巴西专利BR102019003180B1 hydrological analysis and management process and system for basins

专利PDF首页>>巴西专利

专利附录

专利说明

权利要求

类似技术

同族专利

引用文献

法律状态

优先权

专利摘要:
HYDROLOGICAL ANALYSIS AND MANAGEMENT PROCESS AND SYSTEM FOR BASINS. Hydrological analysis and management process and system for basins with networks of meteorological stations and artificial drainage systems with management of natural and artificial dams through locks and pumping stations. It assesses the potential for water risk in each area and analyzes a priori, through simulations, the possible consequences of future rainfall. For this simulation, hydrographs are calculated for each sub-basin, canals and rivers in the basin. It simulates the behavior of the basin under different scenarios corresponding to different operation management of the gates and/or pumps, and compares its results in flooded area loss, economic loss of each sector, loss by flooding of inhabited area, etc. The simulation optimization through AI (Artificial Intelligence, meta-heuristic algorithms, neural networks, etc.) allows to serve as a search engine to find better solutions, the best resource management configuration that allows to minimize the socioeconomic impact on the basin.
公开号:BR102019003180B1
申请号:R102019003180-8
申请日:2019-02-15
公开日:2021-05-25
发明作者:Enrique Menotti Pescarmona
申请人:Enrique Menotti Pescamona；
IPC主号:

专利说明:

TECHNICAL FIELD
[001] Hydrological analysis and management process and system through the use of Artificial Intelligence (AI) for basins with networks of meteorological stations and artificial drainage systems with management of natural and artificial dams through locks and pumping stations. The invention combines aspects from different technical fields such as meteorology, hydrology, engineering, process control, Artificial Intelligence and Machine Learning, among others. TECHNICAL STATUS
[002] Existing basic systems for hydrological analysis of a basin work by analyzing atmospheric data results, measuring rainfall by average precipitation, humidity detectors, etc.; flow, temperature, solarimetry and others; which are collected and used to make predictions manually with classical calculation techniques. Data collection is handled by cable by different private and government entities that tend to differentiate between meteorological and hydrological data. There are still, for example in the United States of America, cases in which a single entity records all measured data, whether meteorological or hydrological.
[003] These known systems are based on mathematical models that perform meteorological and basin state forecasts, usually as described below.
[004] A hydrological analysis system for basins with networks of meteorological stations of the current state of the art uses calculation models, among which stands out the meteorological model and the model of the particular basin to be analyzed.
[005] This set of models usually simulates together the three main processes of a flood event, which are precipitation, infiltration and surface runoff, obtaining in a short time information on flooded areas, channel flooding levels, retained and evacuated volumes, and other relevant data. This current technique only receives the information and does not fix the existing problem. In other words, they do not control water flows or generate recommendations or coordination of actions or steps to follow.
[006] Meteorological models are necessary to evaluate and interpolate the precipitation that falls on each of the sub-basins, from the data that are provided by the pluviometric stations. The meteorological information has two main sources, the pluviometers that give information on the amount of water accumulated in time increments and the meteorological radar response. Both sources have to be calibrated and adjusted manually and in situ to obtain a reliable measurement of the output hydrograph.
[007] Known meteorological models assess rainfall over a given area by using the relative position of the respective rain gauges in the area. These areas are called Thiessen polygons (the range of the rain gauge as half the distance between two consecutive rain gauges).
[008] The isohyet method, on the other hand, determines lines of equal height of precipitation.
[009] With both results, these known models calculate, following standard procedures and general recommendations, the precipitation height, temporal distribution of rain and spatial distribution of rain. They provide information about the phenomenon, but do not have an option to solve the problem.
[010] On the other hand, the model of the basin itself, includes a scheme of the drainage network, the existing connections between the different sub-basins, as well as the physiological properties of each sub-basin. This known basin model also comprises the chosen calculation methods and the hydraulic parameters associated with them.
[011] In each sub-basin runoff is calculated by the SCS method (by its acronym in English, Soil Conservation Service) which allows separating surface runoff from infiltrated. Infiltration is a variable that changes according to the characteristics of the terrain, amount of water absorbed, type of vegetation, etc. For this, you must have geophysical information about it.
[012] From the surface current is obtained the hydrograph of the sub-basin. Each sub-basin has calibrated the relation of rainfall with runoff in relation to the CN (curve number) which is a percentage between 0% and 100% of water absorption from the land. In addition, CN depends on pre-rain humidity conditions. There will also be a horizontal input from underground flow, which depends on the slope of the land and the porosity of the land.
[013] As output variables, there are evaporation and evapotranspiration. The first variable measures in mm/time the evaporation occurring in reservoirs or flooded areas. It is measured at recording stations with evaporimeter tanks. On the other hand, the evapotranspiration variable depends on the type of crop/vegetation, which can be an important strategy in this water balance to reduce the permanence times of flooded soils.
[014] From here it is estimated how much runs, its hydrograph from the sub-basin surface to canals and rivers. Adding the contributions of each sub-basin, it is possible to calculate the flooding in channels and reservoirs, according to calculation methods and standard codes for general use.
[015] The possible distribution flows are also calculated. This is usually done in known systems with models that are based on equations such as Chezy/Manning that inversely depend on the Manning coefficient (depends on the roughness/vegetation and can range from 0.013 for channels lined to 0.045 with vegetation) and directly proportional to the root square of the hydraulic radius of the channel (calculated as the quotient between wet surface and perimeter) and the slope of the channel:
[016]

[017] In which
[018] C = Chezy's Roughness Coefficient

[019] R = Radio Hydraulic [m]
[020] n = Manning roughness

[021]

[022] Substituting the roughness calculated in the Chezy equation, the Manning equation is obtained:
[023]

[024] In what
[025]

[026]

[027] S = Tilt Friction

[028] Another variable used in the analysis systems of the State of the Art is the vertical natural output variable, which is percolation in the ground. This will depend on the type of terrain and its saturation level given by the water table. The PULS method is used and the MUSKINGUM method relates the hydrographs at the entrance and exit of each dam/reservoir, excess flow, maximum levels of the dam, size of the excess works and determination of the maximum extraordinary water level. For this, the profiles of natural or artificial channels and the bathymetric profile of the dams must be known, in order to be able to determine the storage capacity and the relation of area vs. reserved volume. In other words, these technical systems depend mainly on the reliability of previously acquired data.
[029] In each part of the current models there are fixed coefficients and other variables that change with time, such as ambient moisture, soil moisture, change in the evapotranspiration coefficient by changing the sown crop, solar radiation, etc. The model is parameterized with these variables. Some of these variables are direct or indirect measures. The current model of a basin with many sub-basins and reservoirs has an order of hundreds of parameters, depending on the size of model discretization.
[030] A common aspect in the hydrological analysis systems of the prior art is that they comprise a control center. In this control center, some data such as water volumes and where excess water volumes must be mobilized are known to obtain control over floods, but they cannot make advances, because there are no solutions for these problems.
[031] Currently available tools issue a reserved prediction that does not have an associated mitigation strategy. They warn about the hydrological problem without intervening in solutions with a socioeconomic impact. These alerts are disclosed to interested organizations (Stake Holders) and the solutions are implemented in isolation, without optimizing the resources available in the basin in question. Similarly, there are water volume management systems in basins that do not have intelligence associated with the forecasts received. In general, the implementation of an integral system that would solve these multidisciplinary problems would require a large number of qualified and prepared people to take decisions.
[032] The patent US 9,726,982 B2 is also known from the State of the Art, which seeks to solve similar problems by implementing methods, systems and storage devices for computer programs to generate a response to a flood.
[033] This invention describes the implementation of two general subsystems. A computer-implemented subsystem to generate a hydro-meteorological forecast of floods, with feedback from meteorological information and data from ground stations. A decision support subsystem, which generates a report with recommendations and feedbacks on the quality of measurements to the previous subsystem. As an example, the patent describes the procedure for detected and validated flood events. The system can report and perform the following:
[034] a) Location of resources.
[035] b) Process the information from the decision support subsystem to generate a list of recommendation actions in case of flooding.
[036] c) Provide a list of actions for responding to floods that include: communicating the message of recommendations to all agents involved in the flood; evacuation of the population; secure transport routes; secure critical supplies; isolation of affected infrastructure, sending qualified people and teams to key locations.
[037] d) Continuously communicate the actions taken and the progress of the event.
[038] However, it does not result from this priority as both the calculation and the evaluation of estimates of the hydrometeorological event as the hydrographs are performed, because the method used for the forecast could have greater error.
[039] Neither does this priority include the integration of reservoir control over mitigation and integral planning of the basin, such as floodable areas, areas of greater productivity, critical areas, etc. Nor does it use aerial and satellite images to this effect to estimate flooded areas, water volumes, soil quality and vegetation information, etc.
[040] Finally, the State of the Art has not yet provided an intelligent system that works online with machine learning tools that can be used to feed back forecasts and dynamically adjust event estimates and hydrological models, in which the constant System learning can strengthen decision-making in future events, depending on the continued use of the system, to solve problems quickly and improve the situation of populations in areas of implementation.
[041] Nor has the State of the Art provided a hydrological analysis and management system that uses aerial and satellite images to estimate flooded areas, water volumes, soil quality and vegetation information; not even an intelligent system that encompasses the possible generation of electrical energy within an integral plan, taking advantage of artificial reservoirs in a distributed and optimized way, with electric pumps and turbine pumps. BRIEF DESCRIPTION OF THE INVENTION
[042] It is an object of the present invention to provide a process and system for hydrological analysis and management for basins with networks of meteorological stations and artificial drainage systems with management of natural and artificial dams through locks and pumping stations, in which the potential water threat in each area and the possible consequences of future rainfall are analyzed a priori through previous simulations carried out using an Artificial Intelligence system; in which, for said simulation, the calculation of the hydrographs for each sub-basin, channels and rivers of the basin in question is carried out and subsequently the behavior of the basin is simulated under different scenarios corresponding to different controls of operations of the gates and/or pumps and their results are compared in terms of loss of floodable area, economic loss of each sector, loss due to flooding of inhabited area, etc.; in which the simulation optimization through Artificial Intelligence (meta-heuristic algorithms, neural networks, etc.) allows its use as a search engine to find the best possible solutions and the best resource management configuration to allow minimizing the socioeconomic impact in the basin before the hydrological phenomenon.
[043] It is another objective of the present invention to provide a hydrological analysis and management system whose resolution consists of the analysis of a mass balance originated by the input and output volume, which allows retaining said volume in each cell under study into which it is subdivided the system; in which said volumes of water can be mobilized to the destination in order to obtain flood control, in accordance with the chosen protection criteria, through a control and decision-making center that quickly has all the information of online form, as long as you order and direct the activation of gates and pumping stations available for drainage control.
[044] It is another object of the present invention to provide a process and system of analysis and hydrological management for basins in which it is possible to generate, by Artificial Intelligence, strategies to protect the populations, the most productive areas, the best lands and the strategic places of the basin, allowing the storage of water in areas not useful for later use as for irrigation against future droughts and also allowing, through the management of available water resources, to obtain a by-product of strategic importance such as the generation of electricity, using systems conversion advances such as low-speed generators and by-river turbines. BRIEF DESCRIPTION OF THE DRAWINGS
[045] Figure 1 is a basic schematic of the invention, showing mainly the collection and management of data.
[046] Figure 2 is a simplified flow diagram that shows the distinct steps of the claimed process.
[047] Figure 3 shows a simplified scheme of the generation of hydrographs as input variables to the management subsystem.
[048] Figure 4 represents a Linear Recurrent Neural Network scheme.
[049] Figure 5 shows a scheme of a three-layer LDDN network.
[050] Figure 6 is an example of a flood map as a result of the artificial neural network prognosis.
[051] Figure 7 shows a scheme of input layers to the decision system assisted by Artificial Intelligence, and
[052] Figure 8 represents a scheme that reveals the steps and combination of means for managing water volumes, according to the claimed invention. DETAILED DESCRIPTION OF THE INVENTION
[053] The invention in its preferred mode of realization comprises a process and system of analysis and hydrological management through the use of Artificial Intelligence (AI) for basins with networks of meteorological stations and artificial drainage systems with management of natural dams and artificial through locks and pumping stations, which provide real-time responses to the problem as it occurs, thus optimizing flood control in the basin to which the invention applies. It must be understood by real time, the response time of the system, within the order of time of evolution of the meteorological phenomenon.
[054] Both the process and the system combine means from different disciplines. Through the analysis of meteorological events and their hydrological consequences, an integral hydrological model is trained that is capable of describing the hydrological state of the basin and making forecasts of currents, floods and droughts. This model, comprised of a large number of multi-parameter equations, calculates scenarios of the hydrological state of the basin for different meteorological forecasts. In this way, you can prepare a large number of situations, prior to their appearance. Based on the occurrence of a specific weather event, the system automatically selects the calculated situation that best describes what was measured. In this way, in just a few hours, you have a hydrological forecast.
[055] The training of the supposed scenario generates as a result basin behavior curves. These curves are the transfer function (FT) of the system between the input (rain and current situation of the basin) and the output (post-rain scenario). It is a non-linear function that depends on hundreds of factors, some stochastic and others deterministic that vary with time, humidity, weather conditions, temperature, etc.
[056] BEHAVIOR CURVE

[057] The behavior curve is a complex linkage of H=f hydrographs (surface, DEM, soil structure, etc.), E=f runoff (H, infiltration, moisture, soil, etc.) and A=f floods ( E, Bathymetric profile), level of reservoirs NR=F (Volume, Qe inflow, Qs outflow, evaporation) of each of the sub-basins, flows, reserves and channels of the basin, and that is why the training of the supposed scenario generates hundreds of behavior curves.
[058] Thus, when an occurrence resembles the supposed rainfall scenario, the system already recognizes the expected behavior of the basin instantly. The hydrological analysis and management system is configured to generate an initial alert, and notify the community. The information delivered will depend on the forecast conditions in each region of the basin. Relevant bodies (Stake Holders) will be delivered to the interested party through automatic applications via cell phone, emails and "smart" applications or something similar, together with a reservoir control management plan, which determines, for example, the pumping of water from one sub-basin to another.
[059] While this occurs, the operating status of the drainage system is fed back to the expert system using the measurement stations on land, which, through a set of sensors, instruments and computers that have a closed communication system, permanently reclassify the analysis conditions for the purpose of maintaining the basin's mass balance, ie, pumping and closing or opening the doors to carry water from one reservoir to another.
[060] Over time, more information, trends and behavior curves of the watershed, for different rainfall scenarios, are collected in the databases. Furthermore, each behavior curve is parameterized with the variables of the models of each sub-basin. These parameters change over time, such as soil structure, seeded fields, temperature, humidity, etc. This growing database feeds AI-assisted systems to anticipate the system about deterministic and stochastic events.
[061] The model calibration is continuous and is carried out for each sub-basin in particular and proximity using adaptive filtering adjustment algorithms with the measured data, which will be detailed later.
[062] As it is an AI and Machine Learning system, the response time and accuracy of the results improve with the use and learning that is obtained as the experience is added.
[063] The hydrological analysis and management process in its preferred mode of execution comprises three fundamental steps for its development, as illustrated in the simplified diagram in Figure 2.
[064] The first step involves measuring and collecting historical data, in the second step, forecasting systems are distinguished using machine learning methods and water modeling, and said forecasts must be validated by the information collected in the monitoring stations of channel flow and reservoir level, while in the third phase, mitigation strategies are generated to minimize the impact of the meteorological event on the information bank collected by the forecasting sub-module and its constant validation, added to the socioeconomic data of production of the basin, dynamically collected.
[065] A detailing of each of the said phases is developed below, according to the preferred mode of execution of the invention.
[066] The first stage of the process is intended to measure and collect historical data, such as the characterization of the basin, with an emphasis on the logistics of data management and distribution of monitoring stations. The constant support must guarantee the online functioning of this first stage, which feeds the later stages, and that is why its importance in the collective performance of the process is essential. A schematic of this first step is illustrated in figure 1.
[067] Distributed in the basin and in areas of contribution to the basin, meteorological stations are installed with data acquisition and online data transmission systems, via satellite, which communicate with the Acquisition Center and Database. Some of the sensors involved in the system are:
[068] Sensor 1: Rain sensor.
[069] Sensor 2: Evapotranspiration Sensor.
[070] Sensor 3: Wind speed.
[071] Sensor 4: Ambient and soil moisture.
[072] Sensor 5: Ambient temperature.
[073] Sensor 6: Flow meters in canals and rivers.
[074] Sensor 7: Level height of ponds, dams, canals and rivers.
[075] Sensor 8: Permeability and percolation.
[076] Other entries to the database are information via satellite, which is obtained through an automatic system access tool on the server or database of the company that provides the service.
[077] These satellite data are used to indicate the state of the basin as a whole, estimate flooded areas, dry areas and monitoring and storm characteristics. They also allow for the classification of crops to estimate the degree of water consumption according to their characteristics.
[078] Some of the satellite data are images. Two types of satellite images can be used:
[079] - SAR (Synthetic Aperture Radar) images to monitor the flooded area, and
[080] - RGB and Multispectral Images, for the calculation of Green Index, crop classification, etc.
[081] In the second stage of the hydrological analysis and management process, according to the preferred mode of execution of the invention, the forecasting systems using machine learning methods and the water modeling are distinguished. The forecasts are validated by the information collected at the channel flow and reservoir level monitoring stations.
[082] In this second stage of the process, it is essential to activate the forecast or forecaster module.
[083] The predictor module consists of training a dynamic artificial neural network ANN (Artificial Neural Network) to generate predictions of possible flooded areas. These types of networks, with recurrent topology, use pre-processed data collected by stations on land, via satellites, radar and information from distributed meteorological systems, as well as the network forecasts carried out on past occasions. Dynamic networks are generally more powerful than static networks (with greater training difficulty). When memory is available, dynamic networks can be trained to learn sequential or time-varying patterns. This feature has applications in different areas such as forecasting financial markets, equalization of communication channels, phase detection in power systems, fault detection, language recognition and weather forecasting, among others.
[084] To be able to predict a temporal pattern, an ANN requires two distinct components: a memory and an associator.
[085] The memory is generated by a delayed time unit (shift register) that constitutes the "tapped delay line" (TDL) and stores relevant previous information, used to improve the prediction.
[086] The associator will be, for example, a multilayer perceptron network, efficient for non-linear complex static mappings. A network topology that complies with these statements is the so-called "Layered Digital Dynamic Network" (LDDN). Each layer of the network is comprised of the following parts:
[087] - A set of synaptic weight matrices that enter the layer (from which it can be connected from other layers or external inputs), association rules of the weight functions used to combine the weight matrices with the associated inputs and TDL ,
[088] - A bias vector -bias -,
[089] - Rules for the input functions to the network used to combine the outputs of the weight functions with the bias vector to produce the input to the network, and
[090] - A transfer function.
[091] An example of a multilayer LDDN network is illustrated in Figure 5, in which a scheme of a network of three layers (layers) is shown, where the variable LW means the weight of the hidden layers, IW is the weight in the layer of input, B represents the bias unit and f is the interlayer transfer function.
[092] It will then be necessary as many networks as areas to carry out prognosis. Each network must be trained with the available variables of the area and with previously revealed data, determining the existence or not of flooded terrain with water level.
[093] The training of this type of network requires highly skilled personnel in hydrology, neural networks, machine learning, computer systems and associated sciences such as geology, topology, mechanics, electronics and communications, among others; in addition to needing a significant number of computers, so that, connected to each other in vector form, they allow, in a short time, through this analysis, to know the best solutions and with less socioeconomic impact to make the most appropriate decision.
[094] An analysis should be performed to determine the best parameters to improve classification and prognosis. This work is optimizable by parallel computing methods. For each moment analyzed, a map with the flood forecasts is obtained, as shown in Figure 6. The output of the general forecast will then be a mapping of the basin with the flood forecasts. This result will be used to calculate the "basin hydrograph", also taking into account the topographic data of the same. The calculated flows will be validated and this information will be fed back to the entire chain, with the flow measurement points at key points in the basin. In this way, the model and prediction network are constantly updated to improve predictions.
[095] The use of computational complex physical models of the dynamic behavior of the basin serves as a training tool for the prediction system. The multivariable model, optimized with the experience collected, allows the training of the predictor in cases not previously registered, allowing a low response time, due to the fast calculation of neural networks. Thus, the integral forecasting tool involves the calculation times of the physical model and the network's own training.
[096] Finally, in the third stage of the hydrological analysis and management process, according to the preferred mode of execution of the invention, mitigation strategies are created to minimize the impact of the meteorological event due to the information collected by the forecasting sub-module and its constant validation, added to the socioeconomic data of the basin's production, collected dynamically.
[097] The evaluation software is based on a Geographical Information System GIS (Geographical Information System), which uses as input the result of the flood forecast, geographically divided into cells. Each cell contains the water level information present in this assessment area. In this way, the event is analyzed in the form of layers (rasters), which, overlaid, indicate the problem areas and possible areas to carry out the action maneuvers.
[098] The evaluation layers in each cell have an incidence weight value. For example, if the layer represents population density, in the agricultural area the value will be close to zero, while in the urban area it will be much higher.
[099] Furthermore, according to the incidence weight that each layer has, it will be the relevance given to the event. For example, if less weight is given to the livestock activity compared to the agricultural activity, the system will tend to flood more to the agricultural activity area, based on the fact that it will be easier to take the cattle to another higher or safer area than moving a seeded field. An example of this analysis in the form of layers is illustrated in the schematic in Figure 7.
[0100] In this way, each layer will be weighted by a weight, related to socioeconomic priorities.
[0101] The decision module assisted by AI with vector optimization algorithms MCDA (English Multi-Criteria Decision Analysis) is in charge of finding the flood areas, which minimize the socioeconomic impact of the basin. This decision module comprises a distributed and multitasking computing center.
[0102] The result is a single raster with water levels product of the overlap of all input layers. With this result, the decision module performs the simulation of the operation of gates and pumps necessary for the basin to reach the required state in a given time.
[0103] Once it has been determined which volumes of water must be mobilized and where, to achieve flood control in accordance with the chosen protection criteria, the operation report is generated in a control center, from which it is ordered and commands the activation of gates and pumping stations available to control drainage.
[0104] The control center respects a standardized standard, in accordance with international standards of human-machine interface for safety and reliability. In the preferred execution mode, the control center is made up of three sectors to be defined according to the application, but always maintaining the highest security standards.
[0105] A sector is the normal line and backup power supply or uninterrupted battery support. The second sector is the air conditioning equipment room for environmental adjustment, according to the equipment used and the conditions of the area, in the sense that the third sector is comprised of two parts: the system operating room, which is subdivided into display, online calculation, algorithms and modeling; and the server room that contains the communications system.
[0106] In said control and decision-making center, all information will be made available online so that decisions can be taken quickly.
[0107] In this way, strategies will be generated to:
[0108] - Protect populations.
[0109] - Protect the most productive areas.
[0110] - Protect the best, most productive lands and strategic points in the basin.
[0111] - Store water in areas not usable for use in irrigation in future droughts.
[0112] Another advantage of the present invention, compared to the known state of the art, is that the management of available water resources allows the obtainment of a by-product of strategic importance, such as the generation of electricity, using, for example, advanced systems such as low-speed generators and by-river turbines.
[0113] The resolution of the process set in motion by the system, according to the preferred execution mode of the invention, consists of the analysis of a mass balance originated by the input and output volume and retained in each cell under study in which the system is subdivided. Thus, in the domain of each cell, the system solves the following equation:
[0114]

[0115] Where:
[0116] "i" is the cell number that corresponds to the subdivision of the adopted system.
[0117]
is the input volume in cell "i".
[0118]
is the output volume of cell "i".
[0119]
is the volume held by cell "i".
[0120] The choice of the number and surface of each cell under study is defined as a function of the necessary precision of the variables to be determined in each sector of the system. Thus, each system will be divided into cells of larger or smaller size, densifying the number of cells depending on the economic, social and environmental importance of the area under study.
[0121] The input volume v} is determined by the vertical and horizontal inputs. The main vertical input is precipitation, while the horizontal one corresponds to the output volume Vl'+'li of adjacent cells, which is caused by a horizontal flow on the surface and another underground.
[0122] Therefore:
[0123]

[0124] Where:
[0125] V1P is the input volume per precipitation.
[0126]
is the surface volume that enters the cell "i" under study, coming from an "i+n" cell adjacent to the one under study.
[0127]
is the underground volume that enters the cell "i" under study, coming from an "i+n" cell adjacent to the one under study.
[0128] The output volume of each cell originates from vertical and horizontal output volumes.
[0129] Vertical volumes are those caused by evaporation from liquid surfaces as a result of solar irradiation and temperature (mm/time), evapotranspiration (depending on natural vegetation or cultivation) and infiltration (variable with the type of soil and degree of saturation and groundwater level of the existing terrain in each cell under study).
[0130] Horizontal volumes, on the other hand, are determined by the output volume in a given period of time (flow) possible to transfer to adjacent cells through natural or artificial channels.
[0131] The difference between the input V} and output volume of each cell can be positive or negative. If it is positive, that is, the input volume is greater than the output, a volume retained in the cell under study originates, which can be housed or dammed in a controlled manner by means of natural or artificial reservoirs, or in an uncontrolled way, producing unwanted flooding in areas of the cell.
[0132] The dammed volume can be progressively dislodged over time through the distribution of artificial reservoirs, that is, openings of floodgates or spillways, or through pumping stations, located in natural reservoirs defined for this purpose, in this way distributing over time the output volume of each cell for the purposes of alleviating or preventing flooding in cells located downstream.
[0133] If the balance is negative, that is, if the output volume is greater than the input volume V} in periods of low rainfall, then the volume retained in natural or artificial reservoirs, both superficial and underground, to try to balance the equation and solve periods of drought. The management of water volumes in the different reservoirs is carried out by managing the pumps and gates available and interconnected in the basin.
[0134] The foregoing includes examples of one or more execution modes. Of course, it is not possible to describe each of the possible combinations of components or methodologies in order to describe the aforementioned modes of execution, but a mid-level person skilled in the art on the subject may recognize that many other combinations and permutations of various modes of execution are possible. execution. Therefore, the described execution modes are intended to cover the entirety of said changes, modifications and variations that are within the scope of the claims that follow below.

权利要求:
Claims (16)
[0001]
01. A process of hydrological analysis and management for basins, characterized by comprising the steps of: measuring and collecting historical data through meteorological stations previously installed and distributed in the basin, said stations equipped with a series of sensors and means for data acquisition and transmission systems capable of transmitting said data online to an acquisition center and database, predicting the behavior of the basin in the face of a meteorological event through a predictor module comprising a dynamic artificial neural network trained to generate predictions based on the data pre-processed data collected by nearby weather stations and the neural network's own predictions for previous events, and generating and providing mitigation strategies to minimize the impact of the weather event based on the collected information, where generating such strategies comprises: evaluating the data by through a geographic information system that divides the AV area. ally in cells according to the flood forecast and the information on the water level present in said area, analyze the event in the form of superimposed layers, with different values of incidence weight of each determined by socioeconomic information, obtain as a result, by through an AI-assisted decision module with MCDA vector optimization algorithms, a single layer with water levels product of the overlap of all analyzed layers that minimize the socioeconomic impact of the basin, and perform the necessary gates and pumps operation simulation for the basin to reach the required state in a given time, establishing the volumes of water that must be mobilized and in which direction, in order to obtain control over the flood in accordance with the chosen protection criteria; and in which providing said mitigation strategies comprises: generating the operation report and sending it online to a control center, which orders and commands the activation of gates and pumping stations available for drainage control.
[0002]
02. Hydrological analysis and management process for basins according to claim 01, characterized in that the step of measuring and collecting historical data is implemented by means of at least pluviometry, evapotranspiration, wind speed, ambient humidity and soil, ambient temperature, flow meters in canals and rivers, height of the level of lakes, dams, canals and rivers and permeability and percolation, installed in meteorological stations.
[0003]
03. Process of analysis and hydrological management for basins according to claims 01 and 02, characterized in that the step of measuring and collecting historical data is also implemented through satellite images that allow viewing the state of the basin as a whole. images be SAR (Synthetic Aperture Radar) or RGB and multispectral type.
[0004]
04. Hydrological analysis and management process for basins according to claim 01, characterized by the step of predicting the behavior of the basin in the face of a meteorological event, comprising the previous step to train the dynamic neural network of the predictor module to learn sequential or variant patterns in the for this purpose, said network comprising at least one memory and one associator.
[0005]
05. Hydrological analysis and management process for basins according to claims 01 and 04, characterized by the step of predicting the behavior of the basin in the face of a meteorological event, also comprising the step of issuing a mapping of the basin with flood forecasts through the forecast module, and subsequently perform the calculation of the hydrograph of the basin under study, also using its topographic data, comprising the steps of validating the calculated flows and feeding back the entire chain with the flow measurement points at key points of the in order to constantly update the forecast model and the network to improve future forecasts.
[0006]
06. Hydrological analysis and management process for basins according to claim 01, characterized by the step of generating and providing mitigation strategies to minimize the impact of the meteorological event based on the collected information, in which generating such strategies comprises: evaluating the data through a geographic information system that divides the assessment area into cells according to the flood forecast and the information on the water level present in said area, analyze the event in the form of superimposed layers, with different weight values of incidence of each determined by socioeconomic information, obtain as a result, through an AI-assisted decision module with MCDA vector optimization algorithms, a single layer with water levels product of the overlap of all analyzed layers that minimize the socioeconomic impact of the basin, and perform the simulation of the operation of gates and pumps necessary for the basin to reach the required state. taken in a given time, establishing the volumes of water that must be mobilized and in which direction, in order to obtain control over the flood in accordance with the chosen protection criteria; and in which providing said mitigation strategies comprises: generating the operation report and sending it online to a control center, which orders and commands the activation of gates and pumping stations available for drainage control.
[0007]
07. Hydrological analysis and management process for basins according to claim 01, characterized by the step of generating and providing mitigation strategies to minimize the impact of the meteorological event, also comprising the step of analyzing a mass balance caused by the volume of water from input, output and retained in each cell under study in which the evaluation area was subdivided, in which the domain of each of these cells solves the following equation:
[0008]
08. Hydrological analysis and management process for basins according to claim 07, characterized in that the inlet volume is calculated from the following equation:
[0009]
09. Hydrological analysis and management process for basins according to claim 07, characterized in that the output volume of each cell originates from vertical and horizontal output volumes, in which said vertical volumes are originated by the evaporation of liquid surfaces resulting from irradiation solar and temperature, by evapotranspiration and infiltration; while horizontal volumes are determined by the output volume in a given period of time (flow) possible to transfer to adjacent cells through natural or artificial channels.
[0010]
10. Process of analysis and hydrological management for basins according to claim 01, characterized by the management of available water resources, allowing the generation of electrical energy using advanced conversion systems, such as low-speed generators and passing river turbines.
[0011]
11. Hydrological analysis and management system for basins to implement the process of claim 01, characterized by comprising: a series of meteorological stations distributed in the basin configured to acquire data and an online data transmission system via satellite that communicates with an acquisition center and database; a predictor module comprising a dynamic ANN artificial neural network, trained to generate predictions of possible flooded areas by means of previously collected pre-processed data, wherein said prediction conforms to the basin mapping mode; a computer assessment device configured to: receive the mapping generated by the forecaster, divide the assessment area into cells, where each cell contains the water level information in each assessment area, and analyze the event in the form of superimposed layers, with different incidence weight values for each determined by socioeconomic information; an AI-assisted decision module with MCDA vector optimization algorithms configured to: find the flood areas that minimize the socioeconomic impact of the basin, define as a result a single layer with water levels resulting from the overlap of all input layers, and perform the simulation of the operation of gates and pumps necessary to bring the basin to the required state in a previously determined time; and a control center equipped with means of permanent energy supply, servers with a communication system and a series of computers configured to have all the information online, generate the operation report and order and direct the activation of floodgates and pump stations available for drainage control; and a series of gates and pumping stations arranged across the basin for this purpose.
[0012]
12. Hydrological analysis and management system for basins according to claim 11, characterized in that the meteorological stations distributed in the basin comprise at least sensors for rainfall, evapotranspiration, wind speed, ambient and soil moisture, temperature environment, flow meters in canals and rivers, height of the level of lakes, dams, canals and rivers and permeability and percolation.
[0013]
13. Hydrological analysis and management system for basins according to claim 11, characterized in that this predictor module with a dynamic ANN artificial neural network comprises a memory and an associator, wherein said memory is generated by a delayed time unit (register of displacement) that constitutes the "tapped delay line" (TDL) and stores relevant previous information, used to improve the prediction, while the associator can be, among others, a multilayer network of the perceptron type.
[0014]
14. Hydrological analysis and management system for basins according to claim 13, characterized in that each layer of the network comprises at least one set of synaptic weight matrices that enter the layer, association rules of the weight functions used to combine the weight matrices with the inputs and associated TDL, a bias influence vector, rules for the input network functions used to combine the outputs of the weight functions with the bias vector to produce the input to the network, and a function of transfer.
[0015]
15. Hydrological analysis and management system for basins according to claim 13, characterized in that this forecaster module is configured to be trained by means of a multivariable model.
[0016]
16. Hydrological analysis and management system for basins according to claim 11, characterized in that said AI-assisted decision module comprises a distributed and multitasking computing center.

类似技术:

公开号 | 公开日 | 专利标题

BR102019003180B1|2021-05-25|hydrological analysis and management process and system for basins

Krajewski et al.2017|Real-time flood forecasting and information system for the state of Iowa

Merkuryeva et al.2015|Advanced river flood monitoring, modelling and forecasting

Looper et al.2012|An assessment of distributed flash flood forecasting accuracy using radar and rain gauge input for a physics-based distributed hydrologic model

Mujumdar et al.2012|Floods in a changing climate: hydrologic modeling

Kazeminezhad et al.2005|Application of fuzzy inference system in the prediction of wave parameters

Karimi et al.2013|Neuro-fuzzy and neural network techniques for forecasting sea level in Darwin Harbor, Australia

Hung et al.2009|An artificial neural network model for rainfall forecasting in Bangkok, Thailand

Coulibaly et al.2005|Downscaling precipitation and temperature with temporal neural networks

Levy2005|Multiple criteria decision making and decision support systems for flood risk management

Pradhan et al.2011|A 100‐year maximum flood susceptibility mapping using integrated hydrological and hydrodynamic models: Kelantan River Corridor, Malaysia

Bruen et al.2010|Visualizing flood forecasting uncertainty: some current European EPS platforms—COST731 working group 3

Khatami et al.2014|Benefits of GIS application in hydrological modeling: A brief summary

Ghose et al.2013|Prediction and optimization of runoff via ANFIS and GA

Ebrahim et al.2013|Downscaling technique uncertainty in assessing hydrological impact of climate change in the Upper Beles River Basin, Ethiopia

Huang et al.2016|Spatial combination modeling framework of saturation-excess and infiltration-excess runoff for semihumid watersheds

Van Leeuwen et al.2016|Physically based hydrological modelling of inland excess water

Pan et al.2017|Estimating root zone soil moisture at continental scale using neural networks

Liu et al.2021|Mapping the risk zoning of storm flood disaster based on heterogeneous data and a machine learning algorithm in Xinjiang, China

Endreny et al.2000|Satellite‐derived digital elevation model accuracy: hydrological modelling requirements

Nam et al.2015|Future fluvial flood risks in C entral V ietnam assessed using global super‐high‐resolution climate model output

Cloke et al.2013|Operational European flood forecasting

Subramanian2012|Implementation of neural networks in flood forecasting

Hapuarachchi et al.2008|A review of methods and systems available for flash flood forecasting

Dewelde et al.2014|Real-time flood forecasting systems in Flanders

同族专利:

公开号 | 公开日

PT3534187T|2022-01-25|

JP7001626B2|2022-01-19|

AR109623A1|2019-01-09|

US20190354873A1|2019-11-21|

US20210326715A1|2021-10-21|

CO2018010363A1|2019-09-30|

BR102019003180A2|2019-09-10|

ES2894877T3|2022-02-16|

JP2019194424A|2019-11-07|

EP3534187A2|2019-09-04|

EP3534187A3|2020-07-01|

EP3534187B1|2021-10-20|

引用文献:

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

WO2006002687A2|2004-06-30|2006-01-12|Swiss Reinsurance Company|Method and system for automated location-dependent recognition of flood risks|

US7136756B1|2004-11-02|2006-11-14|Vieux And Associates, Inc.|Method for determining runoff|

JP5289474B2|2005-02-25|2013-09-11|株式会社東芝|Rainwater drainage support control device|

MX2012008316A|2010-01-19|2012-10-05|Swiss reinsurance co ltd|Method and system for automated location dependent natural disaster forecast.|

JP6207889B2|2013-06-21|2017-10-04|株式会社東芝|Inundation prediction system, inundation prediction method and program|

KR20160046016A|2014-10-17|2016-04-28|삼성디스플레이 주식회사|Maskless exposure device and method for compensating cumulative of illumination using the same|CN107292527B|2017-06-30|2020-07-31|安徽工业大学|Urban drainage system performance evaluation method|

CN110232471B|2019-05-08|2021-07-02|中国水利水电科学研究院|Rainfall sensor network node layout optimization method and device|

CN110599003B|2019-08-22|2022-03-04|武汉大学|Flood storage and detention distinguishing flood distribution method based on complete information dynamic non-cooperative master-slave game|

CN110717232B|2019-09-11|2020-09-29|中国水利水电科学研究院|Sub-basin confluence simulation method considering influence of terrace check dam|

WO2021071985A1|2019-10-07|2021-04-15|Rsa Protective Technologies, Llc|Surface water management system including weir wall and detention pond|

CN111027175B|2019-11-06|2021-09-24|中国地质大学（武汉）|Method for evaluating social and economic influences of flood based on coupling model integrated simulation|

US20210230016A1|2020-01-29|2021-07-29|EmNet, LLC|Systems and methods relating to effective management of fluid infrastructure|

CN111461421B|2020-03-28|2021-02-09|华中科技大学|Cascade reservoir risk assessment method and system based on mutual feedback relationship analysis|

CN111539597B|2020-04-01|2022-02-22|河海大学|Gridding drainage basin social and economic drought assessment method|

CN111967763B|2020-08-17|2021-05-14|水利部交通运输部国家能源局南京水利科学研究院|River and lake water system communication water safety guarantee demand adaptation method and system|

CN112649898A|2020-11-11|2021-04-13|中国气象局广州热带海洋气象研究所（广东省气象科学研究所）|Weather drought refined monitoring method|

CN112700141B|2020-12-30|2021-11-19|哈尔滨工业大学|Online analysis method for municipal drainage pipe network|

CN113155102B|2021-02-24|2021-11-19|时代云英科技有限公司|River comprehensive treatment system|

CN112906252B|2021-04-27|2021-12-07|中国水利水电科学研究院|Sub-basin dividing method considering lake reservoir range|

CN113392489A|2021-06-07|2021-09-14|广州珠科院工程勘察设计有限公司|Calculation method of distributed hydrodynamic model of river system|

CN113641733B|2021-10-18|2022-01-07|长江水利委员会水文局|Real-time intelligent estimation method for river cross section flow|

法律状态:
2019-09-10| B03A| Publication of a patent application or of a certificate of addition of invention [chapter 3.1 patent gazette]|

2021-05-04| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|

2021-05-25| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 15/02/2019, OBSERVADAS AS CONDICOES LEGAIS. |

优先权:

申请号 | 申请日 | 专利标题

ARP180100382A|AR109623A1|2018-02-16|2018-02-16|PROCESS AND SYSTEM OF ANALYSIS AND HYDROLOGICAL MANAGEMENT FOR BASINS|

AR20180100382|2018-02-16|

[返回顶部]