Cascade hydropower station output prediction method and system

Abstract

The invention discloses a cascade hydropower station output prediction method and system, and the method comprises the steps: collecting various historical data of a cascade hydropower station, analyzing the load condition of the hydropower station, and obtaining the influence relation of upstream and downstream outputs; carrying out clustering analysis on factors which have the maximum influence on the output of the hydropower station according to seasonal changes, and determining an optimal scheduling period for output prediction; constructing a load prediction model, and predicting the future output condition of the cascade hydropower station in the optimal scheduling period; and predicting a future flow water level, and associating the future flow water level with the output prediction result to obtain a cascade hydropower station operation scheduling result. According to the method, the historical condition of the output load of the cascade power station is specifically analyzed, and the load rate change trend of the cascade power station is effectively predicted through clustering analysis and establishment of the hydropower station load prediction model, so that an important reference is provided for operation and scheduling of the hydropower station, and the stability and reliability of power supply are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower station output prediction, and in particular to a cascade hydropower station output prediction method and system. Background Art

[0002] The existing output forecast analysis of cascade hydropower stations often focuses on a single or a few explicit factors, such as inflow, reservoir regulation performance, etc., but ignores the in-depth analysis of other potential factors (such as equipment aging, maintenance status, seasonal influences, etc.). The research on the interaction and comprehensive impact of various factors is not in-depth enough, making it difficult to comprehensively and accurately evaluate the overall output capacity of the hydropower station.

[0003] The monitoring equipment and technical means are relatively backward, resulting in incomplete and inaccurate data collection, which cannot provide reliable data support for in-depth analysis. The data processing and analysis capabilities are limited, making it difficult to extract valuable information from massive data to support decision-making.

[0004] Existing output prediction models are often based on simplified assumptions, which make it difficult to fully reflect the complex situations in actual operation, resulting in inaccurate prediction results, imperfect model updating and verification mechanisms, and inability to make timely adjustments and optimizations based on new data and new situations. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for predicting the output of cascade hydropower stations to solve the above problems.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for predicting the output of a cascade hydropower station, comprising: collecting various historical data of the cascade hydropower station, analyzing the load conditions of the hydropower station, and obtaining the influence relationship between the upstream and downstream outputs;

[0009] Based on the load situation analysis, the factors that have the greatest impact on the output of the hydropower station are clustered according to seasonal changes to obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling cycle for output prediction;

[0010] Construct a load forecasting model to predict the future output of the cascade hydropower station during the optimal dispatching period and obtain the output forecast results;

[0011] Based on the historical data, the future flow and water level are predicted and associated with the output prediction result to obtain the operation and dispatch result of the cascade hydropower station.

[0012] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the analysis of the load conditions of the hydropower station includes:

[0013] The historical data includes the time period, water level, flow rate and output load data of the hydropower station, and the data is pre-processed;

[0014] Calculate the load rate of the hydropower station in different seasons or months, which is the ratio of the actual output of the power system to the rated output within a certain period of time;

[0015] By drawing a time series graph, the daily load value changes of each hydropower station are displayed, and the daily average load value of each hydropower station is analyzed and compared with the load fluctuation threshold;

[0016] If the daily average load value is greater than the load fluctuation threshold, the time period or date corresponding to the daily average load value is marked;

[0017] Compare the load characteristics between different hydropower stations to determine whether there are similar load fluctuation patterns.

[0018] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the method for determining the optimal scheduling period for output prediction includes:

[0019] Performing cluster analysis on the historical data to obtain a solution, averaging multiple columns of each feature in the historical data, and selecting a group average as a measurement method for the distance between groups;

[0020] The clustering solution results are visualized in three-dimensional space, where the position of each point is determined by the factor that has the greatest impact on the output of the hydropower station. Each cluster is represented by a different color, and the clustering pattern of the data points in three-dimensional space is obtained;

[0021] The K-means clustering in three-dimensional space shows the relationship between multiple features of the factors that have the greatest impact on the output of the hydropower station. The data points are grouped according to the similarity of the features to obtain the optimized scheduling period for output prediction.

[0022] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the cluster analysis solution includes:

[0023] Step 1: randomly select k objects from N sample data as the initial clustering centers;

[0024] Step 2: Calculate the distance from each sample to each cluster center, and then assign the object to the cluster with the closest distance;

[0025] Step 3: After all objects are assigned, recalculate the centers of the k clusters;

[0026] Step 4: Compare the k cluster centers calculated last time. If the cluster center changes, proceed to step 2 again; otherwise, proceed directly to step 5.

[0027] Step 5: When the centroid does not change, stop and output the clustering results.

[0028] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the construction of the load prediction model includes:

[0029] A differential equation is established for the discrete sequence, a first-order differential equation model and a discrete and prediction formula are constructed, and the load forecasting model is obtained by derivative definition and form conversion as follows:

[0030]

[0031] in, Indicates the value of the sequence generated at time point k+1, x (1) (1) represents the value of the cumulative generated sequence at the initial time point k=1, represents the attenuation factor, Represents the steady-state value of the model.

[0032] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the first-order differential equation model and the discrete and predictive formulas are constructed as follows:

[0033] The first-order differential equation model is expressed as:

[0034]

[0035] in, represents the first-order derivative of the dependent variable x with respect to time t, that is, the rate of change of x with time t, x represents the dependent variable, that is, the predicted system state or data sequence, t represents the time variable, u represents the exogenous variable, that is, the external influencing factor, and a represents the model parameter, that is, the resistance coefficient of the change of x;

[0036] The discrete and predictive formulas are expressed as:

[0037]

[0038] Among them, x (1) (k+1) means the value of the generated sequence is accumulated at time point k+1.

[0039] As a preferred solution of the method for predicting the output of cascade hydropower stations described in the present invention, the accuracy verification of the load prediction model includes:

[0040] Assume that the load forecasting model has been obtained and will Do a cumulative conversion to It is expressed as:

[0041]

[0042] in, Indicates the value at the nth time point after the cumulative-subtraction conversion;

[0043] Calculate the residual e(k), expressed as:

[0044]

[0045] Among them, x (0) (k) represents the original data sequence at the kth time point;

[0046] The original sequence x (0) And the variance of the residual sequence E is It is expressed as:

[0047]

[0048] in, Represents the original sequence x (0) The average value over n time frames, represents the corresponding residual sequence mean;

[0049] The average values ​​are expressed as:

[0050]

[0051] The calculation of the posterior difference ratio C is expressed as:

[0052]

[0053] In a second aspect, the present invention provides a cascade hydropower station output prediction system, comprising:

[0054] The acquisition module is used to collect various historical data of cascade hydropower stations, analyze the load conditions of the hydropower stations, and obtain the influence relationship between upstream and downstream output;

[0055] An analysis module is used to analyze the load situation, perform cluster analysis on the factors that have the greatest impact on the output of the hydropower station based on seasonal changes, obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling period for output prediction;

[0056] The prediction module is used to build a load prediction model, predict the future output of the cascade hydropower station during the optimal dispatching period, and obtain the output prediction results;

[0057] The scheduling module is used to predict the future flow and water level based on the historical data, and associate it with the output prediction result to obtain the operation scheduling result of the cascade hydropower station.

[0058] In a third aspect, the present invention provides an electronic device, comprising:

[0059] Memory and processor;

[0060] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the cascade hydropower station output prediction method are implemented.

[0061] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for predicting the output of cascade hydropower stations.

[0062] Compared with the prior art, the present invention has the following beneficial effects: the present invention realizes a specific analysis of the historical output load of cascade power stations through the analysis of historical data, and through cluster analysis, the various operating conditions of the hydropower station can be inferred from the relevant factors affecting the output of the hydropower station, such as flow rate and water level, and a load prediction model for the hydropower station can be established, which can effectively predict the changing trend of the future load rate, provide an important reference for the operation and scheduling of the hydropower station, and help ensure the stability and reliability of the power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0064] Figure 1 It is a schematic diagram of the overall process of a method for predicting the output of a cascade hydropower station according to an embodiment of the present invention;

[0065] Figure 2 A schematic diagram of load rate fluctuation of a typical output curve of power station A in a method for predicting output of cascade hydropower stations according to an embodiment of the present invention;

[0066] Figure 3A load rate fluctuation diagram of a typical output curve of a power station B in a method for predicting the output of a cascade hydropower station according to an embodiment of the present invention;

[0067] Figure 4 A load rate fluctuation diagram of a typical output curve of a power station C in a method for predicting the output of a cascade hydropower station according to an embodiment of the present invention;

[0068] Figure 5 A load rate fluctuation diagram of a typical output curve of a power station D in a method for predicting the output of a cascade hydropower station according to an embodiment of the present invention;

[0069] Figure 6 A load rate fluctuation diagram of a typical output curve of a power station E in a method for predicting the output of a cascade hydropower station according to an embodiment of the present invention;

[0070] Figure 7 A load rate fluctuation diagram of a typical output curve of power station A during flood season of a cascade hydropower station output prediction method according to an embodiment of the present invention;

[0071] Figure 8 A load rate fluctuation diagram of a typical output curve of power station A during dry season of a cascade hydropower station output prediction method according to an embodiment of the present invention;

[0072] Fig. 9 A cluster diagram of the dry season of power station A of the cascade hydropower station output prediction method according to an embodiment of the present invention;

[0073] Fig.10 A cluster diagram of power station A during flood season in a method for predicting the output of cascade hydropower stations according to an embodiment of the present invention;

[0074] Fig.11 A schematic diagram of load prediction results of power station A of a cascade hydropower station output prediction method according to an embodiment of the present invention;

[0075] Fig.12 This is a load prediction curve diagram of power station A of a cascade hydropower station output prediction method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0077] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0078] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0079] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0080] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0081] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0082] Example 1

[0083] Reference Figure 1-Figure 8 , is an embodiment of the present invention, and provides a method for predicting the output of a cascade hydropower station, such as Figure 1 As shown, including:

[0084] S100, collect various historical data of cascade hydropower stations, analyze the load conditions of the hydropower stations, and obtain the influence relationship between upstream and downstream output.

[0085] Preferably, analyzing the load condition of the hydropower station includes:

[0086] The historical data includes the time period, water level, flow rate and output load data of the hydropower station, and the data is pre-processed;

[0087] Calculate the load rate of hydropower stations in different seasons or months. The load rate is the ratio of the actual output of the power system to the rated output within a certain period of time.

[0088] By drawing a time series graph, the daily load value changes of each hydropower station are displayed, and the daily average load value of each hydropower station is analyzed and compared with the load fluctuation threshold;

[0089] If the daily average load value is greater than the load fluctuation threshold, the time period or date corresponding to the daily average load value is marked;

[0090] Compare the load characteristics between different hydropower stations to determine whether there are similar load fluctuation patterns.

[0091] It should be noted that the dispatching data of hydropower stations can be roughly divided into two categories: one is the time category, such as the seasonal characteristics of the hydropower station. For example, the rainfall pattern in the province where the X hydropower station is located significantly affects the load demand of the hydropower station. In the flood season, due to the abundant water resources, the hydropower station usually operates at full load and hardly needs to participate in peak regulation. In the dry season, due to the reduction of flow, the hydropower station will maintain a higher water level to maximize the power generation potential. Although the power generation is small, they play an important role in peak regulation and help smooth the overall load curve of the power grid. Therefore, in order to accurately reflect the differences in dispatching strategies in different seasons, model training should be carried out separately according to the flood season and the dry season. The other category is numerical data, such as water level, flow, and power station output load. When collecting these data, sensors are usually relied on, but data integrity is often challenged. For example, water level data is recorded once an hour. If the sensor fails, it may result in zero value records for several consecutive days, or sudden changes at the end of the measurement. In the data cleaning stage, we need to fill in missing values ​​by comparing similar historical situations, and make reasonable inferences and corrections for sudden water level changes. The processing methods of flow data and power station output load are similar to this and will not be described in detail here.

[0092] Specifically, Figure 2-Figure 8 The analysis of a cascade hydropower station from 1 to 12 months is to gain a deeper understanding of the changing trend of the load rate between different seasons or months. The load rate refers to the ratio of the load of the power system to its rated capacity within a certain period of time. It is one of the important indicators for measuring the operating status and supply and demand balance of the power system. In different seasons and months, the load rate is often affected by various factors, such as temperature changes, seasonal demand, industrial production activities, etc. Therefore, analyzing the load rate in different months can help us better understand the impact of these influencing factors on the operation of the power system.

[0093] It should be noted that, firstly, the change of load factor in different months reflects the difference of seasonal demand. For example, in summer, due to the rising temperature, people usually use electrical equipment such as air conditioners, which leads to an increase in electricity load demand and a corresponding increase in load factor; while in winter, the increase in the use of heating equipment will also have an impact on the load factor, but the trend may be opposite to that in summer. Therefore, analyzing the changes in load factor in different months helps us understand the impact of seasonal demand on the power system. Secondly, analyzing the load factor in different months can help with power system planning and management. By understanding the changing trend of load factor in different months, energy suppliers, government agencies and enterprises can formulate corresponding energy production and use strategies to meet the electricity demand in different months and optimize the utilization efficiency of energy resources. In addition, the analysis of load factor can also help predict future energy demand and provide a reference for the operation and dispatch of the power system to ensure the stability and reliability of power supply.

[0094] By drawing a time series graph, we can intuitively display the daily load value changes of each hydropower station, analyze the average daily load value of each hydropower station, find out the time periods or dates with large load fluctuations, compare the load characteristics between different hydropower stations, and check whether there are similar load fluctuation patterns.

[0095] In the attached figure, the table title “Load Rate Variation” indicates that this is a monitoring of load rate variation. Load rate refers to the ratio of actual load to maximum load, which is an important indicator for measuring the operating efficiency of a power plant. The X-axis represents the time of day (hours), ranging from 0 to 24 hours, while the Y-axis represents the load rate, ranging from 0 to 1.

[0096] Figure 2 The load factor of the A hydropower station starts from 0.6 and shows a slight upward trend over time, reaching a high point of nearly 0.7 at noon, and then stabilizes, which shows that the A hydropower station maintains a high operating load for most of the time, which may reflect a relatively stable power demand during the day; Figure 3 The load factor of the B power station fluctuates around 0.5, with a slight increase for a period of time before noon, and a decrease and stabilization in the afternoon. The load factor of the B power station fluctuates less, which may mean that its power output is relatively balanced and has better adaptability to load changes; Figure 4 The load factor of the C power station remained flat at around 0.5, with a sudden drop at noon and then a rapid recovery, which reflects some kind of scheduling operation or a short-term change in load demand. Overall, the load factor of the C power station is relatively stable. Figure 5The load rate of the D power station starts from a low level, gradually increases over time, reaches a peak in the evening, and then drops sharply. This pattern shows that the D station is greatly affected by the daily variation, gradually increasing its output during the day to meet the peak demand, while the demand decreases at night; the load rate of the E hydropower station in the figure always remains between 0.6 and 0.7, which shows that its load variation is very small and has a high load demand stability. Among all the power stations, the E power station shows the most stable operation mode.

[0097] Taking power station A as an example, the load rate variation curves for 24 days in the flood season and dry season are shown in Figure 2. Figure 7 , Figure 8 , Figure 7 The load factor fluctuation diagram of the typical output curve of Power Station A during the flood season is shown. From 0 o'clock to 4 o'clock, the load factor remains at around 0.57, which reflects the relatively stable electricity demand at night. From 4 o'clock to 10 o'clock, the load factor decreases slightly to about 0.55. From 10 o'clock to 12 o'clock, there is a small increase in the load factor, which may indicate that residential electricity consumption has reached the peak in the afternoon, showing that the daytime electricity demand is relatively stable, which may be related to commercial and industrial activities. From 12 o'clock to 16 o'clock, the load factor decreases slightly again, which may be due to the increase in electricity demand during the evening peak. From 16 o'clock to 24 o'clock, the load factor gradually increases, eventually reaching a level of about 0.58, which may be due to the increase in electricity demand during the evening peak. Overall, the fluctuation of the load factor shows that the overall fluctuation of electricity demand at different stages of the day is small, and it is basically stable at around 0.56.

[0098] Figure 8 The figure shows the load rate fluctuation diagram of the typical output curve of Power Station A during the dry season. From 0 o'clock to 6 o'clock, the load rate is maintained at around 0.28. The low load rate during this period may be due to the low electricity demand of residents at night and the cessation of most industrial and commercial activities, reflecting the static electricity consumption pattern at night. Starting from 8 o'clock, the load rate increased significantly, first a small increase to 0.33, and then reached 0.38 at 10 o'clock. The increase in this stage shows that residents began their daily activities and commercial and industrial activities gradually started, resulting in a sharp increase in electricity demand. Between 10 o'clock and 24 o'clock, the load rate basically remained at 0.38, indicating the stability and peak of daytime electricity demand, which may be related to industrial and commercial activities on weekdays and the widespread use of electrical appliances such as air conditioners.

[0099] Figure 7 and Figure 8 The load factor of power station A in different seasons is shown. Although the overall load factor trends in the two figures show some similarities, there are obvious differences in the load factor level, change range and specific time points. Figure 8 The initial load rate is Figure 7is low, but the peak values ​​of the two during the day are similar, indicating that although the demand is low in the early stage, Figure 8 The highest demand between China and Japan Figure 6 Similar to, in addition, Figure 8 The change of medium load factor is faster, especially at the beginning of the morning and the end of the evening, which indicates that the power station needs to have faster response capability and more flexible dispatch strategy to adapt to such rapid changes. Figure 7 The load factor in the Figure 8 It reaches the peak value quickly in a shorter time and remains stable, which reflects the seasonal operation and demand patterns of the two different hydropower stations.

[0100] S200, based on load situation analysis, the factors that have the greatest impact on the output of the hydropower station are clustered according to seasonal changes to obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling period for output prediction.

[0101] Preferably, determining the optimal scheduling period for output prediction includes:

[0102] Perform cluster analysis on historical data, find the average of multiple columns of each feature in the historical data, and select the group average as the measure of the distance between groups;

[0103] The clustering solution results are visualized in three-dimensional space, where the position of each point is determined by the factor that has the greatest impact on the output of the hydropower station. Each cluster is represented by a different color, and the clustering pattern of the data points in three-dimensional space is obtained;

[0104] The K-means clustering in three-dimensional space shows the relationship between multiple features of the factors that have the greatest impact on the output of the hydropower station. The data points are grouped according to the similarity of the features to obtain the optimized scheduling period for output prediction.

[0105] Preferably, the cluster analysis solution includes:

[0106] Step 1: randomly select k objects from N sample data as the initial clustering centers;

[0107] Step 2: Calculate the distance from each sample to each cluster center, and then assign the object to the cluster with the closest distance;

[0108] Step 3: After all objects are assigned, recalculate the centers of the k clusters;

[0109] Step 4: Compare the k cluster centers calculated last time. If the cluster center changes, proceed to step 2 again; otherwise, proceed directly to step 5.

[0110] Step 5: When the centroid does not change, stop and output the clustering results.

[0111] It should be noted that the K-means clustering algorithm (K-means) is a typical non-hierarchical clustering algorithm based on distance. On the basis of minimizing the error function, the data is divided into a predetermined number of clusters K, and the silhouette coefficient is used as the evaluation index of K-means, that is, the closer the distance between two objects, the greater the similarity. In K-means, the Euclidean distance is used to measure the similarity between samples, and the distance between the sample and the cluster can be the distance d (e i ,x), the distance between clusters can be expressed by the distance between cluster centers d(e i ,e j ).

[0112] The clustering results may depend on the random selection of the initial cluster centers, causing the results to seriously deviate from the global optimal classification. In practice, in order to obtain better results, different initial cluster centers are usually selected and the K-means algorithm is run multiple times. After all objects are assigned, the centers of the k clusters are recalculated; for continuous data, the cluster center takes the average value of the cluster data; when some attributes of the sample are categorical variables, the mean may not be defined. In this case, the k-mode can be used to determine the cluster center.

[0113] Regarding the measurement of data types and similarity, for continuous attributes, each attribute value must first be 0-mean normalized, and then the distance is calculated, and P attributes are used to represent the data matrix of n samples, expressed as:

[0114]

[0115] The Euclidean distance is expressed as:

[0116]

[0117] The squared error sum (SSE) is used as the objective function to measure the clustering quality. For two different clustering results, the classification result with the smaller squared error sum is selected.

[0118] The objective function under continuous attributes (SSE) is expressed as:

[0119]

[0120] Among them, cluster E i The cluster center e i It is expressed as:

[0121]

[0122] Among them, K is the number of clusters, E i represents the i-th cluster, x1,x2,…,x nrepresents n samples, n i Represents the number of samples in the i-th cluster.

[0123] Specifically, the load of a hydropower station tends to be stable during the flood season, while the load fluctuates greatly during the dry season. In this embodiment, k-means clustering is performed on the output curves of a group of hydropower stations with poor upstream regulation performance during the dry season, taking the output curves of the dry season from January to March of the following year as an example. When clustering, the group average method is selected as the measurement method for the distance between groups.

[0124] The results are visualized in three-dimensional space, where the position of each point is determined by its water level (x-axis), inflow (y-axis), and load rate (z-axis). Each cluster is represented by a different color, so that the clustering pattern of data points in three-dimensional space can be observed. The relationship between the three features of water level, inflow, and load rate is displayed through K-means clustering in three-dimensional space. Five colors represent five different clusters, indicating that the algorithm groups data points according to the similarity of the features. Each cluster may represent a different operating state or environmental condition. For example, some clusters may correspond to a specific water level range or inflow level.

[0125] In order to simplify the dimension of the data, this embodiment uses the average value to facilitate clustering and visualization. Each feature in the original data (water level, inflow flow, load value) has multiple columns. If all these columns are directly used for clustering, the data dimension will be very high, which increases the calculation complexity and makes visualization difficult. Therefore, by averaging multiple columns of each feature, the dimension can be reduced, making clustering and visualization more intuitive.

[0126] The spatial distribution of clusters reveals the centralization trend and dispersion pattern of data points, and points out the changes in data at different time points. Overall, the records are successfully divided into groups with internal similarities based on data from multiple dimensions, providing valuable insights for further analysis of how to manage water levels and inflows. The three-dimensional clustering diagram reveals the complex relationship between water levels, inflows, and load rates, where each cluster may correspond to a different state in system operation. This visualization helps us identify possible patterns and anomalies, and provides insights for water resource management and process optimization. The range of the "water level" axis is around 1130, representing the corresponding water level, the range of the active activity of the "flow" axis is between 0 and 1200, representing the inflow at the corresponding time, and the range of the "load rate" axis is between 0 and 1, representing the output at the corresponding time.

[0127] S300, constructing a load prediction model, predicting the future output of the cascade hydropower station during the optimal dispatching period, and obtaining the output prediction result.

[0128] Preferably, building a load forecasting model includes:

[0129] The differential equation is established for the discrete sequence, and the first-order differential equation model and the discrete and prediction formula are constructed. By the derivative definition and form conversion, the load forecasting model is expressed as:

[0130]

[0131] in, Indicates the value of the sequence generated at time point k+1, x (1) (1) represents the value of the cumulative generated sequence at the initial time point k=1, represents the attenuation factor, that is, the effect that the rate of change of the system state gradually decreases as the reaction time k increases. represents the steady-state value of the model, when the attenuation factor is 1 (i.e., when k = 0), or when k tends to infinity ( tends to 0), which is the long-term equilibrium state or stable value of the system.

[0132] Preferably, constructing a first-order differential equation model and a discretization and prediction formula includes:

[0133] The first-order differential equation model is expressed as:

[0134]

[0135] in, represents the first-order derivative of the dependent variable x with respect to time t, that is, the rate of change of x with time t, x represents the dependent variable, that is, the predicted system state or data sequence, t represents the time variable, indicating the continuity of time, u represents the exogenous variable, that is, the external influencing factor, which can be a control variable or external interference, which has a direct impact on the system state x, a represents the model parameter, which represents the resistance coefficient of x change (that is, the development grayness), and is a parameter that reflects the influence of internal factors of the system on the system behavior;

[0136] The discrete and predictive formulas are expressed as:

[0137]

[0138] Among them, x (1) (k+1) means the value of the generated sequence is accumulated at time point k+1.

[0139] Exemplarily, the first-order differential equation model is defined by the derivative:

[0140]

[0141] Among them, x(t+Δt) represents the value of x after the time difference Δt at time t, Indicates the limit value when the time difference Δt tends to 0;

[0142] When Δt is very small and takes a very small unit of 1, it can be approximately expressed as:

[0143]

[0144] Converted into discrete form:

[0145]

[0146] because Involves accumulating x (1) The values ​​of two moments, therefore, x (1) (i) It is more reasonable to take the average of the two moments before and after, that is, x (i) (i) be replaced by:

[0147]

[0148] Δ (1) (x (1) (k+1))=x (1) (k+1)-x (1) (k) = x (0) (k+1)

[0149]

[0150] Δ (1) (x (1) (k+1))+a(x(k+1))=u

[0151] The last 4 equations above are sorted as follows:

[0152]

[0153] The converted matrix form is:

[0154]

[0155] Among them, x (0) (i) represents the original data sequence at the i-th time point, that is, the initial observation data without any accumulation or transformation. M is used to refer to the multinomial expression omitted in the brackets. For example, M on the left represents the data from x (0) (3) to x (0) (N) All x in the range (0) (i), the same applies to M on the right side.

[0156] Define y and B as:

[0157] y=(x (0) (2),x (0) (3),L,x (0) (N)T

[0158]

[0159] but

[0160] In summary, the load forecasting model is obtained, which is expressed as:

[0161]

[0162] It should be noted that in response to the changes in the load rate of the water station, we analyzed the influencing factors to derive the changing rules, further understand the operation of the hydropower station, establish a hydropower station load rate change prediction model based on historical data, and predict its monthly load rate.

[0163] Based on the GM (1.1) model, the solution is obtained. Compared with the statistical model, the gray model has high accuracy even when a small amount of data is used. Secondly, from the perspective of its mechanism, the closer the gray model is to the current time point, the higher the accuracy will be. Therefore, the prediction function of the gray model is better than that of the statistical model. Gray system modeling is actually a method of finding numbers by numbers, finding the changing relationship of the system from one or several discrete series of the system, and trying to establish a continuous change model of the system.

[0164] Preferably, verifying the accuracy of the load forecasting model includes:

[0165] Assume that the load forecasting model has been obtained and will Do a cumulative conversion to It is expressed as:

[0166]

[0167] in, Indicates the value at the nth time point after the cumulative-subtraction conversion;

[0168] Calculate the residual e(k), expressed as:

[0169]

[0170] Among them, x (0) (k) represents the original data sequence at the kth time point, that is, the initial observation data without any accumulation or transformation;

[0171] The original sequence x (0) And the variance of the residual sequence E is It is expressed as:

[0172]

[0173] in, Represents the original sequence x (0) The average value over n time frames, represents the corresponding residual sequence mean;

[0174] The average values ​​are expressed as:

[0175]

[0176] The calculation of the posterior difference ratio C is expressed as:

[0177]

[0178] It should be noted that after obtaining the prediction formula, the model needs to be tested for accuracy. Only the model that passes the test can be used for prediction. There are generally three methods for model accuracy testing: relative error size test method, correlation test method and posterior difference test method. This embodiment adopts the posterior difference test method. Among them, the index and are two important indicators of the posterior difference test. The smaller the index, the better. The smaller the index, the larger the original data variance, that is, the original data has a large degree of dispersion. The smaller the residual variance, the smaller the residual dispersion. The smaller the residual variance, the smaller the residual dispersion. The smaller the residual dispersion ...

[0179] A grayscale prediction model for the monthly load rate changes of hydropower stations is established. This model is based on the regularity analysis of historical data and can effectively predict the changing trend of future load rates. By sorting and analyzing historical data, we can establish a corresponding prediction model and use this model to predict future changes in monthly load rates. This prediction result will provide an important reference for the operation and scheduling of hydropower stations, and help ensure the stability and reliability of power supply.

[0180] S400, based on historical data, predict the future flow and water level, and associate it with the output prediction result to obtain the operation and dispatch result of the cascade hydropower station.

[0181] Specifically, output forecasts are made for all cascade hydropower stations, and a set of scheduling plans are formed in chronological order. Cluster analysis methods are used to introduce water level and flow data. Output forecast values ​​are used to provide basic data for future scheduling and are verified based on relevant historical data. Adjustments can be made at any time based on meteorological conditions, especially when the weather is changeable during the flood season, so the optimal scheduling plan can be made.

[0182] It should be noted that through the analysis of historical data, the historical situation of the output load of the cascade power station is specifically analyzed. For the change of its load rate, through cluster analysis, the various operating states of the hydropower station can be inferred from the flow and water level, which are related factors affecting the output of the hydropower station. For example, high water level and high flow indicate that the power generation potential has reached its peak, while lower water level and smaller flow may correspond to the maintenance period or the period of low power generation. By comparing the cluster diagrams of the dry season and the flood season, the change pattern of flow and water level can be observed. The flow and water level in the flood season are significantly higher than those in the dry season, which is an important consideration for the operation and management of the hydropower station. Especially in the flood season, the hydropower station needs to deal with possible extreme weather conditions and reasonably dispatch to ensure safety and stable power supply. Finally, with the help of the grayscale prediction method, the prediction curve is generated to make a more reliable load forecast for the future dispatch. The seasonal factors are used to make it closer to the actual situation, which provides a basis for the subsequent reporting of power generation plans and dispatch adjustments. It can make a load forecast for multiple power stations in the basin for a period of time, and propose a new method for the dispatch optimization of cascade power stations.

[0183] The above is a schematic scheme of a cascade hydropower station output prediction method of this embodiment. It should be noted that the technical scheme of the cascade hydropower station output prediction system and the technical scheme of the cascade hydropower station output prediction method described above belong to the same concept, and the details of the technical scheme of the cascade hydropower station output prediction system not described in detail in this embodiment can be referred to the description of the technical scheme of the cascade hydropower station output prediction method described above.

[0184] The output prediction system of cascade hydropower stations in this embodiment includes:

[0185] The acquisition module is used to collect various historical data of cascade hydropower stations, analyze the load conditions of the hydropower stations, and obtain the influence relationship between upstream and downstream output;

[0186] The analysis module is used to analyze the load situation, cluster the factors that have the greatest impact on the output of the hydropower station according to seasonal changes, obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling cycle for output prediction;

[0187] The prediction module is used to build a load prediction model, predict the future output of the cascade hydropower station during the optimal dispatching period, and obtain the output prediction results;

[0188] The scheduling module is used to predict future flow and water levels based on historical data, and associate it with the output prediction results to obtain the operation scheduling results of cascade hydropower stations.

[0189] This embodiment also provides an electronic device, which is applicable to the case of output prediction of cascade hydropower stations, including:

[0190] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method for predicting the output of cascade hydropower stations as proposed in the above embodiment.

[0191] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for predicting the output of cascade hydropower stations proposed in the above embodiment is implemented.

[0192] The storage medium proposed in this embodiment and the method for realizing output prediction of cascade hydropower stations proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0193] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0194] Example 2

[0195] Reference Figure 9-12 , which is an embodiment of the present invention, provides a method for predicting the output of cascade hydropower stations. In order to verify its beneficial effects, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0196] Taking power station A as an example, cluster analysis is performed on historical data. The cluster analysis results are as follows: Fig. 9 , 10 , and gives the prediction results, such as Fig.11 , 12 , which is the load rate value for each day in the next 30 days.

[0197] Depend on Fig. 9 , Fig.10As shown in the figure, the load of the training data (blue line) gradually increased from a load value of nearly 60 to more than 100 during the period from 0 to about 40 days, and then the load value remained basically at around 100, showing a certain periodic fluctuation. The test data (orange line) started from about 60, and the test data showed a sharp decline from 100 to below 80, and then continued to decline after a short period of stability. The forecast (red dotted line) forecast data started from the end of the test data, predicting that the future load will continue to decline, and tried to simulate a similar fluctuation pattern. The forecast curve is relatively smooth.

[0198] In addition, the root mean square error (RMSE) of the prediction model was calculated as an indicator to evaluate the accuracy of the model. According to the value of RMSE, it can be concluded that the prediction model is relatively accurate and makes a good prediction of the load factor change of power station A. The mean square error is about 2.1333.

[0199] The load forecasting model of a hydropower station, based on the regularity analysis of historical data, can effectively predict the changing trend of future load rates, which will provide an important reference for the operation and scheduling of hydropower stations and help ensure the stability and reliability of power supply.

[0200] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for predicting the output of a cascade hydropower station, characterized in that: include: Collect various historical data of cascade hydropower stations, analyze the load conditions of the hydropower stations, and obtain the influence relationship between upstream and downstream output; Based on the load situation analysis, the factors that have the greatest impact on the output of the hydropower station are clustered according to seasonal changes to obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling cycle for output prediction; Construct a load forecasting model to predict the future output of the cascade hydropower station during the optimal dispatching period and obtain the output forecast results; Based on the historical data, the future flow and water level are predicted and associated with the output prediction result to obtain the operation and dispatch result of the cascade hydropower station.

2. The method for predicting the output of cascade hydropower stations according to claim 1, characterized in that: The analysis of the load situation of the hydropower station includes: The historical data includes the time period, water level, flow rate and output load data of the hydropower station, and the data is pre-processed; Calculate the load rate of the hydropower station in different seasons or months, which is the ratio of the actual output of the power system to the rated output within a certain period of time; By drawing a time series graph, the daily load value changes of each hydropower station are displayed, and the daily average load value of each hydropower station is analyzed and compared with the load fluctuation threshold; If the daily average load value is greater than the load fluctuation threshold, the time period or date corresponding to the daily average load value is marked; Compare the load characteristics between different hydropower stations to determine whether there are similar load fluctuation patterns.

3. The method for predicting the output of cascade hydropower stations according to claim 1 or 2, characterized in that: The determination of the optimal scheduling period for output prediction includes: Performing cluster analysis on the historical data to obtain a solution, averaging multiple columns of each feature in the historical data, and selecting a group average as a measurement method for the distance between groups; The clustering solution results are visualized in three-dimensional space, where the position of each point is determined by the factor that has the greatest impact on the output of the hydropower station. Each cluster is represented by a different color, and the clustering pattern of the data points in three-dimensional space is obtained; The K-means clustering in three-dimensional space shows the relationship between multiple features of the factors that have the greatest impact on the output of the hydropower station. The data points are grouped according to the similarity of the features to obtain the optimized scheduling period for output prediction.

4. The method for predicting the output of cascade hydropower stations according to claim 3, characterized in that: The cluster analysis solution includes: Step 1: randomly select k objects from N sample data as the initial clustering centers; Step 2: Calculate the distance from each sample to each cluster center, and then assign the object to the cluster with the closest distance; Step 3: After all objects are assigned, recalculate the centers of the k clusters; Step 4: Compare the k cluster centers calculated last time. If the cluster center changes, proceed to step 2 again; otherwise, proceed directly to step 5. Step 5: When the centroid does not change, stop and output the clustering results.

5. The method for predicting the output of cascade hydropower stations according to claim 3, characterized in that: The construction of the load forecasting model comprises: A differential equation is established for the discrete sequence, a first-order differential equation model and a discrete and prediction formula are constructed, and the load forecasting model is obtained by derivative definition and form conversion as follows: in, Indicates the value of the sequence generated at time point k+1, x (1) (1) represents the value of the cumulative generated sequence at the initial time point k=1, represents the attenuation factor, Represents the steady-state value of the model.

6. The method for predicting the output of cascade hydropower stations according to claim 5, characterized in that: Constructing the first-order differential equation model and the discretization and prediction formulas include: The first-order differential equation model is expressed as: in, represents the first-order derivative of the dependent variable x with respect to time t, that is, the rate of change of x with time t, x represents the dependent variable, that is, the predicted system state or data sequence, t represents the time variable, u represents the exogenous variable, that is, the external influencing factor, and a represents the model parameter, that is, the resistance coefficient of the change of x; The discrete and predictive formulas are expressed as: Among them, x (1) (k+1) means the value of the generated sequence is accumulated at time point k+1.

7. The method for predicting the output of cascade hydropower stations according to claim 5 or 6, characterized in that: Verifying the accuracy of the load forecasting model includes: Assume that the load forecasting model has been obtained and will Do a cumulative conversion to It is expressed as: in, Indicates the value at the nth time point after the cumulative-subtraction conversion; Calculate the residual e(k), expressed as: Among them, x (0) (k) represents the original data sequence at the kth time point; The original sequence x (0) And the variance of the residual sequence E is It is expressed as: in, Represents the original sequence x (0) The average value over n time frames, represents the corresponding residual sequence mean; The average values ​​are expressed as: The calculation of the posterior difference ratio C is expressed as:

8. A cascade hydropower station output prediction system, characterized in that: Applying the method for predicting the output of cascade hydropower stations as claimed in any one of claims 1 to 7, include, The acquisition module is used to collect various historical data of cascade hydropower stations, analyze the load conditions of the hydropower stations, and obtain the influence relationship between upstream and downstream output; An analysis module is used to analyze the load situation, perform cluster analysis on the factors that have the greatest impact on the output of the hydropower station based on seasonal changes, obtain the relationship between the factors that have the greatest impact on the output and the output, and determine the optimal scheduling period for output prediction; The prediction module is used to build a load prediction model, predict the future output of the cascade hydropower station during the optimal dispatching period, and obtain the output prediction results; The scheduling module is used to predict the future flow and water level based on the historical data, and associate it with the output prediction result to obtain the operation scheduling result of the cascade hydropower station.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method for predicting the output of cascade hydropower stations described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: It stores computer executable instructions, which, when executed by a processor, can implement the steps of the method for predicting the output of cascade hydropower stations as described in any one of claims 1 to 7.

Images