CN117996838A - Distributed photovoltaic recognition device based on non-invasive load monitoring - Google Patents
Distributed photovoltaic recognition device based on non-invasive load monitoring Download PDFInfo
- Publication number
- CN117996838A CN117996838A CN202410390116.4A CN202410390116A CN117996838A CN 117996838 A CN117996838 A CN 117996838A CN 202410390116 A CN202410390116 A CN 202410390116A CN 117996838 A CN117996838 A CN 117996838A
- Authority
- CN
- China
- Prior art keywords
- window
- distributed photovoltaic
- data
- power generation
- photovoltaic power
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 24
- 238000010248 power generation Methods 0.000 claims abstract description 74
- 238000005070 sampling Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 22
- 238000012706 support-vector machine Methods 0.000 claims abstract description 17
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 230000007613 environmental effect Effects 0.000 claims abstract description 9
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 14
- 238000000034 method Methods 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 4
- 238000013500 data storage Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000012545 processing Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
- Photovoltaic Devices (AREA)
Abstract
The invention provides a distributed photovoltaic identification device based on non-invasive load monitoring, and relates to the technical field of smart grids. Firstly, an acquisition module acquires data, and a preprocessing module performs feature selection and sliding window overlapping sampling preprocessing on the acquired data; the training module builds a distributed photovoltaic power generation recognition model based on a support vector machine by using window data after window sampling pretreatment, and trains the acquired data; the identification module inputs the data actually collected by the intelligent ammeter into the trained model for calculation, and identifies the distributed photovoltaic power generation system in actual operation. Aiming at the distributed photovoltaic power generation system, the invention adopts a non-invasive load monitoring technology, utilizes minute-level acquisition data of an intelligent electric energy meter, environmental data and the like, and decomposes the distributed photovoltaic power generation from a main feeder line of a load by using a sliding window with low complexity and a support vector machine technology.
Description
Technical Field
The invention relates to the technical field of smart grids, in particular to a distributed photovoltaic identification device based on non-invasive load monitoring.
Background
The application of new energy represented by the distributed photovoltaic power generation system also brings great burden to the power system, the 'considerable, measurable and controllable' requirements of the photovoltaic system cannot be really met, the randomness and the volatility of the distributed photovoltaic power generation bring difficulty to the power grid absorption and control, meanwhile, the problems of voltage violations, frequency fluctuation, overload, reverse tide and the like which possibly occur danger also have serious influence on the low-voltage power distribution network, and how to monitor and identify the distributed photovoltaic power generation system so as to achieve the aim of controllability becomes a research key point, and in order to achieve the aim, the key point is to acquire domestic electrical load information and conduct intelligent analysis.
As an important load monitoring means, non-invasive load monitoring is a main direction and means for load monitoring, which can grasp and intelligently analyze the power consumption load information without entering the inside of a resident household. However, the non-invasive load monitoring technology at the present stage is mainly focused on identifying conventional loads in the aspect of clients, such as electric equipment, like household appliances, and compared with the conventional household appliances, the distributed photovoltaic power generation system has the characteristics of randomness, reverse power and the like, and is still in a starting stage for monitoring the distributed photovoltaic, and an effective non-invasive monitoring identification system aiming at the distributed photovoltaic power generation system is lacking.
At present, the conventional non-invasive load monitoring method needs to be realized by modifying or adding a monitoring device, but still has some problems such as more investment, higher cost, wide related range, great popularization difficulty and the like. In addition, the traditional non-invasive load monitoring algorithm needs high-frequency data acquisition, and the accuracy and timeliness of intelligent analysis cannot meet the actual requirements. For the distributed photovoltaic power generation system of stock, on one hand, sampling equipment is additionally added to meet the high-frequency data, so that the cost of load identification is increased. On the other hand, the high-frequency sampling brings about improvement of algorithm complexity, and the distributed photovoltaic power generation system is operated for a long time and frequently fluctuates to increase identification complexity.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a distributed photovoltaic identification device based on non-invasive load monitoring, which adopts a non-invasive load monitoring technology for a distributed photovoltaic power generation system, utilizes minute-level acquisition data of an intelligent electric energy meter, environmental data and the like, and utilizes a low-complexity sliding window and support vector machine technology to decompose distributed photovoltaic power generation from a main feeder line of a load.
In order to solve the technical problems, the invention adopts the following technical scheme:
A distributed photovoltaic identification device based on non-invasive load monitoring, comprising a memory storing a computer program to be executed and a processor executing the stored computer program;
The computer program stored in the memory includes:
The acquisition module is used for acquiring data, wherein the acquired data comprise operation electrical data of the distributed photovoltaic power generation system and environment data of the place where the distributed photovoltaic system is located;
the data storage module is used for storing various data acquired by the acquisition module;
The preprocessing module is used for carrying out feature selection on the acquired data and carrying out sliding window overlapping sampling preprocessing;
the training module is used for building a distributed photovoltaic power generation recognition model based on a support vector machine by using window data obtained after window sampling pretreatment, and training the acquired data by using the model;
and the identification module is used for inputting the data actually acquired by the intelligent ammeter into the trained model for calculation by using the training result, and identifying the distributed photovoltaic power generation system in actual operation.
Further, the specific implementation method for data acquisition by the acquisition module comprises the following steps:
The data acquisition comprises the operation electrical data of the distributed photovoltaic power generation system and the environmental data of the place where the distributed photovoltaic power generation system is located;
establishing a data acquisition architecture of a distributed power generation system, wherein the data acquisition architecture comprises the following steps:
The acquisition module is required to be connected with one end of an inverter of the distributed photovoltaic power generation system, and is simultaneously connected with a connection point of the home bus and the power distribution network, and data required to be acquired comprise current and power; the current of the distributed photovoltaic power generation system is I PV, and the power of the distributed photovoltaic power generation system is P PV; the total household load current I load and the total household power P load; the household payload current flowing into the distribution network is I nl, and the household payload power is P nl;
collecting environmental data of a place where a distributed photovoltaic power generation system is located, wherein the environmental data comprise temperature T and illumination intensity Q;
Each piece of collected data is marked in time;
in order to match the sampling frequency of the intelligent electric energy meter, the sampling frequency of the data acquisition is 1 minute;
And filtering and denoising the acquired data.
Further, the specific implementation method of the preprocessing module for feature selection and sliding window overlapping sampling preprocessing comprises the following steps:
The relation among the distributed photovoltaic power generation system power P PV, the total household power P load and the household payload power P nl is shown in the formula (1):
Pnl(t)= Pload(t)– PPV(t) （1）；
Wherein t represents a time stamp;
The power relation described in the formula (1) relates to two forms, wherein the first form P nl (t) >0 represents that the generated energy of the distributed photovoltaic power generation system is smaller than the total load of a household, and a power distribution network is required to transmit power to the inside of the household; the second form P nl (t) <0 shows that the generated energy of the distributed photovoltaic power generation system is larger than the total load of the family, and the redundant electric quantity is fed back to the power distribution network;
the characteristic data is subjected to sliding window sampling to generate a characteristic data window matrix, n represents window length, and a represents sliding step length; the measured values are grouped in windows of n samples, and the window moving distance is half of the size of the selected window, so that the window not only can cover the changes in the grouped samples, but also can include the changes at the edge of each window;
Calculating the mean value, the median, the variance, the standard deviation and the root mean square value of each characteristic data, and reducing the window width to 5 samples;
The sum of all the measured values from 1 to n is calculated and then divided by the total number of samples in the window n to obtain the mean value Wherein n is equal to the window size; mean/>As shown in formula (2);
（2）；
Variance is defined as the mean of square deviation from the mean, and the mathematical equivalent is defined as shown in formula (3);
（3）；
Arranging a group of characteristic data window vectors from lowest to highest, wherein the median represents the median value of a given range;
the standard deviation is the square root of the variance, as shown by equation (4):
（4）；
root mean square values are defined as the effective power values that vary within each time window, and their descriptive values are shown in equation (5):
（5）；
Classifying the characteristic data set windows, evaluating each created window, and classifying the window into a window generating distributed photovoltaic power, namely a PV window, or a window not generating distributed photovoltaic power, namely a non-PV window;
The specific classification method comprises the following steps: setting a threshold for the distributed photovoltaic power generation power, and classifying the window as a PV window if at least half of the window values indicate that the distributed photovoltaic power generation power is higher than the set threshold, and otherwise, classifying the window as a non-PV window; the percentage of non-PV windows and PV windows are compared so that the number of windows of each type is the same.
Further, the specific implementation method for building the distributed photovoltaic power generation recognition model and model training based on the support vector machine by the training module comprises the following steps:
Selecting a training set and a testing set: selecting windows with classification completion, and respectively selecting 75% of the windows with the classification completion as training sets and 25% of the windows with the classification completion as test sets;
Constructing a distributed power generation system identification objective function by using a support vector machine algorithm, wherein the identification objective function is shown in a formula (6);
f(x)=w×k(x)+c=0 （6）；
Where k (x) represents a vector of x mapped to a new feature space, called a kernel function, w is a weight vector, and c is a bias;
The support vector machine algorithm adopts a Gaussian kernel function, as shown in a formula (7);
（7）；
Where x i represents the ith sample vector, x j represents the jth sample vector, and γ is a parameter of the gaussian kernel function.
The beneficial effects of adopting above-mentioned technical scheme to produce lie in: the distributed photovoltaic identification device based on non-invasive load monitoring provided by the invention can effectively utilize minute-level acquisition data of the intelligent electric energy meter, and identify a distributed photovoltaic power generation system in household payload power. The non-invasive load identification is completed without additionally adding power distribution system equipment, so that the identification efficiency is improved, and the cost is reduced. Meanwhile, a thought can be provided for the identification of more new energy power generation and energy storage equipment in the future, the load control capacity is improved, and the construction of a novel power system is effectively promoted.
Drawings
Fig. 1 is a functional schematic diagram of a distributed photovoltaic identification device based on non-invasive load monitoring according to an embodiment of the present invention;
Fig. 2 is a flowchart of an implementation algorithm of a distributed photovoltaic identification device based on non-invasive load monitoring according to an embodiment of the present invention;
Fig. 3 is a diagram of a distributed photovoltaic system data acquisition architecture according to an embodiment of the present invention;
Fig. 4 is a schematic diagram of a sliding window procedure according to an embodiment of the present invention.
Detailed Description
The following describes in further detail the embodiments of the present invention with reference to the drawings and examples. The following examples are illustrative of the invention and are not intended to limit the scope of the invention.
A distributed photovoltaic identification device based on non-invasive load monitoring, a memory storing a computer program to be executed, and a processor executing the stored computer program.
As shown in fig. 1, the computer program stored in the memory includes: the acquisition module is used for acquiring data, wherein the acquired data comprise operation electrical data of the distributed photovoltaic power generation system and environmental data of the place where the distributed photovoltaic system is located; the data storage module is used for storing various data acquired by the acquisition module; the preprocessing module is used for carrying out feature selection on the acquired data and carrying out sliding window overlapping sampling preprocessing; the training module is used for building a distributed photovoltaic power generation recognition model based on a support vector machine by using window data obtained after window sampling pretreatment, and training the acquired data by using the model; the recognition module is used for inputting data actually collected by the intelligent ammeter into a trained model for calculation by using a training result, and recognizing the distributed photovoltaic power generation system in actual operation.
The device may be a device including a processor and a memory, such as a single-chip microcomputer including a central processing unit. And, the processor is for executing the computer program stored in the memory.
The Processor may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), off-the-shelf Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory comprises a memory program area and a memory data area, wherein the memory program area can store an operating system, application programs required by at least one function and the like; the storage data area may store data created from the sensors, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart memory card (SMART MEDIA CARD, SMC), secure Digital (SD) card, flash memory card (FLASH CARD), at least one disk storage device, flash memory device, or other volatile solid-state storage device. The computer program comprises computer program code which may be in source code form, object code form, executable file or in some intermediate form, etc.
As shown in fig. 2, the specific implementation method of the apparatus computer program of this embodiment is as follows.
S1, data acquisition is carried out by an acquisition module, and the acquired data comprise operation electric data of the distributed photovoltaic power generation system and environment data of the place where the distributed photovoltaic power generation system is located.
S1.1, establishing a data acquisition architecture of the distributed power generation system.
As shown in fig. 3, the collection module needs to be connected to one end of an inverter of the distributed photovoltaic power generation system, and is simultaneously connected to a connection point of the home bus and the power distribution network, and data to be collected includes current and power. The current of the distributed photovoltaic power generation system is I PV=[IPV1,IPV2,…,IPVn, and the power of the distributed photovoltaic power generation system is P PV=[PPV1,PPV2,…,PPVn; the total household load current is I load=[Iload1,Iload2,…,Iloadn, and the total household power is P load=[Pload1,Pload2,…,Ploadn;
S1.2, collecting environment data of a distributed power generation system installation place, wherein the environment data comprise temperature T= [ T 1,T2,…,Tn ], illumination intensity Q= [ Q 1,Q2,…,Qn ] and the like;
And S1.3, each data acquired in the steps S1.2 and S1.3 is provided with a time mark, the sampling time interval is 1 minute, n represents the number of sampling points, and the operation electrical data of the distributed power generation system corresponds to the environmental information.
S1.4, filtering and denoising the acquired data, wherein the filtering and denoising of the data is not limited in the embodiment.
S2, a preprocessing module performs feature selection on the data acquired in the step S1 and performs sliding window overlapping sampling preprocessing.
S2.1, the relation among the distributed photovoltaic power generation system power P PV, the total household power P load and the household payload power P nl in the step S1 is shown in the formula (1):
Pnl(t)= Pload(t)– PPV(t) （1）；
Wherein t represents a time stamp.
The power relation described in step S2.2, step S2.1, in this embodiment relates to two forms. The first form P nl (t) >0 represents that the generated energy of the distributed photovoltaic power generation system is smaller than the total load of a household, and the power distribution network is required to transmit power to the inside of the household; and the second form P nl (t) <0 shows that the generated energy of the distributed photovoltaic power generation system is larger than the total load of the household, and the redundant electric quantity is fed back to the power distribution network.
S2.3, a characteristic data window matrix generated by sliding window sampling of the characteristic data is represented by n, the window length is represented by a, and the sliding step length is represented by a. In this embodiment, the measurements are grouped in windows of 50 samples, the window movement distance selecting half the window size, which can cover not only the variations within the grouped samples, but also the variations at the edges of each window. The sliding window procedure is shown in fig. 4.
S2.4, calculating the mean value, the median, the variance, the standard deviation and the root mean square value of each characteristic data, and reducing the window width to 5 samples. This not only helps to reduce measurement noise, but also optimizes algorithm processing time.
The sum of all the measured values from 1 to n is calculated and then divided by the total number of samples in the window n to obtain the mean valueWhere n is equal to the window size. Mean/>As shown in formula (2);
（2）。
Variance is defined as the mean of square deviation from the mean, and the mathematical equivalent is defined as shown in formula (3);
（3）。
Arranging a group of characteristic data window vectors from lowest to highest, wherein the median represents the median value of a given range;
the standard deviation is the square root of the variance, as shown by equation (4):
（4）；
root mean square values are defined as the effective power values that vary within each time window, and their descriptive values are shown in equation (5):
（5）；
S2.5, classifying the characteristic data set windows: each created window is evaluated and classified as either a window that generates distributed photovoltaic power generation (PV window) or a window that does not generate distributed photovoltaic power generation (non-PV window). To this end, a threshold is set for the distributed photovoltaic power generation to eliminate extremely low and noisy measurements of raw data. In this case, if the value of at least half of the window shows that the distributed photovoltaic power generation power is higher than the set threshold value, the window is classified as distributed photovoltaic power generation.
The percentage of windows classified as non-PV windows and PV are then compared, making the number of windows of each type the same. Part of the window data may need to be discarded in this process to provide a balanced data set.
And S3, a training module uses window data obtained after window sampling pretreatment to establish a distributed photovoltaic power generation recognition model based on a support vector machine, and the model is used for training the acquired data.
S3.1, selecting a training set and a testing set: the windows for which classification was completed were selected, 75% as training set and 25% as test set were selected for both types of windows, respectively.
S3.2, constructing a distributed power generation system identification objective function by using a support vector machine algorithm, wherein the identification objective function is shown in a formula (6):
f(x)=w×k(x)+c=0 （6）；
where k (x) represents the vector of x mapped to the new feature space, called the kernel, w is the weight vector, and c is the bias.
S3.3, the support vector machine algorithm adopts a Gaussian kernel function, as shown in a formula (7);
（7）；
Where x i represents the ith sample vector, x j represents the jth sample vector, and γ is a parameter of the gaussian kernel function.
And S4, evaluating the identification effect of the distributed power generation system by using the test set data.
S4.1, evaluating the identification effect of the distributed power generation system by adopting an accuracy rate (Pr), an accuracy rate (Acc), a recall rate (Re) and an F1 fraction (F1), wherein the formulas are as follows:
(8)；
(9)；
(10)；
(11)。
In step S4.2 and step S4.1, for a sample window to be identified, the model identification result is compared with the actual situation, and the actual case (TP), the true counterexample (TN), the false positive case (FP), and the false counterexample (FN). In this embodiment, the true example and the true counterexample may be described as successfully identifying that the distributed photovoltaic power generation system is running and that the distributed photovoltaic power generation system is shutdown within a certain period of time, where the false positive example is that the identification result is running when the distributed photovoltaic power generation system is not running, and the false counterexample indicates that the identification result is shutdown when the distributed photovoltaic power generation system is running. The meaning of the parameters is shown in Table 1.
TABLE 1 meanings of parameters
| Classification | TP | TN | FP | FN |
| Actual situation | 1 | 0 | 1 | 0 |
| Recognition result | 1 | 0 | 0 | 1 |
In order to verify the validity of the device of this embodiment, the device computer program of this embodiment was verified using laboratory data. Firstly, adopting characteristic data measurement quantity as the output of a support vector machine algorithm directly, namely adopting a window of 50 samples and a label thereof to train and test the support vector machine; in a second method, the mean, median, variance, standard deviation and root mean square value of each window are calculated as described in step S2.4. These variables are used as inputs to the classifier, reducing the dimension of the input data from 50 to 5. Both experiments were performed in the same processor. The results obtained are shown in Table 2.
Table 2 experimental results
| Classification | Actual situation | Recognition result |
| TP | 78234 | 78476 |
| TN | 82887 | 83160 |
| FP | 1031 | 761 |
| FN | 5591 | 5502 |
| Pr | 98.75% | 99.03% |
| Acc | 96.05% | 96.25% |
| Re | 94.42% | 93.46% |
| F1 | 95.95% | 96.16% |
| Window size | 50 | 5 |
| Processing time | 2.04 Ms/window | 0.63 Ms/window |
According to the non-invasive type distributed photovoltaic system identification device provided by the embodiment, aiming at the characteristics of a distributed photovoltaic system, the identification of the non-invasive type distributed photovoltaic system based on the data of the electric energy meter is realized by adopting low-frequency electric data and environment data.
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 corresponding technical solutions, which are defined by the scope of the appended claims.
Claims (4)
1. A distributed photovoltaic identification device based on non-invasive load monitoring, characterized in that: comprising a memory storing a computer program to be executed and a processor executing the stored computer program;
The computer program stored in the memory includes:
The acquisition module is used for acquiring data, wherein the acquired data comprise operation electrical data of the distributed photovoltaic power generation system and environment data of the place where the distributed photovoltaic system is located;
the data storage module is used for storing various data acquired by the acquisition module;
The preprocessing module is used for carrying out feature selection on the acquired data and carrying out sliding window overlapping sampling preprocessing;
the training module is used for building a distributed photovoltaic power generation recognition model based on a support vector machine by utilizing window data obtained by the sliding window overlapping sampling pretreatment, and training the acquired data by using the model;
and the identification module is used for inputting the data actually acquired by the intelligent ammeter into the trained model for calculation by using the training result, and identifying the distributed photovoltaic power generation system in actual operation.
2. A distributed photovoltaic identification device based on non-invasive load monitoring according to claim 1, characterized in that: the specific implementation method for the data acquisition by the acquisition module comprises the following steps:
establishing a data acquisition architecture of a distributed power generation system, wherein the data acquisition architecture comprises the following steps:
The acquisition module is required to be connected with an inverter of the distributed photovoltaic power generation system, and is simultaneously connected with a connection point of the home bus and the power distribution network, and data required to be acquired comprise current and power; the current of the distributed photovoltaic power generation system is I PV, and the power of the distributed photovoltaic power generation system is P PV; the total household load current I load and the total household power P load; the household payload current flowing into the distribution network is I nl, and the household payload power is P nl;
collecting environmental data of a place where a distributed photovoltaic power generation system is located, wherein the environmental data comprise temperature T and illumination intensity Q;
Each piece of collected data is marked in time;
in order to match the sampling frequency of the intelligent electric energy meter, the sampling frequency of the data acquisition is 1 minute;
And filtering and denoising the acquired data.
3. A distributed photovoltaic identification device based on non-invasive load monitoring according to claim 2, characterized in that: the specific implementation method for the preprocessing module to perform feature selection and sliding window overlapping sampling preprocessing comprises the following steps:
The relation among the distributed photovoltaic power generation system power P PV, the total household power P load and the household payload power P nl is shown in the formula (1):
Pnl(t)= Pload (t)– PPV (t) （1）；
Wherein t represents a time stamp;
The power relation described in the formula (1) relates to two forms, wherein the first form P nl (t) >0 represents that the generated energy of the distributed photovoltaic power generation system is smaller than the total load of a household, and a power distribution network is required to transmit power to the inside of the household; the second form P nl (t) <0 shows that the generated energy of the distributed photovoltaic power generation system is larger than the total load of the family, and the redundant electric quantity is fed back to the power distribution network;
The characteristic data is subjected to sliding window overlapping sampling pretreatment to generate a characteristic data window matrix, n represents the window length, and a represents the sliding step length; the measured values are grouped in windows of n samples, and the window moving distance is half of the size of the selected window, so that the window not only can cover the changes in the grouped samples, but also can include the changes at the edge of each window;
Calculating the mean value, the median, the variance, the standard deviation and the root mean square value of each characteristic data, and reducing the window width to 5 samples;
The sum of all the measured values from 1 to n is calculated and then divided by the total number of samples in the window n to obtain the mean value Wherein n is equal to the window size; mean/>As shown in formula (2);
（2）；
Variance is defined as the mean of square deviation from the mean, and the mathematical equivalent is defined as shown in formula (3);
（3）；
Arranging a group of characteristic data window vectors from lowest to highest, wherein the median represents the median value of a given range;
the standard deviation is the square root of the variance, as shown by equation (4):
（4）；
root mean square values are defined as the effective power values that vary within each time window, and their descriptive values are shown in equation (5):
（5）；
classifying the characteristic data windows, evaluating each created window, and classifying the window into a window generating distributed photovoltaic power, namely a PV window, or a window not generating distributed photovoltaic power, namely a non-PV window;
The specific classification method comprises the following steps: setting a threshold for the distributed photovoltaic power generation power, and classifying the window as a PV window if at least half of the window values indicate that the distributed photovoltaic power generation power is higher than the set threshold, and otherwise, classifying the window as a non-PV window; the percentage of non-PV windows and PV windows are compared so that the number of windows of each type is the same.
4. A distributed photovoltaic identification device based on non-invasive load monitoring according to claim 3, characterized in that: the specific implementation method for building the distributed photovoltaic power generation recognition model and model training based on the support vector machine by the training module comprises the following steps:
Selecting a training set and a testing set: selecting windows with classification completion, and respectively selecting 75% of the windows with the classification completion as training sets and 25% of the windows with the classification completion as test sets;
Constructing a distributed power generation system identification objective function by using a support vector machine algorithm, wherein the identification objective function is shown in a formula (6);
f(x)=w×k(x)+c=0 （6）；
Where k (x) represents a vector of x mapped to a new feature space, called a kernel function, w is a weight vector, and c is a bias;
The support vector machine algorithm adopts a Gaussian kernel function, as shown in a formula (7);
（7）；
Where x i represents the ith sample vector, x j represents the jth sample vector, and γ is a parameter of the gaussian kernel function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410390116.4A CN117996838B (en) | 2024-04-02 | 2024-04-02 | Distributed photovoltaic recognition device based on non-invasive load monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410390116.4A CN117996838B (en) | 2024-04-02 | 2024-04-02 | Distributed photovoltaic recognition device based on non-invasive load monitoring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117996838A true CN117996838A (en) | 2024-05-07 |
| CN117996838B CN117996838B (en) | 2024-06-28 |
Family
ID=90893496
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410390116.4A Active CN117996838B (en) | 2024-04-02 | 2024-04-02 | Distributed photovoltaic recognition device based on non-invasive load monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117996838B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112198385A (en) * | 2020-09-30 | 2021-01-08 | 国网山西省电力公司晋中供电公司 | Non-invasive load monitoring method |
| CN112434751A (en) * | 2020-12-04 | 2021-03-02 | 国网江苏省电力有限公司宿迁供电分公司 | Resident cell distributed photovoltaic power generation identification method |
| CN115935225A (en) * | 2022-11-15 | 2023-04-07 | 广东电网有限责任公司 | Non-invasive load identification method and device |
| CN116937565A (en) * | 2023-07-24 | 2023-10-24 | 国网北京市电力公司 | Distributed photovoltaic power generation power prediction method, system, equipment and medium |
| CN117477536A (en) * | 2023-10-20 | 2024-01-30 | 国网四川省电力公司营销服务中心 | Resident non-invasive load decomposition method, system and medium considering photovoltaics |
-
2024
- 2024-04-02 CN CN202410390116.4A patent/CN117996838B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112198385A (en) * | 2020-09-30 | 2021-01-08 | 国网山西省电力公司晋中供电公司 | Non-invasive load monitoring method |
| CN112434751A (en) * | 2020-12-04 | 2021-03-02 | 国网江苏省电力有限公司宿迁供电分公司 | Resident cell distributed photovoltaic power generation identification method |
| CN115935225A (en) * | 2022-11-15 | 2023-04-07 | 广东电网有限责任公司 | Non-invasive load identification method and device |
| CN116937565A (en) * | 2023-07-24 | 2023-10-24 | 国网北京市电力公司 | Distributed photovoltaic power generation power prediction method, system, equipment and medium |
| CN117477536A (en) * | 2023-10-20 | 2024-01-30 | 国网四川省电力公司营销服务中心 | Resident non-invasive load decomposition method, system and medium considering photovoltaics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117996838B (en) | 2024-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | A hybrid forecasting approach applied to wind speed time series | |
| WO2022067562A1 (en) | Method and device for diagnosing fault arc, and computer-readable storage medium | |
| CN106992523B (en) | Load flow calculation method for power system with photovoltaic and thermal loads | |
| CN110991263B (en) | Non-invasive load identification method and system for resisting background load interference | |
| CN113011481B (en) | Electric energy meter function abnormality assessment method and system based on decision tree algorithm | |
| Rehman et al. | Low complexity non-intrusive load disaggregation of air conditioning unit and electric vehicle charging | |
| CN113076985B (en) | Power consumer load identification method and device and computer readable medium | |
| CN112365090A (en) | Deep learning-based non-invasive electrical load identification method and device | |
| Hernandez et al. | Development of a non-intrusive load monitoring (nilm) with unknown loads using support vector machine | |
| CN109921462B (en) | New energy consumption capability assessment method and system based on LSTM | |
| Han et al. | Non‐intrusive load monitoring by voltage–current trajectory enabled asymmetric deep supervised hashing | |
| CN114139820A (en) | Improved modal decomposition method for non-invasive electric energy load prediction | |
| CN117996838B (en) | Distributed photovoltaic recognition device based on non-invasive load monitoring | |
| CN111950752A (en) | Photovoltaic power station generating capacity prediction method, device and system and storage medium thereof | |
| CN114118401A (en) | Neural network-based power distribution network flow prediction method, system, device and storage medium | |
| CN115983507B (en) | Method and system for predicting broadband oscillation risk of section of power grid of transmitting end source | |
| CN116861316A (en) | Electrical appliance monitoring method and device | |
| CN116502149A (en) | Low-voltage power distribution network user-transformation relation identification method and system based on current characteristic conduction | |
| CN111199014A (en) | Time sequence based seq2point NILM method and device | |
| CN111126645A (en) | Wind power prediction algorithm based on data mining technology and improved support vector machine | |
| CN114358160A (en) | Data anomaly detection method in power system | |
| Behzadi et al. | Dual identification of multi-complex and non-stationary power quality disturbances using variational mode decomposition in hybrid modern power systems | |
| Behzadi et al. | Identification of combined power quality disturbances in the presence of distributed generations using variational mode decomposition and K-nearest neighbors classifier | |
| CN112993977A (en) | HHT and RVM based active power distribution network electric energy quality disturbance identification method | |
| CN116188199A (en) | Non-invasive load decomposition method and system |
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 |