CN109992911B - Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics - Google Patents
Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics Download PDFInfo
- Publication number
- CN109992911B CN109992911B CN201910278837.5A CN201910278837A CN109992911B CN 109992911 B CN109992911 B CN 109992911B CN 201910278837 A CN201910278837 A CN 201910278837A CN 109992911 B CN109992911 B CN 109992911B
- Authority
- CN
- China
- Prior art keywords
- curve
- training
- data
- current
- voltage
- 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
Images
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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Testing Of Individual Semiconductor Devices (AREA)
- Photovoltaic Devices (AREA)
Abstract
The invention relates to a photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics, which comprises the following steps: step S1: collecting IV characteristic curve data of the photovoltaic module under various environmental conditions; s2, acquiring an IV characteristic curve according to the IV characteristic curve data; s3, eliminating abnormal IV curve data to obtain a normal IV curve data set; and step S4: grid extraction is carried out on a normal IV curve data set according to the irradiance of the IV characteristic curve data and the temperature of the assembly back plate, an IV curve data set which uniformly covers various working conditions is obtained, and the IV curve data set is randomly divided into an independent training curve set and an independent testing curve set; step S5 obtains a training data point set and a test data point set from each data point. Step S6: and establishing a photovoltaic module model according to the training data points obtained in the S5. The method can accurately, effectively and quickly model the IV characteristic of the photovoltaic module, has high model training speed, and has good accuracy and generalization capability.
Description
Technical Field
The invention belongs to modeling technologies of a solar cell and a photovoltaic power generation array, and particularly relates to a photovoltaic module rapid modeling method based on an extreme learning machine and IV characteristics.
Background
The accurate modeling of the photovoltaic module and the array plays an important role in optimizing the generating efficiency of the photovoltaic power station. However, the photovoltaic array and the power station are installed and operated in a severe outdoor environment, and are susceptible to various environmental factors such as thermal cycle, humidity, ultraviolet rays, wind excitation and the like during operation, so that faults such as local aging, performance degradation, cracks and the like of the material are caused, the electrical characteristics of the photovoltaic array are greatly influenced, and the power generation efficiency of the photovoltaic power station is further influenced. Therefore, the model of the photovoltaic module/array which is accurate, efficient and reliable is provided, and the model plays an important role in tracking the maximum power point of the photovoltaic array, detecting faults, predicting power and the like. In addition, with the rapid increase of the loading amount of photovoltaic power generation in the world, an accurate, efficient and reliable modeling method for a photovoltaic module/array has been widely researched by scholars at home and abroad.
The existing photovoltaic model modeling methods can be roughly divided into two types, namely a white-box model based on an equivalent circuit and a black-box model based on data driving. The white box model based on the equivalent circuit mainly adopts a single/double diode model, corresponding equivalent circuit equations are found by the models through I-V characteristic curves, when the root mean square error between a fitting curve obtained by finding the equations and an actually measured curve is minimum, the internal parameter value is obtained, and then the obtained optimal parameters are substituted to obtain the accurate photovoltaic model. Although the method can have high accuracy on a standard set, the accuracy of the model of the method greatly depends on the numerical values of parameters in the formula, and the parameters are easily influenced by environmental factors. Therefore, if an accurate model under each working condition is to be obtained, parameters of the I-V characteristic curve under each working condition are extracted, and complexity and low efficiency of photovoltaic modeling are caused. However, if only the model parameters extracted under the characteristic conditions are used as the common model parameters, the model parameters may be inaccurate. In summary, to solve the problems of the existing white-box model, the data-driven black-box model has received much attention, and the common methods include a BP neural network, a multilayer perceptron (MLP), a generalized neural network (GRNN), and the like. Therefore, the photovoltaic module rapid modeling method based on the extreme learning machine and the IV characteristic is provided, and compared with the traditional machine learning algorithm, the method has shorter model training time and higher generalization performance and has higher accuracy.
Disclosure of Invention
In view of the above, the present invention provides a method for quickly modeling a photovoltaic module based on an extreme learning machine and IV characteristics, so as to overcome the defects of the prior art, thereby greatly shortening the model training time and improving the accuracy and generalization capability of the photovoltaic model.
In order to realize the purpose, the invention adopts the following technical scheme:
a photovoltaic module rapid modeling method based on an extreme learning machine and IV characteristics comprises the following steps:
step S1: collecting IV characteristic curve data, irradiance and assembly backboard temperature of the photovoltaic assembly by adopting an IV characteristic curve tester;
s2, carrying out data point interpolation resampling on the IV characteristic curve data of the photovoltaic component to obtain an IV characteristic curve;
s3, based on the single-diode five-parameter model, carrying out curve fitting by adopting an optimization algorithm to obtain fitting root mean square error, judging an abnormal curve according to the fitting root mean square error, and eliminating abnormal IV curve data to obtain a normal IV curve data set;
and step S4: grid extraction is carried out on a normal IV curve data set according to the irradiance of the IV characteristic curve and the temperature of the assembly back plate, an IV curve data set which uniformly covers various working conditions is obtained, and the IV curve data set is randomly divided into an independent training curve set and an independent testing curve set;
step S5: and combining the voltage and the current of the data point of each curve in the training curve set and the test curve set with the irradiance and the temperature to obtain a training data point set and a test data point set which are composed of the voltage, the current, the irradiance and the temperature of each data point.
Step S6: according to the training data point set, with irradiance, assembly backboard temperature and assembly voltage as input and assembly current as output, training a single hidden layer feedforward neural network by adopting an extreme learning machine fitting algorithm, and establishing a photovoltaic assembly model;
step S7: and inputting the test data point set into the extreme learning machine photovoltaic module model obtained by training in the S6, obtaining a current value calculated by the model, and comparing the current value with the actually measured current value to test and evaluate the error of the extreme learning machine photovoltaic module model.
Further, the step S2 specifically includes:
step S21: finding the open-circuit voltage V of the IV characteristic curve data of the photovoltaic module from the IV characteristic curve data oc And short-circuit current I sc ；
Step S22: in [ 0V oc ]Uniformly obtaining the voltages of N resampling points in the voltage interval, wherein the interval voltage of adjacent resampling points is
And recording the resampled voltage vector V 1 ，V 2 ，...，V c ，...，V N ]In which V is c Indicating the voltage at the c-th voltage resampling point.
Step S23: for each resample point V c C value range [1N ]]Finding the voltage and current values of two data points adjacent to the left and right of the resampling point in the original IV characteristic curve data, namely V c-1 And V c+1 And corresponding current values I c - 1 And I c+1 ；
Step S24: calculating each voltage point V by using a linear interpolation method c Corresponding current I c The specific calculation method comprises the following steps:
thereby obtaining N voltage resampling points (V) c ,I c )
Step S25: at [0I sc ]Uniformly obtaining currents of M resampling points in a current interval, wherein the interval between adjacent resampling points is that
And recording the resampled current vector[I 1 ，I 2 ，...，I d ，...，I M ]In which I d Representing the current at the d-th current resampling point.
Step S26: for each current resampling point I d D value range [1M ]]Finding the current value I of two left and right adjacent data points of the resampling point in the original IV characteristic curve data d-1 And I d+1 And corresponding voltage value V d-1 And V d+1 ；
Step S27: calculating the resampling point I of each current by using a linear interpolation method d Corresponding voltage value V d The specific calculation method comprises the following steps:
thereby obtaining M current resampling points (V) d ,I d )。
Step S28: and (4) combining the N voltage resampling points obtained in the step (24) and the M current resampling points obtained in the step (27) in a sequence from small to large according to the voltage to obtain a resampling IV characteristic curve containing M + N data points.
Further, the curve fitting method specifically includes extracting five parameter values of the single-diode photovoltaic model of each curve under each working condition by using a photovoltaic model parameter extraction method based on an adaptive mixed simplex based on an eagle strategy, and calculating a root mean square error RMSE between the curve obtained by fitting under the five parameter condition and an actually measured curve, and the specific calculation method is as follows:
where I is the measured current value of the current,
is the predicted current value, N is the number of sampling points on the whole I-V curve, and RMSE is greater than an acceptable threshold value RMSE T And (4) determining the curve to be an abnormal curve and removing the abnormal curve to obtain normal IV curve data.
Further, step S4 specifically includes:
step S41: setting the extraction range and the sampling interval of grid sampling for the IV characteristic curve data;
step S42: sequentially selecting grids and counting distributed samples in the grids;
step S43: if the number of the samples in the grid is smaller than the maximum sampling point number, randomly selecting 70% of all the samples in the grid as training samples and 30% of all the samples in the grid as test samples, wherein 90% of the training samples are used for actual training and 10% of the training samples are used as a verification set;
step S44: if the number in the grid is larger than the maximum sampling number, selecting the sampling samples with the maximum sampling number from the samples in the grid, and then randomly selecting 70% of all the samples in the grid as training samples, wherein 90% of the training samples are used for actual training, and 10% of the training samples are used as a verification set;
step S45: repeating the steps S42 to S44 until all grids are taken, processing the obtained sample into a data set with input of (G, T, V) and output of I, wherein G and T are irradiance and temperature, and V and I are voltage and corresponding current vectors;
step S46: and returning the obtained test set, the verification set and the training set.
Further, the step S5 specifically includes:
step S51: respectively counting the number of IV curves contained in the training set and the test set obtained in the step S4;
step S52: for each training set IV curve, taking the irradiance, the assembly backboard temperature, each voltage point and each current point as a training data point set;
step S53: and (4) for each test set IV curve, taking the irradiance, the assembly backboard temperature, each voltage point and each current point as a test data point set.
Further, step S6 specifically includes:
step S61: taking the training data point set obtained in the step S5, taking the irradiance of each data point, the temperature of a backboard of the component and the voltage of the component as the input of the training data, and taking the corresponding current as the output of the training data;
step S62: establishing a single hidden layer feedforward neural network model;
step S63: randomly initializing a connection weight w of an input layer and a hidden layer of the neural network and a threshold b of a neuron of the hidden layer, wherein the connection weight between the hidden layer and an output layer is beta, and representing w and beta by using a matrix as follows:
threshold for hidden layer neurons b = [ b ] 1 ，b 2 ，b 3 ，...，b l ] T
Assuming that the hidden layer neuron activation function is g (X), the first output of the input matrix X after passing through the neural network is
It can be seen that when the jth sample is input into the neural network, the predicted value output by the network is ≥ h>
The output matrix of the hidden layer of the neural network has a predicted value of T = [ T ] 1 ，t 2 ，...，t Q ]；
Step S65: calculating weight beta of a hidden layer and an output layer of the single hidden layer feedforward neural network by using an extreme learning machine fitting algorithm, and calculating beta = H by using a least square method + T', thereby obtaining a network model, wherein the parameters are w, b and β.
Further, the step S7 specifically includes: and (5) taking the test data point set obtained in the step (S5), inputting the irradiance, the temperature of the assembly backboard and the voltage value into the extreme learning machine photovoltaic assembly model obtained by training in the step (S6), obtaining a current value predicted by the model, and comparing the current value with an actually measured current value. The error between the actual measured current value and the model predicted current value is used to evaluate the accuracy of the extreme learning machine photovoltaic module model.
Compared with the prior art, the invention has the following beneficial effects:
compared with the traditional machine learning algorithm, the extreme learning machine has shorter model training time, higher accuracy and stronger generalization capability. And a method for uniformly sampling current and voltage at the same time is adopted, and the universality of sampling points is fully considered, so that the model has higher prediction precision. Compared with the existing modeling method, the method greatly shortens the model training time, has better generalization performance and improves the accuracy of model prediction.
Drawings
FIG. 1 is a general flow diagram of the present invention;
FIG. 2 is a flow chart of resampling and exception culling in an embodiment of the invention;
FIG. 3 is a diagram of a single hidden layer feedforward neural network according to an embodiment of the present invention;
FIG. 4 is a flow chart of extreme learning machine-based modeling training in the present invention;
FIG. 5 is a graph of predicted results for a CdTe75669 device in an embodiment of the invention.
Detailed Description
The invention is further explained by the following embodiments in conjunction with the drawings.
Referring to fig. 1, the invention provides a photovoltaic module rapid modeling method based on an extreme learning machine and IV characteristics, a flow chart is shown in fig. 1, and the method specifically includes the following steps:
step S1: an IV characteristic curve tester is adopted to collect IV characteristic curve data, irradiance and assembly backboard temperature of the photovoltaic assembly;
s2, carrying out data point interpolation resampling on the IV characteristic curve data of the photovoltaic component to obtain an IV characteristic curve;
s3, based on the single-diode five-parameter model, performing curve fitting by adopting an optimization algorithm to obtain a fitting root mean square error, judging an abnormal curve according to the fitting root mean square error, and eliminating abnormal IV curve data to obtain a normal IV curve data set;
and step S4: grid extraction is carried out on a normal IV curve data set according to the irradiance of the IV characteristic curve and the temperature of the assembly back plate, an IV curve data set which uniformly covers various working conditions is obtained, and the IV curve data set is randomly divided into an independent training curve set and an independent testing curve set;
step S5: and combining the voltage and the current of the data point of each curve in the training curve set and the test curve set with the irradiance and the temperature to obtain a training data point set and a test data point set which are composed of the voltage, the current, the irradiance and the temperature of each data point.
Step S6: according to the training data point set, with irradiance, assembly backboard temperature and assembly voltage as input and assembly current as output, training a single hidden layer feedforward neural network by adopting an extreme learning machine fitting algorithm, and establishing a photovoltaic assembly model;
step S7: and inputting the test data point set into the extreme learning machine photovoltaic module model obtained by training in the S6, obtaining a current value calculated by the model, and comparing the current value with the actually measured current value to test and evaluate the error of the extreme learning machine photovoltaic module model.
As shown in fig. 2, in order to provide a flow chart of a bilinear sampling method in an embodiment of a photovoltaic module fast modeling method based on an extreme learning machine and IV characteristics, a specific process is divided into two parts, a large amount of actually measured IV characteristic curve data is obtained first, and resampling is performed, and the specific process is as follows:
step S21: finding the open-circuit voltage V of the IV characteristic curve data of the photovoltaic module from the IV characteristic curve data oc And short-circuit current I sc ；
Step S22: in [ 0V oc ]Uniformly obtaining the voltages of N resampling points in the voltage interval, wherein the interval voltage of adjacent resampling points is
And recording the resampled voltage vector V 1 ，V 2 ，...，V c ，...，V N ]In which V is c Indicating the voltage at the c-th voltage resampling point.
Step S23: for each resampling point V c ，cValue range [1N]Finding the voltage and current values of two data points adjacent to the left and right of the resampling point in the original IV characteristic curve data, namely V c-1 And V c+1 And corresponding current values I c-1 And I c+1 ；
Step S24: calculating each voltage point V by using a linear interpolation method c Corresponding current I c The specific calculation method comprises the following steps:
thereby obtaining N voltage resampling points (V) c ,I c )
Step S25: at [0I sc ]Uniformly obtaining currents of M resampling points in a current interval, wherein the interval between adjacent resampling points is
And recording the resampled current vector [ I 1 ，I 2 ，...，I d ，...，I M ]In which I d Representing the current at the d-th current resampling point.
Step S26: resampling point I for each current d D value range [1M ]]Finding the current value I of two left and right adjacent data points of the resampling point in the original IV characteristic curve data d-1 And I d+1 And corresponding voltage value V d-1 And V d+1 ；
Step S27: calculating the resampling point I of each current by using a linear interpolation method d Corresponding voltage value V d The specific calculation method comprises the following steps:
thereby obtaining M current resampling points (V) d ,I d )。
Step S28: and (4) combining the N voltage resampling points obtained in the step (S24) and the M current resampling points obtained in the step (S27) in a sequence from small to large according to the voltage to obtain a resampling IV curve containing M + N data points, and using the resampling IV curve for subsequent photovoltaic module modeling.
In this embodiment, the grid sampling method in the photovoltaic module rapid modeling method based on the extreme learning machine and the IV characteristics includes the following specific processes:
step S41: for the IV characteristic curve data of the component, the decimation range and the sampling interval of the grid sampling are set. The temperature extraction range is [ -10,70]Irradiance extraction range of [80,1000]The sampling interval of the temperature is 1 ℃, and the irradiance extraction range is 50W/m 2
Step S42: sequentially selecting grids and counting distributed samples in the grids;
step S43: if the number of the samples in the grid is smaller than the maximum sampling point number, randomly selecting 70% of all the samples in the grid as training samples and 30% of all the samples in the grid as test samples, wherein 90% of the training samples are used for actual training and 10% of the training samples are used as a verification set;
step S44: if the number in the grid is larger than the maximum sampling number, selecting the sampling samples with the maximum sampling number from the samples in the grid, and then randomly selecting 70% of all the samples in the grid as training samples, wherein 90% of the training samples are used for actual training, and 10% of the training samples are used as a verification set;
step S45: repeating the steps S42 to S44 until all grids are taken, processing the obtained sample into a data set with input of (G, T, V) and output of I, wherein G and T are irradiance and temperature, and V and I are voltage and corresponding current vectors;
step S46: and returning to pass the obtained test set, verification set and training set.
As shown in fig. 3, the structure diagram of the single hidden layer feedforward neural network in the proposed photovoltaic module rapid modeling method based on the extreme learning machine and the IV characteristics is characterized in that the network structure is simple, and has three layers, namely an input layer, a hidden layer and an output layer. Different from other complex neural network structures, the neural network only comprises one hidden layer, the input layer and the hidden layer are all connected, the hidden layer and the output layer are all connected, the connection between each two layers can be endowed with a weight value, and the hidden neuron is also provided with an activation function and a threshold value so as to achieve the purpose of adjusting input data. The working principle of the single hidden layer feedforward neural network is as follows: all parameters are initialized firstly, then the input layer receives input data, the input data reach the output layer through the hidden layer, and the middle of the input data is endowed with two layers of weights, activation functions and thresholds. And the output layer outputs a calculated value of the network, the output layer feeds back a prediction error forwards by comparing the output value with a true value, and finally, the network adjusts the weight and the threshold according to the error fed back by the output layer, namely, the network parameters are updated. Through multiple iterations, the network can achieve higher prediction accuracy.
Fig. 4, modeling photovoltaic modules and predicting IV characteristics by extreme learning machine based method according to claim 1. The modeling process is as follows:
step S61: and (5) taking the training data point set obtained in the step (S5), taking the irradiance of each data point, the temperature of the backboard of the component and the voltage of the component as the input of the training data, and taking the corresponding current as the output of the training data.
Step S62: establishing a single hidden layer feedforward neural network, randomly initializing a connection weight w of an input layer and a hidden layer of the neural network and a threshold b of a neuron of the hidden layer, wherein the connection weight between the hidden layer and an output layer is beta, and expressing w and beta by using a matrix as follows:
threshold for hidden layer neurons b = [ b 1 ，b 2 ，b 3 ，...，b l ] T In this example, the number of hidden layer neurons is 150.
In this embodiment, g (X) = sigmoid (X) is taken for the activation function, and then the first output of the input matrix X after passing through the neural network is the following output
It can be known that when the jth sample is input into the neural network, the predicted value output by the network is &>
The output matrix of the hidden layer of the neural network has a predicted value of T = [ T ] 1 ，t 2 ，...，t Q ]。
Step S64: calculating the weight beta of the hidden layer and the output layer of the single hidden layer feedforward neural network by using an extreme learning machine fitting algorithm, and calculating the weight beta = H by using a least square method + T', thereby obtaining a network model, wherein the parameters are w, b, β.
As shown in table 1, the results of the proposed photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics on the measured photovoltaic module CdTe 75669. Compared with traditional machine learning algorithms such as a BP neural network, a Generalized Regression Neural Network (GRNN), a Support Vector Machine (SVM) and the like, the photovoltaic module rapid modeling method based on the extreme learning machine provided by the embodiment has the advantages that the training speed of a network model is greatly shortened, and the precision and the robustness of the model on a training set, a verification set and a test set are remarkably improved. In addition, the method still has good performance on a test set, which shows that the method has good generalization performance.
Table 1: comparison of respective algorithm performances
As shown in fig. 5, in order to compare the IV characteristic curves modeled by the proposed extreme learning machine-based photovoltaic module rapid modeling method and the generalized regression network GRNN under various working conditions, it can be seen from the graph that the IV characteristic curve predicted by the extreme learning machine-based photovoltaic module rapid modeling method provided by the present invention is closer to the actually measured IV characteristic curve.
The above are preferred embodiments of the present invention, and all changes made according to the technical solutions of the present invention that produce functional effects do not exceed the scope of the technical solutions of the present invention belong to the protection scope of the present invention.
Claims (6)
1. A photovoltaic module rapid modeling method based on an extreme learning machine and IV characteristics is characterized by comprising the following steps:
step S1: an IV characteristic curve tester is adopted to collect IV characteristic curve data, irradiance and assembly backboard temperature of the photovoltaic assembly;
s2, carrying out data point interpolation resampling on the IV characteristic curve data of the photovoltaic component to obtain an IV characteristic curve;
s3, based on the single-diode five-parameter model, performing curve fitting by adopting an optimization algorithm to obtain a fitting root mean square error, judging an abnormal curve according to the fitting root mean square error, and eliminating abnormal IV curve data to obtain a normal IV curve data set;
and step S4: grid extraction is carried out on a normal IV curve data set according to the irradiance of the IV characteristic curve and the temperature of the assembly back plate, an IV curve data set which uniformly covers various working conditions is obtained, and the IV curve data set is randomly divided into an independent training curve set and an independent testing curve set;
step S5: combining the voltage, the current, the irradiance and the temperature of data points of each curve in the training curve set and the test curve set to obtain a training data point set and a test data point set which are composed of the voltage, the current, the irradiance and the temperature of each data point;
step S6: according to the training data point set, with irradiance, assembly backboard temperature and assembly voltage as input and assembly current as output, training a single hidden layer feedforward neural network by adopting an extreme learning machine fitting algorithm, and establishing a photovoltaic assembly model;
the step S6 specifically includes:
step S61: taking the training data point set obtained in the step S5, taking the irradiance of each data point, the temperature of a backboard of the component and the voltage of the component as the input of the training data, and taking the corresponding current as the output of the training data;
step S62: establishing a single hidden layer feedforward neural network model;
step S63: randomly initializing a connection weight w of an input layer and a hidden layer of the neural network and a threshold b of a neuron of the hidden layer, wherein the connection weight between the hidden layer and an output layer is beta, and representing w and beta by using a matrix as follows:
threshold for hidden layer neurons b = [ b 1 ，b 2 ，b 3 ，...，b l ] T ；
Step S64: assuming that the hidden layer neuron activation function is g (X), the first output of the input matrix X after passing through the neural network is
It can be seen that when the jth sample is input into the neural network, the predicted value output by the network is ≥ h>
The predicted value of the network output is A = [ t ] for the hidden layer output matrix of the neural network 1 ，t 2 ，...，t Q ]；
Step S65: calculating a connection weight beta between a hidden layer and an output layer of the single hidden layer feedforward neural network by using an extreme learning machine fitting algorithm, and calculating beta = H by using a least square method + A', thereby obtaining a network model, wherein the parameters are w, b and beta;
step S7: and inputting the test data point set into the extreme learning machine photovoltaic module model obtained by training in S6, obtaining a current value calculated by the model, and comparing the current value with the actually measured current value to test and evaluate the error of the extreme learning machine photovoltaic module model.
2. The extreme learning machine and IV characteristic-based photovoltaic module rapid modeling method according to claim 1, characterized in that: the step S2 specifically comprises the following steps:
step S21: from IV characteristic curve dataFinding the open-circuit voltage V of the IV characteristic curve data of the photovoltaic module oc And short-circuit current I sc ；
Step S22: in [0,V oc ]Uniformly obtaining the voltages of N resampling points in the voltage interval, wherein the interval voltage of adjacent resampling points is
And recording the resampled voltage vector V 1 ，V 2 ，...，V c ，...，V N ]In which V is c Representing the voltage at the c-th voltage resampling point;
step S23: for each resampling point V c And c value range [1,N]Finding the voltage and current values of two data points adjacent to the left and right of the resampling point in the original IV characteristic curve data, namely V c-1 And V c+1 And corresponding current values I c-1 And I c+1 ；
Step S24: calculating each voltage point V by using a linear interpolation method c Corresponding current I c The specific calculation method comprises the following steps:
thereby obtaining N voltage resampling points (V) c ，I c )
Step S25: in [0,I sc ]Uniformly obtaining currents of M resampling points in a current interval, wherein the interval between adjacent resampling points is
And recording the resampled current vector [ I ] 1 ，I 2 ，...，I d ，...，I M ]In which I d Representing the current at the d current resampling point;
step S26: for each current resampling point I d D value range [1,M]Finding the current value I of two left and right adjacent data points of the resampling point in the original IV characteristic curve data d-1 And I d+1 And corresponding voltage value V d-1 And V d+1 ；
Step S27: calculating the resampling point I of each current by using a linear interpolation method d Corresponding voltage value V d The specific calculation method comprises the following steps:
thereby obtaining M current resampling points (V) d ，I d )；
Step S28: and (4) combining the N voltage resampling points obtained in the step (24) and the M current resampling points obtained in the step (27) in a sequence from small to large according to the voltage to obtain a resampling IV characteristic curve containing M + N data points.
3. The extreme learning machine and IV characteristic-based photovoltaic module rapid modeling method according to claim 1, characterized in that: the curve fitting method specifically comprises the steps of extracting five parameter values of a single-diode photovoltaic model of each curve under each working condition by adopting a photovoltaic model parameter extraction method based on an adaptive mixed simplex based on an eagle strategy, and calculating a Root Mean Square Error (RMSE) between the curve obtained by fitting under the condition of the five parameters and an actually measured curve, wherein the specific calculation method comprises the following steps:
where I is the measured current value of the current,
is the predicted current value, N is the number of sampling points on the whole I-V curve, and RMSE is greater than the acceptable threshold value RMSE T And (4) determining the curve to be an abnormal curve and removing the abnormal curve to obtain normal IV curve data.
4. The extreme learning machine and IV characteristic-based photovoltaic module rapid modeling method according to claim 1, characterized in that: the step S4 specifically includes:
step S41: setting the extraction range and the sampling interval of grid sampling for the IV characteristic curve data;
step S42: sequentially selecting grids and counting distributed samples in the grids;
step S43: if the number of the samples in the grid is smaller than the maximum sampling point number, randomly selecting 70% of all the samples in the grid as training samples and 30% of all the samples in the grid as test samples, wherein 90% of the training samples are used for actual training and 10% of the training samples are used as a verification set;
step S44: if the number in the grid is larger than the maximum sampling number, selecting the sampling samples with the maximum sampling number from the samples in the grid, and then randomly selecting 70% of all the samples in the grid as training samples, wherein 90% of the training samples are used for actual training, and 10% of the training samples are used as a verification set;
step S45: repeating the steps S42 to S44 until all grids are taken, processing the obtained sample into a data set with input of (G, te, V) and output of I, wherein G and Te are irradiance and temperature, and V and I are voltage and corresponding current vectors;
step S46: and returning the obtained test set, the verification set and the training set.
5. The extreme learning machine and IV characteristic-based photovoltaic module rapid modeling method according to claim 1, characterized in that: the step S5 specifically comprises the following steps:
step S51: respectively counting the number of IV curves contained in the training set and the test set obtained in the step S4;
step S52: for each training set IV curve, taking the irradiance, the assembly backboard temperature, each voltage point and each current point as a training data point set;
step S53: and taking the irradiance, the assembly backboard temperature, each voltage point and each current point of each test set IV curve as a test data point set.
6. The extreme learning machine and IV characteristic-based photovoltaic module rapid modeling method according to claim 1, characterized in that: the step S6 specifically includes: taking the test data point set obtained in the step S5, inputting the irradiance, the temperature of the assembly backboard and the voltage value into the extreme learning machine photovoltaic assembly model obtained by training in the step S6, obtaining a current value predicted by the model, and comparing the current value with an actually measured current value; the error between the actual measured current value and the model predicted current value is used to evaluate the accuracy of the extreme learning machine photovoltaic module model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910278837.5A CN109992911B (en) | 2019-05-06 | 2019-05-06 | Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910278837.5A CN109992911B (en) | 2019-05-06 | 2019-05-06 | Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109992911A CN109992911A (en) | 2019-07-09 |
| CN109992911B true CN109992911B (en) | 2023-04-07 |
Family
ID=67131021
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910278837.5A Active CN109992911B (en) | 2019-05-06 | 2019-05-06 | Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109992911B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112380765B (en) * | 2020-11-09 | 2022-11-04 | 贵州电网有限责任公司 | Photovoltaic cell parameter identification method based on improved balance optimizer algorithm |
| CN113037213A (en) * | 2021-03-04 | 2021-06-25 | 国网新疆电力有限公司信息通信公司 | Photovoltaic cell model parameter identification method and device based on goblet sea squirt group algorithm |
| CN113779891A (en) * | 2021-09-26 | 2021-12-10 | 中机国能电力工程有限公司 | Photovoltaic module design method based on data deep mining |
| CN114660531B (en) * | 2022-05-24 | 2022-08-26 | 江西西平计量检测有限公司 | Detection method, system and device based on ammeter measurement error compensation |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109409024A (en) * | 2018-12-25 | 2019-03-01 | 福州大学 | Photovoltaic module voltage-current characteristic modeling method based on one-dimensional depth residual error network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160358099A1 (en) * | 2015-06-04 | 2016-12-08 | The Boeing Company | Advanced analytical infrastructure for machine learning |
-
2019
- 2019-05-06 CN CN201910278837.5A patent/CN109992911B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109409024A (en) * | 2018-12-25 | 2019-03-01 | 福州大学 | Photovoltaic module voltage-current characteristic modeling method based on one-dimensional depth residual error network |
Non-Patent Citations (1)
| Title |
|---|
| Intelligent fault diagnosis of photovoltaic arrays based on optimized kernel extreme learning machine and I-V characteristics.;Chen, ZC (Chen, Zhicong) 等;《Applied Energy》;20171231;第204卷;第912-931页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109992911A (en) | 2019-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109873610B (en) | Photovoltaic array fault diagnosis method based on IV characteristic and depth residual error network | |
| CN109992911B (en) | Photovoltaic module rapid modeling method based on extreme learning machine and IV characteristics | |
| Lasheen et al. | Maximum power point tracking using Hill Climbing and ANFIS techniques for PV applications: A review and a novel hybrid approach | |
| CN109842373B (en) | Photovoltaic array fault diagnosis method and device based on space-time distribution characteristics | |
| CN109409024B (en) | Photovoltaic module voltage and current characteristic modeling method based on one-dimensional depth residual error network | |
| Liu et al. | Fault diagnosis approach for photovoltaic array based on the stacked auto-encoder and clustering with IV curves | |
| CN108983749B (en) | Photovoltaic array fault diagnosis method based on K-SVD training sparse dictionary | |
| CN109672406B (en) | Photovoltaic power generation array fault diagnosis and classification method based on sparse representation and SVM | |
| Liu et al. | Intelligent fault diagnosis of photovoltaic array based on variable predictive models and I–V curves | |
| CN111460644A (en) | Photovoltaic array monitoring system with fault diagnosis and power prediction functions | |
| CN114117921A (en) | Intelligent diagnosis method for faults of photovoltaic array | |
| CN115469184A (en) | New energy transmission line fault identification method based on convolutional network | |
| Yoon et al. | Deep learning-based method for the robust and efficient fault diagnosis in the electric power system | |
| Zhu et al. | Photovoltaic failure diagnosis using sequential probabilistic neural network model | |
| Su et al. | Parameter extraction of photovoltaic single-diode model using integrated current–voltage error criterion | |
| Gao | PV array fault detection based on deep neural network | |
| CN114819382B (en) | LSTM-based photovoltaic power prediction method | |
| CN116632823A (en) | Short-term photovoltaic power prediction method based on power conversion model and multi-layer perceptron | |
| CN111177973B (en) | Photovoltaic array online modeling method based on reinforcement learning | |
| CN114710116A (en) | Actual measurement modeling method and system of photovoltaic cell assembly based on fuzzy model | |
| Liu et al. | An improved code-based fault simulation model for PV module | |
| Gong et al. | Fault diagnosis method of photovoltaic array based on random forest algorithm | |
| CN112232598A (en) | Output power combination prediction method and system for progressive photovoltaic power generation system | |
| Liu et al. | Fault diagnosis of photovoltaic array based on XGBoost method | |
| Zare et al. | Fault diagnosis in photovoltaic modules using a straightforward voltage-current characteristics evaluation |
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 |