An energy management method based on the health state of a fuel cell

Abstract

This patent relates to an energy management method based on the health state of a fuel cell. The method includes: 1. Fuel cell health state estimation: Estimate the state of health (SOH) of the fuel cell using an extreme learning machine optimized by a particle swarm algorithm based on the collected parameter data. 2. Update of the fuel cell system efficiency curve: Update the efficiency curve of the fuel cell system according to the estimated SOH value to accurately reflect the performance degradation of the fuel cell. 3. Calculation of the energy distribution scheme: Combine the external power demand and the updated efficiency curve, and use an optimization algorithm to determine the output power schemes of the fuel cell system and the lithium battery, considering the hard constraint conditions. This method incorporates the health state of the fuel cell into the energy management strategy to optimize the power distribution of the multi-source system, improve the comprehensive performance and efficiency of the system, and the application scope covers fuel cell hybrid systems, promoting the application and development of clean energy technologies in the transportation field.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cell system control, and particularly to an energy management method based on the health state of a fuel cell. Background Art

[0002] With the increasing global emphasis on environmental protection and energy sustainability, fuel cell hybrid systems have received significant attention in the field of new energy vehicles due to their high energy efficiency and environmental friendliness. By combining fuel cells with other energy storage devices (such as batteries, supercapacitors) or traditional energy sources (such as internal combustion engines), such systems form an efficient and reliable power supply solution to meet the dual requirements of modern transportation for energy efficiency and environmental protection.

[0003] However, while fuel cell hybrid systems offer unparalleled environmental advantages, they also face significant challenges in real-time energy management and system optimization. One of the most critical issues is how to effectively manage and distribute the energy generated by various energy components to ensure that the vehicle exhibits optimal performance and efficiency under different operating conditions. In addition, the performance degradation problems of fuel cells themselves, including but not limited to electrode aging, proton exchange membrane damage, and reduced catalyst activity, further exacerbate the complexity of energy management. These degradation factors not only affect the output performance of fuel cells but may also lead to an overall decrease in system efficiency, thereby affecting the vehicle's endurance and lifespan.

[0004] Against this background, it is crucial to develop a new method that can effectively grasp the performance degradation of fuel cell stacks and dynamically adjust the power distribution between multi-source systems. This method needs to be able to accurately evaluate the health state of fuel cells and accordingly adjust the energy management strategy to maximize the overall performance and efficiency of the power system. Summary of the Invention

[0005] Object of the Invention: The object of the present invention is to provide a new method for efficiently and dynamically adjusting the power distribution between multi-source systems. This method aims to address the energy management challenges existing in current fuel cell hybrid systems. The method of the present invention focuses on accurately predicting the state of health (SOH) of fuel cells through advanced health state monitoring and prediction methods and optimizing the energy management strategy accordingly. By dynamically updating the performance curve of the fuel cell stack, a more reliable and efficient fuel cell hybrid system can be achieved, promoting the application and development of clean energy technologies in the transportation field.

[0006] Technical Solution: The energy management method based on the health state of a fuel cell according to the present invention includes the following steps:

[0007] (1) Fuel cell state of health estimation: Based on the collected parameter data, use the extreme learning machine optimized by the particle swarm algorithm to estimate the state of health and SOH value of the fuel cell;

[0008] (2) Fuel cell system efficiency curve update: Update the efficiency curve of the fuel cell system according to the estimated SOH value to accurately reflect the performance degradation of the fuel cell;

[0009] (3) Energy allocation scheme calculation: Combine the external power demand and the updated efficiency curve, and use the optimization algorithm to determine the output power schemes of the fuel cell system and the lithium battery.

[0010] Furthermore, in step (1), continuously collect several parameter data that can characterize the state of health of the fuel cell, build and train an extreme learning machine optimized by the particle swarm algorithm, and obtain the estimated value of the SOH of the fuel cell at the current moment.

[0011] Furthermore, after the weights and biases between the input layer and the hidden layer of the extreme learning machine are randomly assigned, the values of the weights and biases are optimized by the particle swarm algorithm, and the activation function of the extreme learning machine is the sin function.

[0012] Furthermore, the structure of the extreme learning machine is:

[0013] Hβ = O,

[0014]

[0015] where, w i is the weight vector between the input layer and the hidden layer, b i is the bias between the input layer and the hidden layer, β i is the weight vector between the hidden layer and the output layer; g(x) is the activation function, X is the input value, and o is the output value.

[0016] Furthermore, the parameters used in the training of the extreme learning machine include: the output current of the fuel cell stack, the output voltage of the fuel cell stack, the temperature of the air at the inlet of the fuel cell stack, the pressure of the air at the inlet of the fuel cell stack, the temperature of the hydrogen at the inlet of the fuel cell stack, the pressure of the hydrogen at the inlet of the fuel cell stack, and the temperature data of the cooling water at the outlet of the fuel cell stack.

[0017] Furthermore, the end of life of the fuel cell stack is the time point when the average single-cell voltage of the fuel cell decays to 90% of the initial value under the rated current. The SOH of the fuel cell stack at the initial moment is 1, and the SOH when the voltage decays to 90% of the initial value is 0.

[0018] Further, in step (2), the efficiency curve of the fuel cell system at the initial moment is collected, and the estimated value of SOH is substituted into the efficiency function of the fuel cell system to update the efficiency curve of the fuel cell system.

[0019] Further, the efficiency curve of the fuel cell system is calculated by substituting the predicted SOH value into the following formula:

[0020]

[0021] In the formula, η sys is the comprehensive efficiency of the fuel cell system, η air is the comprehensive efficiency of the fuel cell system, which is composed of the air utilization efficiency, η conv is the energy conversion efficiency, η elec is the electrical efficiency; E theory is the theoretical output voltage of the fuel cell stack, U fc，ini is the voltage of the fuel cell stack at the initial moment under a certain current state ; η Dc / Dc is the efficiency of DC / DC.

[0022] Further, the objective function of the energy management strategy of the fuel cell hybrid system takes into account the economy and durability of the hybrid system:

[0023]

[0024] In the formula, C hydrogen is the unit price of hydrogen, is the hydrogen consumed by the fuel cell stack, is the equivalent hydrogen consumption of the lithium battery, C bat is the price of the lithium battery, d bat is the relative attenuation degree of the lithium battery, C fc is the price of the fuel cell system, d fc is the relative attenuation degree of the fuel cell system, d fc = 0.1×(1 - SOH).

[0025] Further, in step (3), according to the external power demand at the current moment and the updated efficiency curve of the fuel cell system, etc., by setting hard constraints on the outputs of the fuel cell system and the lithium battery, and using an optimization algorithm to optimize the objective function, the output power schemes of the fuel cell system and the lithium battery at the current moment are obtained.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] 1. The method of the present invention proposes a method for predicting the remaining service life of a fuel cell based on an extreme learning machine (ELM) neural network optimized by a particle swarm optimization algorithm (PSO). This method is efficient, has a small computational burden, and strong adaptability. By training dynamically collected data, it can predict the SOH of the fuel cell stack in real time and be used in the subsequent adjustment and optimization process of the energy management strategy.

[0028] 2. The purpose of the method of the present invention is to provide a method for accurately evaluating the health state of a fuel cell stack, and an energy management method that updates the efficiency curve of the fuel cell stack based on this health state and comprehensively considers the economy and durability of the fuel cell hybrid system.

[0029] 3. The method of the present invention considers the influence of the attenuation of the fuel cell stack during its life cycle on the efficiency of the fuel cell system. By analyzing the conversion process from the energy contained in the hydrogen entering the fuel cell stack to the available energy actually output to the external system, a calculation method for the efficiency of the fuel cell system changing with the health state of the fuel cell stack is given. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flowchart in the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0031] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be further described below.

[0032] The energy management method based on the health state of the fuel cell proposed by the present invention includes the following steps:

[0033] S1: Obtain the load change situation of the fuel cell vehicle in the driving record, and output the external power demand P req to the power system controller of the whole vehicle;

[0034] S2: Before the fuel cell stack operates, measure the polarization curve at the initial moment to obtain the parameters in the empirical model; continuously collect the output current I of the fuel cell stack fc , the output voltage U of the fuel cell stack fc , the temperature t of the air at the inlet of the fuel cell stack air，in , the pressure p of the air at the inlet of the fuel cell stack air，in , the temperature t of the hydrogen at the inlet of the fuel cell stack hydrogen，in , the pressure p of the hydrogen at the inlet of the fuel cell stack hydrogen，in and the temperature t of the cooling water at the outlet of the fuel cell stack water，out data;

[0035] S3: Build an extreme learning machine neural network; the structure of the extreme learning machine is:

[0036] Hβ = O,

[0037]

[0038] where w i is the weight vector between the input layer and the hidden layer, b i is the bias between the input layer and the hidden layer, and β i is the weight vector between the hidden layer and the output layer; g(x) is the activation function, X is the input value, and O is the output value;

[0039] Use the function g(x) = sin x as the activation function, and use PSO to optimize the weights w i and the bias b i between the input layer and the hidden layer of the neural network; Import a number of collected parameter data to train the neural network. The purpose of training is to minimize the mean square error E(x) of the extreme learning machine, that is:

[0040]

[0041] S4: In the state of the trained PSO-ELM neural network, based on the collected real-time output current I fc of the fuel cell stack, the output voltage U fc of the fuel cell stack, the temperature t air，in of the air at the inlet of the fuel cell stack, the pressure p air，in of the air at the inlet of the fuel cell stack, the temperature t hydrogen，in of the hydrogen at the inlet of the fuel cell stack, the pressure p hydrogen，in of the hydrogen at the inlet of the fuel cell stack, and the temperature t water，out of the cooling water at the outlet of the fuel cell stack, predict the SOH of the fuel cell stack at the current moment;

[0042] S5: Use the obtained SOH of the fuel cell stack to update and calculate the comprehensive efficiency η sys of the fuel cell system; The comprehensive efficiency of the fuel cell system consists of the air utilization efficiency η air , the energy conversion efficiency η conv , and the electrical efficiency η elec ;

[0043] The energy conversion efficiency η conv satisfies:

[0044]

[0045] where E theory is the theoretical output voltage of the fuel cell stack, and U fc，ini is the voltage of the fuel cell stack at the initial moment under a certain current state ;

[0046] Electrical efficiency η elec Satisfies:

[0047]

[0048] Wherein, P output Is the actual power that can be output outward by the fuel cell system, η DC / DC Is the efficiency of the DC / DC, P aux Is the power consumed by other auxiliary components;

[0049] Furthermore,

[0050]

[0051] By calculating the efficiency values of the fuel cell system under different current states, a new efficiency-current curve of the fuel cell system can be obtained; input this curve into the vehicle controller;

[0052] S6: Considering the economy and durability of the fuel cell hybrid system, set the objective function f as:

[0053]

[0054] Wherein, C hydrogen Is the unit price of hydrogen, Is the hydrogen consumed by the fuel cell stack, Is the equivalent hydrogen consumption of the lithium battery, C bat Is the price of the lithium battery, d bat Is the relative attenuation degree of the lithium battery, C fc Is the price of the fuel cell system, d fc Is the relative attenuation degree of the fuel cell system, d fc = 0.1×(1 - SOH);

[0055] Here, Is the hydrogen consumed by the fuel cell stack and is related to the output current I of the fuel cell system fc Satisfies n fc Is the number of fuel cells in the fuel cell stack; d bat Is related to the output current I of the lithium battery bat The internal resistance R of the lithium battery, bat The SOC of the lithium battery, the usage time t, the ambient temperature T, etc., and satisfies The remaining values in the formula are fitting parameters; d fc Is related to the output current I of the fuel cell fc Obtain the attenuation rate of the fuel cell under four working conditions from the fuel cell test database; calculate the total attenuation degree d according to the time ratio and attenuation rate of the four working conditionsfc ;

[0056] S7: Set the hard constraints on the output power of the fuel cell stack, the change rate of the output power of the fuel cell stack, the output power of the lithium battery, and the state of charge (SOC) of the lithium battery, and use the particle swarm optimization algorithm to perform constrained optimization search on the objective function;

[0057] S8: The vehicle controller outputs the output power demand P fc of the fuel cell system obtained by optimization and the output power demand P bat result to the fuel cell system controller and the lithium battery controller; complete the output allocation at the current moment.

[0058] The energy management method for the fuel cell hybrid system can dynamically update and adjust the power distribution among multi-source systems based on the health state of the fuel cell stack, which helps to reduce the usage cost of the fuel cell hybrid system.

[0059] The above are only the preferred embodiments of the present invention and do not impose any limitation on the present invention. Any person skilled in the art within the technical field, without departing from the technical solution of the present invention, makes any form of equivalent substitution or modification and other changes to the technical solution and technical content disclosed by the present invention, all of which belong to the content that does not depart from the technical solution of the present invention and still fall within the protection scope of the present invention.

Claims

1. An energy management method based on the health status of a fuel cell, characterized in that: The following steps are involved: (1) Fuel cell health status estimation: Based on the collected parameter data, the health status and SOH value of the fuel cell are estimated using an extreme learning machine optimized by a particle swarm algorithm; (2) Fuel cell system efficiency curve update: The efficiency curve of the fuel cell system is updated according to the estimated SOH value to accurately reflect the fuel cell performance degradation; (3) Energy allocation scheme calculation: Combine external power demand and updated efficiency curves to determine the output power scheme of the fuel cell system and lithium battery using an optimization algorithm; The weights and biases between the input layer and the hidden layer of the extreme learning machine are randomly assigned, and then the values ​​of the weights and biases are optimized by a particle swarm algorithm, and the activation function of the extreme learning machine is a sin function; The structure of the extreme learning machine is: Hβ=O, Among them, w i is the weight vector between the input layer and the hidden layer, b i is the bias between the input layer and the hidden layer, β i is the weight vector between the hidden layer and the output layer; g(x) is the activation function, and X is the input value from X1 to X N , O is the output value O1 to O N ; The parameters used in the extreme learning machine training include: the output current of the fuel cell stack, the output voltage of the fuel cell stack, the temperature of the air at the inlet of the fuel cell stack, the pressure of the air at the inlet of the fuel cell stack, the temperature of the hydrogen at the inlet of the fuel cell stack, the pressure of the hydrogen at the inlet of the fuel cell stack, and the temperature data of the cooling water at the outlet of the fuel cell stack; In step (2), the efficiency curve of the fuel cell system at the initial moment is collected, the estimated value of SOH is substituted into the efficiency function of the fuel cell system, and the efficiency curve of the fuel cell system is updated; The efficiency curve of the fuel cell system is calculated by substituting the predicted SOH value into the following formula: Where η sys is the overall efficiency of the fuel cell system, η air is the air utilization efficiency of the fuel cell system, η conv is the energy conversion efficiency, η elec E is the electrical efficiency; theory is the theoretical output voltage of the fuel cell stack, U fc,ini For a certain current state I f 0 c The voltage of the fuel cell stack at the initial moment; η DC / DC is the efficiency of DC / DC; The objective function of the energy management strategy of the fuel cell hybrid system takes into account the economy and durability of the hybrid system: In the formula, C hydrogen is the unit price of hydrogen, The hydrogen consumed by the fuel cell stack, is the equivalent hydrogen consumption of lithium battery, C bat is the price of lithium battery, d bat is the relative attenuation degree of lithium battery, C fc is the price of the fuel cell system, d fc is the relative attenuation degree of the fuel cell system, d fc =0.1×(1-SOH), here, is the hydrogen consumed by the fuel cell stack and the output current I of the fuel cell system fc Related, satisfying n fc is the number of fuel cells in the fuel cell stack, In step (3), based on the external power demand at the current moment and the efficiency curve of the updated fuel cell system, hard constraints are set on the output of the fuel cell system and the lithium battery, and the objective function is optimized using an optimization algorithm to obtain the output power solution of the fuel cell system and the lithium battery at the current moment.

2. The energy management method based on the health status of a fuel cell according to claim 1, characterized in that: In step (1), a number of parameter data that can be used to characterize the health status of the fuel cell are continuously collected, an extreme learning machine optimized by a particle swarm algorithm is built and trained, and an estimated value of the SOH of the fuel cell at the current moment is obtained.

3. The energy management method based on the health status of a fuel cell according to claim 1, characterized in that: The end of the life of the fuel cell stack is the time point when the average single-cell voltage of the fuel cell decays to 90% of the initial value under rated current. The SOH of the fuel cell stack at the initial moment is 1, and the SOH is 0 when the voltage decays to 90% of the initial value.

Images