A flood control dispatching management method and system based on digital twin basin

Abstract

The present application discloses a flood control dispatching management method and system based on a digital twin basin, which involves water conservancy management and data processing technology, including: pre-establishing a digital basic model of the target basin based on the historical lowest water level data of the target basin; constructing a training water level sequence based on the water level data under historical weather data; extracting the characteristic information of the water level sequence, and training a long short-term memory neural network LSTM model based on the extracted characteristic information; obtaining the current weather data of the target basin; extracting the characteristic information of the water level sequence to be predicted, and outputting the water level information predicted by the array and the flow rate under the corresponding water level based on the trained LSTM model; based on the current water level information presented by the digital basic model, the predicted water level information and the flow rate under the corresponding water level are presented in association. The method of the present application can realize flood control dispatching management of the basin and improve the auxiliary effect on managers.

Description

Technical Field

[0001] The present application relates to the fields of water conservancy management and data processing technology, and in particular to a flood control dispatching management method and system based on a digital twin river basin. Background Art

[0002] At present, most mainstream flood forecasting models are traditional hydrological models based on the physical mechanism of hydrological processes. They often require detailed definitions of various elements involved in the hydrological cycle within the basin and describe the hydrological processes with complex mathematical formulas. However, the uncertainty of hydrological elements and the complexity of hydrological processes have limited the development of traditional hydrological models.

[0003] Thanks to the development of machine learning and big data technology, artificial neural network technology has made great progress in various fields. In the field of hydrological forecasting, some studies focus on recursive neural networks that can learn the state characteristics of time series.

[0004] However, the existing flood control scheduling and management solutions are not mature and complete and cannot achieve the ideal auxiliary effect. Summary of the invention

[0005] The embodiment of the present application provides a flood control scheduling management method and system based on a digital twin river basin, which gradually performs predictions based on digital twin technology and an LSTM model and presents them to users, so as to realize flood control scheduling management of the river basin and improve the auxiliary effect on managers.

[0006] The embodiment of the present application proposes a flood control scheduling management method based on a digital twin watershed, including:

[0007] Establishing a digital basic model of the target basin based on the historical lowest water level data of the target basin in advance, and obtaining current water level information of the target basin to present the current water level information in the digital basic model based on the water level information;

[0008] Acquire historical weather data of the target watershed, construct a training water level sequence based on the water level data under the historical weather data, and add an associated flow label to the constructed training water level sequence;

[0009] Extracting characteristic information of the water level sequence, and training a long short-term memory neural network LSTM model based on the extracted characteristic information;

[0010] Acquire current weather data of the target watershed, the current weather data including weather data of a set time length before and after the current moment, construct a water level sequence to be predicted based on the current weather data, and add associated flow tags to known water level sequences in the water level sequence to be predicted;

[0011] Extracting characteristic information of the water level sequence to be predicted, and outputting array predicted water level information and flow rate at the corresponding water level based on the trained LSTM model;

[0012] Based on the current water level information presented by the digital basic model, the predicted water level information and the flow rate under the corresponding water level are presented in association.

[0013] Optionally, a digital basic model of the target basin is established in advance based on the historical lowest water level data of the target basin, including:

[0014] Acquiring elevation data of the target watershed;

[0015] The historical lowest water level data is used as the water level reference, and the set water level deviation is determined according to the historical highest water level data;

[0016] In the target watershed, based on the water level reference, terrain data between the water level reference and the historical highest water level data including the water level deviation are established as the digital basic model.

[0017] Optionally, also include:

[0018] Acquire several key elevations based on the elevation data of the target watershed, wherein the key elevations include the building elevations and disaster elevations within the target watershed, and after the water level of the watershed reaches the key elevations, a disaster is formed at a related position of the target watershed; and,

[0019] The key elevations are mapped to the digital basic model through water levels, and under normal conditions, the water levels corresponding to the key elevations are controlled to be hidden.

[0020] Optionally, presenting the current water level information in the digital basic model based on the water level information includes:

[0021] Based on the acquired water level information, the current water area representative pixels of the water level information are filled above the water level benchmark of the digital basic model, wherein different representative pixel intervals are pre-configured for the normal water period and the flood season, and the current water area representative pixels are selected based on the corresponding representative pixel intervals.

[0022] Optionally, associated flow labels are added to the constructed training water level sequence including:

[0023] Obtain gate opening and power generation data under the historical weather data;

[0024] The acquired gate opening and power generation data are converted into drainage flow, so as to use the drainage flow as a flow label.

[0025] Optionally, extracting characteristic information of the water level sequence to be predicted, and outputting array predicted water level information and flow at the corresponding water level based on the trained LSTM model includes:

[0026] Extract the characteristic information of the water level sequence to be predicted, and output the water level information at the next moment and the flow rate at the corresponding water level based on the trained LSTM model;

[0027] According to the set step size, marks are added based on the output water level information at the next moment and the flow rate at the corresponding water level to construct the water level sequence to be predicted in the next step, and the predicted water level information and the flow rate at the corresponding water level are output again based on the trained LSTM model until the predicted moment meets the set requirements.

[0028] Optionally, based on the trained LSTM model, the water level information at the next moment and the flow rate at the corresponding water level are output, including:

[0029] Taking each of the key elevations as a constraint, based on extracting characteristic information of the water level sequence to be predicted, outputting multiple sets of predicted water level information and flow relationship for the same subsequent moment;

[0030] Based on the predicted water level information and flow relationship, a prediction node is constructed to form a tree-like node relationship based on each subsequent moment, and any node includes a set of predicted water level information and flow relationship.

[0031] Optionally, based on the current water level information presented by the digital basic model, the associated presentation of the predicted water level information and the flow rate at the corresponding water level includes:

[0032] Based on the current water level information presented by the digital basic model, when the distance between the current water level information and the nearest key elevation is less than a first threshold, presenting the nearest key elevation in the digital basic model; and

[0033] When the distance between the current water level information and the nearest key elevation is less than the second threshold, multiple groups of predicted water level information and the scheduling relationship of the corresponding flow are presented based on the prediction node corresponding to the current moment.

[0034] An embodiment of the present application also proposes a flood control scheduling and management system based on a digital twin river basin, including a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the flood control scheduling and management method based on the digital twin river basin as described above are implemented.

[0035] The method of the embodiment of the present application gradually performs predictions based on digital twin technology and LSTM model and presents them to users, so as to realize flood control scheduling management of the river basin and improve the auxiliary effect for managers.

[0036] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0038] Figure 1 The basic process diagram of the flood control dispatching management method based on the digital twin watershed of this embodiment is shown in FIG. DETAILED DESCRIPTION

[0039] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0040] The present application embodiment proposes a flood control dispatch management method based on a digital twin watershed, such as Figure 1 As shown, the following steps are included:

[0041] In step S101, a digital basic model of the target watershed is established in advance based on the historical minimum water level data of the target watershed, and the current water level information of the target watershed is obtained to present the current water level information in the digital basic model based on the water level information. The digital basic model of the target watershed can be established by using methods such as GIS, and the established digital basic model can be used as a twin model of the target watershed. The model can be updated according to real-time water level data, so that users can view it and assist users in watershed management.

[0042] In step S102, historical weather data of the target watershed is obtained, a training water level sequence is constructed based on the water level data under the historical weather data, and an associated flow tag is added to the constructed training water level sequence. In some specific examples, adding an associated flow tag to the constructed training water level sequence includes:

[0043] Obtain gate opening and power generation data under the historical weather data;

[0044] The gate opening and power generation data obtained are converted into drainage flow, so that the drainage flow is used as a flow label. In a specific example, especially for flood season floods, the corresponding flow label configuration can configure a preferred scheduling management strategy, so that it can be presented based on the digital basic model in subsequent examples, and prompts can be issued to management personnel through the scheduling management strategy.

[0045] In step S103, feature information of the water level sequence is extracted, and a long short-term memory neural network LSTM model is trained based on the extracted feature information. In a specific example, the LSTM model can be effectively applied to the prediction of water level sequences with time memory, and can be used for the prediction task of water level sequences after training.

[0046] In step S104, the current weather data of the target basin is obtained, and the current weather data includes weather data of a set time length before and after the current moment. A water level sequence to be predicted is constructed based on the current weather data, and an associated flow tag is added to the known water level sequence in the water level sequence to be predicted. In a specific example, a subsequent predicted water level sequence is constructed in a cyclic manner, so that the predicted water level information of the required time in the future and the flow at the corresponding water level can be gradually predicted.

[0047] In step S105, the characteristic information of the water level sequence to be predicted is extracted, so as to output the array predicted water level information and the flow rate under the corresponding water level based on the trained LSTM model.

[0048] In step S106, based on the current water level information presented by the digital basic model, the predicted water level information and the flow rate at the corresponding water level are presented in association. In the embodiment of the present application, based on the current water level information presented by the digital basic model, pixels different from the current water level information can be used to present the water level difference between the predicted water level information and the current water level information. In a specific example, the water level difference can also be presented in a flashing manner.

[0049] The method of the embodiment of the present application gradually performs predictions based on digital twin technology and LSTM model and presents them to users, so as to realize flood control scheduling management of the river basin and improve the auxiliary effect for managers.

[0050] In some embodiments, establishing a digital basic model of the target watershed in advance based on historical minimum water level data of the target watershed includes:

[0051] Acquiring elevation data of the target watershed;

[0052] The historical lowest water level data is used as the water level reference, and the set water level deviation is determined according to the historical highest water level data;

[0053] In the target watershed, based on the water level reference, terrain data between the water level reference and the historical highest water level data including the water level deviation are established as the digital basic model.

[0054] Specifically, the embodiment of the present application does not use riverbed elevation data to establish a digital basic model, but uses, for example, the lowest water level of the target basin in the dry season and the highest water level in the flood season, and uses the deviation between the lowest water level and the highest water level to establish a terrain model, thereby using the terrain model as the digital basic model, and other intervals are not processed, thereby reducing the amount of data calculations in the modeling process, and in the specific example.

[0055] In some embodiments, it also includes:

[0056] Based on the elevation data of the target watershed, several key elevations are obtained, wherein the key elevations include the building elevations and disaster elevations within the target watershed. After the water level in the watershed reaches the key elevations, disasters are formed at the associated locations of the target watershed. Specifically, for example, the building elevations of one or several residential buildings closest to the river bank in the target watershed can be obtained and the corresponding elevations can be used as key elevations. Disaster elevations refer to elevations that may cause disasters after the water level reaches this elevation, such as farmland that may be invaded by floods. In this case, the relevant elevations can be used as key elevations. In a specific example, the buildings or fields included in the key elevations may not be modeled, and only the water level data of the key elevations can be retained.

[0057] The key elevations are mapped to the digital basic model through water levels, and under normal conditions, the water levels corresponding to the key elevations are controlled to be hidden. Under normal conditions, the water levels corresponding to the key elevations are controlled to be hidden, so that the watershed presentation based on the basic digital model can be more focused and the user's visual experience can be improved.

[0058] In some embodiments, presenting the current water level information in the digital base model based on the water level information includes:

[0059] Based on the acquired water level information, the current water area representative pixel of the water level information is filled in above the water level reference of the digital basic model, wherein different representative pixel intervals are configured in advance for the normal water period and the flood season, and the current water area representative pixel is selected based on the corresponding representative pixel interval. For example, the corresponding pixel interval can be used for the clear water flow in the normal water period and the turbid water flow in the flood season, and of course, it can also be a user-defined pixel interval, which is not specifically limited.

[0060] In some embodiments, extracting characteristic information of the water level sequence to be predicted, and outputting array predicted water level information and flow at the corresponding water level based on the trained LSTM model includes:

[0061] Extract the characteristic information of the water level sequence to be predicted, and output the water level information at the next moment and the flow rate at the corresponding water level based on the trained LSTM model;

[0062] According to the set step size, based on the output water level information at the next moment and the flow rate at the corresponding water level, a mark is added to construct the next step of the water level sequence to be predicted, and based on the trained LSTM model, the predicted water level information and the flow rate at the corresponding water level are output again until the predicted moment meets the set requirements. That is, this example performs the prediction of water level information and flow rate by stepping back prediction.

[0063] In some embodiments, based on the trained LSTM model, outputting the water level information at the next moment and the flow rate at the corresponding water level includes:

[0064] Taking each of the key elevations as a constraint, based on extracting characteristic information of the water level sequence to be predicted, outputting multiple sets of predicted water level information and flow relationship for the same subsequent moment;

[0065] Based on the predicted water level information and flow relationship, a prediction node is constructed to form a tree-like node relationship based on each subsequent moment, and any node includes a set of predicted water level information and flow relationship.

[0066] In a specific example, through the prediction of the tree-like node relationship, the pre-configured preferred scheduling management strategy can be combined, where the scheduling management strategy can include measures such as power generation and dam gate opening. In this embodiment, a scheduling management suggestion sequence can be formed based on the tree-like nodes, so as to provide managers with multi-faceted scheduling management suggestions. In a specific example, for any subsequent node, after the user selects a strategy, other nodes and recommended strategies at the same level of the corresponding node are eliminated, and subsequent tracking predictions are performed to improve the auxiliary effect.

[0067] In some embodiments, based on the current water level information presented by the digital basic model, the associated presentation of the predicted water level information and the flow rate at the corresponding water level includes:

[0068] Based on the current water level information presented by the digital basic model, when the distance between the current water level information and the nearest key elevation is less than a first threshold, the nearest key elevation is presented on the digital basic model. In a specific example, when the nearest key elevation is presented on the digital basic model, a water level distance prompt can be provided for the key elevation, and the key elevation can be displayed in other colors with high contrast to the representative pixel interval as a warning, such as red.

[0069] When the distance between the current water level information and the nearest key elevation is less than the second threshold, the subsequent multiple sets of predicted water level information and the dispatching relationship of the corresponding flow are presented based on the prediction node corresponding to the current moment. In some examples, the second threshold may be less than the first threshold, so that for example, for flood control dispatching during the flood season, the multiple sets of predicted water level information and the flood control dispatching relationship of the corresponding flow can be directly displayed in the digital basic model in association with the key elevation. In a specific implementation, the prediction step can be shortened, thereby increasing the subsequent prediction points, so as to achieve near real-time feedback of water level information and dispatching strategies based on the digital basic model, thereby improving the auxiliary effect for managers.

[0070] An embodiment of the present application also proposes a flood control scheduling and management system based on a digital twin river basin, including a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the flood control scheduling and management method based on the digital twin river basin as described above are implemented.

[0071] In addition, although exemplary embodiments have been described herein, the scope includes any and all embodiments based on the present disclosure with equivalent elements, modifications, omissions, combinations (e.g., various embodiments intersecting schemes), adaptations or changes. It is not limited to the examples described in this specification or during the implementation of this application, and its examples will be interpreted as non-exclusive.

[0072] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of them) can be used in combination with each other. For example, those of ordinary skill in the art may use other embodiments when reading the above description.

[0073] The above embodiments are merely exemplary embodiments of the present disclosure. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the essence and protection scope of the present disclosure, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present invention.

Claims

1. A flood control dispatching management method based on digital twin watershed, characterized in that: include: Establishing a digital basic model of the target basin based on the historical lowest water level data of the target basin in advance, and obtaining current water level information of the target basin to present the current water level information in the digital basic model based on the water level information; Acquire historical weather data of the target watershed, construct a training water level sequence based on the water level data under the historical weather data, and add an associated flow label to the constructed training water level sequence; Extracting characteristic information of the water level sequence, and training a long short-term memory neural network LSTM model based on the extracted characteristic information; Acquire current weather data of the target watershed, the current weather data including weather data of a set time length before and after the current moment, construct a water level sequence to be predicted based on the current weather data, and add associated flow tags to known water level sequences in the water level sequence to be predicted; Extracting characteristic information of the water level sequence to be predicted, and outputting array predicted water level information and flow rate at the corresponding water level based on the trained LSTM model; Based on the current water level information presented by the digital basic model, the predicted water level information and the flow rate under the corresponding water level are presented in association; The digital basic model of the target basin is established in advance based on the historical lowest water level data of the target basin, including: Acquiring elevation data of the target watershed; The historical lowest water level data is used as the water level reference, and the set water level deviation is determined according to the historical highest water level data; In the target watershed, based on the water level reference, terrain data between the water level reference and the historical highest water level data including the water level deviation is established as the digital basic model; Also includes: Acquire several key elevations based on the elevation data of the target watershed, wherein the key elevations include the building elevations and disaster elevations within the target watershed, and after the water level of the watershed reaches the key elevations, a disaster is formed at a related position of the target watershed; and, Mapping the key elevations to the digital basic model through water levels, and in a normal state, controlling to hide the water levels corresponding to the key elevations; Extracting the characteristic information of the water level sequence to be predicted, and outputting the water level information predicted by the array and the flow rate under the corresponding water level based on the trained LSTM model includes: Extract the characteristic information of the water level sequence to be predicted, and output the water level information at the next moment and the flow rate at the corresponding water level based on the trained LSTM model; According to the set step size, marks are added based on the output water level information at the next moment and the flow rate at the corresponding water level to construct the water level sequence to be predicted in the next step, and the predicted water level information and the flow rate at the corresponding water level are output again based on the trained LSTM model until the predicted moment meets the set requirements.

2. The flood control dispatching management method based on digital twin watershed according to claim 1, characterized in that: Presenting the current water level information in the digital basic model based on the water level information includes: Based on the acquired water level information, the current water area representative pixels of the water level information are filled above the water level benchmark of the digital basic model, wherein different representative pixel intervals are pre-configured for the normal water period and the flood season, and the current water area representative pixels are selected based on the corresponding representative pixel intervals.

3. The flood control dispatching management method based on digital twin watershed according to claim 1, characterized in that: The associated flow labels added to the constructed training water level series include: Obtain gate opening and power generation data under the historical weather data; The acquired gate opening and power generation data are converted into drainage flow, so as to use the drainage flow as a flow label.

4. The flood control dispatching management method based on digital twin watershed according to claim 1, characterized in that: Based on the trained LSTM model, the water level information at the next moment and the flow rate at the corresponding water level are output, including: Taking each of the key elevations as a constraint, based on extracting characteristic information of the water level sequence to be predicted, outputting multiple sets of predicted water level information and flow relationship for the same subsequent moment; Based on the predicted water level information and flow relationship, a prediction node is constructed to form a tree-like node relationship based on each subsequent moment, and any node includes a set of predicted water level information and flow relationship.

5. The flood control dispatching management method based on digital twin watershed according to claim 4, characterized in that: Based on the current water level information presented by the digital basic model, the predicted water level information and the flow rate under the corresponding water level are presented in association, including: Based on the current water level information presented by the digital basic model, when the distance between the current water level information and the nearest key elevation is less than a first threshold, presenting the nearest key elevation in the digital basic model; and When the distance between the current water level information and the nearest key elevation is less than the second threshold, multiple groups of predicted water level information and the scheduling relationship of the corresponding flow are presented based on the prediction node corresponding to the current moment.

6. A flood control dispatching and management system based on digital twin watershed, characterized in that: It includes a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the flood control scheduling management method based on the digital twin river basin as described in any one of claims 1 to 5 are implemented.

Images