Water conservancy project flow dividing type flood control prediction system and method

Abstract

The invention relates to the technical field of distributed data processing, in particular to a hydraulic engineering distributed flood prevention prediction system and method.The system comprises a real-time monitoring module for calculating the mean value and variance of data of a water level sensor every five minutes, and an anomaly detection module for comparing the mean value difference with a threshold value, recognizing a sudden change point and sending the sudden change point to a cloud server; the window adjustment module adjusts the collection frequency according to the abnormal result, the trend analysis module predicts the flood trend by applying moving average and exponential smoothing, and the early warning generation module evaluates the water level and generates an early warning signal. According to the method, the mean value and the variance of the water level monitoring data are calculated in real time and compared with the threshold values, so that the recognition capability of the abrupt change point is improved, the data collection frequency and the window size are adjusted according to the anomaly detection result, the adaptability and efficiency of data collection are enhanced, and the water level trend is predicted through the moving average and exponential smoothing technology; and in combination with a geographic information system, the affected area is accurately evaluated, so that the early warning accuracy and the decision support effectiveness are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of split-type data processing, and particularly to a split-type flood prevention prediction system and method for water conservancy projects. Background Art

[0002] The split-type data processing technology field involves using a split (Stream) computing framework to process continuous data streams, such as data generated from sensors, social media sources, or online transactions. This technology enables data to be captured, analyzed, and stored almost in real time when it is generated, thus achieving fast data processing and decision support. Split-type processing can not only reduce the dependence on storage, but also improve processing efficiency and response speed. It is widely applied in scenarios such as financial transaction monitoring, network security, real-time advertising delivery, and Internet of Things device management.

[0003] In water conservancy projects, a split-type flood prevention prediction system refers to using split-type data processing technology to achieve real-time flood monitoring and early warning. By collecting data streams from rivers, reservoirs, rainfall, and other relevant hydrological and meteorological sensors, the system can analyze and predict possible flood events and issue alerts in a timely manner. The use of such a system in water conservancy projects lies in enhancing the response speed and efficiency of flood prevention measures, helping to prevent and mitigate the potential impact of flood disasters, especially during the rainy or typhoon seasons.

[0004] Existing technologies usually lack a mechanism for dynamically adjusting the acquisition frequency in data processing, resulting in slow response in a rapidly changing environment. For example, in sudden hydrological events, the lack of the ability to update in a timely manner may delay flood warnings and increase the risk of flood response. In addition, traditional methods do not fully integrate geographical information, limiting the accurate judgment of potential risk areas and the geographical positioning accuracy of early warnings, and weakening the overall effectiveness of prediction and disaster prevention. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a split-type flood prevention prediction system and method for water conservancy projects.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: A split-type flood prevention prediction system for water conservancy projects includes:

[0007] A real-time monitoring module receives the water level monitoring data continuously transmitted by a water level sensor, calculates the mean and variance of the collected data every five minutes, and generates a real-time monitoring result;

[0008] An anomaly detection module receives the real-time monitoring result, compares the data mean difference with a preset critical threshold, identifies the mutation points in the data, including the rise or fall of the water level change speed, and integrates and generates an anomaly detection result;

[0009] Based on the anomaly detection result, the window adjustment module adjusts the size of the data window. For the detected data anomalies, it reduces the window and increases the data collection frequency. For the data without anomalies, it increases the window and reduces the collection frequency, generating the adjusted window parameters.

[0010] Using the adjusted window parameters, the trend analysis module applies the moving average and exponential smoothing methods to the collected water level data to predict the flood trend in the next hour, generating the trend analysis result.

[0011] Based on the trend analysis result, the early warning generation module evaluates whether the predicted future water level reaches the flood warning level, combines GIS technology to determine the affected areas, and generates the warning signal.

[0012] The real-time monitoring result includes the mean value and variance. The anomaly detection result specifically refers to the mutation point and the change speed. The adjusted window parameters specifically refer to the window size and the collection frequency. The trend analysis result includes the moving average and exponential smoothing. The warning signal specifically refers to the water level warning level and the affected area.

[0013] As a further solution of the present invention, the steps for obtaining the real-time monitoring result are specifically as follows:

[0014] Receive the water level monitoring data continuously transmitted by the water level sensor, take out the water level data within every five minutes and perform data sorting. Through preprocessing of formatting each group of water level data and removing obvious error data, generate the preliminary water level data set.

[0015] Calculate the mean value and variance for each group of data in the preliminary water level data set, screen the valid data that meets the reasonable change interval in the data set, and eliminate the abnormal data beyond the upper and lower limits, generating the valid water level data set.

[0016] Based on the valid water level data set, calculate the water level mean value every five minutes, analyze the overall water level according to the water level mean value, and use the formula:

[0017]

[0018] Obtain the overall water level change rate and generate the real-time monitoring result.

[0019] Where, R represents the overall water level change rate, e i is the i-th data point in the valid water level data set, μ i is the mean value of the i-th group of water level data, w i is the weight of the i-th data point, N is the number of data groups in the valid water level data set, δ is the adjustment coefficient used to increase the stability of the calculation result, k iIt is the fluctuation correction coefficient of data point i, which is used to reflect the influence degree of the sudden change of water level.

[0020] As a further solution of the present invention, the obtaining step of the abnormal detection result is specifically as follows:

[0021] Receive the real-time monitoring result, compare the water level change speed in the data with the preset critical threshold, identify the significantly rising or falling speed, and generate a preliminary abnormal point set;

[0022] Call the preliminary abnormal point set, analyze each point, screen the actual abnormal points by calculating the difference between the water level change speed of each point and the dynamic threshold, and generate a refined abnormal point set;

[0023] Utilize the refined abnormal point set, comprehensively analyze the water level change amplitude of each abnormal point, compare it with the critical threshold, and adopt the formula:

[0024]

[0025] Calculate and quantify the abnormal degree of each point, obtain the total abnormal degree, and generate the abnormal detection result;

[0026] Among them, D represents the total abnormal degree, v j is the water level change speed of the j-th point in the refined abnormal point set, θ is the preset critical threshold, s j is the sensitivity adjustment factor of the j-th point, which reflects the sensitivity of the abnormal point to the threshold, λ is the weighting factor, which is used to adjust the influence of the logarithmic ratio, and M is the total number of data points in the refined abnormal point set.

[0027] As a further solution of the present invention, the obtaining step of the adjusted window parameter is specifically as follows:

[0028] Receive the abnormal detection result, determine the adjustment requirement of the data window size according to the abnormal degree of each data point, reduce the data window for the detected abnormal data points, and increase the window for the undetected abnormal data points, thereby generating a preliminary adjustment plan;

[0029] Based on the preliminary adjustment plan, optimize the adjustment plan of each data window, dynamically adjust the data window size according to the abnormal duration and intensity of the data points, and generate an optimized window adjustment plan;

[0030] Based on the optimized window adjustment plan, apply a mathematical model to calculate the window size, and adopt the formula:

[0031]

[0032] Adjust the window size to reflect the actual monitoring requirements, and generate the adjusted window parameter;

[0033] Among them, W new represents the adjusted data window size, and W old is the original data window size, and α is a preset adjustment coefficient used to adjust the sensitivity of the window size. E max represents the maximum outlier among all outliers, and E actual is the outlier of the current data point.

[0034] As a further solution of the present invention, the steps for obtaining the trend analysis result are specifically as follows:

[0035] Based on the adjusted window parameters, segment the collected water level data, apply the moving average method, calculate the average water level value within each window, and generate a set of smoothed water level data;

[0036] Extract each data point from the set of smoothed water level data, apply the exponential smoothing method, calculate the weighted average value, assign weights to the observed data to reflect the recent water level change trend, and generate a set of exponentially smoothed data;

[0037] Use the set of exponentially smoothed data to predict the flood trend within the next hour, using the formula:

[0038]

[0039] Calculate the predicted value of the future water level and generate the trend analysis result;

[0040] Among them, F t+1 represents the predicted water level for the next hour, S t is the moving average value at the current moment, E t is the exponentially smoothed value at the current moment, β is the weight of the moving average, γ is the coefficient affecting the exponentially smoothed value E t The influencing coefficient, δ is an additional adjustment coefficient used to adjust the influence of γ, and N is the number of reference time points to strengthen the model's response to new data.

[0041] As a further solution of the present invention, the steps for obtaining the real-time monitoring result are specifically as follows:

[0042] Analyze the trend analysis result to determine whether the water level reaches the warning level within each prediction period, apply threshold comparison to each period to judge whether it exceeds the flood warning standard, mark the warning status for each eligible period, and generate a list of warning periods;

[0043] Use GIS technology to perform spatial analysis on the geographical locations corresponding to each period in the list of warning periods to determine the areas threatened by potential floods, draw an impact map for the areas, and generate a geographical calibration impact map;

[0044] Based on the geographical calibration impact map, evaluate the flood emergency levels of multiple impact areas, apply an early warning model to predict the flood emergency response levels, and use the formula:

[0045]

[0046] Calculate the early warning levels of each area according to the regional risks and generate early warning signals;

[0047] Among them, P alert represents the early warning issuance probability of multiple regions, reflecting the high or low predictability of floods within the region. I risk is the regional risk index obtained by GIS analysis, referring to factors such as terrain and previous water level data. β0 is the intercept parameter of the model, adjusting the baseline early warning probability, and β1 is the slope parameter of the model, adjusting the influence intensity of the regional risk index I risk on the early warning probability.

[0048] A split - type flood prevention prediction method for water conservancy projects. The split - type flood prevention prediction method for water conservancy projects is executed based on the above - mentioned split - type flood prevention prediction system for water conservancy projects, and includes the following steps:

[0049] S1: Receive the water level monitoring data sent by the water level sensor every five minutes, calculate the mean and variance of the water level monitoring data, and generate real - time monitoring results;

[0050] S2: Compare the data mean difference in the real - time monitoring results with a preset critical threshold to identify mutation points, including the rise or fall of the water level change speed, and generate an anomaly detection result;

[0051] S3: According to the anomaly detection result, adjust the data window size. If data anomalies are detected, reduce the window size and increase the data acquisition frequency. If the data is normal, increase the window and reduce the acquisition frequency, and generate adjusted window parameters;

[0052] S4: Use the adjusted window parameters, apply the moving average and exponential smoothing methods to conduct trend analysis on the collected water level data, predict the flood trend in the next hour, generate a trend analysis result, and combine GIS technology to determine the affected areas and generate early warning signals.

[0053] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0054] In the present invention, by calculating the mean and variance of the water level monitoring data in real time and comparing them with the thresholds, the ability to identify mutation points is improved. By adjusting the data acquisition frequency and window size according to the anomaly detection results, the adaptability and efficiency of data acquisition are enhanced. The water level trend is predicted by using moving average and exponential smoothing techniques, and the affected areas are accurately evaluated in combination with the geographic information system, significantly improving the accuracy of early warning and the effectiveness of decision-making support. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is the system flow chart of the present invention;

[0056] Figure 2 is the flow chart of the steps for obtaining the real-time monitoring results of the present invention;

[0057] Figure 3 is the flow chart of the steps for obtaining the anomaly detection results of the present invention;

[0058] Figure 4 is the flow chart of the steps for obtaining the adjusted window parameters of the present invention;

[0059] Figure 5 is the flow chart of the steps for obtaining the trend analysis results of the present invention;

[0060] Figure 6 is the flow chart of the steps for obtaining the early warning signals of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0062] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more unless otherwise specifically defined.

[0063] Embodiment 1

[0064] Please refer to Figure 1 , the present invention provides a technical solution: a diversion type flood prevention prediction system for water conservancy projects includes:

[0065] The real-time monitoring module receives the water level monitoring data continuously transmitted by the water level sensor, calculates the mean and variance of the collected data every five minutes, and generates real-time monitoring results;

[0066] The anomaly detection module receives the real-time monitoring results, compares the data mean difference with a preset critical threshold, identifies the mutation points in the data, including the rise or fall of the water level change speed, and integrates and generates anomaly detection results;

[0067] The window adjustment module adjusts the data window size according to the anomaly detection results. For the detected data anomalies, it reduces the window to increase the data collection frequency, and for the non-anomaly data, it increases the window to reduce the collection frequency, and generates the adjusted window parameters;

[0068] The trend analysis module uses the adjusted window parameters, applies the moving average and exponential smoothing methods to the collected water level data, predicts the flood trend in the next hour, and generates trend analysis results;

[0069] The early warning generation module, based on the trend analysis results, evaluates whether the predicted future water level reaches the flood warning level, combines GIS technology to determine the affected areas, and generates early warning signals.

[0070] The real-time monitoring results include the mean and variance. The anomaly detection results are specifically the mutation points and the change speed. The adjusted window parameters specifically refer to the window size and the collection frequency. The trend analysis results include the moving average and exponential smoothing. The early warning signals are specifically the water level warning level and the affected areas.

[0071] Please refer to Figure 2 , and the specific steps for obtaining the real-time monitoring results are as follows:

[0072] Receive the water level monitoring data continuously transmitted by the water level sensor, extract the water level data within every five minutes and perform data collation. Through preprocessing such as unifying the format of each group of water level data and removing obvious error data, generate a preliminary water level data set;

[0073] The process of receiving the water level monitoring data continuously transmitted by the water level sensor involves the capture and preliminary filtering of real-time data. First, set up data collection nodes to ensure that the data for every five minutes can be completely collected. Thereafter, exclude abnormal sensor readings, such as jump and breakpoint errors, through a preset error data identification algorithm to ensure the preliminary accuracy of the data. Perform formatting processing on each group of water level data, such as unifying all data to the same timestamp accuracy and measurement unit for subsequent analysis. The processed data is temporarily stored in the data buffer for further detailed analysis and calculation, generating a preliminary water level data set.

[0074] Calculate the mean and variance for each group of data in the preliminary water level data set, filter out the valid data within the reasonable change range in the data set, eliminate the abnormal data beyond the upper and lower limits, and generate a valid water level data set;

[0075] Call the preliminary water level data set and start in-depth processing of the data, including calculating statistical indicators for each group of data, such as the mean and variance. These calculations provide a basis for subsequent data screening. The calculation of the mean and variance takes into account the time series characteristics and potential autocorrelation of each data point. By setting thresholds to judge the mean and variance, abnormal data such as surges and data with long-term trends deviating from the normal water level range are identified and eliminated. The filtered valid data is used to generate the valid water level data set E, ensuring that only reasonable and representative water level data is used for further analysis.

[0076] Based on the valid water level data set, calculate the water level mean every five minutes, analyze the overall water level according to the water level mean, and use the formula:

[0077]

[0078] Obtain the overall water level change rate and generate real-time monitoring results;

[0079] where R represents the overall water level change rate, e i is the i-th data point in the valid water level data set, μ i is the mean of the i-th group of water level data, w i is the weight of the i-th data point, N is the number of data groups in the valid water level data set, δ is an adjustment coefficient used to increase the stability of the calculation result, k i is the fluctuation correction coefficient of data point i, which is used to reflect the degree of influence of water level mutations.

[0080] Formula:

[0081]

[0082] The advantage of the formula is that it not only adjusts the weight based on the mean and the deviation of each point, but also considers the environmental factors and possible abnormal fluctuations of each measurement point by introducing the fluctuation correction coefficient k i to improve the calculation accuracy and response sensitivity of the water level change rate R.

[0083] Detailed explanation of the formula and the derivation process of formula calculation:

[0084] Assume that the valid water level data set E contains the following specific values: e1 = 150, μ1 = 145, w1 = 1.05, δ = 0.03, k1 = 1.1, N = 1. Substitute these values into the formula for calculation as follows:

[0085]

[0086]

[0087] R = 4.7619 + 4.571

[0088] R = 9.3329

[0089] The result shows that the calculated water level change rate R is 9.33, which reflects the change range of the water level within a given time period. A higher R value may indicate a larger water level change and requires further attention.

[0090] Please refer to Figure 3 , and the steps for obtaining the anomaly detection results are specifically as follows:

[0091] Receive real-time monitoring results, compare the water level change speed in the data with a preset critical threshold, identify significantly rising or falling speeds, and generate a preliminary set of anomaly points;

[0092] The received real-time monitoring results contain a large number of data points. Each data point represents the water level change speed at a specific time. These data points are used to compare with the preset critical threshold to determine which data points show abnormal speed changes. This process involves accurately measuring and recording the speed value of each data point. Then, each measured speed value is compared with the threshold to identify those speed changes that significantly deviate from the normal range. These marked points are aggregated into a preliminary set of anomaly points. This set lists all possible abnormal water level change points and lays the foundation for the next step of analysis.

[0093] Call the preliminary set of anomaly points, analyze each point, screen the actual anomaly points by calculating the difference between the water level change speed of each point and the dynamic threshold, and generate a refined set of anomaly points;

[0094] According to the preliminary set of anomaly points, during the screening process, the water level change speed of each point is calculated and compared with the dynamic threshold. This dynamic threshold is adjusted according to the past data fluctuations to ensure that only real abnormal changes are identified, thereby reducing false alarms. This step involves complex data processing techniques, including data cleaning, outlier handling, and dynamic threshold adjustment. Each data point identified as abnormal undergoes multiple verifications to ensure that it represents an actual abnormal event and will not be mislabeled due to daily data fluctuations.

[0095] Utilize the refined set of anomaly points, comprehensively analyze the water level change amplitude of each anomaly point, compare and analyze it with the critical threshold, and use the formula:

[0096]

[0097] Calculate and quantify the degree of abnormality of each point, obtain the total abnormality degree, and generate the anomaly detection result;

[0098] where D represents the total abnormality degree, and v j is the water level change speed of the j-th point in the refined anomaly point set, θ is the preset critical threshold, and s j is the sensitivity adjustment factor of the j-th point, reflecting the sensitivity of the anomaly point to the threshold, λ is the weighting factor used to adjust the influence of the logarithmic ratio, and M is the total number of data points in the refined anomaly point set.

[0099] Formula:

[0100]

[0101] The benefit of the formula is that it not only quantifies the absolute difference between the water level change of each anomaly point and the threshold, but also introduces the consideration of relative change through the logarithmic term, increasing the sensitivity of anomaly recognition. Especially when the difference between the anomaly value and the threshold is large, these differences can be more prominent.

[0102] Detailed explanation of the formula and the derivation process of formula calculation:

[0103] Assume that the refined anomaly point set Q contains three data points, and the water level change speeds v are 100, 120, and 90 respectively. The preset critical threshold θ is 110, the sensitivity adjustment factors s are 10, 20, and 10 respectively, and the weighting factor λ is set to 0.5. The calculation is as follows: The first item calculation:

[0104]

[0105] The second item calculation:

[0106]

[0107] The third item calculation:

[0108]

[0109] Calculation of the total abnormality degree D:

[0110] D = 0.978 + 0.541 + 1.900 ≈ 3.419

[0111] This result indicates that the total abnormality degree calculated according to the formula is 3.419, which means that the cumulative influence of the change amplitude and relative change of all anomaly points relative to the threshold is considered. Further analysis of this value can help determine the sensitivity and response degree of anomaly detection.

[0112] Please refer to Figure 4 , and the specific steps for obtaining the adjusted window parameter are as follows:

[0113] Receive the anomaly detection results, determine the adjustment requirements for the data window size according to the anomaly degree of each data point, reduce the data window for the data points detected with anomalies, and increase the window for the data points not detected with anomalies, thereby generating a preliminary adjustment plan;

[0114] After receiving the anomaly detection results, it is necessary to dynamically adjust the data collection frequency and window size based on the anomaly degree of the data points. This requires precisely analyzing the anomaly of each data point. According to the anomaly degree of the data points, for those data points showing significant abnormal behaviors, reduce the data window size to increase the data collection frequency, which helps to more detailedly track and analyze the development trends of these anomalies. For the data points showing normal or low anomaly degree, the data window can be appropriately increased to reduce the data collection frequency. This not only saves resources but also reduces the burden of data processing. This dynamic adjustment strategy enables the monitoring system to more flexibly respond to various data changes and optimize the efficiency of data collection and analysis.

[0115] Based on the preliminary adjustment plan, optimize the adjustment plan for each data window, dynamically adjust the data window size according to the anomaly duration and intensity of the data points, and generate an optimized window adjustment plan;

[0116] Optimizing the window adjustment strategy for each data point is a complex process. This process involves analyzing the anomaly duration and intensity of each data point, and it is necessary to dynamically adjust the window size according to the specific characteristics of these data points. This not only includes identifying the anomaly degree of each anomaly point but also considering the duration of the anomaly. This information helps to decide how to adjust the collection frequency and the size of the data window. The optimized window adjustment plan aims to ensure the real-time and accuracy of data collection, while also considering the resource efficiency and data processing ability of the system. Through such optimization, a large amount of data can be more effectively processed while maintaining a rapid response to key events.

[0117] Based on the optimized window adjustment plan, apply a mathematical model to calculate the window size, using the formula:

[0118]

[0119] Adjust the window size to reflect the actual monitoring requirements and generate the adjusted window parameters;

[0120] Among them, W new represents the adjusted data window size, W old is the original data window size, α is a preset adjustment coefficient used to adjust the sensitivity of the window size, E max represents the maximum anomaly value among all anomaly points, and E actual is the anomaly value of the current data point.

[0121] Formula:

[0122]

[0123] The benefit of the formula is that it can flexibly adjust the size of the data window according to the actual anomalies, enabling the data acquisition system to more sensitively respond to the occurrence of abnormal events, thereby optimizing the data collection and analysis process and improving the efficiency and accuracy of the monitoring system.

[0124] Detailed explanation of the formula and the derivation process of formula calculation:

[0125] Set the original data window size W old = 10 seconds, the abnormal maximum value E max = 50, the actual abnormal value E actual = 20, the adjustment coefficient α = 0.5,

[0126]

[0127] This result indicates that according to the current degree of anomaly, the new data window size should be adjusted to 13 seconds, which shows that the increase in the window size reflects a lower current degree of anomaly, and the system can reduce the data acquisition frequency and optimize resource usage.

[0128] Please refer to Figure 5 , and the specific steps for obtaining the trend analysis result are as follows:

[0129] Based on the adjusted window parameters, segment the collected water level data, apply the moving average method, calculate the average water level value within each window, and generate a set of smoothed water level data;

[0130] Based on the adjusted window parameters, optimize the data acquisition frequency by adjusting the acquisition window size. The specific method includes collecting the original water level data and adjusting the window size according to the change speed and change amplitude within different time windows to ensure the efficiency and accuracy of data acquisition, while reducing unnecessary data noise. Through this method, the key points of water level changes can be captured more accurately, and more targeted data analysis can be carried out accordingly. This process not only optimizes data storage but also improves the efficiency of data processing. The subsequent processing steps are simplified by the moving average method, and the smoothed data set is used for further analysis, providing reliable basic data for flood trend prediction.

[0131] Extract each data point from the smoothed water level data set, apply the exponential smoothing method, calculate the weighted average value, assign weights to the observed data to reflect the recent water level change trend, and generate a set of exponentially smoothed data;

[0132] Extract data from the smoothed data set. The refinement process includes applying the exponential smoothing method to weight the data. This method emphasizes the importance of recent data and gradually reduces the weight of old data to ensure the response speed and sensitivity of the model. This weighted average method is particularly suitable for rapidly changing environments, such as when the water level is rising or falling rapidly. Through this technique, the water level changes in the short term can be predicted more accurately, providing a scientific basis for flood control measures. The exponentially smoothed data set E contains a sensitive response to future water level change trends, laying a foundation for further predictive analysis.

[0133] Using the exponentially smoothed data set, predict the flood trend within the next hour using the formula:

[0134]

[0135] Calculate the predicted value of the future water level and generate the trend analysis result;

[0136] Where, F t+1 represents the predicted water level for the next hour, S t is the moving average at the current moment, E t is the exponentially smoothed value at the current moment, β is the weight of the moving average, γ is the coefficient affecting the exponentially smoothed value E t δ is an additional adjustment coefficient used to adjust the influence of γ, and N is the number of reference time points to strengthen the model's response to new data.

[0137] Formula:

[0138]

[0139] The advantage of the formula is that it combines the advantages of moving average and exponential smoothing, enhancing the model's response speed to new changes while balancing the influence of historical data, and improving the accuracy and adaptability of prediction.

[0140] Detailed explanation of the formula and the derivation process of formula calculation:

[0141] Assume that among the recently observed N data points, the average water level value S t is 150 cm, the exponentially smoothed average value E t is 155 cm, the weight factor β is set to 0.5, the adjustment coefficient γ is 0.3, and the non-linear adjustment parameter δ is 2. Substitute these values into the formula for calculation:

[0142]

[0143] F t+1 = 75 + 77.5 + 0.3·240.25

[0144] Ft+1 = 75 + 77.5 + 72.075

[0145] F t+1 = 224.575 cm

[0146] This result indicates that based on the current and previous water level data predictions, the water level will reach approximately 224.6 cm within the next hour, reflecting an upward trend. This is crucial for activating the early warning system and making timely responses. Accurate predictions help decision-makers respond promptly to avoid potential disaster risks.

[0147] Please refer to Figure 6 , and the specific steps for obtaining the early warning signal are as follows:

[0148] Analyze the trend analysis results, determine whether the water level reaches the early warning level within each prediction period, apply threshold comparison to each period, judge whether it exceeds the flood early warning standard, mark the early warning status for each eligible period, and generate a list of early warning periods;

[0149] Analyze the trend analysis results to determine which periods have predicted water levels exceeding the early warning standard. This requires a detailed comparison of past trend data, evaluating the relationship between water level data and thresholds, and making judgments based on thresholds set by historical data. Specifically, for each prediction period, if its water level data exceeds the set threshold, that period is marked as a high-risk period, and a list containing all high-risk periods is generated. This list will serve as the basic input for GIS analysis to determine specific geographical areas and conduct further spatial analysis. This process not only relies on accurate water level measurements but also requires effective data processing and analysis techniques to ensure the accuracy of threshold comparison; this list becomes the basis for implementing flood early warning and emergency response strategies.

[0150] Using GIS technology, conduct spatial analysis on the geographical locations corresponding to each period in the list of early warning periods to determine the areas threatened by potential floods, draw an impact map for the areas, and generate a geographically calibrated impact map;

[0151] Using GIS technology, conduct spatial positioning on the periods marked with potential flood risks. This step includes combining the data in the list of early warning periods L with geographical location information and performing spatial analysis through GIS software to determine which specific areas may be affected. This requires advanced geographical data processing and analysis techniques, as well as in-depth understanding of regional terrain and historical flood data. GIS technology can provide accurate maps and spatial data analysis to help decision-makers understand the specific areas that may be affected by floods, generate a geographically calibrated impact map, and provide a scientific basis for emergency response deployment. This map provides key information for subsequent emergency response and resource allocation.

[0152] According to the geographical calibration impact map, evaluate the flood emergency levels of multiple impact areas, apply the warning model to predict the flood emergency response levels, and use the formula:

[0153]

[0154] Calculate the warning level for each area based on the regional risk and generate a warning signal;

[0155] Among them, P alert represents the warning issuance probability of multiple regions, reflecting the high or low predictability of floods within the region. I risk is the regional risk index obtained through GIS analysis, referring to factors such as terrain and previous water level data. β0 is the intercept parameter of the model, adjusting the baseline warning probability, and β1 is the slope parameter of the model, adjusting the influence intensity of the regional risk index I risk on the warning probability.

[0156] Formula:

[0157]

[0158] The benefit of the formula is that by introducing the geographical risk score I risk , combined with the response coefficients β0 and β1, it can quantify the flood warning probability for each region, adapt to the risk characteristics of different regions, and provide customized warning information.

[0159] Detailed explanation of the formula and the derivation process of formula calculation:

[0160] Assume that the flood risk index I risk of a certain region is 5, the response coefficients β0 = 0.5, and β1 = 0.3. Then the formula calculation is as follows:

[0161]

[0162] This indicates that the flood warning issuance probability of this region is approximately 88%, indicating that there is a high flood risk in this area in the next period of time.

[0163] The result shows that the warning system can accurately predict and issue high-risk flood warnings, allowing relevant departments to take appropriate disaster prevention and mitigation measures.

[0164] A diversion-type flood prevention prediction method for water conservancy projects. The diversion-type flood prevention prediction method for water conservancy projects is executed based on the above-mentioned diversion-type flood prevention prediction system for water conservancy projects and includes the following steps:

[0165] S1: Receive the water level monitoring data sent by the water level sensor every five minutes, calculate the mean and variance of the water level monitoring data, and generate a real-time monitoring result;

[0166] S2: Compare the data mean difference in the real-time monitoring results with the preset critical threshold to identify mutation points, including the rise or fall of the water level change rate, and generate an anomaly detection result;

[0167] S3: Adjust the data window size according to the anomaly detection result. If data anomalies are detected, reduce the window size and increase the data collection frequency. If the data is normal, increase the window and reduce the collection frequency to generate adjusted window parameters;

[0168] S4: Use the adjusted window parameters to apply the moving average and exponential smoothing methods to analyze the trend of the collected water level data, predict the flood trend in the next hour, generate a trend analysis result, and combine with GIS technology to determine the affected area and generate a warning signal.

[0169] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A diversion-type flood prevention prediction system for water conservancy projects, characterized in that, The system comprises: The real-time monitoring module receives the water level monitoring data continuously transmitted by the water level sensor, calculates the mean and variance of the collected data every five minutes, and generates real-time monitoring results; The anomaly detection module receives the real-time monitoring results, compares the data mean difference with a preset critical threshold, identifies mutation points in the data, including an increase or decrease in the speed of water level change, and integrates to generate anomaly detection results; The window adjustment module adjusts the data window size according to the abnormality detection result. For the detected data abnormality, the window is reduced to increase the data collection frequency. For the data without abnormality, the window is increased to reduce the collection frequency, and the adjusted window parameters are generated. The trend analysis module uses the adjusted window parameters to apply moving average and exponential smoothing methods to the collected water level data to predict the flood trend forecast for the next hour and generate trend analysis results; The early warning module evaluates whether the water level is expected to reach the flood warning level in the future based on the trend analysis results, determines the affected area with the help of GIS technology, and generates an early warning signal.

2. The diversion type flood prevention prediction system for water conservancy projects according to claim 1, characterized in that The real-time monitoring results include mean and variance, the anomaly detection results specifically include mutation point and change speed, the adjusted window parameters specifically include window size and acquisition frequency, the trend analysis results include moving average and exponential smoothing, and the warning signal specifically includes water level warning level and affected area.

3. The water conservancy project diversion type flood prevention prediction system according to claim 2, characterized in that, The steps for obtaining the real-time monitoring results are specifically as follows: Receive the water level monitoring data continuously transmitted by the water level sensor, extract the water level data within five minutes and sort the data, and generate a preliminary water level data set by pre-processing each set of water level data by unifying the format and removing obviously erroneous data; Calculate the mean and variance of each set of data in the preliminary water level data set, screen the validity data in the data set that conform to the reasonable change range, remove the abnormal data that exceeds the upper and lower limits, and generate a validity water level data set; Based on the validity water level data set, the mean water level every five minutes is calculated, and the overall water level is analyzed according to the mean water level, using the formula: Get the overall water level change rate and generate real-time monitoring results; Among them, R represents the overall water level change rate, e i is the i-th data point in the set of valid water level data, μ i is the mean of the i-th group of water level data, w i is the weight of the i-th data point, N is the number of data groups in the set of valid water level data, δ is the adjustment coefficient used to increase the stability of the calculation result, k i is the fluctuation correction coefficient of data point i, which is used to reflect the degree of influence of the sudden change of the water level.

4. The water conservancy project diversion type flood prevention prediction system according to claim 3, characterized in that, The steps for obtaining the abnormal detection result are specifically as follows: Receive the real-time monitoring results, compare the water level change rate in the data with a preset critical threshold, identify the speed of significant rise or fall, and generate a preliminary abnormal point set; Calling the preliminary abnormal point set, analyzing each point, and filtering actual abnormal points by calculating the difference between the water level change speed of each point and the dynamic threshold value to generate a refined abnormal point set; By using the refined abnormal point set, the water level change amplitude of each abnormal point is comprehensively analyzed and compared with the critical threshold value, using the formula: Calculate and quantify the degree of abnormality of each point, obtain the total abnormality, and generate the abnormality detection result; Among them, D represents the total anomaly degree, v j is the water level change rate of the j-th point in the refined anomaly point set, θ is the preset critical threshold, s j is the sensitivity adjustment factor of the j-th point, reflecting the sensitivity of the anomaly point to the threshold, λ is the weighting factor used to adjust the influence of the logarithmic ratio, and M is the total number of data points in the refined anomaly point set.

5. The water conservancy project diversion type flood prevention prediction system according to claim 4, characterized in that, The steps for obtaining the adjusted window parameters are specifically as follows: Receive the anomaly detection result, determine the adjustment requirement of the data window size according to the degree of anomaly of each data point, reduce the data window for the data point where anomaly is detected, and increase the window for the data point where no anomaly is detected, thereby generating a preliminary adjustment plan; Based on the preliminary adjustment plan, optimize the adjustment plan for each data window, dynamically adjust the data window size according to the abnormal duration and intensity of data points, and generate an optimized window adjustment plan; Based on the optimized window adjustment plan, apply a mathematical model to calculate the window size, using the formula: Adjust the window size to reflect the actual monitoring requirements and generate adjusted window parameters; Among them, W new represents the adjusted data window size, W old is the original data window size, and α is a preset adjustment coefficient used to adjust the sensitivity of the window size. E max represents the largest outlier among all outliers, and E actual is the outlier of the current data point.

6. The water conservancy project diversion type flood prevention prediction system according to claim 5, characterized in that, The specific steps for obtaining the trend analysis result are as follows: Based on the adjusted window parameters, segment the collected water level data, apply the moving average method, calculate the average water level value within each window, and generate a smoothed water level data set; Extract each data point from the smoothed water level data set, apply the exponential smoothing method, calculate the weighted average value, assign weights to the observed data to reflect the recent water level change trend, and generate an exponentially smoothed data set; Using the exponentially smoothed data set, predict the flood trend within the next hour, using the formula: Calculate the predicted value of the future water level and generate a trend analysis result; Among them, F t+1 represents the predicted water level in the next hour, S t is the moving average at the current moment, E t is the exponentially smoothed value at the current moment, β is the weight of the moving average, γ is the coefficient affecting the exponentially smoothed value E t The coefficient of influence, δ is an additional adjustment coefficient used to adjust the influence of γ, and N is the number of reference time points to strengthen the model's response to new data.

7. The water conservancy project diversion type flood prevention prediction system according to claim 6, characterized in that, The specific steps for obtaining the warning signal are as follows: Analyze the trend analysis result, determine whether the water level reaches the warning level within each prediction period, apply threshold comparison to each period to judge whether it exceeds the flood warning standard, mark the warning status for each eligible period, and generate a warning period list; Use GIS technology to perform spatial analysis on the geographical locations corresponding to each period in the warning period list, determine the areas threatened by potential floods, draw an impact map for the areas, and generate a geographically calibrated impact map; According to the geographically calibrated impact map, evaluate the flood emergency levels of multiple affected areas, apply a warning model to predict the flood emergency response level, using the formula: Calculate the warning level for each area according to the regional risk and generate a warning signal; Among them, P alert represents the early warning issuance probability of multiple regions, reflecting the high or low predictability of floods within the regions. I risk is the regional risk index obtained from GIS analysis, referring to factors such as terrain and previous water level data. β0 is the intercept parameter of the model, adjusting the baseline early warning probability, and β1 is the slope parameter of the model, regulating the influence intensity of the regional risk index I risk on the early warning probability.

8. A diversion type flood prevention prediction method for water conservancy projects, characterized in that, Execute according to the water conservancy project diversion type flood prevention prediction system described in any one of claims 1-7, including the following steps: Receive the water level monitoring data sent by the water level sensor every five minutes, calculate the mean and variance of the water level monitoring data, and generate a real-time monitoring result; Compare the data mean difference in the real-time monitoring result with a preset critical threshold to identify mutation points, including the rise or fall of the water level change speed, and generate an anomaly detection result; According to the anomaly detection result, adjust the data window size. If data anomalies are detected, reduce the window size and increase the data collection frequency. If the data is normal, increase the window and reduce the collection frequency, and generate adjusted window parameters; Using the adjusted window parameters, apply the moving average and exponential smoothing methods to perform trend analysis on the collected water level data, predict the flood trend in the next hour, generate a trend analysis result, and combine GIS technology to determine the affected areas and generate a warning signal.

Images