Dynamic Early Warning Method and System for Water Conservancy Monitoring Data, Electronic Device, Storage Medium

Abstract

The present invention relates to the technical field of water conservancy data monitoring and early warning, and discloses a dynamic early warning method and system for water conservancy monitoring data, an electronic device, and a storage medium, including: collecting water level data and rainfall data of a reservoir or a river in real time, and calculating the sliding window difference of the water level and rainfall; setting a basic early warning threshold for the water level according to the sliding window difference; dynamically adjusting the early warning threshold by using a regression algorithm model based on a support vector machine in combination with historical data and the water level data and rainfall data collected in real time according to flood seasons and non-flood seasons; performing fuzzification processing on the water level data and rainfall data collected in real time and the adjusted water level height early warning threshold, and making a decision according to fuzzy rules to determine the type of early warning issued. By introducing a regression algorithm model and a decision-making algorithm based on fuzzy logic, the dynamic adjustment of the early warning threshold and the intelligence of the early warning decision are realized, the accuracy and adaptability of the early warning system are improved, and the situations of false alarms and missed alarms are reduced.

Description

Technical Field

[0001] This application relates to the technical field of water conservancy data monitoring and early warning, and particularly relates to a dynamic early warning method and system for water conservancy monitoring data, an electronic device, and a storage medium. Background Art

[0002] Existing water conservancy data monitoring systems, as important tools for water conservancy management and disaster early warning, usually rely on various sensors and monitoring devices to collect key hydrological parameters such as water level, rainfall, flow rate, and water quality in real time or regularly. These systems have improved the collection efficiency and quality of hydrological data to a certain extent and provided valuable decision-making support for water conservancy management departments. However, with the continuous improvement of water conservancy management requirements and the rapid development of technology, a series of problems have gradually emerged in existing systems.

[0003] Specifically, existing water conservancy data monitoring systems have problems with insufficient timeliness in data collection. Due to limitations in sensor performance, data transmission networks, or equipment maintenance, some data cannot be collected in real time, resulting in a time difference between the monitored data and the actual hydrological situation. This not only affects the timeliness of the data but also may cause decision-makers to not be able to obtain the latest hydrological information at critical moments.

[0004] In addition, existing systems also have obvious shortcomings in data processing capabilities. With the continuous increase in the number of hydrological monitoring points and the improvement of monitoring frequencies, the amount of data that the system needs to process has increased sharply. However, the data processing capabilities of many existing systems have not kept up with this trend, resulting in low data processing efficiency and even possible data loss or processing errors. This not only affects the accuracy and integrity of the data but also limits the application value of the data in water conservancy management.

[0005] In terms of the early warning mechanism, existing systems also have imperfections. Currently, many systems still use fixed early warning thresholds for early warning and cannot dynamically adjust the early warning thresholds according to the real-time monitored hydrological data. This static early warning mechanism cannot adapt to the complex and changing hydrological environment, resulting in the accuracy and timeliness of the early warning being affected. Especially in emergency situations such as extreme weather or sudden floods, if the early warning system cannot issue early warning information in a timely and accurate manner, it will pose a serious threat to the lives and property safety of the people.

[0006] In summary, existing water conservancy data monitoring systems have obvious deficiencies in data collection, data processing, and early warning mechanisms. Among them, the imperfection of the early warning mechanism is particularly prominent because the early warning system, as an important part of water conservancy management, its accuracy and timeliness are directly related to the lives and property safety of the people. Summary of the Invention

[0007] In view of this, the purpose of the embodiments of the present application is to provide a dynamic early warning method and system for water conservancy monitoring data, an electronic device, and a storage medium, so as to solve the obvious deficiencies in aspects such as data collection, data processing, and early warning mechanisms in existing water conservancy data monitoring systems.

[0008] According to the first aspect of the embodiments of the present application, a dynamic early warning method for water conservancy monitoring data is provided, including:

[0009] Real-time collect the water level data and rainfall data of the reservoir or river course;

[0010] According to the real-time collected water level data and rainfall data, calculate the sliding window difference of the water level and rainfall;

[0011] According to the sliding window difference and expert experience, set the basic early warning threshold of the water level;

[0012] According to the flood season and non-flood season, combined with historical data and real-time collected water level data and rainfall data, use a regression algorithm model based on support vector machines to dynamically adjust the early warning threshold. Specifically, first, extract the water level data and rainfall data of the flood season and non-flood season from historical data and perform data cleaning to form a training set; then, combined with the initial early warning level provided by the basic early warning threshold, use the regression algorithm model to train the training set to obtain the dynamic adjustment parameters of the early warning threshold, so as to obtain a trained regression algorithm model; finally, input the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height early warning threshold;

[0013] Perform fuzzy processing on the real-time collected water level data, rainfall data, and the adjusted water level height early warning threshold, and make a decision according to the fuzzy rules to judge the type of early warning to be issued.

[0014] According to the second aspect of the embodiments of the present application, a dynamic early warning system for water conservancy monitoring data is provided, including:

[0015] An acquisition module for real-time collecting the water level data and rainfall data of the reservoir or river course;

[0016] A calculation module for calculating the sliding window difference of the water level and rainfall according to the real-time collected water level data and rainfall data;

[0017] A setting module for setting the basic early warning threshold of the water level according to the sliding window difference and expert experience;

[0018] A dynamic threshold module, which is used to dynamically adjust the warning threshold according to flood seasons and non-flood seasons, in combination with historical data and real-time collected water level data and rainfall data, by using a regression algorithm model based on support vector machines. Specifically, it includes: First, extract the water level data and rainfall data of flood seasons and non-flood seasons from historical data and perform data cleaning to form a training set; Then, combine the initial warning water level provided by the basic warning threshold, and use the regression algorithm model to train the training set to obtain the dynamic adjustment parameters of the warning threshold, so as to obtain a trained regression algorithm model; Finally, input the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height warning threshold;

[0019] A warning module, which is used to perform fuzzification processing on the real-time collected water level data, rainfall data, and the adjusted water level height warning threshold, make decisions according to fuzzy rules, and judge the type of warning to be issued.

[0020] According to the third aspect of the embodiments of the present application, there is provided an electronic device, including:

[0021] One or more processors;

[0022] A memory, which is used to store one or more programs;

[0023] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in the first aspect.

[0024] According to the third aspect of the embodiments of the present application, there is provided a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method as described in the first aspect are implemented.

[0025] The technical solutions provided by the embodiments of the present application may include the following beneficial effects:

[0026] As can be seen from the above embodiments, the water level data and rainfall data collected in real time by the present invention provide real-time monitoring capabilities for water conservancy safety. Calculate the sliding window difference of the water level and rainfall according to the water level data and rainfall data collected in real time; set the basic warning threshold of the water level according to the sliding window difference; according to the flood season and non-flood season, combine historical data with the water level data and rainfall data collected in real time, and use a regression algorithm model based on support vector machines to dynamically adjust the warning threshold to adapt to different environmental and condition changes. Fuzzify the water level data, rainfall data collected in real time, and the adjusted water level height warning threshold, and make a decision according to fuzzy rules to determine the type of warning to be issued. This application can automatically adjust the warning threshold according to the hydrological data monitored in real time, and issue warnings for abnormal water levels, abnormal rainfall values, and rainfall mutations based on a decision-making algorithm based on fuzzy logic, thereby improving the accuracy and timeliness of warnings. Through this improvement, the present invention aims to solve the problem that existing systems cannot adapt to complex and changing hydrological environments, and provide more scientific and effective decision-making support for water conservancy management departments.

[0027] By introducing a regression algorithm model based on SVM and a decision-making algorithm based on fuzzy logic, this application realizes the dynamic adjustment of the warning threshold and the intelligence of warning decision-making, improves the accuracy and adaptability of the warning system, and reduces false alarms and missed alarms.

[0028] In addition, the present invention also has an intelligent warning decision-making function. It can not only monitor data, but also fuzzify the water level data, rainfall data collected in real time, and the adjusted water level height warning threshold, and make a decision according to fuzzy rules to determine the type of warning to be issued. The need for human intervention is reduced through the intelligent decision-making process, and the risk of false alarms and missed alarms caused by operational errors is lowered. By reducing false alarms and missed alarms, the present invention can improve the work efficiency of water conservancy safety management personnel, enabling them to focus their energy on issues that truly require attention.

[0029] In summary, by real-time monitoring and intelligent processing of water conservancy data, the present invention significantly improves the accuracy and efficiency of the warning system. It provides more effective protection for water conservancy safety, ensures the rational use of water resources, and the efficient implementation of disaster prevention and mitigation work. By reducing false alarms and missed alarms, the present invention can also save resources and avoid unnecessary emergency responses and economic losses. In short, the present invention has significant beneficial effects in the field of water conservancy safety and has important practical application value for improving water conservancy management levels and ensuring public safety.

[0030] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings here are incorporated into and form a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application.

[0032] Figure 1 is a flowchart of a method for dynamically warning water conservancy monitoring data shown according to an exemplary embodiment.

[0033] Figure 2 is a block diagram of a system for dynamically warning water conservancy monitoring data shown according to an exemplary embodiment.

[0034] Figure 3 is a schematic structural diagram of an electronic device shown according to an exemplary embodiment. Detailed Description of the Invention

[0035] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. On the contrary, they are merely examples of systems and methods consistent with some aspects of this application as detailed in the appended claims.

[0036] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms "a", "the", and "said" used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0037] Figure 1 is a flowchart of a method for dynamically warning water conservancy monitoring data shown according to an exemplary embodiment, as Figure 1 shown, the method is applied to a terminal and may include the following steps:

[0038] Step 1: Real-time collect the water level data and rainfall data of the reservoir or river course.

[0039] Specifically, use sensors to collect real-time water level data and rainfall data every minute, thereby providing real-time monitoring capabilities for water conservancy safety.

[0040] Step 2: Calculate the sliding window differences of the water level and rainfall based on the real-time collected water level data and rainfall data.

[0041] Specifically, set the sliding window size to n minutes; according to the real-time collected water level data, calculate the sliding window difference of the water level, where the water level sliding window difference is the difference between the current water level and the water level of the previous n minutes; according to the real-time collected rainfall data, calculate the sliding window difference of the rainfall, where the rainfall sliding window difference is the difference between the current rainfall and the rainfall of the previous n minutes.

[0042] In this embodiment, the sliding window size is set to 5 minutes (i.e., the collection interval, not limited to 5 minutes and can be set by oneself). For example, the water level sliding window difference is the difference between the current water level and the water level of the previous 5 minutes, and the rainfall sliding window difference is the difference between the current rainfall and the rainfall of the previous 5 minutes.

[0043] Step 3: Set the basic warning threshold of the water level according to the sliding window difference and expert experience.

[0044] Specifically, assume that the water level sliding window difference is the difference between the current water level and the water level of the previous 5 minutes. Based on these differences and expert experience, the basic warning threshold of the water level will be set. For example, the suspicious threshold can be set to a sliding window difference of 0.6m to 1m, and the error threshold is a sliding window difference exceeding 1m.

[0045] Here, the basic warning value is the warning value set by combining the sliding window difference of the real-time data and expert experience, and it is used as a reference value for subsequent model calculations, which can more accurately set the subsequent dynamic warning threshold. The sliding window difference can capture the short-term change trend of the water level, while expert experience provides insights into historical data and long-term trends. This combination ensures that the warning threshold can respond to immediate hydrological changes and also take into account long-term seasonal and periodic changes.

[0046] Step 4: According to the flood season and non-flood season, combine the historical data after data cleaning, the real-time collected water level data, and rainfall data, and use a regression algorithm model based on support vector machine (SVM) to dynamically adjust the warning threshold.

[0047] Specifically, first, extract the water level data and rainfall data of the flood season and non-flood season from the historical data, and perform data cleaning, including filtering and outlier detection, to form a training set; then, combine the initial warning level provided by the basic warning threshold, and use the regression algorithm model to train the training set to obtain the dynamic adjustment parameters of the warning threshold, so as to obtain a trained regression algorithm model; finally, input the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height warning threshold.

[0048] Algorithm formula:

[0049] SVM regression algorithm model:

[0050] Among them, αi is the weight corresponding to the support vector, Ki(x, xi) is the kernel function (Gaussian kernel function), and b is the bias term.

[0051] For example, it is necessary to design a flood warning system for a city located in the Yangtze River Basin. This system needs to dynamically adjust the warning threshold according to water level and rainfall data during the flood season and non-flood season to ensure that an alarm is issued in a timely manner before a flood occurs.

[0052] Data parameters:

[0053] (1) Flood season: from April to October every year.

[0054] (2) Non-flood season: from January to March and from November to December every year.

[0055] (3) Historical data: including water level and rainfall data for the past 5 years, recorded every minute, and the corresponding historical warning thresholds.

[0056] (4) Real-time data: current water level and rainfall data per minute, and calculate the water level difference and rainfall difference in a 5-minute window.

[0057] The specific implementation steps of Step 4 are described in detail as follows:

[0058] 1. Data extraction and cleaning:

[0059] (1) Extract the water level and rainfall data for the flood season and non-flood season in the past 5 years from the historical database.

[0060] (2) Use filtering and outlier detection to clean the data and form a training set.

[0061] (3) Ensure that the data set contains enough sample points. For example, extract 1000 sample points for both the flood season and non-flood season.

[0062] 2. SVM regression algorithm model training:

[0063] (1) Use the extracted data set to train the SVM regression algorithm model.

[0064] Select the Gaussian kernel function (RBF) as the kernel function, that is , where γ is the kernel function parameter.

[0065] (2) Select the best model parameters through methods such as cross-validation, including the penalty parameter C and the kernel function parameter γ.

[0066] Select appropriate combinations of the penalty parameter C and the kernel function parameter γ to obtain the parameter configuration with the highest accuracy and the strongest generalization ability. Here, the penalty parameter is only used in the training and validation process of the model.

[0067] After the training is completed, the weights αi corresponding to the support vectors and the bias term b are obtained.

[0068] The above-mentioned dynamically adjusted parameters include the penalty parameter C, the kernel function parameter γ, the weights αi, and the bias term b.

[0069] 3. Dynamically adjust the warning threshold:

[0070] (1) Input the real-time data into the trained SVM regression algorithm model.

[0071] (2) Calculate the current warning threshold, which is the predicted value output by the model.

[0072] (3) Compare the current warning threshold with the actual 5-minute window water level difference data to determine whether a warning needs to be issued.

[0073] Suppose the 5-minute window water level warning threshold for the current flood season obtained through the SVM model is 0.67 meters, and the real-time monitoring data shows that the current 5-minute window water level difference is 0.51 meters. Then the system will not issue a warning. If the real-time 5-minute window water level difference reaches or exceeds 0.67 meters, the system will issue a corresponding alarm according to the set warning mechanism.

[0074] Example of applying the algorithm formula:

[0075] Suppose the output of the trained SVM model is:

[0076]

[0077] Among them, x1, x2, and x3 are support vectors, K is the Gaussian kernel function, and b is the bias term.

[0078] When the real-time data "x real-time" is input, the following is calculated:

[0079]

[0080] According to the calculation result, if f(x real-time) is greater than or equal to the set warning threshold, a warning is issued.

[0081] The SVM regression algorithm model calculates and outputs the current warning threshold based on the input real-time 5-minute window difference data, combined with the flood season and non-flood season characteristics in the historical data. This warning threshold is obtained through the comprehensive analysis of historical data and real-time data and can reflect the current water situation and seasonal characteristics.

[0082] Early warning judgment: The system compares the calculated early warning threshold with the real-time water level difference data in a 5-minute window. If the real-time water level difference in the 5-minute window reaches or exceeds the early warning threshold, the system will automatically trigger the early warning mechanism and issue corresponding alarms. At the same time, the system can also predict the future water situation according to the changing trend of the early warning threshold.

[0083] To illustrate the calculation process of f(x real) more specifically, assume the following data:

[0084] (1) The specific values of the support vectors x1, x2, and x3 are (water level difference in 5-minute window: 0.49 m, rainfall: 50 mm), (water level difference in 5-minute window: 0.51 m, rainfall: 75 mm), and (water level difference in 5-minute window: 0.54 m, rainfall: 100 mm) respectively.

[0085] (2) The parameter γ of the Gaussian kernel function is 0.1.

[0086] (3) The bias term b is 0.5.

[0087] When the input real-time data "x real" is (water level: 0.43 m, rainfall: 60 mm), the following can be calculated:

[0088]

[0089] Among them, represents the square of the Euclidean distance between the vector (a, b) and the vector (c, d).

[0090] In this way, a specific value of f(x real) can be obtained. If the calculation result is 0.6 m (this value is only an example, and the actual result varies depending on the model parameters and data), and the current real-time 5-minute window water level is 0.43 m, then since f(x real) (0.6 m) is greater than the real-time data (0.43 m), the system will not issue an early warning. However, if the real-time water level continues to rise and reaches or exceeds 0.6 m, the system will issue corresponding alarms according to the set early warning mechanism.

[0091] The system can dynamically adjust the early warning threshold according to the real-time water level and rainfall data, combined with the flood season and non-flood season characteristics of historical data, so as to ensure that alarms are issued in a timely manner before floods occur and provide strong technical support for flood control work.

[0092] Step 5: Fuzzify the real-time collected water level data and rainfall data, make decisions according to fuzzy rules, and judge the type of early warning to be issued.

[0093] In this embodiment, the fuzzification process uses the Gaussian function, and the fuzzy rules are set according to historical data and expert experience.

[0094] Algorithm formula:

[0095] Fuzzification: μA(x) = exp(-((x - μA)² / (2σA²)))

[0096] Where μA is the mean of the fuzzy set A, and σA is the standard deviation of the fuzzy set A. Specific embodiment:

[0098] Suppose a flood warning system is being designed for a city in the Yangtze River Basin. The system needs to dynamically adjust the warning threshold according to water level and rainfall data during the flood season and non-flood season to ensure timely alerts before floods occur. In step 4, the SVM regression algorithm model has been used to dynamically adjust the warning threshold. Now, based on the adjusted dynamic warning threshold, combined with the decision-making algorithm based on fuzzy logic, water level anomaly warnings, rainfall anomaly value warnings, and rainfall mutation warnings are issued.

[0099] Data parameters:

[0100] (1) Flood season: from June to September every year.

[0101] (2) Non-flood season: from January to May and from October to December every year.

[0102] (3) Historical data: including water level and rainfall data for the past 5 years, recorded once per minute.

[0103] (4) Real-time collected data: current water level and rainfall data per minute, as well as calculating the water level difference and rainfall difference in a 5-minute window.

[0104] The specific implementation of step 5 is as follows:

[0105] (1) Fuzzification:

[0106] 1) Fuzzify the real-time collected water level data and rainfall data. The mean and standard deviation of the fuzzy sets are set according to historical data and expert experience, and can be adjusted according to actual hydrological conditions. Specifically, assume that the fuzzy sets we set include three fuzzy sets: "low", "medium", and "high".

[0107] 2) For the water level data, set the mean and standard deviation of the three fuzzy sets as follows:

[0108] Low water level (A1): mean μA1 = 0.35 m, standard deviation σA1 = 0.05 m;

[0109] Medium water level (A2): mean μA2 = 0.45 m, standard deviation σA2 = 0.05 m;

[0110] High water level (A3): mean μA3 = 0.55 m, standard deviation σA3 = 0.05 m;

[0111] 3) For rainfall data, set the means and standard deviations of the three fuzzy sets as follows:

[0112] Light rainfall (B1): mean μB1 = 10 mm, standard deviation σB1 = 5 mm;

[0113] Moderate rainfall (B2): mean μB2 = 30 mm, standard deviation σB2 = 5 mm;

[0114] Heavy rainfall (B3): mean μB3 = 50 mm, standard deviation σB3 = 5 mm;

[0115] (2) Fuzzy rule setting:

[0116] 1) Based on historical data and expert experience, set the following fuzzy rules:

[0117] For water level and rainfall data at the same gauging station. If the water level is "high" and the rainfall is "heavy", issue a water level anomaly warning.

[0118] If the rainfall is "heavy" and the duration exceeds 2 hours, issue a rainfall anomaly value warning.

[0119] If the rainfall suddenly changes from "light" to "heavy" within 1 hour, issue a rainfall mutation warning.

[0120] Table 1 Reference table for fuzzy rule setting implementation examples:

[0121]

[0122] (3) Decision algorithm:

[0123] 1) Fuzzify the real-time data to obtain the membership degrees of each data point for the fuzzy sets.

[0124] 2) Make a decision based on the set fuzzy rules in combination with the membership degrees.

[0125] 3) If the decision result meets the warning conditions, issue the corresponding warning information.

[0126] For example, assume the current real-time water level is 0.52 m and the real-time rainfall is 45 mm. According to the fuzzification formula, calculate the membership degrees of the water level and rainfall fuzzy sets respectively:

[0127] Table 2 Reference table for water level fuzzification processing implementation examples:

[0128]

[0129] Table 3 Reference Table for Rainfall Fuzzification Processing Examples:

[0130]

[0131] According to the set fuzzy rules, decisions are made in combination with the membership degrees. Specifically, if the membership degree of the "high" fuzzy set of water level data exceeds that of other sets, and the membership degree of rainfall data for the "large" fuzzy set also exceeds that of other sets, then a water level anomaly warning is issued. Similarly, other types of warnings can be judged based on the specific values of fuzzy rules and membership degrees.

[0132] Table 4 Reference Table for Intelligent Decision-Making Examples:

[0133]

[0134] In this way, the flood warning system can, based on real-time data and dynamically adjusted warning thresholds, combine with the fuzzy logic decision algorithm to issue corresponding warning information in a timely manner, thereby effectively preventing flood disasters.

[0135] As can be seen from the above embodiments, the water level data and rainfall data collected in real time by the present invention provide real-time monitoring capabilities for water conservancy safety. Calculate the sliding window difference of water level and rainfall based on the water level data and rainfall data collected in real time; set the basic warning threshold of the water level according to the sliding window difference; according to the flood season and non-flood season, combine historical data and the water level data and rainfall data collected in real time, and use a regression algorithm model based on support vector machines to dynamically adjust the warning threshold to adapt to different environmental and condition changes. Fuzzify the water level data and rainfall data collected in real time and the adjusted water level height warning threshold, and make decisions according to fuzzy rules to judge the types of warnings issued. This application can automatically adjust the warning threshold based on the hydrological data monitored in real time, and issue water level anomaly warnings, rainfall anomaly value warnings, and rainfall mutation warnings based on the decision algorithm of fuzzy logic, thereby improving the accuracy and timeliness of warnings. Through this improvement, the present invention aims to solve the problem that the existing system cannot adapt to complex and changeable hydrological environments, and provide more scientific and effective decision-making support for water conservancy management departments.

[0136] This application realizes the dynamic adjustment of the warning threshold and the intelligence of warning decision-making by introducing a regression algorithm model based on SVM and a decision algorithm based on fuzzy logic, improves the accuracy and adaptability of the warning system, and reduces false alarms and missed alarms.

[0137] Corresponding to the embodiments of the dynamic warning method for water conservancy monitoring data described above, this application also provides embodiments of a dynamic warning system for water conservancy monitoring data.

[0138] Figure 2It is a block diagram of a dynamic early warning system for water conservancy monitoring data shown according to an exemplary embodiment. Refer to Figure 2 , the system includes:

[0139] The acquisition module 1 is used to collect the water level data and rainfall data of the reservoir or river channel in real time;

[0140] The calculation module 2 is used to calculate the sliding window difference of the water level and rainfall according to the water level data and rainfall data collected in real time;

[0141] The setting module 3 is used to set the basic early warning threshold of the water level according to the sliding window difference and expert experience;

[0142] The dynamic threshold module 4 is used to dynamically adjust the early warning threshold by combining historical data and the water level data and rainfall data collected in real time during the flood season and non-flood season, using a regression algorithm model based on support vector machines. Specifically, first, extract the water level data and rainfall data of the flood season and non-flood season from the historical data, and perform data cleaning to form a training set; then, combine the initial warning water level provided by the basic early warning threshold, and use the regression algorithm model to train the training set to obtain the dynamic adjustment parameters of the early warning threshold, so as to obtain a trained regression algorithm model; finally, input the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height early warning threshold;

[0143] The early warning module 5 is used to perform fuzzy processing on the water level data, rainfall data collected in real time and the adjusted water level height early warning threshold, and make a decision according to the fuzzy rules to judge the type of early warning to be issued.

[0144] Regarding the system in the above embodiment, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0145] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The system embodiments described above are only illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement it without creative work.

[0146] Correspondingly, the present application further provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the water conservancy monitoring data dynamic early warning method as described above. As Figure 3 shown, it is a hardware structure diagram of a device with any data processing ability where the water conservancy monitoring data dynamic early warning system provided by the embodiment of the present invention is located. Except for Figure 3 the processors and memory shown, any device with data processing ability where the system is located in the embodiment usually includes other hardware according to the actual functions of the device with any data processing ability, which will not be elaborated here.

[0147] Correspondingly, the present application further provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the water conservancy monitoring data dynamic early warning method as described above is implemented. The computer-readable storage medium may be an internal storage unit of any device with data processing ability described in any of the foregoing embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be an external storage device, such as a plug-in hard disk, a Smart Media Card (SMC), an SD card, a Flash Card, etc. equipped on the device. Further, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing ability. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing ability, and may also be used to temporarily store the data that has been output or will be output.

[0148] Those skilled in the art will readily conceive of other implementations of the present application after considering the specification and practicing the content disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include the common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present application are pointed out by the claims.

[0149] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A dynamic early warning method for water conservancy monitoring data, characterized in that: include: Collect water level data and rainfall data of reservoirs or rivers in real time; Calculate the sliding window difference of water level and rainfall based on the real-time collected water level data and rainfall data; According to the sliding window difference and expert experience, a basic warning threshold of the water level is set; According to the flood season and the non-flood season, combined with historical data and real-time collected water level data and rainfall data, a regression algorithm model based on a support vector machine is used to dynamically adjust the warning threshold, specifically including: first, extracting the water level data and rainfall data of the flood season and the non-flood season from the historical data and performing data cleaning to form a training set; then, combining the initial warning water level provided by the basic warning threshold, using the regression algorithm model to train the training set, obtain the dynamic adjustment parameters of the warning threshold, and thus obtain a trained regression algorithm model; finally, inputting the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height warning threshold; The real-time collected water level data, rainfall data and adjusted water level height warning threshold are fuzzy processed, and decisions are made based on fuzzy rules to determine the type of warning to be issued; The fuzzy processing includes: 1) Perform fuzzy processing on the water level and rainfall data in real-time data, and set the fuzzy sets including "low", "medium" and "high"; 2) For the water level data, the means and standard deviations of the three fuzzy sets are set as follows: Low water level (A1): mean μA1 = 0.35 m, standard deviation σA1 = 0.05 m; Medium water level (A2): mean μA2 = 0.45 m, standard deviation σA2 = 0.05 m; High water level (A3): mean μA3 = 0.55 m, standard deviation σA3 = 0.05 m; 3) For rainfall data, the means and standard deviations of the three fuzzy sets are set as follows: Light rainfall (B1): mean μB1 = 10 mm, standard deviation σB1 = 5 mm; Moderate rainfall (B2): mean μB2 = 30 mm, standard deviation σB2 = 5 mm; Heavy rainfall (B3): mean μB3 = 50 mm, standard deviation σB3 = 5 mm; Wherein, the fuzzy rule setting is: If the water level is "high" and the rainfall is "heavy", an abnormal water level warning is issued; If the rainfall is "heavy" and lasts for more than 2 hours, an abnormal rainfall warning will be issued; If the rainfall suddenly changes from "light" to "heavy" within 1 hour, a sudden rainfall warning will be issued; Perform fuzzy processing on real-time data to obtain the membership degree of each data point to the fuzzy set; Fuzzification: μA(x) = exp(-((x - μA)² / (2σA²))) Among them, μA is the mean of fuzzy set A, and σA is the standard deviation of fuzzy set A.

2. The method according to claim 1, characterized in that: Based on the real-time collected water level data and rainfall data, the sliding window difference of water level and rainfall is calculated, including: Set the sliding window size to n minutes; Calculate the sliding window difference of the water level based on the water level data collected in real time, where the sliding window difference of the water level is the difference between the current water level and the water level in the previous n minutes; The sliding window difference of rainfall is calculated based on the rainfall data collected in real time, where the sliding window difference of rainfall is the difference between the current rainfall and the rainfall in the previous n minutes.

3. The method according to claim 1, characterized in that: The regression algorithm model is as follows: Among them, αi is the weight corresponding to the support vector, Ki(x, xi) is the kernel function, x is the data point to be predicted, xi is the support vector in the training set, and b is the bias term.

4. The method according to claim 1, characterized in that: The training process of the regression algorithm model is as follows: (1) Extracting flood season and non-flood season water level and rainfall data from the historical database as training sets, and using the training sets to train the regression algorithm model; selecting the Gaussian kernel function (RBF) as the kernel function in the training, that is, where γ is the kernel function parameter; (2) Select the best model parameters through cross-validation method; (3) After training is completed, the weight αi and bias term b corresponding to the support vector are obtained.

5. The method according to claim 1, characterized in that Make decisions based on fuzzy rules to determine the type of warning to be issued, including: Make decisions based on the set fuzzy rules and membership degree; If the decision result meets the warning conditions, the corresponding warning information will be issued.

6. A dynamic early warning system for water conservancy monitoring data, characterized in that: include: The collection module is used to collect water level data and rainfall data of reservoirs or rivers in real time; A calculation module is used to calculate the sliding window difference of water level and rainfall according to the real-time collected water level data and rainfall data; A setting module, used for setting a basic warning threshold of the water level according to the sliding window difference and expert experience; The dynamic threshold module is used to dynamically adjust the warning threshold according to the flood season and non-flood season, combined with historical data and real-time collected water level data and rainfall data, using a regression algorithm model based on a support vector machine. Specifically, it includes: first, extracting the water level data and rainfall data of the flood season and non-flood season from the historical data and performing data cleaning to form a training set; then, combining the initial warning water level provided by the basic warning threshold, using the regression algorithm model to train the training set, and obtaining the dynamic adjustment parameters of the warning threshold, thereby obtaining a trained regression algorithm model; finally, inputting the real-time data into the trained regression algorithm model to obtain the dynamically adjusted water level height warning threshold; The early warning module is used to perform fuzzy processing on the real-time collected water level data, rainfall data and the adjusted water level height early warning threshold, make decisions based on fuzzy rules, and determine the type of early warning to be issued; The fuzzy processing includes: 1) Perform fuzzy processing on the water level and rainfall data in real-time data, and set the fuzzy sets including "low", "medium" and "high"; 2) For the water level data, the means and standard deviations of the three fuzzy sets are set as follows: Low water level (A1): mean μA1 = 0.35 m, standard deviation σA1 = 0.05 m; Medium water level (A2): mean μA2 = 0.45 m, standard deviation σA2 = 0.05 m; High water level (A3): mean μA3 = 0.55 m, standard deviation σA3 = 0.05 m; 3) For rainfall data, the means and standard deviations of the three fuzzy sets are set as follows: Light rainfall (B1): mean μB1 = 10 mm, standard deviation σB1 = 5 mm; Moderate rainfall (B2): mean μB2 = 30 mm, standard deviation σB2 = 5 mm; Heavy rainfall (B3): mean μB3 = 50 mm, standard deviation σB3 = 5 mm; Wherein, the fuzzy rule setting is: If the water level is "high" and the rainfall is "heavy", an abnormal water level warning is issued; If the rainfall is "heavy" and lasts for more than 2 hours, an abnormal rainfall warning will be issued; If the rainfall suddenly changes from "light" to "heavy" within 1 hour, a sudden rainfall warning will be issued; Perform fuzzy processing on real-time data to obtain the membership degree of each data point to the fuzzy set; Fuzzification: μA(x) = exp(-((x - μA)² / (2σA²))) Among them, μA is the mean of fuzzy set A, and σA is the standard deviation of fuzzy set A.

7. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images