CN114818345B - Photovoltaic module residual life prediction method and prediction system - Google Patents
Photovoltaic module residual life prediction method and prediction system Download PDFInfo
- Publication number
- CN114818345B CN114818345B CN202210482627.XA CN202210482627A CN114818345B CN 114818345 B CN114818345 B CN 114818345B CN 202210482627 A CN202210482627 A CN 202210482627A CN 114818345 B CN114818345 B CN 114818345B
- Authority
- CN
- China
- Prior art keywords
- photovoltaic module
- degradation
- output power
- residual life
- parameter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000006731 degradation reaction Methods 0.000 claims abstract description 193
- 230000015556 catabolic process Effects 0.000 claims abstract description 155
- 230000005653 Brownian motion process Effects 0.000 claims description 41
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000012544 monitoring process Methods 0.000 claims description 12
- 238000009792 diffusion process Methods 0.000 claims description 10
- 239000013598 vector Substances 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005537 brownian motion Methods 0.000 claims description 5
- 230000001186 cumulative effect Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 4
- 230000003121 nonmonotonic effect Effects 0.000 description 4
- 238000010248 power generation Methods 0.000 description 4
- 239000002131 composite material Substances 0.000 description 3
- YTAHJIFKAKIKAV-XNMGPUDCSA-N [(1R)-3-morpholin-4-yl-1-phenylpropyl] N-[(3S)-2-oxo-5-phenyl-1,3-dihydro-1,4-benzodiazepin-3-yl]carbamate Chemical compound O=C1[C@H](N=C(C2=C(N1)C=CC=C2)C1=CC=CC=C1)NC(O[C@H](CCN1CCOCC1)C1=CC=CC=C1)=O YTAHJIFKAKIKAV-XNMGPUDCSA-N 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 2
- 239000000306 component Substances 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000011065 in-situ storage Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 229910021421 monocrystalline silicon Inorganic materials 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Computational Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Mathematical Analysis (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Computer Hardware Design (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Evolutionary Biology (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Probability & Statistics with Applications (AREA)
- Geometry (AREA)
- Algebra (AREA)
- Game Theory and Decision Science (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Photovoltaic Devices (AREA)
Abstract
The invention discloses a method and a system for predicting the residual life of a photovoltaic module, wherein the method comprises the following steps: acquiring the output power degradation amount of the target photovoltaic module at the current moment; updating preset model parameters according to the degradation amount of the output power; inputting the updated model parameters into a residual life prediction model, and outputting a residual life distribution result of the target photovoltaic module at the current moment; according to the method, when the degradation characteristics of the target photovoltaic module are described, the non-monotonicity, the randomness and the individual variability of the degradation of the photovoltaic module are considered, and the real-time output power degradation data can be utilized to update the result, so that the accuracy of the residual life prediction of the photovoltaic module is improved.
Description
Technical Field
The invention belongs to the technical field of photovoltaic power generation, and particularly relates to a method and a system for predicting the residual life of a photovoltaic module.
Background
The service life of the photovoltaic module serving as a core component of the photovoltaic power generation system is an important factor for determining the unit power generation cost of the photovoltaic system. With the increase of service life of the photovoltaic module and the influence of internal and external random factors, the performance of the photovoltaic module gradually declines, and the continuous accumulation of degradation amount can influence the reliability of photovoltaic power generation. Meanwhile, due to the low performance degradation rate of the photovoltaic module, it is difficult to collect long-term data to confirm degradation paths and life. Therefore, it is necessary to build a random degradation model to characterize the instability and blurring characteristics of the degradation of the performance of the photovoltaic module with time, so as to estimate the residual life of the photovoltaic module and improve the operation reliability of the photovoltaic module.
At present, the methods related to the prediction management of the residual life of the photovoltaic module mainly comprise two types: based on degradation mechanisms and based on degradation data driving. The degradation mechanism model-based method is only an empirical model under prior knowledge, and cannot be used for describing the actual degradation condition of the photovoltaic module. In the degradation data driving method, the degradation mechanism of the photovoltaic module is not required to be known in advance, a mathematical model is only required to be constructed by utilizing characteristic quantities related to the performance degradation of the photovoltaic module, and the residual life distribution is predicted based on model parameters. The method based on the degradation data driving mainly comprises an intelligent algorithm method and a data modeling method. The intelligent algorithm method does not need to model the operation process of the photovoltaic module, directly uses monitoring data to carry out machine learning, and then evaluates the field life of the module. However, the method of the intelligent algorithm generally requires a large amount of sample data, and cannot obtain interval estimation of the on-site residual life of the photovoltaic module, so that unstable and fuzzy characteristics of the residual life of the module with time are difficult to express. The data modeling method has the advantages that the photovoltaic module performance degradation process is modeled based on the on-site state monitoring data, and the unstable and fuzzy characteristics of the residual life prediction result can be effectively quantified.
The method for predicting the residual life of the photovoltaic module based on data modeling at present basically adopts a Gamma process to establish a performance degradation model of the photovoltaic module, namely, the degradation process of the photovoltaic module is strictly monotonous is assumed in advance, the difference between the degradation process and the actual degradation process of the photovoltaic module is ignored, and the error of the life prediction of the photovoltaic module can be increased.
Therefore, how to characterize the degradation characteristics of the current photovoltaic module, considering the non-monotonicity, randomness and individual variability presented by the degradation of the photovoltaic module, so as to improve the accuracy of the residual life prediction of the photovoltaic module is a technical problem to be solved currently.
Disclosure of Invention
In view of the above problems, the present invention provides a method and a system for predicting the remaining life of a photovoltaic module, which at least solve some of the above technical problems, by which, when describing degradation characteristics of a current photovoltaic module, non-monotonicity, randomness and individual variability presented by degradation of the photovoltaic module are considered, so as to improve the accuracy and certainty of predicting the remaining life of the photovoltaic module.
In a first aspect, an embodiment of the present invention provides a method for predicting a remaining lifetime of a photovoltaic module, including:
s1, obtaining the output power degradation amount of a target photovoltaic module at the current moment;
s2, updating preset model parameters according to the degradation amount of the output power;
and S3, inputting the updated model parameters into a residual life prediction model, and outputting a residual life distribution result of the target photovoltaic module at the current moment.
Further, the step S1 specifically includes:
s11, acquiring an initial value of the output power of the target photovoltaic module and the output power of the target photovoltaic module at the current moment;
and S12, calculating the output power degradation amount of the target photovoltaic module at the current moment based on the data obtained in the step S11.
Further, the step S2 specifically includes:
s21, carrying out explicit solution on the preset model parameters based on a Bayesian updating algorithm and an expectation maximization algorithm;
s22, setting an initial value for the solved model parameters, and carrying out self-adaptive updating on the preset model parameters according to the output power degradation of the target photovoltaic module at the current moment.
Further, the construction method of the residual life prediction model is as follows:
acquiring the output power degradation amount of the photovoltaic module at a preset moment;
constructing a photovoltaic module Wiener process degradation model according to the output power degradation amount;
and deducing a residual life distribution result of the photovoltaic module at the current moment based on the degradation model of the photovoltaic module Wiener process to obtain a residual life prediction model of the photovoltaic module.
Further, the photovoltaic module Wiener process degradation model is expressed as:
Y(t)＝Y(0)+vt+σB(t) (2)
wherein Y (t) represents the cumulative output power degradation amount of the photovoltaic module; y (0) represents an initial value of the degradation amount of the output power of the photovoltaic module; v represents the drift parameter of the photovoltaic module in the degradation process; sigma represents a diffusion parameter of the photovoltaic module in a degradation process; b (t) represents standard Brownian motion; t represents a preset time, and t is more than or equal to 0.
Further, the deriving a remaining lifetime distribution result of the photovoltaic module at the current moment based on the photovoltaic module Wiener process degradation model specifically includes:
calculating the time T when the output power degradation quantity reaches a failure threshold value for the first time based on the photovoltaic module Wiener process degradation model;
converting the time T when the output power degradation quantity reaches a failure threshold value for the first time into the residual life of the photovoltaic module at the preset moment;
and according to the residual life of the photovoltaic module at the preset time, combining a Wiener process to obtain a residual life distribution result of the photovoltaic module at the preset time.
On the other hand, the embodiment of the invention also provides a system for predicting the residual life of the photovoltaic module, which is applied to the method, and comprises an acquisition module for acquiring the output power degradation of the target photovoltaic module at the current moment;
the processing module is used for updating preset model parameters according to the output power degradation amount;
and the prediction module is used for outputting a residual life distribution result of the target photovoltaic module at the current moment according to the updated model parameters.
Further, the device also comprises a transmission module;
the transmission module is used for transmitting the residual life distribution result of the target photovoltaic module at the current moment to a cloud platform.
Compared with the prior art, the photovoltaic module residual life prediction method and the photovoltaic module residual life prediction system have the following beneficial effects:
on one hand, due to the composite effect of random factors such as natural environment, mechanical stress and the like, the degradation process taking the output power of the photovoltaic module as the degradation characteristic quantity can show non-monotonicity, randomness and individual variability, and compared with the current common data modeling method, the Wiener process can better characterize the degradation characteristic of the photovoltaic module.
On the other hand, the invention solves the requirement of self-adaptive prediction of the residual life of the photovoltaic module. Based on the degradation track of the photovoltaic module, the parameters of the Wiener model are adaptively updated in real time by combining Bayesian updating and an expected maximization algorithm, and the residual life distribution of the photovoltaic module is predicted on the basis. Therefore, the method can effectively solve the defects of the existing data modeling method, improves the accuracy of the residual life prediction of the photovoltaic module, and enables the residual life prediction method to meet the actual application requirements.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
fig. 1 is a flowchart of a method for predicting the residual life of a photovoltaic module according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of degradation of output power of a photovoltaic module according to an embodiment of the present invention.
FIG. 3 is a schematic diagram of a degradation incremental distribution verification provided by an embodiment of the present invention.
Fig. 4 is an iteration schematic diagram of parameters without considering individual variability according to an embodiment of the present invention.
Fig. 5 is an iteration schematic diagram of parameters considering individual variability according to an embodiment of the present invention.
Fig. 6 (a) is a schematic diagram of a residual life prediction result without considering individual variability according to an embodiment of the present invention.
Fig. 6 (b) is a schematic diagram of a residual life prediction result considering individual variability according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The photovoltaic module is mainly installed outdoors, and the performance degradation process of the module can exhibit non-monotonic characteristics in consideration of the influence of natural factors and the like (such as shading and the like). The performance degradation of the photovoltaic module refers to a degradation failure process caused by the composite effect of factors such as natural environment, mechanical stress and the like on the module, namely the performance of the module gradually declines with the service time until the module fails, and the degradation process shows time uncertainty.
Referring to fig. 1, an embodiment of the present invention provides a method for predicting a remaining lifetime of a photovoltaic module, which specifically includes the following steps:
s1, obtaining the output power degradation amount of a target photovoltaic module at the current moment;
s2, updating preset model parameters according to the degradation amount of the output power;
and S3, inputting the updated model parameters into a residual life prediction model, and outputting a residual life distribution result of the target photovoltaic module at the current moment.
In the step S1, firstly, an initial value of the output power of the target photovoltaic module and the output power of the target photovoltaic module at the current moment are obtained; and secondly, calculating the degradation amount of the output power of the target photovoltaic module at the current moment based on the obtained initial value of the output power of the target photovoltaic module and the output power of the target photovoltaic module at the current moment.
Next, the above steps S2 and S3 will be described in detail.
First, a method for constructing a residual life prediction model will be described.
The degradation model for constructing the equipment service process based on the Wiener process is generally established by introducing a function related to the equipment performance degradation process on the basis of the standard Wiener process. If the degradation description in the equipment service process is { Y (t), t is greater than or equal to 0}, a degradation model constructed based on the Wiener process is shown as follows:
Y(t)＝Y(0)+vZ(t,α η )+σB(t) (1)
wherein Y (t) represents the cumulative output power degradation amount of the photovoltaic module; y (0) represents an initial value of the degradation amount of the output power of the photovoltaic moduleThe method comprises the steps of carrying out a first treatment on the surface of the v represents a drift parameter of the photovoltaic module in the degradation process, represents the degradation speed of the photovoltaic module, and can show individual variability of the photovoltaic module in the degradation process; sigma represents diffusion parameters of the photovoltaic module in the degradation process, and represents uncertainty of the photovoltaic module in the degradation process; z (t, alpha) η ) A function representing a degradation trend of the photovoltaic module; b (t) represents standard Brownian motion and represents random dynamic characteristics of the photovoltaic module in a degradation process; t represents a preset time, and t is more than or equal to 0.
The Wiener process has non-monotonic stable and independent Gaussian increment, and is suitable for describing the non-monotonic degradation trend of equipment failure caused by continuous accumulation of degradation. The output power is selected as the characteristic quantity for describing the performance degradation of the photovoltaic module, so that the mechanism of the time-varying service condition of the module can be well explained. The output power degradation amount of the photovoltaic module is in a linear rising trend along with the whole time in the service process, and meanwhile, due to the composite effect of natural uncertain factors and the like, the degradation process of the module can generate variable time uncertainty, so that degradation randomly fluctuates in a small range. Furthermore, the power of the photovoltaic module follows a gaussian distribution. This is consistent with the linear non-monotonic degradation trend described by the Wiener model.
Therefore, in this embodiment, for the degradation process { Y (t), t is greater than or equal to 0} of the photovoltaic module, the built degradation model of the photovoltaic module Wiener process based on the Wiener process, that is, the degradation model of the photovoltaic module based on the Wiener process is expressed as:
Y(t)＝Y(0)+vt+σB(t) (2)
wherein Y (t) represents the cumulative output power degradation amount of the photovoltaic module; y (0) represents an initial value of the degradation amount of the output power of the photovoltaic module; v represents a drift parameter of the photovoltaic module in the degradation process, represents the degradation speed of the photovoltaic module, and can show individual variability of the photovoltaic module in the degradation process; sigma represents diffusion parameters of the photovoltaic module in the degradation process, and represents uncertainty of the photovoltaic module in the degradation process; b (t) represents standard Brownian motion and represents random dynamic characteristics of the photovoltaic module in a degradation process; t represents a preset time, and t is more than or equal to 0.
In the present embodiment of the present invention, in the present embodiment,taking an initial value Y (0) of the degradation amount of the output power of the photovoltaic module as 0; if the degradation amount of the output power of the photovoltaic module is not 0 in the actual monitoring process, translation exchange can be performed on the photovoltaic module first, and the degradation amount at the initial moment is converted into 0. According to the relevant definition of the Wiener process, the degradation increment delta Y (t) of the photovoltaic module at any moment meets the property of Gaussian distribution, namely delta Y (t) to N (vdelta t, sigma) 2 Δt). Δt is the monitoring time interval, Δt=t i -t i-1 The method comprises the steps of carrying out a first treatment on the surface of the Δy is the output power degradation increment, Δy=y i -Y i-1 。
The random change of the health state of the photovoltaic module along with time is a characteristic of the service process of the photovoltaic module, so that the residual life of the module also has random characteristics, and therefore, the residual life prediction requirement of the photovoltaic module is solved by a probability density function of the residual life. The uncertainty of the residual life of the photovoltaic module is characterized by a conditional random variable, which is expressed as that from the actual monitoring moment, the module is degraded to a failure threshold value y for the first time th Time interval elapsed. And predicting the residual life of the photovoltaic module based on the performance degradation process of the photovoltaic module, namely judging whether the module degradation reaches a preset failure threshold value and when the module degradation reaches the failure threshold value. In general, the degradation amount of the output power of the photovoltaic module in service reaches 20% of the initial power to be used as a threshold value of degradation and failure of the module, and in the actual prediction process, the failure threshold value can be set and changed according to the actual engineering requirement of the module in service.
According to a photovoltaic module Wiener process degradation model with a linear drift Wiener process, the service life T of the photovoltaic module is defined as: the output power degradation Y (t) of the component reaches the failure threshold Y for the first time th Time (first time) of (a):
T＝inf{t:Y(t)≥y th |Y(0)＜y th } (3)
wherein Y (t) represents the accumulated output power degradation amount of the target photovoltaic module; y (0) represents an initial value of the output power degradation amount of the target photovoltaic module; t represents a preset time, and the preset time can be set as a current time in the actual monitoring process; y is th Representing a failure threshold.
The first time of drifting Wiener model obeys inverse Gaussian distribution, and the time scale at the first time is obtainedConversion of degree into photovoltaic modules t k Remaining lifetime of time L k ：
L k ＝inf{l k :Y(t k +l k )≥y th |Y(t k )＜y th } (4)
in the formula ,Y(tk +l k )、Y(t k ) Respectively represent t k +l k 、t k The degradation amount of the output power of the photovoltaic module at any moment; y is th Representing a failure threshold; l (L) k Representing a photovoltaic module t k Remaining lifetime of time L k 。
The above formula (4) is further expressed as:
in the formula ,representing the slave time t k A model of the power degradation at the beginning, known from the nature of the standard Wiener process, is +.>The degradation process still characterized by the drift Wiener model is known to be:
thus, the photovoltaic module is at t k The remaining lifetime distribution at the moment is:
the equipment residual life prediction based on the data modeling method mainly comprises the steps of firstly using service degradation data of similar equipment to estimate prior values of model parameters, and then updating residual life distribution. In practice, however, there are differences in degradation between individuals even with similar devicesSex. Based on this, effective residual life prediction relies on the degradation process of the target device itself while taking into account the effect of individual variability on prediction uncertainty. Characterizing individual variability existing in the degradation process of the photovoltaic module by using v random variation in the degradation model of the photovoltaic module Wiener process, and obtaining t based on a full probability formula k Time-dependent degradation sequence Y 0:k ＝[y 0 ,...,y k ]The remaining life prediction model of (2) is:
in the formula ,Y0:k Representing t k Sequence of degradation amount at time, Y 0:k ＝[y 0 ,...,y k ](k≥1)；p(v|Y 0:k ) Representing the posterior distribution of v.
The above matters are the photovoltaic module residual life prediction model provided by the embodiment of the invention.
Markov is available from standard Wiener processes:
then in equation (8):
from equations (7) and (8), it can be shown that updating the posterior distribution of v and Θ= (v, σ) yields a real-time remaining lifetime that depends on the photovoltaic module degradation trajectory.
After the residual life prediction model of the photovoltaic module is obtained, the target photovoltaic module can be predicted through the residual life prediction model. In order to reduce uncertainty of a prediction result, the embodiment of the invention provides that according to the output power degradation amount of a target photovoltaic module at the current moment, model parameters are adaptively updated, so that the whole residual life prediction process is based on the degradation track of the target module; the specific content of the self-adaptive update of the model parameters is as follows:
taking individual variability presented in degradation process of photovoltaic module into consideration, and makingμ v,0 ,/>Is the prior value of the v super parameter; mu (mu) v,k ,/>For the estimated value of v super parameter based on the degradation track, the prior distribution of v is:
under the drive of a standard Wiener process, based on the given model parameter v, a photovoltaic module degradation sequence Y 0:k Is a multiple Gaussian distribution, v|Y 0:k Also in gaussian distribution, i.e.:
wherein Δt represents the monitoring time interval, Δt=t i -t i-1 The method comprises the steps of carrying out a first treatment on the surface of the Δy represents an increase in output power degradation, Δy=y i -Y i-1 。
Calculating a posterior distribution of v based on bayesian update:
wherein ,μv,0 Andrepresenting a priori values of the super parameter v; mu (mu) v,k and />Representing the estimated value of the hyper-parameter v based on the degradation trajectory.
From model parametersIt can be seen that v|Y 0:k Also in gaussian distribution, i.e.:
deriving the formula (13) and the formula (14) to obtain the v-super parameter at t k The time of day is based on an estimate of the degradation trajectory:
from equation (15), it can be shown that the posterior distribution of v can be adaptively updated when new degradation data is monitored.
After obtaining the posterior distribution of v, it is necessary to use the in-situ monitoring data sequence Y 0:k Estimating unknown parameter vectors in a degradation modelTo reduce the dependence of the residual life prediction on a priori information, an expectation maximization algorithm (EM algorithm) is employed to estimate the unknown parameter vector, specifically comprising:
the steps are as follows: calculation of
in the formula ,representing the parameter vector estimate of step j.
Maximizing: calculation of
Derived from the above-described expectation step and maximization step, it is possible to obtain:
wherein ,representing the estimated value of the mean value of the drift parameter in the j+1th step; />Representing the estimated value of the drift parameter variance in the j+1th step; />The estimated value of the diffusion parameter at step j+1 is shown.
And (3) gradually updating the formula (15) and the formula (18) to obtain parameters in the photovoltaic module performance degradation model. The explicit solving process of the parameters is quick and simple, and the updating efficiency of the model parameters can be greatly improved.
According to the residual life prediction model in the formula (8) and the updating results of the model parameters of the following formulas (15) - (18), the whole monitoring data sequence Y of the photovoltaic module can be obtained 0:k At t k Residual lifetime distribution results at time:
the embodiment of the invention provides a photovoltaic module residual life prediction system, which is applied to the method and comprises an acquisition module, a processing module, a prediction module and a transmission module; the acquisition module is used for acquiring the output power degradation amount of the target photovoltaic module at the current moment; the processing module is used for updating preset model parameters according to the degradation amount of the output power; the prediction module is used for outputting a residual life distribution result of the target photovoltaic module at the current moment according to the updated model parameters; the transmission module is used for transmitting the residual life distribution result of the target photovoltaic module at the current moment to the cloud platform.
Finally, the simulation experiment results corresponding to the embodiment of the invention are explained.
And constructing a Wiener model based on 15-year output power degradation data of a monocrystalline silicon photovoltaic module forming a certain 5kW photovoltaic array, and completing self-adaptive updating of the residual life, thereby verifying the feasibility and superiority of the method. The output power degradation data of the photovoltaic module for 15 years is processed first to obtain the output power degradation amount in the service period, and the result is shown in fig. 2. The degradation characteristic quantity increment described according to the Wiener degradation process has Gaussian characteristic, and the output power degradation increment of the photovoltaic module is subjected to Gaussian distribution test, and the result is shown in figure 3. The test result of fig. 3 shows that the power degradation incremental data of the photovoltaic module is basically presented on a straight line, and accords with the gaussian distribution characteristic.
Predicting the real-time remaining life of a photovoltaic module based on a Wiener process first requires real-time updating of model parameters using degradation data of the module. To further verify the effectiveness of the proposed method, two different cases are considered in the embodiments of the present invention:
1) Regardless of individual variability of the photovoltaic module in the degradation process, namely v and sigma are deterministic parameters, the result is shown in fig. 4 by using the model parameter estimation method provided by the embodiment of the invention. The results in fig. 4 show that the iterative process fluctuates greatly and the time for the parameters to stabilize is slightly longer.
2) Considering individual variability of the photovoltaic module in the degradation process, namelyFused bayesian and expectation maximization algorithmThe method updates the model parameters in real time, and the result is shown in fig. 5. The given initial value and the stable value of the model parameter have larger difference, and mainly aim to verify that the prior information of the model parameter has little dependence when the residual life prediction is carried out by adopting the method, and the result of fig. 5 also shows that the whole iterative process basically has no fluctuation and reaches the stable value quickly.
On the basis of obtaining model parameters, the real-time residual life of the photovoltaic module is predicted based on the concept of the first time. The failure threshold value of the photovoltaic module is set as y th =20w, i.e. component degradation failure is considered when the output power degradation increment reaches 20W. The probability distribution of the residual life of the photovoltaic module at each monitoring time point can be obtained according to the real-time estimated value of the model parameter and the probability density function of the residual life, and the point estimation of the residual life can be further obtained, and the result is shown in fig. 6. As can be seen from fig. 6 (a) and 6 (b), the predicted residual life in case 2 matches the actual value more, and the probability density curve of residual life becomes narrower and higher as the degradation data increases, i.e., the predicted result becomes more accurate, and the predicted value becomes closer to the actual value.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (5)
1. The method for predicting the residual life of the photovoltaic module is characterized by comprising the following steps of:
s1, obtaining the output power degradation amount of a target photovoltaic module at the current moment;
s2, updating preset model parameters according to the degradation amount of the output power;
s3, inputting the updated model parameters into a residual life prediction model, and outputting a residual life distribution result of the target photovoltaic module at the current moment;
according to the output power degradation amount, carrying out self-adaptive updating on preset model parameters:
taking individual variability presented in degradation process of photovoltaic module into consideration, and making Is the prior value of the v super parameter; />For the estimated value of v super parameter based on the degradation track, the prior distribution of v is:
under the drive of a standard Wiener process, based on the given model parameter v, a photovoltaic module degradation sequence Y 0:k Is a multiple Gaussian distribution, v|Y 0:k Also in gaussian distribution, i.e.:
wherein sigma represents a diffusion parameter of the photovoltaic module in a degradation process; Δt represents the monitoring time interval, Δt=t i -t i-1 The method comprises the steps of carrying out a first treatment on the surface of the Δy represents an output power degradation increment, Δy=y i -y i-1 ；
Calculating a posterior distribution of v based on bayesian update:
wherein ,μv,0 Andrepresenting a priori values of the super parameter v; mu (mu) v,k and />Representing an estimated value of the super parameter v based on the degradation track;
from model parametersIt can be seen that v|Y 0:k Also in gaussian distribution, i.e.:
deriving the formula (13) and the formula (14) to obtain the v-super parameter at t k The time of day is based on an estimate of the degradation trajectory:
after v posterior distribution is obtained, the degradation sequence Y of the photovoltaic module is utilized 0:k Estimating unknown parameter vectors in a degradation model
The expectation maximization algorithm (EM algorithm) is used to estimate the unknown parameter vector, and specifically includes:
the steps are as follows: calculation of
in the formula ,representing the parameter vector estimation value of the j th step;
maximizing: calculation of
Derived from the above-described expectation step and maximization step, it is possible to obtain:
wherein ,representing the estimated value of the mean value of the drift parameter in the j+1th step; />Representing the estimated value of the drift parameter variance in the j+1th step; />Representing the estimated value of the diffusion parameter in the j+1th step;
the construction method of the residual life prediction model comprises the following steps:
acquiring the output power degradation amount of the photovoltaic module at a preset moment;
constructing a photovoltaic module Wiener process degradation model according to the output power degradation amount;
deducing a residual life distribution result of the photovoltaic module at the current moment based on the degradation model of the photovoltaic module Wiener process to obtain a residual life prediction model of the photovoltaic module;
the photovoltaic module Wiener process degradation model is expressed as:
Y(t)＝Y(0)+vt+σB(t) (2)
wherein Y (t) represents the cumulative output power degradation amount of the photovoltaic module; y (0) represents an initial value of the degradation amount of the output power of the photovoltaic module; v represents the drift parameter of the photovoltaic module in the degradation process; sigma represents a diffusion parameter of the photovoltaic module in a degradation process; b (t) represents standard Brownian motion; t represents a preset time, and t is more than or equal to 0;
the deriving the residual life distribution result of the photovoltaic module at the current moment based on the photovoltaic module Wiener process degradation model specifically comprises the following steps:
calculating the time T when the output power degradation quantity reaches a failure threshold value for the first time based on the photovoltaic module Wiener process degradation model;
converting the time T when the output power degradation quantity reaches a failure threshold value for the first time into the residual life of the photovoltaic module at the preset moment;
and according to the residual life of the photovoltaic module at the preset time, combining a Wiener process to obtain a residual life distribution result of the photovoltaic module at the preset time.
2. The method for predicting the remaining life of a photovoltaic module according to claim 1, wherein S1 specifically comprises:
s11, acquiring an initial value of the output power of the target photovoltaic module and the output power of the target photovoltaic module at the current moment;
and S12, calculating the output power degradation amount of the target photovoltaic module at the current moment based on the data obtained in the step S11.
3. The method for predicting the remaining life of a photovoltaic module according to claim 1, wherein S2 specifically comprises:
s21, carrying out explicit solution on the preset model parameters based on a Bayesian updating algorithm and an expectation maximization algorithm;
s22, setting an initial value for the solved model parameters, and carrying out self-adaptive updating on the preset model parameters according to the output power degradation of the target photovoltaic module at the current moment.
4. A photovoltaic module remaining life prediction system, characterized by applying the method of any of claims 1-3, the system comprising: the system comprises an acquisition module, a processing module and a prediction module;
the acquisition module is used for acquiring the output power degradation amount of the target photovoltaic module at the current moment;
the processing module is used for updating preset model parameters according to the output power degradation amount;
the prediction module is used for outputting a residual life distribution result of the target photovoltaic module at the current moment according to the updated model parameters;
according to the output power degradation amount, carrying out self-adaptive updating on preset model parameters:
taking individual variability presented in degradation process of photovoltaic module into consideration, and making Is the prior value of the v super parameter; />For the estimated value of v super parameter based on the degradation track, the prior distribution of v is:
under the drive of a standard Wiener process, based on the given model parameter v, a photovoltaic module degradation sequence Y 0:k Is a multiple Gaussian distribution, v|Y 0:k Also in gaussian distribution, i.e.:
wherein sigma represents a diffusion parameter of the photovoltaic module in a degradation process; Δt represents the monitoring time interval, Δt=t i -t i-1 The method comprises the steps of carrying out a first treatment on the surface of the Δy represents an output power degradation increment, Δy=y i -y i-1 ；
Calculating a posterior distribution of v based on bayesian update:
wherein ,μv,0 Andrepresenting a priori values of the super parameter v; mu (mu) v,k and />Representing an estimated value of the super parameter v based on the degradation track;
from model parametersIt can be seen that v|Y 0:k Also in gaussian distribution, i.e.:
deriving the formula (13) and the formula (14) to obtain the v-super parameter at t k The time of day is based on an estimate of the degradation trajectory:
after v posterior distribution is obtained, the degradation sequence Y of the photovoltaic module is utilized 0:k Estimating unknown parameter vectors in a degradation model
The expectation maximization algorithm (EM algorithm) is used to estimate the unknown parameter vector, and specifically includes:
the steps are as follows: calculation of
in the formula ,representing the parameter vector estimation value of the j th step;
maximizing: calculation of
Derived from the above-described expectation step and maximization step, it is possible to obtain:
wherein ,representing the estimated value of the mean value of the drift parameter in the j+1th step; />Representing the estimated value of the drift parameter variance in the j+1th step; />Representing the estimated value of the diffusion parameter in the j+1th step;
the construction method of the residual life prediction model comprises the following steps:
acquiring the output power degradation amount of the photovoltaic module at a preset moment;
constructing a photovoltaic module Wiener process degradation model according to the output power degradation amount;
deducing a residual life distribution result of the photovoltaic module at the current moment based on the degradation model of the photovoltaic module Wiener process to obtain a residual life prediction model of the photovoltaic module;
the photovoltaic module Wiener process degradation model is expressed as:
Y(t)＝Y(0)+vt+σB(t) (2)
wherein Y (t) represents the cumulative output power degradation amount of the photovoltaic module; y (0) represents an initial value of the degradation amount of the output power of the photovoltaic module; v represents the drift parameter of the photovoltaic module in the degradation process; sigma represents a diffusion parameter of the photovoltaic module in a degradation process; b (t) represents standard Brownian motion; t represents a preset time, and t is more than or equal to 0;
the deriving the residual life distribution result of the photovoltaic module at the current moment based on the photovoltaic module Wiener process degradation model specifically comprises the following steps:
calculating the time T when the output power degradation quantity reaches a failure threshold value for the first time based on the photovoltaic module Wiener process degradation model;
converting the time T when the output power degradation quantity reaches a failure threshold value for the first time into the residual life of the photovoltaic module at the preset moment;
and according to the residual life of the photovoltaic module at the preset time, combining a Wiener process to obtain a residual life distribution result of the photovoltaic module at the preset time.
5. The photovoltaic module remaining life prediction system of claim 4, further comprising a transmission module;
the transmission module is used for transmitting the residual life distribution result of the target photovoltaic module at the current moment to a cloud platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210482627.XA CN114818345B (en) | 2022-05-05 | 2022-05-05 | Photovoltaic module residual life prediction method and prediction system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210482627.XA CN114818345B (en) | 2022-05-05 | 2022-05-05 | Photovoltaic module residual life prediction method and prediction system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114818345A CN114818345A (en) | 2022-07-29 |
| CN114818345B true CN114818345B (en) | 2023-09-12 |
Family
ID=82510938
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210482627.XA Active CN114818345B (en) | 2022-05-05 | 2022-05-05 | Photovoltaic module residual life prediction method and prediction system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114818345B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113642895B (en) * | 2021-08-16 | 2023-12-12 | 中国科学院电工研究所 | Residual performance evaluation method of off-grid photovoltaic power station |
| CN115270529A (en) * | 2022-09-28 | 2022-11-01 | 华东交通大学 | Photovoltaic module reliability modeling method integrating environmental factors and random initial degradation |
| CN116756505B (en) * | 2023-06-07 | 2024-05-28 | 上海正泰电源系统有限公司 | Photovoltaic equipment intelligent management system and method based on big data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107480440A (en) * | 2017-08-04 | 2017-12-15 | 山东科技大学 | A kind of method for predicting residual useful life for modeling of being degenerated at random based on two benches |
| WO2018076475A1 (en) * | 2016-10-26 | 2018-05-03 | 广东产品质量监督检验研究院 | Photovoltaic assembly accelerated degradation model established based on deep approach of learning, and method for predicting photovoltaic assembly lifetime |
| CN108304685A (en) * | 2018-04-13 | 2018-07-20 | 中国人民解放军火箭军工程大学 | A kind of non-linear degradation equipment method for predicting residual useful life and system |
| WO2019174142A1 (en) * | 2018-03-14 | 2019-09-19 | 山东科技大学 | Multi-mode degradation process modelling and remaining service life prediction method |
| CN112329272A (en) * | 2020-12-14 | 2021-02-05 | 中国人民解放军国防科技大学 | Residual life fitting prediction method integrating photovoltaic module degradation and life data |
| CN112800616A (en) * | 2021-02-05 | 2021-05-14 | 中国人民解放军空军工程大学 | Equipment residual life self-adaptive prediction method based on proportional acceleration degradation modeling |
| CN113033015A (en) * | 2021-04-09 | 2021-06-25 | 中国人民解放军火箭军工程大学 | Degraded equipment residual life prediction method considering two-stage self-adaptive Wiener process |
-
2022
- 2022-05-05 CN CN202210482627.XA patent/CN114818345B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018076475A1 (en) * | 2016-10-26 | 2018-05-03 | 广东产品质量监督检验研究院 | Photovoltaic assembly accelerated degradation model established based on deep approach of learning, and method for predicting photovoltaic assembly lifetime |
| CN107480440A (en) * | 2017-08-04 | 2017-12-15 | 山东科技大学 | A kind of method for predicting residual useful life for modeling of being degenerated at random based on two benches |
| WO2019174142A1 (en) * | 2018-03-14 | 2019-09-19 | 山东科技大学 | Multi-mode degradation process modelling and remaining service life prediction method |
| CN108304685A (en) * | 2018-04-13 | 2018-07-20 | 中国人民解放军火箭军工程大学 | A kind of non-linear degradation equipment method for predicting residual useful life and system |
| CN112329272A (en) * | 2020-12-14 | 2021-02-05 | 中国人民解放军国防科技大学 | Residual life fitting prediction method integrating photovoltaic module degradation and life data |
| CN112800616A (en) * | 2021-02-05 | 2021-05-14 | 中国人民解放军空军工程大学 | Equipment residual life self-adaptive prediction method based on proportional acceleration degradation modeling |
| CN113033015A (en) * | 2021-04-09 | 2021-06-25 | 中国人民解放军火箭军工程大学 | Degraded equipment residual life prediction method considering two-stage self-adaptive Wiener process |
Non-Patent Citations (1)
| Title |
|---|
| 王泽洲 ; 陈云翔 ; 蔡忠义 ; 罗承昆 ; .考虑非线性退化与随机失效阈值的剩余寿命预测.国防科技大学学报.2020,(第02期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114818345A (en) | 2022-07-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114818345B (en) | Photovoltaic module residual life prediction method and prediction system | |
| Zhou et al. | Infinite time horizon maximum causal entropy inverse reinforcement learning | |
| CN110427654B (en) | Landslide prediction model construction method and system based on sensitive state | |
| CN111027732B (en) | Method and system for generating multi-wind power plant output scene | |
| CN109117566B (en) | Mixed maintenance planning method based on product life prediction model | |
| CN113191484A (en) | Federal learning client intelligent selection method and system based on deep reinforcement learning | |
| CN103310051B (en) | Board information terminal Failure Rate Forecasting Method in a kind of life cycle management | |
| CN112068003B (en) | Method and device for predicting service life of cadmium-nickel storage battery based on linear wiener process | |
| CN113269314B (en) | New energy power generation scene data migration method based on generation countermeasure network | |
| CN110910004A (en) | Reservoir dispatching rule extraction method and system with multiple uncertainties | |
| CN112949026B (en) | Age and state dependence considered degradation equipment residual life prediction method | |
| CN109102155B (en) | Ultra-short-term node marginal electricity price probability prediction method and system | |
| CN109961172B (en) | CPS rare event probability prediction method based on statistical model test | |
| CN112884236B (en) | Short-term load prediction method and system based on VDM decomposition and LSTM improvement | |
| CN110795841B (en) | Mathematical modeling method for uncertainty of intermittent energy output | |
| CN116643499A (en) | Model reinforcement learning-based agent path planning method and system | |
| CN115972211A (en) | Control strategy offline training method based on model uncertainty and behavior prior | |
| CN112232570A (en) | Forward active total electric quantity prediction method and device and readable storage medium | |
| Shahraki et al. | Selective maintenance optimization for multi-state systems operating in dynamic environments | |
| CN116701877A (en) | Fan abnormal value repairing method based on conditional diffusion model | |
| CN114859231B (en) | Battery remaining life prediction method based on wiener process and extreme learning machine | |
| CN112379274A (en) | Method for predicting residual life of power battery | |
| CN109189973B (en) | Large-scale image retrieval method and device based on strategy gradient | |
| CN114399901B (en) | Method and equipment for controlling traffic system | |
| CN115879372A (en) | Equipment residual life prediction method and system based on degradation model dynamic calibration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |