Hydrological simulation method and system for simulating hydrological processes of a water body of a hydropower plant

Abstract

The present application provides a hydrological simulation method and system for simulating the hydrological process of the water body of a hydropower station, and relates to the field of data processing. The method comprises the following steps: determining a plurality of inflow influencing factors and a plurality of inflow influencing factors corresponding to the best inflow influencing time lag, a plurality of outflow influencing factors and a plurality of outflow influencing factors corresponding to the best outflow influencing time lag, establishing a hydrological simulator, wherein the residual structure is used to introduce the input cell state and the output cell state of the previous time step, the cell state update layer is used to update the cell state according to the output of the residual structure and the output of the input gate, and the linear layer is used to predict the water level according to the output of the output gate; training the hydrological simulator; obtaining the current input data of the hydropower station to be simulated; and predicting the water level of the hydropower station to be simulated by using the trained hydrological simulator, thereby improving the simulation accuracy of the hydrological process of the water body of the hydropower station.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, and in particular to a hydrological simulation method and system for simulating hydrological processes of a water body of a hydropower station. BACKGROUND

[0002] It is necessary to develop simulation scheduling of a hydropower station reservoir (group) by means of hydrological simulation to give an approximate scientific model of complex hydrological phenomena, processes and external limiting factors, which not only has important theoretical significance, but also has greater economic value. Through this way, economic and reasonable scheduling simulation of runoff processes is carried out, and scheduling feedback optimization of simulation scheduling results can fully utilize water energy resources and water resources, increase power generation of the hydropower station, improve comprehensive utilization efficiency of the hydropower station reservoir, and provide strong support for preventing and responding to water resource disasters under the premise of meeting engineering and production safety.

[0003] In recent years, research on hydrological simulators is mainly based on unit flow characteristic curves and discharge curves rated when a hydropower station is built, and then the outflow time of the upstream reservoir is evaluated by using experience, finally the unit load and gate opening are converted into inflow and outflow, and then the downstream water level change is deduced. This method ignores the complex relationship and time-varying characteristics of the nonlinearity of each control variable of the hydropower station, and the error caused by introducing various curves in the whole simulation process is not accurate.

[0004] Therefore, it is necessary to provide a hydrological simulation method and system for simulating hydrological processes of a water body of a hydropower station to improve the simulation accuracy of the hydrological processes of the water body of the hydropower station. SUMMARY

[0005] The application provides a hydrological simulation method for simulating the hydrological process of a water body of a hydropower station, comprising the following steps: determining a plurality of inflow influencing factors and corresponding inflow influence optimal time lags of the plurality of inflow influencing factors; determining a plurality of outflow influencing factors and corresponding outflow influence optimal time lags of the plurality of outflow influencing factors; establishing a hydrological simulator based on the plurality of inflow influencing factors, the corresponding inflow influence optimal time lags of the plurality of inflow influencing factors, and the plurality of outflow influencing factors and the corresponding outflow influence optimal time lags of the plurality of outflow influencing factors, wherein the hydrological simulator at least comprises a residual structure, an input gate, a cell state update layer, an output gate and a linear layer, the residual structure is used to introduce the input cell state and the output cell state of the previous time step, the input gate is used to calculate the degree of new information accepted and generate a new candidate memory state, the cell state update layer is used to update the cell state according to the output of the residual structure and the output of the input gate, the output gate is used to determine the information in the output cell state of the current time step output by the cell state update layer that needs to be output to the hidden state, and the linear layer is used to perform water level prediction according to the output of the output gate; a plurality of training samples are established; the hydrological simulator is trained through the plurality of training samples to generate a trained hydrological simulator; current input data of a hydropower station to be simulated is obtained based on the plurality of inflow influencing factors, the corresponding inflow influence optimal time lags of the plurality of inflow influencing factors, and the plurality of outflow influencing factors and the corresponding outflow influence optimal time lags of the plurality of outflow influencing factors; and the water level of the hydropower station to be simulated is predicted according to the current input data of the hydropower station to be simulated by the trained hydrological simulator.

[0006] Further, the plurality of inflow influencing factors at least comprise unit load of an upstream hydropower station, gate opening, water level in front of the dam, water level behind the dam and a plurality of environmental factors.

[0007] Further, the corresponding inflow influence optimal time lags of the plurality of inflow influencing factors are determined, comprising: determining first correlation coefficients of the plurality of inflow influencing factors and water level corresponding to a plurality of first candidate time lags; and screening the inflow influence optimal time lags corresponding to the plurality of inflow influencing factors from the plurality of first candidate time lags based on the first correlation coefficients of the plurality of inflow influencing factors and water level corresponding to the plurality of first candidate time lags.

[0008] Further, the plurality of outflow influencing factors at least comprise power generation load plan and gate opening plan of the current hydropower station.

[0009] Further, the determining of the outflow influence optimal time lag corresponding to the plurality of outflow influence factors comprises: determining second correlation coefficients of the plurality of outflow influence factors and the water level corresponding to a plurality of second candidate time lags; and screening the outflow influence optimal time lag corresponding to the plurality of outflow influence factors from the plurality of second candidate time lags based on the second correlation coefficients of the plurality of outflow influence factors and the water level corresponding to the plurality of second candidate time lags.

[0010] Further, the residual structure introduces the input cell state and the output cell state of the previous time step based on the following formula: wherein, is the output of the residual structure, is the output cell state of the previous time step, is the input cell state of the previous time step.

[0011] Further, the cell state update layer updates the cell state based on the output of the residual structure and the output of the input gate according to the following formula: wherein, is the output cell state of the current time step, is a value between 0 and 1 of the input gate output, indicating the degree to which new information is accepted, is a new candidate memory state of the input gate output, is an element-wise multiplication.

[0012] Further, the output gate determines which information in the output cell state of the current time step output by the cell state update layer needs to be output to the hidden state based on the following formula: , wherein, is a value between 0 and 1 of the output of the output gate, used to determine which information in the output cell state of the current time step output by the cell state update layer needs to be output to the hidden state, is a weight matrix of the output gate, is the hidden state of the previous time step, is a feature matrix, is a bias term of the output gate, is a sigmoid function, is a tangent value of the output cell state of the current time step, is a tangent function, is the hidden state of the current time step.

[0013] Further, the linear layer performs water level prediction based on the output of the output gate according to the following formula: wherein, is the predicted water level output by the linear layer, For the weights of the linear layer, This is the bias for the linear layer.

[0014] This invention provides a hydrological simulation system for simulating hydrological processes in a hydropower station, comprising: a parameter determination module for determining multiple inflow influencing factors and the optimal time lag of the inflow influence corresponding to the multiple inflow influencing factors; and for determining multiple outflow influencing factors and the optimal time lag of the outflow influence corresponding to the multiple outflow influencing factors; and a model building module for building a hydrological simulator based on the multiple inflow influencing factors and the optimal time lag of the inflow influence corresponding to the multiple inflow influencing factors and the optimal time lag of the outflow influence corresponding to the multiple outflow influencing factors, wherein the hydrological simulator includes at least a residual structure, an input gate, a cell state update layer, an output gate, and a linear layer. The residual structure is used to introduce the input cell state and output cell state of the previous time step. The input gate is used to calculate the degree of acceptance of new information and generate new candidate memory states. The cell state update layer is used to calculate the degree of acceptance of new information and generate new candidate memory states. A new layer is used to update the cell state based on the output of the residual structure and the output of the input gate. The output gate is used to determine which information in the cell state output at the current time step of the cell state update layer needs to be output to the hidden state. The linear layer is used to predict the water level based on the output of the output gate. A sample establishment module is used to establish multiple training samples. A model training module is used to train the hydrological simulator using the multiple training samples to generate a trained hydrological simulator. A data acquisition module is used to acquire the current input data of the hydropower station to be simulated based on the multiple inflow influencing factors and the corresponding optimal time lag of the inflow influencing factors and the multiple outflow influencing factors and the corresponding optimal time lag of the outflow influencing factors. A hydrological simulation module is used to predict the water level of the hydropower station to be simulated based on the current input data of the hydropower station to be simulated using the trained hydrological simulator.

[0015] Compared with existing technologies, the hydrological simulation method and system for simulating hydrological processes in hydropower station water bodies provided by this invention have at least the following beneficial effects:

[0016] The hydrological simulator can capture complex nonlinear and long-term dependencies in time series data by introducing complex neural network structures such as residual structures, input gates, cell state update layers, output gates, and linear layers. This enables it to more accurately simulate and predict water level changes, especially for complex systems influenced by multiple inflow and outflow factors. Through training, the hydrological simulator can learn the specific effects of different inflow and outflow factors on water level changes and their interactions. This makes the model more adaptable to water level predictions under different conditions, improving its generalization ability. The training process is based on a large amount of sample data, automatically learning and optimizing model parameters, reducing errors caused by human intervention and subjective judgment. With the continuous addition of new data, the model can also perform online learning and updating, further improving prediction performance. Using real-time or recent inflow and outflow data, the model can quickly predict future water level changes, providing timely and accurate references for the management and scheduling of hydropower stations. This helps hydropower stations develop more scientific operation plans and improve water energy utilization efficiency. Accurate water level prediction helps hydropower stations detect potential water level abnormalities or dangerous situations in a timely manner and take appropriate measures to prevent or respond, thereby reducing operational risks and ensuring the safe and stable operation of hydropower stations. By considering multiple inflow and outflow factors, the model can provide more comprehensive and in-depth analysis of water resources. This helps achieve optimal allocation and rational use of water resources, improve water resource utilization efficiency, and promote sustainable development. BRIEF DESCRIPTION OF DRAWINGS

[0017] The present specification will be further illustrated in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same numbers represent the same structures, wherein:

[0018] Figure 1 is a flowchart of a hydrological simulation method for simulating the hydrological process of a hydropower station water body according to some embodiments of the present specification;

[0019] Figure 2 is a schematic diagram of the calculation process of a hydrological simulator according to some embodiments of the present specification;

[0020] Figure 3 is a schematic diagram of the correlation coefficient of inflow factors and water level according to some embodiments of the present specification;

[0021] Figure 4 is a water level and main influencing factor relationship diagram according to some embodiments of the present specification;

[0022] Figure 5is a structural schematic diagram of a hydrological simulator according to some embodiments of the present specification;

[0023] Figure 6 is a schematic diagram of a water level simulation error according to some embodiments of the present specification;

[0024] Figure 7 is a module schematic diagram of a hydrological simulation system for simulating a hydrological process of a water body of a hydropower station according to some embodiments of the present specification. DETAILED DESCRIPTION

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, without paying creative labor, the present specification can also be applied to other similar scenarios according to these drawings. Unless it is obvious from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structure or operation.

[0026] Figure 1 is a flow schematic diagram of a hydrological simulation method for simulating a hydrological process of a water body of a hydropower station according to some embodiments of the present specification, as Figure 1 shown, the hydrological simulation method for simulating a hydrological process of a water body of a hydropower station can include the following flow.

[0027] Step 110, determining a plurality of reservoir inflow influencing factors and a plurality of reservoir inflow influencing factors corresponding to the reservoir inflow influencing optimal time lag.

[0028] The core work of hydrological simulation is to deduce the change of water level in front of the dam of the hydropower station according to the change of the unit of the hydropower station, the gate and the meteorological factor, and its essence is to determine the main factors affecting the water level change and the relationship between each factor and the simulation target; Therefore, it is necessary to first determine the physical model relied on by the water level simulation, to explore the core influencing factors, and to determine the input quantity of the hydrological simulation through the analysis, specifically, given the initial water level , the final water level , the mathematical expression formula is The main factors affecting the future water level are conducted through two variables of the current water level and the future water level change, and based on this, the input of the hydrological simulation is specifically constructed:

[0029] (1) Current water level

[0030] Current water level represents the latest observation value of the water level, but due to the influence of water waves, water level meter installation position and other factors, the real-time water level measurement value cannot represent the true water level value. The observation value in a period of time after noise reduction processing can better represent the true value of the water level. Therefore, the historical water level in the past period of time is introduced as the input of the hydrological simulation. represents the historical length

[0031] (2) water level change value

[0032] The direct factors affecting the water level change value are the inflow and outflow, but since the inflow and outflow data are derived from the load and gate opening combined with the unit flow characteristic (NHQ) curve and the gate discharge curve, errors in the NHQ curve and the discharge curve are introduced. In the Dadu River Basin, the power generation load value, water level and gate opening value are observation data, which have lower systematic errors than the inflow and outflow. Therefore, the following main influencing factors of inflow and outflow are introduced.

[0033] (3) main influencing factors of inflow

[0034] The inflow is mainly affected by the upstream power station generation load and the upstream power station gate opening. Since the upstream dam front water level and the upstream dam back water level imply the movement relationship of the upstream inflow and outflow, combined with the range of the upstream power station flow time, the historical unit load, historical gate opening, historical dam front and dam back water level of the upstream power station are introduced as the input of the simulator. At the same time, considering the natural runoff of the basin, the regional weather forecast data is introduced as the input of the hydrological simulation.

[0035] As a preferred, the multiple inflow influencing factors at least include the unit load, gate opening, dam front water level, dam back water level of the upstream hydropower station and multiple environmental factors, wherein the environmental factors can at least include environmental temperature, environmental wind direction, environmental wind speed, etc.

[0036] As a preferred, the inflow best time lag corresponding to the multiple inflow influencing factors is determined, including:

[0037] The first correlation coefficient of the multiple inflow influencing factors corresponding to multiple first candidate time lags and the water level is determined, for example, the multiple first candidate time lags can include 5 minutes ahead, 10 minutes ahead, 30 minutes ahead, 1 hour ahead, 5 minutes behind, 10 minutes behind, 30 minutes behind, 1 hour behind, etc.

[0038] Based on the first correlation coefficient of the multiple inflow influencing factors corresponding to multiple first candidate time lags and the water level, the inflow best time lag corresponding to the multiple inflow influencing factors is screened from the multiple first candidate time lags.

[0039] Specifically, for each first candidate time lag, the first correlation coefficient between each inflow influencing factor and the water level can be calculated using the following formula:

[0040]

[0041] in, For the corresponding number The first correlation coefficient between the e-th inflow influencing factor of the first candidate time lag and the water level. For the corresponding number The first candidate time lag, the e-th inflow influencing factor, at time t, decreases by the... The value of the first candidate time delay point, the .... The first candidate time delay indicates synchronization, the second... The first candidate delay < 0 indicates lead, and the second... A first candidate delay > 0 indicates a lag. The total number of selected time points. Let be the water level at time t.

[0042] Figure 3 This is a schematic diagram illustrating the correlation coefficients between inflow influencing factors and water level, based on some embodiments of this specification. Figure 3 As shown, the correlation coefficient between the same inflow rate and water level varies under different lag periods.

[0043] For each first candidate time lag, the matching value corresponding to that first candidate time lag can be calculated based on the first correlation coefficient between each inflow influencing factor and the water level using the following formula:

[0044]

[0045] in, For the first The matching value corresponding to the first candidate time delay The total number of factors affecting inbound flow. The weight is the weight corresponding to the e-th inflow factor.

[0046] The first candidate time delay with the largest matching value can be used as the optimal time delay for the inbound impact of multiple inbound flow influencing factors.

[0047] Step 120: Determine multiple outbound flow influencing factors and the optimal time lag of outbound impact corresponding to these factors.

[0048] Preferably, the multiple factors influencing outflow include at least the current hydropower station's planned power generation load and gate opening schedule. Furthermore, considering the potential scheduling patterns that historical power generation loads and gate openings may imply, historical loads and gate openings are introduced as inputs for the hydrological simulation.

[0049] Preferably, the optimal time lag for the outbound flow impact of multiple influencing factors is determined, including:

[0050] Determine the second correlation coefficient between multiple outflow influencing factors and water level corresponding to multiple second candidate time delays;

[0051] Based on the second correlation coefficient between multiple outflow influencing factors and water level corresponding to multiple second candidate time delays, the optimal time delay for outflow influencing factors corresponding to multiple outflow influencing factors is selected from multiple second candidate time delays.

[0052] The method for determining the optimal time lag of the outbound impact corresponding to multiple outbound flow influencing factors is similar to the method for determining the optimal time lag of the inbound impact corresponding to multiple inbound flow influencing factors, and will not be elaborated here.

[0053] Step 130: Based on multiple inflow influencing factors and the optimal time lag of the inflow impact corresponding to the multiple inflow influencing factors, and multiple outflow influencing factors and the optimal time lag of the outflow impact corresponding to the multiple outflow influencing factors, establish a hydrological simulator.

[0054] Based on multiple inflow influencing factors and their corresponding optimal time lags for inflow and outflow influencing factors, step 130 will examine the relationship between these influencing factors and water level to establish a hydrological simulator. Water level simulation requires extrapolating the downstream dam water level based on upstream and downstream factors affecting water level. The relationship between these factors and the downstream dam water level is complex, exhibiting the following characteristics:

[0055] (1) Nonlinear

[0056] Figure 4 This is a diagram illustrating the relationship between water level and key influencing factors, based on some embodiments shown in this specification. Figure 4 As shown, the future water level changes of downstream hydropower stations are generally inversely related to the future load changes, that is, as the future load increases, the water level tends to decrease, but the relationship between the two is not linear. The current future water level changes of hydropower stations are positively correlated with the load changes of upstream hydropower stations, that is, as the power generation load of upstream hydropower stations increases, the water level of downstream hydropower stations tends to increase, but the overall relationship is not linear.

[0057] (2) Long memory

[0058] Due to the influence of upstream flow arrival time, the influence of the discharge of the upstream hydropower station needs a certain time to affect the water level of the downstream hydropower station. From the time sequence, there is a clear time difference correlation between the water level of the downstream hydropower station and the influencing factors, that is, long memory.

[0059] Based on the above research conclusion, the hydrological simulator selects the long short-term memory (LSTM) model, and continues to optimize and improve the model level according to the characteristics of the basin water regime and the characteristics of the model. First, the flow arrival time is affected by the discharge of the upstream hydropower station. The larger the discharge of the upstream hydropower station, the smaller the flow arrival time. The smaller the discharge of the upstream hydropower station, the larger the flow arrival time. That is, the flow arrival time has obvious time-varying characteristics. In the modeling stage of long short-term memory network, the attention mechanism is introduced, which can automatically adjust the information of different time dimensions according to the characteristics of the upstream hydropower station.

[0060] Second, considering that the long short-term memory network itself has more gates and complex operations, the long short-term memory network is improved. The hydrological simulator recorded in the specification book uses residual instead of forget gate to improve the model running speed while retaining the long memory characteristics of the model.

[0061] Figure 2 is a schematic diagram of the calculation process of the hydrological simulator according to some embodiments of the present specification, Figure 5 is a structural schematic diagram of the hydrological simulator according to some embodiments of the present specification, Figure 5 are input data, are output data, are weights, such as Figure 2 and Figure 5 , wherein the hydrological simulator at least includes a residual structure, an input gate, a cell state update layer, an output gate and a linear layer. The residual structure is used to introduce the input cell state and the output cell state of the previous time step. The input gate is used to calculate the degree of acceptance of new information and generate a new candidate memory state. The cell state update layer is used to update the cell state according to the output of the residual structure and the output of the input gate. The output gate is used to determine the information in the output cell state of the current time step output by the cell state update layer that needs to be output to the hidden state. The linear layer is used to perform water level prediction according to the output of the output gate.

[0062] As a preferred, the residual structure introduces the input cell state and the output cell state of the previous time step based on the following formula:

[0063]

[0064] wherein, ​is the output of the residual structure, is the output cell state of the previous time step, is the input cell state of the previous time step.

[0065] The input gate decides which new information needs to be added to the cell state. It is divided into two parts: one part is a sigmoid layer that decides which information needs to be updated, and the other part is a tanh layer that generates a new candidate memory state. The formula is:

[0066]

[0067] wherein is the output of the input gate, a value between 0 and 1, indicating the degree to which new information is accepted. is the weight matrix of the input gate. is the new candidate memory state. is the weight matrix of the candidate memory state. and are the bias terms of the input gate and the candidate memory state, respectively. is the sigmoid function that maps the input to between 0 and 1. tanh is the hyperbolic tangent function. represents the feature matrix. represents the output hidden layer state of the previous time step.

[0068] As preferred, the cell state update layer updates the cell state based on the output of the residual structure and the output of the input gate according to the following formula:

[0069]

[0070] wherein is the output cell state of the current time step, is a value between 0 and 1 output by the input gate, indicating the degree to which new information is accepted, is the new candidate memory state output by the input gate, is an element-wise multiplication.

[0071] As preferred, the output gate decides which information in the output cell state of the current time step output by the cell state update layer needs to be output to the hidden state according to the following formula:

[0072]

[0073] wherein is the output of the output gate, a value between 0 and 1, used to decide which information in the output cell state of the current time step output by the cell state update layer needs to be output to the hidden state, is the weight matrix of the output gate, is the hidden state of the previous time step, is the feature matrix, is the bias term of the output gate, is the sigmoid function, is the tangent value of the output cell state of the current time step, is the tangent function, is the hidden state of the current time step.

[0074] As a preferred, the linear layer makes the water level prediction according to the output of the output gate based on the following formula:

[0075]

[0076] wherein, is the predicted water level of the linear layer output, is the weight of the linear layer, is the bias of the linear layer.

[0077] Step 140, a plurality of training samples are established.

[0078] Step 150, the hydrological simulator is trained through the plurality of training samples, and a trained hydrological simulator is generated.

[0079] The loss function for training the hydrological simulator is:

[0080]

[0081] wherein, is the mean square error, is the total number of training samples, is the prediction result of the i-th training sample, is the label of the i-th training sample.

[0082] The weight matrix (such as ) and the bias term (such as ) in the hydrological simulator are obtained through learning.

[0083] Step 160, based on the plurality of inflow influencing factors and the corresponding inflow influencing best time lag of the plurality of inflow influencing factors, and the plurality of outflow influencing factors and the corresponding outflow influencing best time lag of the plurality of outflow influencing factors, the current input data of the to-be-simulated hydropower station is obtained.

[0084] Only as an example, the current input data of the to-be-simulated hydropower station at least includes the upstream power station feature data determined based on the plurality of inflow influencing factors and the plurality of inflow influencing factors as shown in Table 1.

[0085] Table 1

[0086]

[0087] At step 170, the water level of the to-be-simulated hydropower station is predicted according to the current input data of the to-be-simulated hydropower station by the trained hydrological simulator.

[0088] The technical effects of the hydrological simulation method for simulating the hydrological process of the water body of the hydropower station are described below in combination with experiments.

[0089] The water level of the hydropower station A in the dry season is simulated by using the above water level simulator, and a 90-minute future simulation is performed every 5 minutes. Under the same boundary conditions, the simulation model based on the NHQ curve is compared, and the experimental results are shown in Table 2 and Figure 6 As can be seen from the experimental results, the water level simulator has high accuracy, and the relative true water level error of the 90-minute future simulation is only 0.094m, which is much lower than the simulation error based on the NHQ curve.

[0090] Table 2

[0091]

[0092] Figure 7 The hydrological simulation system for simulating the hydrological process of the water body of the hydropower station according to some embodiments of the present specification is shown in the module schematic diagram as Figure 7 As shown in the module schematic diagram, the hydrological simulation system for simulating the hydrological process of the water body of the hydropower station can include a parameter determination module, a model establishment module, a sample establishment module, a model training module, a data acquisition module, and a hydrological simulation module.

[0093] The parameter determination module can be used to determine a plurality of reservoir inflow influencing factors and a plurality of reservoir inflow influencing factors corresponding to the best inflow influencing time lag; and to determine a plurality of reservoir outflow influencing factors and a plurality of reservoir outflow influencing factors corresponding to the best outflow influencing time lag.

[0094] The model establishment module can be used to establish a hydrological simulator based on a plurality of reservoir inflow influencing factors and a plurality of reservoir inflow influencing factors corresponding to the best inflow influencing time lag and a plurality of reservoir outflow influencing factors and a plurality of reservoir outflow influencing factors corresponding to the best outflow influencing time lag, wherein the hydrological simulator at least includes a residual structure, an input gate, a cell state update layer, an output gate, and a linear layer. The residual structure is used to introduce the input cell state and the output cell state of the previous time step. The input gate is used to calculate the degree to which new information is accepted and generate a new candidate memory state. The cell state update layer is used to update the cell state according to the output of the residual structure and the output of the input gate. The output gate is used to determine which information in the output cell state of the current time step output by the cell state update layer needs to be output to the hidden state. The linear layer is used to perform water level prediction according to the output of the output gate.

[0095] The sample establishing module can be configured to establish a plurality of training samples.

[0096] The model training module can be configured to train the hydrological simulator by the plurality of training samples, to generate a trained hydrological simulator.

[0097] The data obtaining module can be configured to obtain current input data of the to-be-simulated hydropower station based on the plurality of inflow influencing factors, the inflow influencing optimal time lags corresponding to the plurality of inflow influencing factors, the plurality of outflow influencing factors, and the outflow influencing optimal time lags corresponding to the plurality of outflow influencing factors.

[0098] The hydrological simulation module can be configured to predict the water level of the to-be-simulated hydropower station by the trained hydrological simulator according to the current input data of the to-be-simulated hydropower station.

[0099] The hydrological simulation system for simulating the hydrological process of the water body of the hydropower station can be configured to execute the hydrological simulation method for simulating the hydrological process of the water body of the hydropower station. More description of the hydrological simulation system for simulating the hydrological process of the water body of the hydropower station can be referred to the description of the hydrological simulation method for simulating the hydrological process of the water body of the hydropower station, which will not be repeated here.

[0100] Finally, it should be understood that the embodiments described in the specification are only used to illustrate the principles of the embodiments of the specification. Other variations can also belong to the scope of the specification. Therefore, as an example but not limitation, alternative configurations of the embodiments of the specification can be considered consistent with the teachings of the specification. Accordingly, the embodiments of the specification are not limited to the embodiments explicitly introduced and described in the specification.

Claims

1. A hydrological simulation method for simulating a hydrological process of a water body of a hydropower station, characterized in that, The method comprises the following steps: determining a plurality of inflow influencing factors and corresponding inflow optimal time lags of the plurality of inflow influencing factors; determining a plurality of outflow influencing factors and corresponding outflow optimal time lags of the plurality of outflow influencing factors; establishing a hydrological simulator based on the plurality of inflow influencing factors, the corresponding inflow optimal time lags of the plurality of inflow influencing factors, the plurality of outflow influencing factors and the corresponding outflow optimal time lags of the plurality of outflow influencing factors, wherein the hydrological simulator at least comprises a residual structure, an input gate, a cell state update layer, an output gate and a linear layer, the residual structure is used to introduce the input cell state and the output cell state of the previous time step, the input gate is used to calculate the degree of new information accepted and generate a new candidate memory state, the cell state update layer is used to update the cell state according to the output of the residual structure and the output of the input gate, the output gate is used to determine the information in the output cell state of the current time step output by the cell state update layer which needs to be output to the hidden state, and the linear layer is used to predict the water level according to the output of the output gate; establishing a plurality of training samples; training the hydrological simulator through the plurality of training samples to generate a trained hydrological simulator; obtaining current input data of a to-be-simulated hydropower station based on the plurality of inflow influencing factors, the corresponding inflow optimal time lags of the plurality of inflow influencing factors, the plurality of outflow influencing factors and the corresponding outflow optimal time lags of the plurality of outflow influencing factors; predicting the water level of the to-be-simulated hydropower station according to the current input data of the to-be-simulated hydropower station through the trained hydrological simulator.

2. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower plant according to claim 1, characterized in that, The plurality of inflow influencing factors at least comprises the unit load of an upstream hydropower station, the gate opening, the water level in front of the dam, the water level behind the dam and a plurality of environmental factors.

3. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower station according to claim 1, characterized in that, The determination of the corresponding inflow optimal time lags of the plurality of inflow influencing factors comprises: determining first correlation coefficients between the plurality of inflow influencing factors corresponding to a plurality of first candidate time lags and the water level; screening the corresponding inflow optimal time lags of the plurality of inflow influencing factors from the plurality of first candidate time lags based on the first correlation coefficients between the plurality of inflow influencing factors corresponding to the plurality of first candidate time lags and the water level.

4. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower plant according to claim 1, characterized in that, The plurality of outflow influencing factors at least comprises the power generation load plan and the gate opening plan of the current hydropower station.

5. The hydrologic simulation method for simulating hydrologic processes of a reservoir of claim 1, wherein, The determination of the corresponding outflow optimal time lags of the plurality of outflow influencing factors comprises: determining second correlation coefficients between the plurality of outflow influencing factors corresponding to a plurality of second candidate time lags and the water level; screening the corresponding outflow optimal time lags of the plurality of outflow influencing factors from the plurality of second candidate time lags based on the second correlation coefficients between the plurality of outflow influencing factors corresponding to the plurality of second candidate time lags and the water level.

6. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower station according to any one of claims 1-5, characterized in that, The residual structure introduces the input cell state and the output cell state of the previous time step based on the following formula: wherein, is the output of the residual structure, is the output cell state of the previous time step, is the input cell state of the previous time step.

7. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower plant according to claim 6, characterized in that, The cell state update layer updates the cell state according to the output of the residual structure and the output of the input gate based on the following formula: where, is the output cell state for the current time step, is a value between 0 and 1 output by the input gate, indicating the degree to which new information is accepted, is a new candidate memory state output by the input gate, is an element-wise multiplication.

8. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower plant according to claim 7, characterized in that, The output gate determines which information in the output cell state of the current time step output by the cell state update layer needs to be output into the hidden state based on the following formula: wherein, is an output of the output gate, a value between 0 and 1, used to determine which information in the output cell state of the current time step of the cell state update layer needs to be output to the hidden state, is a weight matrix of the output gate, is a hidden state of the previous time step, is a feature matrix, is a bias term of the output gate, is a sigmoid function, is a tangent value of the output cell state of the current time step, is a tangent function, is a hidden state of the current time step.

9. The hydrological simulation method for simulating hydrological processes of a water body of a hydropower plant according to claim 8, characterized in that, The linear layer performs water level prediction according to the output of the output gate based on the following formula: wherein, is the predicted water level output by the linear layer, is the weight of the linear layer, is the bias of the linear layer.

10. A hydrological simulation system for simulating hydrological processes of a water body of a hydropower plant, characterized in that The method comprises the following steps: The parameter determination module is configured to determine a plurality of inflow influencing factors and corresponding inflow optimal time lags of the plurality of inflow influencing factors; The parameter determination module is further configured to determine a plurality of outflow influencing factors and corresponding outflow optimal time lags of the plurality of outflow influencing factors; The model establishment module is configured to establish a hydrological simulator based on the plurality of inflow influencing factors and the corresponding inflow optimal time lags of the plurality of inflow influencing factors and the plurality of outflow influencing factors and the corresponding outflow optimal time lags of the plurality of outflow influencing factors, wherein the hydrological simulator at least comprises a residual structure, an input gate, a cell state update layer, an output gate, and a linear layer, the residual structure is configured to introduce an input cell state and an output cell state of a previous time step, the input gate is configured to calculate a degree to which new information is accepted and generate a new candidate memory state, the cell state update layer is configured to update a cell state according to an output of the residual structure and an output of the input gate, the output gate is configured to determine which information in an output cell state of a current time step output by the cell state update layer needs to be output into a hidden state, and the linear layer is configured to perform water level prediction according to an output of the output gate; The sample establishment module is configured to establish a plurality of training samples; The model training module is configured to train the hydrological simulator through the plurality of training samples to generate a trained hydrological simulator; The data acquisition module is configured to acquire current input data of a to-be-simulated hydropower station based on the plurality of inflow influencing factors, the corresponding inflow optimal time lags of the plurality of inflow influencing factors, and the plurality of outflow influencing factors and the corresponding outflow optimal time lags of the plurality of outflow influencing factors; The hydrological simulation module is configured to predict a water level of the to-be-simulated hydropower station according to the current input data of the to-be-simulated hydropower station through the trained hydrological simulator.

Images