Power station water flow prediction method, device and equipment and storage medium

Abstract

The invention discloses a power station water flow prediction method, device and equipment and a storage medium, and the method comprises the steps: obtaining geographic data and historical data of a plurality of hydropower stations, the historical data comprising reservoir hydrological data, power generation data and meteorological data; extracting a typical water flow curve of each hydropower station according to the historical data; determining a classification label corresponding to each hydropower station based on the geographic data and the historical data; generating a user portrait of each hydropower station according to the typical water flow curve, the classification label and the power generation data; based on the user portrait and historical data, constructing a water flow prediction model based on deep learning; and based on the user portrait of the to-be-predicted hydropower station and the current data, performing prediction by using the water flow prediction model to obtain a water flow prediction result of the to-be-predicted hydropower station. The accuracy of power station water flow prediction can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, equipment and storage medium for predicting water flow in a power station. Background Art

[0002] Water flow is crucial to the operation of hydropower plants. Its changes directly determine the water storage capacity of reservoirs, the operating efficiency of generators, and the stability of power output. Under different hydrological conditions, changes in water flow will affect the operation of power plants, especially during the dry season, when reduced water flow may lead to reduced power generation capacity or even shutdown; while during the flood season, excessive water flow may cause flooding or safety hazards, affecting the normal operation of power plants. Traditional water flow forecasting methods rely on historical data and simple models, and are often unable to cope with complex changes in climate and geographical factors. Especially when dealing with extreme weather events and uncertain environmental conditions, the forecast accuracy needs to be improved. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a method, device, equipment and storage medium for predicting water flow in a power station, which can improve the accuracy and real-time performance of water flow prediction in a power station.

[0004] An embodiment of the present invention provides a method for predicting water flow in a power station, comprising: Obtaining geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data, and meteorological data; Extracting a typical water flow curve of each hydropower station based on the historical data; Based on the geographic data and the historical data, determining a classification label corresponding to each hydropower station; wherein a hydropower station has several classification labels; Generate a user profile of each hydropower station according to the typical water flow curve, the classification label and the power generation data; Based on the user portrait and the historical data, construct a water flow prediction model based on deep learning; The current data of the hydropower station to be predicted is obtained, and based on the user portrait and current data of the hydropower station to be predicted, the water flow prediction model is used to perform prediction to obtain the water flow prediction result of the hydropower station to be predicted.

[0005] As an improvement of the above solution, the water flow prediction model is obtained by combining a convolutional neural network with a long short-term memory network.

[0006] As an improvement of the above solution, the reservoir hydrological data includes: inflow, water volume, reservoir capacity, and reservoir water level; The power generation data includes: power generation flow, output power, and water head; The meteorological data include: actual precipitation in the basin, actual precipitation in the reservoir, temperature, and humidity.

[0007] As an improvement of the above solution, determining the classification label corresponding to each hydropower station based on the geographic data and the historical data includes: Based on the geographic data and the historical data, the hydropower stations are classified according to water flow influencing factors to obtain an initial classification label for each hydropower station; the water flow influencing factors include water flow, season, precipitation, climate type and geographical location; Calculate the number of hydroelectric power plants under each of the initial classification labels; For the initial classification labels whose number of hydropower stations under the initial classification labels is less than a preset first threshold, similar field labels are aggregated to determine the final classification labels and their corresponding hydropower stations; For the initial classification labels whose number of hydropower stations is greater than a preset second threshold, the typical water flow curves of the hydropower stations under the initial classification labels are clustered to obtain clustering results, and the final classification labels and their corresponding hydropower stations are determined based on the clustering results.

[0008] As an improvement of the above solution, determining the final classification label according to the clustering result includes: Calculating the correlation coefficient between the water flow and the meteorological data; Calculating the first Manhattan distance of the typical water flow curves between different seasons; Calculating the second Manhattan distance of the typical water flow curves between different hydropower stations; The clustering result is optimized according to one or more data of the correlation coefficient, the first Manhattan distance, and the second Manhattan distance to obtain a final classification label.

[0009] As an improvement of the above solution, the water flow prediction model based on deep learning is constructed based on the user portrait and the historical data, including: Splicing the historical data with the data in the user portrait to obtain spliced ​​data; The spliced ​​data is input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0010] As an improvement of the above solution, the water flow prediction model based on deep learning is constructed based on the user portrait and the historical data, including: Based on the user portrait, key features that affect water flow prediction are screened from the historical data; wherein the key features are screened by calculating the Gini coefficient; The key features are input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0011] The embodiment of the present invention further provides a water flow prediction device for a power station, comprising: A data acquisition module, used to acquire geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data and meteorological data; A water flow curve extraction module, used for extracting a typical water flow curve of each hydropower station according to the historical data; A classification label determination module, used to determine the classification label corresponding to each hydropower station based on the geographic data and the historical data; wherein a hydropower station has several classification labels; A user portrait generation module, used to generate a user portrait of each hydropower station according to the typical water flow curve, the classification label and the power generation data; A prediction model building module, used to build a water flow prediction model based on deep learning based on the user portrait and the historical data; The water flow prediction module is used to obtain the current data of the hydropower station to be predicted, and based on the user portrait and current data of the hydropower station to be predicted, use the water flow prediction model to perform prediction to obtain the water flow prediction result of the hydropower station to be predicted.

[0012] An embodiment of the present invention also provides a computer device, including a processor and a memory, wherein a computer program is stored in the memory, and the computer program is configured to be executed by the processor, and when the processor executes the computer program, the water flow prediction method for a power plant described in any one of the above items is implemented.

[0013] An embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any of the above-mentioned methods for predicting water flow in a power plant.

[0014] Compared with the prior art, the beneficial effects of a method, device, equipment and storage medium for predicting water flow in a power station provided by an embodiment of the present invention are: by obtaining geographic data and historical data of several hydropower stations, the historical data includes reservoir hydrological data, power generation data and meteorological data; extracting the typical water flow curve of each hydropower station based on the historical data; determining the classification label corresponding to each hydropower station based on the geographic data and historical data; generating a user portrait of each hydropower station based on the typical water flow curve, classification label and power generation data; constructing a water flow prediction model based on deep learning based on the user portrait and historical data, and then using the water flow prediction model to predict water flow. The embodiment of the present invention can automatically learn the complex laws of flow changes, provide accurate water flow prediction, have strong real-time performance, and can still maintain good accuracy for data under extreme climates. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of a method for predicting water flow in a power station provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a water flow prediction device for a power station provided by an embodiment of the present invention; Figure 3 It is a structural schematic diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] See also Figure 1 , Figure 1 The present invention provides a flow chart of a method for predicting water flow in a power station. The method for predicting water flow in a power station comprises: S1: Obtaining geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data and meteorological data.

[0018] As one of the optional embodiments, the reservoir hydrological data includes: inflow, water volume, reservoir capacity, and reservoir water level; The power generation data includes: power generation flow, output power, and water head; The meteorological data include: actual precipitation in the basin, actual precipitation in the reservoir, temperature, and humidity.

[0019] Among them, the inflow refers to the flow of water into the reservoir, which is mainly affected by factors such as upstream precipitation and basin flow; the water volume refers to the amount of water in the reservoir, measured in cubic meters, which indicates the overall situation of the reservoir's water storage; the reservoir capacity refers to the amount of water that the reservoir can store at a specific water level, which also reflects the reservoir's water storage capacity; the reservoir capacity water level refers to the current water level of the reservoir, and the water level directly affects the output of the hydropower station. The power generation flow refers to the flow used for power generation at the hydropower station; the output power refers to the actual power generation of the hydropower station; the head refers to the height difference between the reservoir water level and the hydropower generator unit. The higher the head, the greater the output of the hydropower station. The actual precipitation in the basin refers to the precipitation within the basin; the actual precipitation in the reservoir refers to the precipitation within the reservoir.

[0020] Specifically, historical data is obtained from multiple data sources such as hydrological monitoring stations, meteorological stations, and hydropower stations. The data include factors such as flow, precipitation, temperature, and water level in different time periods. For example, three years of historical data are obtained at time intervals of 15 minutes. These data are all time series data.

[0021] Furthermore, historical data are processed for denoising, missing value filling and outlier correction to ensure data quality.

[0022] S2: Extracting a typical water flow curve of each hydropower station based on the historical data.

[0023] Specifically, the water flow curve describes the change of water flow over time. It is a representative curve extracted by analyzing the historical data of the hydropower station for many years or the water flow changes in multiple periods to reflect the general change law of water flow under different seasons and climate conditions.

[0024] S3: Based on the geographic data and the historical data, determine the classification label corresponding to each hydropower station; wherein one hydropower station has several classification labels.

[0025] As one of the optional embodiments, determining the classification label corresponding to each hydropower station based on the geographic data and the historical data includes: Based on the geographic data and the historical data, the hydropower stations are classified according to water flow influencing factors to obtain an initial classification label for each hydropower station; the water flow influencing factors include water flow, season, precipitation, climate type and geographical location; Calculate the number of hydroelectric power plants under each of the initial classification labels; For the initial classification labels whose number of hydropower stations under the initial classification labels is less than a preset first threshold, similar field labels are aggregated to determine the final classification labels and their corresponding hydropower stations; For the initial classification labels whose number of hydropower stations is greater than a preset second threshold, the typical water flow curves of the hydropower stations under the initial classification labels are clustered to obtain clustering results, and the final classification labels and their corresponding hydropower stations are determined based on the clustering results.

[0026] As one of the optional embodiments, determining the final classification label according to the clustering result includes: Calculating the correlation coefficient between the water flow and the meteorological data; Calculating the first Manhattan distance of the typical water flow curves between different seasons; Calculating the second Manhattan distance of the typical water flow curves between different hydropower stations; The clustering result is optimized according to one or more data of the correlation coefficient, the first Manhattan distance, and the second Manhattan distance to obtain a final classification label.

[0027] Specifically, based on factors such as water flow, season, and precipitation, hydropower stations are classified to obtain initial classification labels for hydropower stations, which can reflect the operating characteristics and climate sensitivity of the power station under different conditions. Since each hydropower station may be affected by multiple factors at the same time, each hydropower station will have multiple classification labels, such as upstream basin labels, seasonal water flow labels, rainy season labels, etc. After obtaining the initial classification labels, use the labels to annotate relevant historical data (such as water flow data, meteorological data, water level data, geographic data, etc.).

[0028] Furthermore, the distribution of the initial classification labels is analyzed, and the number of hydropower stations corresponding to each classification label is used to determine whether the labels need to be subdivided or aggregated. If the number of hydropower stations under a classification label (such as seasonal water flow) is too small, the number of samples under such labels is increased by aggregating labels in similar fields to reduce data sparsity. If the number of hydropower stations under a certain classification label (such as a high flow period) is too large, the label is further subdivided to form a smaller classification label.

[0029] Specifically, more detailed classification labels are obtained by clustering typical water flow curves. Through cluster analysis, hydropower stations with similar water flow patterns are classified into the same category. Each group of data after clustering represents a typical water flow pattern, and the hydropower stations in each category have high similarity in water flow patterns. As a result, the subsequent model can learn a more consistent feature set during training, thereby reducing the impact of noise, improving the stability and accuracy of predictions, and reducing data sparsity, avoiding overfitting or insufficient prediction accuracy due to too few samples.

[0030] In another specific embodiment, cluster analysis may be performed on the load curve of the hydropower station to find areas with similar load characteristics and generate corresponding clustering results.

[0031] Furthermore, the above clustering results are classified based on the similarity of water flow patterns. However, hydropower stations will also be affected by other factors, such as temperature, precipitation, etc. Therefore, by analyzing the water flow characteristics and influencing factors, the clustering results are further classified and optimized to obtain the final classification labels, so that the water flow patterns in each category are more consistent.

[0032] Specifically, analyze the impact of meteorological data (such as temperature, rainfall, humidity, etc.), water level, water supply, population mobility and other factors on water flow, establish a multivariate linear regression model or other correlation analysis model, and quantify the influencing factors. The specific formula is as follows: Pearson Correlation Coefficient: Used to quantify the correlation between water flow and meteorological data (such as temperature, precipitation, etc.).

[0033] in, and Respectively represent variables (such as precipitation) and (such as water flow), and is their mean, is the Pearson correlation coefficient between the two, indicating the strength of the linear relationship between them.

[0034] Manhattan Distance: Used to calculate the differences in water flow curves between different seasons or different hydropower stations.

[0035] in, and For two water flow sequences in The value at a time point, is the total number of time points, is the Manhattan distance between two sequences, indicating their difference.

[0036] The classification labels of the embodiment of the present invention are obtained by classifying the hydropower stations according to factors such as climate type and seasonal changes, so that the model can identify and capture the changing patterns of water flow in different regions and seasons. For example, some hydropower stations may be located in rainy areas or areas with strong seasonal water flow, while others may be located in arid areas or areas with small seasonal flow fluctuations. Through the subdivision of the classification labels, the model can make targeted predictions based on the specific climate characteristics of each power station, thereby providing more accurate flow predictions under extreme weather conditions (such as heavy rains, droughts, etc.).

[0037] S4: Generate a user profile of each hydropower station based on the typical water flow curve, the classification label and the power generation data.

[0038] Specifically, user portraits help build a personalized portrait of each hydropower station by analyzing the historical data of the hydropower station, such as flow changes, meteorological conditions, and power generation, thereby reflecting its sensitivity to meteorological changes. For example, some hydropower stations may be very sensitive to changes in precipitation, while other hydropower stations may be more affected by temperature or water level fluctuations. User portraits provide comprehensive background information for each hydropower station (such as water flow curves, classification labels, historical power generation, etc.), reflecting the water demand, power generation capacity, seasonal changes, etc. of each hydropower station. The information provided by user portraits can help the model understand the characteristics and operation mode of the hydropower station. By establishing user portraits, the prediction model can understand and adapt to the response patterns of different hydropower stations under extreme weather, ensuring that timely and accurate prediction results can be provided when extreme climate events occur.

[0039] S5: Based on the user portrait and the historical data, a water flow prediction model based on deep learning is constructed.

[0040] As one of the optional embodiments, the constructing of a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Splicing the historical data with the data in the user portrait to obtain spliced ​​data; The spliced ​​data is input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0041] Among them, using user portraits as input data of the model can enable the model to extract the most relevant features based on the user portraits, and then build a more accurate water flow prediction model.

[0042] As another optional embodiment, the constructing a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Based on the user portrait, key features that affect water flow prediction are screened from the historical data; wherein the key features are screened by calculating the Gini coefficient; The key features are input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0043] Specifically, based on user portraits, the importance of characteristic data such as water flow, meteorological factors, and seasons is analyzed to screen out key features that affect water flow prediction. The specific formula is as follows: in, For Class In the dataset The probability of is the number of categories, Gini coefficient is the Gini coefficient of the data set, which indicates the purity of the data set. The smaller the Gini coefficient, the stronger the distinguishing ability of the feature. Specifically, for each feature data (such as precipitation, temperature, etc.), the interval is divided according to its value, and the corresponding Gini coefficient is calculated for each interval. Then, the Gini coefficient of each interval is weighted averaged to obtain the Gini coefficient of the feature data. If the Gini coefficient of a feature is small, it means that the feature has a high purity and contributes more to the prediction result. Features with a small Gini coefficient are preferred as key features.

[0044] As one of the optional embodiments, the water flow prediction model is obtained by combining a convolutional neural network and a long short-term memory network.

[0045] Specifically, a CNN-LSTM model is established. The CNN-LSTM combined model combines the advantages of convolutional neural networks (CNN) and long short-term memory networks (LSTM), and can effectively process spatial features and temporal dependencies in time series data. In water flow prediction, CNN extracts local spatiotemporal features, while LSTM captures long-term temporal dependencies. The combination of the two can more accurately predict water flow changes.

[0046] CNN can extract spatial local features from input time series data, and especially when processing complex multidimensional data, it can discover hidden patterns in the data. And CNN can extract spatial local features from input time series data, and especially when processing complex multidimensional data, it can discover hidden patterns in the data.

[0047] CNN consists of convolutional layers and pooling layers.

[0048] The convolution layer is the core of the CNN part, which filters the input data through the convolution kernel to extract local features.

[0049] in, is the convolution output image , Locations, is the convolution output image , Locations, is the weight of the convolution kernel, is the bias term, and is the size of the convolution kernel.

[0050] The purpose of the pooling layer is to reduce the dimension of the feature map, usually using maximum pooling.

[0051] in, is the pooling window size. Max pooling selects the maximum value of the window as output.

[0052] Long short-term memory network (LSTM) can effectively capture the long-term time series dependency in the data, which is particularly important for the prediction of water flow and its impact on power generation. Because water flow is not only affected by factors such as weather and seasons, the impact of these factors on water flow usually has a long time duration, and the change of water flow is directly related to the power generation of the hydropower station. Using LSTM to process these complex time series data can better capture the long-term dependency of flow changes on water flow and power generation, and improve the accuracy of prediction. The structure of LSTM is as follows: Forget gate: The forget gate determines how much information in the memory unit output at the previous moment needs to be forgotten.

[0053] Input gate: The input gate determines how much of the current input information needs to be stored in the memory cell.

[0054] Candidate value: The candidate memory cell calculates the current input information and adjusts the state of the memory cell.

[0055] Cell state update: The state of the memory cell is updated at every moment, integrating the effects of the forget gate, input gate, and candidate memory cells.

[0056] Output gate: The output gate determines the hidden state at the current moment , which is the final output of LSTM.

[0057] in, is the output of the forget gate, is the input gate output, is the candidate cell state, is the current cell state, is the cell state at the previous moment, is the input at the current moment, is the output at the current moment, is the output at the previous moment, , , are all bias terms, , , are weight matrices, is the sigmoid function, and tanh is the hyperbolic tangent function.

[0058] It should be noted that the water flow prediction model can output water flow prediction data, or can output water flow prediction data and power generation prediction data at the same time. The specific output data type of the model can be selected according to specific needs.

[0059] Furthermore, after the model is trained, the accuracy of the water flow prediction model is evaluated by calculating evaluation indicators such as mean square error (MSE) and root mean square error (RMSE); if the prediction accuracy exceeds 85%, it is considered that the water flow prediction model meets the requirements and can be used for water flow prediction.

[0060] The prediction accuracy is calculated based on the deviation between the prediction result and the actual value. For prediction results with an accuracy lower than 85%, the cause is analyzed, which may be data anomalies, model inapplicability, or label problems, and the low-precision model is optimized according to the cause until the model meets the accuracy requirements.

[0061] S6: Acquire current data of the hydropower station to be predicted, and use the water flow prediction model to perform prediction based on the user portrait and current data of the hydropower station to be predicted to obtain a water flow prediction result of the hydropower station to be predicted.

[0062] Specifically, the current data of the hydropower station to be predicted, including reservoir hydrological data, power generation data and meteorological data, is obtained, and the data is input into a trained water flow prediction model to obtain the water flow prediction results output by the model, thereby achieving accurate and real-time prediction of water flow.

[0063] The embodiments of the present invention are intended to overcome the problems of insufficient prediction accuracy and slow response to extreme weather events in the prior art. The embodiments of the present invention integrate multi-source data such as meteorology, geography, and historical flow, and use deep learning algorithms to automatically learn the complex laws of flow changes, thereby improving the robustness and generalization ability of the model, and can provide more accurate flow predictions for hydropower plants, optimize scheduling decisions, and improve power generation efficiency and safety. Some data sets contain data under extreme climates, and under such data, the embodiments of the present invention can still maintain a good accuracy rate.

[0064] Correspondingly, the present invention also provides a water flow prediction device for a power station, which can implement all the processes of the water flow prediction method for a power station in the above embodiment.

[0065] See also Figure 2 , Figure 2 1 is a schematic diagram of a water flow prediction device for a power station provided by an embodiment of the present invention. The water flow prediction device for a power station comprises: The data acquisition module 201 is used to acquire geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data and meteorological data; A water flow curve extraction module 202, for extracting a typical water flow curve of each hydropower station according to the historical data; A classification label determination module 203 is used to determine the classification label corresponding to each hydropower station based on the geographic data and the historical data; wherein a hydropower station has several classification labels; A user portrait generation module 204, configured to generate a user portrait of each hydropower station according to the typical water flow curve, the classification label and the power generation data; A prediction model building module 205, for building a water flow prediction model based on deep learning based on the user portrait and the historical data; The water flow prediction module 206 is used to obtain the current data of the hydropower station to be predicted, and based on the user portrait and current data of the hydropower station to be predicted, use the water flow prediction model to perform prediction to obtain the water flow prediction result of the hydropower station to be predicted.

[0066] Preferably, the water flow prediction model is obtained by combining a convolutional neural network and a long short-term memory network.

[0067] Preferably, the reservoir hydrological data includes: inflow, water volume, reservoir capacity, and reservoir water level; The power generation data includes: power generation flow, output power, and water head; The meteorological data include: actual precipitation in the basin, actual precipitation in the reservoir, temperature, and humidity.

[0068] Preferably, determining the classification label corresponding to each hydropower station based on the geographic data and the historical data includes: Based on the geographic data and the historical data, the hydropower stations are classified according to water flow influencing factors to obtain an initial classification label for each hydropower station; the water flow influencing factors include water flow, season, precipitation, climate type and geographical location; Calculate the number of hydroelectric power plants under each of the initial classification labels; For the initial classification labels whose number of hydropower stations under the initial classification labels is less than a preset first threshold, similar field labels are aggregated to determine the final classification labels and their corresponding hydropower stations; For the initial classification labels whose number of hydropower stations is greater than a preset second threshold, the typical water flow curves of the hydropower stations under the initial classification labels are clustered to obtain clustering results, and the final classification labels and their corresponding hydropower stations are determined based on the clustering results.

[0069] Preferably, determining the final classification label according to the clustering result includes: Calculating the correlation coefficient between the water flow and the meteorological data; Calculating the first Manhattan distance of the typical water flow curves between different seasons; Calculating the second Manhattan distance of the typical water flow curves between different hydropower stations; The clustering result is optimized according to one or more data of the correlation coefficient, the first Manhattan distance, and the second Manhattan distance to obtain a final classification label.

[0070] Preferably, the constructing of a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Splicing the historical data with the data in the user portrait to obtain spliced ​​data; The spliced ​​data is input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0071] Preferably, the constructing of a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Based on the user portrait, key features that affect water flow prediction are screened from the historical data; wherein the key features are screened by calculating the Gini coefficient; The key features are input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

[0072] In specific implementation, the working principle, control process and technical effect achieved by the water flow prediction device for a power station provided in an embodiment of the present invention are the same as those of the water flow prediction method for a power station in the above embodiment, and will not be repeated here.

[0073] See also Figure 3 , Figure 3 3 is a block diagram of a computer device provided in an embodiment of the present invention, wherein the computer device comprises: a processor 301, a memory 302, and a computer program stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program, the steps in the above-mentioned power station water flow prediction method embodiment are implemented. Alternatively, when the processor 301 executes the computer program, the functions of each module / unit in the above-mentioned device embodiments are implemented.

[0074] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 302 and executed by the processor 301 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the computer device.

[0075] The computer device may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that the schematic diagram is merely an example of a computer device and does not constitute a limitation on the computer device, and may include more or fewer components than shown in the diagram, or may combine certain components, or different components, for example, the computer device may also include an input / output device, a network access device, a bus, etc.

[0076] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 301 is the control center of the computer device, and uses various interfaces and lines to connect various parts of the entire computer device.

[0077] The memory 302 can be used to store the computer program and / or module. The processor 301 implements various functions of the computer device by running or executing the computer program and / or module stored in the memory 302 and calling the data stored in the memory 302. The memory 302 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory 302 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0078] Wherein, if the module / unit integrated in the computer device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor 301. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0079] An embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the water flow prediction method for a power plant described in any of the above embodiments.

[0080] The embodiment of the present invention provides a method, device, equipment and storage medium for predicting water flow of a power station, and its beneficial effects are: by obtaining geographic data and historical data of several hydropower stations, the historical data includes reservoir hydrological data, power generation data and meteorological data; extracting the typical water flow curve of each hydropower station according to the historical data; determining the classification label corresponding to each hydropower station based on the geographic data and historical data; generating a user portrait of each hydropower station according to the typical water flow curve, classification label and power generation data; constructing a water flow prediction model based on deep learning based on the user portrait and historical data, and then using the water flow prediction model to predict water flow. The embodiment of the present invention can automatically learn the complex laws of flow changes, provide accurate water flow prediction, strong real-time performance, and can still maintain good accuracy for data under extreme climates.

[0081] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for predicting water flow in a power station, characterized in that: include: Obtaining geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data, and meteorological data; Extracting a typical water flow curve of each hydropower station based on the historical data; Based on the geographic data and the historical data, determining a classification label corresponding to each hydropower station; wherein a hydropower station has several classification labels; Generate a user profile of each hydropower station according to the typical water flow curve, the classification label and the power generation data; Based on the user portrait and the historical data, construct a water flow prediction model based on deep learning; The current data of the hydropower station to be predicted is obtained, and based on the user portrait and current data of the hydropower station to be predicted, the water flow prediction model is used to perform prediction to obtain the water flow prediction result of the hydropower station to be predicted.

2. The method for predicting water flow in a power plant according to claim 1, characterized in that: The water flow prediction model is obtained by combining a convolutional neural network and a long short-term memory network.

3. The method for predicting water flow in a power plant according to claim 1, characterized in that: The reservoir hydrological data include: inflow, water volume, reservoir capacity, and reservoir water level; The power generation data includes: power generation flow, output power, and water head; The meteorological data include: actual precipitation in the basin, actual precipitation in the reservoir, temperature, and humidity.

4. The water flow prediction method for a power station according to claim 1, characterized in that: The determining, based on the geographic data and the historical data, a classification label corresponding to each hydropower station comprises: Based on the geographic data and the historical data, the hydropower stations are classified according to water flow influencing factors to obtain an initial classification label for each hydropower station; the water flow influencing factors include water flow, season, precipitation, climate type and geographical location; Calculate the number of hydroelectric power plants under each of the initial classification labels; For the initial classification labels whose number of hydropower stations under the initial classification labels is less than a preset first threshold, similar field labels are aggregated to determine the final classification labels and their corresponding hydropower stations; For the initial classification labels whose number of hydropower stations is greater than a preset second threshold, the typical water flow curves of the hydropower stations under the initial classification labels are clustered to obtain clustering results, and the final classification labels and their corresponding hydropower stations are determined based on the clustering results.

5. The method for predicting water flow in a power station according to claim 4, characterized in that: Determining a final classification label according to the clustering result includes: Calculating the correlation coefficient between the water flow and the meteorological data; Calculating the first Manhattan distance of the typical water flow curves between different seasons; Calculating the second Manhattan distance of the typical water flow curves between different hydropower stations; The clustering result is optimized according to one or more data of the correlation coefficient, the first Manhattan distance, and the second Manhattan distance to obtain a final classification label.

6. The method for predicting water flow in a power station according to claim 1, characterized in that: The constructing of a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Splicing the historical data with the data in the user portrait to obtain spliced ​​data; The spliced ​​data is input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

7. The method for predicting water flow in a power plant according to claim 1, characterized in that: The constructing of a water flow prediction model based on deep learning based on the user portrait and the historical data includes: Based on the user portrait, key features that affect water flow prediction are screened from the historical data; wherein the key features are screened by calculating the Gini coefficient; The key features are input into a water flow prediction model based on deep learning for model training to obtain a trained water flow prediction model.

8. A water flow prediction device for a power station, characterized in that: include: A data acquisition module, used to acquire geographic data and historical data of several hydropower stations, wherein the historical data includes reservoir hydrological data, power generation data and meteorological data; A water flow curve extraction module, used for extracting a typical water flow curve of each hydropower station according to the historical data; A classification label determination module, used to determine the classification label corresponding to each hydropower station based on the geographic data and the historical data; wherein a hydropower station has several classification labels; A user portrait generation module, used to generate a user portrait of each hydropower station according to the typical water flow curve, the classification label and the power generation data; A prediction model building module, used to build a water flow prediction model based on deep learning based on the user portrait and the historical data; The water flow prediction module is used to obtain the current data of the hydropower station to be predicted, and based on the user portrait and current data of the hydropower station to be predicted, use the water flow prediction model to perform prediction to obtain the water flow prediction result of the hydropower station to be predicted.

9. A computer device, characterized in that: It comprises a processor and a memory, wherein a computer program is stored in the memory and the computer program is configured to be executed by the processor, and when the processor executes the computer program, the method for predicting water flow in a power plant as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the method for predicting water flow in a power plant as described in any one of claims 1 to 7 is implemented.

Images