CN118134290A - Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS - Google Patents
Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS Download PDFInfo
- Publication number
- CN118134290A CN118134290A CN202410553196.0A CN202410553196A CN118134290A CN 118134290 A CN118134290 A CN 118134290A CN 202410553196 A CN202410553196 A CN 202410553196A CN 118134290 A CN118134290 A CN 118134290A
- Authority
- CN
- China
- Prior art keywords
- photovoltaic array
- anfis
- index
- improved
- membership
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000011156 evaluation Methods 0.000 title claims abstract description 44
- 238000003745 diagnosis Methods 0.000 title claims abstract description 21
- 230000035772 mutation Effects 0.000 claims abstract description 31
- 238000010248 power generation Methods 0.000 claims abstract description 23
- 230000036541 health Effects 0.000 claims abstract description 17
- 239000011159 matrix material Substances 0.000 claims abstract description 16
- 238000011002 quantification Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 68
- 230000007613 environmental effect Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 14
- 230000032683 aging Effects 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000010206 sensitivity analysis Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 230000002441 reversible effect Effects 0.000 claims description 5
- 241001494479 Pecora Species 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 238000013178 mathematical model Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 230000001932 seasonal effect Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 description 8
- 238000012360 testing method Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000001537 neural effect Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- General Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Testing Of Individual Semiconductor Devices (AREA)
Abstract
The invention provides a photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS, and belongs to the technical field of photovoltaic power generation; the intelligent operation and safety problem of the existing photovoltaic station is solved; the method comprises the following steps: analyzing and determining factors influencing the power generation of the photovoltaic array, and constructing an operation state evaluation index; building an improved ANFIS model: the membership function of the self-adaptive fuzzy inference system is improved by improving the ANFIS model, a fuzzy set is constructed by combining a mutation theory, and a super matrix is fused to realize multidimensional nonlinear output; comprehensively weighting the membership of the evaluation index based on an entropy weight method, further considering external environment factors and self electrical factors to establish corresponding health indexes, realizing quantification of the operation state of the photovoltaic array, and evaluating the operation state of the photovoltaic array and diagnosing faults according to the health indexes; the invention is applied to a photovoltaic array.
Description
Technical Field
The invention provides a photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS, and belongs to the technical field of photovoltaic power generation.
Background
Along with the promotion of a novel electric power system, photovoltaic power generation is rapidly developed in domestic markets due to the advantages of low price, flexible deployment, green environmental protection and the like. However, the photovoltaic power generation system has the defects that the photovoltaic power generation system is easily influenced by external environment and is difficult to diagnose internal electrical faults and the like, if the operation state of the photovoltaic array can be timely, comprehensively and accurately estimated, abnormal and risk conditions can be timely found, the power generation efficiency of the photovoltaic system can be improved, and the safety operation level of the photovoltaic grid connection and the power system can be improved.
At present, the data analysis and state evaluation research of a photovoltaic power generation system mainly synthesizes various factors such as irradiance, weather and the like to establish a state evaluation model, and generally comprises a direct method and an indirect method. The direct prediction is based on a statistical method, meteorological data and historical power data of a photovoltaic power station are input into an established prediction model to obtain output power, and common methods include a multiple linear regression method, a BP neural network, a support vector machine, a gray theory and the like. The indirect rule is that the radiation intensity of the sun is predicted firstly, then the predicted value of the output power is calculated by utilizing a photoelectric conversion efficiency formula, and the model precision is continuously and iteratively improved by comparing the predicted value with the actual value, and the common methods include Kalman filtering, random time sequence, wavelet analysis, sky image and the like.
Meanwhile, due to the deep development of artificial intelligence and computer technology, the fault diagnosis research of the power generation system fused with the novel intelligent algorithm is rapid. However, researches at the present stage are focused on predicting the photovoltaic power generation output of the existing station or primarily judging whether faults exist, but the photovoltaic power generation output monitoring system does not have the functions of adaptive tracking monitoring and quantitative evaluation of the running state, lacks risk prompts of common fault types such as internal short circuit, disconnection and aging of the photovoltaic array, and is difficult to meet the intelligent operation and safety requirements of the photovoltaic station.
Disclosure of Invention
The invention provides a photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS (advanced network analysis and analysis system) for solving the problem that the existing photovoltaic station is poor in intelligent operation and safety.
In order to solve the technical problems, the invention adopts the following technical scheme: a photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS comprises the following steps:
s1: analyzing and determining external environmental factors and self-electrical factors which influence the power generation of the photovoltaic array, and constructing corresponding photovoltaic array running state evaluation indexes according to the analyzed factors;
S2: building an improved ANFIS model to evaluate the operation state grade of the photovoltaic array: the improved ANFIS model improves membership functions of the self-adaptive fuzzy inference system, establishes fuzzy sets by combining with a mutation theory, and fuses a supermatrix to realize multidimensional nonlinear output, wherein the membership functions adopt a segmented dynamic coding method, and iterative optimization is carried out on membership function parameters by using a heuristic algorithm with the aim of minimizing fuzzy evaluation errors;
S3: and comprehensively weighting the membership of the evaluation index based on an entropy weight method, further taking external environmental factors and self electrical factors into consideration to establish corresponding health indexes, realizing quantification of the operation state of the photovoltaic array, and evaluating the operation state of the photovoltaic array and diagnosing faults according to the health indexes.
S11: according to the characteristics of the solar cell, analyzing a mathematical model of the output power of the photovoltaic array;
s12: the voltage, current and maximum power point change conditions under five running states of open circuit, short circuit, shading and aging are analyzed by combining an I-V characteristic curve of the photovoltaic array;
S13: and constructing two types of photovoltaic array operation state evaluation indexes comprising external environment factors and self electrical factors.
The external environment indexes comprise an irradiation ratio R S, a temperature ratio R T and a humidity ratio R H, and the calculation formulas are as follows:
；
；
；
In the above formula, S is the actual horizontal irradiance, S 0 is the standard light intensity, T is the actual environment temperature, T ref is the standard temperature, H is the actual environment humidity, H s is the average value of the seasonal relative humidity, the irradiation ratio is the forward index, and the temperature ratio and the humidity ratio are the reverse index;
The self electrical indexes comprise a voltage ratio R U, a current ratio R I and a power ratio R P, and the calculation formulas are as follows:
；
；
；
in the above formula: u m-array is the maximum power point voltage, I m-STC is the maximum power point current, P m-array is the maximum power, U is the actual output voltage, I is the actual output current, and P is the actual output power.
The step S2 specifically comprises the following steps:
S21: analyzing an original ANFIS model mechanism;
s22: the ANFIS model is improved through membership function dynamic coding, fuzzy reasoning set mutation and a super-matrix multidimensional output structure, and is used for evaluating the running state level;
s23: and iteratively correcting errors through a heuristic optimization algorithm.
The membership function considers the influences of five states of open circuit, short circuit, shading, aging and normal operation on the maximum power point of the photovoltaic array, obtains a fuzzy set of indexes, marks the fuzzy set, determines parameters of different membership functions, and encodes the membership functions in a segmented mode according to the parameters.
The fuzzy set is constructed by combining a mutation theory, namely, after two types of evaluation indexes are respectively normalized, membership function calculation is carried out to obtain fuzzy set evaluation matrixes of five running states, then one potential function type is selected, and the fuzzy set evaluation matrixes after nonlinear processing are obtained according to the selected potential function.
The step S3 specifically comprises the following steps:
s31: preparing a training sample, acquiring environmental index data through historical meteorological data, and constructing electrical index data on an I-V characteristic curve in a mode of classifying evenly-spaced sampling points;
s32: preprocessing data distortion, data noise and data missing noise pollution existing in sample data;
s33: comprehensive health index and sensitivity analysis, carrying out objective weighting by adopting an entropy weighting method according to fuzzy set judgment matrixes with indexes corresponding to different state grades, and further differentiating the index weights of different state grades;
wherein the membership degree normalization value is as follows:
；
In the above formula: p ij is a membership standard value, m is an index number, mu ij is a membership function, i is an index ordinal number, and j is an operation state level ordinal number;
The information entropy is as follows:
；
In the above formula: e i is the information entropy of the running state, and n is the number of running state grades;
the weights are as follows:
；
In the above formula: w i is the weight of the index corresponding to different running states;
According to the maximum membership principle, the index comprehensive membership is expressed as:
；
the environmental index and the electric index influence are considered, and the comprehensive health index of the running state is constructed as follows:
；
In the above-mentioned method, the step of, To adjust the coefficient,/>∈(0,1)，/>Is the electrical index of the self.
On the basis of improving the ANFIS model, the root mean square error is selected as a sensitivity analysis index of a training result, and the calculation formula is as follows:
；
In the above formula: n, t is the number of samples in the verification set, r i is the level of failure diagnosed by the model, Is the actual failure level.
The judging steps of the fault level are as follows:
Taking points on the I-V curve at equal intervals by adopting Linspace functions, and matching corresponding environmental parameters to obtain a plurality of groups of sample indexes;
comprehensively considering the break point mutation of the triangular function and the smooth transition characteristic of the Gaussian function, respectively selecting two membership functions for fuzzy processing in an input part, and carrying out error correction by adopting iteration of an artificial sheep algorithm;
And (3) performing nonlinear processing on the fuzzy set judgment matrix by using dovetail mutation, forming a hyper-matrix fuzzy rule with a more complex structure on a third layer of the model, outputting defuzzified values of open circuit, short circuit, shading, aging and normal 5 running states after iterative training, and judging the fault level according to a maximum value corresponding principle.
The method for preprocessing the sample data in step S32 is as follows:
1) Distortion data processing: screening and correcting distortion data by a transverse comparison method;
2) Missing data processing: interpolation is carried out by adopting a smooth correction method;
3) Data normalization: and according to the evaluation index property, adopting a Min-Max normalization method.
Compared with the prior art, the invention has the following beneficial effects: based on the I-V characteristic curves of the photovoltaic array in different running states, the invention compares the differences of several typical fault types and constructs running state evaluation indexes comprising external environment factors and self-electric factors; the ANFIS is improved, a membership dynamic coding method and a fuzzy set construction method based on a mutation theory are provided, and the fault tolerance of the model and the accuracy of fault diagnosis results are improved; the entropy weighting method is used for objectively giving weight, so that the comprehensive health index of the running state is constructed, the quantification of the running state of the photovoltaic array is realized, the interference of human factors can be avoided, and effective support is provided for daily operation and maintenance.
Drawings
The invention is further described below with reference to the accompanying drawings:
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is an electrical block diagram of a photovoltaic power generation system;
FIG. 3 is a diagram of a photovoltaic cell single diode model;
FIG. 4 is a graph of I-V characteristics of a photovoltaic array under different operating conditions;
FIG. 5 is a graph of photovoltaic power generation power versus relative humidity on a day;
FIG. 6 is a graph of photovoltaic power generation power variation for different seasons;
FIG. 7 is a graph of photovoltaic array operating state evaluation indicators;
FIG. 8 is a block diagram of an ANFIS;
FIG. 9 is a graph of a common membership function;
FIG. 10 is a graph of dynamic encoding of triangle membership functions;
FIG. 11 is a diagram of an ANFIS multi-dimensional output architecture incorporating a super matrix;
FIG. 12 is a graph of a nuclear density estimate of an evaluation index;
FIG. 13 is a run state tab diagram of a sample dataset;
FIG. 14 is a graph of a dynamic encoding parameter distribution of membership functions;
FIG. 15 is a comparison chart of fault diagnosis results for the verification set;
FIG. 16 is a graph of root mean square error sensitivity analysis results;
Fig. 17 is a statistical diagram of the comprehensive health index of the operation state of the verification set.
Detailed Description
As shown in fig. 1 to 17, the present invention provides a photovoltaic array operation state evaluation and fault diagnosis method based on improved ANFIS, comprising the following steps:
s1: analyzing external environmental factors influencing the power generation of the photovoltaic array and electrical factors thereof, wherein the electrical factors refer to I-V change characteristic curves of the photovoltaic array in different running states, and constructing corresponding running state evaluation indexes of the photovoltaic array according to the analyzed factors, wherein the running state evaluation indexes are as follows:
S11: according to the characteristics of the solar cell, analyzing a mathematical model of the output power of the photovoltaic array;
the photovoltaic array is a power generation unit formed by connecting a plurality of photovoltaic modules in series or in parallel, and is commonly used in a photovoltaic power station or a distributed power generation system as shown in fig. 2.
Depending on the solar cell characteristics, the output power of the photovoltaic array is mainly related to the ambient temperature and the irradiation intensity, and the short-circuit current I S and the open-circuit voltage U O thereof can be expressed as:
；
；
In the above formula: i S-STC is short-circuit current under standard test conditions, S is actual horizontal irradiance, S 0 is standard light intensity, T is actual environment temperature, T ref is standard temperature, and a is a temperature coefficient of the short-circuit current; u O-STC is the open circuit voltage under standard test conditions, b is the open circuit voltage temperature coefficient, k is the boltzmann constant, q is the electron charge, and T 0 is the temperature of the solar cell.
If the standard current of the photovoltaic array at the maximum power point is I m-STC and the voltage is U m-STC, the photovoltaic array formed by the photovoltaic modules of group B series and group a parallel may be expressed as the maximum power point voltage U m-array, the current I m-array and the power P m-array:
；
；
；
the photovoltaic array can utilize solar energy to the maximum extent through a maximum power point tracking technology (Maximum Power Point Tracking, MPPT) and output maximum power.
S12: the I-V characteristic curve of the photovoltaic array is combined, and the voltage, current and maximum power point change conditions of the photovoltaic array in different running states such as open circuit, short circuit, shading and aging are analyzed;
Photovoltaic cell models are typically based on physical characteristics to build an equivalent circuit, with single-diode and dual-diode photovoltaic cell models being common models.
As shown in fig. 3, if a single diode model is used, the expression of the relationship between the voltage and the current of the photovoltaic module is:
；
In the above formula: i and U respectively represent the output current and voltage of the photovoltaic module, I ph is the photo-generated current, I 0 is the reverse saturation current, R s is the equivalent series resistance, R sh is the equivalent parallel resistance, c is the ideal coefficient of the diode, k is the Boltzmann constant, and T a is the absolute temperature.
The I-V change characteristic curves of the photovoltaic array under different operation conditions are simulated, wherein the simulation comprises 5 working conditions of normal operation, a certain short circuit, a certain open circuit, partial shadow and abnormal aging, and the maximum power point and the I-V characteristic curves are shown in figure 4.
According to the maximum power point change of the photovoltaic module, it can be inferred that 4 types of common faults can directly reduce the power generation efficiency, but the photovoltaic array is usually not caused to stop running. And by combining with the comparison of the U-I characteristic curves, the voltage and the current of the photovoltaic module can be correspondingly changed under different working states. When a short circuit fault occurs, the rated voltage U m and the zero sequence voltage U o become smaller; when an open circuit fault occurs, the current instantaneous value I m and the current effective value I s become smaller; when locally occluded, I s is unchanged and U o is slightly reduced, and the curve shows a 'multi-knee' feature; when the power supply is abnormally aged, the open circuit voltage U oc and the short circuit current I sc are kept unchanged, but the inclination angle of the end section of the curve is obviously reduced, and the maximum power point shifts leftwards.
S13: and constructing two types of photovoltaic array operation state evaluation indexes comprising external environment factors and self electrical factors.
From the above analysis, the output power of the photovoltaic array is directly affected by irradiation and temperature.
The photovoltaic power generation power also shows different laws in different weather conditions (such as humidity) and seasons, as shown in fig. 5-6.
In order to realize accurate evaluation of the operation state of the photovoltaic array, two evaluation indexes, namely an external environment index and an electrical index of the photovoltaic array, are established according to main external environment factors and the characteristics of the photovoltaic power generation, as shown in fig. 7.
1) The external environment indexes mainly include: the radiation ratio R S, the temperature ratio R T and the humidity ratio R H are calculated according to the following formulas:
；
；
；
In the above formula, H is the actual ambient humidity, and H s is the seasonal relative humidity average value. As can be seen from fig. 4 to 5, as the irradiation intensity increases, the maximum power point moves upward, so the irradiation ratio is a forward index; as the temperature or humidity increases, the maximum power decreases, so the temperature ratio and humidity ratio are reverse indicators.
2) The self electrical indexes mainly comprise: the voltage ratio R U, the current ratio R I and the power ratio R P are calculated according to the following formula:
；
；
；
In the above formula: u is the actual output voltage, I is the rated actual output current, and P is the actual output power.
As can be seen from fig. 4 to 6, the voltage ratio, the current ratio, and the power ratio are all forward indicators, and the value ranges are [0,1].
S2: the membership function of the self-adaptive neural fuzzy inference system is improved, a fuzzy set is constructed by combining a mutation theory, and the multi-dimensional nonlinear output is realized by fusing a supermatrix, so that the self-adaptive neural fuzzy inference system has better fault tolerance and accuracy, and the self-adaptive neural fuzzy inference system specifically comprises the following steps:
S21: analysis of the original ANFIS model mechanism:
The Adaptive neural Fuzzy inference system (ANFIS, adaptive Network-based Fuzzy INFERENCE SYSTEM) is a combination of an Adaptive Network and a Fuzzy inference system, inherits the interpretability of the Fuzzy inference system and the learning capability of the Adaptive Network, utilizes a learning mechanism of the neural Network to automatically extract rules from input and output sample data to form an Adaptive neural Fuzzy controller, and can automatically adjust system parameters through a hybrid algorithm to enable the output of the system to be closer to actual output.
As shown in FIG. 8, the ANFIS is composed of 5 functional modules of an input layer, a membership layer, a rule layer, a decision layer and an output layer, wherein X 1 and X 2 are input variables and y is output variable, so that 2 if-then rules are formed. Rule 1: if x is a 1 and y is B 1, then f 1=p1x+q1y+r1; rule 2: if x is A 2 and y is B 2, then f 2=p2x+q2y+r2,p1、p2、q1、q2 is the linear coefficient of the input variable and r 1、r2 is the bias term parameter.
The ANFIS layers are described below:
Layer 1: for the self-adaptive node, fuzzy input features are used by using membership functions to obtain membership degree in a section and output The method comprises the following steps:
；
In the above formula: Ai is an associated variable of the adaptive node, and x is an input variable. Taking the gaussian function as an example, the membership function is expressed as:
；
In the above formula: c i and σ i are parameters set for gaussian functions.
Layer 2: is a fixed node, denoted by pi. The triggering intensity omega i of each rule is obtained by multiplying the membership degree of each feature and outputsThe method comprises the following steps:
；
In the above formula: Bi is the associated variable of the fixed node, and y is the output variable.
Layer 3: and classifying by N, and normalizing the triggering intensity omega i of each rule obtained in the previous layer, namely, the use probability of the rule in the whole reasoning process is used for representing the triggering proportion of the rule. Its outputGiven by the formula:
；
In the above formula: To normalize the trigger intensity.
Layer 4: for adaptive nodes, the output of each node is simply the product of the normalized trigger strength and the first order polynomial, and the outputThe method comprises the following steps:
；
In the above formula: p i、qi is the linear coefficient of the input variable, r i is the bias term parameter, and f i is the rule function.
Layer 5: for a fixed node, an accurate output is obtained by deblurring, characterized by Σ. Weighted average is carried out on the result of each rule to obtain the final output of the modelThe method comprises the following steps:
。
ANFIS overcomes the black box characteristic of a simple neural network, the incompleteness and the roughness of an inference rule in the fuzzy inference process, simplifies the data processing to the greatest extent, and has the characteristics of self-adaption, self-organization and self-learning.
S22: the ANFIS model is improved through membership function dynamic coding, fuzzy reasoning set mutation and a super-matrix multidimensional output structure, and is used for evaluating the running state grade;
the performance of the self-adaptive fuzzy reasoning system is mainly characterized by self-adaptive adjustment and error correction capability, and has important relation with membership functions and fuzzy reasoning rules.
1) Membership function dynamic coding method
Membership is a multi-factor decision method widely applied to fuzzy comprehensive judgment, and the evaluation results can be represented in fuzzy set forms with different degrees through membership functions.
Common membership functions are triangular, bell-shaped, trapezoidal, gaussian, etc., as shown in fig. 9. The existing researches focus on the selection of membership function types, and lack active correction of membership function related parameters and distribution shapes. According to the evolution theory, the membership function can be subjected to sectional dynamic coding, the minimum fuzzy evaluation error is taken as a target, heuristic algorithms such as particle swarm and artificial sheep swarm are utilized to carry out iterative optimization on the membership function parameters, and the shape of the membership function curve is accurately controlled.
Taking the triangular function as an example, as shown in fig. 10, considering the effect of the "open", "short", "shade", "aging", and "normal" operation state on the maximum power point, the fuzzy set of indexes may be sequentially marked as { I, II, III, IV, V }, and the function shape may be determined by 11 parameters, and the segment codes are s= { S1, S2,... Similarly, assuming that the gaussian membership functions are uniformly distributed, a single graph requires only standard deviation to describe, then a total of 8 parameters are required for encoding.
2) Fuzzy reasoning rule based on mutation theory
The basic principle of the mutation theory is to describe the state of the system by using a potential function f (x), and the mutation critical point set is represented by a balanced curved surface composed of a solution of differential equation df (x)/dx=0. Common potential function types of mutation systems are cusp mutation, dovetail mutation, butterfly mutation, etc., as shown in table 1.
Table 1 potential functions for different mutation types.
In the table, u, v, w and t are control variables. Based on a balanced curved surface equation and a singular point set equation which are obtained by different potential functions, a control variable and a state variable are controlled in the range of [0,1], and a normalization formula of the sharp point mutation, the dovetail mutation and the butterfly mutation can be obtained through derivation of a bifurcation equation as follows:
。
Taking environmental index as an example, the method comprises normalizing the irradiation ratio R S, the temperature ratio R T and the humidity ratio R H to obtain 、/>、/>Respectively through membership functions/>And calculating to obtain fuzzy set judgment matrixes of the { I, II, III, IV, V } five states.
Taking dovetail mutation as an example, the fuzzy set evaluation matrix R is as follows:
；
in the above formula: mu A、μB、μC、μD、μE is a weight factor.
3) ANFIS multidimensional output structure fusing super matrix
The supermatrix may be represented as a collection of multi-dimensional arrays, a widely used data structure in high-dimensional mathematics. Compared with the common matrix, the super matrix can not only accommodate more confidence, but also describe more complex nonlinear relations.
As shown in FIG. 11, for m input indexes and n types of running states, if the fuzzy and mutation treatment is performed on p types of different membership functions, a three-dimensional super matrix A m×n×p is generated at the 3 rd layer.
Compared with the original structure, the membership function code can be adaptively adjusted in the input part of the model through error correction and optimization iteration, so that more accurate fuzzy boundary is realized. In addition, a more complex supermatrix fuzzy rule is constructed through mutation theory, so that the fault tolerance and error correction capability of the model is improved, and multi-dimensional target state output is realized.
S3: the method comprises the following specific steps of:
s31: preparing a training sample, acquiring environmental index data through historical meteorological data, and constructing electrical index data on an I-V characteristic curve in a mode of classifying evenly-spaced sampling points;
In the actual operation process of the photovoltaic array, the photovoltaic array is in higher power generation efficiency most of the time, and various indexes can be changed under different states such as open circuit, short circuit, shading, aging, normal and the like.
In order to obtain a better training effect, the total number of samples is at least 5 times of the model parameters. In addition, the model requires further testing and verification after training, and the total sample needs to be randomly divided into a training set, a testing set and a verification set.
S32: preprocessing noise pollution caused by abnormal events such as data distortion, data noise, data missing and the like in sample data;
1) Distortion data processing: and screening and correcting the distortion data by adopting a transverse comparison method. Average value of Sum of variances/>Is calculated as follows:
；
；
In the above formula: x i,j is the data of the jth time node of the ith index, and n represents the nth index.
According to the statistical law, when the deviation of a certain type of index data x i,j is more than 3When the distortion data is corrected, the correction formula is as follows:
；
In the above formula: x i,j-1 and x i,j+1 are data of a node before and after the time node, α and β are correction coefficients, and α+β=1 is satisfied.
2) Missing data processing: interpolation is carried out by adopting a smoothing correction method, and the formula is as follows:
；
In the above formula: t 1 is the number of forward acquisition nodes, T 2 is the number of backward acquisition nodes, deltat 1 is the forward acquisition interval, deltat 2 is the backward acquisition interval.
3) Data normalization: according to the index property, adopting a Min-Max normalization method:
Forward index: ；
reverse index: ；
In the above formula: x imin is the minimum value in the sample data and x imax is the maximum value in the sample data.
S33: and (3) comprehensively analyzing the health index and the sensitivity, objectively weighting by adopting an entropy weight method according to fuzzy set judgment matrixes of indexes corresponding to different state grades, and further differentiating the index weights of different state grades.
Membership normalized values are as follows:
；
In the above formula: p ij is a membership standard value, m is an index number, mu ij is a membership function, i is an index ordinal number, and j is an operation state level ordinal number.
The information entropy is as follows:
；
In the above formula: e i is the information entropy of the running state, and n is the running state class number.
The weights are as follows:
；
In the above formula: w i is the weight of the index corresponding to different running states. According to the maximum membership principle, the index comprehensive membership can be expressed as:
。
the environmental index and the electric index influence are considered, and the comprehensive health index of the running state is constructed as follows:
；
In the above-mentioned method, the step of, To adjust the coefficient,/>E (0, 1), the choice of which depends on the influence of both in the comprehensive evaluation. Unless otherwise specified, take/>=0.5.
On the basis of improving the ANFIS model, the root mean square error is selected as a sensitivity analysis index of a training result, and the calculation formula is as follows:
；
In the above formula: n, t is the number of samples in the verification set, r i is the level of failure diagnosed by the model, Is the actual failure level.
The method of the invention is verified according to a specific example, by analysis of the data samples and comparison with the original method, the effectiveness of the proposed method is verified, in particular as follows:
And a certain 10kW small-sized photovoltaic test system is selected for verification, and the photovoltaic array is connected with 6 groups in series and 10 groups in parallel, and 60 photovoltaic group strings are all used. The environmental parameters that can be collected are: ambient temperature, ambient humidity, coplanar irradiance, etc.; the electrical parameters are: string voltage, string current, maximum output power, etc., for a period of 2023 years 9-12 months.
And taking points on the I-V curve at equal intervals by adopting Linspace functions, and matching corresponding environmental parameters to obtain more than 1000 groups of sample indexes.
According to the evaluation index nuclear density estimation curve of FIG. 12, typical values of the irradiation ratio, the temperature ratio and the humidity ratio in the environmental index are slightly less than 1, and the environmental index accords with the climate conditions of autumn and winter in the region; typical values of voltage ratio, current ratio and power ratio indexes in the electrical indexes are distributed between 0.95 and 1, and the generated power is reduced by less than 5%, which indicates that the photovoltaic array has good working condition.
According to the method in step S32, the evaluation index samples are preprocessed by adopting distortion rejection, missing interpolation and standardization, and are divided into a training set 612 group, a testing set 306 group and a verification set 305 group according to the ratio of 6:2:2. The distribution of the number of data set sample tags is shown in fig. 13 using a random grouping method.
Comprehensively considering the mutation of the triangular function 'break point' and the smooth transition characteristic of the Gaussian function 'smooth', respectively selecting two membership functions for fuzzy processing at the input part, and carrying out error correction by adopting a novel heuristic algorithm-artificial sheep algorithm iteration to obtain index membership parameter values as shown in Table 2, wherein the distribution is shown in FIG. 14.
Table 2 membership function coding parameters.
As shown in fig. 14, after the dynamic encoding process, the membership function graphic parameters of different indexes are not the same fixed values, but exhibit the differentiation characteristics. The method can effectively avoid the interference of manually setting membership functions in the original method, and has better independence and adaptability.
And performing nonlinear processing on the fuzzy set judgment matrix by using dovetail mutation, forming a super matrix fuzzy rule with a more complex structure on a third layer of the model, outputting defuzzification values of 5 operating states of open circuit, short circuit, shading, aging and normal after iterative training, respectively marking the defuzzification values as { I, II, III, IV and V }, and then judging the fault level according to a maximum value corresponding principle.
After the modified ANFIS model was trained and tested, 305 samples were selected for validation and compared to the original ANFIS method, the results are shown in FIG. 15. Compared with the original ANFIS model, the improved ANFIS model obviously reduces the degree of deviation of the diagnosis result from the actual state, and the accuracy of the model is effectively improved by demonstrating membership dynamic coding and mutation theory fuzzy set.
To further verify the sensitivity of the method of the present invention, a ten fold cross-validation method was used to re-divide the original samples into training, test and validation sets, and analysis was performed using root mean square error of 10 repeated experimental results, the results of which are shown in fig. 16. The root mean square error of 10 repeated experiments is between 0.002 and 0.012, which shows that the improved ANFIS model has better accuracy, stronger sensitivity to different inputs and certain robustness.
And (3) carrying out comprehensive health index scoring on the running state of the verification set according to objective weighting by an entropy weighting method and combining the standardized value of the input index. When adjusting the coefficientWhen=0.3, the distribution result is shown in fig. 17. The comprehensive health indexes of the operation state of the photovoltaic array are mainly concentrated and distributed in [0.75,0.85], and the sample accounts for about 70%, which indicates that the operation level of the photovoltaic array is generally good. For health indexes below 0.75, four fault conditions are possible. Therefore, the operation and maintenance of the photovoltaic array can be enhanced by combining a fault diagnosis method, and accidents are avoided.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (10)
1. A photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS is characterized in that: the method comprises the following steps:
s1: analyzing and determining external environment factors and self-electric factors which influence the power generation of the photovoltaic array, and constructing corresponding photovoltaic array running state evaluation indexes according to the analyzed factors, wherein the evaluation indexes comprise external environment indexes and self-electric indexes;
S2: building an improved ANFIS model to evaluate the operation state grade of the photovoltaic array: the improved ANFIS model improves the membership function of the self-adaptive fuzzy inference system, constructs a fuzzy set by combining a mutation theory, and fuses a supermatrix to realize multidimensional nonlinear output;
S3: and comprehensively weighting the membership of the evaluation index based on an entropy weight method, further taking external environmental factors and self electrical factors into consideration to establish corresponding health indexes, realizing quantification of the operation state of the photovoltaic array, and evaluating the operation state of the photovoltaic array and diagnosing faults according to the health indexes.
2. The improved ANFIS-based photovoltaic array operating state evaluation and fault diagnosis method according to claim 1, wherein: the step S1 specifically comprises the following steps:
S11: according to the characteristics of the solar cell, analyzing a mathematical model of the output power of the photovoltaic array;
s12: the voltage, current and maximum power point change conditions under five running states of open circuit, short circuit, shading and aging are analyzed by combining an I-V characteristic curve of the photovoltaic array;
S13: and constructing two types of photovoltaic array operation state evaluation indexes comprising external environment factors and self electrical factors.
3. The improved ANFIS-based photovoltaic array operating state evaluation and fault diagnosis method according to claim 2, wherein: the external environment indexes comprise an irradiation ratio R S, a temperature ratio R T and a humidity ratio R H, and the calculation formulas are as follows:
；
；
；
In the above formula, S is the actual horizontal irradiance, S 0 is the standard light intensity, T is the actual environment temperature, T ref is the standard temperature, H is the actual environment humidity, H s is the average value of the seasonal relative humidity, the irradiation ratio is the forward index, and the temperature ratio and the humidity ratio are the reverse index;
The self electrical indexes comprise a voltage ratio R U, a current ratio R I and a power ratio R P, and the calculation formulas are as follows:
；
；
；
in the above formula: u m-array is the maximum power point voltage, I m-STC is the maximum power point current, P m-array is the maximum power, U is the actual output voltage, I is the actual output current, and P is the actual output power.
4. The method for evaluating the operation state and diagnosing faults of a photovoltaic array based on improved ANFIS according to claim 3, wherein the method comprises the following steps of: the step S2 specifically comprises the following steps:
S21: analyzing an original ANFIS model mechanism;
s22: the ANFIS model is improved through membership function dynamic coding, fuzzy reasoning set mutation and a super-matrix multidimensional output structure, and is used for evaluating the running state level;
s23: and iteratively correcting errors through a heuristic optimization algorithm.
5. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method of claim 4, wherein: the membership function adopts a sectional dynamic coding method, aims at minimizing fuzzy evaluation errors, and performs iterative optimization on membership function parameters by using a heuristic algorithm;
The membership function considers the influences of five states of open circuit, short circuit, shading, aging and normal operation on the maximum power point of the photovoltaic array, obtains a fuzzy set of indexes, marks the fuzzy set, determines parameters of different membership functions, and encodes the membership functions in a segmented mode according to the parameters.
6. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method of claim 5, wherein: the fuzzy set is constructed by combining a mutation theory, namely, after two types of evaluation indexes are respectively normalized, membership function calculation is carried out to obtain fuzzy set evaluation matrixes of five running states, then one potential function type is selected, and the fuzzy set evaluation matrixes after nonlinear processing are obtained according to the selected potential function.
7. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method of claim 6, wherein: the step S3 specifically comprises the following steps:
s31: preparing a training sample, acquiring environmental index data through historical meteorological data, and constructing electrical index data on an I-V characteristic curve in a mode of classifying evenly-spaced sampling points;
s32: preprocessing data distortion, data noise and data missing noise pollution existing in sample data;
s33: comprehensive health index and sensitivity analysis, carrying out objective weighting by adopting an entropy weighting method according to fuzzy set judgment matrixes with indexes corresponding to different state grades, and further differentiating the index weights of different state grades;
wherein the membership degree normalization value is as follows:
；
In the above formula: p ij is a membership standard value, m is an index number, mu ij is a membership function, i is an index ordinal number, and j is an operation state level ordinal number;
The information entropy is as follows:
；
In the above formula: e i is the information entropy of the running state, and n is the number of running state grades;
the weights are as follows:
；
In the above formula: w i is the weight of the index corresponding to different running states;
According to the maximum membership principle, the index comprehensive membership is expressed as:
；
the environmental index and the electric index influence are considered, and the comprehensive health index of the running state is constructed as follows:
；
In the above-mentioned method, the step of, To adjust the coefficient,/>∈(0,1)，/>Is the electrical index of the self.
8. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method of claim 7, wherein: on the basis of improving the ANFIS model, the root mean square error is selected as a sensitivity analysis index of a training result, and the calculation formula is as follows:
；
In the above formula: n, t is the number of samples in the verification set, r i is the level of failure diagnosed by the model, Is the actual failure level.
9. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method according to any one of claims 1 to 8, wherein: the judging steps of the fault level are as follows:
Taking points on the I-V curve at equal intervals by adopting Linspace functions, and matching corresponding environmental parameters to obtain a plurality of groups of sample indexes;
comprehensively considering the break point mutation of the triangular function and the smooth transition characteristic of the Gaussian function, respectively selecting two membership functions for fuzzy processing in an input part, and carrying out error correction by adopting iteration of an artificial sheep algorithm;
And (3) performing nonlinear processing on the fuzzy set judgment matrix by using dovetail mutation, forming a hyper-matrix fuzzy rule with a more complex structure on a third layer of the model, outputting defuzzified values of open circuit, short circuit, shading, aging and normal 5 running states after iterative training, and judging the fault level according to a maximum value corresponding principle.
10. The improved ANFIS-based photovoltaic array operating condition assessment and fault diagnosis method of claim 7, wherein: the method for preprocessing the sample data in step S32 is as follows:
1) Distortion data processing: screening and correcting distortion data by a transverse comparison method;
2) Missing data processing: interpolation is carried out by adopting a smooth correction method;
3) Data normalization: and according to the evaluation index property, adopting a Min-Max normalization method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410553196.0A CN118134290B (en) | 2024-05-07 | 2024-05-07 | Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410553196.0A CN118134290B (en) | 2024-05-07 | 2024-05-07 | Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118134290A true CN118134290A (en) | 2024-06-04 |
| CN118134290B CN118134290B (en) | 2024-07-23 |
Family
ID=91236397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410553196.0A Active CN118134290B (en) | 2024-05-07 | 2024-05-07 | Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118134290B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118469406A (en) * | 2024-07-11 | 2024-08-09 | 国网上海市电力公司 | Dynamic evaluation method and system for power substitution implementation index based on variable weight optimization |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB201121695D0 (en) * | 2011-12-16 | 2012-02-01 | London Metropolitan University | Analysis tool and methodology |
| WO2012023138A2 (en) * | 2010-08-18 | 2012-02-23 | Technion Research And Development Foundation Ltd. | Volatile organic compounds for detecting cell dysplasia and genetic alterations associated with lung cancer |
| US20150188415A1 (en) * | 2013-12-30 | 2015-07-02 | King Abdulaziz City For Science And Technology | Photovoltaic systems with maximum power point tracking controller |
| CN109670553A (en) * | 2018-12-25 | 2019-04-23 | 福州大学 | Diagnosing failure of photovoltaic array method based on Adaptive Neuro-fuzzy Inference |
| CN116819251A (en) * | 2023-06-27 | 2023-09-29 | 国网上海市电力公司 | High-voltage switch cabinet fault diagnosis method based on photoelectric monitoring |
| CN116937638A (en) * | 2023-06-13 | 2023-10-24 | 国网浙江省电力有限公司平湖市供电公司 | Energy regulation and control method based on fuzzy logic reasoning |
-
2024
- 2024-05-07 CN CN202410553196.0A patent/CN118134290B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012023138A2 (en) * | 2010-08-18 | 2012-02-23 | Technion Research And Development Foundation Ltd. | Volatile organic compounds for detecting cell dysplasia and genetic alterations associated with lung cancer |
| GB201121695D0 (en) * | 2011-12-16 | 2012-02-01 | London Metropolitan University | Analysis tool and methodology |
| US20150188415A1 (en) * | 2013-12-30 | 2015-07-02 | King Abdulaziz City For Science And Technology | Photovoltaic systems with maximum power point tracking controller |
| CN109670553A (en) * | 2018-12-25 | 2019-04-23 | 福州大学 | Diagnosing failure of photovoltaic array method based on Adaptive Neuro-fuzzy Inference |
| CN116937638A (en) * | 2023-06-13 | 2023-10-24 | 国网浙江省电力有限公司平湖市供电公司 | Energy regulation and control method based on fuzzy logic reasoning |
| CN116819251A (en) * | 2023-06-27 | 2023-09-29 | 国网上海市电力公司 | High-voltage switch cabinet fault diagnosis method based on photoelectric monitoring |
Non-Patent Citations (3)
| Title |
|---|
| ZHANG, Y, 等: "Photovoltaic MPPT algorithm based on adaptive particle swarm optimization neural-fuzzy control", JOURNAL OF INTELLIGENT & FUZZY SYSTEMS, vol. 44, no. 1, 3 March 2023 (2023-03-03), pages 341 - 351 * |
| 王玮茹;: "基于模糊自适应PI控制的光伏发电MPPT", 电网与清洁能源, vol. 27, no. 06, 25 June 2011 (2011-06-25), pages 61 - 64 * |
| 蒋文波: "基于EVA-突变级数法的光伏企业估值探讨——以阳光电源为例", 中国优秀硕士学位论文全文数据库工程科技Ⅱ辑, no. 02, 15 February 2023 (2023-02-15), pages 042 - 396 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118469406A (en) * | 2024-07-11 | 2024-08-09 | 国网上海市电力公司 | Dynamic evaluation method and system for power substitution implementation index based on variable weight optimization |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118134290B (en) | 2024-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11070056B1 (en) | Short-term interval prediction method for photovoltaic power output | |
| CN118134290B (en) | Photovoltaic array running state evaluation and fault diagnosis method based on improved ANFIS | |
| CN109670553B (en) | Photovoltaic array fault diagnosis method based on adaptive neural fuzzy inference system | |
| CN109842373B (en) | Photovoltaic array fault diagnosis method and device based on space-time distribution characteristics | |
| Liu et al. | Fault diagnosis approach for photovoltaic array based on the stacked auto-encoder and clustering with IV curves | |
| CN108205717A (en) | A kind of photovoltaic generation power Multiple Time Scales Forecasting Methodology | |
| CN112101669B (en) | Photovoltaic power interval prediction method based on improved extreme learning machine and quantile regression | |
| KR102230548B1 (en) | Power generation prediction and efficiency diagnosis system of solar power generation facilities using FRBFNN model | |
| CN110503153B (en) | Photovoltaic system fault diagnosis method based on differential evolution algorithm and support vector machine | |
| Nespoli et al. | A selective ensemble approach for accuracy improvement and computational load reduction in ann-based pv power forecasting | |
| CN111626473A (en) | Two-stage photovoltaic power prediction method considering error correction | |
| CN112215478B (en) | Power coordination control method and device for optical storage station and storage medium | |
| CN112994101B (en) | Neural network-based wind power plant generated power post-evaluation and monitoring method | |
| CN112465271B (en) | Energy storage battery type selection method for energy storage stabilizing wind power fluctuation scene | |
| CN115208310A (en) | Monitoring method of photovoltaic power generation system | |
| CN115618922A (en) | Photovoltaic power prediction method and device, photovoltaic power generation system and storage medium | |
| CN111091141B (en) | Photovoltaic backboard fault diagnosis method based on layered Softmax | |
| CN117218425A (en) | Power generation loss analysis method and system for photovoltaic power station | |
| CN112232714B (en) | Deep learning-based risk assessment method for distribution network under incomplete structural parameters | |
| CN108038518A (en) | A kind of photovoltaic generation power based on meteorological data determines method and system | |
| CN113112085A (en) | New energy station power generation load prediction method based on BP neural network | |
| Kaushik et al. | Performance Analysis of Regression Models in Solar PV Forecasting | |
| CN113205214A (en) | Photovoltaic power ultra-short-term prediction method based on parallel dendritic neural network | |
| Pa | Application of Artificial Intelligence in fault detection and classification of solar power plants and prediction of power generation of combined cycled power plants | |
| Duan et al. | Estimation of Solar Power by Combining Ground Measured and Satellite Data Using Artificial Neural Networks |
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 |