Methods, devices, equipment and storage media for joint prediction of power generation of hydropower station clusters

Abstract

This invention relates to the field of power generation prediction technology, and discloses a method, apparatus, equipment, and storage medium for joint prediction of power generation in a hydropower station cluster. The method includes: collecting historical power generation data and multiple hydrological data related to power generation from each hydropower station in the cluster; collecting geographical distance data between hydropower stations; constructing a spatial impact prediction term for power generation based on the power generation matrix time series and geographical distance data; constructing a temporal impact prediction term for power generation based on the historical moment power generation matrix in the power generation matrix time series; constructing a temporal impact prediction term for hydrological data based on the historical moment hydrological matrix in the hydrological matrix time series; combining these with the construction of a matrix-valued time series spatial autoregressive model; training the matrix-valued time series spatial autoregressive model to obtain a joint prediction model for power generation in the hydropower station cluster; and predicting the power generation of the hydropower station cluster. This invention can improve the accuracy of power generation prediction.

Description

Technical Field

[0001] This invention relates to the field of power generation prediction technology, specifically to a method, apparatus, equipment, and storage medium for joint prediction of power generation in a cluster of hydropower stations. Background Technology

[0002] A hydroelectric power station consists of a hydraulic system, a mechanical system, and power generation devices. It is a key water conservancy project that converts water energy into electrical energy. The sustainability of power production requires the uninterrupted utilization of hydropower. Through the construction of the hydroelectric power station's reservoir system, the temporal and spatial distribution of water resources is artificially regulated and altered, achieving sustainable utilization of these resources. To effectively convert the water energy in the reservoir into electrical energy, the power generation capacity of the hydroelectric power station needs to be predicted so that staff can make appropriate allocations and reserve plans, ensuring the continuity of power supply.

[0003] Existing methods for predicting hydropower generation mainly focus on time series analysis and deep learning. Traditional time series analysis builds autoregressive models for individual hydropower stations, using univariate or vector data modeling based on historical power generation data. Deep learning models, such as LSTM and Extreme Learning Machine, also tend to focus on a single hydropower station, combining information such as inflow, water level, and head to create a model. However, current methods cannot model the spatial effects between hydropower station clusters and do not utilize the natural correlation characteristics between hydropower station clusters, resulting in low accuracy in power generation prediction. Summary of the Invention

[0004] In view of this, the present invention provides a method, apparatus, equipment and storage medium for joint prediction of power generation of a hydropower station cluster, so as to improve the accuracy of power generation prediction.

[0005] In a first aspect, the present invention provides a method for joint prediction of power generation in a hydropower station cluster. This method includes: collecting historical power generation data and multiple hydrological data related to power generation from each hydropower station in the cluster; constructing a power generation matrix time series for the hydropower station cluster based on the historical power generation data of each hydropower station; constructing a hydrological matrix time series for the hydropower station cluster based on the hydrological data of each hydropower station; constructing a dataset based on the power generation matrix time series and the hydrological matrix time series; constructing a spatial impact prediction term for power generation based on the power generation matrix time series and geographical distance data; constructing a temporal impact prediction term for power generation based on the historical moment power generation matrix in the power generation matrix time series; constructing a temporal impact prediction term for hydrological data based on the historical moment hydrological matrix in the hydrological matrix time series; constructing a matrix-valued time series spatial autoregressive model by combining the spatial impact prediction term, the temporal impact prediction term, and the temporal impact prediction term; training the matrix-valued time series spatial autoregressive model based on the dataset to obtain a joint prediction model for power generation in the hydropower station cluster; and predicting the power generation of the hydropower station cluster based on the joint prediction model.

[0006] In this implementation, this application takes the perspective of joint prediction of multiple hydropower stations in a hydropower station cluster. Based on the similar power generation trend characteristics of hydropower stations with similar hydrological and meteorological conditions, it fully considers the mutual influence between different hydropower stations and comprehensively considers the impact of historical power generation matrix time series and other hydrological matrix time series on power generation to construct a model. Specifically, based on the spatial effect influence of the natural geographical relationship between multiple hydropower stations, the pure time dynamic effect, and the hydrological variable effect, a matrix-valued time series spatial autoregressive model is constructed. This model not only considers the mutual influence relationship in the time dimension but also the correlation relationship in the spatial dimension, which can improve the accuracy of power generation prediction for hydropower station clusters.

[0007] In one optional implementation, a power generation matrix time series of a hydropower station cluster is constructed based on the historical power generation data of each hydropower station, including: for the daily historical power generation data, defining the matrix rows as each hydropower station and defining each column of the matrix as the power generation of each hydropower station at different times, to obtain the daily corresponding power generation matrix; constructing a power generation matrix time series based on the power generation matrix of consecutive dates; and constructing a hydrological matrix time series of the hydropower station cluster based on the hydrological data of each hydropower station, including: for the daily hydrological data, calculating the mean of each hydrological variable of each hydropower station; defining each row of the matrix as each hydropower station and defining each column of the matrix as the mean of each hydrological variable of each hydropower station, to obtain the daily corresponding hydrological matrix; and constructing a hydrological matrix time series based on the hydrological matrix of consecutive dates.

[0008] In this implementation, the power generation matrix of the hydropower station is constructed using the date as the matrix standard, and multiple hydrological matrices of the hydropower station are constructed. This approach can take into account the spatial relationship of the data and construct a time series according to the time order, thus taking into account the temporal relationship of the data and providing a good dataset environment for subsequent model training, thereby further improving the model training effect.

[0009] In one optional implementation, a spatial impact prediction term for power generation is constructed based on the power generation matrix time series and geographical distance data, including: constructing a spatial weight matrix for hydropower station clusters based on geographical distance data, wherein the larger the geographical distance data, the smaller the corresponding element value in the spatial weight matrix for hydropower station clusters; constructing a concurrent prediction term for the spatial impact of power generation based on the current moment power generation matrix of the power generation matrix time series and the spatial weight matrix for hydropower station clusters; and constructing a lagged prediction term for the spatial impact of power generation based on the historical moment power generation matrix of the power generation matrix time series and the spatial weight matrix for hydropower station clusters.

[0010] In this implementation, a synchronous prediction term and a lagged prediction term of the spatial impact of power generation are introduced into the matrix-valued time series spatial autoregressive model. This takes into account the pure spatial effect of power generation based on the natural geographical relationship between multiple hydropower stations and the spatial effect of power generation time lag. This approach is more suitable for the joint power generation prediction of current hydropower station clusters and improves the accuracy of subsequent model applications.

[0011] In one optional implementation, a contemporaneous prediction term for the spatial impact of power generation is constructed based on the current-time power generation matrix and the spatial weight matrix of the hydropower station cluster from the power generation matrix time series. This includes: fusing the current-time power generation matrix and the spatial weight matrix of the hydropower station cluster to obtain a first fused matrix; adjusting the first fused matrix using a first diagonal matrix to obtain the contemporaneous prediction term for the spatial impact of power generation; and constructing a lagged prediction term for the spatial impact of power generation based on the historical-time power generation matrix and the spatial weight matrix of the hydropower station cluster from the power generation matrix time series. This includes: combining the historical-time power generation matrix with a first lag order and the spatial weight matrix of the hydropower station cluster. The matrices are fused to obtain a second fused matrix; the second fused matrix is ​​adjusted using the second diagonal matrix to obtain the spatial impact lag prediction term of power generation; the time impact prediction term of power generation is constructed based on the historical moment power generation matrix in the time series of the power generation matrix, including: adjusting the historical moment power generation matrix with the second lag order using the first regression coefficient matrix to obtain the time impact prediction term of power generation; the time impact prediction term of hydrological conditions is constructed based on the historical moment hydrological matrix in the time series of the hydrological matrix, including: adjusting the historical moment hydrological matrix with the third lag order using the second regression coefficient matrix to obtain the time impact prediction term of hydrological conditions.

[0012] In one optional implementation, training a matrix-valued time-series spatial autoregressive model based on a dataset to obtain a joint prediction model for the power generation of a hydropower station cluster includes: dividing the dataset into a training set and a test set according to a partitioning ratio; using the training set to estimate the parameters of the matrix-valued time-series spatial autoregressive model to obtain the trained matrix-valued time-series spatial autoregressive model; and testing the trained matrix-valued time-series spatial autoregressive model based on the test set to determine the lag order and obtain the joint prediction model for the power generation of the hydropower station cluster.

[0013] In this implementation, the iterative estimation method of this application can effectively address the endogeneity problem in the model.

[0014] In one optional implementation, parameter estimation is performed on the matrix-valued time series spatial autoregressive model to obtain a trained matrix-valued time series spatial autoregressive model, including: obtaining multiple preset lag order combinations; using a training set, parameter estimation is performed on the matrix-valued time series spatial autoregressive models corresponding to the multiple preset lag order combinations to obtain multiple trained matrix-valued time series spatial autoregressive models; testing the trained matrix-valued time series spatial autoregressive models based on a test set to determine the lag order to obtain a joint prediction model for the power generation of the hydropower station cluster, including: calculating the model error of the multiple trained matrix-valued time series spatial autoregressive models based on the test set; calculating the Bayesian information criterion value based on the model error, and selecting the preset lag order combination with the smallest Bayesian information criterion value to obtain the corresponding joint prediction model for power generation.

[0015] In this implementation, different combinations of orders are verified and selected to determine the optimal combination. This avoids the problem of underfitting the model when the order is too small, while the problem of overfitting the model when the order is too large. This method enables the model to show a good fit on the existing data and has less prediction error.

[0016] Secondly, the present invention provides a joint prediction device for the power generation of a hydropower station cluster. This device includes: a data acquisition module for acquiring historical power generation data of each hydropower station in the cluster and multiple hydrological data related to power generation; acquiring geographical distance data between hydropower stations; a first construction module for constructing a power generation matrix time series of the hydropower station cluster based on the historical power generation data of each hydropower station; constructing a hydrological matrix time series of the hydropower station cluster based on the hydrological data of each hydropower station; and constructing a dataset based on the power generation matrix time series and the hydrological matrix time series; and a second construction module for constructing a dataset based on the power generation matrix time series and the hydrological matrix time series. The system constructs a spatial impact prediction term for power generation based on geographical distance data, a temporal impact prediction term for power generation based on historical moment power generation matrices in the power generation matrix time series, and a temporal impact prediction term for hydrological conditions based on historical moment hydrological conditions in the hydrological matrix time series. It then combines these three terms to construct a matrix-valued time series spatial autoregressive model. A training module is used to train the matrix-valued time series spatial autoregressive model based on the dataset, resulting in a joint prediction model for the power generation of the hydropower station cluster. Finally, a prediction module is used to predict the power generation of the hydropower station cluster based on the joint prediction model.

[0017] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the joint prediction method for power generation of a hydropower station cluster described in the first aspect or any corresponding embodiment thereof.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the joint prediction method for power generation of a hydropower station cluster as described in the first aspect or any corresponding embodiment thereof.

[0019] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the joint prediction method for power generation of a hydropower station cluster as described in the first aspect or any of its corresponding embodiments. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a method for joint prediction of power generation of a hydropower station cluster according to an embodiment of the present invention.

[0022] Figure 2 This is a flowchart illustrating another method for joint prediction of power generation of a hydropower station cluster according to an embodiment of the present invention.

[0023] Figure 3 This is a structural block diagram of a hydropower station cluster power generation joint prediction device according to an embodiment of the present invention;

[0024] Figure 4 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] With the enhancement of information collection capabilities, more and more complex data structures are being stored and recorded systematically, bringing tremendous impact and innovation to statistical thinking and methods, and giving new meaning to the important field of time series analysis. Traditional time series analysis is mostly based on univariate or vector observations; however, due to the inclusion of complex data structures, traditional time series analysis methods cannot be directly applied. In complex time series, a common form is matrix-valued time series, where the observation value at each time point is a matrix. Time series data with this structure exist in many real-world scenarios; for example, the daily power generation data of different hydropower stations with a sampling interval of 15 minutes.

[0027] Hydropower stations within a cluster exhibit certain correlations, significantly impacting power generation forecasting. Furthermore, the spatial effects arising from these correlations vary. Therefore, analyzing the inherent matrix-valued time-series data structure of hydropower generation clusters and comprehensively examining the spatial effects among different hydropower stations is meaningful. This application proposes a joint power generation forecasting method for hydropower clusters, considering both spatial and temporal effects among hydropower stations to improve forecasting accuracy.

[0028] According to an embodiment of the present invention, a method for joint prediction of power generation of a hydropower station cluster is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0029] This embodiment provides a method for joint prediction of power generation of a hydropower station cluster. Figure 1 This is a flowchart of a joint prediction method for the power generation of a hydropower station cluster according to an embodiment of the present invention. It should be noted that if substantially the same result is obtained, this embodiment does not necessarily reflect that result. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, the process includes the following steps:

[0030] Step S101: Collect historical power generation data and multiple hydrological data related to power generation for each hydropower station in the hydropower station cluster, and collect geographical distance data between hydropower stations.

[0031] The hydropower station cluster comprises multiple hydropower stations, each influencing the others. In one implementation, the hydropower station cluster consists of small, cascaded hydropower stations with hydraulic connections.

[0032] Among them, hydrological data related to power generation include inflow, outflow, water level, and head of each hydropower station.

[0033] Specifically, this application collects power generation data of each hydropower station in the hydropower station cluster according to a preset sampling interval to obtain historical power generation data and records the timestamp of each power generation data; it also collects historical hydrological data related to the prediction of hydropower station power generation, such as inflow, outflow, water level, and head of each hydropower station, and records the timestamp of each hydrological data.

[0034] For example, the preset sampling interval is 15 minutes, and the power generation data of each hydropower station in the hydropower station cluster is collected every 15 minutes.

[0035] Meanwhile, this application collects geographical distance data between any two hydropower stations in a hydropower station cluster. The geographical distance between the hydropower stations represents the degree of mutual influence between the two hydropower stations. The closer the geographical distance, the greater the degree of influence.

[0036] Step S102: Construct a power generation matrix time series for the hydropower station cluster based on the historical power generation data of each hydropower station; construct a hydrological matrix time series for the hydropower station cluster based on the hydrological data of each hydropower station; and construct a dataset based on the power generation matrix time series and the hydrological matrix time series.

[0037] The historical power generation data of each hydropower station and multiple hydrological data related to power generation are divided according to unit time intervals. A power generation matrix is ​​constructed from the historical power generation data within each time unit and arranged in chronological order to obtain a power generation matrix time series. Similarly, a hydrological matrix is ​​constructed from the hydrological data within each time unit and arranged in chronological order to obtain a hydrological matrix time series. The power generation matrix time series and the hydrological matrix time series constitute the dataset.

[0038] In one implementation, the unit time interval is one day. Specifically, a power generation matrix is ​​constructed from the historical power generation data observed for each hydropower station daily, and the power generation matrices for different dates form a power generation matrix time series. A hydrological matrix is ​​constructed based on the hydrological data observed at each hydropower station daily. The hydrological matrices for different dates form a hydrological matrix time series. ,in, Represents time.

[0039] Step S103: Construct a spatial impact prediction term for power generation based on the power generation matrix time series and geographical distance data; construct a temporal impact prediction term for power generation based on the historical moment power generation matrix in the power generation matrix time series; construct a temporal impact prediction term for hydrological conditions based on the historical moment hydrological matrix in the hydrological matrix time series; combine the spatial impact prediction term for power generation, the temporal impact prediction term for power generation, and the temporal impact prediction term for hydrological conditions to construct a matrix value time series spatial autoregressive model.

[0040] This application introduces a spatial regression term for the power generation matrix time series, a spatial regression term for the lag period of the power generation matrix time series, and a regression term for the lag period of the hydrological matrix time series to construct a matrix value time series spatial autoregressive model (MSTLAR).

[0041] Specifically, the power generation matrix time series is combined with the geographical distance data between power stations to consider the influence of hydropower station power generation on power generation prediction in the geographical dimension, thus obtaining a spatial influence prediction term for power generation. Based on the lag period, the historical moment power generation matrix in the power generation matrix series is determined, and considering the influence of hydropower station power generation on power generation prediction in the time dimension, thus obtaining a temporal influence prediction term for power generation. Similarly, based on the lag period, the hydrological data power generation matrix in the hydrological data matrix series is determined, and considering the influence of hydrological data related to hydropower station power generation on power generation prediction in the time dimension, thus obtaining a temporal influence prediction term for hydrological data. Combining these multiple prediction regression terms yields the matrix-valued time series spatial autoregressive model of this application.

[0042] In this implementation, the lag order of the historical power generation matrix and the lag order of the historical hydrological matrix are not limited.

[0043] Step S104: Train the matrix-valued time series spatial autoregressive model based on the dataset to obtain the joint prediction model for power generation of the hydropower station cluster.

[0044] In the dataset, the historical power generation matrix is ​​determined in the power generation matrix time series according to the lag order, the historical hydrological matrix is ​​determined in the hydrological matrix time series, and a set of training samples is determined by combining the current power generation matrix, the historical power generation matrix, and the historical hydrological matrix.

[0045] For example, when the lag order of the historical power generation matrix is ​​1 and the lag order of the historical hydrological matrix is ​​2, then... Power generation matrix at any time Historical power generation matrix at any given time and The historical hydrological data matrix at any given time is used as a set of training samples.

[0046] The matrix-valued time series spatial autoregressive model constructed in the above steps is trained using multiple sets of training samples to obtain a joint prediction model for the power generation of the hydropower station cluster.

[0047] Step S105: Predict the power generation of the hydropower station cluster based on the joint power generation prediction model.

[0048] Historical power generation data and multiple hydrological data corresponding to the lag order are obtained, and power generation matrix and hydrological matrix are constructed. These are then input into the joint power generation prediction model of the hydropower station cluster, and the predicted output is the power generation at future time.

[0049] The joint prediction method for power generation of a hydropower station cluster provided in this embodiment is triggered from the perspective of joint prediction of multiple hydropower stations in a cluster. Based on the similar power generation trend characteristics of hydropower stations with similar hydrological and meteorological conditions, it fully considers the mutual influence between different hydropower stations and comprehensively considers the influence of historical power generation matrix time series and other hydrological matrix time series on power generation to construct a model. Specifically, based on the spatial effect influence of the natural geographical relationship between multiple hydropower stations, the pure time dynamic effect, and the influence effect of hydrological data, a matrix value time series spatial autoregressive model is constructed. This model not only considers the mutual influence relationship in the time dimension but also the correlation relationship in the spatial dimension, which can improve the accuracy of power generation prediction for hydropower station clusters.

[0050] This embodiment provides a method for joint prediction of power generation of a hydropower station cluster. Figure 2This is a flowchart of another method for joint prediction of power generation of a hydropower station cluster according to an embodiment of the present invention. It should be noted that if substantially the same result is obtained, this embodiment does not necessarily use it. Figure 2 The illustrated process sequence is limited. For example... Figure 2 As shown, the process includes the following steps:

[0051] Step S201: Collect historical power generation data and multiple hydrological data related to power generation for each hydropower station in the hydropower station cluster, and collect geographical distance data between hydropower stations.

[0052] Specifically, a sampling time interval is set, such as 15 minutes. Based on the data acquisition and monitoring system of each hydropower station, historical power generation data and its timestamps are collected from each hydropower station in the hydropower station cluster in real time. Based on the reservoir hydrological monitoring system, hydrological data related to the prediction of hydropower station power generation, such as inflow, outflow, water level, and head, and their timestamps are collected from each hydropower station in the hydropower station cluster. All of the data collected in this application are time series data.

[0053] Furthermore, the collected variables related to the prediction of hydropower generation are subjected to correlation screening, and the hydrological variables with high correlation are selected to obtain the corresponding hydrological data.

[0054] In one implementation, Pearson correlation coefficients are calculated between multiple hydrological variables related to hydropower generation prediction and their corresponding historical generation data. A predetermined number of hydrological variables with high correlation are selected as inputs for subsequent model training, and the corresponding hydrological data are extracted. The five hydrological variables with the highest correlation are then selected, and their corresponding hydrological data are extracted.

[0055] Furthermore, the collected data undergoes data preprocessing.

[0056] Specifically, outliers in the collected data are identified using clustering algorithms and the 3-σ principle. Then, missing values ​​are filled and outliers are replaced using moving averages or exponential smoothing. The trend, seasonality, and periodicity of the time series data are identified by combining statistical methods such as unit root tests. Non-stationary time series are transformed into stationary time series using methods such as differencing, deterministic detrending, and structural change analysis. This method can ensure the stationarity of the input data for the time series model. Finally, each time series is normalized.

[0057] Meanwhile, this application collects geographical distance data between any two hydropower stations in the hydropower station cluster.

[0058] Step S202: Construct a power generation matrix time series for the hydropower station cluster based on the historical power generation data of each hydropower station; construct a hydrological matrix time series for the hydropower station cluster based on the hydrological data of each hydropower station; and construct a dataset based on the power generation matrix time series and the hydrological matrix time series.

[0059] Specifically, step S202 includes:

[0060] Step S2021: For the daily historical power generation data, define the matrix rows as each hydropower station and the matrix columns as the power generation of each hydropower station at different times to obtain the daily corresponding power generation matrix.

[0061] For example, when the sampling interval is 15 minutes, for the historical power generation data of each hydropower station, each row is defined as a different hydropower station, and each column represents the power generation collected every 15 minutes each day. That is, each row represents the power generation of a certain hydropower station at 96 times a day, resulting in a power generation matrix corresponding to each day. .

[0062] Step S2022: Construct a time series of the power generation matrix based on the power generation matrix for consecutive dates.

[0063] The power generation matrix time series is constructed by arranging the power generation matrices of different dates in chronological order. .

[0064] Step S2023: For the daily hydrological data, calculate the mean of each hydrological variable for each hydropower station.

[0065] For example, when the sampling time interval is 15 minutes and there are 5 hydrological variables, the mean value of the hydrological data collected at 96 times a day for the 5 different hydrological variables of each hydropower station is calculated.

[0066] Step S2024: Define each row of the matrix as each hydropower station, and define each column of the matrix as the mean of each hydrological variable of each hydropower station, to obtain the daily hydrological matrix.

[0067] For example, for the hydrological data of each hydropower station, the rows are defined as different hydropower stations, and the columns are the averages of five different hydrological variables, resulting in a hydrological matrix for each day. .

[0068] Step S2025: Construct a time series of the hydrological matrix based on the hydrological matrix for consecutive dates.

[0069] Hydrological data matrices for different dates are constructed in chronological order to obtain a hydrological data matrix time series. .

[0070] Step S2026: Construct a dataset based on the power generation matrix time series and the hydrological matrix time series.

[0071] In this implementation, a matrix of power generation from multiple hydropower stations is constructed using dates as the matrix standard, and a hydrological matrix of multiple hydropower stations is also constructed. This approach can take into account the spatial relationships of the data and construct a time series according to the time order, thus providing a good dataset environment for subsequent model training and further improving the model training effect.

[0072] Step S203: Construct a spatial impact prediction term for power generation based on the power generation matrix time series and geographical distance data; construct a temporal impact prediction term for power generation based on the historical moment power generation matrix in the power generation matrix time series; construct a temporal impact prediction term for hydrological conditions based on the historical moment hydrological matrix in the hydrological matrix time series; combine the spatial impact prediction term for power generation, the temporal impact prediction term for power generation, and the temporal impact prediction term for hydrological conditions to construct a matrix value time series spatial autoregressive model.

[0073] The spatial weight matrix of a hydropower station cluster reflects the correlation or association distance between related data of two hydropower stations. Generally, in spatial econometrics, it is assumed that the spatial weight matrix of a hydropower station cluster is known. There are many ways to set the spatial weight matrix, such as the common economic distance spatial weight matrix and the geographical distance spatial weight matrix.

[0074] In this implementation, a spatial weight matrix for hydropower station clusters is constructed based on geographical distance data. The larger the geographical distance data, the smaller the corresponding element value in the spatial weight matrix for hydropower station clusters.

[0075] Specifically, the spatial weight matrix of hydropower station clusters Each element ,in, Represented as the first The hydropower station and the first The geographical distance between the hydropower stations, and .

[0076] In one implementation, the spatial impact prediction of power generation includes a concurrent spatial impact prediction of power generation and a lagged spatial impact prediction of power generation.

[0077] Specifically, the current-time power generation matrix of the power generation matrix time series is combined with the spatial weight matrix of the hydropower station cluster between power stations. The influence of the current hydropower station power generation on the power generation forecast in the geographic spatial dimension is considered to obtain the synchronous forecast term of the spatial influence of power generation. The historical-time power generation matrix of the power generation matrix time series is combined with the spatial weight matrix of the hydropower station cluster between power stations. The influence of the power generation of the lag hydropower station with a certain lag period on the power generation forecast in the geographic spatial dimension is considered to obtain the lag forecast term of the spatial influence of power generation.

[0078] Furthermore, in this implementation method, a matrix-valued time series spatial autoregressive model is constructed by combining the synchronous prediction term of the spatial impact of power generation, the lagged prediction term of the spatial impact of power generation, the temporal impact prediction term of power generation, and the temporal impact prediction term of hydrological conditions.

[0079] Specifically, the constructed matrix-valued time series spatial autoregressive model The specific form is as follows:

[0080] .

[0081] in, The model equation is a time series representation of the power generation matrix, with the left-hand side representing the power generation matrix at the current moment. for A 3D matrix, representing day A hydroelectric power station Observations of power generation at different times.

[0082] The first term on the right side of the model equation is the same-period prediction term of the spatial impact of power generation, which represents the pure spatial effect of power generation.

[0083] Specifically, the current power generation matrix and the spatial weight matrix of the hydropower station cluster are fused to obtain a first fused matrix; the first fused matrix is ​​then adjusted using the first diagonal matrix to obtain the concurrent prediction term for the spatial impact of power generation. yes A 3D spatial weight matrix for a hydropower station cluster, with all elements on the main diagonal being zero, can describe the dependencies between different hydropower stations. and They represent and The first diagonal matrix of dimension . Where, for matrix . The Middle element The prediction, the contribution of the first item No. The concurrent spatial effect of the column It is the first Spatial influence coefficient of each location It is the first The adjustment coefficient of each indicator to spatial effects.

[0084] The second term on the right side of the model equation is the prediction term for the time impact of power generation, which represents the pure dynamic effect of power generation, i.e., the time lag effect.

[0085] Specifically, the power generation matrix at the second lag order is adjusted using the first regression coefficient matrix to obtain the power generation time-related prediction term. Among these, and They are and The first regression coefficient matrix of dimension 1, and They are respectively and The OK, This is the second lag order. Wherein, The Each element can be represented as This indicates that the second term on the right side of the model is important for prediction. The contribution lies in the matrix A linear combination of these represents the lag prediction relationship over time.

[0086] The third term on the right side of the model equation is the spatial impact prediction term of power generation, which represents the spatial effect of power generation time lag.

[0087] Specifically, the historical power generation matrix with the first lag order and the spatial weight matrix of the hydropower station cluster are fused to obtain a second fused matrix; the second fused matrix is ​​then adjusted using the second diagonal matrix to obtain the spatial impact lag prediction term for power generation. Among these, and They represent and The second diagonal matrix of dimension, This is the first lag order. Where, for the matrix... The Middle element The prediction, the contribution of the third item No. The lag space effect of the column It is the first Spatial influence coefficient of each location It is the first The adjustment coefficient of each indicator to spatial effects.

[0088] The fourth term on the right side of the model equation is the prediction term for the time impact of hydrological conditions, representing the time impact effect of hydrological variables.

[0089] Specifically, the historical hydrological data matrix with the third lag order is adjusted using the second regression coefficient matrix to obtain the hydrological time impact prediction term. Among these, The hydrological data matrix is ​​a time series. for A 3D matrix, representing day A hydroelectric power station Observed values ​​of several hydrological variables. This indicates the length of the collected data time series. and They are and The second regression coefficient matrix of dimension . It is the third lag order.

[0090] The fifth term on the right side of the model equation is the error term. It is A dimensional white noise matrix that satisfies ,in , .

[0091] In one implementation, when the lag order is 1... The specific form of the model is as follows:

[0092] (1)

[0093] Step S204: Train the matrix-valued time series spatial autoregressive model based on the dataset to obtain the joint prediction model for power generation of the hydropower station cluster.

[0094] In some optional implementations, step S204 above includes:

[0095] Step S2041: Divide the dataset into training set and test set in time series according to the partition ratio.

[0096] Specifically, the dataset is divided into training and testing sets according to a preset ratio, such as 7:3.

[0097] Step S2042: Using the training set, perform parameter estimation on the matrix-valued time series spatial autoregressive model to obtain the trained matrix-valued time series spatial autoregressive model.

[0098] In one implementation, a one-order lag is used. Taking the model as an example, parameter estimation is performed. Because... It exists on both sides of the model formula, therefore and This correlation will lead to endogeneity problems. In this case, directly using the least squares method... and Regression can lead to inconsistent estimates. Therefore, this paper proposes an estimation method based on the Yule-Walker equation to address the endogeneity problem. The Yule-Walker equation describes the relationship between autoregressive sequence parameters and their covariance function. Since... Therefore, the Yule-Walker equation does not contain information about the error term, which can effectively eliminate the endogeneity problem in the original regression model.

[0099] The MSTLAR model proposed in this application consists of three parts: pure spatial effects, pure dynamic effects (time lag effects), and time-lag spatial effects. Estimation methods for each of these three parts are given, and then an iterative calculation method is constructed. Specifically, given some parameters, the remaining parameters are solved, and then the solution is obtained iteratively. The estimation process mainly consists of the following steps:

[0100] Step 1: Estimate the time-related impact of power generation.

[0101] Specifically, given a matrix , , , , and The estimation result is , , , , and .

[0102] definition At this point, equation (1) is transformed into The model does not contain endogeneity issues. In this case, the least squares method can be used to obtain... and The parameter optimization problem for the consistency estimation can be expressed as:

[0103] (2)

[0104] To optimize the objective function given in equation (2), the matrix Frobenius norm is first transformed into the form of the matrix trace. Then, according to the rule of differentiation of the trace with respect to the matrix, the following steps are performed: and Taking the partial derivative and setting it to 0, we get:

[0105] (3)

[0106] For matrix and The calculation can be performed using the following iterative estimation: given a matrix Estimate Based on equation (3), the matrix can be obtained. Updated estimates :

[0107] , (4)

[0108] Similarly, given a matrix Estimate Based on equation (3), the matrix can be obtained. Updated estimates :

[0109] (5)

[0110] Step 2: Estimate the time-related impact of water conditions on the prediction.

[0111] Specifically, similar to step 1, in a given matrix , , , , and The estimation result is , , , , and Afterwards, since the model does not have an endogeneity problem, the parameter estimation results can be obtained using the least squares method:

[0112] , ;

[0113] .

[0114] in, .

[0115] Step 3: Estimate the spatial impact lag forecast term of power generation.

[0116] Specifically, given a matrix , , and The estimation result is , , and Solve the parameter matrix and The objective function is defined as:

[0117] (6)

[0118] in, Similarly, since there are no endogenous problems in the parameter solution at this time, an iterative estimation method is used to solve the problem. and The least squares estimate.

[0119] Given Estimate ,definition The minimization problem given in equation (6) can be rewritten as follows: Regularly extract all items related to... The relevant formulas, at this time regarding The objective function is:

[0120] (7)

[0121] in, and They are respectively and The Okay. In order to solve The least squares estimate of the objective function (7) is given by... Taking the derivative and setting its partial derivatives to 0, we get:

[0122] (8)

[0123] In estimation Next, normalization is required. This makes its Frobenius norm 1.

[0124] Similarly, given the parameter matrix Estimate You can get The updated estimate is:

[0125] (9)

[0126] in, , and They are respectively and The List.

[0127] Step 4: Estimate the spatial impact of power generation on the same period forecast.

[0128] Specifically, given the parameter matrix , , and The estimate is , , and Solving the parameter matrix based on the Yule–Walker equation and For any ,definition Based on model (3.1), the following Yule–Walker equation can be obtained:

[0129] (10)

[0130] in, yes A dimensional identity matrix. Regularly select elements that are... The relevant lines,

[0131] , .

[0132] in, Represents the spatial weight matrix The OK, Represents the unit array The OK, for The Okay. Because Use the sample covariance matrix instead and ,

[0133] , .

[0134] Then, based on the least squares method, the following objective function is solved by minimizing it. :

[0135] .

[0136] definition By analyzing the above objective function with respect to

[0137] Taking the partial derivative and setting it to zero, we can obtain... The generalized Yule–Walker estimate is:

[0138] (11)

[0139] In estimation Next, normalization is required. This makes its Frobenius norm 1.

[0140] Similarly, regularly extracting from equation (10) the elements related to... Relevant lines:

[0141] , .

[0142] It can be obtained The generalized Yule–Walker estimate is:

[0143] (12)

[0144] in, .

[0145] In summary, the overall iterative solution process can be summarized as follows: ① Given a matrix , , , , and initial value , , , , and Using the matrix given in step 1 and initial value and ;②For the first In the next iteration, the current estimation result is updated using steps 1-3. ③ Repeat step ② until convergence, and give the final parameter estimate. This paper sets the number of iterations to 100. The iteration stopping condition is set as follows: the number of iterations reaches 100, or the sum of the absolute differences of all parameter estimates in two consecutive iterations is less than 0.001. The parameter estimation process of model (1) can be summarized as an iterative algorithm based on the Yule-Walker equation, as follows:

[0146] The algorithm input is: random initial value. .

[0147] The algorithm outputs: parameter estimates. .

[0148] The algorithm steps are as follows:

[0149] 1. , .

[0150] .

[0151] 2. , .

[0152] .

[0153] 3. , .

[0154] .

[0155] 4. , .

[0156] .

[0157] 5. Repeat steps 1-3 until the iterative estimation of each coefficient matrix converges, and output the estimated values:

[0158] .

[0159] In this implementation, the spatial impact prediction term of power generation, the time impact prediction term of power generation, the spatial impact prediction term of power generation, and the time impact prediction term of hydrological conditions are introduced into the matrix-valued time series spatial autoregressive model. This not only considers the impact of historical data and hydrological data, but also the spatial effect impact of different hydropower stations, making it more suitable for the joint power generation prediction of the current hydropower station cluster and improving the accuracy of subsequent model applications.

[0160] In some optional implementations, step S2042 above includes:

[0161] Step a1: Obtain multiple preset lag order combinations.

[0162] Generally, when performing time series based on a matrix value... When fitting a model, it is not possible to know in advance... , and The optimal order. , and If the value is chosen to be too small, the model may underfit. , and A large value may indicate overfitting in the model. To ensure the model achieves a good fit on the existing data while maintaining a low prediction error, it is necessary to adjust the model accordingly. , and The order of lag needs to be carefully selected. Therefore, this application presupposes multiple combinations of lag orders.

[0163] Step a2: Using the training set, perform parameter estimation on the matrix-valued time series spatial autoregressive models corresponding to multiple preset lag order combinations to obtain multiple trained matrix-valued time series spatial autoregressive models.

[0164] Based on the corresponding preset lag order combinations, a training set is determined. Then, using the Yule-Walker equation and least squares method described above, parameters are estimated for the matrix-valued time series spatial autoregressive model corresponding to each preset lag order combination, resulting in multiple trained matrix-valued time series spatial autoregressive models. Specific parameter estimation methods are described above and will not be repeated here.

[0165] Step S2043: Test the trained matrix-valued time series spatial autoregressive model based on the test set to determine the lag order and obtain the joint prediction model for power generation of the hydropower station cluster.

[0166] In some optional implementations, step S2043 above includes:

[0167] Step b1: Calculate the model error of multiple trained matrix-valued time-series spatial autoregressive models based on the test set.

[0168] Among them, for The model is constructed to estimate the covariance of the residuals as the error function. :

[0169] .

[0170] in, , These are the test values ​​for the trained matrix-valued time-series spatial autoregressive model. These are the actual values ​​for the test set.

[0171] Step b2: Calculate the Bayesian information content criterion value based on the error function.

[0172] Specifically, the Bayesian Information Criterion (BIC) value is:

[0173] .

[0174] in, express The value of the determinant.

[0175] Step b3: Select the preset lag order combination with the minimum Bayesian information criterion value, and determine the lag order to obtain the corresponding joint prediction model for power generation.

[0176] The optimal combination of lag orders is obtained by minimizing the Bayesian information criterion value. This leads to the corresponding joint prediction model for power generation.

[0177] In this implementation, different combinations of orders are verified and selected to determine the optimal combination. This avoids the problem of underfitting the model when the order is too small, and overfitting the model when the order is too large. This method enables the model to show a good fit on the existing data and has less prediction error.

[0178] Step S205: Predict the power generation of the hydropower station cluster based on the joint power generation prediction model.

[0179] Historical power generation data and multiple hydrological data corresponding to the lag order are obtained, and power generation matrix and hydrological matrix are constructed. These are then input into the joint power generation prediction model of the hydropower station cluster, and the predicted output is the power generation at future time.

[0180] The joint prediction method for power generation of a hydropower station cluster provided in this embodiment can predict the power generation of a hydropower station cluster while maintaining the matrix value data structure. It comprehensively considers the hydrological influencing factors of the hydropower station cluster and the correlation characteristics between hydropower stations. It not only considers the mutual influence relationship in the time dimension, but also the correlation relationship in the spatial dimension between hydropower station clusters. In terms of parameter estimation, iterative estimation by combining the Yule-Walker equation and the least squares method can solve the endogeneity problem of the model. In addition, the modeling process generates a large number of parameters to be estimated. This application solves the problem that too many parameters may lead to parameter inflation and overfitting under the same sample data.

[0181] This embodiment also provides a joint prediction device for the power generation of a hydropower station cluster. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0182] This embodiment provides a joint prediction device for the power generation of a hydropower station cluster, such as... Figure 3 As shown, it includes:

[0183] The data acquisition module 301 is used to collect historical power generation data and multiple hydrological data related to power generation for each hydropower station in the hydropower station cluster, as well as geographical distance data between hydropower stations.

[0184] The first construction module 302 is used to construct a power generation matrix time series of the hydropower station cluster based on the historical power generation data of each hydropower station; construct a hydrological matrix time series of the hydropower station cluster based on the hydrological data of each hydropower station; and construct a dataset based on the power generation matrix time series and the hydrological matrix time series.

[0185] The second construction module 303 is used to construct a spatial impact prediction term for power generation based on the power generation matrix time series and the geographical distance data, a temporal impact prediction term for power generation based on the historical moment power generation matrix in the power generation matrix time series, and a temporal impact prediction term for hydrological conditions based on the historical moment hydrological matrix in the hydrological matrix time series; and to construct a matrix value time series spatial autoregressive model by combining the spatial impact prediction term for power generation, the temporal impact prediction term for power generation, and the temporal impact prediction term for hydrological conditions.

[0186] Training module 304 is used to train the matrix-valued time series spatial autoregressive model based on the dataset to obtain a joint prediction model for the power generation of the hydropower station cluster.

[0187] The prediction module 305 is used to predict the power generation of the hydropower station cluster based on the joint power generation prediction model.

[0188] In some alternative implementations, the first building module 302 includes:

[0189] The first matrix construction unit is used to define the matrix rows as each hydropower station and the matrix columns as the power generation of each hydropower station at different times for the daily historical power generation data, so as to obtain the corresponding daily power generation matrix.

[0190] The first sequence construction unit is used to construct a time series of the power generation matrix based on the power generation matrix of consecutive dates.

[0191] The calculation unit is used to calculate the mean of each hydrological variable for each hydropower station based on daily hydrological data.

[0192] The second matrix construction unit is used to define each row of the matrix as each hydropower station and each column of the matrix as the mean of each hydrological variable of each hydropower station, so as to obtain the corresponding hydrological matrix for each day.

[0193] The second sequence construction unit is used to construct a time series of hydrological data based on a continuous date hydrological data matrix.

[0194] In some alternative implementations, the second building module 303 includes:

[0195] The first model construction unit is used to construct a spatial weight matrix of hydropower station clusters based on the geographical distance data. The larger the geographical distance data, the smaller the corresponding element value in the spatial weight matrix of hydropower station clusters. It is used to construct a synchronous prediction term of the spatial impact of power generation based on the current moment power generation matrix of the power generation matrix time series and the spatial weight matrix of hydropower station clusters. It is also used to construct a lag prediction term of the spatial impact of power generation based on the historical moment power generation matrix of the power generation matrix time series and the spatial weight matrix of hydropower station clusters.

[0196] In some alternative implementations, the second building module 303 includes:

[0197] The first model construction unit is used to fuse the current-time power generation matrix and the hydropower station cluster spatial weight matrix to obtain a first fusion matrix; and to adjust the first fusion matrix using a first diagonal matrix to obtain the synchronous prediction term of the spatial impact of power generation. The unit then fuses the historical-time power generation matrix with a first lag order and the hydropower station cluster spatial weight matrix to obtain a second fusion matrix; and to adjust the second fusion matrix using a second diagonal matrix to obtain the lag prediction term of the spatial impact of power generation.

[0198] The second model construction unit is used to adjust the historical power generation matrix of the second lag order using the first regression coefficient matrix to obtain the power generation time impact prediction term.

[0199] The third model construction unit is used to adjust the historical time hydrological matrix of the third lag order using the second regression coefficient matrix to obtain the hydrological time impact prediction term.

[0200] In some alternative implementations, training module 304 includes:

[0201] A partitioning unit is used to divide a dataset into a training set and a test set in time series according to a partitioning ratio.

[0202] The training unit is used to estimate the parameters of the matrix-valued time series spatial autoregressive model using the training set, and to determine the lag order to obtain the trained matrix-valued time series spatial autoregressive model.

[0203] The test unit is used to test the trained matrix-valued time series spatial autoregressive model based on the test set, so as to obtain the joint prediction model of power generation of the hydropower station cluster.

[0204] In some alternative implementations, the training unit includes:

[0205] The sub-unit is used to obtain multiple preset lag order combinations, which include a preset first lag order, a preset second lag order, and a preset third lag order.

[0206] The training subunit is used to estimate the parameters of matrix-valued time series spatial autoregressive models corresponding to multiple preset lag order combinations using the training set, and determine the lag order to obtain multiple trained matrix-valued time series spatial autoregressive models.

[0207] In some alternative implementations, the test unit includes:

[0208] The first computational subunit is used to calculate the model error of multiple trained matrix-valued time-series spatial autoregressive models based on the test set.

[0209] The second calculation subunit is used to calculate the Bayesian information criterion value based on the error function, and select the preset lag order combination when the Bayesian information criterion value is minimized to obtain the corresponding joint prediction model for power generation.

[0210] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0211] In this embodiment, the hydropower station cluster power generation joint prediction device is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0212] This invention also provides a computer device having the above-described features. Figure 4 The device shown is a joint prediction device for the power generation of a cluster of hydropower stations.

[0213] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 4As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 4 Take a processor 10 as an example.

[0214] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0215] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0216] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0217] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0218] The computer device also includes an input device 30 and an output device 40. The processor 10, memory 20, input device 30, and output device 40 can be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0219] Input device 30 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the computer device, such as a touchscreen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 40 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The aforementioned display devices include, but are not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touchscreen.

[0220] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0221] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0222] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for joint prediction of power generation in a cluster of hydropower stations, characterized in that, The method includes: Collect historical power generation data for each hydropower station in the hydropower station cluster and multiple hydrological data related to power generation prediction; collect geographical distance data between hydropower stations; A power generation matrix time series of the hydropower station cluster is constructed based on the historical power generation data of each hydropower station; a hydrological matrix time series of the hydropower station cluster is constructed based on the hydrological data of each hydropower station; a dataset is constructed based on the power generation matrix time series and the hydrological matrix time series. Based on the power generation matrix time series and the geographical distance data, a spatial impact prediction term for power generation is constructed; based on the historical moment power generation matrix in the power generation matrix time series, a temporal impact prediction term for power generation is constructed; based on the historical moment hydrological matrix in the hydrological matrix time series, a temporal impact prediction term for hydrological conditions is constructed; combining the spatial impact prediction term for power generation, the temporal impact prediction term for power generation, and the temporal impact prediction term for hydrological conditions, a matrix value time series spatial autoregressive model is constructed. The matrix-valued time series spatial autoregressive model is trained based on the dataset to obtain the joint prediction model for the power generation of the hydropower station cluster. The power generation of the hydropower station cluster is predicted based on the aforementioned joint power generation prediction model. The construction of a spatial impact prediction term for power generation based on the power generation matrix time series and the geographical distance data includes: A spatial weight matrix for hydropower station clusters is constructed based on the geographical distance data. The larger the geographical distance data, the smaller the corresponding element value in the spatial weight matrix for hydropower station clusters. Based on the current moment's power generation matrix of the power generation matrix time series and the spatial weight matrix of the hydropower station cluster, a synchronous prediction term for the spatial impact of power generation is constructed. Based on the historical moment power generation matrix of the power generation matrix time series and the spatial weight matrix of the hydropower station cluster, a spatial impact lag prediction term for power generation is constructed.

2. The method for joint prediction of power generation of a hydropower station cluster according to claim 1, characterized in that, The construction of the power generation matrix time series of the hydropower station cluster based on the historical power generation data of each hydropower station includes: For the historical power generation data for each day, the matrix rows are defined as each hydropower station, and each column of the matrix is ​​defined as the power generation of each hydropower station at different times, thus obtaining the power generation matrix corresponding to each day; Based on the power generation matrix for consecutive dates, construct a time series of the power generation matrix; The construction of the hydrological matrix time series of the hydropower station cluster based on the hydrological data of each hydropower station includes: For the daily hydrological data, calculate the mean value of each hydrological variable for each hydropower station; Each row of the matrix is ​​defined as each hydropower station, and each column of the matrix is ​​defined as the mean of each hydrological variable of each hydropower station, thus obtaining the daily hydrological matrix. Based on the hydrological information matrix with consecutive dates, a time series of the hydrological information matrix is ​​constructed.

3. The method for joint prediction of power generation of a hydropower station cluster according to claim 1, characterized in that, The construction of a synchronous prediction term for the spatial impact of power generation based on the current-time power generation matrix of the power generation matrix time series and the spatial weight matrix of the hydropower station cluster includes: The current power generation matrix and the hydropower station cluster spatial weight matrix are fused to obtain the first fusion matrix; The first fusion matrix is ​​adjusted using the first diagonal matrix to obtain the synchronous prediction term of the spatial impact of power generation; The construction of a spatial impact lag prediction term for power generation based on the historical moment power generation matrix of the power generation matrix time series and the spatial weight matrix of the hydropower station cluster includes: The historical moment power generation matrix of the first lag order and the spatial weight matrix of the hydropower station cluster are fused to obtain the second fused matrix; The second fusion matrix is ​​adjusted using the second diagonal matrix to obtain the spatial impact lag prediction term of power generation; The step of constructing a power generation time impact prediction term based on the historical time-series power generation matrix in the power generation matrix time series includes: The power generation matrix at the historical time with the second lag order is adjusted using the first regression coefficient matrix to obtain the power generation time impact prediction term; The construction of the hydrological time impact prediction term based on the historical time hydrological matrix in the hydrological matrix time series includes: The historical hydrological information matrix with the third lag order is adjusted using the second regression coefficient matrix to obtain the hydrological time impact prediction term.

4. The method for joint prediction of power generation of a hydropower station cluster according to claim 3, characterized in that, The step of training the matrix-valued time-series spatial autoregressive model based on the dataset to obtain the joint prediction model for the power generation of the hydropower station cluster includes: The dataset is divided into a training set and a test set in time series according to the division ratio; Using the training set, the matrix-valued time series spatial autoregressive model is used to estimate parameters, resulting in a trained matrix-valued time series spatial autoregressive model. The trained matrix-valued time series spatial autoregressive model is tested based on the test set to determine the lag order and obtain the joint prediction model for the power generation of the hydropower station cluster.

5. The method for joint prediction of power generation of a hydropower station cluster according to claim 4, characterized in that, The step of estimating the parameters of the matrix-valued time series spatial autoregressive model to obtain the trained matrix-valued time series spatial autoregressive model includes: Obtain multiple preset lag order combinations, wherein the preset lag order combinations include a preset first lag order, a preset second lag order, and a preset third lag order; Using the training set, parameter estimation is performed on the matrix-valued time series spatial autoregressive models corresponding to multiple combinations of the preset lag orders to obtain multiple trained matrix-valued time series spatial autoregressive models; The step of testing the trained matrix-valued time-series spatial autoregressive model based on the test set to determine the lag order and obtain the joint prediction model for the power generation of the hydropower station cluster includes: The model error of the multiple trained matrix-valued time series spatial autoregressive models is calculated based on the test set. The Bayesian information criterion value is calculated based on the model error, and the preset lag order combination with the smallest Bayesian information criterion value is selected to obtain the corresponding joint prediction model for power generation.

6. A joint prediction device for the power generation of a hydropower station cluster, characterized in that, The device includes: The data acquisition module is used to collect historical power generation data and multiple hydrological data related to power generation from each hydropower station in the hydropower station cluster; and to collect geographical distance data between hydropower stations. The first construction module is used to construct a power generation matrix time series of the hydropower station cluster based on the historical power generation data of each hydropower station; construct a hydrological matrix time series of the hydropower station cluster based on the hydrological data of each hydropower station; and construct a dataset based on the power generation matrix time series and the hydrological matrix time series. The second construction module is used to construct a spatial impact prediction item for power generation based on the power generation matrix time series and the geographical distance data; a temporal impact prediction item for power generation based on the historical moment power generation matrix in the power generation matrix time series; and a temporal impact prediction item for hydrological conditions based on the historical moment hydrological matrix in the hydrological matrix time series. It then combines the spatial impact prediction item for power generation, the temporal impact prediction item for power generation, and the temporal impact prediction item for hydrological conditions to construct a matrix value time series spatial autoregressive model. The construction of the spatial impact prediction item for power generation based on the power generation matrix time series and the geographical distance data includes: constructing a hydropower station cluster spatial weight matrix based on the geographical distance data, where the larger the geographical distance, the smaller the corresponding element value in the hydropower station cluster spatial weight matrix; constructing a concurrent spatial impact prediction item for power generation based on the current moment power generation matrix of the power generation matrix time series and the hydropower station cluster spatial weight matrix; and constructing a lag prediction item for power generation spatial impact based on the historical moment power generation matrix of the power generation matrix time series and the hydropower station cluster spatial weight matrix. The training module is used to train the matrix-valued time series spatial autoregressive model based on the dataset to obtain the joint prediction model of the power generation of the hydropower station cluster. The prediction module is used to predict the power generation of the hydropower station cluster based on the joint power generation prediction model.

7. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the joint prediction method for power generation of a hydropower station cluster as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the joint prediction method for power generation of a hydropower station cluster as described in any one of claims 1 to 5.

9. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the joint prediction method for power generation of a hydropower station cluster as described in any one of claims 1 to 5.

Images