30 results about "Flash flood" patented technology

The invention discloses a flash flood early warning method based on deep learning. According to the method, an RNN (recurrent neural network) model with an LTSM (long short-term memory) function is adopted. Problems of gradient disappearance and gradient explosion in the RNN can be effectively solved by adding the LTSM to a model algorithm, prediction of small watersheds or flash flood areas can be performed more effectively, reliable flash flood disaster early warning and predictionare achieved, and prediction time is greatly shortened. The model is only required to focus on input data, so that the model has the characteristics of being convenient to operate and high in operation speed, and facilitating value shifting to other areas. Moreover, with the practical application of the model,data is accumulated continuously, no secondary parameter adjustment is required, parameters can be optimized and adjusted actively by the model, and accordingly, the model is simpler and more convenient to use.

The invention provides a flash flood water level monitoring and pre-warning system, a monitor sensor of which is a buoy position sensor. The buoy position sensor is a digitally coded wireless electric module, wherein a conductive suspended hammer or a conductive rod, a conductive ring and a conductive contact are arranged in the buoy; generally, the conductive suspended hammer or conductive rod is not contacted with the conductive ring, and the sensor does not consume electricity; and when the buoy is inclined to a certain angle due to water level rising, the conductive suspended hammer or the conductive rod is contacted with the conductive ring, the wireless electric module is electrified to work, a coded wireless electrical signal is sent out and received by a monitoring center and a warning station of a protected region through a relay station, and an alarm is sent out. When the weight completely floats in the buoy position sensor, the conductive ring is separated from the conductive contact, the circuit is disconnected, and the sensor does not consume electricity, so the system can work repeatedly. The water level monitoring and pre-warning system can warn the absolute height of water level, and warn the instant strength of flood.

The invention discloses a flood forecasting method based on a multiple ratio algorithm, which includes: step 1, calculating the current flow value of the target small watershed; step 2, calculating the flow value in the upstream forecast period according to the actual measured grid point rainfall data; step 3, Compare the current flow value of the target small watershed with the flow value in the foreseeable period; step 4, when the flow value in the foreseeable period exceeds the current flow value of the target small watershed by more than twice the current flow value of the target small watershed, an early warning message is sent through the flood control forecasting system; it solves the problem of the prior art Mountain torrent disaster early warning is based on the monitoring of the current rainfall by the rainfall station, and it is determined whether to issue early warning information compared with the critical rainfall; because the upper reaches of many small watersheds have experienced light to moderate rain, the resulting changes in rainfall and flow are not enough to trigger the flood warning system. When the instantaneous flow rate suddenly doubled, it caused flash floods, and tourists in the river channel were not prevented enough to cause casualties and other problems.

The present invention relates to a mountain torrent disaster early warning and monitoring system based on edge computing, including a perception layer, a convergence layer, a transmission layer and an application layer: the perception layer is used to collect sensor data; the convergence layer participates in global collaboration at the edge of the sensor network , used to fuse different locations and different types of sensor data; the transmission layer is used to realize the communication link between the aggregation layer and the application layer; the application layer is used for real-time monitoring and emergency command of flash flood disaster-causing elements. The invention can provide low power consumption, low cost, distributed computing, mass connection, and real-time credible data services for disaster prevention and mitigation.

The invention discloses a flash flood early warning method based on deep learning. The invention adopts the RNN model with LTSM function, and adding the long short-term memory (LTSM) method to the model algorithm can effectively solve the problems of gradient disappearance and gradient explosion in RNN, and more Effectively forecast small watersheds or mountain torrent areas, realize reliable mountain torrent disaster early warning and forecast, greatly improve the forecast time, the model only needs to pay attention to the input data, has the characteristics of easy operation, fast calculation speed, and easy transfer to other areas, and With the actual application of the model and the continuous accumulation of data, there is no need to manually adjust the parameters twice. The model will actively optimize and adjust the parameters to make its application easier and more convenient.

The invention discloses a simple mountain torrent early-warning device based on an excess infiltration runoff principle. The simple mountain torrent early-warning device comprises a water collecting cup and an information display unit installed in the water colleting cup. A water penetrating port is formed in the bottom wall of the water collecting cup. An adjusting turn button used for adjusting a leakage speed is installed at the water penetrating port. The information display unit comprises a yellow precaution device, an orange precaution device and a red precaution device which are successively from a low water level to a high water level. The simple mountain torrent early-warning device has the advantages that the structure is simple, the excess infiltration runoff principle is used as the basis, and the excess infiltration runoff process is modeled and visualized; by means of the adjusting and setting the adjusting turn button, mountain torrents are accurately simulated; the device is high in adjustment, applicable to most mountainous areas, and high in popularization; in addition, when the device is in use, the forming and eruption of the mountain torrents can be grasped in real time only by noticing the prompts of the precaution devices and buzzer alarms, and the monitoring and early warning effects of the mountain torrents are further improved, so that the life and property safety of people in mountainous areas is ensured.

A concrete face rockfill dam water pre-settling construction method. After the empty steel gabions are neatly arranged and welded into one body, stones are filled in the gabions; the steel gabions are stacked along the downstream side slope of the rockfill dam body. , are connected to each other with steel bars; threaded steel bars are used as anchor rods, and the steel gabion is anchored with the mountains on both sides. After the water flows through the dam, the pre-settling of the water is completed. The concrete face rockfill dam water-passing pre-settling construction method provided by the present invention can withstand the test of repeated mountain torrent erosion after adopting reinforced gabion protection construction for the downstream side slope of the rockfill dam body; The settlement of the rockfill dam surface due to water passing through passively improves the compactness of the rockfill dam body and accelerates the settlement and deformation of the rockfill dam body. Water passing is equivalent to releasing the settlement of the rockfill dam body in advance, which is beneficial to Settlement deformation control of rockfill dam body and safety and stability of dam body.

The invention discloses a critical rainfall early warning method for mountain torrent disasters in a multi-scenario mode, which comprises the steps of: collecting disaster prevention object data with a small watershed as a unit; determining the soil state in the early stage; determining the type of possible local rainstorms to carry out rainfall time course allocation design; scenarios Mode combination; determine the critical rainfall under different scenario modes; select the corresponding critical rainfall grading warning for different weather conditions. The present invention considers the flash flood disaster outbreak conditions under different scenarios under natural conditions, focuses on the design rain pattern and the influence of the uncertainty of the soil state in the early stage, and combines the measured data of the control section of the small watershed to try to calculate different critical rainfall values, according to the actual situation The corresponding mountain torrent early warning was issued, and through the test of the measured storm and flood process, it was proved that the mountain torrent disaster early warning mode under the multi-scenario mode is feasible, which improved the fit between the mountain torrent disaster early warning and natural conditions, and further improved the mountain torrent disaster early warning. reliability, credibility and applicability.

The invention discloses an energy-saving and power-saving binary output type mountain torrent sensing device which comprises an independent water flow sensor and an independent water level sensor. The water flow sensor consists of a water flow impact sensing device and a tension measuring device, and the water level sensor comprises a water level follow-up device consisting of an outer floater, a crown block and inner and outer connection pull ropes. Short circuit plate electrodes are mounted in the water flow sensor and the water level sensor respectively and provided with electrodes with the minimum dimension delta or delta', measurement binary output range 2N and measurement length output range delta*2N, and all electrodes with equal weights are in short circuit connection. The short circuit plate electrodes in the water flow sensor include a high-order short circuit plate electrode and a low-order short circuit plate electrode. Numbers are outputted according to short circuit relation between short circuit plates and short circuit connecting lines of short circuit plate electrodes. Energy is saved by external arrangement of a rain starting switch device and a regular brief starting switch device. The energy-saving and power-saving binary output type mountain torrent sensing device which adopts binary output is simple and reliable in structure and easy to popularize.

The invention discloses a dynamic risk assessment method for mountain torrent disasters coupled with hyperspectral monitoring and GIS. It includes the following steps: Step 1: Construct a hyperspectral remote sensing technology application system; Step 2: Extract information related to mountain torrent disasters; Step 3: Simulate the evolution and disaster process of single and compound disasters; Step 4: Build a stepwise regression equation ; Step 5: Determine the index system of the risk of mountain torrent disaster; Step 6: Obtain the risk map of mountain torrent disaster and the result map of mountain torrent disaster vulnerability; Step 7: After normalization, obtain the risk map of mountain torrent disaster at multiple spatial scales; Step 8 : To divide the flash flood disaster risk zoning in a specific area. The invention has the advantages of realizing the accuracy, reliability and timeliness of the basic data in areas prone to mountain torrent disasters.

The invention discloses a regional mountain torrent risk prediction method and system, wherein the prediction method includes: S1, predicting the weather information of all hills in the region based on a meteorological model; S2, screening out the mountains that need risk prediction based on the weather information Hill; S3. Divide the hills that need risk prediction into different risk prediction levels according to their basic information; different risk prediction levels correspond to different risk prediction periods; The preliminarily determined disaster-causing factors are screened to obtain the main disaster-causing factors affecting mountain torrents; S6. Training and generating multiple mountain torrent risk prediction models based on the main disaster-causing factors; S7. Selecting several risk prediction models with the best performance to form the final Risk forecasting model; S8. Perform flash flood risk forecasting for the hills to be forecasted. The invention realizes risk prediction for regional hills, realizes low cost, wide coverage and high processing efficiency, and improves the safety of hills.

The present invention provides a method for early warning of torrent and debris flow based on steep slope confluence and cross-section flow calculation, comprising steps: 1. By processing the high-precision DEM terrain data and remote sensing data of the single ditch of the debris flow and its watershed, each grid unit is obtained 2. According to the actual geological conditions of different debris flow ditches, combined with continuous long-duration rainfall data, the rainwater infiltration of each grid unit is obtained; 3. Using the calculation method based on steep slope confluence, calculate the debris flow ditches and The surface confluence of the watershed where it is located; 4. Early warning through the submerged range of mountain torrents and debris flows corresponding to the rainfall in different time periods. A flash flood and mudslide early warning system is also provided. The invention is applied to the early warning of mountain torrents and valley-type mud-rock flows, and has an obvious effect of improving the early warning efficiency for the early formation of mountain torrent mud-rock flows early warning results. It has great practical significance for ensuring personnel safety, emergency avoidance, and reducing the loss of people's lives and property.

The invention relates to the field of geotechnical engineering for geographical environment control, and specifically discloses a torrent and debris flow sand retaining dam with the function of generating electricity and its application. In the present invention, by adding a splitter device upstream of the sand control dam, the turbine water filter net and the debris flow grid in the device can filter and separate water and stones, and can use the separated water for water storage, and use the height difference to make the sand control dam It has the function of generating electricity. The torrent and mud-rock flow sand retaining dam of the present invention not only reduces the disaster of torrent and mud-rock flow, but also utilizes the advantage of the natural height difference of the sand retaining dam to generate power, so that the sand retaining dam has certain economic utility.

The invention discloses a real-time mountain floodv early warning method based on critical rainfall of rainfall time-history distribution. The rainfall data in a T period before a flood peak appears in each previous flooding process are firstly extracted, and the cumulative rainfall in the T period is calculated; then, a Lorenz curve is constructed based on the rainfall data in the T period beforea flood peak appears, the Gini coefficient of the Lorenz curve is calculated, and based on the Gini coefficient, the rainfall in the T period before the flood peak appears is divided into three types; the tension water content saturation, that is, initial soil water content saturation, at the T time before the flood peak appears is then counted; whether each previous flood is an over-alarm floodis judged, cumulative rainfall-initial soil water content saturation scatter diagrams for rainfall in even, uneven, and concentrated types are made respectively, points corresponding to the over-alarmfloods are marked, and a critical rainfall line is drawn in combination of the points corresponding to the over-alarm floods; and real-time early warning is finally carried out according to the critical rainfall line.

The invention discloses a mountain torrent early warning method considering the spatial heterogeneity of rainfall, which belongs to the field of hydrological forecasting. The method includes: acquiring soil saturation and composite rainfall within a set time period before the occurrence of multiple historical flood peaks in the watershed to be early warning; wherein , the combined rainfall takes into account both the amount of rainfall and the spatial distribution of rainfall; construct a mountain torrent early warning model based on the above soil saturation and compound rainfall; collect the soil saturation of the watershed to be warned, and obtain the critical composite rainfall index corresponding to the current soil saturation according to the mountain torrent early warning model, By comparing the current composite rainfall and the critical composite rainfall index in the basin to be warned, it is judged whether the flood caused by rainfall is a super-warning flood. The method of the invention can specifically quantify the unevenness of the spatial distribution of rainfall and the distribution of the center of the rainstorm, and the calculated critical composite rainfall index not only considers the size of the rainfall, but also considers the spatial distribution of the rainfall, thereby improving the accuracy of mountain torrent warning.

The invention provides a real-time accumulated rainfall rainstorm mountain torrent early warning method based on mountainous area small watershed rainfall runoff response. The method comprises the following steps: (1) selecting a small watershed river reach of a mountainous area needing rainstorm mountain torrent early warning as a target river reach, and searching flood characteristic data of n-stage rainstorm mountain torrent of the target river reach and rainfall characteristic data of stage rainstorm corresponding to the n-stage rainstorm mountain torrent; Determining a flood abrupt rise time point of each field rainstorm mountain flood according to the flood characteristic data; (2) analyzing the n times of rainstorm mountain torrent; (3) determining a critical cumulative rainfall index for sending out rainstorm mountain torrent early warning; (4) 1, when rainstorm occurs at the upstream of the target river reach, collecting time period rainfall data of field rainstorm, and calculating accumulated rainfall from the beginning of rainfall to different time period rainfall data collection time points in real time; And 2, when the accumulated rainfall capacity in the step (1) reaches a critical accumulated rainfall capacity index for giving out rainstorm mountain torrent early warning, immediately giving out the rainstorm mountain torrent early warning.

The present invention relates to the field of water conservancy measurement, particularly to a field work measurement method for flash flood disaster investigation and evaluation. According to the field work measurement method, a triangle reference station network is constructed from three reference stations, wherein the distance between the reference stations is 40-60 km; the latitude information, the longitude information and the elevation information of each position in the triangular area are measured by using the reference station network; after the measurement of the one triangular area is completed, the one reference station moves while the other two reference stations do not move so as to construct a triangular reference station network again; and the previous steps are repeatedly performed until the measured triangular areas cover all the areas to be measured. According to the present invention, the triangular mobile networking technology is used, and the CORS network is established; the application radius is 50 Km, and the precision is 1-3 cm; the channel cross section and the topography of the village along the river are efficiently, reliably and economically measured; and the working efficiency is 10-100 times the working efficiency of the traditional measurement.

The invention discloses a school mountain torrent early warning terminal station which comprises a display unit, an alarming unit, an RCU remote control unit, a monitoring computer and a power supply unit, wherein the power supply unit can provide electricity for the early warning terminal station, the RCU remote control unit is connected with the alarming unit and the monitoring computer respectively, and the monitoring computer is connected with the display unit. The school mountain torrent early warning terminal station disclosed by the invention has rich display information types, can refresh and display received possession periphery water level rainfall monitoring station information, possession weather information, mountain torrent early warning information, disaster emergency plan, on-campus registered teacher and student data registration information, notice and announcement information and the like on a display module in real time and enables local early warning objects to comprehensively and timely know local related early warning information.

The invention relates to a method for constructing and updating a mountain torrent disaster database based on web crawler and semantic recognition. It uses web crawler technology and semantic analysis method to grab information web pages and extract key information of mountain torrent disasters, and packs the extracted information into The standard related information is stored in the mountain torrent disaster database, and it is updated and supplemented in real time. This method makes full use of network information resources, supplements and collates the flash flood information at the site, describes and analyzes the occurrence and development process of flood events more comprehensively and accurately, and solves the problem that the current flash flood disaster network data resources are large in number, variety, and source, and it is difficult to accurately The technical problem of efficiently extracting useful data information provides a useful method for the rapid construction and update of the mountain torrent disaster database.

The invention provides a GIS mountain torrent prevention and early warning system based on big data, which is used to solve the problems of limited monitoring range, insufficient reference data and low prediction accuracy of the existing torrent forecasting system, including a cloud computing platform, an environmental monitoring module, and a terrain monitoring module , image processing module, neural network module, data query module, data storage module, alarm driver module and intelligent terminal module; the present invention is provided with a neural network module, which is designed to build a predictive model based on existing data, which can improve the overall Prediction accuracy; the present invention is provided with environmental monitoring module, terrain monitoring module and image processing module, provides abundant data for the present invention, has guaranteed the mountain torrent early warning precision of the present invention; Data storage module is provided in the present invention, and this design is to make up for cloud computing The shortcomings of the platform's storage capabilities make the system more complete and stable.