River and Lake Status Assessment Methods and Systems Based on Big Data
By processing multi-source heterogeneous data and using an extreme gradient boosting tree regression model, the distance threshold of the spatial clustering algorithm is dynamically adjusted, which solves the problem of delayed state assessment in water management systems under extreme weather conditions and enables real-time dynamic scheduling and proactive response of river and lake status.
Patent Information
- Application Number
- CN202611086336.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-21
- Publication Date
- 2026-11-17
AI Technical Summary
Existing spatial clustering algorithms cannot dynamically adjust distance thresholds under sudden extreme weather conditions, resulting in delays in the status assessment of water management systems and physical scheduling instructions, making it impossible to detect water resource shortages in advance.
By collecting and preprocessing multi-source heterogeneous data, and combining it with an extreme gradient boosting tree regression model, the environmental evaporation driving force and agricultural water demand jump response index are dynamically calculated. The distance threshold of the spatial clustering algorithm is adjusted to achieve real-time dynamic assessment and scheduling of river and lake status.
Under extreme weather conditions, the water management system can keenly detect hidden water shortage risks and output scheduling and control signals in advance, avoiding delays in status assessment and physical scheduling, and providing advanced physical response and stable load scheduling.
Smart Images

Figure CN123113644A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of water conservancy information data processing technology. More specifically, this invention relates to a method and system for assessing the state of rivers and lakes based on big data. Background Technology
[0002] In modern water management systems, the assessment of the carrying capacity of river and lake water resources scheduling usually relies heavily on the comprehensive analysis of historical operational data and meteorological data. The systematic assessment methods in existing technologies typically introduce density-based spatial clustering algorithms. These algorithms mainly perform cluster analysis on long-term historical monitoring data of river and lake cross-sections and flow data. During the real-time operation monitoring phase, the water management system calculates the distance from real-time water level characteristic points and flow characteristic points to the center of the aforementioned conventional safe state data point clusters and compares it with a fixed boundary distance threshold to determine whether the current state of the river and lake is still within a safe carrying capacity range.
[0003] This spatial clustering assessment method, based entirely on a fixed distance threshold, does indeed possess a certain degree of statistical rationality under stable climatic conditions. However, in extremely complex and variable real-world environments, when water management systems face sudden high temperatures and strong winds, the actual evaporation and dissipation rates of surface soil and vegetation undergo drastic, nonlinear step changes within a very short period. At this time, the actual agricultural water demand in downstream irrigation areas will show a sharp increase. However, due to the inherent lag in the physical flow of water in nature, the water levels at river and lake sections do not show any significant drop characteristics in the short term. During this period of hidden water shortage, the water management system can only rely solely on the passive drop in water level data for judgment.
[0004] However, the boundary distance threshold of existing spatial clustering algorithms is statically determined based on long-term historical data. This static parameter cannot perceive the rapid changes in external micro-meteorological dynamics and thermodynamic conditions. Modern water management systems rely on a single fixed historical threshold to judge the hydrological status, failing to deeply integrate the driving potential of environmental temperature, wind speed, humidity, etc., on water evaporation and transport with the clustering boundary of the underlying algorithm. As a result, in actual scheduling scenarios, when a severe drought causes a real water shortage, as long as the water level has not fallen significantly below the set static threshold, the water management system still blindly outputs a state safety assessment conclusion. This judgment logic caused by static parameters makes it impossible for the water resource scheduling and control center to obtain water demand response instructions in advance, ultimately resulting in a serious delay in the issuance of physical scheduling instructions, exposing a serious state assessment lag in practice. Summary of the Invention
[0005] To address the technical problems of existing spatial clustering algorithms, which suffer from severe delays in state assessment and physical scheduling command issuance under sudden weather conditions due to static parameters, this invention provides solutions in the following aspects.
[0006] In a first aspect, the present invention provides a method for assessing the state of rivers and lakes based on big data, comprising: real-time acquisition and preprocessing of multi-source heterogeneous hydrological data and micro-meteorological data to obtain hydrological state characteristics and environmental micro-meteorological characteristics; acquiring the calendar time code, historical average temperature, and historical average precipitation for the current period and inputting them into a pre-trained extreme gradient boosting tree regression model for calculation to predict the baseline water demand forecast for the current period; extracting an environmental evaporation driving force index based on the environmental micro-meteorological characteristics and combining it with the soil moisture content in the hydrological state characteristics to extract an agricultural water demand jump response index; using the environmental evaporation driving force index and the agricultural water demand jump response index to perform dynamic constraint calculation on the basic distance threshold of the spatial clustering algorithm to calculate the dynamic clustering distance threshold at the current moment; performing state logic judgment based on the comparison result between the abnormal distance from the current real-time water level characteristics to the preset normal state cluster center and the dynamic clustering distance threshold, and calculating the final river and lake comprehensive scheduling carrying capacity index based on the baseline water demand forecast; and outputting a physical scheduling control signal when the river and lake comprehensive scheduling carrying capacity index is lower than the preset safety threshold lower limit to achieve a safety carrying capacity assessment of the river and lake state.
[0007] This invention utilizes an extreme gradient boosting tree regression model to separate a baseline load unaffected by sudden climate changes. It then extracts environmental evaporation driving forces and agricultural water demand jump response indices, deeply integrating the driving potential of water evaporation transport with the clustering boundaries of the underlying algorithm. By calculating a dynamic clustering distance threshold using the agricultural water demand jump response index, the water management system can sensitively detect and narrow the distance threshold to trigger anomaly detection during periods of hidden water scarcity, such as when sudden droughts cause a real water shortage but water levels have not yet dropped significantly. Finally, it calculates the river and lake comprehensive scheduling carrying capacity index based on the comparison results and outputs physical signals in advance, effectively solving the problems of delayed state assessment and delayed physical scheduling commands caused by static parameters in existing spatial clustering algorithms.
[0008] Preferably, the real-time acquisition and preprocessing of multi-source heterogeneous hydrological data and micrometeorological data to obtain hydrological state characteristics and environmental micrometeorological characteristics includes: synchronously acquiring on-site hydrological data and micrometeorological data through IoT sensing devices to obtain corresponding time series data. The hydrological data includes the current water level, current outflow, current soil volumetric moisture content, and target soil moisture content of the river / lake control section; the micrometeorological data includes the current absolute temperature, current relative humidity, and current average wind speed; median filtering is applied to the acquired time series data, and interpolation is used to fill in missing data frames caused by communication delays to obtain hydrological state characteristics and environmental micrometeorological characteristics.
[0009] This invention removes impulse noise caused by environmental interference and fills in missing frames by collecting and preprocessing multi-source heterogeneous data, thereby providing high-precision and time-continuous basic information support for subsequent state assessment.
[0010] Preferably, the obtained calendar time code, historical average temperature, and historical average precipitation for the current period are input into a pre-trained extreme gradient boosting tree regression model for calculation. The prediction of the baseline water demand for the current period includes: extracting the calendar time code, historical average temperature, and historical average precipitation for the same period in history as the independent variable feature set, using the total water demand actually scheduled in history as the dependent variable, and training the extreme gradient boosting tree regression model; during prediction, the calendar time code, historical average temperature, and historical average precipitation corresponding to the current period are simultaneously obtained from the historical database and used as the target input feature vector and input into the extreme gradient boosting tree regression model. The baseline water demand prediction value is obtained by summing the output weights of all leaf nodes, and the baseline water demand prediction value is limited to be greater than a preset minimum ecological baseflow constant.
[0011] This invention improves the adaptive forward propagation of the tree regression model through extreme gradients, which can accurately separate the basic water demand load of the water operation and management system that is not affected by sudden climate changes, thus ensuring the physical effectiveness of the subsequent calculation system.
[0012] Preferably, the environmental evaporation driving force index satisfies the following relationship:
[0013] ; in, This indicates the environmental evaporation driving force index. This indicates the absolute temperature of the current environment. This indicates the set standard reference absolute temperature. This indicates the average wind speed in the current environment. This indicates the set standard reference wind speed. This indicates the relative humidity of the current environment.
[0014] Preferably, the extraction of the agricultural water demand jump response index includes: when the current soil volumetric moisture content is less than the target soil moisture content, calculating the difference between the target soil moisture content and the current soil volumetric moisture content, dividing the calculated difference by the target soil moisture content, and multiplying the resulting ratio by the environmental evaporation driving force index to obtain the agricultural water demand jump response index.
[0015] Preferably, the calculation of the dynamic clustering distance threshold at the current moment includes: retrieving the basic distance threshold of the normal state clusters pre-trained based on historical stable hydrological data, and using the agricultural water demand jump response index to dynamically penalize the basic distance threshold, outputting the dynamic clustering distance threshold at the current moment, wherein the dynamic clustering distance threshold satisfies the following relationship: ; in, This represents the dynamic clustering distance threshold. This represents the basic distance threshold. This represents an exponential function with the natural constant as its base. This represents the preset sensitivity constant coefficient. This represents the agricultural water demand jump response index.
[0016] This invention introduces physical evolution parameters of the natural environment to dynamically and adaptively adjust the spatial clustering distance threshold boundary, which can effectively eliminate the state perception sluggishness caused by fixed parameters, enabling the water operation and management system to dynamically shrink the safety boundary under external environmental stress.
[0017] Preferably, the real-time water level feature is converted into a feature vector, and the Euclidean distance from the feature vector to the center of the trained normal state cluster is calculated as the abnormal distance from the current real-time water level feature to the center of the normal state cluster. If the abnormal distance is less than or equal to the dynamic clustering distance threshold, the current feature point is marked as a node within the normal state cluster. If the abnormal distance is greater than the dynamic clustering distance threshold, the current feature point is forcibly marked as an out-of-cluster abnormal node.
[0018] Preferably, the river and lake integrated scheduling carrying capacity index satisfies the following relationship: ; in, This represents the comprehensive river and lake scheduling carrying capacity index. This indicates the current outflow rate. This represents the baseline water demand forecast value. Represents the natural logarithm function. This indicates the abnormal distance from the current real-time water level feature to the center of the normal state cluster.
[0019] Preferably, the output physical scheduling control signal includes: the bus automatically sending pulse signals to the programmable logic controller to control the stepper motor of the reservoir water regulating gate to increase the opening position and perform physical water supply replenishment of the entity.
[0020] This invention combines a preceding dynamic logic judgment mechanism with the calculation of the comprehensive scheduling carrying capacity index and the output of automated pulse control, enabling the water affairs operation and management system to achieve advanced physical response and stable carrying capacity scheduling.
[0021] Secondly, the present invention provides a river and lake status assessment system based on big data, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned river and lake status assessment method based on big data is implemented.
[0022] By adopting the above technical solution, the above-mentioned big data-based river and lake status assessment method is generated into a computer program and stored in a memory so that it can be loaded and executed by a processor. In this way, a terminal device can be made based on the memory and processor for convenient use.
[0023] The beneficial effects of this invention are as follows: This invention introduces environmental evaporation driving force index and agricultural water demand jump response index, deeply integrating the dynamics and thermodynamics of external microclimate into the underlying spatial clustering algorithm, and realizes dynamic adaptive adjustment of distance threshold. This dynamic penalty mechanism ensures that even in extremely complex and changeable natural environments, even if the physical water level has not dropped significantly due to hysteresis, the water management system can still capture the hidden water shortage risk caused by microclimate changes.
[0024] Furthermore, this dynamic evaluation system combines real-time preprocessing of multi-source heterogeneous high-precision data with baseline load prediction of extreme gradient boosting tree regression model to construct a comprehensive river and lake scheduling carrying capacity index evaluation system. When faced with sudden extreme weather conditions, the water affairs operation management system no longer blindly outputs erroneous conclusions of state safety, but instead triggers the de-clustering anomaly logic in advance, and then automatically issues physical water supply replenishment instructions to the programmable logic controller. This not only alleviates the serious lag in state assessment and physical scheduling instructions in traditional technology, but also provides modern water affairs operation management systems with advanced physical response and stable and reliable safety carrying capacity scheduling guarantee. Attached Figure Description
[0025] Figure 1 This is a flowchart of a big data-based method for assessing the state of rivers and lakes. Figure 2 This is a comparison chart of the evaluation results of the embodiments of the present invention and the prior art based on fixed boundaries; Figure 3It is a dynamic adaptive evaluation result diagram that integrates the laws of natural physics. Detailed Implementation
[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0028] This invention discloses a method for assessing the state of rivers and lakes based on big data, referring to... Figure 1 This includes steps S1-S4: S1. Real-time acquisition and preprocessing of multi-source heterogeneous hydrological and micro-meteorological data to obtain hydrological state characteristics and environmental micro-meteorological characteristics.
[0029] In one optional embodiment, an IoT sensor network is deployed at the river and lake control section and the main downstream irrigation area to synchronously collect on-site hydrological and micro-meteorological data at a set 15-minute time interval, obtaining corresponding time series data; the current water level and current outflow of the river and lake control section are obtained through ultrasonic water level gauges and Doppler acoustic profilers; the current absolute temperature, current relative humidity, and current average wind speed are obtained through temperature and humidity transmitters and anemometers deployed in a downwind standard meteorological Stevenson screen; and the current soil volumetric water content is obtained through frequency domain reflectance soil moisture sensors buried at a fixed depth in the downstream irrigation area, while the target soil moisture content is obtained based on the calibrated water content of crops during the growing season published by the local agricultural management department.
[0030] The collected time series data were filtered by median to remove impulse noise caused by environmental interference, and interpolation was used to fill in missing data frames caused by communication delay. In order to avoid the abnormal situation that the mathematical denominator is zero when the subsequent calculation is in a completely dry state, the effective lower limit of the target soil moisture content was set to a constant of 0.1.
[0031] Thus, through the collection and rigorous preprocessing of multi-source heterogeneous data, high-precision and time-continuous basic information support can be provided for subsequent state assessment.
[0032] S2. Obtain the calendar time code, historical average temperature and historical average precipitation for the current period and input them into the pre-trained extreme gradient boosting tree regression model for calculation to predict the baseline water demand forecast for the current period.
[0033] In an optional embodiment, the water operation management system extracts the calendar time code, historical average temperature, and historical average precipitation of the same period as the independent variable feature set, and uses the total water demand actually scheduled in history as the dependent variable. The extreme gradient boosting tree regression model is trained in the background server. During the forward propagation calculation of the extreme gradient boosting tree regression model, the extreme gradient boosting tree regression model constructs a deep multi-layer decision tree by calculating the node splitting information gain of each feature dimension. Through gradient optimization, the extreme gradient boosting tree regression model gradually converges to a stable state. At this time, the feature splitting threshold and leaf node weights inside the extreme gradient boosting tree regression model are fixed.
[0034] During the real-time forecasting phase, the water management system does not input the real-time collected hydrological status features and environmental micro-meteorological features into the model. Instead, it extracts the calendar time code corresponding to the current period, historical average temperature, and historical average precipitation from the historical meteorological and time database. These historical features are used as target input feature vectors and input into the trained extreme gradient boosting tree regression model. The baseline water demand forecast value is obtained by summing the output weights of all leaf nodes inside the extreme gradient boosting tree regression model.
[0035] For example, during a stable weather period, the weighted sum of the outputs of each leaf node calculates the baseline water demand forecast for the current period as 10 cubic meters per second. This baseline water demand forecast represents the theoretical basis load for the water management system to maintain normal operation under the condition of excluding extreme weather changes. In order to ensure the physical effectiveness of the subsequent calculation system, the water management system forcibly limits this baseline water demand forecast to be greater than the minimum ecological baseflow constant set by the system.
[0036] Thus, by using extreme gradients to enhance the adaptive forward propagation of the tree regression model, the basic water demand load of the water management system, which is not affected by sudden climate changes, can be accurately separated.
[0037] S3. Based on the environmental micro-meteorological characteristics, extract the environmental evaporation driving force index and combine it with the soil moisture content in the hydrological state characteristics to extract the agricultural water demand jump response index. Using the environmental evaporation driving force index and the agricultural water demand jump response index, perform dynamic constraint calculation on the basic distance threshold of the spatial clustering algorithm to calculate the dynamic clustering distance threshold at the current moment.
[0038] In an optional embodiment, the environmental evaporation driving force index is equal to the ratio of the current ambient absolute temperature to the standard reference absolute temperature, multiplied by the ratio of the current average ambient wind speed to the standard reference wind speed, the square root of the multiplication result, and then multiplied by the difference between the number 1 and the current ambient relative humidity. The environmental evaporation driving force index satisfies the following relationship: ; in, Indicators representing the driving force of environmental evaporation. This indicates the current absolute temperature of the environment. This indicates the set standard reference absolute temperature. This indicates the current average wind speed. This indicates the set standard reference wind speed. This indicates the current relative humidity.
[0039] For example, in the current situation of sudden high temperature accompanied by hot and dry winds, the current absolute temperature is 303.15 Kelvin, the set standard reference absolute temperature is constant at 273.15 Kelvin, the current average wind speed is 4 m / s, the set standard reference wind speed is constant at 1 m / s, and the current relative humidity has decreased to 0.4. Substituting these values into the formula, the following calculation is performed: ; The final calculated environmental evaporation driving force index is 1.264. In this process, all ratios are divisions of physical quantities with the same dimension, which ensures the overall dimensionless property of the environmental evaporation driving force index.
[0040] As an optional interpretation of this embodiment, the construction of the above-mentioned environmental evaporation driving force index is a nonlinear dimensionality reduction approximation of the classical Penman-Montis evapotranspiration equation. By extracting the geometric mean of the ratio of absolute temperature to wind speed and combining it with the macroscopic characterization of saturated water vapor pressure difference, the potential meteorological suction capacity of the atmospheric boundary layer on water bodies is accurately characterized at the algorithm level.
[0041] Then, the water management system calculates the agricultural water demand jump response index based on soil moisture content and environmental evaporation driving force index. Specifically, when the current soil volumetric moisture content is less than the target soil moisture content, the difference between the target soil moisture content and the current soil volumetric moisture content is calculated and divided by the target soil moisture content. The calculated ratio is then multiplied by the environmental evaporation driving force index to obtain the agricultural water demand jump response index.
[0042] For example, the current soil volumetric moisture content has decreased to 0.2, while the target soil moisture content set by the water management system is 0.3. Since the current moisture content is less than the target moisture content, the water management system performs the following calculation: subtract the current soil moisture content from the target soil moisture content to obtain a difference of 0.1; divide the calculated difference of 0.1 by the target moisture content of 0.3 to obtain a dimensionless ratio of 0.333; multiply this dimensionless ratio by the previously obtained environmental evaporation driving force index of 1.264; and finally calculate the agricultural water demand jump response index as 0.421.
[0043] When the current soil volumetric moisture content is greater than or equal to the target soil moisture content, the value of the agricultural water demand jump response index is directly set to zero.
[0044] Finally, the water management system retrieves the basic distance threshold of the normal state cluster, which is pre-trained based on historical stable hydrological data, and uses the agricultural water demand jump response index to dynamically penalize the basic distance threshold, outputting the dynamic clustering distance threshold at the current moment. The dynamic clustering distance threshold satisfies the following relationship: ; in, Indicates the dynamic clustering distance threshold. Indicates the base distance threshold. This represents an exponential function with the natural constant as its base. This represents the preset sensitivity constant coefficient. This indicates the responsiveness of agricultural water demand to sudden changes.
[0045] For example, the base distance threshold is 5, and the preset sensitivity constant coefficient is 1. Substituting these values into the formula, the calculation is as follows: ; This indicates that due to external environmental pressures, the water management system dynamically reduced the original basic threshold from 5 to 3.28.
[0046] As an optional interpretation of this embodiment, the dynamic clustering distance threshold introduces a negative exponential decay term, the physical meaning of which is to map the drought stress memory effect of the watershed ecosystem. When the water jump response caused by a sudden change in soil moisture intensifies, the boundary of the water management system tolerating hydrological anomalies shrinks exponentially.
[0047] Thus, by introducing physical evolution parameters of the natural environment to dynamically and adaptively adjust the spatial clustering distance threshold boundary, the state perception sluggishness caused by fixed parameters can be effectively eliminated.
[0048] S4. Based on the comparison results of the abnormal distance from the current real-time water level characteristics to the preset normal state cluster center and the dynamic clustering distance threshold, perform state logic judgment and calculate the final river and lake comprehensive scheduling carrying capacity index in combination with the baseline water demand prediction value. When the river and lake comprehensive scheduling carrying capacity index is lower than the preset safety threshold lower limit, output the physical scheduling control signal to realize the safety carrying capacity assessment of the river and lake status.
[0049] In an optional embodiment, if the abnormal distance from the current real-time water level feature to the center of the normal state cluster is less than or equal to the dynamic clustering distance threshold, the current feature point is marked as a node within the normal state cluster. If the abnormal distance from the current real-time water level feature to the center of the normal state cluster is greater than the dynamic clustering distance threshold, the current feature point is forcibly marked as an abnormal node that has detached from the cluster. For example, assuming the abnormal distance from the current real-time water level feature to the center of the normal state cluster is 4, which is greater than the contracted dynamic clustering distance threshold of 3.28, the state logic determination is triggered, and the current feature point is forcibly marked as an abnormal node that has detached from the cluster.
[0050] Then, the comprehensive river and lake scheduling carrying capacity index is calculated, with the following formula: ; in, Indicates the comprehensive carrying capacity index of rivers and lakes. This indicates the current outflow rate. This represents the baseline water demand forecast. Represents the natural logarithm function. This represents the abnormal distance from the current real-time water level feature to the center of the normal state cluster. Since the abnormal distance from the current real-time water level feature to the center of the normal state cluster is always greater than or equal to zero, the internal value of the natural logarithm function is always greater than or equal to 1, ensuring that the logarithm calculation result is non-negative, thereby avoiding the problem of the denominator being zero or negative during the calculation process.
[0051] For example, the current outflow rate is 50 cubic meters per second, the known baseline water demand is predicted to be 10 cubic meters per second, and the anomaly distance from the current real-time water level characteristic to the center of the normal state cluster is 4. Substituting these values into the formula, we obtain: ; If the lower limit of the safety threshold set by the water operation management system is 2.5, and the comprehensive scheduling carrying capacity index of 1.916 is far below the safety bottom line, the water operation management system bus will automatically send a pulse signal to the programmable logic controller to control the stepper motor of the reservoir water diversion gate to increase the opening position and perform physical water supply replenishment.
[0052] The above-mentioned comprehensive scheduling carrying capacity index The supply-demand ratio of macroeconomic regulation The logarithmic penalty term is introduced by orthogonally fusing isolated penalty factors with the microscopic data space. This is used to nonlinearly amplify the impact of abnormal distances on bearing capacity, so that when the hydrological state deviates significantly from the historical stable state, the bearing capacity index can respond more sensitively, thereby triggering dispatch instructions in advance.
[0053] Thus, by comprehensively calculating the carrying capacity index and outputting automated pulse control, the water affairs operation and management system can achieve advanced physical response and stable carrying capacity scheduling.
[0054] like Figure 2 As shown, the image presents a two-dimensional rectangular coordinate system, with the horizontal axis parameter representing the measured time span and the vertical axis parameter representing the cross-sectional water level height. In the existing spatial clustering algorithm, a fixed boundary distance threshold map is used as a fixed tolerance boundary. In the latter half of the time span, the environmental parameters have undergone significant abrupt changes. However, due to the physical cumulative delay of the water level drop, the previous monitoring data points have not fallen below the fixed tolerance boundary. During this period, the algorithm still judges the state as normal and outputs scattered points until the water level crosses the fixed tolerance boundary, at which point it outputs warning scattered points. The chart exposes the time difference defect of assessment lag.
[0055] like Figure 3 As shown, the dynamic clustering distance threshold calculated by this invention is used as the physical dynamic contraction boundary. This physical dynamic contraction boundary abandons the flat shape and exhibits a trajectory characteristic that adaptively contracts upwards along the vertical axis as the environmental evaporation stress parameter increases. Based on this dynamic constraint mechanism, the original... Figure 2 The slightly decreased lattice in the middle was judged to be normal. Figure 3 The data points intersect with the dynamic boundary in advance, instantly transforming into a risk warning state scattering. The chart, through the displacement changes of the distribution, confirms that the present invention effectively eliminates the warning lag period of the underlying algorithm.
[0056] This invention also discloses a river and lake status assessment system based on big data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the river and lake status assessment method based on big data according to this invention is implemented.
[0057] The aforementioned big data-based river and lake status assessment system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces. Their settings and functions are known in the art and will not be described in detail here.
[0058] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.
Claims
1. A method for assessing the state of rivers and lakes based on big data, characterized in that, include: Real-time acquisition and preprocessing of multi-source heterogeneous hydrological and micrometeorological data are performed to obtain hydrological state characteristics and environmental micrometeorological characteristics. Obtain the calendar time code, historical average temperature and historical average precipitation for the current period and input them into a pre-trained extreme gradient boosting tree regression model to calculate and predict the baseline water demand forecast for the current period. Based on the environmental evaporation driving force index extracted from the environmental micro-meteorological characteristics and the agricultural water demand jump response index extracted from the soil moisture content in the hydrological state characteristics, the environmental evaporation driving force index and the agricultural water demand jump response index are used to perform dynamic constraint calculation on the basic distance threshold of the spatial clustering algorithm to calculate the dynamic clustering distance threshold at the current moment. Based on the comparison between the abnormal distance from the current real-time water level characteristics to the preset normal state cluster center and the dynamic clustering distance threshold, state logic judgment is performed, and the final river and lake comprehensive scheduling carrying capacity index is calculated in combination with the baseline water demand prediction value. When the river and lake comprehensive scheduling carrying capacity index is lower than the preset safety threshold lower limit, a physical scheduling control signal is output to realize the safety carrying capacity assessment of the river and lake status.
2. The river and lake status assessment method based on big data according to claim 1, characterized in that, The real-time acquisition and preprocessing of multi-source heterogeneous hydrological and micrometeorological data to obtain hydrological state characteristics and environmental micrometeorological characteristics includes: Hydrological and micro-meteorological data are collected synchronously from the site using IoT sensing devices to obtain corresponding time-series data. The hydrological data includes the current water level, current outflow, current soil volumetric moisture content, and target soil moisture content of the river and lake control section. The micro-meteorological data includes the current absolute temperature, current relative humidity, and current average wind speed. The collected time series data were subjected to median filtering and interpolation was used to fill in missing data frames caused by communication delays, thus obtaining hydrological state characteristics and environmental micrometeorological characteristics.
3. The river and lake status assessment method based on big data according to claim 1, characterized in that, The process of obtaining the calendar time code, historical average temperature, and historical average precipitation for the current time period and inputting them into a pre-trained extreme gradient boosting tree regression model for calculation, and predicting the baseline water demand forecast for the current time period, includes: extracting the calendar time code, historical average temperature, and historical average precipitation for the same period in history as the independent variable feature set, using the total water demand actually scheduled in history as the dependent variable, and training the extreme gradient boosting tree regression model; during prediction, simultaneously obtaining the calendar time code, historical average temperature, and historical average precipitation corresponding to the current time period from the historical database, using them as the target input feature vector and inputting them into the extreme gradient boosting tree regression model, and summing the output weights of all leaf nodes to obtain the baseline water demand forecast value, and restricting the baseline water demand forecast value to be greater than the preset minimum ecological baseflow constant.
4. The river and lake status assessment method based on big data according to claim 2, characterized in that, The environmental evaporation driving force index satisfies the following relationship: ; in, This indicates the environmental evaporation driving force index. This indicates the absolute temperature of the current environment. This indicates the set standard reference absolute temperature. This indicates the average wind speed in the current environment. This indicates the set standard reference wind speed. This indicates the relative humidity of the current environment.
5. The river and lake status assessment method based on big data according to claim 4, characterized in that, The extraction of the agricultural water demand jump response index includes: when the current soil volumetric moisture content is less than the target soil moisture content, calculating the difference between the target soil moisture content and the current soil volumetric moisture content, dividing the calculated difference by the target soil moisture content, and multiplying the resulting ratio by the environmental evaporation driving force index to obtain the agricultural water demand jump response index.
6. The river and lake status assessment method based on big data according to claim 5, characterized in that, The calculation of the dynamic clustering distance threshold at the current moment includes: retrieving the basic distance threshold of the normal state clusters pre-trained based on historical stable hydrological data, and using the agricultural water demand jump response index to perform dynamic penalty calculation on the basic distance threshold, outputting the dynamic clustering distance threshold at the current moment, wherein the dynamic clustering distance threshold satisfies the following relationship: ; in, This represents the dynamic clustering distance threshold. This represents the basic distance threshold. This represents an exponential function with the natural constant as its base. This represents the preset sensitivity constant coefficient. This represents the agricultural water demand jump response index.
7. The river and lake status assessment method based on big data according to claim 1, characterized in that, The current real-time water level feature is converted into a feature vector, and the Euclidean distance from the feature vector to the preset normal state cluster center is calculated as the abnormal distance from the current real-time water level feature to the normal state cluster center. If the abnormal distance is less than or equal to the dynamic clustering distance threshold, the current feature point is marked as a node within a normal state cluster. If the abnormal distance is greater than the dynamic clustering distance threshold, the current feature point will be forcibly marked as an abnormal node that has been removed from the cluster.
8. The river and lake status assessment method based on big data according to claim 7, characterized in that, The comprehensive river and lake carrying capacity index satisfies the following relationship: ; in, This represents the comprehensive river and lake scheduling carrying capacity index. This indicates the current outflow rate. This represents the baseline water demand forecast value. Represents the natural logarithm function. This indicates the abnormal distance from the current real-time water level feature to the center of the normal state cluster.
9. The river and lake status assessment method based on big data according to claim 1, characterized in that, The output physical scheduling control signal includes: the bus automatically sends pulse signals to the programmable logic controller to control the stepper motor of the reservoir water regulating gate to increase the opening position and perform physical water supply replenishment of the entity.
10. A river and lake status assessment system based on big data, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the big data-based river and lake status assessment method according to any one of claims 1-9.
