Photovoltaic module service life prediction method and system

Abstract

The invention provides a photovoltaic module residual life prediction method and system by fusing Gaussian process regression and a dynamic Gamma process. The prediction method comprises the steps of determining a residual life distribution function of a photovoltaic module at a to-be-measured moment based on a dynamic Gamma process degradation model of the photovoltaic module; acquiring degradation data of the photovoltaic module, and performing interpolation processing by using a GPR (General Purpose Register); performing parameter real-time updating on the residual life distribution function of the photovoltaic module at the to-be-measured moment in combination with a sliding window and a maximum likelihood estimation method according to the degradation data of the photovoltaic module after interpolation; and carrying out life prediction on the photovoltaic module by using the residual life distribution model after parameter updating. According to the method, the degradation characteristic data of the target photovoltaic module can be fully utilized, the uncertainty and nonlinearity presented in the degradation process are considered, and more accurate life prediction is realized in combination with the interpolated data, so that the precision and reliability of the residual life prediction of the photovoltaic module are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic module life prediction, and more particularly to a photovoltaic module life prediction method and system. Background Art

[0002] Photovoltaic (PV) modules, as core components of photovoltaic (PV) power generation systems, gradually decline in performance over time due to increasing service life and the influence of random factors. This cumulative degradation impacts the reliability of PV power generation. Furthermore, due to the low degradation rate of PV modules, it is difficult to collect long-term data to confirm the degradation path and lifespan. Therefore, it is necessary to develop a stochastic degradation model to characterize the unstable and fuzzy characteristics of PV module performance degradation over time, in order to estimate the remaining lifespan of PV modules and improve their operational reliability.

[0003] At present, the current method of predicting the remaining life of photovoltaic modules based on data modeling basically adopts random processes to establish the performance degradation model of photovoltaic modules. This method first requires sufficient degradation data. Taking into account some irresistible factors, it is difficult to obtain continuous and complete degradation data in actual projects; secondly, the model selected to describe the module degradation process is not suitable for the entire life cycle of photovoltaic modules.

[0004] Therefore, how to solve the above technical problems is an urgent problem that those skilled in the art need to solve. Summary of the Invention

[0005] In view of this, in order to at least partially solve the above technical problems, the present invention provides a photovoltaic module life prediction method and system, which can not only adapt to the entire life cycle of the photovoltaic module, but also overcome the problem of sparse or missing degradation data in the actual operation of the photovoltaic module.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a method for predicting the life of a photovoltaic module, the steps comprising:

[0008] Based on the dynamic Gamma process degradation model of photovoltaic modules, the remaining life distribution function of the photovoltaic modules at the time of testing is determined;

[0009] Obtain PV module degradation data and perform interpolation processing using GPR;

[0010] Based on the interpolated PV module degradation data, the sliding window and maximum likelihood estimation methods are combined to update the parameters of the remaining life distribution function of the PV module at the time of testing in real time.

[0011] The remaining life distribution model with updated parameters is used to predict the life of photovoltaic modules.

[0012] In a second aspect, the present application provides a photovoltaic module life prediction system. This system applies the photovoltaic module life prediction method described above, and includes:

[0013] A remaining life distribution function determination module, configured to determine the remaining life distribution function of the photovoltaic module at the moment to be measured based on the dynamic Gamma process degradation model of the photovoltaic module;

[0014] A degradation data acquisition and processing module, configured to acquire the degradation data of the photovoltaic module and perform interpolation processing using GPR;

[0015] A life distribution function parameter update module, configured to perform real-time parameter update on the remaining life distribution function of the photovoltaic module at the moment to be measured according to the interpolated degradation data of the photovoltaic module, in combination with a sliding window and the maximum likelihood estimation method;

[0016] A remaining life prediction module, which uses the remaining life distribution model with updated parameters to predict the life of the photovoltaic module.

[0017] It can be seen from the above technical solutions that the present invention provides a photovoltaic module life prediction method and system by integrating GPR and the dynamic Gamma process, which can take into account both prediction accuracy and real-time performance, adapt to the entire life cycle of photovoltaic modules, thereby effectively solving the deficiencies of current data modeling methods, reducing the limitations of current remaining life prediction of photovoltaic modules, and making the remaining life prediction method more meet the actual application requirements. Compared with the prior art, the advantages of the present application specifically include:

[0018] 1) The present application constructs a remaining life prediction model at the moment to be predicted based on the dynamic Gamma process, and can still achieve relatively accurate remaining life estimation in the scenario of lacking long-term continuous data;

[0019] 2) Interpolate the degradation data based on GPR, and construct a covariance function based on the anisotropic squared exponential (RBF) kernel and the kernel; which is used to ensure the smoothness of interpolation while retaining the mutation information in the degradation process, so as to achieve the purpose of enhancing prediction flexibility, improving interpolation accuracy, adapting to multi-scale degradation behavior, and facilitating marginal likelihood optimization learning;

[0020] 3) Based on the degradation trajectory of the photovoltaic module, combine a sliding window and the maximum likelihood estimation method to perform real-time update on the parameters of the Gamma degradation model, and on this basis, predict the remaining life of the photovoltaic module in real time. Among them, the sliding window mechanism improves the sensitivity of the model to short-term dynamic changes, and the maximum likelihood estimation realizes local optimal parameter estimation within each window, so as to achieve real-time dynamic update of parameters and enhanced tracking modeling ability, and improve the adaptability and prediction accuracy of the model in non-stationary degradation scenarios.

[0021] Other features and advantages of the present invention will be described in the following specification, and the objectives and other advantages of the present invention can be achieved and obtained by the structures specifically pointed out in the written specification, claims, and drawings.

[0022] The technical solutions of the present invention will be further described in detail below with reference to the drawings and embodiments. Description of the Drawings

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0024] Figure 1 It is a flowchart of the method for predicting the lifespan of the photovoltaic module of the present invention;

[0025] Figure 2 It is an example diagram of the process for predicting the lifespan of the photovoltaic module of the present invention;

[0026] Figure 3 It is a schematic diagram of the hardware composition for applying the method for predicting the lifespan of the photovoltaic module of the present invention;

[0027] Figure 4(a) is a schematic diagram of the process for processing the cumulative power degradation amount of the photovoltaic module in the first monitoring of the present invention;

[0028] Figure 4(b) is a schematic diagram of the process for processing the cumulative power degradation amount of the photovoltaic module in the second monitoring of the present invention;

[0029] Figure 4(c) is a schematic diagram of the process for processing the cumulative power degradation amount of the photovoltaic module in the third monitoring of the present invention;

[0030] Figure 5 It is a schematic diagram of the goodness-of-fit test provided by the example of the present invention;

[0031] Figure 6 It is a diagram of the real-time update process of the degradation model parameters provided by the example of the present invention;

[0032] Figure 7 It is the real-time monitoring process and results of the degradation process of the photovoltaic module provided by the example of the present invention;

[0033] Figure 8 It is the probability distribution of the lifespan failure time of the photovoltaic module provided by the example of the present invention. Detailed Embodiments

[0034] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0035] Many specific details are set forth in the following description in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0036] In one embodiment, the present invention discloses a method for predicting the life of a photovoltaic module, as Figure 1 , the steps include:

[0037] Based on the dynamic Gamma process degradation model of the photovoltaic module, determine the remaining life distribution function of the photovoltaic module at the time to be measured;

[0038] Obtain the degradation data of the photovoltaic module and perform interpolation processing using GPR;

[0039] According to the interpolated degradation data of the photovoltaic module, jointly use the sliding window and the maximum likelihood estimation method to update the parameters of the remaining life distribution function of the photovoltaic module at the time to be measured in real time;

[0040] Use the remaining life distribution model with updated parameters to predict the life of the photovoltaic module.

[0041] In this embodiment, constructing the life distribution and obtaining the component degradation data can be carried out simultaneously, or first obtain the data and then construct the life distribution function;

[0042] In one embodiment, first, based on the dynamic Gamma process degradation model of the photovoltaic module, determine the remaining life distribution function of the photovoltaic module at the time to be measured;

[0043] This embodiment establishes a function related to the device performance degradation process based on the standard Gamma process. When the degradation during the device service process is described as , the degradation model constructed based on the Gamma process is shown as the following formula:

[0044]

[0045] Among them, represents the degradation increment of the Gamma process; k represents the shape parameter; represents the scale parameter; a represents the time variable, which is used to describe the service time of the photovoltaic module, and a ≥ 0.

[0046] The Gamma process assumes that the degradation process of the photovoltaic module is strictly monotonic and is suitable for characterizing the monotonic degradation trend in which the continuous accumulation of decline causes the device to fail. In this application, the cumulative power degradation amount is selected as the characteristic quantity for describing the performance degradation of the photovoltaic module, which can well interpret the mechanism of the service condition of the module changing with time. Specifically, the cumulative power degradation amount is calculated based on the initial value of the output power of the target photovoltaic module obtained and the output power of the target photovoltaic module at the current moment.

[0047] During the service process of the photovoltaic module, the cumulative power degradation amount of the photovoltaic module will generally show an upward trend over time. At the same time, during the data preprocessing process, the time uncertainty of the change in the degradation process of the module caused by the combined effects of uncertain factors such as the natural environment and humans is comprehensively considered. Generally, when the cumulative power degradation amount of the service photovoltaic module reaches 20% of the initial power, it is used as the threshold for module degradation failure. In the actual prediction process, the failure threshold can be set and changed according to the actual engineering needs of the module service.

[0048] Furthermore, the performance failure of the photovoltaic module is defined as a first-passage failure. Once the degradation process X(t) of the module reaches the failure threshold ξ, it can be considered that it has failed. The failure time is defined as the moment when the sample degradation trajectory of X(t) first exceeds ξ, and the lifetime T is the length of the time from when the device starts to be newly put into use until it first reaches the failure threshold. Then the following formula can be obtained:

[0049]

[0050] t represents the moment corresponding to the time T when the cumulative power degradation amount first reaches the failure threshold, and s represents any moment before the moment t.

[0051] When X(t) is a Gamma process, according to the monotonicity of the sample trajectory of the degradation process, the lifetime distribution function is:

[0052]

[0053] In the formula, represents the probability that the lifetime random variable T is less than or equal to a certain time point t, k(t) represents the shape parameter of the dynamic Gamma process at the moment t, ξ represents the failure threshold, and k represents the shape parameter; represents the scale parameter, and γ(k, x) represents the upper incomplete Gamma function, which is defined as:

[0054]

[0055] Where k is the shape parameter of the Gamma process, reflecting the acceleration degree of the degradation process. The larger it is, the faster the degradation. x is the normalized time, reflecting the relative remaining degradation time of the current component until failure, which is related to the current degradation degree and the scale parameter.

[0056] Furthermore, combined with the measured time t0, the remaining life distribution function of the photovoltaic module is determined. Based on the remaining life distribution function constructed by the dynamic Gamma process model and combined with the real-time estimation results of the model parameters, the expected value of the life is further derived.

[0057] That is, if the component has worked until the current time t0, its remaining life distribution is calculated according to its life distribution function:

[0058]

[0059] Where represents the cumulative probability of the photovoltaic module failing before time t, represents the cumulative probability of the component failing before the current time t0, represents the conditional probability that the component fails before the future time t given that it has not failed at time t0.

[0060] For the performance degradation process described by the Gamma process, the distribution function of the time when the degradation process first reaches the failure threshold ξ can be established according to the performance degradation value x0 = X(t0) < ξ of the component at time t0 and the independent increment characteristic of the Gamma process:

[0061]

[0062] When t0 = 0, the remaining life of the photovoltaic module is equivalent to its service life.

[0063] After obtaining the above remaining life prediction model of the photovoltaic module, the target photovoltaic module is predicted through this remaining life prediction model. To solve the problems of sparsity and fragmentation of the input degradation data, the embodiment of the present invention proposes to adaptively update the model parameters according to the processed data of the cumulative power degradation amount of the target photovoltaic module at the current time, so that the entire remaining life prediction process is based on the degradation trajectory of the target component itself, and in the data preprocessing stage, the influence of random effects on the component degradation process is considered.

[0064] In one embodiment, the degradation data of the photovoltaic module is obtained, and for the missing or sparse degradation data, GPR is used for interpolation processing;

[0065] The core idea of GPR is to estimate the distribution of the missing data by constructing a covariance function. For a given data set in the following form , for any new time point Predict its degradation metric X( ):

[0066]

[0067] wherein, represents the new time point of interpolation, represents the time point corresponding degradation data, D represents the degradation data of the photovoltaic module, represents the covariance function of the new point and the degradation data D, C represents the covariance matrix of the degradation data, y represents the degradation value of the degradation data, , represents the degradation value corresponding to the nth time point ; represents the sampling moment of the nth training data point, and n represents the total number of training data points.

[0068] In this embodiment, in order to ensure that the interpolation can smooth the global trend and retain local mutations, the present invention preferably weights and superimposes the anisotropic squared exponential (RBF) kernel and kernel to construct a covariance function:

[0069]

[0070] wherein, represents two arbitrary time input points, belonging to the training sample time set, and is used to construct the kernel function in the Gaussian process regression, is the amplitude hyperparameter, is the characteristic length scale, and is adaptively learned by maximizing the marginal log-likelihood so that the model can achieve the optimal interpolation performance in the full sample space. Among them,

[0071] maximizing the marginal log-likelihood is:

[0072]

[0073] wherein, K represents the covariance matrix constructed from the training data, and its elements are the kernel function values between each sampling point, n represents the number of training data points, that is, the total number of observed degradation data points, represents the logarithmic function with the natural constant e as the base.

[0074]

[0075] wherein, represents the amplitude hyperparameter of the radial basis function (RBF) kernel, and is used to control the overall contribution intensity of the kernel function to the covariance matrix; represents the characteristic length scale of the radial basis function kernel, which is used to measure the influence range of the similarity between data on the covariance. represents the amplitude hyperparameter of the kernel, indicating its contribution intensity; represents the length scale of the kernel, which is used to control the smoothness and perception range of the kernel function, represents the standard deviation of the noise term, which reflects the intensity of the observation error in the degraded data and is often added to the diagonal as the "noise kernel" in the covariance function.

[0076] In this way, GPR can interpolate the missing data and smooth the influence of noise at the same time.

[0077] Furthermore, on this basis, the present application gives a closed-form expression of the prediction variance to quantify the uncertainty of the interpolation result, and its calculation formula is as follows:

[0078]

[0079] In the formula, is the covariance of the new sample itself. As can be seen from the above formula, if there are few observation points near it or the sample noise is large, then the prediction variance increases; otherwise, it decreases, which provides a theoretical basis for the subsequent construction of the confidence interval and active data supplementation.

[0080] In one embodiment, the joint sliding window and maximum likelihood estimation method are used to adaptively update the parameters of the remaining life distribution function of the photovoltaic module at the moment to be measured in real time;

[0081] As a local fitting method, the sliding window fitting segments the data set and separately fits the target distribution parameters within each window, so as to obtain the dynamic change law of the data with time or other variables. It has the characteristics of being simple and easy to implement, strong in dynamics, and high in robustness. In the present application, the sliding window method is used to dynamically fit the cumulative power decay data of the photovoltaic module to obtain a better decay trend.

[0082] Furthermore, for the data within the sliding window, the maximum likelihood estimation method is used to estimate the parameters k and of the Gamma process, including:

[0083] Determine the maximum likelihood function of the data within the sliding window;

[0084] Take the logarithm to obtain the log-likelihood function;

[0085] Maximize the log-likelihood function to update the parameters in real time.

[0086] Assume that the degraded data within the time window is , and the weight of each data point is w(ti ), then the maximum likelihood function is as follows:

[0087]

[0088] In the formula, m is the number of data in the sliding window, i is the data index in the sliding window, represents the difference between adjacent degradation data in the sliding window, , represents the weight of the i-th data point, represents the sampling time of the i-th degradation data point in the sliding window, k and respectively represent the shape parameter and scale parameter in the remaining life distribution function.

[0089] Taking the logarithm can obtain the log-likelihood function. By maximizing the log-likelihood function, the estimated values of k and are obtained.

[0090] In an optional embodiment, for the estimation of parameters k and , the partial derivatives are respectively taken and set to zero, and the following equations are obtained:

[0091]

[0092] In the formula: represents the logarithmic function with the natural constant e as the base, ; represents the Digamma function, and its definition is: .

[0093] By solving the above equations, the parameter estimation values and of the degradation process at the current time t0 can be obtained.

[0094] The and dynamically updated using the maximum likelihood estimation method can be used to substitute into the original model to achieve the purpose of real-time predicting the remaining life of the photovoltaic module.

[0095] The expected value and variance of the remaining life R(t0) are:

[0096]

[0097]

[0098] The above two formulas are used to quantify the statistical characteristics of the remaining life of the component at the current time t0, and are the standard derivation results based on the dynamic Gamma process. Their functions are as follows:

[0099] E represents the expected remaining life of the PV module from the current time t0 until failure;

[0100] Var represents the uncertainty of the remaining life prediction, that is, the fluctuation range of the confidence interval.

[0101] These two indicators provide quantitative support for maintenance decisions and can be used to set the optimal preventive maintenance time window or adjust thresholds in intelligent operation and maintenance strategies.

[0102] In an exemplary embodiment, the overall photovoltaic module life prediction process is referred to Figure 2 ,include:

[0103] S1. Input degradation data;

[0104] S2, GPR interpolation processing; Gaussian smoothing to remove noise;

[0105] S3. Check whether the degradation increment is passed. If not, return to re-interpolation processing; if so, execute S4. In this embodiment, the degradation increment check is to check the temporal monotonicity and physical consistency of the degradation sequence obtained by GPR regression interpolation. This includes eliminating or correcting non-physical jumps or regressions caused by overfitting, edge effects, etc. in the interpolation, ensuring that the interpolated data conforms to the mathematical assumptions of Gamma process modeling.

[0106] S4. Set the sliding window and initialize the Gamma distribution parameters;

[0107] S5, MLE real-time update parameters, determine whether there is new data added, if so, return to S2 and re-enter the interpolation process; if not, execute S6;

[0108] S6, output the latest model parameters;

[0109] S7. Perform real-time life prediction.

[0110] To facilitate the real-time implementation of the life prediction solution of this application, an embodiment discloses a photovoltaic module life prediction system, which applies any of the photovoltaic module life prediction methods described above, including:

[0111] The remaining life distribution function determination module is used to determine the remaining life distribution function of the photovoltaic module at the time of testing based on the dynamic Gamma process degradation model of the photovoltaic module; the degradation data acquisition and processing module is used to obtain the degradation data of the photovoltaic module and perform interpolation processing using GPR; the life distribution function parameter update module is used to update the parameters of the remaining life distribution function of the photovoltaic module at the time of testing in real time based on the interpolated photovoltaic module degradation data, combined with the sliding window and maximum likelihood estimation method; the remaining life prediction module uses the remaining life distribution model after parameter update to predict the life of the photovoltaic module.

[0112] Preferably, it further includes a transmission module for transmitting the life prediction result of the photovoltaic module at the moment to be measured to the cloud platform.

[0113] Or provide a memory with a processor installed inside. The processor is used to execute the above-mentioned photovoltaic module life prediction method, and transmit the prediction result to the operation and maintenance center control room through a transmitter based on a wireless network (4G / 5G). For details, refer to Figure 3 .

[0114] To verify the effect of this application, it is illustrated by the following simulation experiments:

[0115] Based on the output power degradation data of the photovoltaic modules in a certain photovoltaic array within 12 years, a life prediction model of photovoltaic modules based on GPR and dynamic Gamma process is constructed, and the remaining life is adaptively updated to verify the feasibility of the proposed method. It is known that its remaining service life is 7.4 years.

[0116] In order to simulate the sparsification and fragmentation of data records in the actual operation process, the recording interval of the recession data is artificially extended to 0.25 years. That is, at the same time, the prediction results of the first 5 years, the first 6 years, and the first 7 years are used to simulate the real-time update process of the life prediction failure, which are respectively recorded as the first monitoring, the second monitoring, and the third monitoring. The real-time processing and fitting results of the cumulative power recession data at different monitoring times are shown in Figure 4(a), Figure 4(b), and Figure 4(c). The fitting verification of the final monitoring data is as Figure 5 shown. Figure 5 The results show that in the inspections of each segment, the sample data points are all close to a straight line, and the degradation data meets the requirements of the selected model for data.

[0117] Subsequently, the remaining service life of the photovoltaic module is predicted in real time by using the model. First, the model parameters are updated in real time by using the degradation data of the module. The parameter update process is as Figure 6 shown. In the model, the shape parameter and the scale parameter respectively characterize the cumulative characteristics of the degradation process and the change of the degradation intensity per unit time. Specifically, the shape parameter reflects the degradation speed and trend. An increase indicates an acceleration of degradation, reflecting the long-term stability of the system; the scale parameter describes the amplitude of the degradation amount per unit time. An increase indicates an increase in the degradation intensity, reflecting the influence of the external environment or operating conditions. The dynamic adjustment of the two enables the model to flexibly adapt to different stages of the degradation process, accurately describe the complex degradation behavior of the photovoltaic module, and provide a reliable basis for health state assessment and remaining life prediction.

[0118] Based on the obtained model parameters, the real-time remaining life of the photovoltaic module is predicted based on the concept of first passage time. The failure threshold of the photovoltaic module is set to ξ = 20%, that is, when the cumulative power degradation rate reaches 20%, the module is considered to have degraded and failed. According to the real-time estimated values of the model parameters and the probability density function of the remaining life, the real-time monitoring results of the degradation process of the photovoltaic module at each monitoring time point and the probability distribution results of the remaining life can be obtained as Figure 7 and Figure 8 shown. In the process of predicting the remaining life, if t0 = 0, the obtained prediction result is the actual remaining service life; otherwise, combined with the current monitoring time, the actual remaining service life of the photovoltaic module can also be obtained.

[0119] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, please refer to the description in the method section.

[0120] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the lifespan of a photovoltaic module, characterized in that, Including: Based on the dynamic Gamma process degradation model of photovoltaic modules, determine the remaining life distribution function of the photovoltaic modules at the moment to be measured; Obtain the degradation data of the photovoltaic modules and perform interpolation processing using GPR; According to the interpolated degradation data of the photovoltaic modules, jointly use the sliding window and the maximum likelihood estimation method to update the parameters of the remaining life distribution function of the photovoltaic modules at the moment to be measured in real time; Use the remaining life distribution model with updated parameters to predict the life of the photovoltaic modules.

2. The photovoltaic module life prediction method according to claim 1, characterized in that Based on the dynamic Gamma process degradation model of photovoltaic modules, determine the remaining life distribution function of the photovoltaic modules, including: Based on the dynamic Gamma process degradation model of photovoltaic modules, calculate the time T when the cumulative power degradation amount first reaches the failure threshold; Determine the life distribution function of the photovoltaic modules according to the time T; Combined with the moment to be measured t0, determine the remaining life distribution function of the photovoltaic modules.

3. The method for predicting the lifespan of a photovoltaic module according to claim 1 or 2, characterized in that, The expression of the remaining life distribution function of the photovoltaic modules is: ； Wherein, t represents the moment corresponding to the time T when the cumulative power degradation amount first reaches the failure threshold, t0 represents the moment of the remaining life to be measured, γ(k, x) represents the upper incomplete Gamma function, ξ represents the failure threshold, and k represents the shape parameter; represents the scale parameter.

4. The method for predicting the lifespan of a photovoltaic module according to claim 3, wherein, The expression of γ(k, x) is: ； In the formula, k is the shape parameter of the Gamma process, and x is the normalized time.

5. The photovoltaic module life prediction method according to claim 1, wherein, The interpolation method is: ； In the formula, represents the new time point of interpolation, represents the time point corresponding to the degraded data, D represents the degradation data of the photovoltaic module, represents the new point and the covariance function of the degraded data D, C represents the covariance matrix of the degraded data, y represents the degradation value of the degraded data, , represents the nth time point corresponding to the degradation value, represents the sampling moment of the nth training data point, and n represents the total number of training data points.

6. The method for predicting the service life of a photovoltaic module according to claim 1, wherein Based on the anisotropic squared exponential kernel and the kernel is weighted and superimposed to construct a covariance function: ； In the formula, denote two arbitrary time input points, belonging to the training sample time set, and are used to construct the kernel function in Gaussian process regression. is the amplitude hyperparameter. is the characteristic length scale and is adaptively learned by maximizing the marginal log-likelihood. denotes the amplitude hyperparameter of the radial basis function kernel. denotes the characteristic length scale of the radial basis function kernel. denotes the amplitude hyperparameter of the kernel. denotes the length scale of the kernel.

7. The method for predicting the service life of a photovoltaic module according to claim 1, characterized in that The parameter real-time update includes: Determine the maximum likelihood function of the data within the sliding window; Take the logarithm to obtain the log-likelihood function; Maximize the log-likelihood function to update the parameters in real time.

8. The photovoltaic module life prediction method according to claim 7, characterized in that Maximum likelihood function is as follows: ； where m is the number of data in the sliding window, and i is the data index in the sliding window. represents the difference between adjacent degradation data in the sliding window. represents the weight of the i-th data point. represents the sampling time of the i-th degradation data point in the sliding window, and k and represent the shape parameter and scale parameter in the remaining life distribution function, respectively.

9. A photovoltaic module life prediction system, characterized in that, Applying the photovoltaic module life prediction method according to any one of claims 1-8, including: A remaining life distribution function determination module, configured to determine the remaining life distribution function of the photovoltaic modules at the moment to be measured based on the dynamic Gamma process degradation model of the photovoltaic modules; A degradation data acquisition and processing module, configured to obtain the degradation data of the photovoltaic modules and perform interpolation processing using GPR; A life distribution function parameter update module, configured to jointly use the sliding window and the maximum likelihood estimation method to update the parameters of the remaining life distribution function of the photovoltaic modules at the moment to be measured in real time according to the interpolated degradation data of the photovoltaic modules; A remaining life prediction module, which uses the remaining life distribution model with updated parameters to predict the life of the photovoltaic modules; 10. The photovoltaic module life prediction system according to claim 9, characterized in that, It further includes a transmission module, configured to transmit the life prediction result of the photovoltaic modules at the moment to be measured to the cloud platform.

Images