Multi-scale comprehensive prediction method and system for generating power of radial flow type small hydropower station

Abstract

The invention is suitable for the technical field of multi-scale prediction, and provides a multi-scale comprehensive prediction method and system for runoff type small hydropower generation power, and the method comprises the steps: obtaining meteorological data, hydrological data and power data of a preset region from a plurality of channels, and carrying out the data preprocessing; carrying out runoff nonuniformity analysis, runoff concentration degree analysis, runoff change amplitude analysis and runoff inter-annual characteristic analysis on the basis of the preprocessed data to obtain runoff characteristics of the target drainage basin; calculating runoff volume and precipitation volume according to runoff characteristics of the target drainage basin; performing small hydropower station cluster power generation load curve characteristic analysis based on the runoff characteristics, the runoff volume and the precipitation volume of the target drainage basin, and screening input variables according to a small hydropower station cluster power generation load curve characteristic analysis result; inputting the input variable into the trained prediction model; the prediction accuracy is effectively improved, and meanwhile, an accurate and valuable data basis is provided for operation management, power grid dispatching and the like of the small hydropower station.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-scale prediction, and particularly relates to a multi-scale comprehensive prediction method and system for the power generation of a runoff small hydropower station. Background Art

[0002] Large and medium-sized hydropower stations generally have a certain reservoir capacity regulation ability. It is common to predict their power generation capacity according to the power generation capacity calculation formula based on the runoff prediction results and conduct orderly scheduling. Different from the mature prediction models of large and medium-sized hydropower stations, the research on the power generation capacity or power prediction of small hydropower starts relatively late and there are many difficulties. For example, small hydropower stations have small installed capacities, are widely distributed, mostly runoff power stations, have poor regulation ability, and are disorderly and lagging in management, making it difficult to accurately predict their power generation capacity. At the same time, there are many influencing factors for the power generation capacity or power of small hydropower. Considering all factors not only involves a huge workload, but also may affect the prediction effect due to excessive noise brought by the input. It is necessary to select reasonable forecasting factors to achieve efficient and accurate prediction of the power generation capacity or power of small hydropower.

[0003] In the prior art, some research proposes a trend-oriented extreme learning machine prediction model. By extracting the power change trend from the historical power output curve of small hydropower and using the trend as the new feature input of the extreme learning machine model, the prediction accuracy of the power output of small hydropower groups is improved. However, this method depends on the size of the small hydropower group and the similarity of historical output rules, and does not consider the impact of meteorological conditions on power generation output, so the applicable scenarios are limited. In addition, there is also research on analyzing the power generation characteristics of small hydropower and predicting the daily power generation load of small hydropower based on typical days through algorithms. However, this method has certain uncertainties for small hydropower with poor regularity, and the prediction results have certain fluctuations, resulting in inaccurate prediction results. Summary of the Invention

[0004] The embodiments of the present application provide a multi-scale comprehensive prediction method and system for the power generation of a runoff small hydropower station to solve the problem of inaccurate prediction results.

[0005] The first aspect of the embodiments of the present application provides a multi-scale comprehensive prediction method for the power generation of a runoff small hydropower station, including:

[0006] Obtain meteorological data, hydrological data, and power data of a preset area from multiple channels, and perform data preprocessing based on the obtained data;

[0007] Based on the preprocessed data, conduct runoff non-uniformity analysis, runoff concentration degree analysis, runoff change amplitude analysis, and runoff inter-annual characteristic analysis to obtain the runoff characteristics of the target basin;

[0008] Calculate the runoff and precipitation of the preset area according to the runoff characteristics of the target basin;

[0009] Based on the runoff characteristics of the target basin, the runoff and precipitation of the preset area, analyze the characteristics of the small hydropower cluster generation load curve, and screen the input variables according to the results of the analysis of the small hydropower cluster generation load curve characteristics;

[0010] Input the input variables into the trained prediction model;

[0011] Generate the predicted value of the small hydropower generation power of the preset area.

[0012] Furthermore, conduct runoff unevenness analysis, runoff concentration degree analysis, runoff variation range analysis and runoff inter-annual characteristic analysis based on the preprocessed data to obtain the runoff characteristics of the target basin, including:

[0013] Runoff unevenness analysis:

[0014]

[0015] Among them: C v is the coefficient of uneven distribution of annual runoff, R is the monthly average runoff, σ is the standard deviation of runoff, R i is the runoff of the i-th month, is the multi-year average monthly runoff, n = 12, n is the number of months, C r is the coefficient of complete regulation of annual runoff distribution, R max and R min are the maximum monthly average runoff and the minimum monthly average runoff respectively.

[0016] Furthermore, conduct runoff unevenness analysis, runoff concentration degree analysis, runoff variation range analysis and runoff inter-annual characteristic analysis based on the preprocessed data to obtain the runoff characteristics of the target basin, including:

[0017] Runoff concentration degree analysis:

[0018] Arrange the percentages of the runoff of 12 months in a year in chronological order, denoted as P1, P2, …, P 12 ;

[0019]

[0020] Among them: L is an intermediate calculation variable, used to comprehensively consider the proportion of monthly runoff and its chronological order in a year, P i is the percentage of the runoff of the i-th month in the annual runoff, and Cd is the runoff concentration degree.

[0021] Furthermore, runoff unevenness analysis, runoff concentration degree analysis, runoff variation range analysis, and runoff inter-annual characteristic analysis are performed on the preprocessed data to obtain the runoff characteristics of the target basin, including:

[0022] Runoff variation range analysis includes the ratio of the maximum monthly average runoff R max to the annual average runoff and the ratio of the minimum monthly average runoff R min to the annual average runoff .

[0023] Furthermore, runoff unevenness analysis, runoff concentration degree analysis, runoff variation range analysis, and runoff inter-annual characteristic analysis are performed on the preprocessed data to obtain the runoff characteristics of the target basin, including:

[0024] Runoff inter-annual characteristic analysis:

[0025]

[0026] Where: and are the within-group variance and between-group variance respectively after grouping the sequence R1, R2, …, R of annual runoff series by k years, n n is the number of data in the i-th group, R i is the j-th data in the i-th group, ij is the average value of annual runoff, is the average value of the i-th group.

[0027] Furthermore, calculating the runoff and precipitation of the preset area according to the runoff characteristics of the target basin includes:

[0028] Selecting a target calculation method according to the runoff characteristics of the target basin to calculate the runoff of each period in the preset area;

[0029] Collecting the evaporation and basin demand in the preset area, and calculating the precipitation of each period in the preset area by combining the collected evaporation and basin demand with the calculated runoff, or collecting the runoff coefficient in the preset area and calculating the precipitation of each period in the preset area by combining the collected runoff coefficient with the calculated runoff.

[0030] Furthermore, analyzing the characteristics of the power generation load curve of the small hydropower cluster based on the runoff characteristics of the target basin, the runoff and precipitation of the preset area, and screening the input variables according to the results of the analysis of the characteristics of the power generation load curve of the small hydropower cluster, including:

[0031] Determining the power generation load curve based on the runoff characteristics of the target basin, the runoff and precipitation of the preset area;​

[0032] Calculate the seasonal characteristics, intra-day characteristics, and inter-annual characteristics of the small hydropower cluster power generation according to the power generation load curve;

[0033] Calculate the correlation between the runoff characteristics of the target basin, the runoff and precipitation data of the preset area, and the small hydropower cluster power generation;

[0034] Screen the input variables respectively based on the calculated seasonal characteristics, intra-day characteristics, and inter-annual characteristics of the small hydropower cluster power generation, and the calculated correlation between the runoff characteristics of the target basin, the runoff and precipitation data of the preset area, and the small hydropower cluster power generation.

[0035] Furthermore, the calculating the seasonal characteristics, intra-day characteristics, and inter-annual characteristics of the small hydropower cluster power generation according to the power generation load curve includes:

[0036] Seasonal characteristics:

[0037]

[0038] Where: is the average power generation per quarter, P i′ is the power generation at the i-th time point in a quarter, n season is the number of time points in a quarter, P j′ is the power generation at the j-th time point in a quarter, n seor is the number of time points in a year;

[0039] Intra-day characteristics:

[0040]

[0041] Where: σ day is the standard deviation of the intra-day power generation, P i ″ is the power generation at the i-th time point within a day, is the average intra-day power generation, n day is the number of time points in a day, and the standard deviation reflects the degree of fluctuation of the intra-day power generation;

[0042] Inter-annual characteristics:

[0043] y = a + bx

[0044] Where: y is the average annual power generation, x is the year, a is the initial year, and b is the slope.

[0045] Furthermore, the calculating the correlation between the runoff characteristics of the target basin, the runoff and precipitation data of the preset area, and the small hydropower cluster power generation includes:

[0046] Calculate the correlation coefficients between the runoff characteristics of the target basin, the runoff and precipitation data of the preset area, and the power generation power of the small hydropower cluster respectively using the Pearson correlation coefficient.

[0047] The second aspect of the embodiments of the present application provides a multi-scale comprehensive prediction system for the power generation power of runoff small hydropower, including:

[0048] A data acquisition and preprocessing unit, configured to acquire meteorological data, hydrological data, and power data of a preset area from multiple channels, and perform data preprocessing based on the acquired data;

[0049] A target basin runoff characteristic determination unit, configured to perform runoff non-uniformity analysis, runoff concentration degree analysis, runoff change amplitude analysis, and annual runoff characteristic analysis based on the preprocessed data to obtain the runoff characteristics of the target basin;

[0050] A runoff and precipitation calculation unit, configured to calculate the runoff and precipitation of the preset area according to the runoff characteristics of the target basin;

[0051] An input variable screening unit, configured to perform analysis on the characteristics of the power generation load curve of the small hydropower cluster based on the runoff characteristics of the target basin, the runoff and precipitation of the preset area, and screen input variables according to the results of the analysis on the characteristics of the power generation load curve of the small hydropower cluster;

[0052] An input variable input unit, configured to input the input variables into the trained prediction model;

[0053] A predicted value generation unit for the power generation power of small hydropower, configured to generate a predicted value of the power generation power of small hydropower in the preset area.

[0054] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0055] By collecting meteorological data, hydrological data, and power data of a preset region from multiple channels, the present invention can cover, to the greatest extent, various key factor information affecting small hydropower generation, avoiding inaccurate analysis results caused by the limitations of a single data source or data missing; based on the preprocessed data, runoff non-uniformity analysis, runoff concentration degree analysis, runoff change range analysis, and runoff inter-annual characteristic analysis are carried out to obtain the runoff characteristics of the target basin, which can comprehensively and multi-dimensionally analyze the runoff characteristics of the target basin, providing accurate basin background information for subsequent accurate calculation of runoff, precipitation, and analysis of the characteristics of the small hydropower generation load curve; calculate the runoff and precipitation of the preset region according to the runoff characteristics of the target basin; carry out analysis on the characteristics of the small hydropower cluster generation load curve based on the runoff characteristics, runoff, and precipitation of the target basin, and screen input variables according to the results of the analysis on the characteristics of the small hydropower cluster generation load curve. By screening input variables based on the results of the characteristic analysis, irrelevant or redundant variables are avoided from being included in the prediction model, making the input of the prediction model more refined, thereby improving the operation efficiency and prediction accuracy of the model, and at the same time reducing the complexity and overfitting risk of the model; input the screened input variables into the trained prediction model, and finally generate the predicted value of the small hydropower generation power in the preset region, providing accurate and valuable data basis for the operation management of small hydropower, power grid dispatching, power market trading, etc.

[0056] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be learned from the practice of the present invention. Brief Description of the Drawings

[0057] Figure 1 It is a schematic flowchart of an embodiment of a multi-scale comprehensive prediction method for the power generation of a runoff small hydropower in the present invention. Detailed Embodiments

[0058] The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims, and drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "correspond to" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0059] Embodiment 1

[0060] In this embodiment, the implementation method can be realized in the system, can be realized on the server, or can be realized on the terminal, and no specific limitation is made. Next, from the perspective of system implementation, the multi-scale comprehensive prediction method for the power generation power of the radial-flow small hydropower will be introduced in this application. Please refer to Figure 1 , the method provided in the embodiment of this application includes the following steps:

[0061] S11. Obtain meteorological data, hydrological data, and power data of a preset area from multiple channels, and perform data preprocessing based on the obtained data;

[0062] The meteorological data includes rainfall data, temperature data, wind speed, and humidity data. The rainfall data includes historical rainfall records, such as annual rainfall, monthly rainfall, daily rainfall, and even hourly rainfall data at different time scales; the temperature data can indirectly affect factors such as evaporation, thereby affecting the runoff; the wind speed and humidity data affect the evaporation process. The hydrological data includes runoff data, water level data, and water quality data. The runoff data includes historical measured runoff records. By analyzing the runoff data at different time periods, the variation law of the runoff can be understood; the water level data reflects the dynamic change of the water volume in the basin. The power data includes historical power generation power data and on-grid electricity price data. The historical power generation power data is used to analyze the variation law of the power generation power, such as the volatility, intermittency, peak-valley characteristics, etc. of the power generation power; the on-grid electricity price data affects the power generation strategy of the power station. The peak-valley electricity price policy will guide the power generation behavior of the hydropower station at different time periods, so it is necessary to obtain the electricity price information at different time periods in order to analyze the impact of the electricity price on the power generation load curve. The data preprocessing includes data missing value processing, outlier processing, data format conversion, and data standardization.

[0063] S12. Based on the preprocessed data, perform runoff non-uniformity analysis, runoff concentration degree analysis, runoff change amplitude analysis, and annual runoff characteristics analysis to obtain the runoff characteristics of the target basin;

[0064] Collect the monthly runoff data of the target basin for many years, calculate the proportion of the monthly runoff in the annual runoff, and perform non-uniformity analysis based on the proportion data:

[0065]

[0066] Among them: C v is the coefficient of non-uniformity of the annual runoff distribution, R is the monthly average runoff, σ is the standard deviation of the runoff, R i is the runoff of the i-th month, is the average monthly runoff for many years, n = 12, n is the number of months, C r is the coefficient of complete regulation of the annual runoff distribution, R maxand R min They are the maximum monthly average runoff and the minimum monthly average runoff respectively.

[0067] Based on the monthly runoff data over the years, calculate the percentage of the monthly runoff in the annual runoff, and determine the concentration degree of runoff by analyzing these percentage data:

[0068] Arrange the percentages of the monthly runoff in the annual runoff for the 12 months of a year in chronological order, denoted as P1, P2, …, P 12 ;

[0069]

[0070] where: L is an intermediate calculation variable used to comprehensively consider the proportion of the monthly runoff and its chronological order in a year, P i is the percentage of the runoff in the i-th month in the annual runoff, and Cd is the runoff concentration degree. The value of Cd is between 0 and 1. The larger the value of Cd, the more concentrated the runoff.

[0071] Sort out the monthly runoff data of the target basin over the years, find out the maximum monthly average runoff and the minimum monthly average runoff, and calculate their proportional relationship with the annual average runoff: The runoff variation range analysis includes the ratio of the maximum monthly average runoff R max to the annual average runoff and the ratio of the minimum monthly average runoff R min to the annual average runoff .

[0072] Collect the annual runoff data of the target basin over the years, and use statistical methods to analyze the variation trend and periodicity of the runoff over the years:

[0073]

[0074] where: and are the within-group variance and between-group variance after grouping the sequence by k years respectively, n n is the number of data in the i-th group, R i is the j-th data in the i-th group, ij is the annual runoff average value, is the average value of the i-th group. is the average value of the i-th group.

[0075] The above analysis of runoff non-uniformity can clearly present the distribution differences of runoff volume in different periods of the year, clarify the degree of imbalance in runoff changes, and help to grasp the characteristics of water resource volume changes in the preset area in different months or seasons. For example, it can identify the runoff difference patterns in wet seasons and dry seasons. The analysis of runoff concentration degree determines the time characteristics when runoff volume appears concentrated in a year, and determines whether runoff is more inclined to converge in certain specific time periods. This has important reference value for subsequent analysis of the concentrated change trend of small hydropower generation power over time. For example, it can judge whether the power generation peak coincides with the runoff concentration period. The analysis of runoff change amplitude shows the intensity of runoff fluctuations within a year, reflects the stability of water resource volume, and can further assist in evaluating the potential fluctuation risks and uncertainties faced by small hydropower generation power. The analysis of annual runoff characteristics grasps the trend and periodic change laws of runoff on a longer time scale, providing a basis for understanding the evolution trend of small hydropower generation power over the years.

[0076] S13. Calculate the runoff volume and precipitation of the preset area according to the runoff characteristics of the target basin;

[0077] In this embodiment, step S13 includes the following:

[0078] 1. Select a target calculation method according to the runoff characteristics of the target basin to calculate the runoff volume of each period in the preset area;

[0079] According to characteristics such as the unevenness and concentration degree of the annual distribution of runoff, judge whether a simple average method can be used or a more complex time series model needs to be considered. If the annual distribution of runoff is relatively uniform, use the annual average runoff to estimate the runoff volume of each period; if the unevenness is large, it is necessary to calculate in combination with the runoff distribution ratio of each month.

[0080] Specifically, the simple average method is as follows:

[0081]

[0082] Among them: Q is the runoff volume of the calculation period, is the average runoff volume, which can be obtained by averaging the measured runoff volume data over the years. t is the calculation period, Q i is the annual runoff volume of the i-th year, and n is the number of years.

[0083] Considering the situation of uneven annual distribution:

[0084]

[0085] Among them: Q m is the runoff volume of the m-th month, is the annual average runoff volume, P mis the proportion of the runoff in the m-th month to the annual runoff, obtained by statistical analysis of multi-year monthly runoff data, and t m is the time length of the m-th month.

[0086] For the interannual variation of runoff with trends:

[0087]

[0088] where: Q y is the annual runoff of the predicted year y, is the average annual runoff of the reference year y0, and b1 is the linear trend coefficient obtained by the interannual trend analysis of runoff, obtained by methods such as linear regression analysis.

[0089] 2. Collect the evaporation and basin water demand of the preset area, and calculate the precipitation of each time period in the preset area by combining the collected evaporation and basin water demand with the calculated runoff, or collect the runoff coefficient of the preset area, and calculate the precipitation of each time period in the preset area by combining the collected runoff coefficient with the calculated runoff.

[0090] According to the water balance equation, when the runoff, evaporation and change of basin water storage are known, the precipitation can be calculated. First, it is necessary to collect the evaporation data of the basin and the data of the change of basin water storage. Specifically as follows:

[0091] P = Q + E + ΔS

[0092] where: P is the precipitation, Q is the runoff, E is the evaporation, and ΔS is the change of basin water storage, calculated by the water level change of water bodies such as reservoirs and lakes in the basin and the corresponding storage - water level curve.

[0093] The runoff coefficient refers to the ratio of the surface runoff of a certain catchment area to the rainfall. It is necessary to obtain the runoff coefficient of the target basin first, which is obtained by statistical analysis of measured rainfall - runoff data, or can be estimated according to factors such as land use type, soil type, and vegetation cover of the basin. At the same time, accurate runoff data is required. Specifically as follows:

[0094]

[0095] where: C is the runoff coefficient, which is a dimensionless ratio.

[0096] This step accurately calculates the runoff volume of a preset area based on the runoff characteristics of the target basin, enabling the knowledge of the actual water volume available for small hydropower generation at different time scales. This is the core material basis for evaluating the potential of small hydropower generation and predicting the power generation capacity. The calculation of precipitation further improves the water balance analysis of the basin, understanding the specific situation of this water source replenishment. At the same time, the precipitation data and runoff volume data are mutually verified and analyzed, which helps to more deeply understand the cycle and transformation process of water resources in the basin, providing a quantitative basis for considering the impact of precipitation changes on power generation in subsequent power generation capacity predictions.

[0097] S14. Analyze the characteristics of the power generation load curve of the small hydropower cluster based on the runoff characteristics of the target basin, the runoff volume and precipitation of the preset area, and screen the input variables according to the results of the analysis of the characteristics of the power generation load curve of the small hydropower cluster;

[0098] In this embodiment, step S14 includes the following:

[0099] 1. Determine the power generation load curve based on the runoff characteristics of the target basin, the runoff volume and precipitation of the preset area;

[0100] Taking time as the horizontal axis and the power generation data as the vertical axis, plot the power generation load curve, and mark the water resource situation at different time periods on the curve according to the runoff volume and precipitation data, so as to visually observe the change of power generation with time and its relationship with the water resource volume.

[0101] 2. Calculate the seasonal characteristics, intra-day characteristics and inter-annual characteristics of the power generation of the small hydropower cluster according to the power generation load curve;

[0102] Calculation of seasonal characteristics:

[0103] Divide a year by season or month, for example, into four seasons of spring, summer, autumn and winter, or take each month as a time period, calculate statistical indicators such as the average power generation, maximum power and minimum power of each season or month, and analyze the proportion of power generation in each season as follows:

[0104]

[0105] Among them: is the average power generation of a quarter, P i′ is the power generation at the i-th time point in a quarter, n season is the number of time points in a quarter, P j′ is the power generation at the j-th time point in a quarter, n seor is the number of time points in a year.

[0106] Describe the seasonal characteristics through these proportions and statistical indicators, and determine the influence law of the wet season and dry season on the power generation.

[0107] Calculation of intraday characteristics:

[0108] Taking one day as a cycle, analyze the change of power generation within a day, and calculate the standard deviation of the intraday power generation:

[0109]

[0110] Where: σ day is the standard deviation of the intraday power generation, P i″ is the power generation at the i-th time point within a day, is the average power generation within a day, n day is the number of time points in a day. The standard deviation reflects the fluctuation degree of the intraday power generation.

[0111] Determine the peak and trough periods of the intraday power generation. By setting power thresholds or using clustering analysis and other methods, find the time periods with higher and lower power, and analyze the relationships between these periods and factors such as the electricity demand of human activities and runoff changes.

[0112] Calculation of interannual characteristics:

[0113] For the power generation data of multiple years, calculate the change trend of the annual average power generation. Here, the linear regression method is used as follows:

[0114] y = a + bx

[0115] Where: y is the annual average power generation, x is the year, a is the initial year, and b is the slope, indicating the change trend of the interannual power generation. If b>0, it means the power generation shows an upward trend, otherwise it shows a downward trend.

[0116] 3. Calculate the correlation between the runoff characteristics of the target basin, the runoff and precipitation data of the preset area, and the power generation of the small hydropower cluster;

[0117] Use the Pearson correlation coefficient to calculate the correlation coefficients between the data corresponding to the runoff characteristics of the target basin, the runoff and precipitation of the preset area, and the power generation of the small hydropower cluster respectively.

[0118]

[0119] Where: X is the runoff, Y is the power generation, n is the number of data points, X i and Y i are the i-th observations of X and Y respectively, and are the averages of X and Y respectively.

[0120] Meanwhile, to detect possible non - linear relationships, the Spearman rank correlation coefficient is used as a supplement, and its calculation formula is:

[0121]

[0122] Where: is the rank difference between X and Y, that is, the difference in the rankings of the i - th data point in X and Y.

[0123] Calculate the Pearson correlation coefficient and the Spearman rank correlation coefficient between the runoff characteristic indexes, runoff volume, precipitation, and the power generation of the small - hydropower cluster respectively. For example, when calculating the correlation between the runoff volume and the power generation, substitute the runoff volume data as the X variable and the power generation data as the Y variable into the above formula for calculation, obtaining two correlation coefficient values, and comprehensively judge the degree and type of the correlation between the two.

[0124] 4. Screen the input variables respectively based on the calculated seasonal characteristics, intra - day characteristics, and inter - annual characteristics of the small - hydropower cluster power generation, as well as the calculated correlation between the runoff characteristics of the target basin, the runoff volume and precipitation in the preset area, and the small - hydropower cluster power generation.

[0125] According to the results of the seasonal characteristic analysis, determine the factors that have a significant impact on the power generation in different seasons. For example, if in the rainy season, the correlations between the runoff volume and precipitation and the power generation are relatively high, and the fluctuations of the power generation in this season are mainly affected by these two factors, then in the prediction model related to the rainy season, the runoff volume and precipitation are preferentially selected as input variables. Based on the intra - day characteristic analysis, screen the key variables for different time periods. During peak electricity consumption periods, factors such as grid dispatching instructions, industrial and residential electricity demands are closely related to the power generation, and these variables are used as important input variables for predicting the power generation during peak periods. Identify variables that have a long - term and stable impact on the power generation, such as variables related to the basic characteristics of the basin's topography, soil type, etc. Since they are relatively stable over the years and have a long - term correlation with the power generation, they should be retained as important input variables. According to the results of the correlation analysis, set a correlation threshold. For example, for the Pearson correlation coefficient, variables with an absolute value greater than 0.5 are regarded as strongly correlated variables, and these variables are preferentially considered to be included in the input variable set. For variables with multicollinearity, methods such as calculating the variance inflation factor VIF are used for judgment and processing. If the VIF between two variables is greater than 5, it indicates strong multicollinearity. Select the variable with a higher correlation with the power generation and more interpretability in terms of hydropower principles among these two variables as the final input variable to avoid problems such as overfitting of the model and improve the stability and prediction accuracy of the model.

[0126] S15. Input the input variables into the trained prediction model;

[0127] S16. Generate a predicted value of the small hydropower generation power in a preset area.

[0128] First, identify the selected input variables. These variables are obtained through a series of previous analysis processes, based on seasonal characteristics, intraday characteristics, interannual characteristics, and correlation analysis, such as runoff, precipitation, some meteorological factors, and variables related to basin characteristics. Ensure that the data formats of these variables are unified and complete, and the time series corresponds accurately. Organize the historical small hydropower generation power data corresponding to the input variables as a reference standard for the model prediction results, and also ensure that its time scale is consistent with the input variable data, such as data recorded in hours, days, or months as the time unit.

[0129] To treat input variables of different magnitudes equally in the model and avoid poor model training effects or prediction deviations caused by excessive differences in data magnitudes, perform normalization or standardization operations on the data. Divide the preprocessed data into a training set, a validation set, and a test set, and determine an appropriate time step. In addition to the dimension adjustment related to the time step, determine whether the dimension positions corresponding to each input variable are correct. Load the trained long short-term memory network (LSTM) prediction model from local storage or the corresponding model management system. This model is a version with relatively good performance after multiple iterative trainings and hyperparameter adjustments on the validation set. Input the organized and format-matched input variable data into the loaded prediction model. The LSTM model processes the input data according to the laws and mapping relationships it has learned internally. Finally, use evaluation metrics to measure the accuracy and reliability of the prediction results. Based on the results of these evaluation metrics, it can be judged whether the prediction performance of the model meets the actual requirements. If the metrics are not ideal, the reasons can be further analyzed, such as considering whether it is necessary to readjust the model structure, optimize the input variables, or add more high-quality data for retraining, etc., to continuously improve the prediction accuracy.

[0130] The above steps can input the selected input variables into the trained prediction model, generate a predicted value of the small hydropower generation power in the preset area, and at the same time effectively evaluate and analyze the prediction results, improve the prediction accuracy, and provide valuable reference for related work such as the operation management of small hydropower and power grid dispatching.

[0131] Embodiment 2

[0132] An embodiment of a multi-scale comprehensive prediction system for the generation power of a run-of-river small hydropower in the present invention includes the following steps:

[0133] A data acquisition and preprocessing unit, configured to acquire meteorological data, hydrological data, and power data of a preset area from multiple channels, and perform data preprocessing based on the acquired data;

[0134] A target basin runoff characteristic determination unit, configured to perform runoff unevenness analysis, runoff concentration degree analysis, runoff variation range analysis, and runoff interannual characteristic analysis based on the preprocessed data to obtain the target basin runoff characteristics;

[0135] A runoff and precipitation calculation unit, configured to calculate the runoff and precipitation of the preset area according to the target basin runoff characteristics;

[0136] An input variable screening unit, configured to perform small hydropower cluster power generation load curve characteristic analysis based on the target basin runoff characteristics, the runoff and precipitation of the preset area, and screen input variables according to the results of the small hydropower cluster power generation load curve characteristic analysis;

[0137] An input variable input unit, configured to input the input variables into the trained prediction model;

[0138] A predicted value generation unit for small hydropower generation power, configured to generate a predicted value of the small hydropower generation power of the preset area.

[0139] For the specific limitations of the system, reference may be made to the limitations on the method in the foregoing text, which will not be elaborated herein. Each module in the above system can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above respective modules.

[0140] Those of ordinary skill in the art can realize that the units of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components of the examples have been generally described according to their functions in the above description. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0141] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc. In addition, in each embodiment of the present invention, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0142] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs that can store program codes.

[0143] It can be understood that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the various embodiments of the present invention, and they should all be covered by the scope of the claims and the description of the present invention.

Claims

1. A multi-scale comprehensive prediction method for run-of-river small hydropower power generation, characterized in that: include: Obtain meteorological data, hydrological data, and power data for a preset area from multiple channels, and perform data preprocessing based on the acquired data; Based on the preprocessed data, runoff unevenness analysis, runoff concentration analysis, runoff variation analysis and runoff inter-annual characteristics analysis are carried out to obtain the runoff characteristics of the target basin; Calculate the runoff and precipitation in the preset area according to the runoff characteristics of the target watershed; Based on the runoff characteristics of the target watershed, the runoff volume and precipitation of the preset area, a small hydropower cluster power generation load curve characteristic analysis is performed, and input variables are selected according to the results of the small hydropower cluster power generation load curve characteristic analysis; inputting the input variables into the trained prediction model; Generate a predicted value of small hydropower generation power in the preset area.

2. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 1 is characterized in that: The runoff unevenness analysis, runoff concentration analysis, runoff variation analysis and runoff inter-annual characteristics analysis are performed based on the pre-processed data to obtain the runoff characteristics of the target basin, including: Analysis of runoff non-uniformity: Where: C v is the annual runoff distribution unevenness coefficient, R is the monthly average runoff, σ is the standard deviation of runoff, and R i is the runoff in the ith month, is the multi-year average monthly runoff, n = 12, n is the number of months, C r The total regulation coefficient for runoff distribution within a year, R max and R min They are the maximum and minimum monthly average runoff respectively.

3. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 2 is characterized in that: The runoff unevenness analysis, runoff concentration analysis, runoff variation analysis and runoff inter-annual characteristics analysis are performed based on the pre-processed data to obtain the runoff characteristics of the target basin, including: Analysis of runoff concentration: Arrange the percentage of runoff in 12 months of a year to the annual runoff in chronological order and record them as P1, P2, ..., P 12 ; Among them: L is the intermediate calculation variable, which is used to comprehensively consider the proportion of runoff in each month and its time sequence in a year, P i is the percentage of the runoff in the ith month to the annual runoff, and Cd is the runoff concentration.

4. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 2 is characterized in that: The runoff unevenness analysis, runoff concentration analysis, runoff variation analysis and runoff inter-annual characteristics analysis are performed based on the pre-processed data to obtain the runoff characteristics of the target basin, including: The analysis of runoff variation includes the maximum monthly average runoff R max The annual average runoff The ratio and the minimum monthly average runoff R min The annual average runoff Ratio.

5. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 2 is characterized in that: The runoff unevenness analysis, runoff concentration analysis, runoff variation analysis and runoff inter-annual characteristics analysis are performed based on the pre-processed data to obtain the runoff characteristics of the target basin, including: Analysis of interannual characteristics of runoff: in: and The annual runoff series are R1, R2, ..., R n The variance within and between groups after the series is grouped into k years, n i is the number of data in the i-th group, R ij is the jth data in the i-th group, is the average annual runoff value, is the average value of group i.

6. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 1 is characterized in that: The step of calculating the runoff volume and precipitation of the preset area according to the runoff characteristics of the target watershed includes: Select a target calculation method according to the runoff characteristics of the target watershed to calculate the runoff volume of each time period in the preset area; Collect the evaporation and watershed demand of the preset area, combine the collected evaporation and watershed demand with the calculated runoff to calculate the precipitation in each time period of the preset area, or collect the runoff coefficient of the preset area, combine the collected runoff coefficient with the calculated runoff to calculate the precipitation in each time period of the preset area.

7. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 1 is characterized in that: The analysis of the load curve characteristics of the small hydropower cluster based on the runoff characteristics of the target basin, the runoff volume and precipitation of the preset area, and the screening of input variables according to the results of the analysis of the load curve characteristics of the small hydropower cluster, include: Determine a power generation load curve based on the runoff characteristics of the target watershed, the runoff volume and precipitation in the preset area; Calculate seasonal, intra-day and inter-annual characteristics of power generation of the small hydropower cluster according to the power generation load curve; Calculate the correlation between the runoff characteristics of the target watershed, the runoff volume and precipitation data of the preset area and the power generation of the small hydropower cluster; Input variables are screened based on the calculated seasonal characteristics, intra-day characteristics and inter-annual characteristics of power generation of the small hydropower cluster, the calculated runoff characteristics of the target watershed, the data corresponding to the runoff and precipitation in the preset area and the correlation with the power generation of the small hydropower cluster.

8. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 7 is characterized in that: The calculation of seasonal characteristics, intra-day characteristics and inter-annual characteristics of power generation of the small hydropower cluster according to the power generation load curve includes: Seasonal characteristics: in: is the average power generation in one quarter, P i′ is the power generation at the i-th time point in a quarter, n season is the number of time points in a quarter, P j′ is the power generation at the jth time point in a quarter, n seor is the number of time points throughout the year; Intraday features: Where: day is the standard deviation of the daily power generation, P i″ is the power generation at the i-th time point in the day, is the average daily power generation, n day is the number of time points in a day, and the standard deviation reflects the fluctuation degree of power generation during the day; Interannual characteristics: y=a+bx Among them: y is the annual average power generation, x is the year, a is the initial year, and b is the slope.

9. The multi-scale comprehensive prediction method for run-of-river small hydropower generation power according to claim 7 is characterized in that: The calculation of the runoff characteristics of the target watershed, the correlation between the data corresponding to the runoff and precipitation in the preset area and the power generation of the small hydropower cluster includes: The Pearson correlation coefficient is used to calculate the correlation coefficient between the runoff characteristics of the target watershed, the data corresponding to the runoff and precipitation in the preset area, and the power generation capacity of the small hydropower cluster.

10. A multi-scale comprehensive prediction system for run-of-river small hydropower generation, characterized in that: The multi-scale comprehensive prediction method of run-of-river small hydropower power generation power according to any one of claims 1 to 9 comprises: A data acquisition and preprocessing unit, used to acquire meteorological data, hydrological data and power data of a preset area from multiple channels, and perform data preprocessing based on the acquired data; The target watershed runoff characteristic determination unit is used to perform runoff non-uniformity analysis, runoff concentration analysis, runoff variation range analysis and runoff inter-annual characteristic analysis based on the pre-processed data to obtain the runoff characteristics of the target watershed; A runoff and precipitation calculation unit, used to calculate the runoff and precipitation of the preset area according to the runoff characteristics of the target watershed; An input variable screening unit, used to perform a small hydropower cluster power generation load curve characteristic analysis based on the target basin runoff characteristics, the runoff volume and precipitation in the preset area, and screen input variables according to the results of the small hydropower cluster power generation load curve characteristic analysis; An input variable input unit, used to input the input variable into the trained prediction model; The predicted value generating unit of the small hydropower generation power is used to generate the predicted value of the small hydropower generation power in the preset area.

Images