PEM electrolyzer accelerated stress testing system and method based on multi-stress coupling

Abstract

The present invention discloses a PEM electrolyzer accelerated stress test system and method based on multi-stress coupling, and relates to the technical field of electrolyzer accelerated test. The steps of the method include: obtaining initial performance and initial proton membrane safety state before the test begins; combining proton membrane surface deformation and electrolyzer hydrogen production efficiency, real-time monitoring of proton membrane safety state; setting an adaptive stress application algorithm, and adjusting the step size and residence time of stress change in real time according to the proton membrane safety state; establishing a life prediction model, statistically analyzing the performance degradation trend under different stress conditions, and training the life prediction model through the performance degradation trend to evaluate the service life of the PEM electrolyzer under different stress conditions. The present invention solves the impact of the proton membrane safety state on the test cost in the accelerated test of the PEM electrolyzer, and realizes the test response when the hydrogen production efficiency drops sharply.

Description

Technical Field

[0001] The invention relates to the technical field of electrolytic cell accelerated testing, and in particular to a PEM electrolytic cell accelerated stress testing system and method based on multi-stress coupling. Background Art

[0002] The performance degradation of PEM electrolyzers is a complex process involving multiple factors such as catalyst failure, proton membrane decomposition, and metal oxide formation. In order to deeply understand the performance changes of PEM electrolyzers under different stress conditions, accelerated stress testing has become an important research method. By simulating extreme operating conditions in actual operation, accelerated stress testing can quickly reveal the failure mechanism of PEM electrolyzers and provide data support for optimizing design and improving system durability.

[0003] For example, the existing Chinese patent application with publication number CN118246197A discloses a method and device for modeling the attenuation of a PEM electrolyzer under random variable power operation, including using a drift Wiener random process to describe the random volatility and cumulative growth law of the performance degradation rate of a PEMWE battery stack under variable power accelerated stress, and establishing an accelerated performance degradation model; based on the accelerated performance degradation model, considering the influence of power change frequency, change amplitude and average operating power on the degradation rate, a multi-stress parameter acceleration model is established; based on the multi-stress parameter acceleration model, the life of the battery stack under different cyclic stresses is estimated; the invention can deduce the life of the battery stack under the influence of factors such as input power change amplitude, change frequency and operating power level.

[0004] Although the above method can deduce the life of the battery stack under different cyclic stresses, the accuracy of long-term predictions remains a challenge. Due to the randomness and nonlinearity of the degradation process of the PEM electrolyzer, there may be great uncertainty in the long-term prediction results. In addition, as the operating time increases, the uncertainty of the physical and chemical reactions inside the PEM electrolyzer and the state of the proton membrane in the electrolyzer will also be affected, further affecting the accuracy of the prediction results, and increasing the time and labor costs of the test conditions, and lacking a test response to a sudden drop in hydrogen production efficiency. To this end, the present invention provides a PEM electrolyzer accelerated stress testing system and method based on multi-stress coupling. Summary of the invention

[0005] The object of the present invention is to provide a PEM electrolyzer accelerated stress testing system and method based on multi-stress coupling to solve the existing problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a PEM electrolyzer accelerated stress testing system based on multi-stress coupling, comprising:

[0007] A test preparation module is used to install a transparent window outside the electrolyzer to allow an external light source and an imaging device to observe the interior through the window, and an image sensor is installed outside the transparent window to collect a proton membrane image in real time, obtain an initial performance and an initial proton membrane safety state before the test starts, and adjust the stress condition of the PEM electrolyzer;

[0008] The proton membrane safety status monitoring module is used to monitor the proton membrane safety status in real time by combining the proton membrane surface deformation and the hydrogen production efficiency of the electrolyzer;

[0009] An adaptive stress application module, used for adjusting the step size and dwell time of stress change in real time according to the safety state of the proton membrane;

[0010] The life prediction model establishment module is used to statistically analyze the performance degradation trend under different stress conditions, and train the life prediction model through the performance degradation trend to evaluate the service life of the PEM electrolyzer under different stress conditions.

[0011] A further improvement of the present invention is that the test preparation module includes an initial performance calculation unit and an initial stress condition setting unit. The initial performance calculation unit is used to obtain the initial performance of the equipment and to obtain the initial proton membrane safety state by extracting the initial spectral image of the proton membrane before the test starts and calculating the initial spectral impact value.

[0012] A further improvement of the present invention is that the test preparation module includes establishing a stress database, and the stress database includes temperature, pressure, current and voltage.

[0013] A further improvement of the present invention is that the proton membrane safety status monitoring module includes collecting the proton membrane surface image at the current moment every time the stress changes, and listing the collected images into a proton membrane status image sequence in the order of collection time, and extracting the proton membrane status value sequence corresponding to the proton membrane status image sequence.

[0014] On the other hand, the present invention provides a PEM electrolyzer accelerated stress testing method based on multi-stress coupling, comprising the following steps:

[0015] S1. Obtain the initial performance and initial proton membrane safety status before the test begins;

[0016] S2, setting an initial stress step length and an initial stress dwell time, and adjusting the stress condition of the PEM electrolyzer according to the initial stress step length and the initial stress dwell time;

[0017] S3, combining the proton membrane surface deformation and the hydrogen production efficiency of the electrolyzer to monitor the safety status of the proton membrane in real time;

[0018] S4, setting an adaptive stress application algorithm to adjust the step size and dwell time of stress change in real time according to the safety state of the proton membrane;

[0019] S5. Establish a life prediction model, count the performance degradation trends under different stress conditions, and train the life prediction model through the performance degradation trends to evaluate the service life of the PEM electrolyzer under different stress conditions.

[0020] A further improvement of the present invention is that the initial performance is obtained by calculating the initial hydrogen production efficiency during the initial residence time, and the initial hydrogen production efficiency is obtained by the ratio of the total amount of hydrogen produced during the initial residence time to the initial residence time.

[0021] A further improvement of the present invention is that the initial proton membrane safety state includes extracting the initial spectral image of the proton membrane before the test starts, and calculating the initial spectral impact value of the initial spectral image through a damage spectrum combination strategy. At this time, the initial proton membrane state value is recorded as 0. The damage spectrum combination strategy is obtained by calculating the ratio of the difference and the sum of the visible light band reflectance and the ultraviolet light band reflectance of the spectral image at the current moment.

[0022] The present invention is further improved in that the monitoring of the safety state of the proton membrane includes real-time acquisition of the proton membrane surface image when the proton membrane is working by the image sensor, firstly, the rough positioning of the proton membrane crack is performed by the Roberts operator to obtain the rough defect edge, then the n-order m-order orthogonal matrix of the pixels in the rough defect edge is calculated, and the rough defect edge is rotated counterclockwise again. The pixel orthogonal matrix within the rough defect edge after rotation is obtained, and the absolute value of the difference between the ratio of the orthogonal matrix before and after rotation and 1 is calculated, which is recorded as the proton membrane loss parameter.

[0023] A further improvement of the present invention is that the monitoring of the safety status of the proton membrane also includes extracting the corresponding spectrum of the rough defect edge, recorded as the first spectral data, graying the first spectral data, setting a grayscale threshold, setting the grayscale value of pixels in the first spectral data whose grayscale value is less than or equal to the grayscale threshold to 0, and extracting pixels in the first spectral data whose grayscale value is greater than the grayscale threshold, recorded as the second spectral data, calculating the spectral influence value of the second spectral data through the damage spectrum combination strategy, extracting the difference between the spectral influence value of the second spectral data and the initial spectral influence value, and calculating the product of the difference and the proton membrane loss parameter, recorded as the proton membrane status value.

[0024] A further improvement of the present invention is that the adaptive stress application algorithm includes defining an objective function , and define the particle dimension as 2, that is, each particle consists of two values, including the step size and residence time of stress change; introduce the particle swarm algorithm, after adjusting the stress condition of the PEM electrolyzer according to the initial stress step size and the initial stress residence time, update the particle velocity and particle position according to the proton membrane state value, and then randomly select the individual optimal position of any particle as the initial global optimal position, and update the individual optimal position and the global optimal position after evaluating the fitness of each particle according to the objective function. When there is a particle whose individual optimal position has a fitness greater than the fitness of the initial global optimal position, update the individual optimal position of the particle as the new global optimal position, and output the stress change step size and residence time corresponding to the particle.

[0025] A further improvement of the present invention is that the objective function is expressed as ,in , and represents the weight factor, , and represents the sub-objective function, Indicates a negative impact on the safety status of the proton membrane, represents the energy consumption of the electrolyzer, represents the maintenance cost, , E represents the minimum energy consumption expectation, C represents the minimum maintenance cost expectation, represents the stress variation step, Indicates the dwell time, represents the proton membrane state value, , and Represents the coefficient.

[0026] The present invention is further improved in that the updating of particle velocity and particle position includes setting the particle position initialization range, the proton membrane safety state threshold and the initial inertia weight, and setting the inertia weight interval and proton membrane safety range , extract the current proton membrane state value in the proton membrane state value sequence, and the proton membrane state value at the previous moment, update the inertia weight to update the particle group speed; when the difference between the current proton membrane state value and the proton membrane state value at the previous moment is less than or equal to the proton membrane safety state threshold, the update formula of the inertia weight is ; When the difference between the current proton membrane state value and the proton membrane state value at the previous moment is greater than the proton membrane safety state threshold, the inertia weight is updated by calculating the ratio of the proton membrane safety state threshold to the difference; the ratio of the proton membrane safety state threshold to the difference between the current proton membrane state value and the proton membrane state value at the previous moment is extracted as the position update coefficient, and when updating the particle position, the position update coefficient is added to update the particle position.

[0027] A further improvement of the present invention is that the life prediction model includes obtaining the corresponding proton membrane state value and performance under different stress conditions, the performance is obtained by calculating the hydrogen production efficiency during the residence time under the stress condition, and the hydrogen production efficiency is obtained by calculating the ratio of the total hydrogen production during the residence time under the stress condition to the residence time under the stress condition; the performance degradation index is calculated by multiplying the corresponding proton membrane state value and performance under different stress conditions, the growth trend of the performance degradation index is analyzed by an exponential decay model, and the performance decay coefficient K is obtained, a support vector machine model is introduced, and a life prediction model is established, the performance degradation index is used as training data, the current stress is used as an input parameter, and the time in the output exponential decay model is the expected life of the PEM electrolyzer.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. The present invention first collects the surface image of the proton membrane in real time, uses the Roberts operator to roughly locate the crack, then calculates the precise defect edge position through the orthogonal matrix, monitors the proton membrane loss parameter by calculating the change of the spectral influence value of the rough defect edge, and calculates the proton membrane state value accordingly, which helps to timely discover the damage of the proton membrane and ensure the safe operation of the proton membrane;

[0030] 2. By defining the objective function and using the particle swarm optimization algorithm, the step size and residence time of stress change are adjusted in real time according to the safety status of the proton membrane. The speed and position of the particles are dynamically adjusted by introducing the proton membrane state value. In particular, the search space is explored more carefully by adjusting the inertia weight. It is possible to optimize according to the actual working state of the proton membrane and thus find the optimal stress conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a framework diagram of the PEM electrolyzer accelerated stress testing system based on multi-stress coupling of the present invention;

[0032] Figure 2 This is an example diagram of a PEM electrolyzer device of the present invention;

[0033] Figure 3 The present invention is a flow chart of the accelerated stress testing method of a PEM electrolyzer based on multi-stress coupling;

[0034] Figure 4 The flow chart of the proton membrane safety status monitoring of the present invention is shown;

[0035] Figure 5 The flowchart of the adaptive stress application algorithm of the present invention is shown.

[0036] Reference numerals: 100, PEM electrolyzer; 200, transparent window; 300, image sensor; 400, proton membrane. DETAILED DESCRIPTION

[0037] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0038] The term "and / or" is only a description of the association relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " generally indicates that the related objects are in an "or" relationship.

[0039] Example 1

[0040] Figure 1 The framework diagram of the PEM electrolyzer accelerated stress test system based on multi-stress coupling disclosed in this embodiment is shown, including:

[0041] Test preparation module, Figure 2 The example diagram of the PEM electrolyzer device of this embodiment is shown; a transparent window 200 is installed on the outside of the PEM electrolyzer 100 to allow an external light source and an imaging device to observe the inside through the transparent window 200, and an image sensor 300 is installed outside the transparent window 200 to collect the image of the proton membrane 400 in real time, obtain the initial performance and the initial proton membrane safety state before the test starts, and adjust the stress condition of the PEM electrolyzer 100; the test preparation module includes an initial performance calculation unit and an initial stress condition setting unit, and the initial performance calculation unit is used to obtain the initial performance of the device and calculate the initial spectrum impact value to obtain the initial proton membrane safety state by extracting the initial spectrum image of the proton membrane 400 before the test starts. The test preparation module includes establishing a stress database, and the stress database includes temperature, pressure, current and voltage.

[0042] The proton membrane safety status monitoring module is used to monitor the proton membrane safety status in real time by combining the surface deformation of the proton membrane 400 and the hydrogen production efficiency of the electrolyzer; the proton membrane safety status monitoring module includes collecting the surface image of the proton membrane 400 at the current moment every time the stress changes, and listing the collected images in the order of collection time into the proton membrane status image sequence, and extracting the proton membrane status value sequence corresponding to the proton membrane status image sequence.

[0043] An adaptive stress application module, used for adjusting the step size and dwell time of stress change in real time according to the safety state of the proton membrane;

[0044] The life prediction model establishment module is used to count the performance degradation trends under different stress conditions, and train the life prediction model through the performance degradation trends to evaluate the service life of the PEM electrolyzer 100 under different stress conditions, so that when the operator inputs the stress state of the PEM electrolyzer 100, the system will output the predicted life according to the model experience.

[0045] Example 2

[0046] Figure 3 The flowchart of the accelerated stress testing method of a PEM electrolyzer based on multi-stress coupling of the present invention is shown. Based on the same inventive concept as Example 1, the present invention provides an accelerated stress testing method of a PEM electrolyzer based on multi-stress coupling, comprising the following steps:

[0047] S1. Obtaining initial performance and initial proton membrane safety status before the test begins; the initial performance is obtained by calculating the initial hydrogen production efficiency within the initial residence time, and the initial hydrogen production efficiency is obtained by the ratio of the total hydrogen production within the initial residence time to the initial residence time.

[0048] The initial proton membrane safety state includes extracting the initial spectral image of the proton membrane 400 before the test begins. Since the proton membrane 400 is more sensitive to the visible light band and the ultraviolet light band, the initial spectral impact value of the initial spectral image is calculated by setting a damage spectrum combination strategy for the visible light band and the ultraviolet light band. At this time, the initial proton membrane state value is recorded as 0. The damage spectrum combination strategy is obtained by calculating the ratio of the difference and the sum of the visible light band reflectance and the ultraviolet light band reflectance of the spectral image at the current moment.

[0049] S2, setting an initial stress step length and an initial stress dwell time, and adjusting the stress condition of the PEM electrolyzer 100 according to the initial stress step length and the initial stress dwell time;

[0050] S3. Since the proton membrane 400 is the core component of the PEM electrolyzer 100, the stress test of the electrolyzer will have a great impact on the proton membrane 400 or even damage it. In order to ensure that the proton membrane 400 remains in a safe state during the test, the surface deformation of the proton membrane 400 and the hydrogen production efficiency of the electrolyzer should be combined during the stress change test to monitor the safety state of the proton membrane in real time;

[0051] Figure 4 The flow chart of the proton membrane safety status monitoring of the present invention is shown. The monitoring of the proton membrane safety status includes real-time acquisition of the proton membrane surface image of the proton membrane 400 when the proton membrane is working by the image sensor 300. First, the crack of the proton membrane is roughly located by the Roberts operator to obtain the rough defect edge. Since the orthogonal matrix can realize spatial transformation while keeping the length and direction relationship of the vector unchanged, the edge position of the defect can be accurately located at the sub-pixel level. This is particularly important for a fine structure such as the proton membrane 400. Therefore, the n-order m-order orthogonal matrix of the pixels in the rough defect edge is calculated, and the rough defect edge is rotated counterclockwise again. The orthogonal matrix of pixels within the rough defect edge after rotation. Since the larger the difference between the ratio of the orthogonal matrix and 1, the greater the change before and after rotation, the absolute value of the difference between the ratio of the orthogonal matrix before and after rotation and 1 is calculated and recorded as the proton membrane loss parameter.

[0052] The monitoring of the safety status of the proton membrane also includes extracting the corresponding spectrum of the rough defect edge, recording it as the first spectrum data, graying the first spectrum data, setting a grayscale threshold, setting the grayscale value of pixels in the first spectrum data whose grayscale value is less than or equal to the grayscale threshold to 0, and extracting pixels in the first spectrum data whose grayscale value is greater than the grayscale threshold, recording it as the second spectrum data, calculating the spectral influence value of the second spectrum data through the damage spectrum combination strategy, extracting the difference between the spectral influence value of the second spectrum data and the initial spectral influence value, and calculating the product of the difference and the proton membrane loss parameter, recording it as the proton membrane status value.

[0053] S4, setting an adaptive stress application algorithm to adjust the step size and dwell time of stress change in real time according to the safety state of the proton membrane;

[0054] Figure 5 The flow chart of the adaptive stress application algorithm of the present invention is shown, and the adaptive stress application algorithm includes defining the objective function , and define the particle dimension as 2, that is, each particle consists of two values, including the step size and residence time of stress change; introduce a particle swarm algorithm, after adjusting the stress condition of the PEM electrolyzer 100 according to the initial stress step size and the initial stress residence time, update the particle velocity and particle position according to the proton membrane state value, and then randomly select the individual optimal position of any particle as the initial global optimal position, and update the individual optimal position and the global optimal position after evaluating the fitness of each particle according to the objective function. When there is a particle whose individual optimal position has a fitness greater than the fitness of the initial global optimal position, update the individual optimal position of the particle as the new global optimal position, and output the stress change step size and residence time corresponding to the particle.

[0055] The objective function is expressed as ,in , and represents the weight factor, , and represents the sub-objective function, Indicates a negative impact on the safety status of the proton membrane, represents the energy consumption of the electrolyzer, represents the maintenance cost, , E represents the minimum energy consumption expectation, C represents the minimum maintenance cost expectation, represents the stress variation step, Indicates the dwell time, represents the proton membrane state value, , and Represents the coefficient.

[0056] The updating of the particle velocity and particle position includes setting the particle position initialization range, for example, the temperature change step size can be from 0 to the maximum allowable change range (for example, 5°C), and the residence time can be from the minimum allowable time to the maximum allowable time (for example, from 1 minute to 10 minutes). The updating of the particle velocity and particle position also includes setting the proton membrane safety state threshold and the initial inertia weight; when the proton membrane state value increases, it means that the proton membrane safety level is low, which means that the proton membrane 400 may be in an unstable state, and it is necessary to find a better temperature change step size and residence time to improve safety. Therefore, the particles should be made more dependent on the guidance of the individual optimal position and the global optimal position. At this time, by setting the proton membrane safety state threshold and the initial inertia weight, and setting the inertia weight interval and proton membrane safety range , extract the current proton membrane state value in the proton membrane state value sequence, and the proton membrane state value at the previous moment, update the inertia weight to update the particle group speed; when the difference between the current proton membrane state value and the proton membrane state value at the previous moment is less than or equal to the proton membrane safety state threshold, the update formula of the inertia weight is ; At this time, when the proton membrane state value near When , when the proton membrane state value near When , when the difference between the current proton membrane state value and the proton membrane state value at the previous moment is greater than the proton membrane safety state threshold, the inertia weight is updated by calculating the ratio of the proton membrane safety state threshold to the difference, that is, multiplying the ratio by the minimum inertia weight;

[0057] In order to make the particles more inclined to the individual optimal position and the global optimal position, the ratio of the proton membrane safety state threshold to the difference between the current proton membrane state value and the proton membrane state value at the previous moment is extracted as the position update coefficient. When updating the particle position, the position update coefficient is added to update the particle position. It should be noted that each particle can remember the optimal position it has found (individual optimal, pBest), which is the best solution found by the particle itself during the search process. At the same time, the optimal position found in the entire particle swarm is called the global optimal (gBest), which is the best solution found by all particles together. During the operation of the algorithm, the particles update their speed and position by tracking these two "extreme values": the individual extreme value (pBest) is the best position experienced by the particle itself during the search process, and the global extreme value (gBest) is the best position currently found by the entire particle swarm. In this way, the particle swarm algorithm can find the global optimal solution in the search space. This guidance through individual optimality and global optimality enables the particle swarm to effectively approach the optimal solution of the problem, reflecting the biological principle of the "synchronization effect".

[0058] S5. Establish a life prediction model, count the performance degradation trends under different stress conditions, and train the life prediction model through the performance degradation trends to evaluate the service life of the PEM electrolyzer 100 under different stress conditions. The life prediction model includes obtaining the corresponding proton membrane state value and performance under different stress conditions, the performance is obtained by calculating the hydrogen production efficiency during the residence time under the stress condition, and the hydrogen production efficiency is obtained by calculating the ratio of the total hydrogen production during the residence time under the stress condition to the residence time under the stress condition; calculate the performance degradation index by multiplying the proton membrane state value and performance under different stress conditions, analyze the growth trend of the performance degradation index through the exponential decay model, and obtain the performance decay coefficient K, introduce the support vector machine model, establish the life prediction model, use the performance degradation index as training data, use the current stress as an input parameter, and output the time in the exponential decay model as the expected life of the PEM electrolyzer 100.

[0059] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0060] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. 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 a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0061] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0062] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0063] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. PEM electrolyzer accelerated stress test system based on multi-stress coupling, Features: include: A test preparation module is used to install a transparent window outside the electrolyzer to allow an external light source and an imaging device to observe the interior through the window, and an image sensor is installed outside the transparent window to collect a proton membrane image in real time, obtain an initial performance and an initial proton membrane safety state before the test starts, and adjust the stress condition of the PEM electrolyzer; The proton membrane safety status monitoring module is used to monitor the proton membrane safety status in real time by combining the proton membrane surface deformation and the hydrogen production efficiency of the electrolyzer; An adaptive stress application module, used for adjusting the step size and dwell time of stress change in real time according to the safety state of the proton membrane; The life prediction model building module is used to count the performance degradation trends under different stress conditions, and train the life prediction model through the performance degradation trends to evaluate the service life of the PEM electrolyzer under different stress conditions; The test preparation module includes an initial performance calculation unit and an initial stress condition setting unit. The initial performance calculation unit is used to obtain the initial performance of the device and to obtain the initial proton membrane safety state by extracting the initial spectrum image of the proton membrane before the test starts and calculating the initial spectrum impact value; The initial performance is obtained by calculating the initial hydrogen production efficiency during the initial residence time, and the initial hydrogen production efficiency is obtained by the ratio of the total amount of hydrogen produced during the initial residence time to the initial residence time.

2. The PEM electrolyzer accelerated stress testing system based on multi-stress coupling according to claim 1, Features: The test preparation module includes establishing a stress database, which includes temperature, pressure, current and voltage.

3. The PEM electrolyzer accelerated stress testing system based on multi-stress coupling according to claim 2, Features: The proton membrane safety status monitoring module includes collecting the current proton membrane surface image at each stress change, listing the collected images into a proton membrane status image sequence in the order of collection time, and extracting the proton membrane status value sequence corresponding to the proton membrane status image sequence.

4. A PEM electrolyzer accelerated stress testing method based on multi-stress coupling, implemented based on a PEM electrolyzer accelerated stress testing system based on multi-stress coupling as claimed in any one of claims 1 to 3, Features: The following steps are involved: S1. Obtain the initial performance and initial proton membrane safety status before the test begins; S2, setting an initial stress step length and an initial stress dwell time, and adjusting the stress condition of the PEM electrolyzer according to the initial stress step length and the initial stress dwell time; S3, combining the proton membrane surface deformation and the hydrogen production efficiency of the electrolyzer to monitor the safety status of the proton membrane in real time; S4, setting an adaptive stress application algorithm to adjust the step size and dwell time of stress change in real time according to the safety state of the proton membrane; S5. Establish a life prediction model, count the performance degradation trends under different stress conditions, and train the life prediction model through the performance degradation trends to evaluate the service life of the PEM electrolyzer under different stress conditions.

5. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 4, Features: The initial proton membrane safety state includes extracting the initial spectral image of the proton membrane before the test starts, and calculating the initial spectral impact value of the initial spectral image through a damaged spectrum combination strategy. At this time, the initial proton membrane state value is recorded as 0. The damaged spectrum combination strategy is obtained by calculating the ratio of the difference and the sum of the visible light band reflectance and the ultraviolet light band reflectance of the spectral image at the current moment.

6. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 5, Features: The monitoring of the safety status of the proton membrane includes collecting the surface image of the proton membrane in real time when the proton membrane is working through the image sensor. First, the crack of the proton membrane is roughly located by the Roberts operator to obtain the rough defect edge, and then the n-order m-order orthogonal matrix of the pixels in the rough defect edge is calculated, and the rough defect edge is rotated counterclockwise again. The pixel orthogonal matrix within the rough defect edge after rotation is obtained, and the absolute value of the difference between the ratio of the orthogonal matrix before and after rotation and 1 is calculated, which is recorded as the proton membrane loss parameter.

7. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 6, Features: The monitoring of the safety status of the proton membrane also includes extracting the corresponding spectrum of the rough defect edge, recording it as the first spectrum data, graying the first spectrum data, setting a grayscale threshold, setting the grayscale value of pixels in the first spectrum data whose grayscale value is less than or equal to the grayscale threshold to 0, and extracting pixels in the first spectrum data whose grayscale value is greater than the grayscale threshold, recording it as the second spectrum data, calculating the spectral influence value of the second spectrum data through the damage spectrum combination strategy, extracting the difference between the spectral influence value of the second spectrum data and the initial spectral influence value, and calculating the product of the difference and the proton membrane loss parameter, recording it as the proton membrane status value.

8. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 7, Features: The adaptive stress application algorithm includes defining an objective function , and define the particle dimension as 2, that is, each particle consists of two values, including the step size and residence time of stress change; introduce the particle swarm algorithm, after adjusting the stress condition of the PEM electrolyzer according to the initial stress step size and the initial stress residence time, update the particle velocity and particle position according to the proton membrane state value, and then randomly select the individual optimal position of any particle as the initial global optimal position, and update the individual optimal position and the global optimal position after evaluating the fitness of each particle according to the objective function. When there is a particle whose individual optimal position has a fitness greater than the fitness of the initial global optimal position, update the individual optimal position of the particle as the new global optimal position, and output the stress change step size and residence time corresponding to the particle.

9. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 8, Features: The objective function is expressed as ,in , and represents the weight factor, , and represents the sub-objective function, Indicates a negative impact on the safety status of the proton membrane, represents the energy consumption of the electrolyzer, represents the maintenance cost, , E represents the minimum energy consumption expectation, C represents the minimum maintenance cost expectation, represents the stress variation step, Indicates the dwell time, represents the proton membrane state value, , and Represents the coefficient.

10. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 9, Features: The updating of particle velocity and particle position includes setting the particle position initialization range, proton membrane safety state threshold and initial inertia weight, and setting the inertia weight interval and proton membrane safety range , extract the current proton membrane state value in the proton membrane state value sequence and the proton membrane state value at the previous moment, update the inertia weight, and update the particle group speed; When the difference between the current proton membrane state value and the proton membrane state value at the previous moment is less than or equal to the proton membrane safety state threshold, the update formula of the inertia weight is: ； When the difference between the current proton membrane state value and the proton membrane state value at the previous moment is greater than the proton membrane safety state threshold, the inertia weight is updated by calculating the ratio of the proton membrane safety state threshold to the difference; the ratio of the proton membrane safety state threshold to the difference between the current proton membrane state value and the proton membrane state value at the previous moment is extracted as the position update coefficient, and when updating the particle position, the position update coefficient is added to update the particle position.

11. The PEM electrolyzer accelerated stress testing method based on multi-stress coupling according to claim 10, Features: The life prediction model includes obtaining the corresponding proton membrane state value and performance under different stress conditions, wherein the performance is obtained by calculating the hydrogen production efficiency within the residence time under the stress condition, and the hydrogen production efficiency is obtained by calculating the ratio of the total hydrogen production within the residence time under the stress condition to the residence time under the stress condition; calculating the performance degradation index by multiplying the corresponding proton membrane state value and performance under different stress conditions, analyzing the growth trend of the performance degradation index through an exponential decay model, and obtaining the performance decay coefficient K, introducing a support vector machine model, establishing a life prediction model, using the performance degradation index as training data, and using the current stress as an input parameter, and the time in the output exponential decay model is the expected life of the PEM electrolyzer.

Images