Reservoir dynamic scheduling method, device and equipment and storage medium

Abstract

The invention discloses a reservoir dynamic scheduling method, device and equipment and a storage medium, and relates to the technical field of water conservancy and hydropower, and the reservoir dynamic scheduling method can rapidly predict possible reservoir runoff information according to real-time typhoon information, and then carries out reservoir dynamic scheduling according to the reservoir runoff prediction information. The method does not need to reduce the water level of the target reservoir below the flood control water level in the whole flood season and further reduce the water level of the reservoir before the typhoon comes, does not need to depend on the subjective experience of dispatchers, has prediction accuracy and a runoff prediction model optimization mechanism, and gives consideration to efficient utilization of water resources and stable and accurate flood prevention effects.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy and hydropower, and particularly to a reservoir dynamic scheduling method, device, equipment and storage medium. Background Art

[0002] As a weather phenomenon with strong seasonality, high uncertainty and great harmfulness, typhoons often bring extreme precipitation during a specific period, and this period is often also a key stage for reservoir water storage. Therefore, how to balance the flood control safety and beneficial utilization of a reservoir under the influence of typhoons is of crucial significance for promoting the positive interaction between high-quality development and high-level safety.

[0003] In terms of water resource management technology, during the typhoon landing period, the existing reservoir scheduling technology faces the following challenges: A traditional and relatively conservative scheduling method is to keep the reservoir water level below the flood control limit level during the entire flood season. Once a typhoon is forecast to come, the reservoir water level is further lowered to vacate the storage capacity. Although this scheduling method is cautious, it is prone to waste of water resources, reduce the power generation benefit of the reservoir, and may also affect the water supply benefit during the dry season. Relatively speaking, a more advanced and efficient reservoir scheduling method is dynamic control, that is, to keep the reservoir operating above the normal storage level during the flood season, and when a typhoon is predicted to come, pre-discharge in a timely manner to lower the water level below the flood control limit level. However, currently this scheduling method still mainly relies on the prediction of typhoons and their rainfall. Once a typhoon or rainfall is predicted, the reservoir water level is quickly lowered below the flood control limit level. This qualitative judgment control process lacks accuracy and optimization; in addition, this scheduling method needs to rely on the experience of dispatchers for reservoir scheduling during typhoon rainstorms. Although it has a certain degree of flexibility, it is difficult to form a fixed operating procedure for subsequent reference, has strong subjectivity, and is greatly affected by the experience and level of dispatchers. Therefore, there are still technical problems to be solved in the related technologies. Summary of the Invention

[0004] Embodiments of the present invention provide a reservoir dynamic scheduling method, device, equipment and storage medium, which can dynamically schedule a reservoir according to real-time typhoon information, taking into account the efficient utilization of water resources and stable and accurate flood control effects.

[0005] In a first aspect, an embodiment of the present invention provides a reservoir dynamic scheduling method, and the method includes:

[0006] Obtain the historical typhoon information and corresponding historical inflow runoff information of the target reservoir, and form a first variable set according to the historical typhoon information. The first variable set includes characteristic data for each time period, and the characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and the typhoon wind circles at all levels or the typhoon center point, and the azimuth of the target reservoir with respect to the typhoon center point;

[0007] For each time period in the first variable set, it is respectively corresponded with the characteristic data n time periods before (where n is greater than or equal to 2) to obtain new variables, and a second variable set is formed.

[0008] Variables with a maximum information coefficient greater than a first preset value with the historical inflow runoff information are screened from the second variable set to obtain a third variable set.

[0009] According to the magnitudes of the characteristic data in the third variable set, the top L principal components are screened and duplicate data is removed to obtain a fourth variable set.

[0010] Taking the fourth variable set as the input factor set and the historical inflow runoff information as the output value, a runoff prediction model is constructed, and the runoff prediction model is trained and optimized.

[0011] In response to receiving real-time typhoon information, the real-time typhoon information is input into the runoff prediction model, the predicted inflow runoff information is output, and the target reservoir is scheduled based on the predicted inflow runoff information and a preset scheduling strategy.

[0012] In a second aspect, an embodiment of the present invention provides a reservoir dynamic scheduling device, and the reservoir dynamic scheduling device includes:

[0013] A historical data acquisition unit, configured to acquire the historical typhoon information and the corresponding historical inflow runoff information of the target reservoir, and form a first variable set according to the historical typhoon information. The first variable set includes the characteristic data of each time period, and the characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and the typhoon wind circles at all levels or the typhoon center point, and the azimuth of the target reservoir with respect to the typhoon center point.

[0014] A variable extraction unit, configured to, for each time period in the first variable set, respectively correspond with the characteristic data n time periods before (where n is greater than or equal to 2) to obtain new variables, and form a second variable set.

[0015] A first screening unit, configured to screen variables with a maximum information coefficient greater than a first preset value with the historical inflow runoff information from the second variable set to obtain a third variable set.

[0016] A second screening unit, configured to screen the top L principal components according to the magnitudes of the characteristic data in the third variable set and remove duplicate data to obtain a fourth variable set.

[0017] A model training unit, configured to take the fourth variable set as the input factor set and the historical inflow runoff information as the output value, construct a runoff prediction model, and train and optimize the runoff prediction model.

[0018] A dynamic scheduling unit is configured to input real-time typhoon information into a runoff prediction model in response to receiving the real-time typhoon information, output predicted reservoir inflow information, and perform scheduling on a target reservoir based on the predicted reservoir inflow information and a preset scheduling strategy.

[0019] In a third aspect, an embodiment of the present invention further provides a reservoir dynamic scheduling device, including a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to execute the steps of any one of the reservoir dynamic scheduling methods provided by the embodiments of the present invention.

[0020] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium. The computer-readable storage medium includes a computer program, and when the computer program runs on an electronic device, the computer program is used to cause the electronic device to execute the steps of any one of the reservoir dynamic scheduling methods provided by the embodiments of the present invention.

[0021] The beneficial effects of the present invention are as follows:

[0022] The reservoir dynamic scheduling method of the present invention obtains historical typhoon information and corresponding historical reservoir inflow information of a target reservoir, forms a first variable set according to the historical typhoon information, and then for each time period in the first variable set, it corresponds to the characteristic data 1 to n time periods before to obtain new variables, forming a second variable set, and screening out the variables in the second variable set whose maximum information coefficient with the historical reservoir inflow information is greater than a first preset value to obtain a third variable set; further, according to the magnitudes of the characteristic data in the third variable set, screening out the top L principal components and removing duplicate data to obtain a fourth variable set, using the fourth variable set as an input factor set and the historical reservoir inflow information as an output value, constructing a runoff prediction model, training and optimizing the runoff prediction model, and finally when receiving real-time typhoon information, inputting the real-time typhoon information into the runoff prediction model, outputting predicted reservoir inflow information, and performing scheduling on the target reservoir based on the predicted reservoir inflow information and a preset scheduling strategy. Therefore, the present invention can quickly predict possible reservoir inflow information according to real-time typhoon information, and then perform reservoir dynamic scheduling based on the predicted reservoir inflow information, without having to lower the water level of the target reservoir below the flood limit water level throughout the flood season and further lower the reservoir water level before the typhoon arrives, nor relying on the subjective experience of schedulers. At the same time, it has prediction accuracy and a runoff prediction model optimization mechanism, taking into account the efficient utilization of water resources and a stable and accurate flood control effect. Description of the Drawings

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0024] Figure 1 is a schematic flowchart of the reservoir dynamic scheduling method provided in the embodiments of the present invention;

[0025] Figure 2 is a schematic diagram of the specific scheduling effect of the reservoir dynamic scheduling method provided in the embodiments of the present invention;

[0026] Figure 3 is a schematic structural diagram of the reservoir dynamic scheduling device provided in the embodiments of the present invention;

[0027] Figure 4 is a schematic structural diagram of the reservoir dynamic scheduling equipment provided in the embodiments of the present invention. Specific embodiments

[0028] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention. At the same time, in the description of the embodiments of the present invention, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more features. In the description of the embodiments of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0029] The embodiments of the present invention provide a reservoir dynamic scheduling method, device, equipment and storage medium.

[0030] Specifically, this embodiment will be described from the perspective of the reservoir dynamic scheduling device. This reservoir dynamic scheduling device can be specifically integrated in the reservoir dynamic scheduling equipment, that is, the reservoir dynamic scheduling method in the embodiments of the present invention can be executed by the reservoir dynamic scheduling equipment.

[0031] The following will be described in detail with reference to the accompanying drawings respectively. In this embodiment, the execution entity is the reservoir dynamic scheduling equipment as an example. It should be noted that the description order of the following embodiments does not limit the preferred order of the embodiments. Although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from that shown in the accompanying drawings.

[0032] In terms of water resources management technology, during the typhoon landing period, the existing reservoir operation technology faces the following challenges: A traditional and relatively conservative operation method is to keep the reservoir water level below the flood limit level throughout the flood season. Once a typhoon is predicted to come, the reservoir water level is further lowered to vacate the storage capacity. Although this operation method is cautious, it is easy to cause waste of water resources, reduce the power generation efficiency of the reservoir, and may also affect the water supply efficiency during the dry season. Relatively speaking, a more advanced and efficient reservoir operation method is dynamic control, that is, to keep the reservoir operating above the normal storage level during the flood season. When a typhoon is predicted to come, pre-discharge is carried out in a timely manner to lower the water level below the flood limit level. However, at present, this operation method still mainly relies on the prediction of typhoons and their rainfall. Once a typhoon or rainfall is predicted, the reservoir water level is quickly lowered below the flood limit level. This qualitative judgment control process lacks accuracy and optimization; in addition, this operation method needs to rely on the experience of the operation personnel to operate the reservoir during the typhoon rain period. Although it has a certain degree of flexibility, it is difficult to form a fixed operation procedure for subsequent reference, has strong subjectivity, and is greatly affected by the experience and level of the operator. Therefore, there is an urgent need for a reservoir dynamic operation method that can efficiently utilize water resources in the face of typhoons and achieve stable and accurate flood control effects.

[0033] To solve the above problems, the embodiment of the present invention discloses a reservoir dynamic operation method. This reservoir dynamic operation method obtains the historical typhoon information and the corresponding historical inflow runoff information of the target reservoir, forms a first variable set according to the historical typhoon information, and then for each time period in the first variable set, corresponds to the characteristic data n time periods ago to obtain new variables, forms a second variable set, and screens out the variables in the second variable set whose maximum information coefficient with the historical inflow runoff information is greater than the first preset value to obtain a third variable set; further, according to the magnitudes of the characteristic data of the third variable set, screens out the top L principal components and eliminates duplicate data to obtain a fourth variable set. The fourth variable set is used as the input factor set, and the historical inflow runoff information is used as the output value to construct a runoff prediction model, and the runoff prediction model is trained and optimized. Finally, when receiving real-time typhoon information, the real-time typhoon information is input into the runoff prediction model, the inflow runoff prediction information is output, and the target reservoir is operated based on the inflow runoff prediction information and the preset operation strategy. Thus, the embodiment of the present invention can quickly predict the possible inflow runoff information according to the real-time typhoon information, and then carry out reservoir dynamic operation according to the inflow runoff prediction information, without lowering the water level of the target reservoir below the flood limit level throughout the flood season and further lowering the reservoir water level before the typhoon comes, nor relying on the subjective experience of the operation personnel, and at the same time has prediction accuracy and a runoff prediction model optimization mechanism, taking into account the efficient utilization of water resources and stable and accurate flood control effects.

[0034] Please refer toFigure 1 , the specific process of the reservoir dynamic scheduling method can be as follows in steps S101 to S106, where:

[0035] Step S101, obtain the historical typhoon information and corresponding historical inflow runoff information of the target reservoir, and form a first variable set according to the historical typhoon information.

[0036] Among them, the target reservoir is the reservoir for which the reservoir dynamic scheduling method of the embodiment of the present invention performs dynamic scheduling. The target reservoir can be one or more.

[0037] It can be understood that the historical typhoon information and corresponding historical inflow runoff information of the target reservoir refer to the typhoon information that has occurred in history and affected the target reservoir, as well as the inflow runoff information of the target reservoir under the influence of this typhoon information. Specifically, the historical inflow runoff information includes time (year, month, day, hour) and inflow discharge, recorded as Y = y(t), where t represents time and y(t) represents the inflow discharge at time t. The historical typhoon information specifically includes the typhoon occurrence time (year, month, day, hour), moving path, intensity (such as air pressure, wind speed), etc. In the embodiment of the present invention, a first variable set composed of multiple variables is formed according to the historical typhoon information, recorded as X 1 ={X 1 , X 2 ,..., X n , X n+1}.

[0038] In the embodiment of the present invention, the first variable set includes the characteristic data of each time period. The characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and each typhoon wind circle or the typhoon center point, and the azimuth of the target reservoir with respect to the typhoon center point. It should be noted that the distance between the target reservoir and each typhoon wind circle includes multiple variables, that is, the distance between each wind circle and the target reservoir corresponds to one variable.

[0039] It should be noted that each variable in the first variable set is composed of a certain characteristic data of the historical typhoon information in each time period.

[0040] Step S102, for each time period in the first variable set, respectively correspond to the characteristic data of 1 to n time periods before to obtain new variables, and form a second variable set.

[0041] Among them, n is greater than or equal to 2.

[0042] Specifically, for each time period in the first variable set, it is respectively corresponded with the characteristic data of 1 to n time periods before to obtain new variables. That is, the characteristic data of time period t is corresponded with that of time period t - n to obtain new variables, and each newly formed variable is composed of a certain characteristic data of time period t - m corresponding to each time period t, where m = [1, n].

[0043] Exemplarily, in some embodiments, each time period is 1 hour. Assuming n = 8, the characteristic data of time period t is respectively corresponded with that of t - 1, t - 2, t - 3, t - 4, t - 5, t - 6, t - 7, and t - 8. For each variable, 8 new variables will be obtained in the above - mentioned manner. For example, the characteristic data of 9:00 is corresponded with that of 8:00, 7:00, 6:00, 5:00, 4:00, 3:00, 2:00, and 1:00 respectively.

[0044] Exemplarily, for the first variable set X 1 ={X 1 ,X 2 ,...,X n ,X n+1}, the characteristic data of each time period is respectively corresponded with that of 1 to 2 time periods before to obtain new variables, which form the second variable set.

[0045] S103. Screen out the variables in the second variable set whose maximum information coefficient with the historical stored - in - reservoir runoff information is greater than the first preset value to obtain the third variable set.

[0046] Specifically, in the embodiment of the present invention, the maximum information coefficient between each variable in the second variable set and the stored - in - reservoir runoff information is calculated, and based on the calculation results, the variables whose maximum information coefficient is greater than the first preset value are screened out from them to obtain the variables whose correlation with the stored - in - reservoir runoff information meets the requirements, forming the third variable set X 3 .

[0047] More specifically, in some embodiments, the maximum information coefficient between each variable in the second variable set and the stored - in - reservoir runoff information is calculated by the following formula:

[0048]

[0049] Among them, I * (D, x, y) represents the maximum mutual information found in the mutual information of all possible distributions D| G on the x×y grid; B(n) is the upper bound of the partition of the x×y grid, and ω(1) < B(n) ≤ ο(n 1-ε ) holds for any 0 < ε < 1.

[0050] S104. According to the magnitudes of the respective characteristic data of the third variable set, screen out the top L principal components and eliminate duplicate data to obtain a fourth variable set.

[0051] As can be seen from steps S102 and S103, each characteristic data in the third variable set may include at least one vector composed of data from several time periods before the corresponding time period t. Therefore, in step S104 of the reservoir dynamic scheduling method according to the embodiments of the present invention, the variables in the third variable set are further screened. For each characteristic data, according to the magnitude of the characteristic data, screen out the top L principal components and eliminate duplicate data. For example, if a certain characteristic data contains two variables composed of the data of time periods t - 8 and t - 9 corresponding to each time period t, then retain the variable composed of the data of time period t - 8 corresponding to each time period t. At the same time, since there may be duplicate data in some characteristic data, the embodiments of the present invention also eliminate duplicate data. Thus, in step S104 of the embodiments of the present invention, a fourth variable set X composed of factors with higher information content is obtained. 4 Where L can be any value, and the specific value is not limited here.

[0052] Optionally, in some embodiments, the process of eliminating duplicate data may include:

[0053] Standardize the third variable set using the principal component analysis method.

[0054] Specifically, in the embodiments of the present invention, the third variable set is standardized using the principal component analysis method, the correlation coefficient matrix between the column vectors (i.e., variables) is calculated, the eigenvalues and eigenvectors are obtained by solving the characteristic equation, and the top p principal components are selected according to the magnitudes of the eigenvalues to form the fourth variable set X. 4 .

[0055] It can be understood that the fourth variable set X 4 is a factor set that plays a key role in simulating the inflow of the target reservoir under the influence of typhoons, and is a high-information factor set further screened on the basis of the high-correlation factor set (the third variable set X 3 ), that is, a high-quality factor set.

[0056] Optionally, in some embodiments, the time distance between the variable of the characteristic data in the fourth variable set and the current time period is used as the lead time. It can be understood that since the maximum information coefficient between the characteristic data of a certain time period and the historical inflow runoff information after the lead time is greater than the first preset value, the real-time typhoon information of the current time period can be used subsequently to predict the inflow runoff information of the target reservoir after the lead time, improving the initiative and timeliness of flood control.

[0057] Furthermore, it can be understood that the embodiments of the present invention can further improve the prediction period of the target reservoir operation according to the typhoon forecast information, which plays an important role in protecting the safety of the target reservoir dam itself and the flood control downstream, and can further improve the initiative and timeliness of flood control.

[0058] S105. Use the fourth variable set as the input factor set, and use the historical inflow runoff information as the output value to construct a runoff prediction model, and train and optimize the runoff prediction model.

[0059] It can be understood that the variables in the fourth variable set are the key variables affecting the historical inflow runoff information of the target reservoir. In the embodiments of the present invention, the fourth variable set is used as the input of the runoff prediction model, and the historical inflow runoff information is used as the output of the runoff prediction model to train and optimize the runoff prediction model, so that the trained runoff prediction model can be used to predict the inflow runoff information of the target reservoir based on real-time typhoon information in the future.

[0060] Optionally, in some embodiments, the runoff prediction model is a fusion model of a Deep Belief Networks (DBN) based on deep learning and a Long Short-Term Memory (LSTM) network. Training and optimizing the runoff prediction model may include:

[0061] 1) Use the input factor set as the input, and train the deep belief network with the first model parameters.

[0062] 2) Input the input factor set into the trained deep belief network for feature extraction to obtain a sequence of feature vectors.

[0063] 3) Use the sequence of feature vectors as the input, and train the long short-term memory network with the second model parameters.

[0064] 4) Input the feature vectors into the trained long short-term memory network to obtain the historical inflow runoff prediction information.

[0065] 5) Calculate the prediction accuracy according to the historical inflow runoff prediction information and the historical inflow runoff information.

[0066] 6) If the prediction accuracy is less than the second preset value, adjust the first model parameters and the second model parameters, and return to the step of training the deep belief network with the first model parameters until the prediction accuracy is greater than the second preset value.

[0067] It can be understood that a deep belief network is a generative model composed of multiple restricted Boltzmann machines (RBMs) stacked on top of each other, with the top layer connected to a classifier (usually softmax regression). A long short-term memory network is a special type of recurrent neural network (RNN) designed to address the problem of vanishing or exploding gradients that traditional RNNs encounter when dealing with long-range dependencies.

[0068] Among them, the first model parameters can include restricted Boltzmann machine (RBM) parameters, hierarchical structure parameters, pre-training parameters, and fine-tuning parameters. Among them, the RBM parameters include the weights between the visible layer neurons and the hidden layer neurons that determine how the input data is transformed into a higher-level feature representation, the biases of the visible layer neurons that affect the activation probability of the visible layer neurons, and the biases of the hidden layer neurons that affect the activation probability of the hidden layer neurons; the hierarchical structure parameters include the number of RBM layers that determine the depth of feature extraction and the number of neurons in each RBM layer that affect the complexity and ability of feature extraction; the pre-training parameters include the number of iterations of contrastive divergence (CD) that affects the degree of parameter convergence and the quality of feature extraction, and the learning rate that determines the weight update step size; the fine-tuning parameters include the parameters of the classifier, such as the weights and biases of the fully connected layer, and the learning rate of fine-tuning. The second model parameters can include the number of LSTM layers and the number of neurons in each layer.

[0069] Furthermore, in some embodiments, using the first model parameters to train the deep belief network with the input factor set as the input may include:

[0070] Setting the deep belief network with the first model parameters; inputting the input factor set into the deep belief network, and performing unsupervised pre-training and supervised fine-tuning (layer-by-layer greedy training) in sequence.

[0071] Furthermore, in some embodiments, using the second model parameters to train the long short-term memory network with the feature vector sequence as the input may include:

[0072] Setting the long short-term memory network with the second model parameters; inputting the feature vector sequence into the long short-term memory network for network training.

[0073] It is understandable that for long sequence data (time series data) such as typhoon information, the deep belief network gradually learns the hierarchical feature representation of the data through multiple RBM layers. Each layer can learn different levels of abstraction of the data, so that the high-level layer can capture the complex patterns and long-term dependencies in time series data such as typhoon information, and can automatically discover and extract key features and nonlinear relationships in time series data. At the same time, the deep belief network is pre-trained through unsupervised learning, which means that it can learn the intrinsic structure of the data without labels, and is also suitable for time series data such as typhoon information, and saves the cost of obtaining labeled data and reduces the difficulty of data acquisition. In addition, the long short-term memory network can capture the long-term dependencies in time series data and maintain a stable gradient in a long sequence, so that the network can learn from long-term data, which is very important for predicting future runoff. At the same time, the long short-term memory network has a nonlinear activation function, which can also capture complex nonlinear relationships, so that it can accurately describe the nonlinear changes in runoff. The embodiment of the present invention adopts a fusion model of a deep belief network based on deep learning and a long short-term memory network as a runoff prediction model, which can more accurately extract the characteristic sequence of typhoon information and ultimately output the predicted value of the inflow runoff of the target reservoir.

[0074] S106, in response to receiving the real-time typhoon information, inputting the real-time typhoon information into the runoff prediction model, outputting the inflow runoff prediction information, and scheduling the target reservoir based on the inflow runoff prediction information and the preset scheduling strategy.

[0075] Among them, the preset scheduling strategy is a scheduling strategy pre-set for the target reservoir based on the inflow runoff prediction information.

[0076] Specifically, when receiving real-time typhoon information, such as observing typhoon characteristic data or obtaining typhoon characteristic data from an external network, the embodiment of the present invention inputs the real-time typhoon information into the runoff prediction model trained in step S105, and outputs the inflow runoff prediction information. Then, the embodiment of the present invention can dispatch the target reservoir based on the inflow runoff prediction information and the preset dispatching strategy.

[0077] Optionally, in some embodiments, before scheduling the target reservoir based on the inflow runoff prediction information and the preset scheduling strategy, the reservoir dynamic scheduling method may further include:

[0078] 1) In response to receiving demand information, an optimization scheduling objective function is established according to the demand information.

[0079] 2) Calculate the outbound route based on the optimization scheduling objective function and the inflow runoff prediction information.

[0080] 3) Generate a preset scheduling strategy based on the outbound route.

[0081] Among them, the demand information can be the specific demand for the target reservoir operation input by the user. When the embodiment of the present invention receives the demand information, it can establish a corresponding optimal operation objective function according to the demand information.

[0082] Exemplarily:

[0083] Objective 1: The flood control storage capacity occupied by the target reservoir is minimized.

[0084] The corresponding optimal operation objective function is:

[0085]

[0086] Among them, V 1,t represents the water storage of the target reservoir at time t, with the unit of m 3 .

[0087] Objective 2: The peak shaving rate of the downstream section is maximized.

[0088] The corresponding optimal operation objective function is:

[0089]

[0090] Among them, R’ 1,t is the flow rate at the downstream section evolved from the average inflow of the target reservoir at time t, with the unit of m 3 / s; Q’ 1,t is the flow rate at the downstream section evolved from the average outflow of the target reservoir at time t, with the unit of m 3 / s; q t is the flow rate of the interval between the target reservoir and the downstream, with the unit of m 3 / s; T represents the number of scheduling periods.

[0091] Thus, the embodiment of the present invention can calculate the outflow route according to the specific optimal operation objective function in combination with the inflow runoff prediction information, and then generate a preset scheduling strategy based on the outflow route. Among them, the preset scheduling strategy can include a series of reservoir optimization control strategies to support the formulation of the reservoir optimal operation mode.

[0092] Optionally, in some embodiments, after scheduling the target reservoir based on the inflow runoff prediction information and the preset scheduling strategy, the reservoir dynamic scheduling method may further include:

[0093] Regarding the real-time typhoon information as historical typhoon information, return to step S101.

[0094] That is, the embodiments of the present invention can, in real time, regard past typhoon information as historical typhoon information, and based on this historical typhoon information and the corresponding reservoir inflow information, repeat the target reservoir operation process of steps S101 - S106, thereby realizing the dynamic adjustment of the entire operation process, and further improving the accuracy and reliability of the prediction of the target reservoir inflow and the operation effect and quality of the target reservoir.

[0095] A specific embodiment of the present invention is as follows:

[0096] 1) Obtain the historical typhoon information and the corresponding historical reservoir inflow information of the target reservoir, and form a first variable set according to the historical typhoon information. The first variable set includes characteristic data for each time period, and the characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and the typhoon wind circles at all levels or the typhoon center point, and the azimuth of the target reservoir relative to the typhoon center point, as shown in Table 1.

[0097] Table 1 First variable set

[0098]

[0099] 2) For each time period in the first variable set, respectively correspond to the characteristic data 1 to n time periods before to obtain new variables, and form a second variable set, as shown in Table 2. Among them, n is greater than or equal to 2.

[0100] Table 2 Second variable set

[0101]

[0102] 3) Select the variables from the second variable set whose maximum information coefficient with the historical reservoir inflow information is greater than the first preset value to obtain a third variable set, as shown in Table 3.

[0103] Table 3 Third variable set

[0104]

[0105] 4) According to the magnitudes of the characteristic data in the third variable set, select the top L principal components and eliminate duplicate data to obtain a fourth variable set, as shown in Table 4.

[0106] Table 4 Fourth variable set

[0107]

[0108] 5) Take the fourth variable set as the input factor set and the historical reservoir inflow information as the output value, construct a fusion model of a deep belief network and a long short - term memory network based on deep learning, and train and optimize this fusion model.

[0109] 6) In response to receiving real-time typhoon information, input the real-time typhoon information into the runoff prediction model to output the predicted reservoir inflow information.

[0110] 7) Dispatch the target reservoir based on the predicted reservoir inflow information and the preset scheduling strategy. Specifically, with the goal of minimizing the downstream inundation loss, establish the corresponding optimized scheduling objective function, calculate the release route according to the optimized scheduling objective function and the predicted reservoir inflow information, and perform optimized scheduling on the reservoir to obtain the optimal release process of the reservoir, as Figure 2 shown. Among them, the average daily inflow is the average daily inflow calculated based on the predicted reservoir inflow information.

[0111] 8) Repeat the above steps to continuously generate the optimized reservoir scheduling plan in a rolling manner.

[0112] In summary, in the embodiment of the present invention, by obtaining the historical typhoon information and the corresponding historical reservoir inflow information of the target reservoir, forming the first variable set according to the historical typhoon information, and then for each time period in the first variable set, corresponding to the characteristic data 1 to n time periods before respectively to obtain new variables, forming the second variable set, and screening the variables in the second variable set whose maximum information coefficient with the historical reservoir inflow information is greater than the first preset value to obtain the third variable set; further, according to the magnitudes of the characteristic data of the third variable set, screening the top L principal components and removing duplicate data to obtain the fourth variable set, using the fourth variable set as the input factor set and the historical reservoir inflow information as the output value, constructing the runoff prediction model, training and optimizing the runoff prediction model, and finally when receiving real-time typhoon information, inputting the real-time typhoon information into the runoff prediction model to output the predicted reservoir inflow information, and dispatching the target reservoir based on the predicted reservoir inflow information and the preset scheduling strategy. Thus, the present invention can quickly predict the possible reservoir inflow information according to the real-time typhoon information, and then perform dynamic reservoir scheduling according to the predicted reservoir inflow information, without having to lower the water level of the target reservoir below the flood limit water level throughout the flood season and further lower the reservoir water level before the typhoon arrives, nor relying on the subjective experience of the dispatcher. At the same time, it has prediction accuracy and a runoff prediction model optimization mechanism, taking into account the efficient utilization of water resources and the stable and accurate flood control effect.

[0113] In the embodiments of the present invention, a correlation is established between typhoon characteristic parameters and reservoir inflow runoff. According to the real-time prediction results of typhoon information, possible inflow runoff with a certain confidence level is dynamically and iteratively generated. Based on the obtained inflow runoff, according to the current reservoir operation rules, the embodiments of the present invention simulate and predict the reservoir operation process to obtain the reservoir outflow discharge and water level change process, and can further obtain the flow process of the downstream flood control section, providing a reference for decision-makers to adjust the plan. Further, if the plan needs to be adjusted, the embodiments of the present invention can also optimize the operation of the target reservoir based on the current water level and operation target of the target reservoir to determine the optimized operation process of the target reservoir.

[0114] This embodiment also provides a reservoir dynamic operation device, which can be specifically integrated in the reservoir dynamic operation equipment. As Figure 3 shown, the reservoir dynamic operation device may include:

[0115] A historical data acquisition unit 301, configured to acquire the historical typhoon information and corresponding historical inflow runoff information of the target reservoir, and form a first variable set according to the historical typhoon information. The first variable set includes characteristic data for each time period, and the characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and the typhoon wind circles at all levels or the typhoon center point, and the azimuth of the target reservoir with respect to the typhoon center point;

[0116] A variable extraction unit 302, configured to, for each time period in the first variable set, respectively correspond to the characteristic data 1 to n time periods before to obtain new variables, and form a second variable set, where n is greater than or equal to 2;

[0117] A first screening unit 303, configured to screen out variables in the second variable set whose maximum information coefficient with the historical inflow runoff information is greater than a first preset value to obtain a third variable set;

[0118] A second screening unit 304, configured to screen out the top L principal components according to the magnitudes of the characteristic data in the third variable set and eliminate duplicate data to obtain a fourth variable set;

[0119] A model training unit 305, configured to use the fourth variable set as an input factor set and the historical inflow runoff information as an output value to construct a runoff prediction model, and train and optimize the runoff prediction model;

[0120] A dynamic operation unit 306 is configured to, upon receiving real-time typhoon information, input the real-time typhoon information into the runoff prediction model, output the inflow runoff prediction information, and operate the target reservoir based on the inflow runoff prediction information and a preset operation strategy.

[0121] As Figure 4 shown, Figure 4Schematic structural diagram of the reservoir dynamic scheduling device provided by the embodiment of the present invention. The reservoir dynamic scheduling device 1100 includes a processor 1101 having one or more processing cores, a memory 1102 having one or more computer-readable storage media, and a computer program stored on the memory 1102 and executable on the processor. Among them, the processor 1101 is electrically connected to the memory 1102. Those skilled in the art can understand that the structural diagram of the reservoir dynamic scheduling device shown in the figure does not constitute a limitation on the reservoir dynamic scheduling device, and it may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0122] The processor 1101 is the control center of the reservoir dynamic scheduling device 1100, connecting various parts of the entire reservoir dynamic scheduling device 1100 through various interfaces and lines. By running or loading software programs and / or units stored in the memory 1102, and calling data stored in the memory 1102, it executes various functions of the reservoir dynamic scheduling device 1100 and processes data, thereby monitoring the reservoir dynamic scheduling device 1100 as a whole. The processor 1101 can be a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), etc., and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention.

[0123] In the embodiment of the present invention, the processor 1101 in the reservoir dynamic scheduling device 1100 will load the instructions corresponding to the processes of one or more application programs into the memory 1102 according to the following steps, and the processor 1101 will run the application programs stored in the memory 1102 to implement various functions. For details, refer to the previous embodiments and will not be elaborated here.

[0124] Optionally, as Figure 4 shown, the reservoir dynamic scheduling device 1100 further includes: a touch display screen 1103, a radio frequency circuit 1104, an audio circuit 1105, an input unit 1106, and a power supply 1107. Among them, the processor 1101 is electrically connected to the touch display screen 1103, the radio frequency circuit 1104, the audio circuit 1105, the input unit 1106, and the power supply 1107 respectively. Those skilled in the art can understand that Figure 4 the structural diagram of the reservoir dynamic scheduling device shown does not constitute a limitation on the reservoir dynamic scheduling device, and it may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0125] The touch display screen 1103 can be used to display a graphical user interface and receive operation instructions generated by a user acting on the graphical user interface. The touch display screen 1103 may include a display panel and a touch panel. Among them, the display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the reservoir dynamic scheduling device. These graphical user interfaces can be composed of graphics, text, icons, videos, and any combination thereof. Optionally, the display panel can be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc. The touch panel can be used to collect touch operations of the user on or near it (such as operations of the user using any suitable object or accessory such as a finger or a stylus on or near the touch panel), and generate corresponding operation instructions, and the operation instructions execute the corresponding program. Optionally, the touch panel can include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the touch orientation of the user, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into contact coordinates, and then sends it to the processor 1101, and can receive and execute the commands sent by the processor 1101. The touch panel can cover the display panel. After the touch panel detects a touch operation on or near it, it is transmitted to the processor 1101 to determine the type of touch event. Subsequently, the processor 1101 provides a corresponding visual output on the display panel according to the type of touch event. In the embodiment of the present invention, the touch panel and the display panel can be integrated into the touch display screen 1103 to implement input and output functions. However, in some embodiments, the touch panel and the touch panel can be implemented as two independent components to implement input and output functions. That is, the touch display screen 1103 can also be used as a part of the input unit 1106 to implement the input function.

[0126] The radio frequency circuit 1104 can be used to transmit and receive radio frequency signals to establish wireless communication with a network device or other reservoir dynamic scheduling devices through wireless communication, and transmit and receive signals with the network device or other reservoir dynamic scheduling devices.

[0127] The audio circuit 1105 can be used to provide an audio interface between the user and the reservoir dynamic scheduling device through a speaker and a microphone. The audio circuit 1105 can transmit the electrical signal converted from the received audio data to the speaker, which converts it into a sound signal for output; on the other hand, the microphone converts the collected sound signal into an electrical signal, which is received by the audio circuit 1105 and then converted into audio data. After the audio data is output and processed by the processor 1101, it is sent through the radio frequency circuit 1104 to, for example, another reservoir dynamic scheduling device, or the audio data is output to the memory 1102 for further processing. The audio circuit 1105 may also include an earphone jack to provide communication between the peripheral earphone and the reservoir dynamic scheduling device.

[0128] The input unit 1106 can be used to receive input digital, character information or user feature information (such as fingerprint, iris, facial information, etc.), and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0129] The power supply 1107 is used to supply power to each component of the reservoir dynamic scheduling device 1100. Optionally, the power supply 1107 can be logically connected to the processor 1101 through a power management system, so as to implement functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 1107 may also include any components such as one or more DC or AC power supplies, a recharge system, a power failure detection circuit, a power converter or inverter, and a power status indicator.

[0130] In the above embodiments, the descriptions of the respective embodiments have their own focuses. For parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0131] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions or by controlling relevant hardware through instructions. The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0132] Therefore, an embodiment of the present invention provides a computer-readable storage medium, in which multiple computer programs are stored. The computer programs can be loaded by a processor to execute any one of the reservoir dynamic scheduling methods provided by the embodiments of the present invention. The computer programs can execute the steps of the aforementioned reservoir dynamic scheduling method. Reference can be made to the previous embodiments, and details are not repeated here.

[0133] Among them, the computer-readable storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), a magnetic disk or an optical disc, etc.

[0134] Since the computer program stored in the computer-readable storage medium can execute any one of the reservoir dynamic scheduling methods provided by the embodiments of the present invention, the beneficial effects achievable by any one of the reservoir dynamic scheduling methods provided by the embodiments of the present invention can be realized. For details, refer to the foregoing embodiments and will not be elaborated herein.

[0135] In the above embodiments of the reservoir dynamic scheduling device, computer-readable storage medium, reservoir dynamic scheduling equipment, and computer program product, the descriptions of the respective embodiments have their own focuses. For parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments. Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes and the beneficial effects brought by the above-described reservoir dynamic scheduling device, computer-readable storage medium, computer program product, reservoir dynamic scheduling equipment, and their corresponding units can refer to the description of the reservoir dynamic scheduling method in the above embodiments and will not be elaborated herein specifically.

[0136] The above has introduced in detail a reservoir dynamic scheduling method, a reservoir dynamic scheduling device, a reservoir dynamic scheduling equipment, a computer-readable storage medium, and a computer program product provided by the embodiments of the present invention. Specific examples are used herein to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those skilled in the art, based on the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A reservoir dynamic dispatching method, characterized in that: The method comprises: Acquire historical typhoon information and corresponding historical inflow runoff information of the target reservoir, and form a first variable set according to the historical typhoon information, wherein the first variable set includes characteristic data of each time period, and the characteristic data includes the intensity of the typhoon center point, the distance between the target reservoir and the typhoon wind circles of each level or the typhoon center point, and the direction of the target reservoir to the typhoon center point; For each time period in the first variable set, a new variable is obtained by corresponding to the feature data 1 to n time periods ago to form a second variable set, wherein n is greater than or equal to 2; A third variable set is obtained by screening the variables in the second variable set whose maximum information coefficient with the historical inflow runoff information is greater than a first preset value; According to the size of each characteristic data of the third variable set, the first L principal components are selected and duplicate data are removed to obtain a fourth variable set; Using the fourth variable set as an input factor set and the historical inflow runoff information as an output value, constructing a runoff prediction model, and training and optimizing the runoff prediction model; In response to receiving real-time typhoon information, the real-time typhoon information is input into the runoff prediction model, the inflow runoff prediction information is output, and the target reservoir is scheduled based on the inflow runoff prediction information and a preset scheduling strategy.

2. The reservoir dynamic dispatching method according to claim 1, characterized in that: Before dispatching the target reservoir based on the inflow runoff prediction information and the preset dispatching strategy, the method further includes: In response to receiving the demand information, establishing an optimized scheduling objective function according to the demand information; Calculate the outbound route according to the optimization scheduling objective function and the inbound runoff prediction information; The preset dispatching strategy is generated according to the outbound route.

3. The reservoir dynamic dispatching method according to claim 1, characterized in that: After scheduling the target reservoir based on the inflow runoff prediction information and the preset scheduling strategy, the method further includes: The step of using the real-time typhoon information as the historical typhoon information, returning the historical typhoon information of the target reservoir and the corresponding historical inflow runoff information, and forming a first variable set according to the historical typhoon information.

4. The reservoir dynamic dispatching method according to claim 1, characterized in that: The runoff prediction model is a fusion model of a deep belief network and a long short-term memory network based on deep learning; The training and optimization of the runoff prediction model includes: Taking the input factor set as input, training the deep belief network using first model parameters; Inputting the input factor set into the trained deep belief network for feature extraction to obtain a feature vector sequence; Taking the feature vector sequence as input, and training the long short-term memory network using second model parameters; Inputting the feature vector into the trained long short-term memory network to obtain historical inflow runoff prediction information; Calculating the prediction accuracy rate based on the historical inflow runoff prediction information and the historical inflow runoff information; If the prediction accuracy is less than the second preset value, the first model parameters and the second model parameters are adjusted, and the step of training the deep belief network using the first model parameters is returned to until the prediction accuracy is greater than the second preset value.

5. The reservoir dynamic dispatching method according to claim 4, characterized in that: The step of taking the input factor set as input and adopting the first model parameter to train the deep belief network comprises: Setting the deep belief network with the first model parameters; The input factor set is input into the deep belief network, and unsupervised pre-training and supervised fine-tuning are performed in sequence.

6. The reservoir dynamic dispatching method according to claim 4, characterized in that: The taking the feature vector sequence as input and using the second model parameter to train the long short-term memory network comprises: Setting the long short-term memory network with the second model parameters; The feature vector sequence is input into the long short-term memory network for network training.

7. The reservoir dynamic dispatching method according to claim 1, characterized in that: The removal of duplicate data includes: The third variable set is standardized using principal component analysis.

8. A reservoir dynamic dispatching device, characterized in that: The reservoir dynamic dispatching device comprises: A historical data acquisition unit, used to acquire historical typhoon information and corresponding historical inflow runoff information of a target reservoir, and form a first variable set according to the historical typhoon information, wherein the first variable set includes characteristic data of each time period, and the characteristic data includes the intensity of the typhoon center, the distance between the target reservoir and typhoon wind circles of various levels or the typhoon center, and the direction of the target reservoir to the typhoon center; A variable extraction unit, configured to obtain new variables by matching each time period in the first variable set with the feature data 1 to n time periods ago, to form a second variable set, wherein n is greater than or equal to 2; A first screening unit, configured to obtain a third variable set from the second variable set by screening the variables whose maximum information coefficient with the historical inflow runoff information is greater than a first preset value; A second screening unit is used to screen the top L principal components according to the size of each characteristic data of the third variable set, and remove duplicate data to obtain a fourth variable set; A model training unit, used to use the fourth variable set as an input factor set and the historical runoff information as an output value to construct a runoff prediction model, and to train and optimize the runoff prediction model; The dynamic scheduling unit is used to respond to receiving real-time typhoon information, input the real-time typhoon information into the runoff prediction model, output the inflow runoff prediction information, and schedule the target reservoir based on the inflow runoff prediction information and a preset scheduling strategy.

9. A reservoir dynamic dispatching device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the reservoir dynamic scheduling method described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a computer program. When the computer program is run on an electronic device, the computer program is used to enable the electronic device to execute the steps of the reservoir dynamic scheduling method according to any one of claims 1 to 7.

Images