Water level prediction method and equipment for water conservancy project

Abstract

The invention discloses a water level prediction method and equipment for a water conservancy project. The method comprises the following steps: reading water level time sequence data of a reservoir by taking preset time as an interval; external influence data of the reservoir level are obtained, feature extraction is performed, external influence features are obtained and stored in the block chain, correlation analysis is performed on the external influence features and the reservoir level, and weights are distributed to the external influence features according to analysis results; inputting the water level time sequence data and the external influence features carrying the weights into a pre-trained water level prediction model to obtain an initial predicted water level; according to the water quality information and geological information of the reservoir entering river, first internal influence data of the reservoir level is calculated through the reservoir entering flow of the reservoir; measuring underground water level data and underground geological data of the reservoir to determine second internal influence data of the reservoir level; and based on the first internal influence data and the second internal influence data, correcting the initial predicted water level through a water level correction model to obtain a final predicted water level.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a water level prediction method and device for water conservancy projects. Background Art

[0002] Water conservancy projects have positive significance for the rational use of water resources, ensuring people's water supply for daily life, industry and agriculture, and reducing the impact of natural disasters such as floods and droughts on people's production and life.

[0003] As an important part of water conservancy projects, reservoirs have multiple functions such as water storage, water flow regulation, power generation and flood control. Accurately predicting reservoir water level fluctuations is of great significance to the scientific and rational use of water resources and reducing the impact of natural disasters such as floods and droughts.

[0004] However, most of the existing reservoir water level prediction schemes are implemented through neural network models, which input the historical water level data and influencing factors of the reservoir into the pre-trained model to output the predicted water level results. However, in this prediction process, on the one hand, in order to improve the prediction accuracy, most of them are scheme improvements at the model level, selecting models with better prediction performance or using combined models, while the selection of model input influencing factors is similar. Most of them only consider single factors such as rainfall, flood discharge, and evaporation. Other factors that may affect the reservoir water level, such as the reservoir groundwater level, the geographical environment factors around the reservoir, or the geographical environment factors around the river flowing into the reservoir, are not considered enough or are not directly taken into consideration. Although these other factors do not have a high degree of influence on the reservoir water level as factors such as rainfall and flood discharge, the reservoir water level will also change to a certain extent under the influence of these other factors. Therefore, incorporating these factors into the consideration range of influencing factors is of certain significance to improving the accuracy of water level prediction. On the other hand, most of the existing solutions for predicting water levels through models extract features of influencing factors within the model, and use the performance of the model to extract features that affect the water level. When the model performance is certain, the feature extraction process is not targeted. At the same time, after the feature extraction is completed, the model generally directly uses the extracted influencing features to predict the water level, which only considers that these features will affect the water level, but does not consider the degree of influence of each feature on the water level, which is bound to affect the accuracy of the prediction model. Summary of the invention

[0005] The present application provides a water level prediction method and device for water conservancy projects, which are used to solve at least one of the above-mentioned technical problems.

[0006] This application adopts the following technical solutions:

[0007] A water level prediction method for a water conservancy project, the water conservancy project at least comprising a reservoir, the method comprising: reading water level time series data of the reservoir at preset time intervals; obtaining external impact data of the reservoir water level, and performing feature extraction on the external impact data to obtain external impact features and store them in a blockchain, then performing correlation analysis on the external impact features and the reservoir water level, and assigning weights to the external impact features according to the analysis results; inputting the water level time series data and the external impact features carrying weights into a pre-trained water level prediction model to obtain an initial predicted water level; calculating first internal impact data of the reservoir water level through the inflow of the reservoir according to water quality information and geological information of the river entering the reservoir; measuring groundwater level data and underground geological data of the reservoir, and determining second internal impact data of the reservoir water level through the measurement results; based on the first internal impact data and the second internal impact data, correcting the initial predicted water level through a water level correction model to obtain a final predicted water level.

[0008] In a possible implementation of the present application, the method also includes: reading the geographic location data of the reservoir, and obtaining the meteorological data of the geographic location data between the current time node and the predicted time node; extracting rainfall data and temperature and humidity data from the meteorological data to perform feature extraction on the rainfall data to obtain the rainfall characteristics corresponding to the reservoir at the predicted time node, and performing feature extraction on the temperature and humidity data, and combining the measured water temperature data to obtain the evaporation characteristics corresponding to the reservoir at the predicted time node; determining the river information associated with the reservoir, and determining the inflow characteristics corresponding to the reservoir based on the river information and the meteorological information corresponding to the river information between the current time node and the predicted time node; obtaining the water scheduling plan of the reservoir between the current time node and the predicted time node, to extract the outflow characteristics corresponding to the reservoir according to the water scheduling plan; and constructing the external influencing characteristics of the reservoir water level based on the rainfall characteristics, the evaporation characteristics, the inflow characteristics and the outflow characteristics.

[0009] In a possible implementation of the present application, the process of extracting / determining the external influence features includes: when the geographic location data is administrative division data, obtaining a dynamic rainfall cloud map of the administrative division data between the current time node and the predicted time node, and dynamically adding the temperature and humidity data to the dynamic rainfall cloud map; determining the reservoir area in the administrative division data, and intercepting the corresponding reservoir area in the dynamic rainfall cloud map to obtain a dynamic rainfall cloud map of the reservoir area; setting a number of sampling areas in the dynamic rainfall cloud map corresponding to the reservoir area according to the spatial position and / or rainfall distribution, and generating a corresponding two-dimensional array for each sampling area, and the first dimension element in the two-dimensional array is used to indicate the corresponding Rainfall time series data, the second dimension element in the two-dimensional array is used to indicate the temperature and humidity time series data corresponding to the current sampling area; extract the first dimension elements in all two-dimensional arrays, and output the rainfall characteristics after superimposing the rainfall data at the same time; extract the second dimension elements in all two-dimensional arrays, and output the temperature and humidity characteristics after averaging the temperature and humidity data at the same time; obtain the current water level data of the reservoir in the water level time series data, and update the current water level data according to the rainfall characteristics; determine the temperature influence coefficient α and the humidity influence coefficient β according to the water temperature data and the temperature and humidity characteristics, and calculate the evaporation characteristics of the reservoir through the following formula through the updated current water level data and the horizontal surface area of ​​the reservoir:

[0010]

[0011] Among them, E i is the evaporation characteristic, α is the temperature influence coefficient, β is the humidity influence coefficient, f(W i is the wind speed function derived from meteorological data, S is the horizontal surface area of ​​the reservoir, Q0 is the current water level data, P0 is the current water vapor pressure, T0 is the water temperature data, T i is the temperature characteristic, U i is the humidity feature, C0 is the adjustment parameter, i is any time node from the current time node to the predicted time node, and n is the number of all time nodes from the current time node to the predicted time node.

[0012] In a possible implementation of the present application, the extraction / determination process of the external impact feature also includes: obtaining the historical storage data and historical meteorological data of the river information in the time period corresponding to the current time node and the predicted time node, the historical storage data is the average of the storage data in the historical time period corresponding to the time period in the previous N2 years, and the historical meteorological data is the meteorological data in the historical time period corresponding to the time period in the previous N1 years, wherein N1≥N2, and the initial value of N2 is 1; performing similarity calculation on the historical meteorological data and the meteorological information, the similarity calculation at least includes rainfall similarity calculation and temperature and humidity similarity calculation; in the similarity When the result of the similarity calculation is greater than or equal to the preset similarity threshold, the historical flow data is determined as the inflow flow characteristic, and when the result of the similarity calculation is less than the preset similarity threshold, N2=N2+1 is executed, and the similarity calculation is performed again until the result of the similarity calculation is greater than or equal to the preset similarity threshold and / or the value of N2 is greater than 5; the outflow flow is identified in the water scheduling plan, and the outflow flow includes at least any one or more of the irrigation water, power generation water, navigation water, urban water supply, discharge flow and leakage loss of the reservoir in the time period corresponding to the current time node and the predicted time node; after the outflow flow is time-series processed, the outflow flow characteristic is obtained.

[0013] In a possible implementation of the present application, a correlation analysis is performed on the external influence characteristics and the reservoir water level, including: in the water level time series data, historical water level data Q[q1,q2,…,qn] corresponding to several time nodes between the current time node and the predicted time node are extracted, and historical rainfall characteristics R[r1,r2,…,rn], historical evaporation characteristics E[e1,e2,…,ei], historical inflow characteristics D[d1,d2,…,dn], and historical outflow characteristics F[f1,f2,…,fn] corresponding to the historical water level data are obtained. ]; calculate the means and variances corresponding to Q, R, E, D, and F respectively, and calculate the covariances between Q and R, Q and E, Q and D, and Q and F based on the means corresponding to Q, R, E, D, and F respectively, and calculate the standard deviations between Q and R, Q and E, Q and D, and Q and F based on the variances corresponding to Q, R, E, D, and F respectively; determine the ratio of the covariance to the standard deviation as the correlation coefficient between Q and R, Q and E, Q and D, and Q and F, and generate a correlation coefficient matrix; after normalizing the elements in the correlation coefficient matrix, assign the value of each element to the weight of the external influencing feature.

[0014] In a possible implementation of the present application, the water quality information includes at least the physical entity content, and the geological information includes at least the river sediment content; the calculation process of the first internal impact data includes: obtaining the historical inflow mean value of the inflow river in the non-rainy season, and calculating the predicted inflow flow of the inflow river in the time period corresponding to the current time node to the preset time node based on the historical inflow mean value; extracting the rainfall data of the corresponding area of ​​the inflow river in the rainfall dynamic cloud map, and accumulating the actual inflow flow based on the rainfall data and the predicted inflow flow; obtaining the vegetation coverage rate within the preset range on both sides of the inflow river; and calculating the vegetation coverage rate based on the vegetation coverage rate. When the coverage rate is higher than the preset coverage rate threshold, the physical entity content corresponding to the inflowing river under the influence of the rainfall data is predicted by the vegetation coverage rate, and the physical entity includes at least any one or more of leaves, branches and trunks, and when the vegetation coverage rate is lower than the preset coverage rate threshold, the physical entity content is taken as zero; according to the first product between the sediment content of the river water and the actual inflow flow, and the second product between the physical entity content and the actual inflow flow, the amount of non-water matter brought into the reservoir by the inflowing river is determined; through the non-water matter amount and the storage capacity of the reservoir, the first impact data of the inflowing river on the water level of the reservoir is calculated.

[0015] In a possible implementation of the present application, the underground geological data at least includes a water absorption index, and the second internal impact data is determined by the following formula: Among them, λ is the second internal influence data, k is the water absorption index, Q0 is the initial predicted water level, Q u is the groundwater level data, P c is the reservoir bottom edge pressure, P u is the groundwater level pressure, and R is the reservoir bottom shape parameter.

[0016] In a possible implementation of the present application, the water level prediction model adopts a time series model and / or a neural network model; correcting the initial predicted water level includes: performing hyperparameter optimization on the water level prediction model using a particle swarm algorithm; updating the water level prediction model based on the optimization results, and retraining the updated water level prediction model to obtain the water level correction model; inputting the initial predicted water level, the first internal influence data, and the second internal influence data into the water level correction model to perform a secondary prediction on the initial predicted water level to obtain the final predicted water level.

[0017] In a possible implementation of the present application, correcting the initial predicted water level also includes: taking the initial predicted water level as a reference, determining the positive influence parameter corresponding to the first internal influence data and determining the negative influence parameter corresponding to the second internal influence data, wherein the influence parameter is used to indicate the degree of influence of the first internal influence data and / or the second internal influence data on the upward and downward fluctuation of the initial predicted water level; calculating the product between the initial predicted water level and the positive influence parameter and the negative influence parameter, and determining the final predicted water level according to the product.

[0018] On the other hand, the present application also provides a water level prediction device for a water conservancy project, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor so that the at least one processor can execute a water level prediction method for a water conservancy project as described in any one of the above embodiments.

[0019] The present application provides a water level prediction method and device for water conservancy projects, which has high accuracy or prediction precision, which is specifically reflected in the following two aspects:

[0020] On the one hand, on the basis of considering factors that directly affect or have a high degree of influence on the reservoir water level, such as rainfall, evaporation, inflow and outflow, other factors that may affect the reservoir water level, such as river meteorological information, river information, groundwater level information and underground geological information, are added to consider, and the selection or consideration of influencing factors is more comprehensive; at the same time, in order not to affect the prediction accuracy of the model, this application chooses to input rainfall, evaporation, inflow and outflow, which are factors with a high degree of influence, for prediction during the initial prediction process, and after obtaining the initial predicted water level, the above-mentioned other possible influencing factors are used to correct or adjust the initial predicted water level. This can ensure that the water level prediction model is not affected by too many input factors and predicts an initial water level with a higher degree of accuracy. At the same time, other possible influencing factors can be taken into consideration, so that the predicted water level can be corrected, thereby further improving the accuracy of the water level prediction results.

[0021] On the other hand, the present application also performs pre-processing on the feature extraction process of the influencing factors, so that more targeted extraction can be performed for different influencing factors. For example, for rainfall, the traditional prediction process is to input the rainfall of the entire area where the reservoir is located as input data into the model. In this way, the input rainfall data is bound to increase relative to the rainfall of the corresponding water level. However, the present application can directly obtain the actual rainfall in the reservoir area through the introduction of a dynamic rainfall cloud map, so that only this part of the rainfall characteristics is input into the model for prediction, making the input of the model more accurate. In this way, the correlation between the data input to the model and the water level can be higher, thereby making the prediction results of the model more accurate; at the same time, after the feature extraction is completed, it is also subjected to a correlation analysis with the water level, so that the features of the input model carry the weights determined according to the results of the correlation analysis, and the degree of influence of each factor on the reservoir water level is characterized and reflected, so that the model can pay attention to the degree of influence of each factor when predicting, which can further improve the accuracy of the model prediction. Through the above two aspects, this application can obtain more accurate prediction results, ensure the prediction accuracy of reservoir water level, and thus provide more valuable decision-making basis for reservoir scheduling, early warning, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0023] Figure 1 A flow chart of a water level prediction method for water conservancy projects provided in this application;

[0024] Figure 2 A schematic diagram of the structure of a water level prediction device for water conservancy projects provided in this application. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in this application will be described clearly and completely below in conjunction with the drawings in this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0026] The present application provides a water level prediction method and device for water conservancy projects. By pre-processing the feature extraction process of external influencing data and correcting the prediction results and / or prediction models, accurate prediction of reservoir water levels in water conservancy projects is achieved, providing a decision-making basis for flood warnings downstream of the reservoir.

[0027] The method in this application is described in detail below with reference to the accompanying drawings.

[0028] Figure 1 A flow chart of a water level prediction method for water conservancy projects provided in this application, such as Figure 1 As shown, the water level prediction method in this application includes at least the following execution steps:

[0029] Step 101: Read the water level time series data of the reservoir at preset time intervals.

[0030] The reservoir water level prediction scheme in this application is implemented based on a time series model or a neural network model, such as an initial prediction model sampling LSTM model, etc. At the same time, since the reservoir water level has a certain correlation in time, this application first reads the reservoir water level time series data when making a prediction.

[0031] In one example, the water level time series data is obtained by extracting the water level data of each historical time node from the current time node as the starting point and the preset time interval as the interval, and then sorting them in the positive order of the time series. For example, if the current time node is t0 day and the preset time interval is 1 day, then the read water level time series data is: [..., t 0-2 ,t 0-1 ,t0], as for t 0-2 How many data exist before is determined by the length of the water level time series data. Of course, you can also directly choose which time node to start reading based on the length requirement of the water level time series data. For example, if you want to read the water level data within 10 days to form the water level time series data, and the preset time interval is 1 day, then the final water level time series data is: [t 0-9 ,...,t 0-2 ,t 0-1 ,t0].

[0032] It should be noted that, if the water level time series data is in days, then preferably, each water level data corresponds to the water level data at the same time point in a day.

[0033] Step 102: Obtain external impact data of the reservoir water level, perform feature extraction on the external impact data, obtain external impact features and store them in the blockchain.

[0034] In a possible implementation of the present application, the external influencing data of the reservoir water level refers to the rainfall data, evaporation data, inflow data and outflow data that have a greater impact on the water level of the reservoir, wherein each data corresponds to rainfall features, evaporation features, inflow features and outflow features after feature extraction. The present application pre-places the feature extraction process of the model through this step, or this step can also be considered as processing the acquired raw data to obtain input data that is more suitable for the model prediction process, and the feature extraction process or processing process of each data is targeted, so that the processed data or extracted features can better show the impact on the reservoir water level.

[0035] In addition, the features proposed in this application can also be stored in the blockchain, using the characteristics of the blockchain to make the data tamper-proof and support the traceability of data affecting the reservoir water level.

[0036] Specifically, the acquisition process and feature extraction process of the aforementioned influencing data are as follows:

[0037] 1) For rainfall data:

[0038] This application first obtains the geographical location data of the reservoir. The geographical location data here can be longitude and latitude data or administrative division data. Then, the meteorological data of the area where the reservoir is located between the current time node and the predicted time node is obtained according to the obtained geographical location. For example, the current time node is the same day, and the predicted time node is 3 days later, then the meteorological data from the same day to the next 3 days is obtained. The acquisition method can be obtained through the meteorological monitoring stations around the reservoir, or the meteorological data on the mobile terminal. Rainfall data and temperature and humidity data are extracted from the acquired meteorological data. The rainfall data corresponds to the rainfall data, and the temperature and humidity data corresponds to the evaporation data. Therefore, the rainfall characteristics are obtained by extracting features from the rainfall data.

[0039] Specifically, when the geographical location data corresponding to the reservoir is administrative division data, a rainfall dynamic cloud map of the administrative division data between the current time node and the predicted time node is obtained, and then based on the corresponding regional position of the reservoir in the administrative division data, the corresponding reservoir area is intercepted in the rainfall dynamic cloud map to obtain a rainfall dynamic cloud map of the reservoir area. For example, the reservoir is in City X, and the rainfall data generally obtained is the rainfall cloud map displayed on the map of City X. In order to ensure the accuracy of the impact of rainfall on the water level of the reservoir, this application only counts the rainfall corresponding to the reservoir area. Therefore, the reservoir area is intercepted on the map of City X, and the rainfall dynamic cloud map of the reservoir is obtained. It should be noted that the dynamic here refers to a certain location point of the reservoir. The cloud map shows all the rainfall information of the location point in the time period between the current time node and the predicted time node. Afterwards, a number of sampling areas are set in the rainfall dynamic cloud map corresponding to the reservoir area according to the spatial position and / or rainfall distribution, and a corresponding two-dimensional array is generated for each sampling area. The first dimension element in the two-dimensional array is used to indicate the rainfall time series data corresponding to the current sampling area. Generally speaking, the rainfall levels displayed on the dynamic rainfall cloud map corresponding to the reservoir area are similar and not much different. Therefore, it is preferred to set the sampling area according to the spatial position. Assuming that there are four sampling points or sampling areas in the reservoir area, four two-dimensional data corresponding to the sampling area can be extracted based on the dynamic rainfall cloud map of the reservoir, and then the first-dimensional elements of the four two-dimensional data are extracted. Each set of first-dimensional elements contains the rainfall data of the sampling area from the current time node to the predicted time node. Therefore, by superimposing the rainfall data corresponding to the same time in the extracted first-dimensional elements, the rainfall characteristics are obtained.

[0040] The above process is visualized as follows: the corresponding dynamic rainfall cloud map is displayed on the upper layer of the reservoir visualization plane map, the analyzed two-dimensional array is displayed on the upper layer of the dynamic rainfall cloud map, the first-dimensional elements in the two-dimensional array are extracted to form a new array, and then the elements corresponding to the same time position in each new data are superimposed in the upper layer, and the dynamic superposition process is displayed, and the top layer displays the rainfall characteristics obtained by superposition.

[0041] It should be noted that even if the rainfall in the reservoir area does not vary much, in order to improve the prediction accuracy of the reservoir water level, the accuracy of the data needs to be guaranteed from the feature or input data level. Therefore, this application sets up sampling areas to superimpose the rainfall in each area to replace the rainfall in the entire area, which is equivalent to dividing the reservoir into finer dimensions to make the rainfall data more accurate.

[0042] 2) Regarding evaporation data:

[0043] The evaporation of the reservoir is mainly related to the temperature and humidity data of the meteorological information corresponding to the reservoir. Therefore, temperature and humidity data are also added to the dynamic cloud map of rainfall in the aforementioned reservoir area, and this temperature and humidity data is added based on time, and the temperature and humidity data with the same corresponding time are added to the rainfall data at the same time. In this way, the second dimension element of the two-dimensional array can be used to represent the temperature and humidity data. In an example, the temperature and humidity data here can be a temperature range and / or humidity range, and in order to make the temperature and humidity characteristics finally obtained able to characterize the temperature and humidity of the entire reservoir area, rather than the temperature and humidity of a certain sampling area in the reservoir, this application will extract the second dimensional elements, and the temperature and humidity data with the same corresponding time are averaged, and the temperature and humidity characteristics are obtained after the processing.

[0044] Furthermore, the current water level data of the current time node is extracted from the water level time series data of the reservoir, and the current water level data is updated by the aforementioned rainfall characteristics. The purpose of this step is to assume that all rainfall falls into the reservoir area, that is, the evaporation of rainwater during the falling process is not considered, but after all rainfall falls, the evaporation after the water level change caused by rainfall is considered. At the same time, the evaporation can also be related to the water temperature data. Therefore, when determining the evaporation characteristics, the present application first calculates the temperature influence coefficient and the humidity influence coefficient based on the water temperature data and the temperature and humidity characteristics. It should be noted that the temperature influence coefficient has a positive impact on the evaporation characteristics of the reservoir, while the humidity influence coefficient has a negative impact on the evaporation characteristics of the reservoir; then, based on the following formula, the evaporation characteristics are calculated:

[0045]

[0046] Among them, E i is the evaporation characteristic, α is the temperature influence coefficient, β is the humidity influence coefficient; f(W i ) is the wind speed function derived from meteorological data, which can be obtained according to the existing algorithm; S is the horizontal surface area of ​​the reservoir, Q0 is the current water level data, P0 is the current water vapor pressure, T0 is the water temperature data, and these quantities are all known quantities; T i is the temperature characteristic, U i is the humidity feature, C0 is the adjustment parameter, which is a constant, i is any time node from the current time node to the predicted time node, and n is the number of all time nodes from the current time node to the predicted time node.

[0047] It should be noted that the calculated evaporation feature is also a time series feature, that is, assuming that the evaporation feature is represented by an array, there is a time series relationship between the elements in the array, and the time corresponding to each element is the time node between the current time node and the predicted time node, and these time nodes correspond to the time nodes in the rainfall feature and run through the calculation process of the evaporation feature.

[0048] 3) For inbound traffic data:

[0049] The inflow flow data in this application only considers the river data that has an inflow relationship with the reservoir, and it is believed that the impact of rainfall around the river and / or rainfall in the area around the reservoir on the reservoir water level is also achieved through the river, that is, the rainfall first falls into the river and is then brought into the water level by the river.

[0050] In a possible implementation of the present application, first determine the river information associated with the reservoir, where the association refers to the river water flowing into the reservoir. Obtain the historical storage data and historical meteorological data of the river information in the time period corresponding to the current time node and the predicted time node. The historical storage data here is the average of the storage data in the historical time period corresponding to the previous N2 years, and the historical meteorological data is the meteorological data in the historical time period corresponding to the previous N1 years, where N1≥N2, and the initial value of N2 is 1; for example, the time period between the current time node and the preset time node is August 3-5, then the historical storage data obtained can be the average of the storage flow from August 3 to 5 of each year in the past year, and the historical meteorological data is also the meteorological data from August 3 to 5 of the previous year. The year is used as the unit here because the meteorological data or flow data at the same time node in different years in the same area are similar, and have a certain reference value compared to other times.

[0051] Furthermore, the historical meteorological data and meteorological information are similarly calculated, and the similarity calculation process includes at least rainfall similarity calculation and temperature and humidity similarity calculation, which are also the two parameters in meteorological data that have the greatest impact on river and reservoir water levels. In addition, the similarity calculation process here can be implemented by existing technologies or algorithms, and this application will not go into details here. If the result of the similarity calculation is greater than or equal to the preset similarity threshold, the historical flow data is determined as the inflow flow feature, and if the result of the similarity calculation is less than the preset similarity threshold, N2=N2+1 is executed, the purpose is to find a year with more similar meteorological data, and the similarity calculation is performed again until the result of the similarity calculation is greater than or equal to the preset similarity threshold and / or the value of N2 is greater than 5, and the inflow flow feature is obtained.

[0052] 4) For outbound traffic data:

[0053] This application obtains the water dispatching plan of the reservoir between the current time node and the predicted time node, and extracts the corresponding outflow flow data of any one or more of irrigation water, power generation water, navigation water, urban water supply, water discharge flow and leakage loss in the water dispatching plan. After that, the outflow flow data of each item with the same corresponding time are superimposed and arranged in positive order according to the time series, so as to obtain the outflow flow characteristics of the reservoir.

[0054] Finally, the external influence characteristics are constructed through the rainfall characteristics, evaporation characteristics, inflow characteristics and outflow characteristics extracted through the above four processes.

[0055] Step 103: perform a correlation analysis on the external impact characteristics and the reservoir water level, and assign weights to the external impact characteristics according to the analysis results.

[0056] Through the above process, it is determined that the external influencing features input into the water level prediction model for water level prediction mainly include rainfall features, evaporation features, inflow features and outflow features. Therefore, when performing correlation analysis, the relationship between these four items and the reservoir water level is also analyzed separately. Since the influence of influencing factors on the reservoir water level has a certain regularity, this application performs the above correlation analysis process through the historical water level data Q[q1,q2,…,qn], historical rainfall features R[r1,r2,…,rn], historical evaporation features E[e1,e2,…,ei], historical inflow features D[d1,d2,…,dn] and historical outflow features F[f1,f2,…,fn] corresponding to several time nodes between the current time node and the prediction time node, to calculate the correlation between the reservoir water level and the external influencing features. The correlation here can be used to characterize the degree of influence of each external influencing feature on the water level of the reservoir. In this way, no matter what model is used for prediction in the water level prediction model in this application, the premise is to ensure that the adopted water level prediction model itself has a certain prediction accuracy, and the degree of influence of each influencing factor can be considered when predicting the future water level of the reservoir, so that the prediction accuracy is further improved, and the applicability of the present application scheme is also improved. It should also be noted that the historical corresponding time node is selected here (for example, the current time is August 2, and the water level on August 5 is predicted, and the historical data of August 3-5 of the previous year is used for correlation analysis), because the data at the same time in different years have a higher similarity than the data at other time points, and the calculated results are more credible.

[0057] The specific correlation analysis process is as follows: first, calculate the means and variances corresponding to Q, R, E, D, and F respectively; then, based on the means corresponding to Q, R, E, D, and F respectively, calculate the covariances between Q and R, Q and E, Q and D, and Q and F; and based on the variances corresponding to Q, R, E, D, and F respectively, calculate the standard deviations between Q and R, Q and E, Q and D, and Q and F; then, determine the ratio between the covariance and the standard deviation as the correlation coefficient between Q and R, Q and E, Q and D, and Q and F, and generate a correlation coefficient matrix with four elements, each element corresponding to the degree of correlation between different external influencing characteristics and reservoir water levels; finally, normalize the elements in the correlation coefficient matrix, and assign the value of each element to the weight of each external influencing characteristic.

[0058] Step 104: input the water level time series data and the external influence features carrying weights into a pre-trained water level prediction model to obtain an initial predicted water level.

[0059] After the above three steps are completed, the water level time series data and the weighted external influence features are input into the pre-trained water level prediction model to output the initial predicted water level. In one example, since the external influence features in this application have been subjected to correlation analysis and different weights have been assigned based on the analysis results to represent different degrees of influence, the water level prediction model here is likely to adopt several types of models commonly used in current technology, which has lower requirements on the model, making this solution more adaptable; at the same time, pre-positioning the feature extraction process can also increase the training time and difficulty of the model, making the model more lightweight.

[0060] Step 105: Calculate the first internal impact data of the reservoir water level according to the water quality information and geological information of the river entering the reservoir and the inflow flow of the reservoir.

[0061] In this application, factors that may affect the water level of the reservoir other than the aforementioned external influencing features are roughly divided into two categories: one is the first internal influencing data that has a positive impact on the water level of the reservoir, that is, the water level of the reservoir may rise; the other is the second internal influencing data that has a negative impact on the water level of the reservoir, that is, the water level of the reservoir may decrease. The initial predicted water level predicted above is corrected by these two types of data, so that a water level prediction result with higher precision or accuracy is obtained on the basis of a prediction result with a certain precision.

[0062] The first internal impact data in this application is obtained through the water quality information, geological information and inflow of the river entering the reservoir. Specifically, the historical inflow mean of the river entering the reservoir in the non-rainy season is obtained, and the predicted inflow of the river entering the reservoir in the time period corresponding to the current time node to the preset time node is calculated based on the historical inflow mean, and then the rainfall data of the corresponding area of ​​the river entering the reservoir is extracted from the rainfall dynamic cloud map, and the actual inflow is obtained by accumulating the rainfall data and the predicted inflow. This step is equivalent to taking the inflow in the non-rainy season as the basic flow, clarifying the flow of the river entering the reservoir without the influence of rainfall, and then applying the rainfall data to the basic inflow to obtain the actual inflow. In this way, compared with directly selecting the inflow in the rainy season or not limiting the non-rainy season, the difference in inflow caused by different rainfall is avoided. Then, the vegetation coverage rate within a preset range on both sides of the inflowing river is obtained, and when the vegetation coverage rate is higher than the preset coverage rate threshold, the vegetation coverage rate is used to predict the corresponding physical entity content of the inflowing river under the influence of rainfall data. The physical entity here includes at least any one or more of leaves, branches and trunks, that is, the vegetation coverage rate is used to predict how many branches, trunks and other physical entities that affect the water level will be carried by the river due to rainfall in a time period; and when the vegetation coverage rate is lower than the preset coverage rate threshold, the physical entity content is taken as zero, and the first product between the sediment content of the river water (which can be calculated by the existing algorithm based on the geological data of the river, or is a known quantity) and the actual inflow flow, and the second product between the physical entity content and the actual inflow flow are accumulated to obtain the non-water mass brought into the reservoir by the inflowing river, that is, the sediment volume and the physical entity mass; finally, the first impact data of the inflowing river on the reservoir water level is calculated through the non-water mass and the storage capacity of the reservoir.

[0063] That is, the first impact data here reflects the content of other substances brought into the reservoir by the inflowing river in addition to the outflow flow that may affect the reservoir water level.

[0064] Step 106: Measure the underground water level data and underground geological data of the reservoir, and determine the second internal influence data of the reservoir water level through the measurement results.

[0065] Depending on the geographical location of the reservoir, the bottom of some reservoirs is composed of materials with poor water absorption capacity such as rocks, while the bottom of some reservoirs is soil with strong water absorption capacity, such as natural reservoirs such as Hu Po. Therefore, the underground conditions of the reservoir will also have a certain impact on the water level of the reservoir. These impacts are characterized by the second internal impact data.

[0066] In a possible implementation of the present application, in order to obtain the second internal impact data, it is first necessary to obtain the groundwater level data Q of the reservoir at the current time node u, and underground geological data, where the underground geological data refers to the water absorption index k of the reservoir underground, and then the second internal impact data is calculated by the following formula:

[0067]

[0068] Among them, λ is the second internal impact data, Q0 is the initial predicted water level, P c is the reservoir bottom edge pressure, P u is the groundwater level pressure, R is the reservoir bottom shape parameter, and it takes different values ​​according to the different shapes of the reservoir bottom.

[0069] Step 107: Based on the first internal influence data and the second internal influence data, the initial predicted water level is corrected by a water level correction model to obtain a final predicted water level.

[0070] After obtaining the first internal data and the second internal data, the present application uses the two data to correct the predicted water level of book publishing, and there are two correction methods:

[0071] Method 1: Correct the water level prediction model. Specifically, the particle swarm optimization algorithm is used to optimize the hyperparameters of the water level prediction model. The hyperparameters here refer to the parameters of the model itself, such as the number of convolution kernels. The water level prediction model is updated through the optimization results, and the updated water level prediction model is trained twice to obtain a water level correction model. The initial predicted water level, the first internal influence data, and the second internal influence data are input into the water level correction model to output the final predicted water level. This process performs hyperparameter optimization and model training again because the process of obtaining the final predicted water level is considered to be a process derived from the prediction of the initial predicted water level, in which the influencing factors are only the first internal influence data and the second internal influence data. Hyperparameter optimization makes the model more adaptable to new influencing factors, and secondary training ensures the prediction accuracy of the new model. The particle swarm algorithm hyperparameter optimization process and the model secondary training process here can be implemented through existing solutions, and this application will not go into details here.

[0072] In one example, the water level prediction model can be implemented using a model with prediction capabilities such as an LSMT model and a Seq2Seq model.

[0073] Method 2: Correct the initial predicted water level. Specifically, the positive influence parameter and the negative influence parameter are obtained by quantifying the influence of the first internal influence data and the second internal influence data on the reservoir water level. The quantification process here can also be implemented by existing technologies or algorithms, and the product between the initial predicted water level and the positive influence parameter and the negative influence parameter is calculated, and finally the final predicted water level is determined according to the product. In this method, there is no need to use a model, but only to adjust the initial predicted water level through the influencing data itself, so that the initial predicted water level rises or falls slightly, and a higher accuracy final predicted water level is obtained.

[0074] By using any of the above two correction methods, the final predicted water level with relatively high precision and accuracy can be obtained.

[0075] Based on the same inventive concept, the present application also provides a water level prediction device for water conservancy projects, the structure of which is as follows: Figure 2 shown.

[0076] Figure 2 This is a schematic diagram of the structure of a water level prediction device for water conservancy projects provided in this application. Figure 2 As shown, the water level prediction device 200 for water conservancy projects in the present application specifically includes: at least one processor 201; and a memory 203 that is communicatively connected to the at least one processor 201 (connected via a bus 202); wherein the memory 203 stores instructions that can be executed by the at least one processor 201, so that the at least one processor 201 can execute a water level prediction method for water conservancy projects as described in the above embodiment.

[0077] In a possible implementation of the present application, the aforementioned processor is used to execute: reading the water level time series data of the reservoir at preset time intervals; obtaining the external impact data of the reservoir water level, and performing feature extraction on the external impact data to obtain the external impact features, and then performing correlation analysis on the external impact features and the reservoir water level, and assigning weights to the external impact features according to the analysis results; inputting the water level time series data and the external impact features carrying weights into a pre-trained water level prediction model to obtain an initial predicted water level; calculating the first internal impact data of the reservoir water level through the inflow flow of the reservoir according to the water quality information and geological information of the river entering the reservoir; measuring the groundwater level data and underground geological data of the reservoir, and determining the second internal impact data of the reservoir water level through the measurement results; based on the first internal impact data and the second internal impact data, correcting the initial predicted water level through the water level correction model to obtain the final predicted water level.

[0078] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0079] The device and method provided in the present application correspond one to one, and therefore, the device also has similar beneficial technical effects as its corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the device will not be repeated here.

[0080] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, devices, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media that include computer-usable program code.

[0081] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0082] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A water level prediction method for a water conservancy project, wherein the water conservancy project at least includes a reservoir, characterized in that: The method comprises: Read the reservoir water level time series data at preset time intervals; Obtain external impact data of the reservoir water level, perform feature extraction on the external impact data, obtain external impact features and store them in the blockchain, then perform correlation analysis on the external impact features and the reservoir water level, and assign weights to the external impact features according to the analysis results; Inputting the water level time series data and the external influence features carrying weights into a pre-trained water level prediction model to obtain an initial predicted water level; According to the water quality information and geological information of the river entering the reservoir, the first internal impact data of the reservoir water level is calculated through the inflow flow of the reservoir; Measure the groundwater level data and underground geological data of the reservoir, and determine the second internal impact data of the reservoir water level through the measurement results; Based on the first internal influence data and the second internal influence data, the initial predicted water level is corrected by a water level correction model to obtain a final predicted water level.

2. A water level prediction method for water conservancy projects according to claim 1, characterized in that: The method further comprises: Reading the geographical location data of the reservoir, and obtaining the meteorological data of the geographical location data between the current time node and the predicted time node; Extracting rainfall data and temperature and humidity data from the meteorological data, performing feature extraction on the rainfall data to obtain rainfall characteristics corresponding to the reservoir at the prediction time node, and performing feature extraction on the temperature and humidity data, and combining the measured water temperature data to obtain evaporation characteristics corresponding to the reservoir at the prediction time node; Determine river information associated with the reservoir, and determine inflow characteristics corresponding to the reservoir based on the river information and meteorological information corresponding to the river information between a current time node and a predicted time node; Obtaining a water dispatching scheme of the reservoir between the current time node and the predicted time node, so as to extract the outflow characteristics corresponding to the reservoir according to the water dispatching scheme; Based on the rainfall characteristics, the evaporation characteristics, the inflow characteristics and the outflow characteristics, the external influence characteristics of the reservoir water level are constructed.

3. A water level prediction method for water conservancy projects according to claim 2, characterized in that: The process of extracting / determining the external impact features includes: When the geographical location data is administrative division data, a dynamic rainfall cloud map of the administrative division data between the current time node and the predicted time node is obtained, and the temperature and humidity data are dynamically added to the dynamic rainfall cloud map; Determine a reservoir area in the administrative division data, and intercept the corresponding reservoir area in the rainfall dynamic cloud map to obtain a rainfall dynamic cloud map of the reservoir area; In the rainfall dynamic cloud map corresponding to the reservoir area, a number of sampling areas are set according to the spatial position and / or rainfall distribution, and a corresponding two-dimensional array is generated for each sampling area, the first dimension element in the two-dimensional array is used to indicate the rainfall time series data corresponding to the current sampling area, and the second dimension element in the two-dimensional array is used to indicate the temperature and humidity time series data corresponding to the current sampling area; Extract the first-dimensional elements from all two-dimensional arrays, and after superimposing the rainfall data with the same corresponding time, output the rainfall characteristics; The second-dimensional elements in all two-dimensional arrays are extracted, and the temperature and humidity data with the same corresponding time are averaged, and the temperature and humidity features are output; Acquire current water level data of the reservoir from the water level time series data, and update the current water level data according to the rainfall characteristics; According to the water temperature data and the temperature and humidity characteristics, the temperature influence coefficient α and the humidity influence coefficient β are determined, and the evaporation characteristics of the reservoir are calculated by the following formula using the updated current water level data and the horizontal surface area of ​​the reservoir: Among them, E i is the evaporation characteristic, α is the temperature influence coefficient, β is the humidity influence coefficient, f(W i ) is the wind speed function derived from meteorological data, S is the horizontal surface area of ​​the reservoir, Q0 is the current water level data, P0 is the current water vapor pressure, T0 is the water temperature data, T i is the temperature characteristic, U i is the humidity feature, C0 is the adjustment parameter, i is any time node from the current time node to the predicted time node, and n is the number of all time nodes from the current time node to the predicted time node.

4. A water level prediction method for water conservancy projects according to claim 3, characterized in that: The process of extracting / determining the external impact features also includes: Obtain the historical storage data and historical meteorological data of the river information in the time period corresponding to the current time node and the predicted time node, wherein the historical storage data is the average of the storage data in the historical time period corresponding to the time period in the previous N2 years, and the historical meteorological data is the meteorological data in the historical time period corresponding to the time period in the previous N1 years, wherein N1≥N2, and the initial value of N2 is 1; Performing similarity calculation on the historical meteorological data and the meteorological information, wherein the similarity calculation at least includes rainfall similarity calculation and temperature and humidity similarity calculation; When the result of the similarity calculation is greater than or equal to the preset similarity threshold, the historical traffic data is determined as the inbound traffic feature, and when the result of the similarity calculation is less than the preset similarity threshold, N2=N2+1 is executed, and the similarity calculation is performed again until the result of the similarity calculation is greater than or equal to the preset similarity threshold and / or the value of N2 is greater than 5; Identify the outflow flow in the water scheduling plan, the outflow flow at least including any one or more of irrigation water, power generation water, navigation water, urban water supply, discharge flow and leakage loss of the reservoir in the time period corresponding to the current time node and the predicted time node; After time series processing of the outbound flow, the outbound flow characteristics are obtained.

5. A water level prediction method for water conservancy projects according to claim 2, characterized in that: Conduct correlation analysis on the external impact characteristics and reservoir water level, including: From the water level time series data, extract the historical water level data Q[q1, q2, ..., qn] corresponding to several time nodes between the current time node and the predicted time node, and obtain the historical rainfall characteristics R[r1, r2, ..., rn], historical evaporation characteristics E[e1, e2, ..., ei], historical inflow characteristics D[d1, d2, ..., dn], and historical outflow characteristics F[f1, f2, ..., fn] corresponding to the historical water level data; Calculate the means and variances of Q, R, E, D, and F respectively, and calculate the covariances between Q and R, Q and E, Q and D, and Q and F based on the means of Q, R, E, D, and F respectively, and calculate the standard deviations between Q and R, Q and E, Q and D, and Q and F based on the variances of Q, R, E, D, and F respectively; Determine the ratio of the covariance to the standard deviation as the correlation coefficient between Q and R, Q and E, Q and D, Q and F, and generate a correlation coefficient matrix; After normalizing the elements in the correlation coefficient matrix, the value of each element is assigned as the weight of the external influence feature.

6. A water level prediction method for water conservancy projects according to claim 1, characterized in that: The water quality information at least includes the physical entity content, and the geological information at least includes the river sediment content; the calculation process of the first internal impact data includes: Obtaining the historical average inflow flow of the inflow river in the non-rainy season, and calculating the predicted inflow flow of the inflow river in the time period corresponding to the current time node to the preset time node based on the historical average inflow flow; Extracting rainfall data of the area corresponding to the inflow river from the rainfall dynamic cloud map, and accumulating the rainfall data and the predicted inflow flow to obtain the actual inflow flow; Obtain the vegetation coverage within the preset range on both sides of the inflowing river; When the vegetation coverage rate is higher than a preset coverage rate threshold, the physical entity content corresponding to the inflow river under the influence of the rainfall data is predicted by the vegetation coverage rate, the physical entity at least including any one or more of leaves, branches and trunks, and when the vegetation coverage rate is lower than the preset coverage rate threshold, the physical entity content is set to zero; Determine the mass of non-water substances brought into the reservoir by the inflowing river according to a first product between the sediment content of the river water and the actual inflowing flow rate, and a second product between the physical entity content and the actual inflowing flow rate; The first impact data of the inflow river on the reservoir water level is calculated based on the mass of the non-water substances and the storage capacity of the reservoir.

7. A water level prediction method for water conservancy projects according to claim 1, characterized in that: The underground geological data at least includes a water absorption index, and the second internal impact data is determined by the following formula: Among them, λ is the second internal influence data, k is the water absorption index, Q0 is the initial predicted water level, Q u is the groundwater level data, P c is the reservoir bottom edge pressure, P u is the groundwater level pressure, and R is the reservoir bottom shape parameter.

8. A water level prediction method for water conservancy projects according to any one of claims 6-7, characterized in that: The water level prediction model adopts a time series model and / or a neural network model; the correction of the initial predicted water level includes: Using a particle swarm algorithm to optimize hyperparameters of the water level prediction model; The water level prediction model is updated based on the optimization result, and the updated water level prediction model is trained again to obtain the water level correction model; The initial predicted water level, the first internal influence data and the second internal influence data are input into the water level correction model to perform a secondary prediction on the initial predicted water level to obtain a final predicted water level.

9. A water level prediction method for water conservancy projects according to any one of claims 6-7, characterized in that: Correcting the initial predicted water level also includes: Taking the initial predicted water level as a reference, determining a positive influence parameter corresponding to the first internal influence data and determining a negative influence parameter corresponding to the second internal influence data, wherein the influence parameter is used to indicate the degree of influence of the first internal influence data and / or the second internal influence data on the upward and downward fluctuation of the initial predicted water level; The product of the initial predicted water level and the positive influence parameter and the negative influence parameter is calculated, and the final predicted water level is determined according to the product.

10. A water level prediction device for water conservancy projects, characterized in that: The device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the water level prediction method for water conservancy projects according to any one of claims 1-9.

Images