CN106355336A - Power generation efficiency evaluation method of photovoltaic power station - Google Patents

Abstract

The invention relates to a power generation efficiency evaluation method of a photovoltaic power station. The method comprises steps as follows: factors affecting the power generation efficiency of the power station are classified into multiple evaluation indexes, relevant parameters affecting each evaluation index is analyzed, and feature vectors of each evaluation index are constructed; a prediction model of each evaluation index is established according to the feature vectors of each evaluation index; a test sample or a real-time sample of the feature vectors of each evaluation index is input into a corresponding prediction model, and a prediction value of each evaluation index is obtained; weight of the prediction value of each evaluation index is determined with an entropy evaluation method, the comprehensive evaluation value of the power generation efficiency of the photovoltaic power station is further obtained and compared with an evaluation index set, and a comprehensive evaluation result is obtained. The evaluation method is comprehensive, objective and accurate, provides a theoretical basis for operation management and optimized design of the photovoltaic power station and has broad engineering application value.

Description

Photovoltaic power station power generation efficiency evaluation method

Technical Field

The invention relates to a method for evaluating the generating efficiency of a photovoltaic power station, and belongs to the technical field of new energy control.

Background

In recent years, with the issue of national approval rights and the emergence of various policies, photovoltaic power generation has come up with a new rapid development period. Along with the change of the construction subsidy mode of the photovoltaic power station, namely, the original power station construction subsidy is changed into the electricity subsidy, a user pays more attention to the generated energy of the photovoltaic power station, the generated energy mainly depends on local solar energy resources and photovoltaic power generation efficiency, the power generation efficiency is a main factor directly influencing the manufacturing cost and the income of a system, and the generated energy and the power generation efficiency serve as the most important indexes for evaluating the economic and social benefits of the photovoltaic power station and are also the reference standard for later-stage operation and maintenance. Therefore, it is necessary to provide an intelligent and efficient power generation efficiency evaluation method.

At present, the research work of evaluating the generating efficiency of the photovoltaic power station in China is still in the starting stage. Most of research objects are efficiency evaluation of single equipment such as a photovoltaic array or an inverter, and the fact that the whole photovoltaic power station is a large system formed by a plurality of photovoltaic arrays, inverters, combiner boxes and other units together is not considered. Chinese patent No. CN103605891A discloses an evaluation method of comprehensive efficiency of outdoor operation of a grid-connected photovoltaic inverter, wherein an inverter efficiency weighting coefficient is determined according to irradiation resources at different power levels to obtain comprehensive efficiency performance of the inverter; chinese patent No. CN 104408562 a discloses a comprehensive evaluation method for photovoltaic system power generation efficiency based on a BP neural network, where training of the BP network requires a large amount of sample data, and the convergence speed is slow and is liable to fall into a local extreme point.

Disclosure of Invention

The invention aims to provide a photovoltaic power station power generation efficiency evaluation method, which is used for solving the problem that an evaluation result is inaccurate in the existing photovoltaic power generation efficiency evaluation process.

In order to solve the technical problem, the invention provides a method for evaluating the power generation efficiency of a photovoltaic power station, which comprises the following steps:

step 1, classifying factors influencing the power generation efficiency of a power station to form a plurality of evaluation indexes, analyzing relevant parameters influencing each evaluation index, and forming a feature vector of each evaluation index;

step 2, constructing a prediction model of each evaluation index according to the feature vector of each evaluation index;

step 3, inputting the test sample or the real-time sample of the feature vector of each evaluation index into a corresponding prediction model to obtain a prediction value of each evaluation index;

and 4, determining the weight of the predicted value of each evaluation index by adopting an entropy method, further obtaining a comprehensive evaluation value of the power generation efficiency of the photovoltaic power station, and comparing the comprehensive evaluation value with the evaluation index set to obtain a comprehensive evaluation result.

Further, the evaluation index in step 1 includes: PV module efficiency, square matrix losses, inverter losses, dc line losses, ac line losses, and transformer losses.

Further, the relevant parameters of each evaluation index are respectively:

PV module efficiency: direct irradiance, diffuse irradiance, total irradiance, temperature, humidity, wind speed, wind direction, voltage temperature coefficient, current temperature coefficient, maximum output power temperature coefficient, component decay rate, and component dispersion;

and (3) square matrix loss: PV assembly efficiency, series-parallel mismatch loss, array inclination angle, azimuth angle, spacing loss, shading loss, temperature rise loss, hot spot power attenuation, subfissure power attenuation and MPPT deviation loss;

inverter losses: a DC input voltage range, a DC input voltage and current, an AC input voltage and current;

direct current line loss: equivalent resistance, resistivity, cable length, cross-sectional area and direct current of the direct current cable;

alternating current line loss: equivalent resistance, cable length, cross-sectional area, AC output line voltage and AC output power of the AC cable;

transformer loss: no-load loss, short-circuit loss, actual input power, rated capacity, high-side rated voltage, and load factor.

Further, the prediction model of each evaluation index in step 2 is constructed based on a least squares support vector machine.

Further, the method for constructing the prediction model of each evaluation index in step 2 includes the following steps:

step 2-1, using a plurality of groups of feature vectors of each evaluation index in the step 2 as training samples, mapping the training samples to a high-dimensional feature space, and constructing an optimal linear decision function and a corresponding target function of the training samples in the high-dimensional feature space;

and 2-2, solving the objective function, determining a coefficient vector and a deviation term corresponding to the optimal linear decision function, and further determining an expression of the optimal linear decision function.

Further, the step 4 of determining the weight of the predicted value of each evaluation index by using an entropy method to obtain a comprehensive evaluation value of the power generation efficiency of the photovoltaic power station includes the following steps:

step 4-1, establishing an evaluation matrix according to the predicted value of each evaluation index;

4-2, performing index syntropy treatment on the evaluation matrix, converting negative indexes in the evaluation matrix into positive indexes, and obtaining an index syntropy matrix;

4-3, carrying out normalization processing on the index syntropy matrix to obtain a standard evaluation matrix;

4-4, calculating an information entropy value corresponding to the predicted value of each evaluation index according to the standard evaluation matrix;

4-5, calculating a weight vector of a predicted value of each evaluation index according to the information entropy of each evaluation index;

and 4-6, carrying out weighted summation on the weight vector of each evaluation index predicted value, and calculating the comprehensive evaluation value of the power generation efficiency of the photovoltaic power station.

The invention has the beneficial effects that:

from the perspective of the whole photovoltaic power station, all factors influencing photovoltaic power generation efficiency are comprehensively analyzed, a prediction model of each evaluation index is constructed, all evaluation index values are predicted, the weight of each evaluation index is determined by an entropy method, the accuracy of an evaluation result is improved, the deviation of artificial subjective factors to the evaluation result is avoided, a theoretical basis is provided for operation management and optimization design of the photovoltaic power station, and the method has wide engineering application value.

In addition, the prediction model of each evaluation index is constructed by adopting the least square support vector, the problems of small samples, nonlinearity and high dimension can be effectively solved, the quadratic programming problem can be converted into a linear equation set, and the problem solving speed and convergence precision are improved.

Drawings

Fig. 1 is a flowchart of a photovoltaic power plant power generation efficiency evaluation method.

Detailed Description

The following describes embodiments of the present invention in further detail with reference to the accompanying drawings.

As shown in fig. 1, the method for evaluating the power generation efficiency of the photovoltaic power station specifically includes the following steps:

1. the method comprises the steps of classifying factors influencing the power generation efficiency of the power station to form a plurality of evaluation indexes, analyzing relevant parameters influencing each evaluation index, and forming a feature vector of each evaluation index.

Wherein, each evaluation index is respectively: PV module efficiency, square matrix losses, inverter losses, dc line losses, ac line losses, transformer losses;

the relevant parameters of each evaluation index and the corresponding formed feature vectors are respectively as follows:

the relevant parameters of PV module efficiency are: direct irradiance, scattered irradiance, total irradiance, temperature, humidity, wind speed, wind direction, voltage temperature coefficient, current temperature coefficient, maximum output power temperature coefficient, component decay rate and component dispersion, and the corresponding eigenvector is recorded as (SV)_ikSS_ikSZ_ikT_ikH_ikFS_ikFX_ikVT_ikIT_ikPT_ikSJ_ikLS_ik) Where i represents the ith PV module.

The relevant parameters of the square matrix loss are respectively as follows: PV component efficiency, series-parallel mismatch loss, array inclination angle, azimuth angle, spacing loss, shielding loss, temperature rise loss, hot spot power attenuation, hidden crack power attenuation and MPPT deviation loss, and the corresponding characteristic vector is recorded as (PV)_kSP_kQJ_kFW_kJS_kZS_kWS_kBS_kLS_kMPS_k)。

The relevant parameters of the inverter loss are respectively: maximum value of DC input voltage, minimum value of DC input voltage, DC input voltage and current, AC input voltage and current, and corresponding eigenvector (UM)_ikUN_ikUZ_ikIZ_ikUJ_ikIJ_ik) Where i represents the ith inverter.

The related parameters of the direct current line loss are respectively as follows: equivalent resistance, resistivity, cable length, cross-sectional area and direct current of the direct current cable, and the corresponding eigenvector is recorded as (ZR)_kZRO_kZL_kZS_kZI_k)。

The relevant parameters of the AC line loss are respectively as follows: equivalent resistance, cable length, cross-sectional area, AC output line voltage and AC output power of the AC cable, and the corresponding eigenvector is expressed as (JR)_kJL_kJS_kJU_kJP_k)。

The relevant parameters of the transformer loss are respectively: no-load loss, short-circuit loss, actual input power, rated capacity, high-voltage-side rated voltage and load factor, and the corresponding eigenvector is recorded as (P0)_ikDL_ikP_ikS_ikUH_ikFX_ik) Where i represents the ith transformer.

In the above feature vector, k is a sample collection number, k is an integer from 1 to N, and each number represents a working state, where N represents the total number of working states. In addition, the evaluation indexes affecting the power generation efficiency of the photovoltaic power plant and the parameters within each evaluation index are not limited to the kinds and numbers specifically listed above.

2. And constructing a prediction model of each evaluation index according to the feature vector of the evaluation index.

In the embodiment, a prediction model of an evaluation index is constructed by using a least squares support vector machine (LS-SVM), and each evaluation index value is predicted.

The construction method of the prediction model of each evaluation index based on the least square support vector machine (LS-SVM) comprises the following steps:

(1) selecting n groups of characteristic vectors of the mth evaluation index as training samples x_mAnd by a functionWill train sample x_mMapping to high-dimensional feature space, and constructing a training sample x_mOptimal linear decision function in the high-dimensional feature space

Wherein, w_mAnd b_mThe coefficient vector and the deviation term of the mth LS-SVM, respectively, can be determined by solving an objective function of the LS-SVM, wherein the objective function is as follows:

$\min Q(w_m,b_m)=\frac{1}{2}\left|\right|w_m|{|}^2+\frac{r_m}{2}\underset{i=1}{\overset{n}{\Σ}}{\left(e_m^i\right)}^2$

the constraint conditions are as follows:

wherein x is_iAnd y_iRespectively an ith feature vector training sample and corresponding output, n is the number of feature vector training samples, r_mIs a penalty coefficient, and r_m＞0，For x in the mth LS-SVM_iAn error variable of

(2) Defining lagrange functions

Wherein,is lagrange multiplier, i is 1,2, …, n.

Taking kernel function

Let L be opposite to w_m、b_m、e_mAnd a_mThe partial derivatives of (A) are all zero, and w is eliminated according to the optimization condition of KKT (Karush-Kuhn-Tucker)_mAnd e_mTwo parameters to obtain a linear equation, and solving the linear equation to obtain a_mAnd b_mThat is, the coefficient vector of LS-SVM can be derived as：

The linear decision function is:

$y\left(x\right)=sign\[\underset{i=1}{\overset{n}{\Σ}}{a}_m^{i}K(x,x_i)+b_m\]$

a prediction model is constructed by adopting a least square support vector machine, the problems of small samples, nonlinearity and high dimension can be effectively solved, the quadratic programming problem can be converted into a linear equation set, and the problem solving speed and convergence precision are improved.

Of course, the method for constructing the prediction model is not limited to the least square support vector mechanism, and other artificial intelligence methods which can achieve the same modeling effect can be adopted.

3. And inputting the test sample or the real-time sample of each evaluation index feature vector into the corresponding prediction model, and respectively calculating the function value of the corresponding decision function so as to obtain the evaluation index value corresponding to the test sample or the real-time sample.

4. And determining the weight of each predicted evaluation index value by adopting an entropy method to obtain a comprehensive evaluation value of the power generation efficiency of the photovoltaic power stations, and comparing the comprehensive evaluation value of each photovoltaic power station with the evaluation index set to further obtain a comprehensive evaluation result of each photovoltaic power station.

The specific steps of determining the weight of each prediction evaluation index value by adopting an entropy method so as to obtain the comprehensive evaluation value of the power generation efficiency of the photovoltaic power station are as follows:

1) in this embodiment, it is assumed that the power generation efficiency of 5 photovoltaic power stations needs to be evaluated, the predicted values corresponding to the 6 evaluation indexes obtained in step 3 are selected, and the evaluation matrix a is constructed as (a)_ij)_5×6＝(A₁A₂A₃A₄A₅A₆)，A₁For PV Module efficiency, A₂For square matrix losses, A₃For inverter losses, A₄For DC line loss, A₅For ac line loss, A₆Is a transformer loss.

2) Index syntropy processing is carried out on the evaluation matrix A, negative indexes in the 6 evaluation index predicted values are converted into positive indexes, and an index syntropy matrix B ═ (B) is obtained_ij)_5×6The conversion formula is:

$\begin{array}{ccc}{b}_{ij}=\frac{1}{0.1+max\left|A_j\right|+a_{ij}}& i=1,2,3,4,5& j=1,2,3,4,5,6\end{array}$

wherein, a_ijMatrix elements representing the evaluation matrix A, b_ijMatrix element, max | A, representing index homologation matrix B_jI represents the index vector A_jThe maximum value of the medium element.

Of course, other suitable conversion formulas in the prior art may also be used to perform index homography processing on the evaluation matrix a.

3) Normalizing the index homologation matrix B to obtain a standard evaluation matrix C ═ C_ij)_5×6Wherein

$c_{ij}=\frac{b_{ij}}{\underset{i=1}{\overset{5}{\Σ}}{b}_{ij}}$

4) And calculating the information entropy value of the predicted value of each evaluation index according to the standard evaluation matrix C, wherein the calculation formula is as follows:

$H_j=-\frac{1}{ln\mathrm{\α}}\underset{j=1}{\overset{6}{\Σ}}{c}_{ij}\·\ln c_{ij}$

where i is 1, …,5, j is 1,2, …,6, α is equal to the number of columns of the standard evaluation matrix C, i.e., α is 6.

5) Calculating a weight vector w (w) corresponding to the predicted value of each evaluation index based on the information entropy of the predicted value of the evaluation index₁，w₂，…，w₆) The calculation formula is as follows:

$\begin{array}{cc}{\mathrm{\ω}}_j=\frac{1-H_j}{\underset{k=1}{\overset{6}{\Σ}}(1-H_k)}& j=1,2,...,6\end{array}$

6) calculating the comprehensive evaluation value of the power generation efficiency of each photovoltaic power station according to the weight of each evaluation index predicted value, wherein the calculation formula is as follows:

$d_i=\underset{j=1}{\overset{6}{\Σ}}{c}_{ij}{\mathrm{\ω}}_j^T$

wherein d is_iThe overall evaluation value of the power generation efficiency of the ith photovoltaic power plant is shown, i is 1,2, …, 5.

The weight of each evaluation index is determined by adopting an entropy method, so that the deviation of evaluation results caused by artificial subjective factors can be effectively avoided, and the power generation efficiency of the photovoltaic power station can be comprehensively and objectively evaluated.

In addition, table 1 specifically shows a certain evaluation index set of the photovoltaic power station.

TABLE 1 evaluation criteria set

The photovoltaic power station power generation efficiency evaluation method adopted in the embodiment performs comprehensive evaluation on the power generation efficiency of 5 photovoltaic power stations, obtains a plurality of evaluation indexes by comprehensively analyzing a plurality of influence factors of the power generation efficiency of the photovoltaic power stations, constructs a prediction model based on a least square support vector machine, predicts each evaluation index, finally obtains a comprehensive evaluation result of the power generation efficiency of each power station, guides operation and maintenance decision and optimization design of the photovoltaic power stations according to the evaluation result, has potential risks, and ensures safe and stable operation of the power stations.

Of course, the method for evaluating the power generation efficiency of the photovoltaic power station can also comprehensively evaluate the power generation efficiency of only one photovoltaic power station or more power stations.

The above embodiments are only used to help understanding the core idea of the present invention, and the present invention is not limited thereby, and any modifications or equivalent substitutions made on the present invention according to the idea of the present invention and the modifications made on the specific embodiments and the application scope should be included in the protection scope of the present invention for those skilled in the art.

Claims (6)

1. A method for evaluating the power generation efficiency of a photovoltaic power station is characterized by comprising the following steps:

step 1, classifying factors influencing the power generation efficiency of a power station to form a plurality of evaluation indexes, analyzing relevant parameters influencing each evaluation index, and forming a feature vector of each evaluation index;

step 2, constructing a prediction model of each evaluation index according to the feature vector of each evaluation index;

step 3, inputting the test sample or the real-time sample of the feature vector of each evaluation index into a corresponding prediction model to obtain a prediction value of each evaluation index;

and 4, determining the weight of the predicted value of each evaluation index by adopting an entropy method, further obtaining a comprehensive evaluation value of the power generation efficiency of the photovoltaic power station, and comparing the comprehensive evaluation value with the evaluation index set to obtain a comprehensive evaluation result.

2. The method according to claim 1, wherein the evaluation index in step 1 comprises: PV module efficiency, square matrix losses, inverter losses, dc line losses, ac line losses, and transformer losses.

3. The method for evaluating the power generation efficiency of the photovoltaic power station as claimed in claim 2, wherein the relevant parameters of each evaluation index are respectively:

PV module efficiency: direct irradiance, diffuse irradiance, total irradiance, temperature, humidity, wind speed, wind direction, voltage temperature coefficient, current temperature coefficient, maximum output power temperature coefficient, component decay rate, and component dispersion;

and (3) square matrix loss: PV assembly efficiency, series-parallel mismatch loss, array inclination angle, azimuth angle, spacing loss, shading loss, temperature rise loss, hot spot power attenuation, subfissure power attenuation and MPPT deviation loss;

inverter losses: a DC input voltage range, a DC input voltage and current, an AC input voltage and current;

direct current line loss: equivalent resistance, resistivity, cable length, cross-sectional area and direct current of the direct current cable;

alternating current line loss: equivalent resistance, cable length, cross-sectional area, AC output line voltage and AC output power of the AC cable;

transformer loss: no-load loss, short-circuit loss, actual input power, rated capacity, high-side rated voltage, and load factor.

4. The method according to any one of claims 1 to 3, characterized in that the prediction model for each evaluation index in step 2 is constructed based on a least squares support vector machine.

5. The method for evaluating the power generation efficiency of the photovoltaic power plant according to claim 4, wherein the method for constructing the prediction model of each evaluation index in the step 2 comprises the following steps:

step 2-1, using a plurality of groups of feature vectors of each evaluation index in the step 2 as training samples, mapping the training samples to a high-dimensional feature space, and constructing an optimal linear decision function and a corresponding target function of the training samples in the high-dimensional feature space;

and 2-2, solving the objective function, determining a coefficient vector and a deviation term corresponding to the optimal linear decision function, and further determining an expression of the optimal linear decision function.

6. The method for evaluating the power generation efficiency of the photovoltaic power station as claimed in claim 1, wherein the step 4 of determining the weight of the predicted value of each evaluation index by using an entropy method so as to obtain the comprehensive evaluation value of the power generation efficiency of the photovoltaic power station comprises the following steps:

step 4-1, establishing an evaluation matrix according to the predicted value of each evaluation index;

4-2, performing index syntropy treatment on the evaluation matrix, converting negative indexes in the evaluation matrix into positive indexes, and obtaining an index syntropy matrix;

4-3, carrying out normalization processing on the index syntropy matrix to obtain a standard evaluation matrix;

4-4, calculating an information entropy value corresponding to the predicted value of each evaluation index according to the standard evaluation matrix;

4-5, calculating a weight vector of a predicted value of each evaluation index according to the information entropy of each evaluation index;

and 4-6, carrying out weighted summation on the weight vector of each evaluation index predicted value, and calculating the comprehensive evaluation value of the power generation efficiency of the photovoltaic power station.