An energy scheduling and energy-saving optimization method for smart grid

Abstract

The present application relates to the technical field of data processing, in particular to a kind of energy scheduling and energy-saving optimization method of smart grid;Information richness is obtained according to the frequency spectrum data of power generation time sequence;Adaptive neighborhood window is obtained according to frequency spectrum data and information richness;Power generation reference weight is obtained according to time interval characteristics;Power generation similarity is obtained according to the power generation difference characteristics of any time and other dates at the same time in adaptive neighborhood window, power generation reference weight;Variation difference degree is obtained according to the power generation change difference characteristics of any time and other times in adaptive neighborhood window, time interval characteristics.The present application obtains the robustness weight of any time according to power generation similarity and variation difference degree and carries out time sequence decomposition, carries out prediction and fusion according to decomposition result, carries out electric energy scheduling according to power generation prediction data, improves the accuracy of data decomposition and the rationality of electric energy scheduling.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to an energy scheduling and energy saving optimization method of smart grid. BACKGROUND

[0002] With the continuous expansion of the scale of clean energy power stations, although it has the advantages of low cost and environmental protection for power supply, there are intermittent and fluctuating conditions in the power generation process, for example, wind power is affected by changes in wind direction and wind speed, and photovoltaic power generation is affected by sunlight intensity; the instability of clean energy factors brings certain challenges to the energy scheduling of the power grid.

[0003] In order to improve the matching stability of the power supply side and the load side in the power grid, and avoid the situation of insufficient and waste of energy supply, the prior art uses STL time series decomposition algorithm to decompose the power generation data of the power station, and uses LSTM long short-term memory neural network to predict the seasonal term, trend term and residual term after decomposition respectively, and fuses the prediction results to obtain more accurate power generation prediction data in the future. According to the prediction data, the power scheduling is carried out, and the matching degree of the power supply side and the load side is improved. The robustness weight value of each data point directly affects the final decomposition result during STL decomposition, but the power generation of clean energy is often affected by environmental factors, resulting in that the power generation data at some time may change greatly. If the compliance degree of the robustness weight at any time and the power generation characteristics is poor, it is easy to cause the accuracy of the decomposition result to be low, and finally affect the rationality of the power generation prediction result and the power scheduling. SUMMARY

[0004] In order to solve the technical problem that the robustness weight of each time of the power generation time series of the clean energy in the STL decomposition affects the accuracy of the decomposition result, thereby causing the rationality of the power generation prediction result and the power scheduling to be low, the purpose of the present application is to provide an energy scheduling and energy saving optimization method of smart grid, and the technical scheme adopted is as follows:

[0005] Obtain the power generation time series of the power station;

[0006] Obtain the information richness according to the amplitude difference characteristics of different frequencies in the frequency spectrum data of the power generation time series; and obtain the adaptive neighborhood window according to the frequency spectrum data and the information richness;

[0007] Obtain the power generation reference weight according to the time interval characteristics of any time and the same time of other dates in the power generation time series; obtain the power generation similarity according to the power generation difference characteristics of the any time and the same time of other dates within the adaptive neighborhood window, and the power generation reference weight of the same time of the other dates; and obtain the change difference degree according to the difference characteristics of the power generation change of the any time and other times within the adaptive neighborhood window, and the time interval characteristics.

[0008] obtaining the robustness weight of the arbitrary moment according to the power generation similarity and the change difference degree; decomposing the robustness weight of all moments in the power generation time sequence through a time sequence decomposition algorithm, and obtaining power generation prediction data through prediction and fusion according to the decomposition result; and performing power scheduling according to the power generation prediction data.

[0009] Further, the step of obtaining the information richness according to the amplitude difference characteristics of different frequencies in the frequency spectrum data of the power generation time sequence comprises:

[0010] In the formula, R represents the information richness, G represents the number of amplitudes whose amplitudes are not a constant 0 in the frequency spectrum data, e represents a natural constant, F max represents the maximum amplitude in the frequency spectrum data, F g represents the gth amplitude whose amplitude is not a constant 0 and is not the maximum value in the frequency spectrum data, represents the information contribution degree of the gth amplitude, H max represents the frequency corresponding to the maximum amplitude in the frequency spectrum data, H g represents the frequency corresponding to the gth amplitude, |H max -H g represents the frequency difference degree.

[0011] Further, the step of obtaining the adaptive neighborhood window according to the frequency spectrum data and the information richness comprises:

[0012] calculating the ratio of the preset sampling frequency of the power generation time sequence to the frequency corresponding to the maximum amplitude in the frequency spectrum data to obtain a range reference; calculating the product of the information richness after negative correlation mapping through a preset function and the range reference and taking the nearest odd number to obtain a data length; and taking an arbitrary analysis moment in the power generation time sequence as the center of the adaptive neighborhood window and taking the data length as the window length of the adaptive neighborhood window to obtain the adaptive neighborhood window of the arbitrary analysis moment.

[0013] Further, the step of obtaining the power generation reference weight according to the time interval characteristics of the arbitrary moment in the power generation time sequence and the same moment of other dates comprises:

[0014] calculating the reciprocal of the time interval of the arbitrary moment and the same moment of any other date to obtain a first value; calculating the sum of all first values corresponding to the arbitrary moment to obtain a second value; and calculating the ratio of the first value to the second value to obtain the power generation reference weight of the same moment of the arbitrary other date to the arbitrary moment.

[0015] Further, the step of obtaining the power generation similarity according to the difference feature of the power generation at the same time of the arbitrary time and other dates in the adaptive neighborhood window, and the power generation reference weight at the same time of the other dates comprises:

[0016] In the formula, Q represents the power generation similarity of the arbitrary time, N represents the number of the same time of the other dates of the arbitrary time, e represents a natural constant, Y represents the power generation time sequence segment in the adaptive neighborhood window of the arbitrary time, X n represents the power generation time sequence segment in the adaptive neighborhood window of the nth same time of the other dates, DTW(Y,X n ) represents the dynamic time warping distance between Y and X n , and represents the change similarity; T n represents the power generation reference weight of the nth same time of the other dates.

[0017] Further, the step of obtaining the change difference degree according to the difference feature of the power generation change of the arbitrary time and other times in the adaptive neighborhood window, and the time interval feature comprises:

[0018] In the formula, W represents the change difference degree, M represents the number of times in the adaptive neighborhood window, e represents a natural constant, K represents the sum value of the power generation difference values of the arbitrary time and adjacent times; L m represents the sum value of the power generation difference values of the mth other time and adjacent times in the adaptive neighborhood window of the arbitrary time; represents the change difference feature value; D m represents the reciprocal of the time interval of the arbitrary time and the mth other time.

[0019] Further, the step of obtaining the robustness weight of the arbitrary time according to the power generation similarity and the change difference degree comprises:

[0020] calculating the sum value of the change difference degrees of all times in the adaptive neighborhood window of the arbitrary time to obtain a fourth value; calculating the ratio of the change difference degree of the arbitrary time and the fourth value to obtain a change difference reference value; calculating the difference value between the constant 1 and the change difference reference value to obtain a change normality; and calculating the product of the power generation similarity and the change normality and normalizing to obtain the robustness weight of the arbitrary time.

[0021] The present application has the following beneficial effects:

[0022] In the present application, the information richness can represent the frequency information richness of the power generation time sequence, and then the adaptive neighborhood window of power generation data analysis can be accurately determined according to the information richness and the frequency data, the local characteristic change of the power generation data in the time sequence can be more accurately analyzed, and the accuracy of the robustness weight in the decomposition process at each time is improved. Since the clean energy power station has seasonality in the power generation process, the power generation reference weight can be obtained to determine the reliability of the similar features of the power generation between different time, and further improve the accuracy of the robustness weight. The power generation similarity can represent the degree to which the power generation characteristics at the arbitrary time meet the normal power generation, so as to determine the size of the robustness weight. The change difference degree can represent the difference degree of the local change characteristics of the power generation at different times in the adaptive neighborhood window, and reduce the error of the robustness weight. Finally, the power generation time sequence is decomposed according to the robustness weight of all times, which can improve the decomposition accuracy and make the power generation prediction data more accurate. The rationality of the electric energy scheduling is improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, below, a brief introduction will be given to the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0024] Figure 1 A flow chart of an energy scheduling and energy saving optimization method of a smart grid provided by an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the specific implementation, structure, features and effects of the energy scheduling and energy saving optimization method of a smart grid according to the present application are described in detail as follows by combining with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0027] The specific scheme of the energy scheduling and energy saving optimization method of a smart grid provided by the present application is specifically described below in combination with the drawings.

[0028] Please refer toFigure 1 Fig. 9 shows a flow chart of a method for energy scheduling and energy saving optimization of a smart grid according to an embodiment of the present application, which comprises the following steps:

[0029] In step S1, the power generation time sequence of the power station is obtained.

[0030] In the embodiment of the present application, the implementation scenario is to perform energy scheduling on the clean energy power station to improve the energy saving and rationality of energy scheduling. First, the power generation time sequence of the power station is obtained, and a history power generation time sequence of any clean energy power station before the current time is obtained. In the embodiment of the present application, 30 days of power generation data is collected, and the collection frequency is once per second. The implementer can determine it according to the implementation scenario.

[0031] In step S2, the information richness is obtained according to the amplitude difference characteristics of different frequencies in the frequency spectrum data of the power generation time sequence, and the adaptive neighborhood window is obtained according to the frequency spectrum data and the information richness.

[0032] When the power grid performs energy scheduling and energy saving optimization, the power generation data of the power station is usually predicted, and then the predicted power generation and the load demand in the power grid are matched and scheduled. In order to improve the rationality of energy scheduling, the accuracy of the power generation prediction result is particularly important. Since the stability of the clean energy power station is low, for example, the photovoltaic power station in different time periods of a day, due to the difference in sunlight intensity, the power generation exists volatility and periodicity, which increases the difficulty of power generation prediction. In order to improve the accuracy of the prediction result, the power generation time sequence can be decomposed by the existing STL time sequence decomposition algorithm to obtain the seasonal term, the trend term and the residual term. Each term is predicted and fused to obtain the final power generation prediction result. In the STL decomposition process, each data point has a corresponding robustness weight, which can reduce the influence of outliers on the decomposition result. The power generation of clean energy is easily affected by environmental factors, resulting in different mutation characteristics in the power generation time sequence. The accuracy of the existing residual method for obtaining the robustness weight of the data point is low, so it is necessary to combine the data characteristics of the power generation time sequence to improve the rationality of the robustness weight of each data point and improve the accuracy of decomposition and prediction.

[0033] Further, when the time series data is decomposed, the change characteristics of the time series data need to be analyzed in connection with the data before and after it, in combination with the change law of the data before and after it. The robustness weight of the data point should also be determined in combination with the change law of the data. Because the change law and trend of the power generation of different power stations are different, when the power generation time series of different power stations is decomposed by STL, the reference range of the robustness weight of the data point also has differences. In order to more accurately obtain the robustness weight of the power station in decomposition, first, a suitable analysis window range is selected according to the information characteristics in the power generation time series of the power station; therefore, the information richness is obtained according to the amplitude difference characteristics of different frequencies in the frequency spectrum data of the power generation time series; it should be noted that the frequency spectrum data of the power generation time series is obtained by fast Fourier transform, and the specific steps are not repeated; preferably, in the embodiment of the present application, the step of obtaining the information richness comprises:

[0034]

[0035] In the formula, R represents information richness, G represents the number of amplitudes whose amplitudes are not a constant 0 in the frequency spectrum data, e represents a natural constant, F max represents the maximum amplitude in the frequency spectrum data, F g represents the gth amplitude whose amplitude is not a constant 0 and is not the maximum value in the frequency spectrum data, represents the information contribution degree of the gth amplitude, H max represents the frequency corresponding to the maximum amplitude in the frequency spectrum data, H g represents the frequency corresponding to the gth amplitude, |H max -H g represents the frequency difference degree. H max is the main frequency of the current signal, which represents the overall change characteristics, and the remaining frequencies represent the local fluctuations superimposed on the overall change characteristics; if the amplitudes of the remaining frequency signals are higher and the difference from the main frequency is greater, that is, the information contribution degree and the frequency difference degree are greater, it means that the frequency information of the current signal is richer, and the fluctuation characteristics are more obvious, and then when analyzing the robustness weight, a smaller window range should be selected, which can more carefully analyze the local characteristic change of the power generation time series.

[0036] Further, when the information richness is greater, it indicates that the frequency information of the power generation time sequence is richer, so that a smaller window range should be selected when analyzing the robustness weight, and the local feature change of the power generation time sequence can be analyzed in more detail. Therefore, the adaptive neighborhood window is obtained according to the spectrum data and the information richness; preferably, in the embodiment of the present application, the step of obtaining the adaptive neighborhood window comprises: calculating the ratio of the preset sampling frequency of the power generation time sequence to the frequency corresponding to the maximum amplitude in the spectrum data to obtain a range reference; in the embodiment of the present application, the preset sampling frequency is 1 Hz, and the range reference represents the number of power generation data points contained in a periodic signal, and then the range reference is taken as a reference value for range determination. After the information richness is negatively correlated and mapped by a preset function, the product of the range reference is calculated and the nearest odd number is taken to obtain the data length; in the embodiment of the present application, the preset function is In the formula, S is the mapping value, e represents the natural constant, R represents the information richness, the mapping range is (0, 3), and the implementer can determine it by himself according to the implementation scene. When the information richness is greater, the window range should be smaller, and the data length should be smaller. Any analysis time in the power generation time sequence is taken as the center of the adaptive neighborhood window, the analysis time is the time for analyzing the robustness weight, the data length is taken as the window length of the adaptive neighborhood window, and the adaptive neighborhood window of the analysis time is obtained. Thus, the adaptive neighborhood window of each time in the power generation time sequence is obtained, and the analysis of the robustness weight can be performed based on the data in the adaptive neighborhood window.

[0037] In step S3, the power generation reference weight is obtained according to the time interval feature of the analysis time and the same time of other dates in the power generation time sequence; the power generation similarity is obtained according to the power generation difference feature of the analysis time and the same time of other dates in the adaptive neighborhood window, and the power generation reference weight of the same time of other dates; and the change difference degree is obtained according to the power generation change difference feature of the analysis time and other times in the adaptive neighborhood window, and the time interval feature.

[0038] Because the power generation features of the daily same period of the clean energy power station have strong correlation rules, for example, the solar irradiation angle at noon is similar when the photovoltaic power generation is performed, so the power generation and the change trend at the noon period are also similar. Then, the power generation features of the daily same period can be analyzed to determine whether the change rules are similar; however, because the natural environment and other factors have seasonal changes in time sequence, the two same periods with a relatively short time interval have a higher reference weight when the power generation features are analyzed; therefore, the power generation reference weight is obtained according to the time interval feature of the analysis time and the same time of other dates in the power generation time sequence.

[0039] Preferably, in the embodiments of the present application, the step of obtaining the power generation reference weight comprises: calculating the reciprocal of the time interval between the time point of the arbitrary time point and the time point of the arbitrary other date, to obtain a first value; the closer the time interval, the greater the first value, meaning that the time interval between the time point of the arbitrary time point and the time point of the arbitrary other date is closer, and the reliability of the power generation amount feature analysis result of the two is higher. Calculate the sum of all first values corresponding to the arbitrary time point, to obtain a second value; the purpose is to make the sum of the power generation reference weights of the time points of all other dates be 1. Calculate the ratio of the first value to the second value, to obtain the power generation reference weight of the time point of the arbitrary other date to the time point of the arbitrary time point, and the sum of all power generation reference weights is 1.

[0040] Further, if the similarity of the change trend of the power generation amount data of the adaptive neighborhood window of the arbitrary time point and the adaptive neighborhood window of the time point of all other dates is high, it means that the power generation amount feature of the arbitrary time point is close to the data in the normal power generation process. Therefore, the power generation similarity can be obtained according to the power generation difference feature of the adaptive neighborhood window of the time point of the arbitrary time point and the time point of other dates, and the power generation reference weight of the time point of other dates; Preferably, in the embodiments of the present application, the step of obtaining the power generation similarity comprises:

[0041]

[0042] In the formula, Q represents the power generation similarity of the arbitrary time point, N represents the number of the time points of other dates of the arbitrary time point, e represents the natural constant, Y represents the power generation time sequence segment in the adaptive neighborhood window of the arbitrary time point, X n represents the power generation time sequence segment in the adaptive neighborhood window of the nth time point of other dates, DTW(Y, X n ) represents the dynamic time warping distance between Y and X n , and it should be noted that the dynamic time warping distance is obtained by the existing dynamic time warping algorithm, and the more similar the change trends of the two sequences, the smaller the dynamic time warping distance. represents the change similarity, and the dynamic time warping distance is negatively correlated by the exponential function, and the greater the change similarity means the more similar the change trends of the two time periods; T n represents the power generation reference weight of the nth time point of other dates, and the greater the power generation reference weight means the closer the time interval of the two same time periods, and the greater the reliability of the change similarity. Further, the greater the power generation similarity means that the power generation amount feature of the arbitrary time point is more consistent with the normal power generation situation.

[0043] Further, the greater the power generation similarity at the arbitrary time means that the power generation feature at the arbitrary time is more normal, and then the greater the robustness weight of the power generation data at the arbitrary time in STL decomposition; but since the power generation similarity is obtained according to the power generation data at all times in the adaptive neighborhood window, if the data at other times in the adaptive neighborhood window of the arbitrary time has a mutation feature, the power generation similarity at the arbitrary time will be low, thereby affecting the accuracy of the robustness weight. Therefore, the difference degree of the power generation local variation feature at the arbitrary time and the power generation local variation feature at other times is obtained, and the robustness weight is further obtained according to the difference degree; and then the difference degree of variation is obtained according to the difference feature of the power generation variation at the arbitrary time and other times in the adaptive neighborhood window and the time interval feature; preferably, in the embodiment of the application, the step of obtaining the difference degree of variation comprises:

[0044]

[0045] In the formula, W represents the difference degree of variation, M represents the number of times in the adaptive neighborhood window, e represents a natural constant, and K represents the sum of the power generation difference values at the arbitrary time and the adjacent times; the greater the K is from 0, the more likely the power generation at the arbitrary time has a mutation feature. m L represents the sum of the power generation difference values at the mth other time in the adaptive neighborhood window of the arbitrary time and the adjacent times; the greater the L is from 0, the more likely the power generation at the other time has a mutation feature. m D represents the difference degree of variation feature value; the greater the difference degree of variation feature value is, the greater the difference between the power generation local variation features of the two times is. m D represents the reciprocal of the time interval between the arbitrary time and the mth other time; since the closer the time interval is, the more similar the power generation variation features are, D m is greater, and the corresponding difference degree of variation feature value has a higher credibility. The difference degree of variation represents the difference degree of the power generation local variation features between the arbitrary time and other times in the adaptive neighborhood window; the greater the difference degree of variation is, the greater the difference between the power generation local variation features is.

[0046] Step S4: obtaining the robustness weight of the arbitrary time according to the power generation similarity and the difference degree of variation; decomposing all times in the power generation time sequence through a time sequence decomposition algorithm according to the robustness weight of the times, and obtaining power generation prediction data according to the decomposition result for prediction and fusion; and performing electric energy scheduling according to the power generation prediction data.

[0047] ​After the power generation similarity and the change difference degree are obtained, the robustness weight of an arbitrary moment can be obtained according to the power generation similarity and the change difference degree; preferably, in the embodiment of the present application, the step of obtaining the robustness weight of the arbitrary moment comprises: calculating the sum value of the change difference degrees of all moments in the adaptive neighborhood window of the arbitrary moment to obtain a fourth numerical value; calculating the ratio of the change difference degree of the arbitrary moment to the fourth numerical value to obtain a change difference reference value. The change difference reference value represents the proportion of the change difference degree of the arbitrary moment in the fourth numerical value, and the smaller the change difference degree of the arbitrary moment is, the smaller the change difference reference value is, which means that the local change feature of the power generation of other moments in the adaptive neighborhood window of the arbitrary moment is more obvious, and there is a certain mutation feature, while the local change feature of the power generation of the arbitrary moment is smaller; therefore, the mutation feature of other moments will cause the power generation similarity and the robustness weight of the arbitrary moment to be smaller, which affects the decomposition accuracy; therefore, the smaller the change difference reference value is, the larger the robustness weight of the arbitrary moment should be. The difference between the constant 1 and the change difference reference value is calculated to obtain a change normality; the smaller the change difference reference value is, the larger the change normality is. The product of the power generation similarity and the change normality is calculated and normalized to obtain the robustness weight of the arbitrary moment; the larger the power generation similarity and the change normality are, which means that the power generation feature of the arbitrary moment is more in line with the normal power generation trend, and the larger the robustness weight of the arbitrary moment is.

[0048] Further, after obtaining the robustness weight of all moments in the power generation time sequence, the robustness weight of all moments in the power generation time sequence can be decomposed by a time sequence decomposition algorithm, prediction and fusion are performed according to the decomposition result to obtain power generation prediction data; it should be noted that the STL time sequence decomposition algorithm belongs to the prior art, and the specific decomposition steps will not be described again; the robustness weight obtained according to the data characteristics of the power generation time sequence is more accurate than the robustness weight obtained according to the residual characteristics, the volatility and periodic variation characteristics of the power generation of the clean energy power station are considered in the process of obtaining the robustness weight, and finally the accuracy of the decomposition result is improved. After decomposition, the seasonal term, the trend term and the residual term are obtained, in the embodiment of the application, the seasonal term, the trend term and the residual term are respectively predicted by using the existing LSTM long short-term memory neural network, so as to improve the prediction accuracy; it should be noted that the LSTM neural network belongs to the prior art, and the specific prediction steps will not be described again. All prediction results are fused to obtain power generation prediction data; finally, power scheduling is performed according to the power generation prediction data, the power generation prediction data and the power consumption of different power grid regions are matched, the region with the closest power generation and power consumption is selected for power transmission, so as to avoid waste of electric energy, and the implementer can set the rules of electric energy scheduling according to the power generation prediction data, which is not limited herein. At this point, by analyzing the historical power generation time sequence, the robustness weight of each moment is obtained, and the accuracy of the power generation time sequence decomposition and prediction result is improved, so that the rationality of the power scheduling is higher.

[0049] In summary, the embodiment of the application provides an energy scheduling and energy saving optimization method for a smart grid; information richness is obtained according to the frequency spectrum data of the power generation time sequence; an adaptive neighborhood window is obtained according to the frequency spectrum data and the information richness; a power generation reference weight is obtained according to the time interval characteristics; a power generation similarity is obtained according to the power generation difference characteristics of any moment and other moments at the same moment in the adaptive neighborhood window, and the power generation reference weight; and a change difference degree is obtained according to the power generation change difference characteristics of any moment and other moments in the adaptive neighborhood window, and the time interval characteristics. The robustness weight of any moment is obtained according to the power generation similarity and the change difference degree, and time sequence decomposition is performed, prediction and fusion are performed according to the decomposition result, power scheduling is performed according to the power generation prediction data, the accuracy of data decomposition and the rationality of power scheduling are improved.

[0050] It should be noted that: the above-mentioned embodiment of the application is only for description, and does not represent the advantages and disadvantages of the embodiment. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0051] The various embodiments described in this specification are presented by way of example, and each embodiment is not inherently more important than any other embodiment.

Claims

1. A method for energy scheduling and energy saving optimization of a smart grid, characterized in that, The method comprises the following steps: Obtaining power generation time sequence of power plant; Obtaining information richness according to amplitude difference characteristics of different frequencies in spectrum data of the power generation time sequence; obtaining adaptive neighborhood window according to the spectrum data and the information richness; Obtaining power generation reference weight according to time interval characteristics of any time and the same time of other dates in the power generation time sequence; obtaining power generation similarity according to power generation difference characteristics of the any time and the same time of other dates in the adaptive neighborhood window, and the power generation reference weight of the same time of the other dates; obtaining change difference degree according to difference characteristics of power generation change of the any time and other times in the adaptive neighborhood window, and time interval characteristics; Obtaining robustness weight of the any time according to the power generation similarity and the change difference degree; decomposing all times in the power generation time sequence by time sequence decomposition algorithm according to the robustness weight, and obtaining power generation prediction data by prediction and fusion according to the decomposition result; scheduling electric energy according to the power generation prediction data; The step of obtaining information richness according to amplitude difference characteristics of different frequencies in spectrum data of the power generation time sequence comprises: , wherein R represents information richness, G represents the number of amplitudes other than the constant 0 in the spectrum data, e represents a natural constant, represents the maximum amplitude in the spectrum data, represents the gth amplitude other than the maximum amplitude in the spectrum data, represents the information contribution degree of the gth amplitude, represents the frequency corresponding to the maximum amplitude in the spectrum data, represents the frequency corresponding to the gth amplitude, represents the frequency difference degree; The step of obtaining adaptive neighborhood window according to the spectrum data and the information richness comprises: Calculating the ratio of preset sampling frequency of the power generation time sequence and the frequency corresponding to the maximum amplitude value in the spectrum data to obtain a range reference; calculating the product of the information richness after negative correlation mapping by a preset function and the range reference, and taking the nearest odd number to obtain a data length; taking any analysis time in the power generation time sequence as the center of the adaptive neighborhood window, and taking the data length as the window length of the adaptive neighborhood window to obtain the adaptive neighborhood window of the any analysis time.

2. The energy scheduling and optimization method of the smart grid according to claim 1, wherein, The step of obtaining power generation reference weight according to time interval characteristics of any time and the same time of other dates in the power generation time sequence comprises: Calculating the reciprocal of the time interval of the any time and the same time of any other date to obtain a first value; calculating the sum of all first values corresponding to the any time to obtain a second value; calculating the ratio of the first value and the second value to obtain the power generation reference weight of the same time of the any other date to the any time.

3. The energy scheduling and saving optimization method of the smart grid according to claim 1, wherein, The step of obtaining power generation similarity according to power generation difference characteristics of the any time and the same time of other dates in the adaptive neighborhood window, and the power generation reference weight of the same time of the other dates comprises: where Q denotes the generation similarity at the arbitrary time, N denotes the number of other-date-same-time at the arbitrary time, e denotes the natural constant, denotes the generation time series segment within the adaptive neighborhood window at the arbitrary time, denotes the generation time series segment within the adaptive neighborhood window at the nth other-date-same-time, denotes the dynamic time warping distance between Y and denotes the change similarity; denotes the change similarity; denotes the generation reference weight at the nth other-date-same-time.

4. The energy scheduling and saving optimization method of the smart grid according to claim 1, wherein, The step of obtaining change difference degree according to difference characteristics of power generation change of the any time and other times in the adaptive neighborhood window, and time interval characteristics comprises: , wherein W represents a variation difference degree, M represents a number of time points in the adaptive neighborhood window, e represents a natural constant, and K represents a sum value of the power generation difference values of the arbitrary time point and adjacent time points before and after the arbitrary time point; , wherein W represents a variation difference degree, M represents a number of time points in the adaptive neighborhood window, e represents a natural constant, and K represents a sum value of the power generation difference values of the arbitrary time point and adjacent time points before and after the arbitrary time point; , wherein W represents a variation difference degree, M represents a number of time points in the adaptive neighborhood window, e represents a natural constant, and K represents a sum value of the power generation difference values of the arbitrary time point and adjacent time points before and after the arbitrary time point; , wherein W represents a variation difference degree, M represents a number of time points in the adaptive neighborhood window, e represents a natural constant, and K represents a sum value of the power generation difference values of the arbitrary time point and adjacent time points before and after the arbitrary time point; 5. The energy scheduling and saving optimization method of the smart grid according to claim 1, wherein, The step of obtaining robustness weight of the any time according to the power generation similarity and the change difference degree comprises: The sum value of the change difference degree of all time points in the adaptive neighborhood window of the arbitrary time point is calculated to obtain a fourth numerical value; the ratio of the change difference degree of the arbitrary time point to the fourth numerical value is calculated to obtain a change difference reference value; the difference between the constant 1 and the change difference reference value is calculated to obtain a change normality; and the product of the power generation similarity and the change normality is calculated and normalized to obtain the robustness weight of the arbitrary time point.

Images