing the influence of humidity deviation on the conductivity of the proton exchange membrane; is the temperature attenuation factor, which is a coefficient describing the influence of temperature deviation on the conductivity of the proton exchange membrane; is the exponential function.

[0084] Collecting multi-source data can comprehensively reflect the operating state of the battery, provide a high-quality basis for subsequent data processing, and ensure the accuracy of the analysis results.

[0085] More specifically, the preset method of a series of pulsed current load sequences is as follows:

[0086] Define the type of the pulse waveform as a rectangular wave, and define the pulse waveform parameters of the rectangular wave, including:

[0087] The maximum current density , which is the current density at the peak part of the pulse waveform, represents the current intensity at the maximum dynamic load, and is usually set according to the limit capacity of the fuel cell to avoid exceeding the rated operating range of the battery. Generally, it is set to 70%-80% of the rated operating density of the fuel cell to reflect the typical dynamic load capacity of the fuel cell and avoid problems such as overheating, drying of the proton exchange membrane, or insufficient fluid supply caused by setting too high;

[0088] The minimum current density , which is the current density at the trough part of the pulse waveform, represents the current intensity at the minimum dynamic load, and is generally set to 20%-30% of the rated operating density of the battery to ensure that the electrochemical reaction is still within the normal range and avoid setting it to 0 (open circuit state), resulting in unstable experimental results or inability to test the dynamic response;

[0089] The duty cycle , which represents the proportion of the duration of the high current load in the pulse waveform, is an important parameter for simulating the actual dynamic load situation, and the calculation formula is: , where is the duration of the high current state, is the pulse period time (the sum of the high and low current state times);

[0090] The pulse frequency , that is, the switching rate of the pulse waveform, is the frequency of the repeated change of the waveform, and the calculation formula is: ;

[0091] Based on the maximum current density , the minimum current density and the active area Calculate the peak current and the valley current , and the calculation formula is: and ;

[0092] Based on the above parameters, calculate the dynamic relationship between the pulse waveform parameters, including the average current density and the average power density , and the calculation formula is:

[0093] ; The average current density represents the average current density of the fuel cell within a pulse period and is used to evaluate the overall ability of the fuel cell under continuous load;

[0094] ; Among them, the average power density represents the average probability output per unit active area of the fuel cell within a pulse period and is the core index to measure the overall output ability and energy utilization efficiency of the fuel cell; is the average voltage actually output by the fuel cell during the pulse period, and the calculation formula is ; represents the output voltage of the fuel cell at any time scale t, and t is the time scale;

[0095] Combine the time scale t and the pulse waveform parameters to construct a pulse current load sequence ;

[0096] Apply a dynamic current to the fuel cell according to a preset series of pulse current load sequences and collect response parameters ;

[0097] Among them, is the dynamic output voltage of the fuel cell, reflecting the instantaneous energy performance and dynamic response efficiency of the fuel cell; is the dynamic output current of the fuel cell, which is an important indicator of gas consumption and is closely related to the dynamic load; is the dynamic change internal resistance of the fuel cell, reflecting the health state of the internal materials of the fuel cell and the dynamic operation stability; is the dynamic operating temperature of the fuel cell, affecting the catalyst performance, proton exchange membrane conductivity and reaction rate; is the dynamic change humidity in the fuel of the fuel cell, affecting the hydration state of the proton exchange membrane; is the dynamic hydrogen supply amount at the anode, related to the chemical reaction amount of the fuel cell; is the dynamic oxygen supply amount at the cathode; is the dynamic membrane conductivity; is the dynamic discharge capacitance of the fuel cell; is the catalyst active area of the fuel cell.

[0098] By applying a pulsed current load sequence, the dynamic response characteristics of the fuel cell under different load conditions can be captured, revealing its performance potential and limitations.

[0099] For the response parameters The ways of performing Fourier transform processing include:

[0100] Extract key time series data from the response parameters including the dynamic output voltage and the dynamic output current , define the sampling frequency and the number of sampling points ;

[0101] Remove the DC offset in the time series data in order to focus on analyzing the dynamic change part. The removal formula is: ; In the formula, is the original time series data; n is the discrete index of the time series data, representing a certain discrete time scale; represents the value of the time series data corresponding to the nth time scale t; represents the DC component of the time series data, that is, the average value, which is used to extract the static and low-frequency bias part in the time series data. The calculation formula is: , is the total length of the time series data; represents the time series data after removing the DC offset, that is, the remaining part of the time series data only represents the dynamic characteristics changing with time, and in the time domain, it shows more concentrated fluctuations around zero than the original data;

[0102] Perform windowing processing on the time series data after removing the DC offset to reduce spectral leakage. The processing formula is: ; In the formula, is the window function, a weight function used to force the time series data boundary to gradually decrease to zero; is the signal after being processed by the window function, that is, the weighted time series data;

[0103] Perform Fourier transform on the weighted time series data. The transform formula is:

[0104] :

[0105] In the formula, The frequency-domain signal output by the fast Fourier transform processing is a complex number, representing the amplitude and phase of the response parameter at discrete frequencies ; The discrete frequency is calculated by the formula: , where k is the frequency index; is the Fourier basis function.

[0106] The method of extracting the frequency-domain characteristics from the frequency-domain signal is as follows:

[0107] Gradually calculate the amplitude of the frequency-domain signal, and the calculation formula is:

[0108] ;

[0109] In the formula, is the square of the real part, representing the intensity based on the cosine wave; is the square of the imaginary part, representing the intensity based on the sine wave;

[0110] Find the frequency component with the largest amplitude in the frequency-domain signal , which usually corresponds to the main dynamic characteristics of the time-series data, such as the pulse frequency in the pulse current load sequence, and the calculation formula is:

[0111] ;

[0112] Extract the amplitude of the harmonic component, which is the integer multiple frequency component of the pulse frequency , and normalize the amplitude of the harmonic component to the ratio of the amplitude of the pulse frequency , that is, the harmonic amplitude , to quantify the nonlinear characteristics; the calculation formula is: ; where, represents the amplitude of the k-th harmonic component;

[0113] Calculate the frequency-domain energy distribution in the frequency component , and the calculation formula is: , where the integration range [a, b] represents the frequency interval for energy calculation; the frequency-domain energy distribution includes low-frequency energy (the energy in the frequency range [0, 1Hz]) and high-frequency energy (the energy in the frequency range [1, 5Hz]), and the calculation formulas are respectively: , ; where, is the discrete frequency interval, and the calculation formula is ; Then based on the low-frequency energy and high-frequency energy Calculate the energy ratio for evaluating the dynamic complexity of time-series data , and the calculation formula is: ;

[0114] Extract other parameters from the response parameters , including the dynamically changing internal resistance of the fuel cell , the dynamic operating temperature , the dynamically changing humidity , the dynamic hydrogen supply , the dynamic oxygen supply and the dynamic membrane conductivity , which are used to explain the frequency-domain characteristics, and the expression of the explanatory relationship includes: , , and ;

[0115] Finally, output the frequency-domain characteristics, including the frequency component with the largest amplitude , the harmonic amplitude , the frequency-domain energy distribution and the energy ratio ; By converting the response parameters into a frequency-domain signal , the dynamic characteristics of the fuel cell can be extracted, and then the frequency-domain characteristics can be extracted from the frequency-domain signal , which can reveal the non-linear behavior and dynamic response ability of the battery, quantify the response ability of the fuel cell to different frequency loads, and provide a basis for performance evaluation.

[0116] The methods for vector-level fusion of frequency-domain characteristics and multi-source data include:

[0117] Calculate the key characteristics of the battery based on the frequency-domain characteristics and multi-source data, including:

[0118] Voltage response characteristics , which reflect the response ability of the fuel cell to load changes under the pulse frequency of the pulse current load sequence, and the calculation formula is: ; When , it indicates that the fuel cell is operating stably;

[0119] Discharge capacity retention rate , which reflects the remaining available capacity of the fuel cell, and the calculation formula is:

[0120] ; In the formula, is the initial discharge capacity. When , it indicates that the battery discharge capacity is completely normal. When Indicates that the battery discharge capacity decays due to aging or fuel shortage;

[0121] Energy efficiency , which directly measures the utilization degree of the input energy (chemical fuel) by the fuel cell, is an important reference for optimizing the operating performance. The calculation formula is: ; In the formula, is the time interval; is the calorific value of hydrogen combustion; When , it indicates that the battery has a high energy utilization rate and a reasonable conversion efficiency. When , it indicates that there is an efficiency loss, usually caused by problems such as too high internal resistance or side reactions;

[0122] Catalyst activity , which is a quantitative value reflecting the health state of the catalyst. The calculation formula is:

[0123] ; In the formula, is the initial catalyst active area, and the value of the catalyst activity characteristic represents the health state of the current activity. When , it indicates that the catalyst is in the initial active state. When , it indicates that the catalyst activity has decreased due to aging or pollution;

[0124] Proton exchange membrane conduction performance , which affects the proton exchange membrane conductivity . The calculation formula is:

[0125] ; The value of the proton exchange membrane conduction performance represents the performance state of the proton exchange membrane conductivity . When , it indicates that the proton exchange membrane has the best conductivity. When , it indicates that the proton exchange membrane conductivity has declined due to drying, degradation or high temperature, etc.;

[0126] Performance decay rate , which reflects the performance decline speed of the fuel cell. The calculation formula is:

[0127] ; In the formula, is the time derivative of the voltage drop, indicating the rate of performance decay of the fuel cell; When , it indicates performance decline. When , it indicates that the performance remains stable;

[0128] Unify the dimension normalization processing of the above parameters and construct them into a fused feature vector , which can comprehensively reflect the state and performance of the fuel cell.

[0129] Fused feature vector It includes the dynamic performance, life characteristics and operating status of the fuel cell, can comprehensively reflect the health status of the battery, and the unified form of the feature vector is convenient for input into the machine learning model to improve the prediction ability of the model and ensure that the model can reflect the latest state of the battery.

[0130] The ways to build a performance and life balance model include:[[]]

[0131] Define the balance index score of the fuel cell performance and life , which is used to comprehensively evaluate the current operating status of the fuel cell, and its calculation formula is ;

[0132] In the formula, The higher the value of, the better the balance between the performance and life of the battery, If the value of is too low, it means that the operating conditions need to be optimized or maintenance is required; is the short-term performance efficiency, which reflects the energy utilization efficiency of the fuel cell under the current operating conditions, and its calculation formula is: ; is the life loss rate, which reflects the degradation rate of the fuel cell; is the adjustment cost of the operating conditions, which quantifies the potential impact of performance optimization on life, and the calculation formula is: , where, , and are the adjustment cost weights of the corresponding parameters, usually determined by providing experimental data; , and are the weight coefficients of the corresponding parameters, which are used to adjust the importance of performance, life and adjustment cost in the comprehensive index, ;

[0133] Life loss rate The acquisition method of is:

[0134] Input the fused feature vector into the pre-trained life loss model to predict the life loss rate ; The training method of the life loss model includes:

[0135] Define a recurrent neural network as the basic structure of the life loss model, and the basic structure includes an input layer, a hidden layer and an output layer;

[0136] The input layer is used to receive the fused feature vector as input, where represents the vector time step; The hidden layer consists of It consists of

[0137] The calculation formula for each neuron is: ; where is the current vector time step of the hidden state vector, is the hidden state vector of the previous vector time step; is the activation function (such as tanh or ReLU); is the weight matrix from the input layer to the hidden layer; is the recurrent weight matrix of the hidden layer; is the bias vector of the hidden layer;

[0138] The output layer calculates the predicted value based on the output of the hidden layer ; where is the activation function of the output layer (such as a linear function or softmax); is the weight matrix from the hidden layer to the output layer; is the bias vector of the output layer.

[0139] Collect the frequency domain characteristics and multi-source data of multiple fuel cells over a fixed past time, and correspondingly construct a fused feature vector, denoted as the historical fused feature vector ; Unfold the historical fused feature vector by time step to form sequence data as the training input of the life loss model, and correspondingly label the sequence data according to the vector time step; The label is the life state index of the fuel cell; The labeling range of the life state index is from 0 to 100%;

[0140] Initialize the network parameters of the life loss model; The network parameters include the weight matrix from the input layer to the hidden layer , the recurrent weight matrix of the hidden layer , the bias vector of the hidden layer , the weight matrix from the hidden layer to the output layer and the bias vector of the output layer ;

[0141] Define the loss function of the life loss model ; where is the length of the sequence data; is the loss weight function weighted according to the life state; is the th life state index of the vector time step; is the quantile loss function; is the th labeled life state index of the vector time step; is the The predicted value of the life loss model for a single vector time step; is the quantile level;

[0142] ; where, is the indicator function; the indicator function is used to represent whether a certain condition holds;

[0143] The threshold interval of the preset life state, the upper limit of the threshold interval is ; the lower limit of the threshold interval is ; if is less than , then is ; if is within the threshold interval, then is 1; if is greater than , then is ; where, and are preset weight coefficients, and is greater than 1, the purpose is to increase the loss weight when the life state is poor, so that the model pays more attention to this state; is less than 1; the purpose is to appropriately reduce the loss weight when the life state is good, to avoid overtracking this state; specifically, the values of and are obtained according to the deviation distribution of the model prediction when the life state is poor or good by analyzing historical data, and according to the actual application scenario, the tolerance degree for the life state being poor or good is determined;

[0144] The setting of the life state threshold interval and the weight coefficient effectively guides the model to pay more attention to the key area and improves the accuracy of the battery life state evaluation.

[0145] The sequence data is passed to the input layer, and then passes through the hidden layer and the output layer in turn to output the predicted value, that is, the life loss rate ; and calculate the value of the corresponding loss function, and backpropagate the error gradient along the output layer to the input layer; according to the error gradient, use an optimization algorithm (such as gradient descent, Adam, etc.) to update the network parameters to reduce the value of the loss function; repeat the training of the sequence data until the life loss model converges or reaches the preset number of iterations in advance, that is, the training of the life loss model is completed; the convergence of the life loss model can be understood as that the value of the loss function no longer changes.

[0146] Compare the balance index score of the fuel cell performance and life with the preset balance threshold interval, the upper limit of the balance threshold interval is , the lower limit of the balance threshold interval is ; If , it indicates that the performance and lifespan of the fuel cell reach the optimal balance state and no additional optimization is required; if , it indicates that the performance and lifespan of the fuel cell are not fully balanced and regulation is needed; if , it indicates that the operating state of the fuel cell has significantly deteriorated and the operating conditions need to be optimized first.

[0147] By predicting the balance index score of the performance and lifespan of each fuel cell through a machine learning model , the latest state of the battery can be dynamically reflected. Among them, the performance parameters provide the immediate operating performance of the battery, the lifespan parameters reveal the long-term health trend, and the balance index score combines the two to achieve a comprehensive assessment of the operating state.

[0148] The method for constructing the fuel cell stack regulation model is as follows:

[0149] Based on the balance index score of the performance and lifespan of each fuel cell calculate its load distribution weight , and the calculation formula is: ; represents the load weight ratio that the th single fuel cell should be allocated. The value of each load distribution weight is a number between 0 and 1. By adjusting the load distribution weight , more load is allocated to the single fuel cell with a better balance of performance and lifespan, reducing the pressure on the single fuel cell with a poorer balance of performance and lifespan, thereby optimizing the consistency and overall lifespan of the fuel cell stack; is the balance index score of the th fuel cell monomer; represents the sum of the balance index scores of all single fuel cells in the fuel cell stack, is the total number of single fuel cells, is the index of the fuel cell;

[0150] Based on the load distribution weight of each single fuel cell calculate its actual load distribution , and the calculation formula is: ; In the formula, is the total output probability of the battery stack, and the calculation formula is: , where and are the total output voltage and total output current of the fuel cell stack respectively;

[0151] Based on the actual load distribution of each single fuel cell update its response parameter to obtain the consistency regulation parameter, including:

[0152] ;

[0153] Wherein, is the actual current of the fuel cell under actual load distribution ; is the operating voltage of the th fuel cell; is the th internal resistance of the fuel cell; is the voltage drop in the ohmic region, obtained by experimental measurement; is the power loss of the fuel cell due to internal resistance; is the actual energy utilization efficiency of the th fuel cell, reflecting how much of the input energy is converted into effective output power; is the input energy of a single fuel cell, and its calculation formula is: , where is the Gibbs free energy of the reaction;

[0154] Return the consistency regulation parameter to the performance and life balance model, and calculate the updated balance index score . Based on the updated balance index score calculate the updated load distribution weight ; and the updated load distribution weight satisfies the following conditions: ;

[0155] Calculate the distribution weight difference of a single fuel cell, and the calculation formula is: ;

[0156] If the distribution weight difference , calculate the excess load distribution , and distribute the excess load distribution to other single fuel cells to ensure that the total power demand of the fuel cell stack is such that while each fuel cell in the fuel cell stack maintains the highest efficiency and longest life, the fuel cell stack maintains the highest efficiency and longest life.

[0157] By calculating the load distribution differences of single fuel cells, dynamically adjusting the load distribution, protecting low-score cells, and preferentially utilizing high-score cells, the operating consistency and overall health of the fuel cell stack can be improved, and the load distribution and operating conditions can be optimized to enhance the overall efficiency and reliability of the fuel cell stack.

[0158] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for balancing fuel cell life and performance, characterized in that: The following steps are involved: Step 1: Collect multi-source data of the fuel cell, apply a preset pulse current load sequence to the fuel cell, and measure the response parameters of the fuel cell output in real time; Step 2: Perform Fourier transform processing on the response parameters to obtain the frequency domain signal of the response parameters; The method of performing Fourier transform processing on the response parameters includes: From the response parameters Extract time series data, including the dynamic output voltage of the fuel cell and dynamic output current , defines the sampling frequency of time series data and number of sampling points ; Remove the DC offset in the time series data. The removal formula is: ; In the formula, is the original time series data; n is the discrete index of the time series data; Indicates the value of the time series data at the nth time scale t; Represents the DC component of time series data, and the calculation formula is: , is the total length of the time series data; represents the time series data after removing the DC offset; After removing the DC offset, the time series data Perform windowing processing, the processing formula is: ; In the formula, is the window function; is the signal processed by the window function, that is, the weighted time series data; Perform Fourier transform on the weighted time series data, and the transformation formula is: ； In the formula, The frequency domain signal output by fast Fourier transform processing; is the discrete frequency, and the calculation formula is: , k is the frequency index; is the Fourier basis function; Step 3: Extract frequency domain characteristics from frequency domain signals, fuse frequency domain characteristics and multi-source data at vector level to obtain fused feature vectors; The frequency domain characteristics include the frequency component with the largest amplitude , harmonic amplitude , frequency domain energy distribution and energy ratio ; The way to extract frequency domain characteristics is: Calculate the amplitude of a frequency domain signal step by step , find the frequency domain signal The frequency component with the largest amplitude ; Normalize the amplitude of the harmonic components to the pulse frequency The ratio of the amplitude, i.e. the harmonic amplitude ;in, represents the amplitude of the kth harmonic component; Calculate frequency components Energy distribution in the mid-frequency domain ; The integral range [a, b] represents the frequency range of energy calculation; frequency domain energy distribution Including low frequency energy and high frequency energy , and then based on the low-frequency energy and high frequency energy Calculating Energy Ratio ; Step 4: Construct a performance and life balance model based on the fusion feature vector to predict the balance index of fuel cell performance and life. ; Step 5: Balance index analysis based on single fuel cell Construct a fuel cell group control model and calculate the load distribution weight of the single fuel cell group , and then calculate the actual load distribution based on the fuel cell load distribution weight , using actual load distribution Its response parameters Update and obtain consistency control parameters; return the consistency control parameters to the performance and life balance model, and calculate the updated balance index score. Repeat the above steps to calculate the distribution weight difference of the single fuel cell ,in Assign weights to the updated load; if the assigned weights are different , calculate excess load distribution , The total output power of the battery pack and the excess load distribution distributed to other fuel cells.

2. A method for balancing fuel cell life and performance as claimed in claim 1, characterized in that: The multi-source data includes: Electrochemical performance parameters: discharge capacity of fuel cells , open circuit voltage and resistor ; Environmental parameters: Temperature and humidity ; Gas supply parameters: oxygen flow rate and hydrogen flow rate ; Core component status: catalyst active area and proton exchange membrane conductivity , where the conductivity of the proton exchange membrane is The calculation formula is: ; In the formula, is the optimal conductivity; is the humidity deviation, is the temperature deviation, is the humidity attenuation factor, is the temperature attenuation factor, is an exponential function.

3. A method for balancing fuel cell life and performance as claimed in claim 2, characterized in that: The pulse current load sequence is preset as follows: Define the pulse waveform type as rectangular wave, and define the pulse waveform parameters of the rectangular wave, including: maximum current density ; Minimum current density ; Duty cycle , the calculation formula is: ,in is the duration of the high current state, is the pulse cycle time; pulse frequency , the calculation formula is: ; Based on the maximum current density , minimum current density and active area Calculating Peak Current and valley current , the calculation formula is: and ; Based on the above parameters, calculate the average current density and average power density , the calculation formula is: ； ; is the average voltage actually output by the fuel cell during the pulse period, and the calculation formula is: ; represents the output voltage of the fuel cell at any time scale t, where t is the time scale; Combine the time scale t and the pulse waveform parameters to construct a pulse current load sequence ; The method of measuring the response parameter output by the fuel cell includes: Apply dynamic current to the fuel cell according to the preset pulse current load sequence and collect response parameters ; In the formula, is the dynamic output voltage of the fuel cell; is the dynamic output current of the fuel cell; is the dynamically changing internal resistance of the fuel cell; is the dynamic operating temperature of the fuel cell; is the dynamically changing humidity within the fuel in the fuel cell; is the dynamic supply of hydrogen to the anode; The dynamic supply of oxygen to the cathode; is the membrane dynamic conductivity; is the dynamic discharge capacitance of the fuel cell; is the catalyst active area of ​​the fuel cell.

4. A method for balancing fuel cell life and performance as claimed in claim 3, characterized in that: The method of vector-level fusion of the frequency domain characteristics and multi-source data includes: Calculate key battery characteristics based on frequency domain characteristics and multi-source data, including: Voltage response characteristics ; Discharge capacity retention rate ; In the formula, is the initial discharge capacity; Energy efficiency ; In the formula, is the time interval; is the calorific value of combustion of hydrogen; Catalyst activity ; In the formula, is the initial catalyst active area; Proton exchange membrane conductivity ; Performance decay rate ; In the formula, is the time derivative of the voltage drop; the above parameters are standardized and dimensionalized, and constructed as a fusion feature vector .

5. A method for balancing fuel cell life and performance as claimed in claim 4, characterized in that: The method of constructing the performance and life balance model includes: Defining the balance between fuel cell performance and life , and its calculation formula is ; In the formula, is the short-term performance efficiency, which is calculated as: ; is the life loss rate; is the adjustment cost of the operating conditions, and the calculation formula is: ,in, , and is the adjustment cost weight of the corresponding parameter; , and is the weight coefficient of the corresponding parameter, .

6. A method for balancing fuel cell life and performance as claimed in claim 5, characterized in that: The life loss rate The way to obtain is: The fused feature vector Input into the pre-trained life loss model to predict the life loss rate ; The training methods of the life loss model include: Define a recursive neural network as the basic structure of the life loss model, which includes an input layer, a hidden layer, and an output layer; Collect the frequency domain characteristics and multi-source data of multiple fuel cells in the past fixed time, and construct the corresponding fusion feature vector, which is recorded as the historical fusion feature vector ,in Represents a vector time step; fuses the history with the feature vector Expand by time step to form sequence data as the training input of the life loss model, and label the sequence data according to the vector time step; the label is the life state index of the fuel cell; the labeling range of the life state index is 0 to 100%; Initialize the network parameters of the life loss model; the network parameters include the weight matrix from the input layer to the hidden layer , the recurrent weight matrix of the hidden layer , the bias vector of the hidden layer , the weight matrix from hidden layer to output layer and the bias vector of the output layer ; Define the loss function for the lifetime loss model ;in, is the length of the sequence data; is the loss weight function weighted according to the life state; It is A vector of life state indices for time steps; is the quantile loss function; It is The life state index is annotated with a vector of time steps; It is The predicted values ​​of the life loss model for vector time steps; is the quantile level; ;in, is the indicative function; The threshold interval of the preset life state, the upper limit of the threshold interval is ; The lower limit of the threshold interval is ;like Less than ,but for ;like If it is within the threshold range, is 1; if Greater than ,but for ;in, and is the preset weight coefficient, and greater than 1, Less than 1; The sequence data is passed to the input layer, and then passes through the hidden layer and the output layer in turn to output the predicted value; the value of the corresponding loss function is calculated, and the error gradient is back-propagated from the output layer to the input layer; according to the error gradient, the network parameters are updated using the optimization algorithm; the sequence data is repeatedly trained until the life loss model converges or reaches the preset number of iterations, that is, the training of the life loss model is completed.

7. A method for balancing fuel cell life and performance as claimed in claim 6, characterized in that: The input layer is used to receive the fused feature vector as input; the hidden layer consists of It is composed of neural units; The calculation formula for each neural unit is: ;in, is the current vector time step The hidden state vector of is the hidden state vector of the previous vector time step; is the activation function; is the weight matrix from the input layer to the hidden layer; is the recurrent weight matrix of the hidden layer; is the bias vector of the hidden layer; The output layer calculates the predicted value based on the output of the hidden layer ;in, is the activation function of the output layer; is the weight matrix from the hidden layer to the output layer; is the bias vector of the output layer.

8. A method for balancing fuel cell life and performance as claimed in claim 7, characterized in that: The method of constructing the fuel cell group control model is: Defining load distribution weights The calculation formula is: ; Indicates The load weight ratio that should be allocated to each single fuel cell; It is The balance index of each fuel cell monomer; It represents the sum of the balance indexes of all the individual fuel cells in the fuel cell stack. is the total number of single fuel cells, is the index of the fuel cell; Defining the actual load distribution The calculation formula is: ; In the formula is the total output probability of the battery pack, and the calculation formula is: ,in and are the total output voltage and total output current of the fuel cell group respectively; The update formula that defines the consistency control parameters includes: ； In the formula, For fuel cells in actual load distribution The actual current under For the The operating voltage of each fuel cell; For the The internal resistance of a fuel cell; is the voltage drop in the ohmic region; is the power loss caused by the internal resistance of the fuel cell; For the The actual energy utilization efficiency of a fuel cell; is the input energy of the single fuel cell, and its calculation formula is: ,in is the Gibbs free energy of the reaction; Calculate the distribution weight difference of a single fuel cell , the calculation formula is: .

9. A method for balancing fuel cell life and performance as claimed in claim 8, characterized in that: The updated load distribution weight The constraints are: ; Excess load distribution The constraints are: .

Images