A method and system for calculating lost power of a photovoltaic string

Abstract

The application provides a photovoltaic string loss power calculation method, historical operation data of a string is acquired; a device state analysis model is established, and the string is classified according to the historical operation data of the string; a plurality of control group strings under different conditions are set, and corresponding control loss power is obtained according to the power generation of the control group and the actual power generation of the string; a loss power calculation model is constructed according to the historical operation data of the string, the state of the string and the control loss power; and the measured operation data of the string is brought into the loss power calculation model to obtain the loss power of the string. A system for implementing the above method is also provided. The application constructs a model for each string, and considers the specific device state when constructing the model, thereby solving the problem of inaccurate loss power calculation in the prior art.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation monitoring technology, and in particular to a method and system for calculating the power loss of photovoltaic strings. Background Technology

[0002] The power generation level of a photovoltaic (PV) power plant during operation is the most critical factor directly affecting its economic efficiency. This level is closely related to the plant's daily maintenance, requiring timely inspection and repair of faulty equipment and its condition. The power loss of the strings is a crucial basis for PV equipment inspection and maintenance. However, analyzing and calculating power loss during actual operation is quite difficult, hindering staff from assessing the overall power generation efficiency of the plant.

[0003] CN202111353709 predicts and analyzes power loss under a single fault model based on dust coverage, but lacks systematic analysis of power loss under various fault scenarios, resulting in poor accuracy in calculating power loss under other fault modes. CN202111210412 uses normal operation data of other similar strings as input variables and employs a neural network model to calculate power loss for abnormal strings. However, different strings have different model parameters, and a single prediction model cannot guarantee the accuracy of power loss calculation. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for calculating the power loss of photovoltaic strings, which solves the problem of insufficient accuracy in calculating power loss in existing technologies.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for calculating the power loss of a photovoltaic string includes the following steps: obtaining historical operating data of the string;

[0007] A device status analysis model is established, and the string status is classified according to the historical operating data of the string. A control group string is set up, and the corresponding control loss power is obtained according to the power generation of the control group and the historical actual power generation of the string. A loss power calculation model is constructed based on the historical operating data of the string, the status of the string, and the control loss power. The string loss power is obtained by substituting the actual operating data of the string into the loss power calculation model.

[0008] This invention also proposes a method for calculating the power loss of a photovoltaic power plant based on the above method, which superimposes the power loss of all strings in the power plant to obtain the power loss and distribution of the photovoltaic power plant.

[0009] The present invention also proposes a photovoltaic string power loss calculation system, including a data acquisition module, a data processing module, and a data analysis module; the data acquisition module is used to collect the string's operating data; the data processing module is used to extract the power generation, the features affecting power generation, and the string image features from the operating data; the data analysis module is used to analyze the string equipment status and calculate the current string power loss.

[0010] Compared with the prior art, the present invention has the following beneficial effects:

[0011] This invention uses historical operating data of photovoltaic (PV) strings as a reference for power loss, obtained by setting up a control group, to construct a model. This invention establishes a separate model for each PV string based on its own historical data, avoiding the problem of model incompatibility caused by differences in individual strings and environmental conditions. When building the model, it considers not only the correspondence between historical data and the reference power loss, but also the equipment state of the PV string at the time of data collection. This ensures that the power loss model, once established, takes into account the changing patterns and inherent relationships between characteristics of the historical operating data, resulting in a more accurate power loss calculation. This provides a precise reference for subsequent maintenance and thus ensures the power generation level of the PV power plant. Other advantages, objectives, and features of this invention will be partly apparent from the following description, and partly understood by those skilled in the art through study and practice of the invention. Attached Figure Description

[0012] Figure 1 This is an overall logic block diagram of an embodiment of the present invention. Detailed Implementation

[0013] To make the technical means, creative features, objectives and effects of the present invention clearer and easier to understand, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0014] like Figure 1 As shown, this invention provides a method for calculating the power loss of a photovoltaic string, comprising the following steps: acquiring historical operating data of the string; establishing a device status analysis model and classifying the string status according to the historical operating data; setting a control group string and obtaining the corresponding control power loss based on the power generation of the control group and the historical actual power generation of the string; constructing a power loss calculation model based on the historical operating data of the string, the status of the string, and the control power loss; and inputting the measured operating data of the string into the power loss calculation model to obtain the power loss of the string.

[0015] To ensure the operation, management, and control of photovoltaic (PV) power plants, data monitoring is currently conducted on a string-by-string basis. This data primarily includes DC voltage, DC current, inverter efficiency of the connected inverters, grid frequency, daily power generation, power factor, DC power, panel orientation, weather conditions, radiation intensity, and ambient temperature. Some of this data can be readily analyzed using existing data analysis models to determine the power generation of the PV system and information on equipment damage and aging. When modeling using this data, features related to voltage, current, power, radiation intensity, and angle are extracted to calculate power generation or influence it. Abnormal data is then removed. Existing technologies have provided relatively complete descriptions of specific modeling methods, differing only in the factors considered. The modeling process can be completed as follows: The PV array is divided into strings, each string is assigned a normal operating physical control group, the sampling frequency and period are determined, and sufficient actual operating data for each string is obtained, forming a historical operating dataset for the strings and a normal operating dataset for the corresponding physical control group at a given time. Data aggregation and dimensionality reduction were performed on datasets one and two, and feature extraction was achieved using statistical machine learning or deep learning methods. A data feature extraction tool was constructed to lay the foundation for subsequent data modeling. Model learning requires the analysis of historical string operation data to form feature quantities, state quantities, and loss quantities. First, based on dataset one of the string's historical operation and the extracted data features, the feature quantities required for model learning were formed. The current operating state of the string was identified using a device state analysis model, mainly including: normal state, foreign object coverage, shadow occlusion, dust accumulation, component aging, etc., forming the state quantities of the string operation data. At the same time, based on the data analysis model, the actual power generation at the current string operation moment was analyzed. Further, based on dataset two of the string physical control group operating in normal state and the extracted data features, the normal power generation of the physical control group was comprehensively evaluated and calculated using the data analysis model and the power generation prediction model. Finally, based on the comparison error between the obtained actual power generation of the string and the normal power generation of the control group, the power loss of the string in the current state can be calculated, forming the loss quantity required for model learning. Based on the feature fusion model, the string state variables and data features are fused, and combined with the string loss variables from the evaluation analysis, machine learning techniques such as decision trees, support vector machines, neural networks, deep learning, and ensemble learning are used to construct a string-level power loss calculation model.

[0016] The purpose of studying string power generation loss is twofold: firstly, to gain a comprehensive understanding and control of string power generation; and more importantly, to provide guidance for subsequent maintenance. In the actual operation of photovoltaic power plants, equipment failure, component aging, foreign object obstruction, shading, and dust accumulation significantly reduce power plant efficiency. However, due to rapid changes in physical parameters such as weather and environment, analyzing and calculating power loss is difficult, hindering staff from assessing the overall power plant efficiency. While weather and shading are difficult to mitigate through maintenance, equipment aging and dust accumulation can be improved through maintenance. Therefore, whether to include power loss in the calculation depends on the specific purpose of the study.

[0017] Due to the large area of ​​photovoltaic power plants, manual inspection is difficult and inconvenient for recording. Therefore, drones are currently mostly used to obtain inspection image data. This invention, through image analysis, can determine the presence of foreign objects and dust accumulation, and deduce their impact on power generation efficiency. Shading exhibits a clear temporal pattern depending on the angle of sunlight at different times of day, ultimately reflecting changes in radiation intensity and location over time, which can be derived from operational data patterns. Equipment failures and component aging are directly reflected in abnormal parameters of the equipment itself. Classifying the equipment status of photovoltaic strings allows for the consideration of historical data development patterns and the inherent relationships between data components during modeling, leading to a more accurate model for calculating power loss. Furthermore, it provides more valuable reference for subsequent maintenance.

[0018] The control group data were obtained under the same temperature, radiation intensity, and wind speed conditions, assuming the equipment was functioning normally, the surface was clean, and there were no obstructions. Analysis revealed that, under normal equipment conditions, the main factor affecting power generation is radiation intensity, with temperature and wind speed also having some influence. Furthermore, these three factors are interconnected; therefore, the control group must maintain consistency.

[0019] A data analysis model is constructed to calculate the actual power generation of the string and the power generation of the control group. A power generation prediction model is also constructed to calculate the power generation of the control group and compare it with the control group's power generation data. A deviation threshold is set between the two models, and data within the threshold range is used as a control. The data analysis model calculates power generation based on the detected data, while the power generation prediction model predicts power generation based on current environmental and equipment conditions. Under normal circumstances, the data obtained from these two methods in the control group should be consistent. By comprehensively evaluating the power generation of the control group using both models, the problem of inaccurate data from the control group itself can be avoided.

[0020] In another embodiment, the present invention proposes a method for calculating the power loss of a photovoltaic power station. This method involves superimposing the power loss of all strings within the power station to obtain the power loss and distribution of the photovoltaic power station, thereby facilitating a better understanding of the overall power generation and equipment status of the photovoltaic power station. Except for some special peripheral locations, the configuration of each string in a photovoltaic power station is generally the same; therefore, the data from the control group is universal, and only routinely collected data is needed.

[0021] In another embodiment, the present invention proposes a photovoltaic string power loss calculation system, including a data acquisition module, a data processing module, and a data analysis module. The data acquisition module typically includes built-in detection devices on the equipment itself and video acquisition devices such as drones and cameras, used to collect string operation data. The data processing module is used to extract features from the operation data to calculate power generation, features affecting power generation, and string image features. The data analysis module is used to analyze the string equipment status and calculate the current string power loss. The data analysis module mainly includes an equipment status analysis module and a power loss calculation module. It also includes a comparison module used to calculate the difference between the actual power generation and the normal power generation of a control group, used when establishing a power loss model.

[0022] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for calculating the power loss of a photovoltaic string, characterized in that, Includes the following steps: Retrieve historical execution data of the string; Establish an equipment status analysis model and classify the equipment status of the strings based on the historical operating data of the strings; Set up a control group string, and calculate the corresponding control loss power based on the power generation of the control group and the historical actual power generation of the string; A power loss calculation model is constructed based on the historical operation data of the string, the state of the string, and the power loss. The actual measured operating data of the string is input into the power loss calculation model to obtain the power loss of the string and the corresponding equipment status.

2. The method for calculating the power loss of a photovoltaic string as described in claim 1, characterized in that, The historical operational data includes inspection image data of the series.

3. The method for calculating the power loss of a photovoltaic string as described in claim 2, characterized in that, The historical operating data includes sampling time, DC voltage, DC current, inverter efficiency of string-connected inverters, inverter grid frequency, inverter power, inverter DC power, string panel orientation, weather conditions, radiation intensity, and ambient temperature.

4. The method for calculating the power loss of a photovoltaic string as described in claim 3, characterized in that, The device status includes at least two of the following: fully normal, covered by foreign objects, shaded, dust accumulation, and component aging.

5. The method for calculating the power loss of a photovoltaic string as described in claim 1, characterized in that, The data for the control group were obtained under the same temperature, radiation intensity, and wind speed conditions, assuming the equipment was functioning normally, the surface was clean, and there were no obstructions.

6. A method for calculating the power loss of a photovoltaic string as described in claim 1 or 5, characterized in that, A data analysis model was constructed, and the actual power generation of the string and the power generation of the control group were calculated using the data analysis model.

7. The method for calculating the power loss of a photovoltaic string as described in claim 6, characterized in that, A power generation prediction model was constructed, and the power generation of the control group was calculated using the power generation prediction model. The power generation data of the control group calculated by the data analysis model was compared with the data of the control group. A deviation threshold was set between the two, and data within the threshold range were used as the control.

8. A method for calculating the power loss of a photovoltaic power plant, characterized in that, The power loss of a single string is calculated by using the method described in any one of claims 1-6, and the power loss of all strings in the photovoltaic power station is superimposed to obtain the power loss and distribution of the photovoltaic power station.

9. A photovoltaic string power loss calculation system, characterized in that, It includes a data acquisition module, a data processing module, and a data analysis module; the data acquisition module is used to collect the operating data of the string; the data processing module is used to extract the power generation, the features affecting power generation, and the string image features from the operating data; the data analysis module is used to analyze the status of the string equipment and calculate the current power loss of the string.

10. The photovoltaic string power loss calculation system as described in claim 9, characterized in that, It also includes a comparison module, which is used to calculate the difference between the actual power generation and the normal power generation.

Images