High-power fuel cell hydrogen fault diagnosis method and system

Abstract

The invention provides a high-power fuel cell hydrogen fault diagnosis method and system, and the method comprises a fault diagnosis classification method based on an improved artificial bee colony algorithm optimization support vector machine, and improves the global search capability of a conventional artificial bee colony algorithm through introducing a Levy flight strategy. Then a penalty factor C and a kernel function parameter g of a support vector machine are optimized through an improved artificial bee colony algorithm, an optimal support vector machine classifier model is constructed, and finally fault diagnosis classification is performed on the fuel cell hydrogen system by using the optimized support vector machine classifier model. According to the invention, the fault of the high-power fuel cell hydrogen system can be effectively monitored.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cell control, and particularly to a high-power fuel cell hydrogen fault diagnosis method and system. Background Art

[0002] In recent years, fuel cell technology with characteristics such as high current density, high efficiency, high energy conversion efficiency, low noise, high reliability, non-corrosiveness, and low emissions has received increasing attention. Fuel cells are considered to be a very effective alternative to fossil fuel energy systems, mainly because they are very environmentally friendly and do not emit harmful substances. A fuel cell is an electrochemical device that generates electricity using hydrogen and oxygen. The biggest difference between it and a battery is that they do not need to be charged and can operate quietly and stably. However, a fuel cell system is a complex non-linear and strongly coupled system with a complex structure. Under operating conditions with variable working conditions and harsh environments, various faults are likely to occur.

[0003] With the rapid development of technologies such as automation and computers and the continuous maturity of fault diagnosis theories, diagnosis strategies based on the fault detection and isolation framework have gradually received attention from scientific researchers in various countries and have gradually been applied to the fault diagnosis of actual fuel cell systems. To eliminate or reduce the harm caused by faults to the operation of the actual system, effective diagnosis technologies are studied for different types of fault situations to achieve real-time monitoring of the health status of the fuel cell system and timely isolation of faults, which is of great significance for ensuring the stable performance of the battery, delaying battery aging, and expanding the commercial development of fuel cells. Summary of the Invention

[0004] The present invention proposes a high-power fuel cell hydrogen fault diagnosis method and system, which can effectively monitor the faults of the hydrogen system of high-power fuel cells.

[0005] The present invention adopts the following technical solutions.

[0006] A high-power fuel cell hydrogen fault diagnosis method, the fault diagnosis method includes a fault diagnosis classification method based on an improved artificial bee colony algorithm to optimize the support vector machine. This method improves the global search ability of the traditional artificial bee colony algorithm by introducing the Levy flight strategy, and then optimizes the penalty factor C and kernel function parameter g of the support vector machine through the improved artificial bee colony algorithm to construct an optimal support vector machine classifier model. Finally, the optimized support vector machine classifier model is used to perform fault diagnosis classification on the fuel cell hydrogen system.

[0007] After the optimized support vector machine classifier model performs fault diagnosis classification on the fuel cell hydrogen system, the fuel cell system controller controls the bottle mouth solenoid valve of the hydrogen supply cylinder through a PWM signal to cut off the hydrogen supply to the fuel cell.

[0008] The described fault diagnosis method includes the following steps:

[0009] Step S1: Establish a support vector machine model;

[0010] Step S2: By introducing the Levy flight strategy into the artificial bee colony algorithm, improve the global search ability, optimize the penalty factor C and the kernel function parameter g in the support vector machine model through the improved artificial bee colony algorithm, use the optimized support vector machine model to conduct fault diagnosis classification on the fuel cell hydrogen system, and construct an optimal support vector machine classifier model for fault diagnosis.

[0011] In step S1, the radial basis kernel function selected in the support vector machine model is:

[0012] K(x i ,y j )=exp(-g||x i -y j || 2 )

[0013] The objective function is:

[0014]

[0015] The corresponding optimal classification function is:

[0016]

[0017] In the formula: g is the kernel function parameter; C is the penalty factor; x i ,x j ∈R n ; α i ,α j is the Lagrange multiplier; α * is the optimal solution;

[0018] In step S1, three diagnostic variables are extracted from the hydrogen system of the fuel cell, namely the hydrogen inlet pressure, the hydrogen inlet temperature, and the hydrogen concentration in the hydrogen cylinder. If the changes in the above three parameters affect the output voltage and current of the fuel cell, then these three variables are selected as fault feature variables, corresponding to three fault types: hydrogen inlet pressure fault, hydrogen inlet temperature fault, and hydrogen concentration fault, and the collected variables are used to establish the original data set for establishing and training the support vector machine model.

[0019] Step S2 specifically includes the following steps:

[0020] Step S21: Initialize the parameters in the artificial bee colony algorithm, including the bee colony size, the number of nectar sources, the maximum number of cycles of the nectar source, the maximum number of iterations, the search range of the penalty factor C, and the search range of the kernel function parameter g. The nectar source is initialized as xij ;

[0021] Step S22: Determine the fitness function in the artificial bee colony algorithm, and the objective function value is the classification accuracy rate:

[0022] In the formula: F i is the fitness value of the i-th group of parameters; f i is the objective function value of the i-th nectar source.

[0023] Step S23: Introduce the Levy formula to obtain the position of the new nectar source, conduct a global search on the current nectar source, calculate the fitness of the new nectar source according to the above formula. If the fitness of the new nectar source is better, replace the original nectar source position, otherwise remain unchanged;

[0024]

[0025] x' ij = x ij + α(x ij - x best )L(α)

[0026] In the formula: α is the characteristic exponent, and its value is 1.5; Γ is the gamma function; x ij ' is the position of the new nectar source.

[0027] Step S24: Determine whether the loop count of a certain nectar source is greater than the maximum loop count. If it is greater, generate a new nectar source according to the following formula:

[0028] x ij = minx ij + rand(0,1)(maxx ij - minx ij )

[0029] Step S25: Record the current optimal nectar source, and determine whether the maximum number of iterations is satisfied. If it is satisfied, go to step S26, otherwise go to step S23;

[0030] Step S26: The obtained global optimal nectar source is the optimal parameters C and g, construct the optimal support vector machine classifier model, and verify the trained support vector machine classifier model through the test set samples.

[0031] The fault diagnosis method further includes step S3, that is: after the fuel cell system controller receives the relevant sensor signals and the fault information obtained in step S2, it controls the bottle mouth solenoid valve through the PWM signal to cut off the hydrogen supply.

[0032] A high-power fuel cell hydrogen fault diagnosis system is used to execute a high-power fuel cell hydrogen fault diagnosis method. The hydrogen fault diagnosis system is integrated at the hydrogen system of the high-power fuel cell. The fuel cell includes multiple stack groups, and each stack group is connected in parallel to achieve power supply. Each stack group contains multiple stack units, and each of the stack units is connected in series one by one.

[0033] The hydrogen system includes two groups of hydrogen cylinders, electric control valves, and manual control valves. The electric control valves are 2 bottle mouth switch valves and 1 proportional solenoid valve, and the manual control valves are 2 bottle mouth needle valves, 1 exhaust valve, and 1 ball valve. For the complex working conditions of high-power fuel cell vehicles during driving, two groups of bottle mouth combined valves are designed at the outlets of the two groups of hydrogen cylinders, which are composed of bottle mouth switch valves, bottle mouth solenoid valves, bottle mouth needle valves, and over-temperature and over-pressure protection devices.

[0034] The hydrogen system also includes multiple various sensors, among which:

[0035] There are 2 hydrogen temperature sensors, which are respectively placed in the two groups of bottle mouth combined valves to detect the hydrogen temperature at the outlet of the hydrogen cylinder;

[0036] There is 1 high-pressure sensor, which is placed in the first group of bottle mouth combined valves to detect the pressure at the outlet of the hydrogen cylinder;

[0037] There is 1 low-pressure sensor, which is placed after the pressure regulator and before the proportional solenoid valve to detect the medium pressure of hydrogen;

[0038] There are 4 hydrogen concentration sensors: two are respectively installed at the mouth and the tail of each of the two groups of hydrogen cylinders;

[0039] There is 1 anode inlet pressure sensor, which is placed after the proportional solenoid valve to detect the hydrogen inlet pressure;

[0040] There is 1 anode inlet temperature sensor, which is placed after the proportional solenoid valve to detect the hydrogen inlet temperature.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. The high-power fuel cell includes multiple stack groups, and each stack group is connected in parallel to achieve power supply. Each stack group contains multiple stack units, and each of the stack units is connected in series one by one, which improves the reliability and stability of the fuel cell system.

[0043] 2. The high-power fuel cell hydrogen system builds a dedicated gas supply pipeline for the two-bottle-group hydrogen system and designs the bottle mouth combined valve structure, which improves the safety of the hydrogen system.

[0044] 3. The high-power fuel cell hydrogen fault diagnosis method extracts three diagnostic variables from the hydrogen system, namely hydrogen inlet pressure, hydrogen inlet temperature, and hydrogen concentration in the hydrogen cylinder. These three variables are selected as model parameters. The changes in the above model parameters will affect the output voltage and current of the fuel cell. The above three variables respectively correspond to three types of faults: hydrogen inlet pressure fault, hydrogen inlet temperature fault, and hydrogen concentration fault. The collected variables are used to establish an original data set, and through normalization processing, a sample data set is established.

[0045] 4. By introducing the Levy flight strategy, the standard artificial bee colony algorithm is improved to enhance the global search ability. Then, the penalty factor C and kernel function parameter g of the support vector machine are optimized through the improved artificial bee colony algorithm to improve the fault diagnosis effect of the support vector machine classifier model.

[0046] 5. After receiving relevant sensor signals and fault information, the fuel cell system controller can control the solenoid valve at the bottle mouth through a PWM signal to cut off the hydrogen supply. Brief Description of the Drawings

[0047] The present invention will be further described in detail below in conjunction with the drawings and specific embodiments:

[0048] Attached Figure 1 is a schematic diagram of the high-power fuel cell hydrogen system of the present invention;

[0049] Attached Figure 2 is a schematic diagram of the fault diagnosis process of the high-power fuel cell hydrogen fault diagnosis method of the present invention;

[0050] Attached Figure 3 is a schematic diagram of the fault diagnosis result of a specific embodiment of the present invention. Detailed Description of the Embodiments

[0051] The following describes the specific embodiments of the present invention with reference to the drawings. To fully understand the present invention, many details are described below. However, for those skilled in the art, the present invention can be implemented without these details. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present application and should not be construed as a limitation of the present application.

[0052] As shown in the figure, a high-power fuel cell hydrogen fault diagnosis method, the fault diagnosis method includes a fault diagnosis classification method based on an improved artificial bee colony algorithm to optimize the support vector machine. This method improves the global search ability of the traditional artificial bee colony algorithm by introducing the Levy flight strategy, and then optimizes the penalty factor C and kernel function parameter g of the support vector machine through the improved artificial bee colony algorithm to construct an optimal support vector machine classifier model. Finally, the optimized support vector machine classifier model is used to perform fault diagnosis and classification on the fuel cell hydrogen system.

[0053] After the optimized support vector machine classifier model classifies the faults of the fuel cell hydrogen system, the fuel cell system controller controls the solenoid valve at the bottle mouth of the hydrogen cylinder for gas supply through a PWM signal to cut off the hydrogen supply of the fuel cell.

[0054] The fault diagnosis method includes the following steps:

[0055] Step S1: Establish a support vector machine model;

[0056] Step S2: By introducing the Levy flight strategy into the artificial bee colony algorithm to improve the global search ability, optimize the penalty factor C and the kernel function parameter g in the support vector machine model through the improved artificial bee colony algorithm, use the optimized support vector machine model to classify the faults of the fuel cell hydrogen system, and construct an optimal support vector machine classifier model for fault diagnosis.

[0057] In step S1, the radial basis kernel function selected in the support vector machine model is:

[0058] K(x i ,y j )=exp(-g||x i -y j || 2 )

[0059] The objective function is:

[0060]

[0061] The corresponding optimal classification function is:

[0062]

[0063] In the formula: g is the kernel function parameter; C is the penalty factor; x i ,x j ∈R n ; α i ,α j is the Lagrange multiplier; α * is the optimal solution;

[0064] In step S1, three diagnostic variables are extracted from the hydrogen system of the fuel cell, namely the hydrogen inlet pressure, the hydrogen inlet temperature, and the hydrogen concentration in the hydrogen cylinder. If the changes in the above three parameters affect the output voltage and current of the fuel cell, these three variables are selected as fault feature variables, corresponding to three fault types: hydrogen inlet pressure fault, hydrogen inlet temperature fault, and hydrogen concentration fault, and the collected variables are used to establish the original data set for establishing and training the support vector machine model.

[0065] Step S2 specifically includes the following steps:

[0066] Step S21: Initialize the parameters in the artificial bee colony algorithm, including the bee colony size, the number of nectar sources, the maximum number of cycles of the nectar source, the maximum number of iterations, the search range of the penalty factor C, and the search range of the kernel function parameter g. The nectar source is initialized as x ij ;

[0067] Step S22: Determine the fitness function in the artificial bee colony algorithm. The objective function value is the classification accuracy rate:

[0068] In the formula: F i is the fitness value of the i-th group of parameters; f i is the objective function value of the i-th nectar source.

[0069] Step S23: Introduce the Levy formula to obtain the position of the new nectar source, conduct a global search for the current nectar source, calculate the fitness of the new nectar source according to the above formula. If the fitness of the new nectar source is better, replace the original nectar source position; otherwise, remain unchanged;

[0070]

[0071] x′ ij =x ij +α(x ij -x best )L(α)

[0072] In the formula: α is the characteristic exponent, with a value of 1.5; Γ is the gamma function; x ij ′ is the position of the new nectar source.

[0073] Step S24: Determine whether the number of cycles of a certain nectar source is greater than the maximum number of cycles. If it is greater, generate a new nectar source according to the following formula:

[0074] x ij =minx ij +rand(0,1)(maxx ij -minx ij )

[0075] Step S25: Record the current optimal nectar source, and determine whether the maximum number of iterations is reached. If it is satisfied, go to Step S26; otherwise, go to Step S23;

[0076] Step S26: The obtained global optimal nectar source is the optimal parameters C and g. Construct the optimal support vector machine classifier model, and verify the trained support vector machine classifier model through the test set samples.

[0077] The fault diagnosis method further includes step S3, that is: after the fuel cell system controller receives the relevant sensor signals and the fault information obtained in step S2, it controls the bottle mouth solenoid valve through a PWM signal to cut off the hydrogen supply.

[0078] A high-power fuel cell hydrogen fault diagnosis system is used to execute a high-power fuel cell hydrogen fault diagnosis method. The hydrogen fault diagnosis system is integrated at the hydrogen system of the high-power fuel cell. The battery includes a plurality of stack groups, and each stack group is connected in parallel to achieve power supply. Each stack group includes a plurality of stack units, and each of the stack units is connected in series one by one.

[0079] The hydrogen system includes two groups of hydrogen cylinders, electric control valves, and manual control valves. The electric control valves are two bottle mouth switch valves and one proportional solenoid valve, and the manual control valves are two bottle mouth needle valves, one evacuation valve, and one ball valve; for the complex working conditions of high-power fuel cell vehicles during driving, two groups of bottle mouth combination valves are designed at the outlets of the two groups of hydrogen cylinders, which are composed of a bottle mouth switch valve, a bottle mouth solenoid valve, a bottle mouth needle valve, and an over-temperature and over-pressure protection device.

[0080] The hydrogen system further includes a plurality of various sensors, among which:

[0081] There are two hydrogen temperature sensors, which are respectively placed in the two groups of bottle mouth combination valves to detect the hydrogen temperature at the outlet of the hydrogen cylinder;

[0082] There is one high-pressure sensor, which is placed in the first group of bottle mouth combination valves to detect the pressure at the outlet of the hydrogen cylinder;

[0083] There is one low-pressure sensor, which is placed behind the pressure regulator and in front of the proportional solenoid valve to detect the medium pressure of hydrogen;

[0084] There are four hydrogen concentration sensors: two are respectively installed at the bottle mouth and the bottle tail of the two groups of hydrogen cylinders;

[0085] There is one anode inlet pressure sensor, which is placed behind the proportional solenoid valve to detect the hydrogen inlet pressure;

[0086] There is one anode inlet temperature sensor, which is placed behind the proportional solenoid valve to detect the hydrogen inlet temperature.

[0087] Embodiment:

[0088] This example provides a high-power fuel cell hydrogen fault diagnosis method, which can effectively monitor the faults of the high-power fuel cell hydrogen system.

[0089] Figure 2 The following is the fault diagnosis flowchart of a high-power fuel cell hydrogen fault diagnosis method in this example, which specifically includes the following steps:

[0090] Step S1: Establish a support vector machine model;

[0091] Step S2: By introducing the Levy flight strategy into the artificial bee colony algorithm, the global search ability is improved. The penalty factor C and the kernel function parameter g in the support vector machine model are optimized by the improved artificial bee colony algorithm, and the optimized support vector machine model is used to perform fault diagnosis and classification on the fuel cell hydrogen system.

[0092] In this embodiment, in step S1, the radial basis kernel function is selected in the support vector machine model:

[0093] K(x i ,y j )=exp(-g||x i -y j || 2 )

[0094] The objective function is

[0095]

[0096] The corresponding optimal classification function:

[0097]

[0098] In the formula: g is the kernel function parameter; C is the penalty factor; x i ,x j ∈R n ; α i ,α j is the Lagrange multiplier; α * is the optimal solution;

[0099] In this embodiment, step S2 specifically includes the following steps:

[0100] Step S21: Initialize the parameters in the artificial bee colony algorithm, mainly including the bee colony size, the number of nectar sources, the maximum number of cycles of the nectar source, the maximum number of iterations, the search range of the penalty factor C, and the search range of the kernel function parameter g. The nectar source is initialized as x ij .

[0101] Step S22: Determine the fitness function in the artificial bee colony algorithm. The objective function value is the classification accuracy:

[0102]

[0103] In the formula: F i is the fitness value of the i-th group of parameters; f i is the objective function value of the i-th nectar source.

[0104] Step S23: Introduce the Levy formula to obtain the position of the new nectar source, conduct a global search for the current nectar source, calculate the fitness of the new nectar source according to the above formula. If the fitness of the new nectar source is better, replace the original nectar source position; otherwise, keep it unchanged.

[0105]

[0106] x′ ij = x ij + α(x ij - x best )L(α)

[0107] In the formula: α is the characteristic exponent, with a value of 1.5; Γ is the gamma function; x ij ′ is the position of the new nectar source.

[0108] Step S24: Determine whether the loop count of a certain nectar source is greater than the maximum loop count. If it is greater, generate a new nectar source according to the following formula:

[0109] x ij = min x ij + rand(0,1)(max x ij - min x ij )

[0110] Step S25: Record the current optimal nectar source, and determine whether the maximum number of iterations is reached. If it is satisfied, go to Step S26; otherwise, go to Step S23.

[0111] Step S26: The obtained global optimal nectar source is the optimal parameters C and g. Construct the optimal support vector machine classifier model, and verify the trained support vector machine classifier model through the test set samples.

[0112] The fault diagnosis classification simulation process of the specific embodiment of the present invention is as follows. Extract three diagnostic variables from the hydrogen system, namely hydrogen inlet pressure, hydrogen inlet temperature, and hydrogen concentration in the hydrogen cylinder. The changes in the above parameters will affect the output voltage and current of the fuel cell. Select these three variables as fault characteristic variables, corresponding to three fault types: hydrogen inlet pressure fault, hydrogen inlet temperature fault, and hydrogen concentration fault. Establish an original data set for the collected variables. In this embodiment, 360 groups of sample data are collected. Divide the sample data into three categories according to the three fault types, with 120 groups of data in the first category, 120 groups of data in the second category, and 120 groups of data in the third category. Normalize each group of data to establish a sample data set, and then divide the data of each category into a training set and a test set according to a 2:1 cross-division.

[0113] By introducing the Levy flight strategy, the standard artificial bee colony algorithm is improved to enhance the global search ability of the algorithm. In the improved artificial bee colony algorithm, the number of bee colonies is 20, the number of nectar sources is 10, the maximum number of searches is 50, and the maximum number of iterations is 100. In the support vector machine model, the search range of the penalty factor C is 0.01 - 40, and the search range of the kernel function parameter g is 0.01 - 40. The penalty factor C and the kernel function parameter g in the support vector machine are optimized by the improved artificial bee colony algorithm to construct an optimal support vector machine classifier model. The model is trained using the optimal penalty factor C and kernel function parameter g, and the trained support vector machine model is verified through the test set samples, such as Figure 3 is the fault diagnosis result of the specific embodiment of the present invention. The actual classification and predicted classification diagrams of the test set are obtained, and the classification accuracy of the test set is 99.2%. At the same time, after the fuel cell system controller receives relevant sensor signals and fault information, it controls the bottle mouth solenoid valve through the PWM signal to cut off the hydrogen supply.

[0114] This application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the invention of this application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1 one process or multiple processes and / or blocks Figure 1 a device for the functions specified in one block or multiple blocks.

[0115] As described above, it is not to limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent inventions with equivalent changes. However, any simple modification, equivalent change, and modification made without departing from the technical solution content of the present invention and based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A high-power fuel cell hydrogen fault diagnosis method, characterized in that: The fault diagnosis method includes a fault diagnosis classification method based on an improved artificial bee colony algorithm to optimize a support vector machine. The method improves the global search capability of a traditional artificial bee colony algorithm by introducing a Levy flight strategy, and then optimizes the penalty factor C and kernel function parameter g of the support vector machine by using the improved artificial bee colony algorithm to construct an optimal support vector machine classifier model. Finally, the optimized support vector machine classifier model is used to perform fault diagnosis classification on a fuel cell hydrogen system.

2. A high-power fuel cell hydrogen fault diagnosis method according to claim 1, characterized in that: After the optimized support vector machine classifier model performs fault diagnosis and classification on the fuel cell hydrogen system, the fuel cell system controller controls the bottle mouth solenoid valve of the hydrogen bottle through a PWM signal to cut off the hydrogen supply to the fuel cell.

3. A high-power fuel cell hydrogen fault diagnosis method according to claim 1, characterized in that: The fault diagnosis method comprises the following steps: Step S1: Establish a support vector machine model; Step S2: By introducing the Levy flight strategy into the artificial bee colony algorithm, the global search capability is improved, the penalty factor C and the kernel function parameter g in the support vector machine model are optimized by the improved artificial bee colony algorithm, and the fuel cell hydrogen system is classified by the optimized support vector machine model to construct the optimal support vector machine classifier model for fault diagnosis.

4. A high-power fuel cell hydrogen fault diagnosis method according to claim 3, characterized in that: In step S1, the radial basis kernel function selected in the support vector machine model is: K(x i ,y j )=exp(-g||x i -y j || 2 ) The objective function is: The corresponding optimal classification function is: Where: g is the kernel function parameter; C is the penalty factor; x i ,x j ∈R n ; α i ,α j is the Lagrange multiplier; α * is the optimal solution; 5. A high-power fuel cell hydrogen fault diagnosis method according to claim 3, characterized in that: In step S1, three diagnostic variables are extracted from the hydrogen system of the fuel cell, namely, hydrogen inlet pressure, hydrogen inlet temperature and hydrogen concentration in the hydrogen bottle. If the changes in the above three parameters will affect the output voltage and current of the fuel cell, these three variables are selected as fault characteristic variables, corresponding to three types of faults: hydrogen inlet pressure fault, hydrogen inlet temperature fault and hydrogen concentration fault. The collected variables are used to establish the original data set for establishing and training the support vector machine model.

6. A high-power fuel cell hydrogen fault diagnosis method according to claim 4, characterized in that: Step S2 specifically includes the following steps: Step S21: Initialize the parameters in the artificial bee colony algorithm, including the colony size, the number of nectar sources, the maximum number of nectar source cycles, the maximum number of iterations, the search range of the penalty factor C, and the search range of the kernel function parameter g. The nectar source is initialized to x ij ; Step S22: Determine the fitness function in the artificial bee colony algorithm, and the objective function value is the classification accuracy: Where: F i is the fitness value of the i-th group of parameters; f i is the objective function value of the i-th nectar source. Step S23: Introduce the Levy formula to obtain the new nectar source position, perform a global search on the current nectar source, and calculate the fitness of the new nectar source according to the above formula. If the fitness of the new nectar source is better, replace the original nectar source position, otherwise it remains unchanged; x' ij =x ij +α(x ij -x best )L(a) Where: α is the characteristic exponent, which is 1.5; Γ is the gamma function; x ij ′ is the location of the new nectar source. Step S24: Determine whether the number of cycles of a certain nectar source is greater than the maximum number of cycles. If so, generate a new nectar source according to the following formula: x ij =minx ij +rand(0,1)(maxx ij -minx ij ) Step S25: record the current optimal nectar source and determine whether the maximum number of iterations has been met. If so, go to step S26; otherwise, go to step S23. Step S26: The obtained global optimal nectar source is the optimal parameters C and g, and the optimal support vector machine classifier model is constructed. The trained support vector machine classifier model is verified through the test set samples.

7. A high-power fuel cell hydrogen fault diagnosis method according to claim 3, characterized in that: The fault diagnosis method further includes step S3, namely: after the fuel cell system controller receives the relevant sensor signal and the fault information obtained in step S2, it controls the bottle mouth solenoid valve through the PWM signal to cut off the hydrogen supply.

8. A high-power fuel cell hydrogen fault diagnosis system, used to perform a high-power fuel cell hydrogen fault diagnosis method, characterized in that: The hydrogen fault diagnosis system is integrated in the hydrogen system of a high-power fuel cell. The battery includes a plurality of stack groups, each of which is connected in parallel to realize power supply. Each of the stack groups includes a plurality of stack units, and each of the stack units is connected in series one by one.

9. A high-power fuel cell hydrogen fault diagnosis system according to claim 8, characterized in that: The hydrogen system includes two groups of hydrogen cylinders, electric control valves and manual control valves. The electric control valves are two bottle mouth switch valves and one proportional solenoid valve. The manual control valves are two bottle mouth needle valves, one drain valve and one ball valve. In view of the complex operating conditions of high-power fuel cell vehicles, two sets of bottle mouth combination valves are designed at the outlets of the two sets of hydrogen bottles, which are composed of bottle mouth switch valves, bottle mouth solenoid valves, bottle mouth needle valves and over-temperature and over-pressure protection devices.

10. A high-power fuel cell hydrogen fault diagnosis system according to claim 8, characterized in that: The hydrogen system also includes a number of various sensors, including: Two hydrogen temperature sensors are placed in two groups of bottle mouth combination valves to detect the hydrogen temperature at the outlet of the hydrogen bottle; One high-pressure sensor is placed in the first group of bottle mouth combination valves to detect the pressure at the outlet of the hydrogen bottle; One low-pressure sensor, placed after the pressure regulator and before the proportional solenoid valve, is used to detect the medium pressure of hydrogen; 4 hydrogen concentration sensors: two are installed at the mouth and tail of the two sets of hydrogen cylinders respectively; One anode inlet pressure sensor, placed after the proportional solenoid valve, used to detect the hydrogen inlet pressure; One anode inlet temperature sensor is placed after the proportional solenoid valve to detect the hydrogen inlet temperature.

Images