TWI832767B - Hydrological data analysis method and hydrological data analysis system - Google Patents

Abstract

A hydrological data analysis method includes implementing following steps by a processor: obtaining hydrological data of a water area of from a sensing module; filtering the hydrological data; determining whether the hydrological data is normal data or abnormal data; when determining the hydrological data is the abnormal data, selecting one of a plurality of data fixing models as a selected data fixing model according to the type of the water area; inputting the abnormal data into the selected data fixing model to fix the abnormal data. By the aforementioned method, the loss data area or the abnormal data points in the hydrological data may be fixed according to the different water areas, thereby improving the accuracy and the completeness of the hydrological data.

Description

Hydrological data analysis method and hydrological data analysis system

The present invention relates to the technical field of hydrological data, in particular to a hydrological data analysis method and a hydrological data analysis system capable of repairing hydrological data.

Natural disasters always bring disasters to human life, and the frequency of extreme disasters has become more frequent in recent years. Therefore, the importance of natural disaster analysis is increasing day by day.

When the amount of wind and rain brought by a typhoon is too large, the amount of floods and sediment increases instantly. After the sediment reaches the starting flow rate, it mixes with water and begins to be transported by gravity. A large amount of sand-containing water enters the reservoir and causes siltation. Since the amount of sediment in the reservoir is too large, sensors in the reservoir area often cannot operate normally due to environmental influences or measurement limitations, resulting in incomplete hydrological data generated by water flow sensors or sediment monitors. Leading to difficulties in later analysis of natural disasters.

In summary, the inventor of the present invention thought about and designed a hydrological data analysis method and system in order to improve the deficiencies of the existing technology and thereby increase the practical application value in industry.

According to the foregoing, the purpose of the present invention is to provide a hydrological data analysis method and a hydrological data analysis system, which can solve the problem of incomplete hydrological data.

Based on the above objectives, the present invention provides a hydrological data analysis method, which includes executing by a processor: obtaining hydrological data of the water area from the sensing module; filtering the hydrological data; and determining whether the hydrological data is Normal data or abnormal data; when the hydrological data is judged to be abnormal data, at least one of the plurality of data repair models is selected as the selected data repair model according to the type of water area; and the abnormal data is input into the selected data repair model to repair the abnormal data.

In an embodiment of the present invention, the hydrological data includes a plurality of hydrological data points, and filtering the hydrological data includes: establishing a trust interval based on the plurality of hydrological data points; determining whether each of the plurality of hydrological data points is within the trust interval; and when the hydrological data points When the data points are not in the confidence interval, delete the hydrological data points.

In an embodiment of the present invention, determining whether the hydrological data is normal data or abnormal data includes: determining whether the hydrological data has at least one data missing area; when it is determined that the hydrological data has at least one data missing area, marking the hydrological data as abnormal data; and when it is determined that the hydrological data has at least one data missing area; It is judged that the hydrological data does not have at least one data loss area, and the hydrological data is marked as normal data.

In an embodiment of the present invention, inputting abnormal data into the selected data repair model to repair the abnormal data includes: repairing at least one data loss area by using the selected data repair model.

In an embodiment of the present invention, the method further includes executing by the processor: obtaining topographic data and disaster data; inputting topography data, disaster data, and repaired abnormal data or normal data into the artificial intelligence model to generate a predicted sediment volume; and marking the predicted sediment volume as one of a plurality of hazard risks.

Based on the above objectives, the present invention provides a hydrological data analysis system, including a sensing module and a processor. The sensing module generates hydrological data corresponding to the water area based on the meteorological conditions, water flow changes, and sediment changes in the water area. The processor is connected to the sensing module, filters the hydrological data and determines whether the hydrological data is normal data or abnormal data. When it is determined that the hydrological data is abnormal data, the processor selects at least one from a plurality of data repair models as the selected data repair model, and Input abnormal data to the selected data repair model to repair abnormal data.

In the embodiment of the present invention, the hydrological data includes a plurality of hydrological data points, and the filtering of the hydrological data performed by the processor includes: establishing a trust interval based on the plurality of hydrological data points; and determining whether each of the plurality of hydrological data points is in trust. interval; and when the hydrological data point is not in the confidence interval, delete the hydrological data point.

In an embodiment of the present invention, determining whether the hydrological data is normal data or abnormal data performed by the processor includes: determining whether the hydrological data has at least one data missing area; when it is determined that the hydrological data has at least one data missing area, marking the hydrological data as abnormal data; and when it is determined that the hydrological data does not have at least one data loss area, the hydrological data is marked as normal data.

In an embodiment of the present invention, inputting the abnormal data to the selected data repair model to repair the abnormal data performed by the processor includes: repairing at least one data loss area through the selected data repair model.

In an embodiment of the present invention, the processor further obtains geological data and disaster data, and inputs the geological data, disaster data and post-repair abnormal data or normal data into the artificial intelligence model. The artificial intelligence model generates a predicted amount of sediment, and the processor Indicates the predicted sediment volume as one of several hazard risks.

Following the above, the hydrological data analysis method and hydrological data analysis system of the present invention can filter the hydrological data and determine whether the hydrological data is normal data or abnormal data, and selectively repair abnormalities based on the judgment results of the hydrological data and the waters corresponding to the hydrological data. data to improve the accuracy and completeness of hydrological data.

10: Sensing module

20: Processor

CI1: confidence interval

DD1: Disaster information

DR1: data loss area

FP1: Repair rainfall data points

GD1: Geographic data

HD1: Hydrological data

S11~S17, S21~S32, S121~S123: steps

Figure 1 is a functional block diagram of a hydrological data analysis system according to an embodiment of the present invention.

Figure 2 shows functional blocks of a hydrological data analysis system according to another embodiment of the present invention. Figure.

Figure 3 is a flow chart of a hydrological data analysis method according to an embodiment of the present invention.

Figure 4 is a flow chart of filtering hydrological data in a hydrological data analysis method according to an embodiment of the present invention.

Figure 5 is a plot of rainfall versus time with confidence intervals according to an embodiment of the present invention.

Figure 6 is a plot of rainfall versus time before data restoration using a hydrological data analysis method according to an embodiment of the present invention.

Figure 7 is a graph of rainfall versus time before data restoration using a hydrological data analysis method according to an embodiment of the present invention.

Figures 8A and 8B are flow charts of a hydrological data analysis method according to another embodiment of the present invention.

The advantages, features and technical methods for achieving the present invention will be described in more detail with reference to the exemplary embodiments and accompanying drawings to be more easily understood. The present invention can be implemented in different forms and should not be understood to be limited to what is described herein. Stated Examples. On the contrary, the provided embodiments will make this disclosure more thorough and complete and fully convey the scope of the invention to those of ordinary skill in the art, and the invention will only be defined by the appended patent scope. .

It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, components, regions, layers and/or sections, these elements, components, regions, layers and/or sections These terms should not be limited. These terms are only used to distinguish one element, component, region, layer and/or section from another element, component, region, layer and/or section.

In addition, the terms "comprises" and/or "comprises" refer to stated features, regions, integers, steps The presence or addition of one or more other features, regions, integers, steps, operations, elements, components and/or combinations thereof is not excluded.

Please refer to Figure 1, which is a functional block diagram of a hydrological data analysis system according to an embodiment of the present invention. As shown in Figure 1 , the hydrological data analysis system of the present invention includes a sensing module 10 and a processor 20 . The sensing module 10 generates hydrological data HD1 corresponding to the water area based on the meteorological conditions, water flow changes, and sediment changes of the water area. The processor 20 is connected to the sensing module 10 and filters the hydrological data HD1 and determines whether the hydrological data HD1 is normal data or abnormal data, and selectively repairs the abnormal data based on the judgment results of the hydrological data HD1 and the waters corresponding to the hydrological data HD1, and filters the hydrological data. The details of the judgment of data HD1, hydrological data HD1 and the repair of abnormal data will be explained in detail in the paragraph of hydrological data analysis method.

In this embodiment, the sensing module 10 includes a rain sensor, a water flow sensor, a sediment volume sensor, and a water level sensor. The rainfall sensor generates multiple rainfall data points according to the meteorological conditions of the water area. The water flow sensor generates multiple water flow data points according to the water flow changes in the water area. The sediment volume sensor generates multiple sediment volume data points based on the changes in sediment. The water level The sensor generates multiple water level data points according to the water level changes in the water area; for example, a rainfall sensor, a water flow sensor, a sediment volume monitor, and a water level sensor can generate a single rainfall data point per unit hour. A single water flow data point, a single sediment volume data point and a single water level data. The hydrological data includes a plurality of hydrological data points, and the plurality of hydrological data points includes a plurality of rainfall data points, a plurality of water flow data points, a plurality of sediment volume data points, and a plurality of water level data points. The processor 20 operates according to the plurality of rainfall data points. , a plurality of water flow data points, a plurality of sediment volume data points and a plurality of water level data points to generate rainfall maps, water flow maps, sediment volume maps and water level maps. It should be noted that the water area can be a water catchment area, a reservoir area or a downstream river channel area, and the sensing modules 10 can be respectively provided in the water catchment area, reservoir area or downstream river channel area.

In this embodiment, the processor 20 may be a central processing unit (CPU), a microcontroller unit (MCU), or other processors with data processing functions. In one implementation, the processor 20 is disposed in a computing device, which may be a computer or a supercomputer; in another implementation, the processor 20 is disposed in a server, and the server is disposed in the cloud or edge. end.

Please refer to Figure 2, which illustrates functional blocks of a hydrological data analysis system according to another embodiment of the present invention. As shown in Figure 2, the hydrological data analysis system of the present invention includes a sensing module 10 and a processor 20. The configurations of the sensing module 10 and the processor 20 are similar to those shown in Figure 1. There are no differences here. Repeat the narrative again. However, the configuration in Figure 2 is still different from the configuration in Figure 1: the processor 20 can obtain the geography data GD1 and disaster data DD1 from the Water Resources Department, for example, and the processor 20 can also obtain the geography data GD1 and disaster data from other public sources. The network platform of data DD1 obtains geographical data GD1 and disaster data DD1, without being limited to the scope stated in the present invention. The processor 20 inputs the geological data GD1, the disaster data DD1, the abnormal data after repair or the normal data to the artificial intelligence model. The artificial intelligence model generates prediction results. The details of the prediction results of the artificial intelligence model will be detailed in the paragraph on the hydrological data analysis method. instruction.

Please refer to Figure 3, which is a flow chart of a hydrological data analysis method according to an embodiment of the present invention. As shown in Figure 3, the hydrological data analysis method of the present invention includes steps S11 to S17. The hydrological data analysis method shown in Figure 3 is suitable for the hydrological data analysis system shown in Figure 1 or 2, but is not limited to this. Steps S11 to S17 are explained below by taking the operation of the hydrological data analysis system shown in Figure 1 as an example.

Step S11: Obtain the hydrological data HD1 of the water area. Specifically, the processor 20 obtains the hydrological data HD1 from the sensing module 10. The hydrological data HD1 includes a plurality of rainfall data points, a plurality of water flow data points, a plurality of sediment volume data points and a plurality of water level data points. The processor 20 obtains the hydrological data HD1 from the sensing module 10. 20 According to the complex Several rainfall data points, a plurality of water flow data points, a plurality of sediment volume data points and a plurality of water level data points generate rainfall maps, water flow maps, sediment volume maps and water level maps.

The following will take the rain gauge as an example to illustrate steps S12 to S17. The data processing details of the water flow map, sediment volume map and water level map in steps S12 to S17 are the same as the data processing details of the rain gauge in steps S12 to S17. Repeat the narrative again.

Step S12: Filter the hydrological data HD1. Specifically, the processor 20 retains the normal rainfall data points in the rain gauge, and removes the abnormal rainfall data points in the rain gauge; wherein, the normal rainfall data points and the abnormal rainfall data points are two categories, which belong to the rainfall data of the normal rainfall data points. The number of points is plural, and the number of rainfall data points belonging to abnormal rainfall data points is at least one.

Furthermore, the processor 20 generates trust intervals based on multiple rainfall data points, distinguishes normal rainfall data points and abnormal rainfall data points based on the trust intervals, and then retains normal rainfall data points and removes abnormal rainfall data points. Please further refer to Figure 4. Figure 4 is a flow chart of filtering hydrological data in a hydrological data analysis method according to an embodiment of the present invention. As shown in Figure 4, the steps of filtering hydrological data include steps S121 to S123. Steps S121 to S123 are described below by taking the operation of the hydrological data analysis system shown in Figure 1 as an example.

Step S121: Establish a confidence interval CI1 based on a plurality of hydrological data points. Specifically, the processor 20 calculates the average value and the standard deviation based on the plurality of rainfall data points, and calculates the confidence interval CI1 based on the average value, the standard deviation, and the number of the plurality of rainfall data points.

Step S122: Determine whether each of the plurality of hydrological data points is within the confidence interval CI1. Specifically, as shown in Figure S, the processor 20 marks the trust interval CI1 on the rain gauge according to the upper limit and lower limit of the trust interval CI1, and determines whether each rainfall data point in the rain gauge is within the trust interval CI1. . Taking a single rainfall data point as an example, when the rainfall data point falls within the confidence interval CI1, the processor 20 The aforementioned rainfall data point is determined to be a normal rainfall data point; when the rainfall data point falls outside the confidence interval CI1, the processor 20 determines the aforementioned rainfall data point to be an abnormal rainfall data point; when the value of the rainfall data point is the upper limit of the confidence interval CI1 When the value or the lower limit value is reached, the processor 20 determines that the aforementioned rainfall data point is a normal rainfall data point. The mechanism for determining whether the remaining rainfall data points are within the confidence interval CI1 is the same as the mechanism for determining whether a single rainfall data point is within the confidence interval CI1, and will not be described again here.

Step S123: When the hydrological data point is not in the trust interval, delete the hydrological data point. Specifically, based on the trust interval CI1, the processor 20 deletes abnormal rainfall data points outside the trust interval CI1.

Step S13: Determine whether the hydrological data HD1 is normal data or abnormal data. Specifically, the processor 20 determines whether there is a data loss area DR1 as shown in FIG. 6 in the rainfall map. When the processor 20 determines that there is no data loss area DR1 in the rain gauge, the processor 20 continues to execute step S14; when the processor 20 determines that there is one or more data loss areas DR1 in the rain gauge, the processor 20 continues to execute step S15. . It should be noted that the data loss area DR1 is an area where the rainfall sensor does not generate rainfall data points in a specific time period. Depending on the water area, meteorological conditions, water flow changes, and sediment changes, the time period in the data loss area DR1 may occur. changes, so the time period generated by the data loss area DR1 is not limited.

Step S14: Mark the hydrological data as normal data. Specifically, the processor 20 marks a plurality of normal rainfall data points and rainfall patterns as normal data.

Step S15: Mark the hydrological data as abnormal data. Specifically, the processor 20 marks a plurality of normal rainfall data points and rainfall patterns as abnormal data.

Step S16: Select at least one from a plurality of data repair models as the selected data repair model according to the type of water area. Specifically, the processor 20 determines that the water area is a water catchment area, a reservoir area, or One of the downstream water channels generates a judgment result, and at least one data repair model is selected as the selected data repair model according to the judgment result.

For example, the number of data repair models is three and they are the first data repair model, the second data repair model and the third data repair model respectively. The following will give an example of the hydrological data repair model as the first data repair model, the second data repair model and the third data repair model. Data patching model or tertiary data patching model.

The hydrological data repair model can select one- and two-dimensional sand transport numerical models that are often used in Taiwan's water conservancy field, and use their simulation results to perform data repair. For the one- and two-dimensional sand transport numerical models in Taiwan's water conservancy field, please refer to the paper "Huang, CC, Lai, YG, Lai, JS, & Tan, YC (2019). Field and numerical modeling study of turbidity current in Shimen Reservoir during typhoon events . Journal of Hydraulic Engineering , 145 (5),05019003".

In one implementation, the processor 20 selects one of the first data repair model, the second data repair model, and the third data repair model as the selected data repair model according to the judgment result. Specifically, when the judgment result is a water catchment area, the processor 20 selects the first data repair model as the selected data repair model; when the judgment result is a reservoir area, the processor 20 selects the second data repair model as the selected data repair model; when the judgment result is a reservoir area, the processor 20 selects the second data repair model as the selected data repair model. The judgment result is the downstream river channel area, and the processor 20 selects the third data repair model as the selected data repair model.

In another implementation, the processor 20 selects the first data repair model, the second data repair model and the third data repair model as the selected data repair model according to the judgment result. In other words, when the judgment result is a water catchment area, a reservoir area or a downstream river channel area, the processor 20 selects the first data repair model, the second data repair model and the third data repair model as the selected data repair model.

Step S16: Input the abnormal data into the selected data repair model to repair the abnormal data. Specifically, the processor 20 inputs abnormal data to the selected data repair model, and the selected data repair model repairs As shown in Figure 6, there are multiple data loss areas DR1 in the abnormal data.

In step S16, as shown in Figure 7, the selected data repair model uses the uninterruptibility and consistency of the numerical model to add a plurality of repaired rainfall data points FP1 to each data loss area DR1 to repair a plurality of data loss areas. DR1. According to the status of each data loss area DR1, the number of restored rainfall data points FP1 changes, and there is no limit on the number of restored rainfall data points FP1 here.

Therefore, when the environment is extremely harsh and causes the sensing module 10 to be interrupted or unable to measure, resulting in a plurality of data loss areas DR1 in the hydrological data, the hydrological data analysis method of the present invention can be used to instantly repair the plurality of data loss areas in the hydrological data. DR1 makes hydrological data more complete and facilitates subsequent natural disaster analysis and assessment.

Please refer to Figure 8A and Figure 8B, which is a flow chart of a hydrological data analysis method according to another embodiment of the present invention. As shown in Figures 8A and 8B, the hydrological data analysis method of the present invention includes steps S21 to S32; wherein steps S21 to S27 are the same as steps S11 to S17 shown in Figure 3, and there are no differences here. Repeat the narrative again. The hydrological data analysis methods shown in Figure 8A and Figure 8B are suitable for the hydrological data analysis system shown in Figure 2, but are not limited to this. Steps S28 to S32 are described below by taking the operation of the hydrological data analysis system shown in Figure 2 as an example.

Step S28: Obtain geography data GD1 and disaster data DD1. Specifically, the processor 20 obtains topographic data GD1 and disaster data DD1 from the Water Resources Bureau; wherein, the topographic data GD1 includes collapse data in the watershed area, siltation data in the reservoir area, and terrain data, and the disaster data DD1 includes disaster events. quantity.

Step S29: Verify the geography data GD1. Specifically, the processor 20 first determines the correlation between collapse data and sedimentation data, and then determines whether the terrain changes in the physiographic data are reasonable.

First, regarding the correlation between collapse data and sedimentation data, the processor 20 calculates the collapse data and the correlation coefficient between the sedimentation data, and compare the correlation coefficient and the threshold to selectively use collapse data and sedimentation data. When the correlation coefficient is greater than the threshold, the processor 20 uses the collapse data and sedimentation data; when the correlation coefficient is less than the threshold, the processor 20 does not use the collapse data and sedimentation data, and re-obtains new collapse data and sedimentation data from the Water Resources Bureau.

In terms of the rationality of the terrain changes of the terrain data, the processor 20 determines whether the terrain changes of the terrain data are caused by natural disasters or caused by artificial development of the natural environment.

Step S30: Check disaster data DD1. Specifically, the processor 20 compares the disaster event amount and the critical range to generate a comparison result, and selectively adopts the disaster data DD1 according to the comparison result. When the comparison result indicates that the amount of disaster events is within the critical value range, the processor 20 uses the disaster data DD1; when the comparison result indicates that the disaster event amount is not within the critical value range, the processor 20 does not use the disaster data DD1 and obtains new data from the Water Resources Bureau. Disaster information DD1. It should be noted that disaster events are obvious disasters (such as landslides) caused in water catchment areas, reservoir areas or downstream waterways. Step S30 is to confirm whether the amount of disaster events defined above is sufficient to avoid excessive disaster events or Too little will affect the predictions of subsequent artificial intelligence models.

Step S31: Input topographic data GD1, disaster data DD1, and post-repair abnormal data or normal data into the artificial intelligence model to generate predicted sediment volume. Specifically, the processor 20 inputs the geological data GD1, the disaster data DD1, and the abnormal data or normal data after repair to the artificial intelligence model. The artificial intelligence model generates predictions based on the geological data GD1, the disaster data DD1, and the abnormal data or normal data after repair. Amount of mud and sand. It should be noted that the artificial intelligence model is a pre-trained model and can estimate the amount of sediment discharged from the reservoir based on the input geological data GD1, disaster data DD1 and hydrological data HD1. The aforementioned artificial intelligence model estimates the amount of sediment discharged by the reservoir. The amount of sediment discharged is the predicted amount of sediment.

Step S32: Mark the predicted amount of sediment as one of a plurality of disaster risks. Specifically, The artificial intelligence model can select one of several disaster risks based on the estimated sediment volume and type of water area, and mark the selected disaster risk on the predicted sediment volume. Among them, the water catchment area, reservoir area and downstream waterway respectively correspond to a plurality of disaster risks, and the plurality of disaster risks include low risk, medium risk and high risk.

Following the above, the hydrological data analysis method and hydrological data analysis system of the present invention filter the hydrological data and determine whether the hydrological data is normal data or abnormal data, and selectively repair abnormalities based on the judgment results of the hydrological data and the waters corresponding to the hydrological data. data to improve the accuracy and completeness of hydrological data.

The above is only illustrative and not restrictive. Any equivalent modifications or changes that do not depart from the spirit and scope of the present invention shall be included in the appended patent scope.

S11~S17: Steps

Claims (8)

A hydrological data analysis method includes executing by a processor: obtaining hydrological data of a water area from a sensing module; filtering the hydrological data; determining whether the hydrological data is normal data or abnormal data; when determining the hydrological data When the data is the abnormal data, select at least one from a plurality of data repair models as a selected data repair model according to the type of the water area; and input the abnormal data into the selected data repair model to repair the abnormal data; wherein each of the A plurality of data repair models include a one-dimensional sand transport numerical model or a two-dimensional sand transport numerical model to use the simulation results to perform data repair; wherein the processor further executes: obtaining a landform data and a disaster data; inputting a landform The text data, a disaster data and the repaired abnormal data or the normal data are added to a trained artificial intelligence model to generate a predicted sediment amount; and mark the predicted sediment amount as one of a plurality of disaster risks, of which the plurality Disaster risk includes low risk, medium risk and high risk. The hydrological data analysis method as described in claim 1, wherein the hydrological data includes a plurality of hydrological data points, and filtering the hydrological data includes: establishing a confidence interval based on the plurality of hydrological data points; Determine whether each of the plurality of hydrological data points is within the trust interval; and when the hydrological data point is not in the trust interval, delete the hydrological data point. According to the hydrological data analysis method described in claim 1, determining whether the hydrological data is the normal data or the abnormal data includes: determining whether the hydrological data has at least one data loss area; when determining whether the hydrology data has the at least one data loss area; , marking the hydrological data as abnormal data; and when it is determined that the hydrological data does not have the at least one data loss area, marking the hydrological data as normal data. According to the hydrological data analysis method described in claim 3, inputting the abnormal data into the selected data repair model to repair the abnormal data includes: repairing the at least one data loss area through the selected data repair model. A hydrological data analysis system includes: a sensing module that generates hydrological data corresponding to the water area based on a meteorological condition, a water flow change, and a sediment change in the water area; a processor connected to the sensing module, And filter the hydrological data and determine whether the hydrological data is normal data or abnormal data. When determining that the hydrological data is the abnormal data, the processor selects at least one from a plurality of data repair models as a selected data repair model, and Input the abnormal data into the selected data repair model to repair the abnormal data; each of the plurality of data repair models includes a one-dimensional sand transport numerical model or a two-dimensional sand transport numerical model to use its simulation results for data repair; The processor further obtains a local language data and a disaster data, and inputs the The geological data, the disaster data and the restored abnormal data or the normal data are combined into a trained artificial intelligence model. The artificial intelligence model generates a predicted sediment amount and marks the predicted sediment amount as one of a plurality of disaster risks, where The plurality of disaster risks include low risk, medium risk and high risk. The hydrological data analysis system as described in claim 5, wherein the hydrological data includes a plurality of hydrological data points, and filtering the hydrological data performed by the processor includes: establishing a trust interval based on the plurality of hydrological data points; determining the Whether each of the plurality of hydrological data points is within the confidence interval; and when the hydrological data point is not in the confidence interval, delete the hydrological data point. The hydrological data analysis system as described in claim 5, wherein the determination performed by the processor whether the hydrological data is the normal data or the abnormal data includes: determining whether the hydrological data has at least one data loss area; when determining whether the hydrological data has at least one data loss area; If the at least one data loss area is present, the hydrological data is marked as abnormal data; and when it is determined that the hydrological data does not have the at least one data loss area, the hydrological data is marked as normal data. The hydrological data analysis system as described in claim 5, wherein inputting the abnormal data to the selected data repair model to repair the abnormal data performed by the processor includes: repairing the at least one data loss area through the selected data repair model. .