Distributed photovoltaic grid-connected regulation and control system

Abstract

The invention relates to the technical field of data processing, in particular to a distributed photovoltaic grid-connected regulation and control system, which comprises a memory, a processor and a computer program which is stored in the memory and runs on the processor, when the processor executes the computer program, the following steps are realized: acquiring an electric power change curve of electric power time sequence data and an illumination intensity change curve of illumination intensity time sequence data of any photovoltaic node in a historical time period; obtaining abnormal wave crest points in the electric power change curve; obtaining the environmental influence degree of the abnormal wave crest point according to a coordinate point corresponding to the abnormal wave crest point belonging to the same sampling moment in the illumination intensity change curve; according to the method, the lag coefficient of the ARIMA model is adaptively corrected according to the environmental influence degree of each abnormal wave crest point, the trained ARIMA model is obtained and used for grid-connected regulation and control of photovoltaic nodes, and the accuracy of electric power data prediction by using the trained ARIMA model is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a distributed photovoltaic grid-connected control system. Background Technology

[0002] Distributed photovoltaic (PV) grid-connected regulation refers to the real-time monitoring, management, and adjustment of the power output of a distributed PV system during its grid connection process, using various technologies and equipment. This ensures a safe and stable connection between the distributed PV system and the grid, and optimizes power generation efficiency. The distributed PV grid-connected regulation system integrates various data sources, including monitoring data from the distributed PV system and environmental parameters. This comprehensive information allows for grid-connected regulation of the distributed PV system, improving its efficiency.

[0003] In existing technologies, ARIMA models are typically used to predict the power data of distributed photovoltaic (PV) power generation systems to obtain corresponding predicted values. Grid connection regulation is then carried out based on these predicted values. However, with the large-scale development and integration of distributed PV, the randomness, volatility, and intermittency of distributed PV power generation systems make it difficult to use ARIMA models for prediction. This introduces significant uncertainty into the load of distributed PV power generation systems, resulting in low confidence levels in the power prediction results and consequently, poor accuracy in grid connection regulation based on the prediction results.

[0004] Therefore, improving the confidence level of power data prediction for distributed photovoltaic power generation systems using the ARIMA model to enhance the accuracy of grid-connected regulation has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a distributed photovoltaic grid-connected control system to address the problem of how to improve the confidence level of predicting power data of distributed photovoltaic power generation systems using ARIMA models, thereby improving the accuracy of grid-connected control.

[0006] This invention provides a distributed photovoltaic grid-connected control system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it performs the following steps:

[0007] In the process of grid-connected control of distributed photovoltaics, the power time-series data and irradiance time-series data of any photovoltaic node in the historical period are obtained, and the power change curve of the power time-series data and the irradiance change curve of the irradiance time-series data are constructed respectively.

[0008] All peaks in the power change curve are obtained. Based on the morphological characteristics and fluctuation range of each peak, the degree of abnormality of the morphological characteristics of each peak is obtained. Based on the degree of abnormality of the morphological characteristics of each peak, abnormal peaks are obtained.

[0009] For any abnormal peak point, the degree of environmental impact of the abnormal peak point is obtained based on the fluctuation difference between the abnormal peak point and other peak points, and the coordinate point in the light intensity change curve that corresponds to the same sampling time as the abnormal peak point.

[0010] The environmental impact level of each abnormal peak point is obtained. During the training of the ARIMA model using the power time series data, the lag coefficient of the corresponding power data in the power time series data is corrected according to the environmental impact level of each abnormal peak point to obtain the trained ARIMA model. The trained ARIMA model is then used to perform grid-connected regulation of the photovoltaic node.

[0011] Preferably, the step of obtaining the degree of morphological anomaly of each wave crest point based on its morphological characteristics and fluctuation range includes:

[0012] For any peak point, obtain the left and right adjacent coordinate points of the peak point, and obtain the absolute value of the difference in the line slope between the peak point and the left adjacent coordinate point and the right adjacent coordinate point.

[0013] In the power change curve, the lowest valley points on both sides of the peak point are obtained respectively. Based on the vertical axis span between the peak point and each of the lowest valley points, the maximum vertical axis span is obtained. The product between the maximum vertical axis span and the absolute value of the difference in the slope of the connecting line is normalized to obtain the corresponding first normalized value.

[0014] Obtain the horizontal span between the two lowest valley points, normalize the horizontal span to obtain the corresponding second normalized value, and use the sum of the first normalized value and the second normalized value as the degree of morphological anomaly of the peak point.

[0015] Preferably, the step of obtaining abnormal peak points based on the degree of abnormality in the morphological characteristics of each peak point includes:

[0016] If the degree of abnormality in the morphological characteristics of any peak point is greater than or equal to a preset threshold for the degree of abnormality in morphological characteristics, then the peak point is regarded as an abnormal peak point.

[0017] Preferably, the step of obtaining the environmental impact degree of the abnormal peak point based on the fluctuation difference between the abnormal peak point and other peak points, and the coordinate points in the illumination intensity change curve that correspond to the same sampling time as the abnormal peak point, includes:

[0018] The degree of fluctuation instability of the abnormal peak point is obtained based on the fluctuation difference between the abnormal peak point and other peak points.

[0019] All peak points in the illumination intensity variation curve are obtained and used as target peak points. Based on the coordinate points in the illumination intensity variation curve that correspond to the same sampling time as the abnormal peak points, the target peak point closest to the coordinate point is determined as the first target peak point. The first vertical axis span between the first target peak point and its left adjacent coordinate point is obtained in the illumination intensity variation curve. The second vertical axis span between the first target peak point and its right adjacent coordinate point is obtained in the illumination intensity variation curve. The absolute value of the difference between the first vertical axis span and the second vertical axis span is calculated.

[0020] The degree of environmental impact of the abnormal peak point is obtained based on the degree of fluctuation instability of the abnormal peak point, the absolute value of the difference in the vertical axis span, and the degree of abnormality in the morphological characteristics.

[0021] Preferably, the step of obtaining the fluctuation instability degree of the abnormal peak point based on the fluctuation difference between the abnormal peak point and other peak points includes:

[0022] In the power change curve, obtain the peak point that has the same peak value as the abnormal peak point, and form a peak point sequence with the abnormal peak point. Calculate the horizontal axis span between every two adjacent peak points in the peak point sequence to obtain the variance of the horizontal axis span.

[0023] The number of peaks in the peak sequence other than the abnormal peaks is counted, and the product of the reciprocal of the number of peaks and the variance of the horizontal axis span is normalized to obtain the degree of fluctuation instability of the abnormal peaks.

[0024] Preferably, the step of obtaining the environmental impact degree of the abnormal wave peak point based on the fluctuation instability degree, the absolute value of the difference in the vertical axis span, and the degree of morphological anomaly of the abnormal wave peak point includes:

[0025] The absolute value of the difference in the vertical axis span is normalized to obtain the corresponding normalized value. The product of the fluctuation instability of the abnormal peak point, the normalized value, and the abnormality of the morphological characteristics is taken as the degree of environmental impact of the abnormal peak point.

[0026] Preferably, the step of correcting the lag coefficient of the corresponding power data in the power time series data according to the degree of environmental impact of each of the abnormal peak points includes:

[0027] For any abnormal peak point, the environmental impact of the peak point is normalized to obtain the corresponding normalization result. The difference between the constant 1 and the normalization result by a preset multiple is obtained. The power data corresponding to the abnormal peak point in the power time series data is determined as the target data. The product of the original hysteresis coefficient of the target data in the ARIMA model and the difference is used as the corrected hysteresis coefficient of the target data.

[0028] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0029] This invention acquires power data and environmental monitoring data from any distributed photovoltaic (PV) node within a historical time period. Based on the peak shape characteristics of each peak in the power change curve corresponding to the power data, it identifies anomalous peaks with a strong predictive impact. Furthermore, based on the distribution characteristics of the power change curve and the continuous change characteristics of the light intensity change curve corresponding to the environmental monitoring data, it obtains the environmental impact level of each anomalous peak. This adaptively adjusts the ARIMA parameters (hysteresis coefficient) for each power data point, increasing the contribution of stable power data to the ARIMA model training and reducing the contribution of power data with random fluctuations in PV power due to environmental influences. This reduces the negative impact of random data from distributed PV on ARIMA model training, thereby improving the robustness of ARIMA model training and making the predictive ability of the ARIMA model more accurate. Ultimately, this enhances the accuracy of power data prediction using the trained ARIMA model. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a flowchart of a distributed photovoltaic grid-connected control method provided in Embodiment 1 of the present invention;

[0032] Figure 2 This is a schematic diagram of a photovoltaic node equipped with a distribution area fusion terminal provided in an embodiment of the present invention. Detailed Implementation

[0033] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0034] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0035] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0036] The specific scenario addressed by this invention is as follows: When performing grid-connected regulation on one of the multiple distributed photovoltaic nodes in a city, the ARIMA model is used to predict the power generation of the photovoltaic node to facilitate subsequent grid-connected regulation and allocation. However, due to the randomness and volatility of distributed photovoltaic power generation caused by weather conditions, the accuracy of the ARIMA model's prediction results is low, resulting in low accuracy of grid-connected regulation using the prediction results.

[0037] This invention provides a distributed photovoltaic grid-connected control system, including a processor and a memory. The processor executes a computer program stored in the memory to implement a distributed photovoltaic grid-connected control method, such as... Figure 1 As shown, the method includes the following steps:

[0038] Step S101: During the distributed photovoltaic grid-connected control process, acquire the power time-series data and irradiance time-series data of any photovoltaic node in the historical period, and construct the power change curve of the power time-series data and the irradiance change curve of the irradiance time-series data, respectively.

[0039] Each photovoltaic node (distributed photovoltaic node) or a group of photovoltaic nodes is equipped with one such as Figure 2 The integrated photovoltaic (PV) grid connection APP (PV grid connection terminal) shown is used to collect and regulate the grid connection of distributed PV in the area. In this embodiment of the invention, it is necessary to obtain the power data from the collected data of each PV node, as well as the environmental monitoring data of the PV node in the integration terminal. It is worth noting that since PV power generation is mainly affected by sunlight, the sunlight intensity data from the environmental monitoring data is used as the main environmental monitoring data.

[0040] In the process of grid-connected control of distributed photovoltaics, for any photovoltaic node, the power data and irradiance data of that photovoltaic node in the historical period are obtained. The acquisition frequency of irradiance data is the same as that of power data, which is 10Hz. There is no restriction here, and it can be set according to the implementation scenario. In this way, the power data collected in the historical period is combined into a power time series data. Similarly, the irradiance data is combined into an irradiance time series data.

[0041] Furthermore, a two-dimensional planar coordinate system was constructed for the time-series data of electrical power and light intensity, respectively. The horizontal axis is time series, and the vertical axis is electrical power and light intensity, respectively. Correspondingly, the electrical power variation curve of the electrical power time-series data and the light intensity variation curve of the light intensity time-series data were obtained in the two-dimensional planar coordinate system.

[0042] Step S102: Obtain all peak points in the power change curve; based on the morphological characteristics and fluctuation range of each peak point, obtain the degree of abnormality of the morphological characteristics of each peak point; and based on the degree of abnormality of the morphological characteristics of each peak point, obtain abnormal peak points.

[0043] Random data fluctuations inevitably produce peaks in the curve. Therefore, the AMPD peak finding algorithm is used to obtain all peak points in the power change curve. The AMPD peak finding algorithm is existing technology and will not be elaborated upon here. Since the power data itself contains small peaks due to slight fluctuations, these slight fluctuations do not easily change the data's morphological distribution. However, peaks with relatively larger fluctuations have a stronger impact on the prediction results of the ARIMA model. Therefore, in this embodiment of the invention, firstly, based on the morphological characteristics and fluctuation range of each peak point, the degree of morphological anomaly of each peak point is obtained. Then, based on the degree of morphological anomaly of each peak point, abnormal peak points are obtained, thereby filtering out peak points with stronger influence for subsequent analysis.

[0044] Specifically, based on the morphological characteristics and fluctuation range of each wave peak, the degree of morphological anomaly of each wave peak is obtained, including:

[0045] For any peak point, obtain the left and right adjacent coordinate points of the peak point, and obtain the absolute value of the difference in the line slope between the peak point and the left adjacent coordinate point and the right adjacent coordinate point.

[0046] In the power change curve, the lowest valley points on both sides of the peak point are obtained respectively. Based on the vertical axis span between the peak point and each of the lowest valley points, the maximum vertical axis span is obtained. The product between the maximum vertical axis span and the absolute value of the difference in the slope of the connecting line is normalized to obtain the corresponding first normalized value.

[0047] Obtain the horizontal span between the two lowest valley points, normalize the horizontal span to obtain the corresponding second normalized value, and use the sum of the first normalized value and the second normalized value as the degree of morphological anomaly of the peak point.

[0048] In one embodiment, taking the q-th peak as an example, the slope of the line connecting the previous adjacent coordinate point of the q-th peak and the q-th peak is obtained in the power change curve. Simultaneously, obtain the slope of the line connecting the next adjacent coordinate point of the q-th wave crest point to the q-th wave crest point. In the power variation curve, find the lowest valleys on both sides of the q-th peak, with one valley corresponding to each side. Based on the vertical span (i.e., the difference in vertical coordinate values) between the q-th peak and each valley, obtain the maximum vertical span, denoted as . Simultaneously calculate the horizontal span (i.e., the difference in horizontal coordinate values) between the two lowest points, denoted as . Then, based on the slope of the connecting line Slope of the connecting line Maximum vertical span and horizontal axis span To obtain the degree of morphological anomaly at the q-th peak, the expression for calculating the degree of morphological anomaly at the q-th peak is:

[0049]

[0050] in, This indicates the degree of morphological anomaly at the q-th peak, where norm() represents the normalization function, and || represents the absolute value sign. This represents the slope of the line connecting the q-th wave crest point to its preceding adjacent coordinate point. This represents the slope of the line connecting the next adjacent coordinate point of the q-th wave crest to the q-th wave crest. Indicates the maximum vertical span. Indicates the span of the horizontal axis.

[0051] It should be noted that, This is used to reflect the difference in slope before and after the qth peak. The larger the absolute value of this difference, the sharper the qth peak is, and the greater the abnormality of the morphological characteristics of the qth peak is. The larger the value, the steeper the q-th peak; the steeper the peak, the stronger the morphological anomaly of the q-th peak. The larger the value, the greater the proportion of the q-th peak in the power change curve, the stronger its impact on prediction, and the stronger the corresponding morphological anomaly.

[0052] Abnormal peak points are identified based on the degree of morphological anomaly of each peak point, including:

[0053] If the degree of abnormality in the morphological characteristics of any peak point is greater than or equal to a preset threshold for the degree of abnormality in morphological characteristics, then the peak point is regarded as an abnormal peak point.

[0054] In one embodiment, the threshold for the degree of morphological feature anomaly is set to 0.8. This is not limited here and can be set according to the implementation scenario. The threshold is set based on the degree of morphological feature anomaly at any peak point. If a peak is considered to have a strong influence on prediction, it is marked as an abnormal peak. Similarly, each peak is analyzed to obtain all abnormal peaks.

[0055] Step S103: For any abnormal peak point, based on the fluctuation difference between the abnormal peak point and other peak points, and the coordinate point in the light intensity change curve that corresponds to the same sampling time as the abnormal peak point, obtain the degree of environmental impact of the abnormal peak point.

[0056] After identifying anomalous peaks that have a strong impact on subsequent predictions, it is considered that these anomalous peaks may be caused by environmental changes (such as changes in solar intensity or cloud cover during cloudy weather), which are random and highly volatile, making them difficult to predict. Alternatively, they may be caused by the periodic on / off switching of loads in household or industrial applications (such as air conditioner compressors or refrigerator compressors), leading to fluctuations in system current and voltage, thus affecting the output power of the distributed photovoltaic system. These fluctuations exhibit certain regularity and have a positive impact on prediction. Therefore, in this embodiment of the invention, the environmental impact of each anomalous peak can be analyzed based on the differences in the distribution characteristics of the anomalous peaks in the power curve and in conjunction with the changing characteristics of the solar intensity curve.

[0057] In cloudy weather, due to cloud cover, the intensity of sunlight received by photovoltaic systems fluctuates, leading to significant random fluctuations in electrical power. This results in varying peak values ​​in the power change curve, with different intervals between peaks caused by changes in sunlight. Conversely, power changes caused by the switching of a fixed load are uniform, resulting in peak values ​​of similar magnitudes in the corresponding power change curve, and similar intervals between peaks caused by periodic load switching. Therefore, for any abnormal peak, the more peaks with the same peak value and the higher the similarity of intervals between these peaks, the more likely it is a peak point with a stable load. Conversely, the fewer peaks with the same peak value and the lower the similarity of intervals between abnormal peaks with the same peak value, the more likely it is a peak point caused by unstable environmental changes. The degree of instability of the abnormal peak point can then be determined based on the fluctuation differences between it and other peaks.

[0058] The method for obtaining the fluctuation instability of abnormal peak points is as follows: obtain peak points with the same peak value as the abnormal peak points in the power change curve, and form a peak point sequence with the abnormal peak points. Calculate the horizontal axis span between every two adjacent peak points in the peak point sequence to obtain the variance of the horizontal axis span.

[0059] The number of peaks in the peak sequence other than the abnormal peaks is counted, and the product of the reciprocal of the number of peaks and the variance of the horizontal axis span is normalized to obtain the degree of fluctuation instability of the abnormal peaks.

[0060] In one embodiment, taking the p-th abnormal peak as an example, the number of peaks in the power change curve that have the same peak value as the p-th abnormal peak is obtained and denoted as . These peak points are then combined with the p-th abnormal peak point to form a peak point sequence. The horizontal axis interval distance (difference of horizontal coordinate values) between every two adjacent peak points in the sequence is obtained sequentially, and the variance of the horizontal axis interval distance is denoted as . According to the number of peaks variance of the distance between the horizontal axis and the horizontal axis To obtain the fluctuation instability degree of the p-th abnormal peak point, the expression for calculating the fluctuation instability degree of the p-th abnormal peak point is:

[0061]

[0062] in, This indicates the degree of instability of the fluctuation at the p-th abnormal peak point. This represents the number of peaks that have the same peak value as the p-th abnormal peak. The variance of the horizontal axis interval between any two adjacent peaks in the peak sequence is represented by norm(), which represents the normalization function.

[0063] It should be noted that, The smaller the value, the stronger the randomness of the fluctuation of the p-th abnormal peak point caused by environmental changes, and the greater the instability of the fluctuation of the p-th abnormal peak point. The larger the value, the greater the difference in the interval distance between the p-th abnormal peak point and other peak points with the same peak value. This indicates that the p-th abnormal peak point has stronger fluctuation characteristics caused by environmental changes and stronger instability.

[0064] To more accurately determine whether anomalous peaks are caused by environmental changes and to avoid reducing the predictive contribution of anomalous peaks not caused by environmental changes, this study combines the characteristics of environmental changes with the degree of instability of each anomalous peak to obtain the final environmental impact of each peak. Since the fluctuations in electrical power data caused by environmental changes, specifically changes in light intensity, occur when clouds obstruct the data, the same time as the power fluctuations is accompanied by changes in light intensity data collected by the environmental monitor. For anomalous peaks with high instability, the greater the change in light intensity at the same time, the more likely they are unstable peaks caused by abnormal environmental interference. Conversely, for peaks with high stability, the greater the change in light intensity at the same time, they may also be peaks similar to load peaks caused by light changes. Therefore, for any abnormal peak point, considering the slight time delay between the change in irradiance monitored by the irradiance monitoring device and the power output of the photovoltaic system caused by the change in irradiance, all peak points in the irradiance change curve are first acquired and used as target peak points. Based on the coordinate points in the irradiance change curve that correspond to the same sampling time as the abnormal peak point, the target peak point closest to the original coordinate point is determined as the first target peak point. The first vertical axis span between the first target peak point and its left adjacent coordinate point is obtained in the irradiance change curve. The second vertical axis span between the first target peak point and its right adjacent coordinate point is also obtained in the irradiance change curve. The absolute value of the difference between the first and second vertical axis spans is calculated and denoted as [the value is missing here]. .

[0065] Then, based on the degree of fluctuation instability of the abnormal peak point, the absolute value of the difference in the vertical axis span, and the degree of abnormality in the morphological features, the degree of environmental impact of the abnormal peak point is obtained. Specifically, the absolute value of the difference in the vertical axis span is normalized to obtain the corresponding normalized value, and the product of the degree of fluctuation instability of the abnormal peak point, the normalized value, and the degree of abnormality in the morphological features is taken as the degree of environmental impact of the abnormal peak point.

[0066] In one embodiment, the expression for calculating the environmental impact of the p-th abnormal peak point is:

[0067]

[0068] in, This indicates the degree of environmental impact of the p-th abnormal peak point. This represents the absolute value of the difference in the vertical span between the p-th anomalous peak point and its left and right adjacent coordinate points in the illumination intensity variation curve. This indicates the degree of instability of the fluctuation at the p-th abnormal peak point. This indicates the degree of morphological abnormality of the p-th abnormal peak.

[0069] It should be noted that, The larger the value, the greater the change in light intensity at the p-th anomalous peak point, and the greater the environmental impact of the p-th anomalous peak point; the greater the abnormality of the morphological characteristics and the greater the instability of the fluctuations at the p-th anomalous peak point, the greater the environmental impact of the p-th anomalous peak point and the greater the influence of light intensity.

[0070] Step S104: Obtain the environmental impact degree of each of the abnormal peak points. During the training of the ARIMA model using the power time series data, the lag coefficient of the corresponding power data in the power time series data is corrected according to the environmental impact degree of each of the abnormal peak points to obtain the trained ARIMA model. The trained ARIMA model is then used to perform grid-connected regulation of the photovoltaic node.

[0071] Following the method described above for obtaining the environmental impact level of the p-th abnormal peak point, the environmental impact level of each abnormal peak point in the power variation curve is obtained. The greater the environmental impact level, the greater the impact of the corresponding abnormal peak point on the prediction results of the ARIMA model. Therefore, when training the ARIMA model using power time series data, the lag coefficient of each abnormal peak point is corrected according to the environmental impact level of each abnormal peak point. The specific correction method is as follows:

[0072] For any abnormal peak point, the environmental impact of the peak point is normalized to obtain the corresponding normalization result. The difference between the constant 1 and the normalization result by a preset multiple is obtained. The power data corresponding to the abnormal peak point in the power time series data is determined as the target data. The product of the original hysteresis coefficient of the target data in the ARIMA model and the difference is used as the corrected hysteresis coefficient of the target data.

[0073] In one embodiment, the corrected expression for the lag coefficient is:

[0074]

[0075] in, This represents the hysteresis coefficient after correction for the p-th abnormal peak. This represents the hysteresis coefficient before correction for the p-th anomalous peak, which is also the original hysteresis coefficient in the ARIMA model. 1 indicates a constant, 0.5 indicates a preset multiple, and norm() represents the normalization function. This indicates the degree of environmental impact of the p-th abnormal peak point.

[0076] It should be noted that, The linear normalization control magnitude is in the range [0, 1]. The correction level is controlled at 0.5, but can be adjusted according to the specific needs of the scenario. The stronger the environmental impact of the p-th abnormal peak point, the worse the stability of the corresponding p-th abnormal peak point, and the stronger the interference with the prediction. Therefore, the greater the reduction in the lag coefficient of the corresponding p-th abnormal peak point.

[0077] The lag coefficient is a model parameter in the ARIMA model training process. The method for obtaining it and the training process of the ARIMA model are existing technologies. However, in the traditional ARIMA model training process, the environmental influence of abnormal peak points in the power time series data is used to correct the original lag coefficient of the power data corresponding to the abnormal peak points in the ARIMA model training process. For normal data points in the power time series data, the original lag coefficient is retained and no correction is required, thereby obtaining a trained ARIMA model.

[0078] After obtaining the trained ARIMA model, the grid-connected control of photovoltaic nodes can be carried out using the trained ARIMA model. Specifically, the trained ARIMA model is used to predict the power data in real time, and the predicted power value is obtained within a certain time range. The grid-connected output of the photovoltaic nodes is adjusted using the predicted power value. When the predicted power is higher, more power is input into the grid, and when the predicted power is lower, the power output is reduced.

[0079] It should be noted that the focus of this embodiment of the invention is on how to obtain a trained ARIMA model. Using a trained ARIMA model to perform grid-connected regulation of photovoltaic nodes is an existing technology and will not be described in detail here.

[0080] In summary, this embodiment of the invention acquires the power data and environmental monitoring data of any distributed photovoltaic node within a historical time period. Based on the peak shape characteristics of each peak point in the power change curve corresponding to the power data, it obtains abnormal peak points with a strong influence on prediction. Furthermore, based on the distribution characteristics of the power change curve and the continuous change characteristics of the light intensity change curve corresponding to the environmental monitoring data, it obtains the degree of environmental impact of each abnormal peak point. This allows for adaptive adjustment of the ARIMA parameters (hysteresis coefficient) of each power data point, increasing the contribution of stable power data to the training of the ARIMA model and reducing the contribution of power data with random fluctuations in photovoltaic power due to environmental influences to the training of the ARIMA model. This reduces the negative impact of random data from distributed photovoltaics on the training of the ARIMA model, thereby improving the robustness of the ARIMA model training and making the prediction ability of the ARIMA model more accurate. Ultimately, this improves the accuracy of power data prediction using the trained ARIMA model.

[0081] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A distributed photovoltaic grid-connected control system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it performs the following steps: In the process of grid-connected control of distributed photovoltaics, the power time-series data and irradiance time-series data of any photovoltaic node in the historical period are obtained, and the power change curve of the power time-series data and the irradiance change curve of the irradiance time-series data are constructed respectively. All peaks in the power change curve are obtained. Based on the morphological characteristics and fluctuation range of each peak, the degree of abnormality of the morphological characteristics of each peak is obtained. Based on the degree of abnormality of the morphological characteristics of each peak, abnormal peaks are obtained. For any abnormal peak point, the degree of environmental impact of the abnormal peak point is obtained based on the fluctuation difference between the abnormal peak point and other peak points, and the coordinate point in the light intensity change curve that corresponds to the same sampling time as the abnormal peak point. The environmental impact level of each abnormal peak point is obtained. During the training of the ARIMA model using the power time series data, the lag coefficient of the corresponding power data in the power time series data is corrected according to the environmental impact level of each abnormal peak point to obtain the trained ARIMA model. The trained ARIMA model is then used to perform grid-connected regulation of the photovoltaic node.

2. The distributed photovoltaic grid-connected control system according to claim 1, characterized in that, The step of obtaining the degree of morphological anomaly of each wave crest point based on its morphological characteristics and fluctuation range includes: For any peak point, obtain the left and right adjacent coordinate points of the peak point, and obtain the absolute value of the difference in the line slope between the peak point and the left adjacent coordinate point and the right adjacent coordinate point. In the power change curve, the lowest valley points on both sides of the peak point are obtained respectively. Based on the vertical axis span between the peak point and each of the lowest valley points, the maximum vertical axis span is obtained. The product between the maximum vertical axis span and the absolute value of the difference in the slope of the connecting line is normalized to obtain the corresponding first normalized value. Obtain the horizontal span between the two lowest valley points, normalize the horizontal span to obtain the corresponding second normalized value, and use the sum of the first normalized value and the second normalized value as the degree of morphological anomaly of the peak point.

3. The distributed photovoltaic grid-connected control system according to claim 1, characterized in that, The step of obtaining abnormal peak points based on the degree of abnormality in the morphological characteristics of each peak point includes: If the degree of abnormality in the morphological characteristics of any peak point is greater than or equal to a preset threshold for the degree of abnormality in morphological characteristics, then the peak point is regarded as an abnormal peak point.

4. A distributed photovoltaic grid-connected control system according to claim 1, characterized in that, The step of obtaining the environmental impact degree of the abnormal peak point based on the fluctuation difference between the abnormal peak point and other peak points, and the coordinate points in the illumination intensity change curve corresponding to the same sampling time as the abnormal peak point, includes: The degree of fluctuation instability of the abnormal peak point is obtained based on the fluctuation difference between the abnormal peak point and other peak points. All peak points in the illumination intensity variation curve are obtained and used as target peak points. Based on the coordinate points in the illumination intensity variation curve that correspond to the same sampling time as the abnormal peak points, the target peak point closest to the coordinate point is determined as the first target peak point. The first vertical axis span between the first target peak point and its left adjacent coordinate point is obtained in the illumination intensity variation curve. The second vertical axis span between the first target peak point and its right adjacent coordinate point is obtained in the illumination intensity variation curve. The absolute value of the difference between the first vertical axis span and the second vertical axis span is calculated. The degree of environmental impact of the abnormal peak point is obtained based on the degree of fluctuation instability of the abnormal peak point, the absolute value of the difference in the vertical axis span, and the degree of abnormality in the morphological characteristics.

5. A distributed photovoltaic grid-connected control system according to claim 4, characterized in that, The step of obtaining the degree of fluctuation instability of the abnormal peak point based on the fluctuation difference between the abnormal peak point and other peak points includes: In the power change curve, obtain the peak point that has the same peak value as the abnormal peak point, and form a peak point sequence with the abnormal peak point. Calculate the horizontal axis span between every two adjacent peak points in the peak point sequence to obtain the variance of the horizontal axis span. The number of peaks in the peak sequence other than the abnormal peaks is counted, and the product of the reciprocal of the number of peaks and the variance of the horizontal axis span is normalized to obtain the degree of fluctuation instability of the abnormal peaks.

6. A distributed photovoltaic grid-connected control system according to claim 4, characterized in that, The step of obtaining the environmental impact degree of the abnormal wave peak point based on the fluctuation instability degree, the absolute value of the difference in the vertical axis span, and the degree of morphological anomaly of the abnormal wave peak point includes: The absolute value of the difference in the vertical axis span is normalized to obtain the corresponding normalized value. The product of the fluctuation instability of the abnormal peak point, the normalized value, and the abnormality of the morphological characteristics is taken as the degree of environmental impact of the abnormal peak point.

7. A distributed photovoltaic grid-connected control system according to claim 1, characterized in that, The step of correcting the lag coefficient of the corresponding power data in the power time series data according to the degree of environmental impact of each of the abnormal peak points includes: For any abnormal peak point, the environmental impact of the peak point is normalized to obtain the corresponding normalization result. The difference between the constant 1 and the normalization result by a preset multiple is obtained. The power data corresponding to the abnormal peak point in the power time series data is determined as the target data. The product of the original hysteresis coefficient of the target data in the ARIMA model and the difference is used as the corrected hysteresis coefficient of the target data.

Images