Reservoir operation and maintenance and risk early warning method and system
By acquiring reservoir data and utilizing simulation analysis and crack prediction models, the problem of lagging monitoring of cracks in reservoir dams was solved, enabling timely early warning and ensuring reservoir safety.
Patent Information
- Application Number
- CN202510914608.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-03
- Publication Date
- 2025-11-21
AI Technical Summary
Existing technologies cannot provide early warning information before cracks appear in reservoir dams, resulting in monitoring delays.
By acquiring basic data, runoff data, and precipitation data of the reservoir, and using simulation analysis models and crack prediction models, the future deformation and cracking of the dam are predicted, and the results are compared with warning thresholds to issue early warning information.
This enables timely warnings before cracks occur, ensuring the safe operation of the reservoir.
Smart Images
Figure CN120994988A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of reservoir monitoring, in particular to a reservoir operation and risk early warning method and system. BACKGROUND
[0002] A reservoir is an important water conservancy project. By building a dam on a river, water can be collected in a certain area, thereby forming a reservoir. Reservoirs have the functions of reducing flood peaks, storing water for irrigation, generating electricity, etc., and play a very important role in economic development and people's safety.
[0003] A dam is an important part of a reservoir. For medium and large reservoirs, a concrete dam is generally used, while small reservoirs use earth dams. Whether it is a concrete dam or an earth dam, under the influence of factors such as geological subsidence, water level changes, and temperature changes, cracks may occur. Cracks occurring on the dam are divided into surface cracks and internal cracks, and surface cracks are further divided into dry shrinkage cracks, transverse cracks, and longitudinal cracks, etc. due to different factors. Regardless of the cause of the crack, the reservoir management personnel need to be informed in the first time so that appropriate preventive measures can be taken according to the specific circumstances of the crack to avoid serious consequences caused by the expansion of the crack.
[0004] Currently, the monitoring of cracks is mainly carried out by crack meters or manual patrols. Crack meters are installed in positions where cracks may occur or have already occurred through pre-embedding or post-installation to monitor the occurrence and evolution of cracks, while manual patrols mainly involve on-site viewing by personnel or using devices such as drones to visually check the condition of the cracks. Regardless of which method is used, it can only be monitored after the crack occurs, and this monitoring method is lagging behind and cannot provide warning information in advance of the occurrence of the crack. SUMMARY
[0005] The present application provides a reservoir operation and risk early warning method and system to solve the problem that warning information cannot be obtained in advance of the occurrence of a crack in the prior art.
[0006] In one aspect, the present application provides a reservoir operation and risk early warning method, comprising:
[0007] obtaining basic data, confluence data, and precipitation data of a reservoir;
[0008] predicting predicted water level data at multiple different times in the future according to the basic data, the confluence data, and the precipitation data;
[0009] introducing each predicted water level data into a simulation analysis model to obtain deformation data of the dam;
[0010] obtaining structure data of the dam, and selecting a crack prediction model matching the structure data;
[0011] input the deformation data and the structure data into the crack prediction model to obtain a crack prediction result of the dam, the crack prediction result indicating a position, a length and a depth of the crack;
[0012] compare the crack prediction result with a warning threshold prepared in advance, and if the crack prediction result exceeds the warning threshold, issue a warning information.
[0013] In another aspect, the application also provides a reservoir operation and risk warning system, comprising:
[0014] a data acquisition module configured to acquire basic data, confluence data and precipitation data of the reservoir;
[0015] a water level prediction module configured to predict predicted water level data at multiple different time instants in a future period according to the basic data, the confluence data and the precipitation data;
[0016] a deformation analysis module configured to introduce each of the predicted water level data into a simulation analysis model to obtain deformation data of the dam;
[0017] a model selection module configured to acquire structure data of the dam and select a crack prediction model matched with the structure data;
[0018] a crack prediction module configured to input the deformation data and the structure data into the crack prediction model to obtain a crack prediction result of the dam, the crack prediction result indicating a position, a length and a depth of the crack;
[0019] a warning module configured to compare the crack prediction result with a warning threshold prepared in advance, and if the crack prediction result exceeds the warning threshold, issue a warning information.
[0020] The reservoir operation and risk warning method and system have the following advantages:
[0021] The water level in the future period is predicted by using multiple models, the crack is predicted based on the predicted water level, and thus a more accurate warning result is generated, the management personnel are informed before the crack occurs, and timely prevention is made, thereby ensuring the safe operation of the reservoir. BRIEF DESCRIPTION OF DRAWINGS
[0022] In order to more clearly illustrate the technical solutions in the application embodiments or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.
[0023] Figure 1A flowchart of a reservoir operation and risk early warning method provided by an embodiment of the present application. DETAILED DESCRIPTION
[0024] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.
[0025] Figure 1 A flowchart of a reservoir operation and risk early warning method provided by an embodiment of the present application. The embodiment of the present application provides a reservoir operation and risk early warning method, which comprises:
[0026] S100, acquiring basic data, confluence data and precipitation data of the reservoir.
[0027] Exemplarily, the basic data comprises initial water level, three-dimensional shape and temperature and humidity data of the reservoir, and the confluence data comprises inflow data and outflow data. The initial water level is the real-time water level at the beginning of early warning of the reservoir, which can be acquired by a hydrological station arranged near the dam, the three-dimensional shape is a three-dimensional model established after measurement of the entire reservoir, and the temperature and humidity data comprises real-time temperature and humidity data at the beginning of early warning and historical temperature and humidity data acquired from operation data.
[0028] The inflow data is the water quantity flowing into the reservoir from natural water bodies such as rivers and lakes, and a flow rate sensor can be arranged at an inlet of the reservoir to acquire real-time flow rate at the inlet, and then the water quantity flowing into the inlet per unit time can be calculated in combination with the longitudinal cross-sectional area of the inlet.
[0029] The outflow data is the water quantity discharged from the sluice of the dam, and a flow rate sensor can be arranged in the sluice to acquire real-time flow rate at the sluice, and then the water quantity flowing out of the sluice per unit time can be calculated in combination with the longitudinal cross-sectional area of the sluice.
[0030] After the inflow data and the outflow data are determined, in subsequent prediction work, the inflow data and the outflow data can be assumed to be unchanged, or the variation law of the inflow data and the outflow data in the operation data can be extracted, and then applied to prediction of the inflow data and the outflow data in a future period of time.
[0031] The precipitation data and the temperature and humidity data are similar, and comprise real-time precipitation data and historical precipitation data acquired from operation data.
[0032] S110, predicting the predicted water level data at multiple different times in a future period of time according to the basic data, the confluence data and the precipitation data.
[0033] Illustratively, after obtaining the basic data, the water area is determined according to the three-dimensional shape and the initial water level, the predicted evaporation amount is determined in combination with the water area and the temperature and humidity data, the water level drop amount is determined according to the outflow data and the predicted evaporation amount, the water level rise amount is determined according to the inflow data and the precipitation data, and the real-time predicted water level data is determined in combination with the initial water level, the water level drop amount and the water level rise amount.
[0034] Specifically, the three-dimensional shape of the reservoir presents a shape of large at the top and small at the bottom, so the area of the water surface, i.e., the water area, will also be different at different water levels. After obtaining the initial water level at a specific position, the size of the cross section of the three-dimensional shape at the initial water level can be calculated in combination with the three-dimensional shape, and the area of the cross section is the water area.
[0035] The water in the reservoir will decrease not only because of the water lock but also because of evaporation, and the larger the water area, the greater the evaporation amount. Therefore, the evaporation amount per unit area under different temperature and humidity conditions can be determined in advance through experiments, the evaporation amount curve is fitted, the evaporation amount corresponding to the predicted temperature and humidity data is taken on the evaporation amount curve, and then the product of the evaporation amount and the water area is calculated to obtain the predicted evaporation amount.
[0036] Further, when obtaining the temperature and humidity data, a plurality of environment monitoring sensors are arranged in the reservoir, each of which is used to collect real-time environment data, and the real-time environment data is input into an environment prediction model to predict the temperature and humidity data at multiple different times in a future period of time.
[0037] Specifically, the environment monitoring sensors are a collection of multiple different types of sensors. In the embodiments of the present application, the environment monitoring sensors at least include temperature and humidity sensors, precipitation sensors, wind speed sensors and air pressure sensors. These sensors collect temperature and humidity, precipitation, wind speed and air pressure data of each hydrological station and other areas in the reservoir, which are transmitted in real time to the operation and maintenance center for storage and analysis.
[0038] Further, one environment monitoring sensor is arranged in each different area of the reservoir, and each environment monitoring sensor is used to collect real-time environment data of the corresponding area in the reservoir. After collecting the real-time environment data of each area, the temperature and humidity data of the corresponding area are obtained by inputting the real-time environment data into an environment prediction model, the regional predicted evaporation amount of the corresponding area is determined in combination with the water area and the temperature and humidity data of the corresponding area, and the sum of the regional predicted evaporation amounts of all areas is the predicted evaporation amount.
[0039] Specifically, for some medium and large reservoirs, due to the large water area, there are certain differences in environmental data in different areas due to the influence of vegetation, geographical location and the like. Therefore, in order to accurately understand the real situation of each area, at least one environmental monitoring sensor can be arranged in an area that is significantly different from other areas to collect real-time environmental data in the area. After dividing the entire reservoir into multiple areas, at least one environmental monitoring sensor needs to be arranged in each area to monitor the temperature, humidity, precipitation and other data in each area.
[0040] It should be understood that the evaporation amount is a cumulative value, that is, the evaporation amount at each time is the sum of the evaporation amount at the time and all evaporation amounts before the time, so when calculating the predicted evaporation amount, the evaporation amount at each time before the prediction time also needs to be predicted, and the predicted evaporation amount is obtained by accumulation. Moreover, since the water area needs to be known when predicting the evaporation amount, the water area will change with the water level, so the water level at the previous time also needs to be predicted when predicting the evaporation amount. Similarly to the evaporation amount, the water level is also related to the historical water level at the previous time, so each time of water level needs to be predicted in order, and then the predicted water level data is calculated based on the water level at the previous time and the water level drop and water level rise.
[0041] Further, when obtaining the precipitation data, the real-time environmental data is input into the precipitation prediction model to predict the precipitation data at multiple different times in the future.
[0042] Specifically, the environmental prediction model and the precipitation prediction model are both established based on the LSTM (Long Short Term Memory) network. The LSTM network is trained using historical environmental data to obtain the environmental prediction model, and the LSTM network is trained using historical precipitation data to obtain the precipitation prediction model.
[0043] The LSTM network can use the environmental or precipitation feature information learned from the historical data to predict the future temperature and humidity and precipitation based on the current real-time environmental data. For temperature and humidity prediction, historical environmental data including temperature, humidity, air pressure, wind speed and precipitation data are needed. For precipitation prediction, in addition to historical precipitation data, temperature, humidity, wind speed and air pressure data are also needed. Whether it is an environmental prediction model or a precipitation prediction model, a training set, a test set and a validation set are needed for training, testing and validation, respectively. Only when the prediction accuracy meets the expectation can it be put into actual prediction work.
[0044] In temperature and humidity prediction, historical temperature and humidity data needs to be prepared first. This data then needs to be labeled and normalized. Next, an LSTM model is built using Keras. The labeled and normalized historical environmental data is then used to train, test, and validate the LSTM model. After completing this series of steps, an environmental prediction model is obtained. Inputting real-time environmental data collected by environmental monitoring sensors into the environmental prediction model will yield temperature and humidity data for a future period.
[0045] The process of building and utilizing precipitation prediction models is similar to the process of temperature and humidity prediction described above, except that PyTorch can be used to build LSTM models.
[0046] S120, each predicted water level data is introduced into the simulation analysis model to obtain the dam deformation data.
[0047] For example, the simulation analysis model can use the finite element software ANSYS. Because the pressure on the dam varies at different water levels—for instance, at lower water levels, the dam crest experiences no pressure, and the pressure in the middle and bottom of the dam is also relatively small—while as the water level approaches the dam crest, the crest experiences some pressure, and the middle and bottom of the dam experience greater pressure. The dam will undergo corresponding deformation under different water pressures, and this deformation is directly proportional to the pressure; that is, the greater the pressure, the greater the deformation.
[0048] S130: Obtain the structural data of the dam and select a crack prediction model that matches the structural data.
[0049] For example, the structural data of a dam includes the type, size, shape, construction parameters, etc. Dams with different structural data will produce different deformation and crack occurrence results for the same water level. Therefore, dams with each type of structural data can be simulated in experiments to establish crack prediction models corresponding to different structural data.
[0050] Furthermore, the crack prediction model can adopt the GM(1,1) model. Since the causes and mechanisms of cracks in dams are not very clear, and the GM(1,1) model is a prediction technique based on grey system theory, this model does not require all information to be known in advance. It predicts cracks on the surface or inside of the dam through this unpredictable grey information and the known white information.
[0051] Further, the water in the reservoir is not static, the above prediction of cracks by pressure is to approximate the reservoir as a static water area, but in the wet season, especially when heavy rainfall occurs upstream, the water often contains a large amount of sand, fallen trees, and even some destroyed buildings, these impurities are mixed into the reservoir with water at a faster speed, and the impact force on the dam is very large, which even exceeds the static pressure of the water. At this time, the scouring effect under this condition needs to be considered.
[0052] In the embodiments of the present application, a plurality of unmanned aerial vehicles can be deployed at the inlet of the reservoir to control the unmanned aerial vehicles to take off at fixed times and shoot images of the water surface of the inlet. After analyzing the images, the number of various impurities in the water is identified. The impurities in the water are roughly divided into sand, trees and building debris in the embodiments of the present application. Since the water containing a large amount of impurities flows at a fast speed, the impurities are roughly uniformly distributed in the water. Therefore, the number of trees and building debris contained in the water flowing through the inlet can be determined by identifying the number of trees and building debris in the image of the water surface, and the content of sand can be obtained by identifying the color of the water in the image of the water surface.
[0053] After obtaining the number of impurities in the water, the number is input into a fluid simulation software to simulate the impact force of the water containing impurities on the dam. Specifically, the fluid simulation software can use FLOW-3D. By setting the types, sizes and numbers of impurities in the water, the impact force that the dam may be subjected to in the future can be obtained.
[0054] After simulation, the obtained impact force contains numerical values of the impact force size at each position of the dam. By combining the numerical values with the pressure obtained according to the predicted water level data, the pressure that the dam is subjected to at each position under the dual action of the static pressure of the water body and the dynamic scouring can be obtained, and then the deformation data of the dam is obtained by using ANSYS simulation.
[0055] S140, input the deformation data and the structure data into the crack prediction model to obtain the crack prediction result of the dam, and the crack prediction result represents the position, length and depth of the crack.
[0056] For example, after analyzing the predicted water level data by using the simulation analysis model, the deformation data of each position of the dam is obtained. In the crack prediction result, the crack occurs at the position with the largest deformation data on the dam. It should be understood that under severe or long-term deformation, the cracks that the dam may produce will not be only one. Therefore, when determining the position of the crack, a deformation threshold can be set. If the change amount of a certain deformation data in a unit time exceeds the deformation threshold, it indicates that the deformation of the dam at the position is relatively severe, and it can be considered that a crack will occur at the position. Therefore, the positions of the deformation data whose change amount in a unit time exceeds the deformation threshold can be considered as the positions where the cracks occur.
[0057] After the location of the crack is determined, the length and depth of the crack can be predicted using a crack prediction model.
[0058] S150, compare the crack prediction result with a pre-prepared warning threshold, and if the crack prediction result exceeds the warning threshold, issue a warning information.
[0059] Exemplarily, after the location of the crack is determined, the warning threshold corresponding to the location is obtained, the length data and depth data in the crack prediction result are compared with the warning threshold, and whether to issue a warning information is determined according to the comparison result.
[0060] Since the risk level of the crack is related to the location thereof, the warning threshold corresponding to each location needs to be pre-set, and if the length data and depth data of the crack at a certain location both exceed the length data and depth data in the corresponding warning threshold, the operation and maintenance center can issue a warning information.
[0061] Further, even if the length data and depth data of the crack do not exceed the corresponding warning threshold, the operation and maintenance center can also issue a reminder information to make the reservoir operation and management personnel know the occurrence of the crack.
[0062] The embodiment of the application also provides a reservoir operation and risk early warning system, which comprises:
[0063] A data acquisition module is configured to acquire basic data, confluence data and precipitation data of the reservoir.
[0064] A water level prediction module is configured to predict predicted water level data at multiple different time points in a future period of time according to the basic data, the confluence data and the precipitation data.
[0065] A deformation analysis module is configured to introduce each predicted water level data into a simulation analysis model to obtain deformation data of the dam.
[0066] A model selection module is configured to acquire structure data of the dam and select a crack prediction model matched with the structure data.
[0067] A crack prediction module is configured to input the deformation data and the structure data into the crack prediction model to obtain a crack prediction result of the dam, the crack prediction result indicating a location, a length and a depth of the crack.
[0068] A warning module is configured to compare the crack prediction result with a pre-prepared warning threshold, and issue a warning information if the crack prediction result exceeds the warning threshold.
[0069] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that such additions and modifications be included within the scope of the application. It is the following claims, including any amendments thereto, which define the scope of the application.
[0070] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.
Claims
1. A method for reservoir operation and maintenance and risk early warning, characterized in that, include: Acquire basic data, runoff data, and precipitation data of the reservoir; Based on the aforementioned basic data, the confluence data, and the precipitation data, predict the water level data at multiple different times in the future period; Each of the predicted water level data is introduced into the simulation analysis model to obtain the dam deformation data; Obtain the structural data of the dam and select a crack prediction model that matches the structural data; The deformation data and the structural data are input into the crack prediction model to obtain the crack prediction results of the dam. The crack prediction results include the location, length and depth of the cracks. The crack prediction result is compared with a pre-prepared warning threshold. If the crack prediction result exceeds the warning threshold, a warning message is issued.
2. The reservoir operation and maintenance and risk early warning method according to claim 1, characterized in that, The basic data includes the reservoir's initial water level, three-dimensional shape, and temperature and humidity data. The confluence data includes inflow and outflow data. After acquiring the basic data, the water area is determined based on the three-dimensional shape and the initial water level. The predicted evaporation is determined by combining the water area and the temperature and humidity data. The water level drop is determined based on the predicted evaporation and the outflow data. The water level rise is determined based on the inflow data and the precipitation data. The real-time predicted water level data is determined by combining the initial water level, the water level drop, and the water level rise.
3. The reservoir operation and maintenance and risk early warning method according to claim 2, characterized in that, When acquiring the temperature and humidity data, the reservoir is equipped with multiple environmental monitoring sensors. Each environmental monitoring sensor is used to collect real-time environmental data, and the real-time environmental data is input into an environmental prediction model to predict the temperature and humidity data at multiple different times in the future.
4. The reservoir operation and maintenance and risk early warning method according to claim 3, characterized in that, An environmental monitoring sensor is installed in different areas of the reservoir. Each environmental monitoring sensor is used to collect real-time environmental data of the corresponding area in the reservoir. After the real-time environmental data of each area is collected, it is input into the environmental prediction model to obtain the temperature and humidity data of the corresponding area. The corresponding area's predicted evaporation is determined by combining the water area and the temperature and humidity data of the corresponding area. The sum of the predicted evaporation of all areas is the predicted evaporation.
5. The reservoir operation and maintenance and risk early warning method according to claim 3, characterized in that, When acquiring the precipitation data, the real-time environmental data is input into the precipitation prediction model to predict the precipitation data at multiple different times in the future.
6. The reservoir operation and maintenance and risk early warning method according to claim 5, characterized in that, Both the environmental prediction model and the precipitation prediction model are based on LSTM networks. The environmental prediction model is obtained by training the LSTM network using historical environmental data, and the precipitation prediction model is obtained by training the LSTM network using historical precipitation data.
7. The reservoir operation and maintenance and risk early warning method according to claim 1, characterized in that, The crack prediction model adopts the GM(1,1) model.
8. The reservoir operation and maintenance and risk early warning method according to claim 1, characterized in that, After analyzing the predicted water level data using the simulation analysis model, the deformation data of each location of the dam is obtained. In the crack prediction results, the crack occurs at the location of the largest deformation data on the dam.
9. A reservoir operation and maintenance and risk early warning method according to claim 8, characterized in that, After determining the location of the crack, the warning threshold corresponding to that location is obtained. The length and depth data in the crack prediction result are compared with the warning threshold, and the warning information is determined based on the comparison result.
10. A system applying the reservoir operation and maintenance and risk early warning method according to any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire basic data, runoff data, and precipitation data of the reservoir. The water level prediction module is used to predict water level data at multiple different times in the future based on the basic data, the runoff data and the precipitation data; The deformation analysis module is used to input each of the predicted water level data into the simulation analysis model to obtain the deformation data of the dam; The model selection module is used to acquire the structural data of the dam and select a crack prediction model that matches the structural data. The crack prediction module is used to input the deformation data and the structural data into the crack prediction model to obtain the crack prediction result of the dam, wherein the crack prediction result indicates the location, length and depth of the crack. The early warning module is used to compare the crack prediction result with a pre-prepared warning threshold. If the crack prediction result exceeds the warning threshold, an early warning message is issued.