A sluice gate water level early warning system based on rainfall

Abstract

The application discloses a kind of sluice water level early warning systems based on rainfall, it is related to water level prediction technical field, setting water level information collection module real-time acquisition water level information of flowing through each sluice;Setting water level information sharing module carries out information sharing to each sluice passage flow and sluice downstream water level data;Setting rainfall information collection module monitors the rainfall after each sluice;Setting model training module generates the deep neural network model of predicting downstream sluice water level change according to upstream and downstream rainfall, upstream and downstream water level and upstream passage flow;Setting water level prediction module uses deep neural network model to predict the water level change of downstream sluice in real time according to upstream passage flow, upstream and downstream water level and rainfall;Setting water level early warning module according to the predicted height of water level, early warning is carried out to the sluice possibly exceeding water level warning line;Solve the problem of predicting water level change according to rainfall, early warning.

Description

Technical Field

[0001] This invention belongs to the field of sluice gate water level measurement and involves deep learning technology. Specifically, it is a sluice gate water level early warning system based on rainfall. Background Technology

[0002] The main functions of sluice gates on natural river channels, used to regulate upstream water levels and control downstream flow, are: during the dry season, closing the gates to raise upstream water levels to meet the requirements of navigation, irrigation, power generation, and urban water supply; and during the flood season, opening the gates to release floodwaters, ensuring that the upstream water supply level does not exceed the flood control limit, while simultaneously controlling the downstream flow to ensure it does not exceed the safe discharge capacity of the downstream river channel. On sediment-laden rivers, sluice gates also serve the task of desilting and releasing sediment, striving to maintain the original water-sediment balance of the riverbed.

[0003] Currently, the opening degree of the upstream and downstream sluice gates is estimated based on the flow rate and current water level data, but the following problems exist in the actual gate opening process:

[0004] 1. A water level monitoring point was set up upstream and downstream of each sluice gate, but information sharing between upstream and downstream water level monitoring points was not achieved, resulting in redundant arrangement of water level monitoring points.

[0005] 2. Failure to consider the impact of upstream rainfall on the downstream water level of the sluice gate resulted in the downstream sluice gate being unable to plan the degree of sluice gate opening in advance based on the rainfall situation.

[0006] To address this, a sluice gate water level early warning system based on rainfall is proposed. Summary of the Invention

[0007] This invention aims to at least solve one of the technical problems existing in the prior art. To this end, this invention proposes a sluice gate water level early warning system based on rainfall. This system includes: a water level information collection module to acquire real-time water level information of each sluice gate; a water level information sharing module to share the flow rate and downstream water level data of each sluice gate; a rainfall information collection module to monitor the rainfall downstream of each sluice gate; a model training module to generate a deep neural network model to predict the water level changes of downstream sluice gates based on upstream and downstream rainfall, upstream and downstream water levels, and upstream flow rate; a water level prediction module to predict the water level changes of downstream sluice gates in real-time using the deep neural network model based on upstream flow rate, upstream and downstream water levels, and upstream and downstream rainfall; and a water level early warning module to issue early warnings for sluice gates that may exceed the water level warning line based on the predicted water level height. This solves the problem of predicting water level changes based on rainfall and providing early warnings.

[0008] To achieve the above objectives, according to an embodiment of the first aspect of the present invention, a sluice gate water level early warning system based on rainfall is proposed, comprising a water level information collection module, a water level information sharing module, a rainfall information collection module, a model training module, a water level prediction module, and a water level early warning module; wherein the modules are connected to each other via electrical and / or wireless network means;

[0009] The river sluice gates are numbered sequentially from upstream to downstream along the river's flow. The numbering method is numerical, with the sluice gates numbered 1, 2, ..., N, where N is the total number of sluice gates on the river, and n represents each sluice gate.

[0010] The water level information collection module is mainly used to acquire water level information flowing through each sluice gate in real time.

[0011] The water level information collection module consists of water level monitoring points deployed after each sluice gate. The location of the water level monitoring points should avoid the influence of the gradual section and turbulence. The flow rate through the gate is calculated using the free outflow method, and the water level after each sluice gate n is obtained through a water level gauge. The flow rate through sluice gate n is marked as Fnt, and the water level after the gate is marked as Wnt. The water level information collection module sends the flow rate Fnt and the water level Wnt after sluice gate n to the water level information sharing module. Here, t is time.

[0012] The water level information sharing module is mainly used to share information such as the flow rate through the upstream sluice gate and the water level data after the gate with the downstream sluice gate.

[0013] The water level information sharing module pre-installs a wireless signal transceiver device in each sluice gate; the water level information sharing module transmits the correspondence between the sluice gate number n, the flow rate Fnt, and the water level Wnt after the gate to the water level prediction module in real time through the wireless signal transceiver device.

[0014] The rainfall information collection module is mainly used to monitor the rainfall behind each sluice gate in real time.

[0015] The rainfall information collection module includes a rainfall monitoring device installed after each sluice gate; the rainfall monitoring device monitors the rainfall after each sluice gate in real time; the rainfall can be expressed in millimeters per minute; the rainfall after sluice gate n is marked as Rnt; the rainfall information collection module sends the number of each sluice gate and its corresponding rainfall to the water level prediction module;

[0016] The model training module is mainly used to generate a deep neural network model to predict the water level changes after the downstream sluice gate based on upstream and downstream rainfall, downstream water level, upstream sluice gate flow and upstream water level.

[0017] The model training module generates a deep neural network model for each sluice gate n to predict the water level change after the downstream sluice gate n+1, including the following steps:

[0018] Step S1: The model training module pre-collects the flow rate Fnt, water level Wnt, and rainfall Rnt after each sluice gate n at time t over several days, as well as the water level W(n+1)t and rainfall R(n+1)t after sluice gate n+1; and estimates the time required to travel from sluice gate n to sluice gate n+1 based on the water flow velocity; this time is marked as ft; the model training module obtains the water level W(n+1)(t+ft) after time t+ft; and calculates C(n+1)t = W(n+1)(t+ft) - W(n+1)t as the water level change of sluice gate n+1;

[0019] Step S2: Normalize and combine the flow rate Fnt, the water level Wnt and W(n+1)t downstream of the gate, the rainfall Rnt and R(n+1)t downstream of the gate into a feature vector; label this feature vector as Vn;

[0020] Step S3: Input the feature vector Vn into the deep neural network for training; the deep neural network outputs the water level change value; the actual water level change value C(n+1)t of the sluice gate n+1 is used as the target value; when the water level change prediction accuracy of the trained deep neural network reaches 95%, training is stopped; the trained deep neural network model is labeled as Mn.

[0021] The model training module sends all trained deep neural network models Mn to the water level prediction module;

[0022] The water level prediction module is mainly used to predict the water level changes of the downstream sluice gate in real time based on the upstream sluice gate flow, upstream and downstream water levels and rainfall, using a deep neural network model M.

[0023] The water level prediction module predicts the water level change of the downstream sluice gate (n+1) for each sluice gate n, including the following steps:

[0024] Step P1: The water level prediction module receives the flow rate through each sluice gate n, the water level after sluice gate n and sluice gate n+1 sent by the water level information sharing module; and the rainfall after sluice gate n and sluice gate n+1 sent by the rainfall information collection module; and normalizes each data point to form a feature vector.

[0025] Step P2: Input the feature vector into the deep neural network model M to obtain the predicted water level change value of sluice gate n+1; label the predicted water level change value as P(n+1)t;

[0026] Step P3: Based on the water flow velocity after sluice gate n, estimate the time ft required for the water flow to travel from sluice gate n to sluice gate n+1; then, after the predicted time t+ft, the water level height after sluice gate n+1 is W(n+1)(t+ft)=W(n+1)t+P(n+1)t.

[0027] The water level prediction module sends the predicted water level height W(n+1)(t+ft) for each sluice gate n+1 to the water level early warning module in real time;

[0028] The water level early warning module is mainly used to provide early warnings for sluice gates that may exceed the water level warning line in real time based on the predicted water level height, and to notify monitoring personnel to control the gate opening size in advance to reduce the flow of water through the gate.

[0029] The water level early warning module pre-sets a warning water level line W(n+1) for each sluice gate according to the actual situation of the sluice gate; when the predicted water level height W(n+1)(t+ft) of sluice gate n+1 is greater than W(n+1), the water level early warning information is sent to the monitoring personnel responsible for sluice gate n+1 through voice broadcast.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] This invention includes a water level information collection module to acquire real-time water level information passing through each sluice gate; a water level information sharing module to share data on the flow rate and downstream water level of each sluice gate; a rainfall information collection module to monitor the rainfall downstream of each sluice gate; a model training module to generate a deep neural network model to predict the water level changes of downstream sluice gates based on upstream and downstream rainfall, upstream and downstream water levels, and upstream flow rate; a water level prediction module to predict the water level changes of downstream sluice gates in real-time using the deep neural network model based on upstream flow rate, upstream and downstream water levels, and upstream and downstream rainfall; and a water level early warning module to issue warnings for sluice gates that may exceed the water level warning line based on the predicted water level height. This invention solves the problem of predicting water level changes based on rainfall and providing early warnings. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0033] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] like Figure 1As shown, a sluice gate water level early warning system based on rainfall includes a water level information collection module, a water level information sharing module, a rainfall information collection module, a model training module, a water level prediction module, and a water level early warning module; wherein, the various modules are connected to each other via electrical and / or wireless network means;

[0035] It is understandable that sluice gates on rivers are built sequentially at intervals along the river; most sluice gates are equipped with hydraulic structure measuring stations to monitor the flow rate of water passing through the gate and the water level of the river near the sluice gate.

[0036] The river sluice gates are numbered sequentially from upstream to downstream along the river's flow. The numbering method is numerical, with the sluice gates numbered 1, 2, ..., N, where N is the total number of sluice gates on the river, and n represents each sluice gate.

[0037] The water level information collection module is mainly used to acquire water level information flowing through each sluice gate in real time.

[0038] In a preferred embodiment, the water level information collection module consists of water level monitoring points deployed after each sluice gate. The location of the water level monitoring points should avoid the influence of the gradual section and turbulence. The flow rate through the gate is calculated using a free outflow method, and the water level after each sluice gate n is obtained through a water level gauge. The flow rate through sluice gate n is marked as Fnt, and the water level after the gate is marked as Wnt. The water level information collection module sends the flow rate Fnt and the water level Wnt of sluice gate n to the water level information sharing module. Wherein, t is time.

[0039] The water level information sharing module is mainly used to share the flow rate through the upstream sluice gate and the water level data after the gate with the downstream sluice gate.

[0040] In a preferred embodiment, the water level information sharing module pre-installs a wireless signal transceiver device at each sluice gate; the water level information sharing module transmits the correspondence between the sluice gate number n, the flow rate Fnt, and the water level Wnt after the gate to the water level prediction module in real time via the wireless signal transceiver device.

[0041] The rainfall information collection module is mainly used to monitor the rainfall behind each sluice gate in real time.

[0042] In a preferred embodiment, the rainfall information collection module includes a rainfall monitoring device installed after each sluice gate; the rainfall monitoring device monitors the rainfall after each sluice gate in real time; the rainfall can be expressed in millimeters per minute; it is understood that, due to the long river length, the locations of different sluice gates may be far apart, therefore, at the same time, the rainfall monitored by the rainfall monitoring devices of different sluice gates may be different; the rainfall after sluice gate n is marked as Rnt; the rainfall information collection module sends the number of each sluice gate and its corresponding rainfall to the water level prediction module;

[0043] The model training module is mainly used to generate a deep neural network model to predict the water level changes after the downstream sluice gate based on upstream and downstream rainfall, downstream water level, upstream sluice gate flow and upstream water level.

[0044] In a preferred embodiment, the model training module generates a deep neural network model for each sluice gate n to predict the water level change after the downstream sluice gate n+1, comprising the following steps:

[0045] Step S1: The model training module pre-collects the flow rate Fnt, water level Wnt, and rainfall Rnt after each sluice gate n at time t over several days, as well as the water level W(n+1)t and rainfall R(n+1)t after sluice gate n+1; and estimates the time required to travel from sluice gate n to sluice gate n+1 based on the water flow velocity; this time is marked as ft; the model training module obtains the water level W(n+1)(t+ft) after time t+ft; and calculates C(n+1)t = W(n+1)(t+ft) - W(n+1)t as the water level change of sluice gate n+1;

[0046] Step S2: Normalize and combine the flow rate Fnt, the water level Wnt and W(n+1)t downstream of the gate, the rainfall Rnt and R(n+1)t downstream of the gate into a feature vector; label this feature vector as Vn;

[0047] Step S3: Input the feature vector Vn into the deep neural network for training; the deep neural network outputs the water level change value; the actual water level change value C(n+1)t of the sluice gate n+1 is used as the target value; when the water level change prediction accuracy of the trained deep neural network reaches 95%, training is stopped; the trained deep neural network model is labeled as Mn.

[0048] The model training module sends all trained deep neural network models Mn to the water level prediction module;

[0049] The water level prediction module is mainly used to predict the water level changes of the downstream sluice gate in real time based on the upstream flow rate, upstream and downstream water levels and upstream and downstream rainfall, using a deep neural network model M.

[0050] In a preferred embodiment, the water level prediction module predicts the water level change of the downstream sluice gate n+1 for each sluice gate n, including the following steps:

[0051] Step P1: The water level prediction module receives the flow rate through each sluice gate n, the water level after sluice gate n and sluice gate n+1 sent by the water level information sharing module; and the rainfall after sluice gate n and sluice gate n+1 sent by the rainfall information collection module; and normalizes each data point to form a feature vector.

[0052] Step P2: Input the feature vector into the deep neural network model M to obtain the predicted water level change value of sluice gate n+1; label the predicted water level change value as P(n+1)t;

[0053] Step P3: Based on the water flow velocity after sluice gate n, estimate the time ft required for the water flow to travel from sluice gate n to sluice gate n+1; then, after the predicted time t+ft, the water level height after sluice gate n+1 is W(n+1)(t+ft)=W(n+1)t+P(n+1)t.

[0054] The water level prediction module sends the predicted water level height W(n+1)(t+ft) for each sluice gate n+1 to the water level early warning module in real time;

[0055] The water level early warning module is mainly used to provide early warnings for sluice gates that may exceed the water level warning line in real time based on the predicted water level height, and to notify monitoring personnel to control the gate opening size in advance to reduce the flow of water through the gate.

[0056] In a preferred embodiment, the water level early warning module pre-sets a warning water level line W(n+1) for each sluice gate according to the actual situation of the sluice gate; when the predicted water level height W(n+1)(t+ft) of sluice gate n+1 is greater than W(n+1), the water level early warning information is sent to the monitoring personnel responsible for sluice gate n+1 by voice broadcast.

[0057] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A rainfall-based sluice water level warning system, characterized by, The water level information collection module is used for real-time acquisition of water level information flowing through each water gate; the water level information collection module sends the gate-passing flow and the water level behind the gate of the water gate to the water level information sharing module; The water level information sharing module is used for information sharing of the gate-passing flow and the water level data behind the gate of the upstream water gate with the downstream water gate; the water level information sharing module sends the corresponding relationship between the water gate number and the gate-passing flow and the water level behind the gate to the water level prediction module in real time; The rainfall information collection module is used for real-time monitoring of the rainfall behind each water gate; the rainfall information collection module sends the number of each water gate and its corresponding rainfall to the water level prediction module; The model training module is used for generating a deep neural network model for predicting the water level change behind the downstream water gate according to the upstream and downstream rainfall, the downstream water level, the upstream gate-passing flow and the upstream water level; the model training module sends all the trained deep neural network models to the water level prediction module; The water level prediction module is used for real-time prediction of the water level change of the downstream water gate according to the upstream gate-passing flow, the upstream and downstream water levels and the upstream and downstream rainfall by using the deep neural network model; the water level prediction module sends the water level prediction height of each water gate to the water level warning module in real time; The water level warning module is used for real-time warning of the water gate that may exceed the water level warning line according to the predicted height of the water level. The water level information collection module is a water level monitoring point arranged behind each water gate; the establishment position of the water level monitoring point should avoid the influence of the gradual section and the turbulence; and the free outflow method is used to calculate the gate-passing flow, and the water level meter is used to obtain the water level behind each water gate n; the gate-passing flow of the water gate n is marked as Fnt; and the water level behind the gate is marked as Wnt; wherein, t is time.

2. The rainfall-based sluice water level warning system according to claim 1, characterized in that, The rainfall information collection module includes a rainfall monitoring device installed behind each water gate; the rainfall monitoring device monitors the rainfall behind each water gate in real time; the rainfall behind the water gate n is marked as Rnt.

3. The rainfall-based sluice water level warning system according to claim 1, wherein The model training module generates a deep neural network model for predicting the water level change behind the downstream water gate n+1 for each water gate n, including the following steps:

4. The rainfall-based sluice water level warning system according to claim 1, wherein ​ Step S1: The model training module pre-collects the flow Fnt after each sluice n, the water level Wnt after the sluice, the rainfall Rnt after the sluice, the water level W(n+1)t after the sluice n+1, and the rainfall R(n+1)t after the sluice; and estimates the time required for the water flow from the sluice n to the sluice n+1 according to the flow velocity; marks the time as ft; the model training module obtains the water level W(n+1)(t+ft) after the sluice n+1 at time t+ft; and calculates C(n+1)t=W(n+1)(t+ft)-W(n+1)t as the water level change of the sluice n+1; Step S2: The flow Fnt after the sluice, the water level Wnt and W(n+1)t after the sluice, and the rainfall Rnt and R(n+1)t after the sluice are normalized and combined into a feature vector; the feature vector is marked as Vn; Step S3: The feature vector Vn is input into the deep neural network for training; the deep neural network takes the water level change value as the output; the actual water level change value C(n+1)t of the sluice n+1 is taken as the target value; when the prediction accuracy of the water level change of the trained deep neural network reaches 95%, the training is stopped; and the trained deep neural network model is marked as Mn.

5. The rainfall-based sluice water level warning system according to claim 1, wherein The water level prediction module predicts the water level change of the downstream sluice n+1 for each sluice n, including the following steps: Step P1: The water level prediction module receives the flow after each sluice n, the water level after the sluice n and the sluice n+1, and the rainfall after the sluice n and the sluice n+1 sent by the water level information sharing module; and the rainfall information collection module; each data is normalized and combined into a feature vector; Step P2: The feature vector is input into the deep neural network model M to obtain the predicted water level change value of the sluice n+1; the predicted water level change value is marked as P(n+1)t; Step P3: According to the water flow velocity after the sluice n, the time ft required for the water flow to reach the sluice n+1 from the sluice n is estimated; then the predicted water level height W(n+1)(t+ft) of the sluice n+1 at time t+ft is W(n+1)t+P(n+1)t.

6. The rainfall-based sluice water level warning system according to claim 1, wherein The water level warning module sets a warning water level line W(n+1) for each sluice according to the actual situation of the sluice; when the predicted water level height W(n+1)(t+ft) of the sluice n+1 is greater than W(n+1), the water level warning information is sent to the monitoring personnel in charge of the sluice n+1 through voice broadcast.

Images