WO2023087569A1

WO2023087569A1 - Photovoltaic string communication abnormality identification method and system based on xgboost

Info

Publication number: WO2023087569A1
Application number: PCT/CN2022/078431
Authority: WO
Inventors: 王振荣; 曾谁飞; 祝金涛; 王青天; 赵鹏程; 王�华; 任鑫; 赵斌; 李靖
Original assignee: 中国华能集团清洁能源技术研究院有限公司
Priority date: 2021-11-17
Filing date: 2022-02-28
Publication date: 2023-05-25
Also published as: CN114298084A

Abstract

A photovoltaic string communication abnormality identification method and system based on XGBoost. The method comprises: acquiring current values of all photovoltaic strings in a photovoltaic power station within a time period to be subjected to detection, current values of all the photovoltaic strings in the photovoltaic power station within a preset time period prior to said time period, and corresponding communication abnormality label data, and normalizing the current values; training an XGBoost model on the basis of generated data of a VaDE model, so as to obtain a trained XGBoost model; and inputting, into the trained XGBoost model, normalized current values of all the photovoltaic strings in the photovoltaic power station within said time period, and identifying a photovoltaic string, which has a current value communication abnormality, in the photovoltaic power station within said time period.

Description

An XGBoost-based photovoltaic string communication abnormality identification method and system

Cross References to Related Applications

This application claims priority to Chinese Patent Application No. 202111362748.2 filed in China on November 17, 2021, the entire contents of which are incorporated herein by reference.

technical field

The present disclosure relates to the field of abnormal identification of photovoltaic string communication, and in particular to a method and system for identifying abnormal communication of photovoltaic strings based on XGBoost.

Background technique

With the rapid development of science and technology, photovoltaic power generation technology has been widely used at home and abroad. Its application forms are various and its application places are widely distributed. Building integration, photovoltaic street lights, etc. In practical applications, due to the large number of photovoltaic strings in the photovoltaic power station, there is no way to find and deal with the abnormal communication of photovoltaic strings in a timely manner. False detections and missed detections occur in the string abnormality diagnosis model, which reduces the reliability of correlation analysis. Therefore, it is urgent to propose a method and system that can accurately identify communication abnormalities in photovoltaic strings in photovoltaic power plants.

Contents of the invention

The present disclosure provides an XGBoost-based photovoltaic string communication abnormality identification method and system to at least solve the technical problem of inaccurate identification of photovoltaic string communication abnormality in a photovoltaic power station in the related art.

The embodiment of the first aspect of the present disclosure proposes an XGBoost-based photovoltaic string communication abnormality identification method, the method including:

Obtain the current value of each photovoltaic string in the photovoltaic power station during the period to be tested and the current value of all photovoltaic strings in the photovoltaic power station in the preset period before the period to be tested and the corresponding communication abnormal tag data, and normalize the current value ;

Record the normalized current values of all photovoltaic strings and corresponding abnormal communication label data in the photovoltaic power station within the preset period before the date to be tested as sample set A, and record the current values of the faulty photovoltaic strings in A and the corresponding The sample set composed of abnormal communication labels is denoted as B;

Input the current value of the faulty photovoltaic string in the sample set B into the pre-trained VaDE model to obtain a sample set of abnormal communication of the current value of the faulty photovoltaic string generated by the VaDE model, and denote the sample set as C, Merge sample set A and sample set C to obtain a combined sample set D;

Use the sample set D to train the XGBoost model to obtain the trained XGBoost model;

Input the normalized value of the current of each photovoltaic string in the photovoltaic power station during the period to be tested into the trained XGBoost model, and identify the photovoltaic strings with abnormal current value communication in the photovoltaic power station during the period to be tested.

The embodiment of the second aspect of the present disclosure proposes an XGBoost-based photovoltaic string communication abnormality identification system, the system includes:

The obtaining module is used to obtain the current value of each photovoltaic string in the photovoltaic power station during the period to be tested and the current value of all photovoltaic strings in the photovoltaic power station during the preset period before the period to be tested and the corresponding abnormal communication tag data, and send the Current value normalization;

The first sample module is used to record the normalized current values of all photovoltaic strings of the photovoltaic power plant and the corresponding abnormal communication tag data within the preset period before the date to be tested as a sample set A, and record the faulty photovoltaics in A The sample set composed of the current value of the string and the corresponding abnormal communication label is denoted as B;

The second sample module is used to input the current value of the faulty photovoltaic string in the sample set B into the pre-trained VaDE model, obtain the sample set of abnormal communication of the current value of the faulty photovoltaic string generated by the VaDE model, and send the The above sample set is denoted as C, and the combined sample set A and sample set C are combined to obtain the combined sample set D;

The training module is used to use the sample set D to train the XGBoost model to obtain the trained XGBoost model;

The identification module is used to input the normalized value of the current of each photovoltaic string in the photovoltaic power station into the trained XGBoost model during the period to be tested, and identify the photovoltaic strings with abnormal current value communication in the photovoltaic power station during the period to be tested .

The embodiment of the third aspect of the present disclosure proposes an electronic device, including: a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the program, the implementation of the first aspect above is realized. The XGBoost-based photovoltaic string communication abnormality identification method described in the example.

The embodiment of the fourth aspect of the present disclosure proposes a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the XGBoost-based photovoltaic string communication abnormality identification method described in the embodiment of the first aspect above is implemented. .

The embodiment of the fifth aspect of the present disclosure proposes a computer program product, the computer program product includes computer program code, when the computer program code is run on the computer, to execute the XGBoost-based An abnormal identification method for photovoltaic string communication.

The embodiment of the sixth aspect of the present disclosure proposes a computer program, the computer program includes computer program code, when the computer program code is run on the computer, so that the computer executes the XGBoost-based A photovoltaic string communication abnormal identification method.

The technical solutions provided by the embodiments of the present disclosure bring at least the following beneficial effects:

The present disclosure provides an XGBoost-based photovoltaic string communication abnormality identification method and system, the method comprising: obtaining the current value of each photovoltaic string in the photovoltaic power station within the period to be tested and the photovoltaic power station within the preset period before the period to be tested The current values of all photovoltaic strings and the corresponding abnormal communication tag data, and normalize the current values; the normalized current values and corresponding The data of the abnormal communication tag in the sample set is recorded as sample set A, and the sample set composed of the current value of the faulty photovoltaic string in A and the corresponding abnormal communication tag is recorded as B; the current value of the faulty photovoltaic string in the sample set B is Input the pre-trained VaDE model to obtain the sample set of faulty photovoltaic string current value communication abnormality generated by the VaDE model, and record the sample set as C, and combine sample set A and sample set C to obtain the combined sample set D. Use the sample set D to train the XGBoost model to obtain the trained XGBoost model; input the normalized value of the current of each photovoltaic string in the photovoltaic power plant into the trained XGBoost model during the period to be tested, and identify the PV strings with abnormal current value communication in the PV power plant during the measurement period. The technical solution provided by this disclosure trains the XGBoost model based on the generated data of the VaDE model to obtain a trained XGBoost model, and uses the trained model to identify whether the current communication of each string in the photovoltaic power station is abnormal, which can improve Accuracy of XGBoost model in identifying abnormal PV strings in PV power plants.

Additional aspects and advantages of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.

Description of drawings

The above and/or additional aspects and advantages of the present disclosure will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, wherein:

Fig. 1 is a flow chart of an XGBoost-based photovoltaic string communication abnormality identification method provided according to an embodiment of the present disclosure;

Fig. 2 is a structural diagram of an XGBoost-based photovoltaic string communication abnormality identification system according to the present disclosure.

Detailed ways

Embodiments of the present disclosure are described in detail below, examples of which are illustrated in the drawings, in which the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary and are intended to explain the present disclosure and should not be construed as limiting the present disclosure.

An XGBoost-based photovoltaic string communication abnormality identification method and system proposed in the present disclosure, the method includes: obtaining the current value of each photovoltaic string in the photovoltaic power station within the period to be tested and the photovoltaic power station within the preset period before the period to be tested The current values of all photovoltaic strings and the corresponding abnormal communication tag data, and normalize the current values; the normalized current values and corresponding The data of the abnormal communication tag in the sample set is recorded as sample set A, and the sample set composed of the current value of the faulty photovoltaic string in A and the corresponding abnormal communication tag is recorded as B; the current value of the faulty photovoltaic string in the sample set B is Input the pre-trained VaDE model to obtain the sample set of faulty photovoltaic string current value communication abnormality generated by the VaDE model, and record the sample set as C, and combine sample set A and sample set C to obtain the combined sample set D. Use the sample set D to train the XGBoost model to obtain the trained XGBoost model; input the normalized value of the current of each photovoltaic string in the photovoltaic power plant into the trained XGBoost model during the period to be tested, and identify the PV strings with abnormal current value communication in the PV power plant during the measurement period. The technical solution provided by this disclosure trains the XGBoost model based on the generated data of the VaDE model to obtain a trained XGBoost model, and uses the trained model to identify whether the current communication of each string in the photovoltaic power station is abnormal, which can improve Accuracy of XGBoost model in identifying abnormal PV strings in PV power plants.

Example 1

Fig. 1 is a flow chart of an XGBoost-based photovoltaic string communication abnormality identification method provided in this embodiment. As shown in Fig. 1, the method includes:

Step 1: Obtain the current value of each photovoltaic string in the photovoltaic power station during the period to be tested and the current value of all photovoltaic strings in the photovoltaic power station in the preset period before the period to be tested and the corresponding communication abnormal tag data, and store the current value Normalized;

In one example, the current value of each photovoltaic string in the photovoltaic power plant on the day of the detection date and the current value of all photovoltaic strings in the photovoltaic power plant within 30 days before the detection date and the corresponding abnormal communication tag data are obtained, wherein, within the 30 days Contains PV strings with abnormal current value communication.

In an example, the calculation formula for normalizing the current value of the photovoltaic string m at the jth moment in the period to be tested is as follows:

In the formula,

is the normalized current value of the photovoltaic string m at the jth moment in the period to be tested, x _m,j is the current value of the photovoltaic string m at the jth moment in the period to be tested, x _m,k is the current value of the photovoltaic string m to be tested The current value of the photovoltaic string m at the kth moment in the measurement period, h is the set of all moments in the measurement period of the acquired data, and k is any moment in h.

Step 2: Record the normalized current values of all photovoltaic strings and the corresponding communication abnormal tag data in the preset period before the date to be tested as sample set A, and record the current value of the faulty photovoltaic strings in A And the sample set composed of corresponding abnormal communication labels is denoted as B;

Step 3: Input the current value of the faulty photovoltaic string in the sample set B into the pre-trained VaDE model, obtain the sample set of abnormal communication of the current value of the faulty photovoltaic string generated by the VaDE model, and record the sample set For C, combine sample set A and sample set C to obtain the combined sample set D;

It should be noted that the pre-trained VaDE model includes: a first neural network layer, a sampling layer and a second neural network layer;

The training procedure of described VaDE model comprises:

Obtain the current value data of the faulty photovoltaic string in the photovoltaic power plant within the normalized historical period;

The data input that obtains is in initial VaDE model first neural network layer, sampling layer and the second neural network layer, with marginal likelihood lower bound as the loss function of model, utilize adaptive matrix estimation Adam optimization algorithm to carry out described model Training to get the trained VaDE model.

Step 4: Use the sample set D to train the XGBoost model to obtain the trained XGBoost model;

Step 5: Input the normalized value of the current of each photovoltaic string in the photovoltaic power station during the period to be tested into the trained XGBoost model, and identify the photovoltaic strings with abnormal current value communication in the photovoltaic power station during the period to be tested.

In the disclosed embodiment, the XGBoost model is trained using the sample set D to obtain a trained XGBoost model, including:

Obtain a sample set D, where D={(xi _, y _i )}(|D|=n, _xi ∈ R ^m , y _i ∈ R), and _xi is the index value of the i-th sample at each time of the day , y _i is the tag value of whether the indicator has abnormal communication in the i-th sample in one day, y _i ∈ {0, 1}, 0 indicates normal communication, 1 indicates abnormal communication, n is the number of samples, and m is the number of sample features;

Based on the sample set D, construct a gradient boosting tree, use the CART regression tree as the subtree model of the model, increase the CART regression tree iteratively, and merge all the CART regression trees together to obtain the trained XGBoost model.

It should be noted that when adding the CART regression tree, the greedy algorithm is used to first find the segmentation point with the largest profit based on each sample feature, and then find the feature with the largest profit based on all sample features. In the double-layer cycle of features and segmentation points Pick out the split point with the largest gain, and split and increase the CART regression tree based on the split point.

Specifically, define the XGBoost model to predict y in D. First, construct a gradient boosting tree: accumulate trees based on the sample set D, build a tree each time, and use the CART regression tree as the subtree model of the model. Secondly, construct The loss function of the regularization item at the tth iteration:

l is the real value y _i and predicted value specified inside the algorithm

The loss function of is a differentiable convex function, f _t ( _xi ) is the weight of the i-th sample at the leaf node in the t-th iteration, Ω(f _t ) is a regular term, and the leaf node of the t-th iteration tree The number of nodes and the weight of leaf nodes are represented, and then, for L ^(t) in

The second-order Taylor expansion is performed at the place, and the weight is determined by minimizing the loss function. Finally, all the CART regression trees are merged together to obtain the trained XGBoost model.

It should be noted that the index value of the photovoltaic string obtained here is the current value, and other index values of the photovoltaic string can also be obtained, which is not limited here.

To sum up, the embodiment of the present disclosure provides an XGBoost-based photovoltaic string communication anomaly identification method, which trains the XGBoost model based on the generated data of the VaDE model to obtain a trained XGBoost model, and uses the trained model to Identifying whether the current communication of each string in the photovoltaic power station is abnormal can improve the accuracy of the XGBoost model in identifying abnormal photovoltaic strings in the photovoltaic power station.

Example 2

Fig. 2 is a structural diagram of an XGBoost-based photovoltaic string communication abnormality identification system provided by an embodiment of the present disclosure. As shown in Fig. 2 , the system includes:

In the embodiment of the present disclosure, the normalized calculation formula of the current value of the photovoltaic string m at the jth moment in the period to be tested is as follows:

In the formula,

In an embodiment of the present disclosure, the pre-trained VaDE model includes: a first neural network layer, a sampling layer and a second neural network layer;

The training procedure of described VaDE model comprises:

In an embodiment of the present disclosure, the training module includes:

The acquisition unit is used to acquire the sample set D, where D={( _xi , y _i )}(|D|=n, _xi ∈ R ^m , y _i ∈ R), and _xi is the i-th sample day The index value at each moment, y _i is the label value of whether the index has abnormal communication in the i-th sample in one day, y _i ∈ {0, 1}, 0 means normal communication, 1 means abnormal communication, n is the number of samples, m is number of sample features;

The training unit is used to construct a gradient boosting tree based on the sample set D, using the CART regression tree as the subtree model of the model, increasing the CART regression tree through iteration, and merging all the CART regression trees together to obtain the trained XGBoost model .

To sum up, the XGBoost-based photovoltaic string communication anomaly recognition system provided by the present disclosure trains the XGBoost model based on the generated data of the VaDE model to obtain a trained XGBoost model, and uses the trained model to identify photovoltaic strings. Identifying whether the current communication of each string in the power station is abnormal can improve the accuracy of the XGBoost model in identifying abnormal photovoltaic strings in the photovoltaic power station.

An embodiment of the present disclosure also proposes an electronic device, including: a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the program, the method based on XGBoost's photovoltaic string communication anomaly identification method.

The embodiment of the present disclosure also proposes a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the XGBoost-based photovoltaic string communication abnormality identification method described in Embodiment 1 is implemented.

The embodiment of the present disclosure also proposes a computer program product, the computer program product includes computer program code, when the computer program code is run on the computer, to execute the XGBoost-based photovoltaic string communication described in Embodiment 1 Exception identification method.

The embodiment of the present disclosure also proposes a computer program, the computer program includes computer program code, when the computer program code is run on the computer, so that the computer executes the XGBoost-based photovoltaic string communication exception described in Embodiment 1 recognition methods.

In the description of this specification, descriptions referring to the terms "one embodiment", "some embodiments", "example", "specific examples", or "some examples" mean that specific features described in connection with the embodiment or example , structure, material or characteristic is included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the described specific features, structures, materials or characteristics may be combined in any suitable manner in any one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples and features of different embodiments or examples described in this specification without conflicting with each other.

Any process or method descriptions in flowcharts or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing custom logical functions or steps of a process , and the scope of preferred embodiments of the present disclosure includes additional implementations in which functions may be performed out of the order shown or discussed, including substantially concurrently or in reverse order depending on the functions involved, which shall It is understood by those skilled in the art to which the embodiments of the present disclosure pertain.

Although the embodiments of the present disclosure have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limitations on the present disclosure, and those skilled in the art can understand the above-mentioned embodiments within the scope of the present disclosure. The embodiments are subject to changes, modifications, substitutions and variations.

Claims

An XGBoost-based photovoltaic string communication abnormal identification method, characterized in that the method comprises:

Obtain the current value of each photovoltaic string in the photovoltaic power station during the period to be tested and the current value of all photovoltaic strings in the photovoltaic power station in the preset period before the period to be tested and the corresponding communication abnormal tag data, and normalize the current value ;

Record the normalized current values of all photovoltaic strings and corresponding abnormal communication label data in the photovoltaic power station within the preset period before the date to be tested as sample set A, and record the current values of the faulty photovoltaic strings in A and the corresponding The sample set composed of abnormal communication labels is denoted as B;

Input the current value of the faulty photovoltaic string in the sample set B into the pre-trained VaDE model to obtain a sample set of abnormal communication of the current value of the faulty photovoltaic string generated by the VaDE model, and denote the sample set as C, Merge sample set A and sample set C to obtain a combined sample set D;

Use the sample set D to train the XGBoost model to obtain the trained XGBoost model;

Input the normalized value of the current of each photovoltaic string in the photovoltaic power station during the period to be tested into the trained XGBoost model, and identify the photovoltaic strings with abnormal current value communication in the photovoltaic power station during the period to be tested.
The method according to claim 1, characterized in that, the calculation formula for normalizing the current value of the photovoltaic string m at the jth moment in the period to be measured is as follows:

In the formula,
is the normalized current value of the photovoltaic string m at the jth moment in the period to be tested, x m,j is the current value of the photovoltaic string m at the jth moment in the period to be tested, x m,k is the current value of the photovoltaic string m to be tested The current value of the photovoltaic string m at the kth moment in the measurement period, h is the set of all moments in the measurement period of the acquired data, and k is any moment in h.
The method according to claim 1 or 2, wherein said pre-trained VaDE model comprises: a first neural network layer, a sampling layer and a second neural network layer;

The training procedure of described VaDE model comprises:

Obtain the current value data of the faulty photovoltaic string in the photovoltaic power plant within the normalized historical period;

The data input that obtains is in initial VaDE model first neural network layer, sampling layer and the second neural network layer, with marginal likelihood lower bound as the loss function of model, utilize adaptive matrix estimation Adam optimization algorithm to carry out described model Training to get the trained VaDE model.
The method according to any one of claims 1 to 3, wherein the training of the XGBoost model by using the sample set D to obtain the trained XGBoost model includes:

Obtain a sample set D, where D={(xi , y i )}(|D|=n, xi ∈ R m , y i ∈ R), and xi is the index value of the i-th sample at each time of the day , y i is the tag value of whether the indicator has abnormal communication in the i-th sample in one day, y i ∈ {0, 1}, 0 indicates normal communication, 1 indicates abnormal communication, n is the number of samples, and m is the number of sample features;

Based on the sample set D, construct a gradient boosting tree, use the CART regression tree as the subtree model of the model, increase the CART regression tree iteratively, and merge all the CART regression trees together to obtain the trained XGBoost model.
The method according to claim 4, wherein when increasing the CART regression tree, a greedy algorithm is first used to find the segmentation point with the largest profit based on each sample feature, and then to find the feature with the largest profit based on all sample features. In the double-layer cycle of the split point and the split point, the split point with the largest gain is selected, and the CART regression tree is increased by splitting based on the split point.
An XGBoost-based photovoltaic string communication abnormality identification system, characterized in that the system includes:

The obtaining module is used to obtain the current value of each photovoltaic string in the photovoltaic power station during the period to be tested and the current value of all photovoltaic strings in the photovoltaic power station during the preset period before the period to be tested and the corresponding abnormal communication tag data, and send the Current value normalization;

The first sample module is used to record the normalized current values of all photovoltaic strings of the photovoltaic power plant and the corresponding abnormal communication tag data within the preset period before the date to be tested as a sample set A, and record the faulty photovoltaics in A The sample set composed of the current value of the string and the corresponding abnormal communication label is denoted as B;

The second sample module is used to input the current value of the faulty photovoltaic string in the sample set B into the pre-trained VaDE model, obtain the sample set of abnormal communication of the current value of the faulty photovoltaic string generated by the VaDE model, and send the The above sample set is denoted as C, and the combined sample set A and sample set C are combined to obtain the combined sample set D;

The training module is used to use the sample set D to train the XGBoost model to obtain the trained XGBoost model;

The identification module is used to input the normalized value of the current of each photovoltaic string in the photovoltaic power station into the trained XGBoost model during the period to be tested, and identify the photovoltaic strings with abnormal current value communication in the photovoltaic power station during the period to be tested .
The system according to claim 6, wherein the calculation formula for normalizing the current value of the photovoltaic string m at the jth moment in the period to be measured is as follows:

In the formula,
is the normalized current value of the photovoltaic string m at the jth moment in the period to be tested, x m,j is the current value of the photovoltaic string m at the jth moment in the period to be tested, x m,k is the current value of the photovoltaic string m to be tested The current value of the photovoltaic string m at the kth moment in the measurement period, h is the set of all moments in the measurement period of the acquired data, and k is any moment in h.
system as claimed in claim 6 or 7, is characterized in that, described pre-trained VaDE model comprises: the first neural network layer, sampling layer and the second neural network layer;

The training procedure of described VaDE model comprises:

Obtain the current value data of the faulty photovoltaic string in the photovoltaic power plant within the normalized historical period;

The data input that obtains is in initial VaDE model first neural network layer, sampling layer and the second neural network layer, with marginal likelihood lower bound as the loss function of model, utilize adaptive matrix estimation Adam optimization algorithm to carry out described model Training to get the trained VaDE model.
The system according to any one of claims 6 to 8, wherein the training module includes:

The acquisition unit is used to acquire the sample set D, where D={( xi , y i )}(|D|=n, xi ∈ R m , y i ∈ R), and xi is the i-th sample day The index value at each moment, y i is the label value of whether the index has abnormal communication in the i-th sample in one day, y i ∈ {0, 1}, 0 means normal communication, 1 means abnormal communication, n is the number of samples, m is number of sample features;

The training unit is used to construct a gradient boosting tree based on the sample set D, using the CART regression tree as the subtree model of the model, increasing the CART regression tree through iteration, and merging all the CART regression trees together to obtain the trained XGBoost model .
The system according to claim 9, wherein when increasing the CART regression tree, a greedy algorithm is used to first find the segmentation point with the largest profit based on each sample feature, and then find the feature with the largest profit based on all sample features. In the double-layer cycle of the split point and the split point, the split point with the largest gain is selected, and the CART regression tree is increased by splitting based on the split point.
An electronic device, characterized in that it comprises:

A memory, a processor, and a computer program stored in the memory and operable on the processor, wherein the processor implements the method according to any one of claims 1-5 when executing the program.
A computer-readable storage medium, on which a computer program is stored, wherein, when the program is executed by a processor, the method according to any one of claims 1-5 is realized.
A computer program product, characterized in that the computer program product includes computer program code, and when the computer program code is run on a computer, the method according to any one of claims 1-5 is executed.
A computer program, characterized in that the computer program includes computer program code, and when the computer program code is run on a computer, the computer executes the method according to any one of claims 1-5.