CN115587642B - BP neural network-based photovoltaic system fault alarm method - Google Patents

BP neural network-based photovoltaic system fault alarm method Download PDF

Info

Publication number
CN115587642B
CN115587642B CN202210712863.6A CN202210712863A CN115587642B CN 115587642 B CN115587642 B CN 115587642B CN 202210712863 A CN202210712863 A CN 202210712863A CN 115587642 B CN115587642 B CN 115587642B
Authority
CN
China
Prior art keywords
neural network
photovoltaic system
network model
distributed photovoltaic
input
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
Application number
CN202210712863.6A
Other languages
Chinese (zh)
Other versions
CN115587642A (en
Inventor
贾炳军
刘彦鹏
程永卓
王永宏
唐宏芬
公冶令沛
李红涛
董颖华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Datang Sanya Future Energy Research Institute Co ltd
New Energy Branch Of Datang Hainan Energy Development Co ltd
Datang Hainan Energy Development Co ltd
China Electric Power Research Institute Co Ltd CEPRI
Original Assignee
Datang Sanya Future Energy Research Institute Co ltd
New Energy Branch Of Datang Hainan Energy Development Co ltd
Datang Hainan Energy Development Co ltd
China Electric Power Research Institute Co Ltd CEPRI
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Datang Sanya Future Energy Research Institute Co ltd, New Energy Branch Of Datang Hainan Energy Development Co ltd, Datang Hainan Energy Development Co ltd, China Electric Power Research Institute Co Ltd CEPRI filed Critical Datang Sanya Future Energy Research Institute Co ltd
Priority to CN202210712863.6A priority Critical patent/CN115587642B/en
Publication of CN115587642A publication Critical patent/CN115587642A/en
Application granted granted Critical
Publication of CN115587642B publication Critical patent/CN115587642B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/185—Electrical failure alarms
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Development Economics (AREA)
  • General Engineering & Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Emergency Management (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention discloses a photovoltaic system fault warning method based on a BP neural network, which is characterized in that training data are screened out by establishing the BP neural network of a distributed photovoltaic system, and further, abnormal output characteristics are warned in the operation process of the distributed photovoltaic system in an area.

Description

BP neural network-based photovoltaic system fault alarm method
Technical Field
The invention relates to the field of fault diagnosis of a distributed photovoltaic system, in particular to a photovoltaic system fault alarming method based on a BP neural network.
Background
The BP neural network has arbitrary complex pattern classification capability and excellent multidimensional function mapping capability, and solves Exclusive OR (XOR) and other problems which cannot be solved by a simple perceptron. Structurally, a BP network has an input layer, a hidden layer and an output layer; basically, the BP algorithm uses the square of the network error as an objective function, and uses a gradient descent method to calculate the minimum value of the objective function. In the prior art, the input data of the photovoltaic system fault alarm method is single, the accuracy requirement on the acquisition of the early-stage data is high, when running data is absent, fault alarm is easy to miss or misinformation, in the prior art, CN201510567882 discloses a distributed photovoltaic system fault diagnosis method based on a neural network, the power station is diagnosed only through historical input and output data of the diagnosed power station and meteorological data monitored by the power station, the diagnosis precision is low, meanwhile, the residual error between estimation output and actual output of the system is adopted in the prior art to judge, and the error is large.
Disclosure of Invention
Aiming at the defects of the prior art, the invention discloses a photovoltaic system fault alarming method based on a BP neural network, theoretical output data of an estimated photovoltaic power station is obtained through output data and meteorological data of each photovoltaic system in the whole county, the theoretical output data and actual output data of the estimated photovoltaic power station are judged by adopting Pearson correlation coefficients, further, the fault alarming condition of the estimated photovoltaic power station is obtained, the fault alarming of the photovoltaic power station is realized, and the alarming accuracy is high.
The technical scheme of the application specifically comprises the following steps:
a photovoltaic system fault alarming method based on BP neural network is characterized in that training data are screened out by establishing a distributed photovoltaic system BP neural network, and then alarming is carried out on abnormal output characteristics in the running process of an intra-area distributed photovoltaic system.
The invention discloses a photovoltaic system fault alarm method based on a BP neural network, which comprises the following steps:
s1, acquiring output data of a distributed photovoltaic system in a regional photovoltaic system and meteorological data of the regional photovoltaic system;
s2, carrying out normalization processing on output data and meteorological data of the distributed photovoltaic system;
s3, based on the output data and the meteorological data of the distributed photovoltaic system after normalization processing, establishing a BP neural network model input matrix x of the distributed photovoltaic system t-i * Inputting BP neural network model into matrix x t-i * Inputting an implicit layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area;
s4, normalizing output data of the evaluated distributed photovoltaic system, and inputting matrix x under the same time stamp t-i * And output vector Y t * Performing correlation judgment to obtain filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000011
And Meteorological data->
Figure GDA0004265776800000012
S5, distributing the filtered mixtureOutput data of photovoltaic system
Figure GDA0004265776800000021
And Meteorological data->
Figure GDA0004265776800000022
As the input training parameters of the BP neural network model of the current area photovoltaic system, training the BP neural network model of the current area photovoltaic system, and outputting a trained neural network system;
s6, applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model, and obtaining an output reference vector
Figure GDA0004265776800000023
Calculating an output reference vector +.>
Figure GDA0004265776800000024
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure GDA0004265776800000025
And judging whether the estimated distributed photovoltaic power station has suspicious faults according to the correlation.
A photovoltaic system fault alarm method based on BP neural network is characterized in that,
step S1, specifically comprising the following steps:
101 Determining available distributed photovoltaic systems in the regional environment range of the regional photovoltaic system, wherein the regional photovoltaic system comprises more than 1 other distributed photovoltaic systems besides the evaluated photovoltaic power generation system, and acquiring output data of the distributed photovoltaic systems;
102 For the photovoltaic system being evaluated in the area and other n distributed photovoltaic systems, n>=1, respectively obtaining output data at t time, wherein the output data of the estimated distributed photovoltaic system at t time is recorded as a vector Y t Output data matrix is formed by output data of n distributed photovoltaic systemsP t-i ，i＝1,2，…，n；P t-i The output data of the ith distributed photovoltaic system at the t moment is represented;
103 Acquiring meteorological data of a regional photovoltaic system at the moment t, wherein the meteorological data comprise data of irradiance, temperature and wind speed of 3 traditional parameters, and the meteorological data of the tropical region also comprise humidity data to form a meteorological data vector W at the moment t t-j ，j＝1,2，3，4…，m，W t-j The j-th meteorological data vector at time t is represented, in this embodiment, m=4.
Preferably, the step S2 specifically includes the following steps:
201 Based on the installed capacity of n distributed photovoltaic systems, normalizing the output data of the distributed photovoltaic systems, taking the normalized parameters as input training parameters of the BP neural network model, wherein the total number of all the input parameters of the BP neural network model is n, and the normalized output data matrix at the moment t is
Figure GDA0004265776800000026
202 When available meteorological data exists in the environmental range of the regional photovoltaic system, the regional photovoltaic meteorological data is normalized by adopting irradiance of 1000W/square meter, environmental temperature of 45 ℃ and wind speed of 10m/s and humidity of 100% as the basis, and a normalized meteorological matrix W at the moment t is formed t-j * The normalized output data matrix is taken as
Figure GDA0004265776800000027
And weather matrix W t-j * As the input training parameters of the BP neural network model, n+4 total input parameters of the BP neural network model are provided.
Preferably, the step S3 specifically includes the following steps:
301 Building an input matrix x of BP neural network model of regional photovoltaic system at t moment t-i * Is of formula (1):
Figure GDA0004265776800000031
302 According to the input matrix x) t-i * Determining the number of hidden layers of the BP neural network model according to the dimension of the BP neural network model and the requirement on the precision of the prediction result;
303 Using M-layer unipolar Sigmoid function as an hidden layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area;
the hidden layer activation function f (v) of the BP neural network model is expressed as formula (2):
Figure GDA0004265776800000032
wherein v is an input parameter, which is the multi-layer input of the hidden layer of the BP neural network model, and the initial value of the input parameter is an input matrix x t-i * The first layer is the input matrix x t-i * The input v of the second layer is the result after the calculation of the first layer, the v domain is (- ≡, ++ infinity A kind of electronic device.
Preferably, the step S4 specifically includes the following steps:
401 Normalizing the output data of the evaluated distributed photovoltaic system according to the installed capacity of the evaluated distributed photovoltaic system in the region, wherein the output vector of the normalized output data of the evaluated distributed photovoltaic system is Y t * ；
402 For input matrix x at time t) t-i * And output vector Y t * Screening, and judging the relevance of the input and output functions by using the Pearson correlation coefficient;
Figure GDA0004265776800000033
wherein: c is sample x t-i * And sample Y t * Correlation coefficients; cov (x) t-i * ,Y t * ) For sample x t-i * And sample Y t * Co-ordination of correlation coefficientsDifference, var (x t-i * )、Var(Y t * ) Fractional sample x t-i * And sample Y t * Is a variance of (2);
403 Setting the first correlation coefficient threshold value as the input matrix x with the filtered correlation coefficient c smaller than the first correlation coefficient threshold value t-i * And the output vector is Y t * Obtaining filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000034
And Meteorological data->
Figure GDA0004265776800000035
Preferably, the step S5 specifically includes the following steps:
filtered output data of distributed photovoltaic system at t moment
Figure GDA0004265776800000036
And Meteorological data->
Figure GDA0004265776800000037
As the input training parameters of the BP neural network model of the current area photovoltaic system; hidden layer activation function f (v) based on BP neural network model * ) Training the neural network, and optimizing the neural network; using the filtered input and output matrices, and the layer 2 hidden layer activation function f (v * ) Training the neural network to form a multi-layer neural network to obtain a trained neural network system; the method comprises the steps of carrying out a first treatment on the surface of the
Figure GDA0004265776800000041
v * Is the filtered input parameter of BP neural network model, in this step, v * The initial value of (1) is the filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000042
And Meteorological data->
Figure GDA0004265776800000043
Preferably, the step S6 specifically includes the following steps:
601 Applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model in real time, and obtaining corresponding output reference vectors
Figure GDA0004265776800000044
Outputting the reference vector
Figure GDA0004265776800000045
Is obtained by inputting an input vector into a multi-layer neural network composed of the formula (2).
602 Calculating an output reference vector using Pearson correlation coefficients
Figure GDA0004265776800000046
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure GDA0004265776800000047
Correlation, denoted as c Y ：
Figure GDA0004265776800000048
603 Setting a second correlation coefficient threshold, when c y And when the correlation coefficient is smaller than the second correlation coefficient threshold value of 0.95, the estimated distributed photovoltaic power station is considered to have suspicious faults, and alarming is carried out.
Preferably, the first correlation coefficient threshold is 0.95.
The second correlation coefficient threshold is 0.95.
Compared with the prior art, the invention has the following beneficial effects:
the invention discloses a photovoltaic system fault alarming method based on a BP neural network, which alarms according to the fault condition of a distributed photovoltaic system in a tropical region, further improves the operation maintenance level of the distributed photovoltaic system in the region, and improves the generated energy of the distributed photovoltaic system.
According to the invention, a BP neural network method is adopted, theoretical output data of the photovoltaic power station to be evaluated is obtained through output data and meteorological data of each photovoltaic system in the county, the theoretical output data and actual output data of the photovoltaic power station to be evaluated are judged by adopting Pearson correlation coefficients, so that fault alarm conditions of the photovoltaic power station to be evaluated are obtained, and fault alarm of the photovoltaic power station is realized.
The input quantity of the photovoltaic power station fault diagnosis alarm system not only comprises the input and output data and meteorological data of the diagnosed power station, but also comprises the power generation data of surrounding photovoltaic power stations, and when certain operation data is absent, the fault diagnosis alarm of the photovoltaic power stations can be realized through more input data.
The Pearson correlation coefficient is adopted for fault judgment, and the judgment method is simple and high in precision.
Drawings
FIG. 1 is a flow chart of a photovoltaic system fault warning method based on a BP neural network;
FIG. 2 is a regional photovoltaic fault alert neural network diagram;
and (3) alarming the fault of the photovoltaic system in the area of figure 3.
Detailed Description
The following description of the embodiments of the present invention will be made more apparent and fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be made by one of ordinary skill in the art without undue burden on the person of ordinary skill in the art based on embodiments of the present invention, are within the scope of the present invention.
As shown in fig. 1, in the photovoltaic system fault warning method based on the BP neural network, training data are screened out by establishing the distributed photovoltaic system BP neural network, so that the warning is performed on abnormal output characteristics in the running process of the distributed photovoltaic system in the region (the region is divided geographically and can refer to the whole county, and the patent is used for advancing the photovoltaic in the whole county), and the running and maintenance capabilities of the distributed photovoltaic system are improved.
A photovoltaic system fault alarm method based on BP neural network comprises the following steps:
s1, acquiring output data of a distributed photovoltaic system in a regional photovoltaic system and meteorological data of the regional photovoltaic system;
s2, carrying out normalization processing on output data and meteorological data of the distributed photovoltaic system;
s3, based on the output data and the meteorological data of the distributed photovoltaic system after normalization processing, establishing a BP neural network model input matrix x of the distributed photovoltaic system t-i * Inputting BP neural network model into matrix x t-i * Inputting an implicit layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area;
f (v) is a common function of the neural network, and the function f (v) is trained by using known input and output data through two steps S4 and S5, so that the exclusive neural network for alarming the power station fault can be obtained
S4, normalizing output data of the evaluated distributed photovoltaic system, and inputting matrix x under the same time stamp t-i * And output vector Y t * Performing correlation judgment to obtain filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000051
And weather data
Figure GDA0004265776800000052
S5, outputting the filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000053
And Meteorological data->
Figure GDA0004265776800000054
As the input training parameters of the BP neural network model of the current area photovoltaic system, training the BP neural network model of the current area photovoltaic system, and outputting a trained neural network system;
s6, applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model, and obtaining an output reference vector
Figure GDA0004265776800000055
Calculating an output reference vector +.>
Figure GDA0004265776800000056
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure GDA0004265776800000057
And judging whether the estimated distributed photovoltaic power station has suspicious faults according to the correlation.
Step S1, specifically comprising the following steps:
101 Determining available distributed photovoltaic systems in the regional environment range of the regional photovoltaic system, wherein the regional photovoltaic system comprises more than 1 other distributed photovoltaic systems besides the evaluated photovoltaic power generation system, and acquiring output data of the distributed photovoltaic systems;
102 For the photovoltaic system being evaluated in the area and other n distributed photovoltaic systems, n>=1, respectively obtaining output data at t time, wherein the output data of the estimated distributed photovoltaic system at t time is recorded as a vector Y t The output data of n distributed photovoltaic systems form an output data matrix P t-i ，i＝1,2，…，n；P t-i The output data of the ith distributed photovoltaic system at the t moment is represented;
103 Acquiring meteorological data of a regional photovoltaic system at t time, wherein the meteorological data comprise data of irradiance, temperature and wind speed of 3 traditional parameters, and the meteorological data of the tropical region also comprise humidity dataMeteorological data vector W at t moment t-j ，j＝1,2，3，4…，m，W t-j And represents the j-th meteorological data vector at the t moment.
The step S2 specifically comprises the following steps:
201 Based on the installed capacity of n distributed photovoltaic systems, normalizing the output data of the distributed photovoltaic systems, taking the normalized parameters as input training parameters of the BP neural network model, wherein the total number of all the input parameters of the BP neural network model is n, and the normalized output data matrix at the moment t is
Figure GDA0004265776800000061
202 When available meteorological data exists in the environmental range of the regional photovoltaic system, the regional photovoltaic meteorological data is normalized by adopting irradiance of 1000W/square meter, environmental temperature of 45 ℃ and wind speed of 10m/s and humidity of 100% as the basis, and a normalized meteorological matrix W at the moment t is formed t-j * The normalized output data matrix is taken as
Figure GDA0004265776800000062
And weather matrix W t-j * As the input training parameters of the BP neural network model, n+4 total input parameters of the BP neural network model are provided.
The step S3 specifically comprises the following steps:
301 Building an input matrix x of BP neural network model of regional photovoltaic system at t moment t-i * Is of formula (1):
Figure GDA0004265776800000063
302 According to the input matrix x) t-i * Determining the number of hidden layers of the BP neural network model according to the dimension of the BP neural network model and the requirement on the precision of the prediction result;
the number of hidden layers is determined according to an empirical value, the more the number of layers is, the more complex the calculation is, the fewer the number of layers is, and the calculation accuracy is not enough, and 2 layers are adopted in the embodiment.
303 Using M-layer unipolar Sigmoid function as an hidden layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area; the BP neural network model of the current area photovoltaic system is shown in figure 2;
the hidden layer activation function f (v) of the BP neural network model is expressed as formula (2):
Figure GDA0004265776800000071
wherein v is an input parameter, which is the multi-layer input of the hidden layer of the BP neural network model, and the initial value of the input parameter is an input matrix x t-i * The first layer is the input matrix x t-i * The input v of the second layer is the result after the calculation of the first layer, the v domain is (- ≡, ++ infinity A kind of electronic device.
The step S4 specifically comprises the following steps:
401 Normalizing the output data of the evaluated distributed photovoltaic system according to the installed capacity of the evaluated distributed photovoltaic system in the region, wherein the output vector of the normalized output data of the evaluated distributed photovoltaic system is Y t * ；
402 For input matrix x at time t) t-i * And output vector Y t * Screening, and judging the relevance of the input and output functions by using the Pearson correlation coefficient;
Figure GDA0004265776800000072
wherein: c is sample x t-i * And sample Y t * Correlation coefficients; cov (x) t-i * ,Y t * ) For sample x t-i * And sample Y t * Covariance of correlation coefficient, var (x t-i * )、Var(Y t * ) Fractional sample x t-i * And sample Y t * Is a variance of (2);
403 Setting the first correlation coefficient threshold value to 0.95, filtering the input matrix x with the correlation coefficient c smaller than the first correlation coefficient threshold value t-i * And the output vector is Y t * Obtaining filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000073
And weather data
Figure GDA0004265776800000074
The step S5 specifically comprises the following steps:
filtered output data of distributed photovoltaic system at t moment
Figure GDA0004265776800000075
And Meteorological data->
Figure GDA0004265776800000076
As the input training parameters of the BP neural network model of the current area photovoltaic system; hidden layer activation function f (v) based on BP neural network model * ) Training the neural network, and optimizing the neural network; using the filtered input and output matrices, and the layer 2 hidden layer activation function f (v * ) Training the neural network to form a multi-layer neural network to obtain a trained neural network system; the method comprises the steps of carrying out a first treatment on the surface of the
Figure GDA0004265776800000077
v * Is the filtered input parameter of BP neural network model, in this step, v * The initial value of (1) is the filtered output data of the distributed photovoltaic system
Figure GDA0004265776800000078
And Meteorological data->
Figure GDA0004265776800000079
The step S6 specifically comprises the following steps:
601 Applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model in real time, and obtaining corresponding output reference vectors
Figure GDA0004265776800000081
Outputting the reference vector
Figure GDA0004265776800000082
Is obtained by inputting an input vector into a multi-layer neural network composed of the formula (2).
602 Calculating an output reference vector using Pearson correlation coefficients
Figure GDA0004265776800000083
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure GDA0004265776800000084
Correlation, denoted as c Y ：
Figure GDA0004265776800000085
603 Setting the second correlation coefficient threshold to 0.95, when c y When the correlation coefficient is smaller than the second correlation coefficient threshold value of 0.95, the estimated distributed photovoltaic power station is considered to have suspicious faults, alarming is carried out, the domain distributed photovoltaic fault alarming result is shown in fig. 3, and the prediction alarming is consistent with the actual result.
The steps S1-S4 are to build a neural network model, train the neural network model through known and definite input and output data, and enable the trained model to have the capability of realizing the steps S5-S6. That is, the model which is just built is a general model which cannot be recognized, and the model can be targeted to the power station of the local area only through the training of S1-S4, so that the steps S5-S6 can be realized.
In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be construed as reflecting the intention that: i.e., the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.
Those skilled in the art will appreciate that the modules or units or groups of devices in the examples disclosed herein may be arranged in a device as described in this embodiment, or alternatively may be located in one or more devices different from the devices in this example. The modules in the foregoing examples may be combined into one module or may be further divided into a plurality of sub-modules.
Those skilled in the art will appreciate that the modules in the apparatus of the embodiments may be adaptively changed and disposed in one or more apparatuses different from the embodiments. The modules or units or groups of embodiments may be combined into one module or unit or group, and furthermore they may be divided into a plurality of sub-modules or sub-units or groups. Any combination of all features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or units of any method or apparatus so disclosed, may be used in combination, except insofar as at least some of such features and/or processes or units are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.
Furthermore, some of the embodiments are described herein as methods or combinations of method elements that may be implemented by a processor of a computer system or by other means of performing the functions. Thus, a processor with the necessary instructions for implementing the described method or method element forms a means for implementing the method or method element. Furthermore, the elements of the apparatus embodiments described herein are examples of the following apparatus: the apparatus is for carrying out the functions performed by the elements for carrying out the objects of the invention.
The various techniques described herein may be implemented in connection with hardware or software or, alternatively, with a combination of both. Thus, the methods and apparatus of the present invention, or certain aspects or portions of the methods and apparatus of the present invention, may take the form of program code (i.e., instructions) embodied in tangible media, such as floppy diskettes, CD-ROMs, hard drives, or any other machine-readable storage medium, wherein, when the program is loaded into and executed by a machine, such as a computer, the machine becomes an apparatus for practicing the invention.
In the case of program code execution on programmable computers, the computing device will generally include a processor, a storage medium readable by the processor (including volatile and non-volatile memory and/or storage elements), at least one input device, and at least one output device. Wherein the memory is configured to store program code; the processor is configured to execute the evaluation method of the invention in accordance with instructions in said program code stored in the memory.
By way of example, and not limitation, computer readable media comprise computer storage media and communication media. Computer-readable media include computer storage media and communication media. Computer storage media stores information such as computer readable instructions, data structures, program modules, or other data. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. Combinations of any of the above are also included within the scope of computer readable media.
As used herein, unless otherwise specified the use of the ordinal terms "first," "second," "third," etc., to describe a general object merely denote different instances of like objects, and are not intended to imply that the objects so described must have a given order, either temporally, spatially, in ranking, or in any other manner.
While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of the above description, will appreciate that other embodiments are contemplated within the scope of the invention as described herein. Furthermore, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the appended claims. The disclosure of the present invention is intended to be illustrative, but not limiting, of the scope of the invention, which is defined by the appended claims.

Claims (1)

1. The photovoltaic system fault alarming method based on the BP neural network is characterized in that training data are screened out by establishing the BP neural network of the distributed photovoltaic system, and then alarming is carried out on abnormal output characteristics in the operation process of the distributed photovoltaic system in an area;
the method specifically comprises the following steps:
s1, acquiring output data of a distributed photovoltaic system in a regional photovoltaic system and meteorological data of the regional photovoltaic system;
s2, carrying out normalization processing on output data and meteorological data of the distributed photovoltaic system;
s3, based on the output data and the meteorological data of the distributed photovoltaic system after normalization processing, establishing a BP neural network model input matrix x of the distributed photovoltaic system t-i * Inputting BP neural network model into matrix x t-i * Inputting an implicit layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area;
s4, normalizing output data of the evaluated distributed photovoltaic system, and inputting matrix x under the same time stamp t-i * And output vector Y t * Performing correlation judgment to obtain filtered output data of the distributed photovoltaic system
Figure QLYQS_1
And Meteorological data->
Figure QLYQS_2
S5, outputting the filtered output data of the distributed photovoltaic system
Figure QLYQS_3
And Meteorological data->
Figure QLYQS_4
As the input training parameters of the BP neural network model of the current area photovoltaic system, training the BP neural network model of the current area photovoltaic system, and outputting a trained neural network system;
s6, applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model, and obtaining output parametersTest vector
Figure QLYQS_5
Calculating an output reference vector +.>
Figure QLYQS_6
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure QLYQS_7
The correlation is used for judging whether the evaluated distributed photovoltaic power station has suspicious faults or not according to the correlation;
201 Based on the installed capacity of n distributed photovoltaic systems, normalizing the output data of the distributed photovoltaic systems, taking the normalized parameters as input training parameters of the BP neural network model, wherein the total number of all the input parameters of the BP neural network model is n, and the normalized output data matrix at the moment t is
Figure QLYQS_8
202 When available meteorological data exists in the environmental range of the regional photovoltaic system, carrying out normalization processing on the meteorological data of the regional photovoltaic system to form a normalized meteorological matrix W at the moment t t-j * The normalized output data matrix is taken as
Figure QLYQS_9
And weather matrix W t-j * As the input training parameters of the BP neural network model, n+m input parameters are used for the BP neural network model;
301 Building an input matrix x of BP neural network model of regional photovoltaic system at t moment t-i * Is of formula (1):
Figure QLYQS_10
302 According to the input matrix x) t-i * Dimension and accuracy of prediction resultsSolving, namely determining the number of hidden layers of the BP neural network model;
303 Using M-layer unipolar Sigmoid function as an hidden layer activation function f (v) of the BP neural network model to obtain the BP neural network model of the photovoltaic system in the current area;
the hidden layer activation function f (v) of the BP neural network model is expressed as formula (2):
Figure QLYQS_11
wherein v is an input parameter, which is the multi-layer input of the hidden layer of the BP neural network model, and the initial value of the input parameter is an input matrix x t-i * The first layer is the input matrix x t-i * The input v of the second layer is the result after the calculation of the first layer, the v domain is (- ≡, ++ infinity a) is provided;
the step S4 specifically comprises the following steps:
401 Normalizing the output data of the evaluated distributed photovoltaic system according to the installed capacity of the evaluated distributed photovoltaic system in the region, wherein the output vector of the normalized output data of the evaluated distributed photovoltaic system is Y t * ；
402 For input matrix x at time t) t-i * And output vector Y t * Screening, and judging the relevance of the input and output functions by using the Pearson correlation coefficient;
Figure QLYQS_12
wherein: c is sample x t-i * And sample Y t * Correlation coefficients; cov (x) t-i * ,Y t * ) For sample x t-i * And sample Y t * Covariance of correlation coefficient, var (x t-i * )、Var(Y t * ) Fractional sample x t-i * And sample Y t * Is a variance of (2);
403 Set up)Defining a first correlation coefficient threshold as an input matrix x with filtered correlation coefficients c less than the first correlation coefficient threshold t-i * And the output vector is Y t * Obtaining filtered output data of the distributed photovoltaic system
Figure QLYQS_13
And Meteorological data->
Figure QLYQS_14
Step S1, specifically comprising the following steps:
101 Determining available distributed photovoltaic systems in the regional environment range of the regional photovoltaic system, wherein the regional photovoltaic system comprises more than 1 other distributed photovoltaic systems besides the evaluated photovoltaic power generation system, and acquiring output data of the distributed photovoltaic systems;
102 For the photovoltaic system being evaluated in the area and other n distributed photovoltaic systems, n>=1, respectively obtaining output data at t time, wherein the output data of the estimated distributed photovoltaic system at t time is recorded as a vector Y t The output data of n distributed photovoltaic systems form an output data matrix P t-i ，i＝1,2，…，n；P t-i The output data of the ith distributed photovoltaic system at the t moment is represented;
103 Acquiring meteorological data of a regional photovoltaic system at the moment t to form a meteorological data vector W at the moment t t-j ，j＝1,2，3，4…，m，W t-j Representing the j-th meteorological data vector at the t moment;
the step S5 specifically comprises the following steps: filtered output data of distributed photovoltaic system at t moment
Figure QLYQS_15
And weather data
Figure QLYQS_16
As the input training parameters of the BP neural network model of the current area photovoltaic system; hidden layer activation function f (v) based on BP neural network model * ) For a pair ofTraining the neural network, and optimizing the neural network; using the filtered input matrix, output matrix and hidden layer activation function f (v * ) Training the neural network to form a multi-layer neural network to obtain a trained neural network system;
Figure QLYQS_17
v * is the filtered input parameter of BP neural network model, v * The initial value of (1) is the filtered output data of the distributed photovoltaic system
Figure QLYQS_18
And Meteorological data->
Figure QLYQS_19
The step S6 specifically comprises the following steps:
601 Applying the trained neural network system to fault warning of the regional photovoltaic system, monitoring all input parameter vectors of the BP neural network model in real time, and obtaining corresponding output reference vectors
Figure QLYQS_20
Outputting the reference vector
Figure QLYQS_21
Is obtained by inputting an input vector into a multi-layer neural network formed by the formula (6);
602 Calculating an output reference vector using Pearson correlation coefficients
Figure QLYQS_22
The actual output of the photovoltaic power station to be evaluated under the corresponding time stamp>
Figure QLYQS_23
Correlation, noted cY:
Figure QLYQS_24
603 Setting a second correlation coefficient threshold, when c y When the correlation coefficient is smaller than the second correlation coefficient threshold value of 0.95, the estimated distributed photovoltaic power station is considered to have suspicious faults, and alarming is carried out;
the first correlation coefficient threshold is 0.95;
the second correlation coefficient threshold is 0.95.
CN202210712863.6A 2022-06-22 2022-06-22 BP neural network-based photovoltaic system fault alarm method Active CN115587642B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210712863.6A CN115587642B (en) 2022-06-22 2022-06-22 BP neural network-based photovoltaic system fault alarm method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210712863.6A CN115587642B (en) 2022-06-22 2022-06-22 BP neural network-based photovoltaic system fault alarm method

Publications (2)

Publication Number Publication Date
CN115587642A CN115587642A (en) 2023-01-10
CN115587642B true CN115587642B (en) 2023-07-11

Family

ID=84771597

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210712863.6A Active CN115587642B (en) 2022-06-22 2022-06-22 BP neural network-based photovoltaic system fault alarm method

Country Status (1)

Country Link
CN (1) CN115587642B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106570593A (en) * 2016-11-10 2017-04-19 甘肃省电力公司风电技术中心 Photovoltaic power station output data repairing method based on weather information
CN109150100A (en) * 2018-09-13 2019-01-04 国网电子商务有限公司 Fault detection method, device, equipment and the storage medium of photovoltaic plant
CN110751326A (en) * 2019-10-17 2020-02-04 江苏远致能源科技有限公司 Photovoltaic day-ahead power prediction method and device and storage medium
CN111177127A (en) * 2019-12-09 2020-05-19 国网甘肃省电力公司电力科学研究院 Photovoltaic power station output data restoration method based on BP neural network
CN111680820A (en) * 2020-05-08 2020-09-18 国网河南省电力公司电力科学研究院 Distributed photovoltaic power station fault diagnosis method and device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106570593A (en) * 2016-11-10 2017-04-19 甘肃省电力公司风电技术中心 Photovoltaic power station output data repairing method based on weather information
CN109150100A (en) * 2018-09-13 2019-01-04 国网电子商务有限公司 Fault detection method, device, equipment and the storage medium of photovoltaic plant
CN110751326A (en) * 2019-10-17 2020-02-04 江苏远致能源科技有限公司 Photovoltaic day-ahead power prediction method and device and storage medium
CN111177127A (en) * 2019-12-09 2020-05-19 国网甘肃省电力公司电力科学研究院 Photovoltaic power station output data restoration method based on BP neural network
CN111680820A (en) * 2020-05-08 2020-09-18 国网河南省电力公司电力科学研究院 Distributed photovoltaic power station fault diagnosis method and device

Also Published As

Publication number Publication date
CN115587642A (en) 2023-01-10

Similar Documents

Publication Publication Date Title
Cheng et al. Machine health monitoring using adaptive kernel spectral clustering and deep long short-term memory recurrent neural networks
CN110766059A (en) Transformer fault prediction method, device and equipment
WO2019080367A1 (en) Method for evaluating health status of mechanical device
Kankanala et al. AdaBoost $^{+} $: An Ensemble Learning Approach for Estimating Weather-Related Outages in Distribution Systems
CN109581871B (en) Industrial control system intrusion detection method of immune countermeasure sample
CN108197648A (en) A kind of Fault Diagnosis Method of Hydro-generating Unit and system based on LSTM deep learning models
CN109163913A (en) A kind of Diagnosis method of automobile faults and relevant device
Alrifaey et al. Hybrid deep learning model for fault detection and classification of grid-connected photovoltaic system
CN110334865B (en) Power equipment fault rate prediction method and system based on convolutional neural network
CN110006552B (en) Method for detecting abnormal temperature of equipment
Goharnejad et al. Vulnerability assessment of southern coastal areas of Iran to sea level rise: evaluation of climate change impact
Wang et al. Study on wavelet neural network based anomaly detection in ocean observing data series
CN110163075A (en) A kind of multi-information fusion method for diagnosing faults based on Weight Training
CN114355240A (en) Power distribution network ground fault diagnosis method and device
Morteza et al. A dagging‐based deep learning framework for transmission line flexibility assessment
Daş Forecasting the energy demand of Turkey with a NN based on an improved Particle Swarm Optimization
Hammoudi et al. Review on maintenance of photovoltaic systems based on deep learning and internet of things
Wang et al. Uncertainty quantification of bearing remaining useful life based on convolutional neural network
CN113758652B (en) Oil leakage detection method and device for converter transformer, computer equipment and storage medium
CN115587642B (en) BP neural network-based photovoltaic system fault alarm method
Takianngam et al. Discrete Wavelet Transform and One-Class Support Vector Machines for anomaly detection in wireless sensor networks
Ultsch et al. Self-organizing feature maps predicting sea levels
CN112419243B (en) Power distribution room equipment fault identification method based on infrared image analysis
CN117310389A (en) CNN-based power distribution network fault positioning method and device and related equipment
Ahmadi et al. Hybrid model for weather forecasting using ensemble of neural networks and mutual information

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