Method and device for predicting the lifetime of a variable power operated pem electrolyser

Abstract

This invention provides a method and apparatus for predicting the lifespan of PEM electrolyzers operating at variable power. First, considering the non-monotonic nature of PEM electrolyzer degradation data, a polynomial fitting is used to fit the trajectory of the degradation data. The upper and lower limits of a certain confidence interval of the fitting error distribution are then superimposed to establish an interval-based performance degradation model for the PEM electrolyzer. Based on this, a Hyper-CuboidalVolume accelerated model is employed to describe the impact of power fluctuation amplitude and fluctuation time scale on the electrolyzer degradation process. Then, an optimization-based parameter estimation method for the accelerated model is proposed. Finally, a statistical method for time-scale fluctuation segments of random power sequences is proposed to predict the lifespan of PEM electrolyzers operating at variable power. This invention incorporates the non-monotonic characteristics of PEM electrolyzer degradation data into its lifespan prediction, making it closer to real-world models and resulting in more refined and accurate lifespan predictions.

Description

Technical Field

[0001] This invention belongs to the field of power system equipment life prediction, specifically relating to a method and device for predicting the life of a PEM electrolyzer operating at variable power. Background Technology

[0002] Faced with the challenges of large-scale, high-proportion renewable energy grid integration, insufficient balancing capacity of new power systems, and difficulties in ensuring power supply under extreme conditions, the demand for flexible resources has surged. Proton exchange membrane (PEM) electrolyzers, with their second-level response (hot start-up from standby to rated production in 1-5 seconds, ramp-up rate of 100% per second) and wide-range regulation (10%-200% regulation range for electro-hydrogen systems), can provide auxiliary services such as frequency regulation and peak shaving for the power grid. Their large-scale grid integration provides considerable regulation capacity for the system, supporting the stable and economical operation of new power systems. However, long-term operation or frequent cyclic operation of PEM electrolyzers leads to performance degradation, which includes both reversible and irreversible aspects. Microscopic degradation phenomena include cation contamination, proton exchange membrane thinning, catalyst corrosion, and titanium-based component oxidation; macroscopic manifestations include abnormal operating temperature, flooding, gas leakage, low hydrogen purity, and increased operating voltage (i.e., an overall upward shift in the polarization curve), ultimately resulting in electrolyzer performance failure. Many factors influence the performance and lifespan of PEM electrolyzers, primarily including intrinsic factors, system factors, and environmental factors. Intrinsic factors involve the equipment's structure and design parameters, such as the loading of platinum and iridium catalysts, the thickness of the proton exchange membrane, and electrode materials, which are related to the electrolyzer's manufacturing process and selection. System factors mainly include improper parameter settings or subsystem failures in the multi-energy flow management subsystems (water, heat, electricity, gas, etc.). Environmental factors mainly involve variations in the PEM electrolyzer's operating temperature, pressure, and input power. While operating temperature and pressure are controllable parameters, the input power of PEM electrolyzers directly connected to fluctuating power sources like wind and solar power often varies with the power supply. Therefore, establishing performance degradation and lifespan prediction models for variable-power operating electrolyzers is of great significance.

[0003] Currently, many studies have discussed the impact mechanism of variable power operation on the working performance and lifespan of PEM electrolyzers. However, none of the current studies have taken into account the non-monotonic decline of PEM electrolyzer performance into their lifespan prediction models. The lifespan prediction model for variable power PEM electro-hydrogen production systems needs to be further refined. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, this invention proposes a method and device for predicting the lifespan of a PEM electrolyzer operating with variable power, thereby making the prediction more accurate and precise by more closely matching the actual model.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for predicting the lifespan of a PEM electrolyzer operating at variable power includes the following steps:

[0007] 1) Obtain the performance degradation trajectory of PEM electrolyzers by fitting degradation data with a polynomial;

[0008] 2) Based on the performance degradation trajectory of the PEM electrolyzer obtained in step 1), the upper and lower confidence limits of the fitting error distribution at a certain confidence level are superimposed to establish a performance degradation model of the PEM electrolyzer with interval characterization.

[0009] 3) Based on the PEM electrolyzer performance degradation model obtained in step 2), describe the influence of dual stresses of power fluctuation amplitude and fluctuation time scale on the electrolyzer degradation process, and establish a Hyper-CuboidalVolume accelerated model.

[0010] 4) Based on the Hyper-CuboidalVolume acceleration model obtained in step 3), calculate the Hyper-CuboidalVolume acceleration model parameters using the optimization method;

[0011] 5) Based on the PEM electrolyzer performance degradation model, Hyper-CuboidalVolume acceleration model and parameters obtained in steps 2), 3) and 4), calculate the lifespan of the PEM electrolyzer under different stresses.

[0012] Furthermore, step 1) includes the following steps:

[0013] The changes in degradation data at different time periods are characterized by the Iton stochastic process, as shown in equation (1):

[0014] dA=μ(A,θ)dt+δ(A,θ)dW t (1)

[0015] In equation (1), A represents the rate of change of the degradation data; the drift term μ(A,θ)dt represents the trend of the average degradation rate of the electrolytic cell; δ(A,θ)dW t Let W be the diffusion term, representing the random uncertainty of electrolyzer degradation; θ represents the random variable, W t This is a standard Wiener process.

[0016] By fitting the experimental degradation data with the polynomial function shown in equation (2), the parameters for calculating the drift term are derived:

[0017]

[0018] In equation (2), M represents the highest order of the polynomial, w j For xj The coefficient of the term, where x is the time variable.

[0019] Furthermore, step 2) includes the following steps:

[0020] The fitting error sequence ε is generated according to equation (3), where y0 is the actual measured data; it is assumed that the fitting error follows a normal distribution with mean u and variance m:

[0021] ε i =y0(x i )-y(x i ,w) (3)

[0022]

[0023]

[0024] In the formula, I represents the number of fitting points.

[0025] The lower limit c1 and upper limit c2 of a certain confidence interval of the error distribution are superimposed on the fitted curve, as shown in Equation (6), to generate the trajectory of the performance degradation of the electrolyzer characterized by the interval:

[0026]

[0027]

[0028] In equation (7), y- represents the lower limit of the fitting lifespan interval of the electrolytic cell, and y+ represents the upper limit of the fitting lifespan interval of the electrolytic cell; when calculating the confidence interval for a large sample, r is a constant determined based on the confidence interval.

[0029] Furthermore, step 3) includes the following steps:

[0030] The Hyper-Cuboidal Volume acceleration model is shown in Equation (7):

[0031]

[0032] In the formula, A(S1,S2,…,S) n ) for the product under accelerated stress S1~S n Lifespan a under combined action i (i = 0, 1, ..., n) are undetermined coefficients; P(S i ) represents a function of each accelerating stress. This is the symbol for consecutive multiplication.

[0033] Taking n = 2, we simplify equation (8) by taking the logarithm of both sides to equation (9), which is easier to solve:

[0034] ln(A(S1,S2))=a1+a2lnS1+a3lnS2 (9)

[0035] In equation (9), a1, a2 and a3 are the acceleration model parameters to be estimated.

[0036] Furthermore, step 4) includes the following steps:

[0037] With the goal of minimizing the difference between the actual and estimated values, a parameter optimization model for the Hyper-CuboidalVolume accelerated model is constructed, namely equations (10) to (11), and solved using a genetic algorithm:

[0038] a1+a2 ln(s 1,n )+a3 ln(s 2,n )-ln(y n )=d n (10)

[0039]

[0040] In equation (10), y n The lifetime under different stresses is obtained from the performance degradation model, a i,max This represents the upper limit for each parameter. s 1,n With s 2,n These are the values ​​of stress 1 and stress 2 corresponding to the nth set of degenerate data.

[0041] Furthermore, step 5) includes the following steps:

[0042] Lifetime assessment is performed by measuring the decay rate of each fluctuation within a precise operating cycle. The specific steps are as follows:

[0043] Step 1: Define the fluctuation segment: Let the power at time t be Pt, and if the power at the next time moment is P t+1 If the power is greater than or equal to Pt, then the period from time t to time t+1 is defined as the power rising period; otherwise, Pt is defined as the power falling period. If the power continues to rise from time t to time t+n, and then begins to fall after time t+n, then the period from t to t+n is defined as a rising fluctuation segment. If the power continues to fall from time t to time t+n, then the period from t to t+n is defined as a falling fluctuation segment.

[0044] Step 2: Input the renewable energy output sequence and remove segments where the power output is continuously zero; calculate the duration and amplitude of the rising or falling segments in the renewable energy output sequence.

[0045] Step 3: Calculate the difference Δdg,i between the average attenuation rate at the beginning and end of each wave segment according to Equation (12), and then calculate the attenuation amount dg,i within the wave segment, as shown in Equation (13).

[0046] Δd g,i =|D / (A(T) i,t V i,t0 ))-D / A(T i,t V i,tend )|(12)

[0047] d g,i =k·Δd g,i (13)

[0048] In equation (11), D is the degradation failure threshold, and T i,t V is the cycle period of the i-th fluctuation. i,t0 With V i,tend denoted as the operating voltage at the initial and final moments of the i-th fluctuation; k represents the duration of the fluctuation period in hours.

[0049] Step 4: Calculate the lifespan of the PEM electrolyzer as shown in equation (14):

[0050]

[0051] In equation (14), I represents the number of upward and downward fluctuation segments.

[0052] The present invention also provides a PEM electrolyzer lifetime prediction device for variable power operation, comprising:

[0053] The degradation trajectory fitting module is used to obtain the performance degradation trajectory of the PEM electrolyzer by fitting the degradation data with a polynomial.

[0054] The accelerated model parameter optimization calculation module is used to superimpose the upper and lower confidence limits of the fitting error distribution at a certain confidence level to establish a PEM electrolyzer performance degradation model with interval characterization; it describes the influence of dual stresses of power fluctuation amplitude and fluctuation time scale on the electrolyzer degradation process and establishes a Hyper-Cuboidal Volume accelerated model.

[0055] The lifetime prediction module is used to calculate the parameters of the Hyper-Cuboidal Volume accelerated model based on the optimization method, and to calculate the lifetime of the PEM electrolyzer under different stresses based on the obtained PEM electrolyzer performance degradation model, Hyper-Cuboidal Volume accelerated model and parameters.

[0056] The present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the above-described evaluation method.

[0057] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the various steps of the above-described evaluation method.

[0058] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can 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. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0059] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0060] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0061] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0062] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0063] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0064] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0065] Beneficial effects:

[0066] This invention incorporates the non-monotonic characteristics of PEM electrolyzer degradation data into its lifetime prediction, making it closer to real-world models and resulting in more refined and accurate lifetime predictions. This invention is of great significance for widespread application in the electrolytic hydrogen production fields of new energy sources such as wind power and photovoltaics. Compared with existing PEM electrolyzer lifetime assessment methods, this invention has the following advantages: 1) It considers the impact of the non-monotonic decrease in electrolyzer performance degradation data on its performance degradation trajectory, making the lifetime prediction assessment results more accurate; 2) The proposed PEM electrolyzer lifetime prediction model can be directly used for PEM electrolyzer lifetime assessment in wind farms and photovoltaic power plants, enabling reasonable replacement of electrolyzers and better ensuring the safety of the power system where the electrolyzers are located; 3) The proposed PEM electrolyzer lifetime prediction model can directly guide the economic planning of PEM electrolyzers in various application scenarios, promoting their application and development. Attached Figure Description

[0067] Figure 1 This is a flowchart of the method for predicting the lifetime of a PEM electrolyzer operating under variable power conditions according to the present invention.

[0068] Figure 2 This is a schematic diagram illustrating the degradation principle of a variable power PEM electrolyzer. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0070] This invention provides a method and apparatus for predicting the lifespan of PEM electrolyzers operating at variable power. First, considering the non-monotonic nature of PEM electrolyzer degradation data, a polynomial-fitted degradation trajectory is used as the basic degradation trajectory. The upper and lower confidence limits of the fitting error distribution at a certain confidence level are superimposed to establish an interval-characterized PEM electrolyzer performance degradation model. Based on this, considering the multi-timescale fluctuation characteristics of random input power, a Hyper-Cuboidal Volume acceleration model is adopted to describe the influence of dual stresses—power fluctuation amplitude and fluctuation time scale—on the electrolyzer degradation process. Then, combining the principle of accelerated model parameter estimation, an optimization-based accelerated model parameter estimation method is proposed. Finally, a statistical method for time-scale fluctuation segments of random power sequences is proposed to predict the lifespan of PEM electrolyzers operating at variable power. This invention incorporates the non-monotonic characteristics of PEM electrolyzer degradation data into its lifespan prediction, making it closer to the real-world model and resulting in more refined and accurate lifespan predictions. The above scheme is described in detail below.

[0071] like Figure 1 As shown, the steps of the method for predicting the lifetime of a PEM electrolyzer operating under variable power conditions according to the present invention are as follows:

[0072] Step 1) Considering the non-monotonic nature of PEM electrolyzer degradation data, the performance degradation trajectory of PEM electrolyzer is obtained by fitting the degradation data with a polynomial.

[0073] Step 2) Based on the performance degradation trajectory of the PEM electrolyzer obtained in Step 1), the upper and lower confidence limits of the fitting error distribution at a certain confidence level are superimposed to establish a performance degradation model of the PEM electrolyzer with interval characterization.

[0074] Step 3) Based on the PEM electrolyzer performance degradation model obtained in Step 2), describe the influence of dual stresses of power fluctuation amplitude and fluctuation time scale on the electrolyzer degradation process, and establish the Hyper-CuboidalVolume accelerated model.

[0075] Step 4) Based on the Hyper-CuboidalVolume acceleration model obtained in Step 3), calculate the Hyper-CuboidalVolume acceleration model parameters using the optimization method;

[0076] Step 5) Based on the PEM electrolyzer performance degradation model, Hyper-Cuboidal Volume acceleration model and parameters obtained in Steps 2), 3) and 4), calculate the lifespan of the PEM electrolyzer under different stresses.

[0077] Specifically, step 1) of fitting the performance degradation trajectory of the PEM electrolyzer operating at variable power includes the following steps:

[0078] Observation of accelerated degradation data from PEM electrolyzers revealed that the degradation process is nonlinear. The changes in degradation data at different time periods do not exhibit a strictly monotonically increasing or decreasing pattern, but can be characterized by an Itō stochastic process, as shown in equation (1).

[0079] dA=μ(A,θ)dt+δ(A,θ)dW t (1)

[0080] In the formula, A represents the rate of change of the degradation data; the drift term μ(A,θ)dt represents the trend of the average degradation rate of the electrolytic cell; δ(A,θ)dW t Let W be the diffusion term, representing the random uncertainty of electrolyzer degradation; θ represents the random variable, W t This is a standard Wiener process.

[0081] The parameters of the drift term can be derived by fitting the experimental degradation data with a polynomial function as shown in equation (2).

[0082]

[0083] In the formula, M represents the highest order of the polynomial, and w j For x j The coefficient of the term, where x is the time variable.

[0084] Step 2) involves establishing a performance degradation model for the PEM electrolyzer based on interval representation, which includes the following steps:

[0085] Generate the fitting error sequence ε according to formula (3), where y0 is the actual measurement data; assume that the fitting error follows a normal distribution with mean u and variance m.

[0086] ε i =y0(x i )-y(x i ,w) (3)

[0087]

[0088]

[0089] In the formula, I represents the number of fitting points.

[0090] The lower limit c1 and upper limit c2 of a certain confidence interval of the error distribution are superimposed on the fitted curve, as shown in Equation (6), to generate the trajectory of the performance degradation of the PEM electrolyzer represented by the interval.

[0091]

[0092]

[0093] In the formula, y- represents the lower limit of the fitted lifespan range of the electrolytic cell, and y+ represents the upper limit of the fitted lifespan range of the electrolytic cell; when calculating the confidence interval for a large sample, r is a constant determined based on the confidence interval.

[0094] Step 3) involves establishing the Hyper-Cuboidal Volume acceleration model, which includes the following steps:

[0095] Hyper-Cuboidal Volume accelerated models are often used to describe the combined effects of multiple stresses on the product performance degradation process, as shown in Equation (7).

[0096]

[0097] In the formula, A(S1,S2,…,S) n ) for the product under accelerated stress S1~S n Lifespan under combined action, a i (i = 0, 1, ..., n) are undetermined coefficients; P(S i ) represents a function of each accelerating stress. The sign represents continuous multiplication. Since the variable power operation of the PEM electrolytic cell is mainly affected by two variables, the cycle period and the cycle voltage, here n=2, so we can take the logarithm of both sides of equation (8) to simplify it into equation (9), which is convenient for solving.

[0098] ln(A(S1,S2))=a1+a2lnS1+a3lnS2 (9)

[0099] In the formula, a1, a2 and a3 are the acceleration model parameters to be estimated.

[0100] Step 4) involves calculating the Hyper-CuboidalVolume acceleration model parameters using an optimization method, which includes the following steps:

[0101] Based on the principle of accelerated model parameter estimation, an accelerated model parameter optimization model is constructed with the goal of minimizing the difference between the actual value and the estimated value, as shown in equations (10) to (11).

[0102] a1+a2ln(s 1,n )+a3ln(s 2,n )-ln(y n )=d n (10)

[0103]

[0104] In equation (10), y n The lifetime under different stresses is obtained from the performance degradation model, a i,max This represents the upper limit for each parameter. s1,n With s 2,n These are the values ​​of stress 1 and stress 2 corresponding to the nth set of degenerate data.

[0105] Step 5) involves calculating the lifespan of the PEM electrolyzer under different stresses, including the following steps:

[0106] Since the performance degradation of PEM electrolyzers is affected by the cycle period and fluctuation amplitude, in order to achieve a more accurate life assessment, it is necessary to refine the degradation rate of each fluctuation within the operating cycle. The specific steps are as follows:

[0107] Step 1: Define the fluctuation segment. Let the power at time t be Pt. If the power at the next time point Pt+1 is greater than or equal to Pt, then the period from time t to time t+1 is defined as the power rising period; otherwise, Pt is defined as the power falling period. If the power continues to rise from time t to time t+n, and then begins to fall after time t+n, then the period from t to t+n is defined as a rising fluctuation segment. If the power continues to fall from time t to time t+n, then the period from t to t+n is defined as a falling fluctuation segment.

[0108] Step 2: Input the new energy power output sequence and remove segments where the power output is continuously zero; count the duration and amplitude of the rising or falling fluctuation segments in the new energy power output sequence.

[0109] Step 3: Calculate the difference Δdg,i between the average attenuation rate at the beginning and end of each wave segment according to Equation (12), and then calculate the attenuation amount dg,i within the wave segment, as shown in Equation (13).

[0110]

[0111] d g,i =k·Δd g,i (13)

[0112] In the formula, D is the degradation failure threshold, Ti,t is the cycle period of the i-th fluctuation, Vi,t0 and Vi,tend are the working voltages at the beginning and end of the i-th fluctuation, respectively; k is the duration of the fluctuation period in hours.

[0113] Step 4: Calculate the lifespan of the PEM electrolyzer as shown in equation (14).

[0114]

[0115] In the formula, I represents the number of upward and downward fluctuation segments.

[0116] The present invention also provides a PEM electrolyzer lifetime prediction device for variable power operation, comprising:

[0117] The degradation trajectory fitting module is used to obtain the performance degradation trajectory of PEM electrolyzers by fitting degradation data with a polynomial.

[0118] The accelerated model parameter optimization calculation module is used to superimpose the upper and lower confidence limits of the fitting error distribution at a certain confidence level to establish a PEM electrolyzer performance degradation model with interval characterization; it describes the influence of dual stresses of power fluctuation amplitude and fluctuation time scale on the electrolyzer degradation process and establishes a Hyper-Cuboidal Volume accelerated model.

[0119] The lifetime prediction module is used to calculate the parameters of the Hyper-Cuboidal Volume accelerated model based on the optimization method, and to calculate the lifetime of the PEM electrolyzer under different stresses based on the obtained PEM electrolyzer performance degradation model, Hyper-Cuboidal Volume accelerated model and parameters.

[0120] The present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the above-described evaluation method.

[0121] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the various steps of the above-described evaluation method.

[0122] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0123] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can 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. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0124] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0125] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0127] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0128] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for predicting the lifetime of a variable power operated PEM electrolyzer, characterized in that, Comprising the following steps: 1) Obtain the performance degradation trajectory of the PEM electrolyzer by polynomial fitting of the degradation data, comprising the following steps: The change of degradation data in different time periods is characterized by an Itô random process, as shown in equation (1): (1) In formula (1), A represents the rate of change of degradation data; μ(A, θ)dt drift term, representing the change trend of the average decay rate of the electrolytic cell degradation; δ(A, θ) dW t is a diffusion term, representing the random uncertainty of the electrolytic cell degradation; θ represents a random variable, W t is a standard Wiener process; The experimental degradation data is fitted by a polynomial function as shown in equation (2), and the parameters of the drift term are derived and calculated: (2) In formula (2), M represents the highest order of the polynomial, w j is x j the coefficient of the term, x is the time variable; 2) Based on the performance degradation trajectory of the PEM electrolyzer obtained in step 1), superimpose the upper and lower limits of the confidence interval of the fitting error distribution at a certain confidence level to establish an interval representation of the PEM electrolyzer performance degradation model; 3) Based on the PEM electrolyzer performance degradation model obtained in step 2), describe the influence of power fluctuation amplitude and fluctuation time scale double stress on the degradation process of the electrolyzer, and establish a Hyper-Cuboidal Volume acceleration model; 4) Based on the Hyper-Cuboidal Volume acceleration model obtained in step 3), calculate the Hyper-Cuboidal Volume acceleration model parameters based on the optimization method; 5) Based on the PEM electrolyzer performance degradation model, Hyper-Cuboidal Volume acceleration model and parameters obtained in steps 2), 3) and 4), calculate the life of the PEM electrolyzer under different stresses.

2. A method of predicting the lifetime of a variable power operated PEM electrolyzer according to claim 1, characterized in that: The step 2) comprises the following steps: Generate the fitting error sequence ε according to equation (3), y0 is the actual measurement data; assume that the fitting error follows a normal distribution with mean u and variance m: (3) (4) (5) In the formula, I is the number of fitting points; Superimpose the lower limit c1 and the upper limit c2 of the error distribution at a certain confidence interval on the fitting curve respectively, as shown in equation (6), to generate the interval representation of the electrolyzer performance degradation change trajectory: (6) (7) In equation (7), y- represents the lower limit value of the electrolyzer fitting life interval, y+ represents the upper limit value of the electrolyzer fitting life interval; when calculating the confidence interval of a large sample, r is a constant according to the confidence interval.

3. A method of predicting the lifetime of a variable power operated PEM electrolyzer according to claim 2, characterized in that: The step 3) comprises the following steps: The Hyper-Cuboidal Volume acceleration model is as shown in equation (7): (8) where A(S1, S2, …, S n ) is the life a of the product under the combined action of the stresses S1~S n i (i=0, 1, …, n) are undetermined coefficients; P(S i ) represents a function of each stress, is the continuous multiplication symbol;​ Take n=2, and simplify equation (8) to equation (9) by taking the logarithm of both sides, which is convenient for solving: (9) In equation (9), a1, a2 and a3 are the acceleration model parameters to be estimated.

4. A method of predicting the lifetime of a variable power operated PEM electrolyzer according to claim 3, characterized in that: The step 4) comprises the following steps: Take the minimum difference between the actual value and the estimated value as the optimization objective, construct the parameter optimization model of the Hyper-Cuboidal Volume acceleration model, that is, equations (10)-(11), and solve it using genetic algorithm: (10) (11) In equation (10), y n The lifetime under different stresses is obtained from the performance degradation model, a i,max s represents the upper limit of each parameter; 1,n With s 2,n These are the values ​​of stress 1 and stress 2 corresponding to the nth set of degenerate data.

5. A method of predicting the lifetime of a variable power operated PEM electrolyzer according to claim 4, characterized in that: The step 5) comprises the following steps: Life assessment is carried out through the decay rate of each fluctuation in the fine operation period, and the specific steps are as follows: Step 1: define fluctuation segment: set time t power as Pt, if next time power P t+1 greater than or equal to Pt, define as t time ~ t+1 time define as power rise period, otherwise Pt define as power drop period; if power continuously rises from t time to t+n time, after t+n time power begins to drop, define t~t+n period as an rising fluctuation segment; if power continuously drops from t time to t+n time, define t~t+n period as a falling fluctuation segment; Step 2: Input the new energy output sequence, and remove the segment with zero power; count the duration and fluctuation amplitude of the rising or falling fluctuation segment of the new energy output sequence; Step 3: Calculate the difference Δdg,i between the average decay rates at the initial and final times of each fluctuation segment according to equation (12), and then calculate the decay amount dg,i in the fluctuation segment as shown in equation (13); (12) (13) In equation (11), D is the degradation failure threshold, and T i,t V is the cycle period of the i-th fluctuation. i,t0 With V i,tend Let be the operating voltage at the beginning and end of the i-th fluctuation; k is the duration of the fluctuation period in hours. Step 4: Calculate the life of the PEM electrolyzer as shown in equation (14): (14) In formula (14), I is the number of the up fluctuation segment and the down fluctuation segment.

6. A device for predicting the lifetime of a variable power operated PEM electrolyzer, characterized by, Comprise: The degeneration trajectory fitting module is used for fitting the degeneration data with a polynomial to obtain a performance degeneration trajectory of the PEM electrolyzer; The change amount of the degeneration data in different time periods is represented by an It random process, as shown in formula (1): (1) In formula (1), A represents the rate of change of degradation data; μ(A, θ)dt drift term, representing the change trend of the average decay rate of the electrolytic tank degradation; δ(A, θ) dW t is a diffusion term, representing the random uncertainty of the electrolytic tank degradation; θ represents a random variable, W t is a standard Wiener process; The experimental degeneration data is fitted by a polynomial function as shown in formula (2), and the parameters of the drift term are derived and calculated: (2) In formula (2), M represents the highest order of the polynomial, w j is x j the coefficient of the term, x is the time variable The accelerated model parameter optimization calculation module is used for superimposing the upper and lower limits of the confidence level of the fitting error distribution to establish an interval representation PEM electrolyzer performance degeneration model; the effects of power fluctuation amplitude and fluctuation time scale double stresses on the electrolyzer degeneration process are described, and a Hyper-Cuboidal Volume accelerated model is established; The life prediction module is used for calculating the Hyper-Cuboidal Volume accelerated model parameters based on an optimization method, and calculating the life of the PEM electrolyzer under different stresses based on the obtained PEM electrolyzer performance degeneration model, Hyper-Cuboidal Volume accelerated model and parameters.

7. An electronic device comprising: The memory, the processor and the computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the method for predicting the life of the variable-power PEM electrolyzer according to any one of claims 1 to 5 is implemented.

8. A readable storage medium, having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement each step of the method for predicting the life of the variable-power PEM electrolyzer according to any one of claims 1 to 5.

Images