Distributed photovoltaic voltage out-of-limit prediction method and device and storage medium

Abstract

The invention relates to a distributed photovoltaic voltage out-of-limit prediction method and device, and a storage medium. The method comprises the following steps: collecting distributed photovoltaic historical operation data; constructing a voltage out-of-limit prediction model, and inputting distributed photovoltaic historical operation data into the voltage out-of-limit prediction model for training; acquiring real-time operation data of the distributed photovoltaic system, and inputting the real-time operation data into the trained voltage out-of-limit prediction model to obtain a voltage out-of-limit prediction probability of the distributed photovoltaic system; the model has higher prediction precision, and effectively improves the reliability and timeliness of voltage out-of-limit early warning while keeping the conservative property.

Description

Technical Field

[0001] This invention relates to a method, device, and storage medium for predicting voltage overruns in distributed photovoltaic systems, belonging to the field of voltage overrun prediction technology. Background Technology

[0002] Distributed photovoltaic (PV) power generation, with its advantages of being green, environmentally friendly, and readily available for local consumption, has ushered in unprecedented development opportunities. A large number of distributed PV power stations have been connected to the distribution network, becoming an important part of the energy supply system. However, while the large-scale grid connection of distributed PV brings many benefits, it also poses serious challenges to the safe and stable operation of the distribution network. Because the output power of distributed PV is significantly intermittent and fluctuating due to the influence of natural conditions such as sunlight intensity and temperature, and user-side load changes are also relatively frequent, the combined effect of these factors makes the power flow distribution of the distribution network complex and volatile, easily leading to voltage exceeding limits. Once voltage exceeds limits, it not only reduces power quality and affects the user's electricity experience, but may also threaten the safe operation of power equipment and even cause grid failures.

[0003] Therefore, researching an accurate and effective method for predicting voltage overruns in distributed photovoltaic systems, and gaining early insight into voltage change trends, is of significant theoretical and practical value for ensuring the safe, stable, and efficient operation of the distribution network and promoting the sustainable development of the distributed photovoltaic industry.

[0004] Currently, there are many research methods for voltage over-limit prediction. Traditional methods, such as time series analysis and regression analysis, have certain advantages in dealing with simple linear relationships, but their prediction accuracy is limited for complex nonlinear time series data such as distributed photovoltaic power output.

[0005] In the task of predicting voltage limits, existing models perform well when dealing with the phase of gradual voltage change, demonstrating high prediction accuracy and stability. However, when the voltage fluctuates drastically, the prediction results show significant lag. At the same time, when the voltage drops sharply, the predicted value is significantly lower than the actual value, causing some limit-breaking situations to be unpredictable. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention proposes a method, device, and storage medium for predicting voltage over-limit in distributed photovoltaic systems.

[0007] The technical solution of the present invention is as follows: On the one hand, the present invention provides a method for predicting voltage over-limit in distributed photovoltaic systems, comprising the following steps: Collect historical operating data of distributed photovoltaic power generation; A voltage over-limit prediction model is constructed, and historical operating data of distributed photovoltaic power is input into the voltage over-limit prediction model for training. Real-time operating data of distributed photovoltaic power is collected and input into a trained voltage over-limit prediction model to obtain the voltage over-limit prediction probability of distributed photovoltaic power.

[0008] Preferably, the historical operating data of the distributed photovoltaic system includes distributed photovoltaic power output data, distributed photovoltaic load data, and distributed photovoltaic voltage monitoring data.

[0009] Preferably, the voltage over-limit prediction model includes a preprocessing layer, a hidden layer, an attention layer, a fully connected layer, and a feedback layer.

[0010] Preferably, the preprocessing layer is used to preprocess historical operating data of distributed photovoltaic systems.

[0011] Preferably, the hidden layer is constructed based on a gated loop unit; Preliminary feature extraction is performed on the preprocessed historical operation data of distributed photovoltaic power generation through a hidden layer, and the time-series features of the historical operation data of distributed photovoltaic power generation are output.

[0012] Preferably, the attention layer is used to assign attention weights to the output of the hidden layer and output a context feature vector.

[0013] Preferably, the fully connected layer is used to perform a linear transformation on the output of the attention layer to output the voltage over-limit prediction probability of the distributed photovoltaic system.

[0014] Preferably, during the training of the voltage over-limit prediction model, the feedback layer adjusts the parameters of the voltage over-limit prediction model based on the output results of the fully connected layer in each round of training using the quantile loss function.

[0015] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the present invention.

[0016] In another aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described in the present invention.

[0017] The present invention has the following beneficial effects: 1. In the model training process, this invention selects different quantiles for different changing trends to obtain prediction intervals under different quantiles. By comparing errors, a voltage change trend-model parameter mapping record is formed. During prediction, the corresponding model parameters in the vector table are queried by analyzing the trend and different predictions are made. This model has higher prediction accuracy than traditional prediction models. At the same time, while retaining conservatism, it effectively improves the reliability and timeliness of voltage over-limit warning. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0021] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0022] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0023] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.

[0024] See Figure 1 A method for predicting voltage exceedance in distributed photovoltaic systems includes the following steps: Collect historical operating data of distributed photovoltaic power generation; A voltage over-limit prediction model is constructed, and historical operating data of distributed photovoltaic power is input into the voltage over-limit prediction model for training. Real-time operating data of distributed photovoltaic power is collected and input into a trained voltage over-limit prediction model to obtain the voltage over-limit prediction probability of distributed photovoltaic power.

[0025] In some embodiments, the historical operating data of the distributed photovoltaic system includes distributed photovoltaic power output data, distributed photovoltaic load data, and distributed photovoltaic voltage monitoring data.

[0026] In one specific embodiment, the historical operating data of the distributed photovoltaic system includes photovoltaic output data at 5-minute intervals, meteorological data (sunlight intensity, temperature, etc.), and voltage monitoring data.

[0027] In some embodiments, the voltage over-limit prediction model includes a preprocessing layer, a hidden layer, an attention layer, a fully connected layer, and a feedback layer.

[0028] In some embodiments, the preprocessing layer is used to preprocess historical operating data of distributed photovoltaic systems.

[0029] In one specific embodiment, the preprocessing layer preprocesses the historical operating data of distributed photovoltaic systems, specifically as follows: For missing values ​​in the historical operation data of distributed photovoltaic systems, smoothing is performed, and the average of the upper and lower values ​​of the missing value is taken as the filling value for the missing value. Normalize the historical operating data of distributed photovoltaic power generation after filling in missing values, and scale the data to [a specific scale]. Within the range, to improve the training efficiency and prediction accuracy of the model; The final preprocessing layer outputs preprocessed historical operating data of distributed photovoltaic systems.

[0030] In some embodiments, the hidden layer is constructed based on a gated loop unit; Preliminary feature extraction is performed on the preprocessed historical operation data of distributed photovoltaic power generation through a hidden layer, and the time-series features of the historical operation data of distributed photovoltaic power generation are output.

[0031] In one specific embodiment, the gated loop unit is set to a width of 40 and a depth of 1; In some embodiments, the attention layer is used to assign attention weights to the output of the hidden layer and output a context feature vector.

[0032] In some embodiments, the fully connected layer is used to perform a linear transformation on the output of the attention layer to output the voltage over-limit prediction probability of distributed photovoltaic.

[0033] In some embodiments, during the training of the voltage over-limit prediction model, the feedback layer adjusts the parameters of the voltage over-limit prediction model based on the output results of the fully connected layer in each round of training using a quantile loss function.

[0034] In a specific embodiment, the feedback layer adjusts the voltage over-limit prediction model parameters based on the output of each round of training of the fully connected layer using the quantile loss function. The specific steps are as follows: Based on the voltage monitoring data of the input voltage over-limit prediction model trained in this round, the voltage change trend is calculated, as shown in the following formula: ; in: This indicates the voltage change trend corresponding to the current voltage monitoring data; This indicates the voltage value at the end of the collected voltage monitoring data; This indicates the voltage value at the start of the collected voltage monitoring data; Multiple continuous voltage change ranges are preset. In this embodiment, based on the general voltage change range of ±0.2 (20%), five ranges are preset, as shown in the following formula: ; The quantile set is pre-defined as a combination of multiple different quantiles. In this embodiment, the specific formula for selecting quantiles to construct the quantile set is as follows: ; For the voltage change trend of the voltage monitoring data corresponding to this round of training, the voltage limit prediction probability loss is calculated for each quantile in the quantile set using the quantile loss function. The minimum voltage limit prediction probability loss is selected and it is determined whether the voltage limit prediction probability loss reaches the preset loss threshold. If it does not reach the threshold, the voltage limit prediction model parameters are adjusted and the training is performed again. If the preset loss threshold is reached, the current voltage change trend interval, quantile value, and the corresponding parameters of the voltage limit prediction model are recorded to construct a mapping record. Once the preset number of training rounds is reached, training stops, and the trained voltage over-limit prediction model is obtained.

[0035] In a specific embodiment, the specific steps for predicting voltage limits using the trained voltage limit prediction model are as follows: Collect real-time photovoltaic power output data, meteorological data, and voltage monitoring data of distributed photovoltaic systems; The voltage change trend is calculated based on voltage monitoring data, and the corresponding model parameters are queried from all mapping records according to the range of the voltage change trend and applied to the current model. After applying the model parameters to the current model, real-time photovoltaic power output data, meteorological data, and voltage monitoring data of distributed photovoltaic power are input into the model for prediction, and the voltage over-limit prediction probability of distributed photovoltaic power is obtained.

[0036] In some embodiments, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method as described in any embodiment of the present invention.

[0037] In some embodiments, a computer-readable storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the method as described in any embodiment of the present invention.

[0038] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0039] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0040] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0041] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0042] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for predicting voltage exceedance in distributed photovoltaic systems, characterized in that, Includes the following steps: Collect historical operating data of distributed photovoltaic power generation; A voltage over-limit prediction model is constructed, and historical operating data of distributed photovoltaic power is input into the voltage over-limit prediction model for training. Real-time operating data of distributed photovoltaic power is collected and input into a trained voltage over-limit prediction model to obtain the voltage over-limit prediction probability of distributed photovoltaic power.

2. The method for predicting voltage exceedance in distributed photovoltaic systems according to claim 1, characterized in that, The historical operating data of the distributed photovoltaic system includes distributed photovoltaic power output data, distributed photovoltaic load data, and distributed photovoltaic voltage monitoring data.

3. The distributed photovoltaic voltage over-limit prediction method according to claim 1, characterized in that, The voltage over-limit prediction model includes a preprocessing layer, a hidden layer, an attention layer, a fully connected layer, and a feedback layer.

4. The distributed photovoltaic voltage over-limit prediction method according to claim 1, characterized in that, The preprocessing layer is used to preprocess historical operating data of distributed photovoltaic systems.

5. The method for predicting voltage exceedance in distributed photovoltaic systems according to claim 1, characterized in that, The hidden layer is constructed based on a gated loop unit; Preliminary feature extraction is performed on the preprocessed historical operation data of distributed photovoltaic power generation through a hidden layer, and the time-series features of the historical operation data of distributed photovoltaic power generation are output.

6. The distributed photovoltaic voltage over-limit prediction method according to claim 1, characterized in that, The attention layer is used to assign attention weights to the output of the hidden layer and output a context feature vector.

7. The method for predicting voltage exceedance in distributed photovoltaic systems according to claim 1, characterized in that, The fully connected layer is used to perform a linear transformation on the output of the attention layer, and output the voltage over-limit prediction probability of distributed photovoltaic.

8. The distributed photovoltaic voltage over-limit prediction method according to claim 1, characterized in that, During the training of the voltage over-limit prediction model, the feedback layer adjusts the parameters of the voltage over-limit prediction model based on the output results of the fully connected layer in each round of training using the quantile loss function. The specific steps are as follows: Based on the voltage monitoring data of the input voltage over-limit prediction model trained in this round, the voltage change trend is calculated; Multiple consecutive voltage change ranges are preset; The set of quantiles is predefined as a combination of multiple different quantiles. For the voltage change trend of the voltage monitoring data corresponding to this round of training, the voltage limit prediction probability loss is calculated for each quantile in the quantile set using the quantile loss function. The minimum voltage limit prediction probability loss is selected and it is determined whether the voltage limit prediction probability loss reaches the preset loss threshold. If it does not reach the threshold, the voltage limit prediction model parameters are adjusted and the training is performed again. If the preset loss threshold is reached, the current voltage change trend interval, quantile value, and the corresponding parameters of the voltage limit prediction model are recorded to construct a mapping record. Once the preset number of training rounds is reached, training stops, and the trained voltage limit prediction model is obtained. The specific steps for predicting voltage limits using the trained voltage limit prediction model are as follows: Collect real-time photovoltaic power output data, meteorological data, and voltage monitoring data of distributed photovoltaic systems; The voltage change trend is calculated based on voltage monitoring data, and the corresponding model parameters are queried from all mapping records according to the range of the voltage change trend and applied to the current model. After applying the model parameters to the current model, real-time photovoltaic power output data, meteorological data, and voltage monitoring data of distributed photovoltaic power are input into the model for prediction, and the voltage over-limit prediction probability of distributed photovoltaic power is obtained.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images