A hydrological data real-time monitoring and early warning method and system based on data analysis

Abstract

The application discloses a hydrological data real-time monitoring and early warning method and system based on data analysis, relates to the technical field of hydrological data processing, and starts from the time response relationship among water level, rainfall and flow rate based on a multi-source hydrological synchronous time series data set Syn, gradually constructs a water level change response feature set Lrf, a rainfall change response feature set Rrf and a flow rate change response feature set Vrf, and further forms a hydrological data time consistency offset Tco capable of reflecting the time collaborative change state of the multi-source hydrological data. By comparing and analyzing the hydrological data time consistency offset Tco with a normal hydrological response consistency reference interval Ref, effective early warning can be given before a flood peak is formed when the water level has not reached an alarm value, but obvious time response mismatch has occurred among the water level change, the rainfall change and the flow rate change, and the reliability of monitoring and early warning in actual basin management is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrological data processing, in particular to a hydrological data real-time monitoring and early warning method and system based on data analysis. BACKGROUND

[0002] With the continuous improvement of the complexity of the hydrological environment of the basin, relying on data analysis means to monitor and warn hydrological data in real time has become an important technical direction in water resources management, flood control and disaster reduction, and safe operation of the basin. In the existing hydrological monitoring system, through continuous collection and analysis of multi-source hydrological data such as water level, rainfall, flow rate, etc., the dynamic change process of the water level is realized.

[0003] In the existing hydrological early warning method based on data analysis, the early warning judgment usually takes whether the water level exceeds the warning value or whether the rainfall reaches the set threshold as the main basis. Although this method has simple implementation logic, it often ignores the time dimension of the coordinated response relationship between multi-source hydrological data such as water level, rainfall, and flow rate. When the rainfall process just occurs and the basin has not yet appeared obvious water level rise, the time response consistency between multi-source hydrological data has changed, but since each single indicator has not reached the early warning threshold, the existing method is difficult to identify the potential risk in time. In this case, if there is no effective analysis of the time response synchronization of multi-source hydrological data, it is easy to miss the opportunity to identify the abnormal hydrological evolution process in advance. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a hydrological data real-time monitoring and early warning method and system based on data analysis, which solves the problems mentioned in the background art.

[0005] To achieve the above purpose, the present application is realized by the following technical scheme: a hydrological data real-time monitoring and early warning method based on data analysis, comprising the following steps:

[0006] S1, collecting water level data, rainfall data and flow rate data in the monitoring basin to generate a multi-source hydrological synchronous time series data set Syn;

[0007] S2, based on the multi-source hydrological synchronous time series data set Syn, respectively extracting water level change response features, rainfall change response features and flow rate change response features to form a water level change response feature set Lrf, a rainfall change response feature set Rrf and a flow rate change response feature set Vrf;

[0008] S3, based on the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf, constructing a hydrological data time consistency offset Tco reflecting the time response consistency of multi-source hydrological data;

[0009] S4, compare the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref, generate an abnormal trend determination result Alm, and issue through the instruction signaling device.

[0010] Preferably, the S1 comprises S11;

[0011] S11, based on the existing hydrological data acquisition system, acquiring water level data, rainfall data and flow rate data in the monitoring basin;

[0012] Among them, for water level data, by periodically sampling the height change of water body at the monitoring section, the water level observation value at the corresponding time point is obtained;

[0013] For rainfall data, by continuously recording the precipitation per unit time in the monitoring area, the rainfall observation value at the corresponding time point is obtained;

[0014] For flow rate data, by periodically sampling the flow state of water body at the monitoring section, the flow rate observation value at the corresponding time point is obtained;

[0015] Arrange the water level observation value in time sequence to form a water level time series data set Lev;

[0016] Arrange the rainfall observation value in time sequence to form a rainfall time series data set Rai;

[0017] Arrange the flow rate observation value in time sequence to form a flow rate time series data set Vel.

[0018] Preferably, the S1 further comprises S12;

[0019] S12, based on a unified time reference, correct the collection time of each data in the water level time series data set Lev, the rainfall time series data set Rai and the flow rate time series data set Vel, so that the data of different sources are correlated under the same time scale;

[0020] Among them, when there is data missing or sampling inconsistency at any time point in the water level time series data set Lev, the rainfall time series data set Rai and the flow rate time series data set Vel, the missing data is adjusted by time mapping method;

[0021] Integrate the processed water level time series data set Lev, rainfall time series data set Rai and flow rate time series data set Vel to generate a multi-source hydrological synchronous time series data set Syn.

[0022] Preferably, the S2 comprises S21;

[0023] S21, based on the multi-source hydrological synchronous time series data set Syn, respectively, the water level data, rainfall data and flow rate data of the same time node are calculated;

[0024] For water level data, the water level change between adjacent time nodes is calculated according to the water level value difference between adjacent time nodes, and the water level change sequence is formed;

[0025] For rainfall data, the rainfall change between adjacent time nodes is calculated according to the rainfall value difference between adjacent time nodes, and the rainfall change sequence is formed;

[0026] For flow rate data, the flow rate change between adjacent time nodes is calculated according to the flow rate value difference between adjacent time nodes, and the flow rate change sequence is formed.

[0027] Preferably, the S2 further comprises S22;

[0028] S22, based on the water level change sequence, selecting an analysis time period composed of a plurality of continuous time nodes, and performing direction determination, amplitude statistics and continuity determination on the water level change sequence in the analysis time period;

[0029] Based on the change sign distribution of the water level change sequence in the analysis time period, the direction determination is performed to determine the water level change direction;

[0030] Wherein, the change sign is determined by the change amount in the water level change sequence, when the change amount is increased, the change sign bit is determined as positive, when the change amount decreases, the change sign is determined as negative;

[0031] At the same time, the water level change amplitude is obtained based on the difference between the maximum value and the minimum value of the water level change sequence in the analysis time period;

[0032] Based on the length of the continuous change sign segment of the water level change sequence in the analysis time period, the continuity determination is performed to determine the water level change continuity;

[0033] The water level change direction, water level change amplitude and water level change continuity are combined to form a water level change response feature set Lrf;

[0034] Based on the rainfall change sequence, an analysis time period composed of a plurality of continuous time nodes is selected, and intensity statistics, persistence determination and concentration degree determination are performed on the rainfall change sequence in the analysis time period;

[0035] Based on the peak value and the average value of the rainfall change sequence in the analysis time period, the rainfall change intensity is obtained;

[0036] Based on the length of the continuous positive value segment of the rainfall variation amount sequence in the analysis time period, persistence determination is performed to obtain rainfall variation persistence;

[0037] Based on the time node distribution of the positive value in the rainfall variation amount sequence in the analysis time period, concentration degree determination is performed to obtain rainfall variation persistence;

[0038] The rainfall variation intensity, rainfall variation persistence, and rainfall variation concentration degree are combined to form a rainfall variation response feature set Rrf;

[0039] Based on the flow rate variation amount sequence, an analysis time period composed of a plurality of continuous time nodes is selected, and direction determination, rate statistics, and stationarity determination are performed on the flow rate variation amount sequence in the analysis time period;

[0040] Based on the sign distribution of the flow rate variation amount sequence in the analysis time period, direction determination is performed to obtain flow rate variation direction;

[0041] Based on the mean and peak value of the flow rate variation amount sequence in the analysis time period, flow rate variation rate description information is obtained to obtain flow rate variation rate;

[0042] Based on the dispersion degree of the flow rate variation amount sequence in the analysis time period, stationarity determination is performed to obtain flow rate variation stationarity;

[0043] The flow rate variation direction, flow rate variation rate, and flow rate variation stationarity are combined to form a flow rate variation response feature set Vrf.

[0044] Preferably, the S3 comprises S31;

[0045] S31, by reading the feature information corresponding to the time period in the water level variation response feature set Lrf, the rainfall variation response feature set Rrf, and the flow rate variation response feature set Vrf;

[0046] The feature information corresponding to the water level variation response feature set Lrf is subjected to significant change determination to obtain the water level variation significant change time position;

[0047] The feature information corresponding to the rainfall variation response feature set Rrf is subjected to significant change determination to obtain the rainfall variation significant change time position;

[0048] The feature information corresponding to the flow rate variation response feature set Vrf is subjected to significant change determination to obtain the flow rate variation significant change time position;

[0049] The significant change determination is performed by comparing the change amplitudes of the feature information corresponding to adjacent time nodes, and when the change amplitudes of the feature information between adjacent time nodes reach a preset significant change proportion condition, the corresponding time nodes are determined as significant change time positions.

[0050] The water level change significant change time position, the rainfall change significant change time position, and the flow rate change significant change time position are aligned under a time reference to form a time dimension offset relationship of multi-source hydrological change responses at each time point.

[0051] The feature information change proportion is used to represent a proportion relationship of a change amount of feature information between adjacent time nodes relative to a historical change range or a current analysis time period change range.

[0052] Preferably, the S3 further includes S32.

[0053] The S32 calculates a time difference based on the obtained water level change significant change time position, the rainfall change significant change time position, and the flow rate change significant change time position, including:

[0054] A time difference between the water level change significant change time position and the rainfall change significant change time position is calculated to form a water level rainfall time difference result Lrt.

[0055] A time difference between the water level change significant change time position and the flow rate change significant change time position is calculated to form a water level flow rate time difference result Lvt.

[0056] A time difference between the rainfall change significant change time position and the flow rate change significant change time position is calculated to form a rainfall flow rate time difference result Rvt.

[0057] The water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt, and the rainfall flow rate time difference result Rvt are proportionally processed according to a time length of an analysis time period to obtain normalized time difference results: the water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt, and the rainfall flow rate time difference result Rvt, and are summarized to obtain a hydrological data time consistency offset Tco.

[0058] Preferably, the S4 includes S41.

[0059] The S41 performs interval comparison analysis on each normalized time difference result in the obtained hydrological data time consistency offset Tco and a pre-established normal hydrological response consistency reference interval Ref.

[0060] The interval comparison analysis is performed by judging whether the water level-rainfall time difference result Lrt, the water level-flow rate time difference result Lvt and the rainfall-flow rate time difference result Rvt in the hydrological data time consistency offset Tco fall within the normal hydrological response consistency reference interval Ref.

[0061] When at least one normalized time difference result in the hydrological data time consistency offset Tco exceeds the normal hydrological response consistency reference interval Ref, it is determined that the consistency of the multi-source hydrological data in the time response dimension is abnormal.

[0062] When no normalized time difference result in the hydrological data time consistency offset Tco exceeds the normal hydrological response consistency reference interval Ref, it is determined that the consistency of the multi-source hydrological data in the time response dimension is not abnormal.

[0063] At the same time, according to the number of normalized time difference results exceeding the normal hydrological response consistency reference interval Ref, an abnormal trend determination result Alm is generated.

[0064] Preferably, the S4 further comprises S42.

[0065] S42, based on the abnormal trend determination result Alm, confirming the warning level of the current hydrological change process, specifically generating a first warning, a second warning and a third warning according to the number in the abnormal trend determination result Alm, and forming hydrological abnormal warning information War corresponding to the warning level;

[0066] Synchronously converting the obtained hydrological abnormal warning information War into warning instruction information, and issuing it through an instruction signaling device to send an abnormal warning instruction to the corresponding hydrological monitoring or management system.

[0067] A hydrological data real-time monitoring and warning system based on data analysis, comprising a hydrological data acquisition module, a time sequence feature extraction module, a consistency offset analysis module and a monitoring and warning module.

[0068] The hydrological data acquisition module acquires water level data, rainfall data and flow rate data in the monitored river basin, and generates a multi-source hydrological synchronous time sequence data set Syn.

[0069] The time sequence feature extraction module extracts water level change response features, rainfall change response features and flow rate change response features based on the multi-source hydrological synchronous time sequence data set Syn, respectively, to form a water level change response feature set Lrf, a rainfall change response feature set Rrf and a flow rate change response feature set Vrf.

[0070] The consistency offset analysis module constructs a hydrological data time consistency offset Tco reflecting the time response consistency of the multi-source hydrological data based on the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf;

[0071] The monitoring and early warning module compares and analyzes the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref, generates an abnormal trend determination result Alm, and issues through the instruction signaling device.

[0072] The present application provides a kind of based on data analysis's hydrological data real-time monitoring and early warning method and system, with following beneficial effects:

[0073] (1) based on multi-source hydrological synchronous time series data set Syn, starting from the time response relationship of water level, rainfall and flow rate, gradually constructing water level change response feature set Lrf, rainfall change response feature set Rrf and flow rate change response feature set Vrf, and further forming hydrological data time consistency offset Tco capable of reflecting the time collaborative change state of multi-source hydrological data. By comparing and analyzing the hydrological data time consistency offset Tco with the normal hydrological response consistency reference interval Ref, effective early warning can be given before the formation of flood peak when the water level has not reached the warning value, but the water level change, rainfall change and flow rate change have appeared obvious time response mismatch.

[0074] (2) by calculating and segmenting the multi-source hydrological synchronous time series data set Syn, the originally numerical and instantaneous hydrological observation data are converted into water level change response feature set Lrf, rainfall change response feature set Rrf and flow rate change response feature set Vrf capable of reflecting the hydrological evolution law, thereby effectively solving the problem of only looking at the current value and not looking at the change process in the prior art. By constructing water level change amount sequence, rainfall change amount sequence and flow rate change amount sequence, and introducing direction determination, amplitude or intensity statistics, continuity or persistence determination and stability or concentration degree determination in the analysis period, each type of hydrological element not only has the judgment result of "whether change", but also has the multidimensional description ability of "whether the change direction is consistent, whether the change is continuous, whether the change is concentrated or intense".

[0075] (3) By significant change determination of feature information in the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow velocity change response feature set Vrf under the same time reference, respectively extracting the water level change significant change time position, the rainfall change significant change time position and the flow velocity change significant change time position, so that the "key change moment" of water level, rainfall and flow velocity can be clearly positioned; On this basis, further construct the water level rainfall time difference result Lrt, the water level flow velocity time difference result Lvt and the rainfall flow velocity time difference result Rvt, and eliminate the influence of different analysis time length on the time difference result through proportionalization processing, and finally summarize the hydrological data time consistency offset Tco, which can truly reflect the collaborative destruction process of multi-source hydrological elements in the flood gestation process, and is an important supplement and substantial improvement to the existing hydrological warning technology. BRIEF DESCRIPTION OF DRAWINGS

[0076] Figure 1 It is a step schematic diagram of the hydrological data real-time monitoring and early warning method based on data analysis of the present application;

[0077] Figure 2 It is a block diagram schematic diagram of the hydrological data real-time monitoring and early warning system based on data analysis of the present application;

[0078] Figure 3 It is a time change schematic diagram of the hydrological data time difference result. DETAILED DESCRIPTION

[0079] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0080] Embodiment 1: The present application provides a hydrological data real-time monitoring and early warning method based on data analysis, please refer to Figure 1 , including the following steps:

[0081] S1, collecting water level data, rainfall data and flow velocity data in the monitoring basin, and generating a multi-source hydrological synchronous time series data set Syn;

[0082] S2, based on the multi-source hydrological synchronous time series data set Syn, respectively extracting water level change response features, rainfall change response features and flow velocity change response features, forming a water level change response feature set Lrf, a rainfall change response feature set Rrf and a flow velocity change response feature set Vrf;

[0083] S3, constructing a hydrological data time consistency offset Tco reflecting the time response consistency of the multi-source hydrological data based on the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf;

[0084] S4, comparing and analyzing the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref to generate an abnormal trend determination result Alm, and issuing the result through an instruction signaling device.

[0085] In this embodiment, based on the multi-source hydrological synchronous time series data set Syn, the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf are constructed step by step from the time response relationship of water level, rainfall and flow rate, and further the hydrological data time consistency offset Tco reflecting the time collaborative change state of multi-source hydrological data is formed. By comparing and analyzing the hydrological data time consistency offset Tco with the normal hydrological response consistency reference interval Ref, the abnormal trend determination result Alm can be generated in advance and the warning instruction can be issued when the water level has not reached the warning value, but the time response mismatch has occurred between the water level change, the rainfall change and the flow rate change. For example, in the case of short-term heavy rainfall in the upstream, the flow rate starts to rise in advance while the water level is still in the normal interval, the traditional threshold-based warning method often cannot respond in time, but this method can identify the time offset of the rainfall change response feature set Rrf and the flow rate change response feature set Vrf relative to the water level change response feature set Lrf through the hydrological data time consistency offset Tco, thereby giving effective warning before the flood peak is formed, significantly making up for the deficiency of "no pre-warning when not exceeding the threshold and pre-warning lag" in the prior art, and improving the foresight and reliability of hydrological monitoring and warning in actual river basin management and flood control scheduling.

[0086] Embodiment 2: Specifically, the S1 includes S11;

[0087] S11, based on the existing hydrological data acquisition system, acquiring water level data, rainfall data and flow rate data in the monitored river basin respectively;

[0088] Among them, for the water level data, the water level observation value at the corresponding time point is obtained by periodically sampling the water height change at the monitoring section;

[0089] For rainfall data, the rainfall observation value at the corresponding time point is obtained by continuously recording the precipitation per unit time in the monitoring area;

[0090] For flow rate data, the flow rate observation value at the corresponding time point is obtained by periodically sampling the water flow state at the monitoring section;

[0091] arranging the water level observation values in sequence of acquisition time to form a water level time series data set Lev;

[0092] arranging the rainfall observation values in sequence of acquisition time to form a rainfall time series data set Rai;

[0093] arranging the flow rate observation values in sequence of acquisition time to form a flow rate time series data set Vel;

[0094] The water level time series data set Lev, the rainfall time series data set Rai, and the flow rate time series data set Vel all contain corresponding relationships between corresponding data values and collection times.

[0095] The S1 further comprises S12;

[0096] S12, based on a unified time reference, corrects the collection times corresponding to each data in the water level time series data set Lev, the rainfall time series data set Rai, and the flow rate time series data set Vel, so that data from different sources are correlated at the same time scale;

[0097] When there is missing data or inconsistent sampling at any time point in the water level time series data set Lev, the rainfall time series data set Rai, and the flow rate time series data set Vel, the missing data is adjusted through time mapping;

[0098] The processed water level time series data set Lev, the rainfall time series data set Rai, and the flow rate time series data set Vel are integrated to generate a multi-source hydrological synchronous time series data set Syn for subsequent joint analysis of multi-source hydrological data;

[0099] The unified time reference is determined by the following method: the latest collection time in the water level time series data set Lev, the rainfall time series data set Rai, and the flow rate time series data set Vel is determined respectively; the maximum value of the three latest collection times is determined as the alignment termination reference time, and according to the time granularity of the unified time reference, the continuous time nodes are generated starting from the starting point of the unified time reference and increasing until covering the alignment termination reference time, thereby forming the time node sequence of the unified time reference.

[0100] In this embodiment, the unified time processing is obtained by normalizing the water level time series data set Lev, the rainfall time series data set Rai and the flow rate time series data set Vel, which solves the basic problem of "different collection methods, inconsistent time scales and data that cannot be directly aligned" in existing hydrological monitoring. Specifically, the water level observation value, the rainfall observation value and the flow rate observation value are respectively time-sequenced according to their actual collection methods, and a unified time reference is introduced on this basis to correct and map the data of different sampling frequencies and different sampling times, so that the generated multi-source hydrological synchronous time series data set Syn has a one-to-one correspondence in the time dimension. For example, in actual basin monitoring, rainfall data is usually recorded continuously at a minute level, while water level and flow rate data are mostly periodically sampled, and traditional methods often need to artificially interpolate or simply discard part of the data, which is easy to introduce errors; and the method generates a time node sequence covering the latest collection time through a unified time reference, so that even if there is data missing or sampling inconsistency at individual time points, data alignment can be completed through time mapping, thereby ensuring the integrity and continuity of the multi-source hydrological synchronous time series data set Syn. This effect provides a reliable data basis for subsequent construction of the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf, and significantly reduces the risk of misjudgment caused by time inconsistency, which is an actual improvement effect that traditional methods relying on a single data source or simple time interpolation methods cannot achieve.

[0101] In a specific embodiment, the S2 comprises S21.

[0102] S21, based on the multi-source hydrological synchronous time series data set Syn, respectively calculates the change amount of water level data, rainfall data and flow rate data at the same time node;

[0103] For water level data, the change amount of water level between adjacent time nodes is calculated according to the difference between the water level values of adjacent time nodes, and a water level change amount sequence is formed.

[0104] For rainfall data, the change amount of rainfall between adjacent time nodes is calculated according to the difference between the rainfall values of adjacent time nodes, and a rainfall change amount sequence is formed.

[0105] For flow rate data, the change amount of flow rate between adjacent time nodes is calculated according to the difference between the flow rate values of adjacent time nodes, and a flow rate change amount sequence is formed.

[0106] The S2 further comprises S22.

[0107] S22, based on the water level change amount sequence, selects an analysis time period composed of a plurality of consecutive time nodes, and performs direction determination, amplitude statistics and continuity determination on the water level change amount sequence in the analysis time period.

[0108] The direction determination is performed based on the change sign distribution of the water level change amount sequence in the analysis time period, and the water level change direction is determined, specifically: when the proportion of positive values in the water level change amount sequence is higher than the proportion of negative values, it is determined that the water level change direction is upward; when the proportion of negative values in the water level change amount sequence is higher than the proportion of positive values, it is determined that the water level change direction is downward; when the water level change direction is not determined as upward and downward, it is determined that the water level change direction is oscillation direction;

[0109] The change sign is determined by the change amount in the water level change amount sequence, and when the change amount increases, the change sign bit is determined as positive value, and when the change amount decreases, the change sign is determined as negative value;

[0110] At the same time, the water level change amplitude is obtained based on the difference between the maximum value and the minimum value of the water level change amount sequence in the analysis time period, and the mean value and the dispersion degree of the water level change amount sequence in the analysis time period are counted and analyzed, and then combined with the water level change amplitude to form the amplitude description information representing the water level change intensity;

[0111] The continuity determination is performed based on the length of the continuous change sign segment of the water level change amount sequence in the analysis time period, and the water level change continuity is determined, specifically: the length of the continuous change sign segment represents the longest continuous segment length of the water level change amount sequence which is continuously positive or negative, when the longest continuous segment length reaches a preset continuous segment threshold, it is determined that the water level change continuity is high continuity; when the longest continuous segment length does not reach the preset continuous segment threshold and the positive value and the negative value appear alternately, it is determined that the water level change continuity is low continuity;

[0112] The water level change direction, the water level change amplitude and the water level change continuity are combined to form a water level change response feature set Lrf;

[0113] Based on the rainfall change amount sequence, an analysis time period composed of a plurality of continuous time nodes is selected, and the intensity statistics, the continuity determination and the concentration degree determination of the rainfall change amount sequence are performed in the analysis time period;

[0114] The rainfall change intensity is obtained based on the peak value and the mean value of the rainfall change amount sequence in the analysis time period, and the proportion of positive values of the rainfall change amount sequence in the analysis time period is counted to obtain the intensity description information representing the rainfall change sudden increase characteristic;

[0115] determining persistence based on a length of a continuous positive value segment of the rainfall variation amount sequence in the analysis time period, obtaining rainfall variation persistence, the length of the continuous positive value segment refers to a longest continuous segment length of the rainfall variation amount sequence being continuously positive, when the longest continuous segment length reaches a preset continuous segment threshold, determining that the rainfall variation persistence is high persistence, when the longest continuous segment length does not reach the preset continuous segment threshold, determining that the rainfall variation persistence is low persistence;

[0116] determining concentration degree based on a time node distribution of the positive value in the rainfall variation amount sequence in the analysis time period, obtaining rainfall variation persistence, when the positive value is mainly concentrated in a few continuous time nodes of the analysis time period, determining that the rainfall variation concentration degree is high concentration degree, when the positive value is scattered and there is no obvious concentration interval, determining that the rainfall variation concentration degree is low concentration degree;

[0117] combining the rainfall variation intensity, the rainfall variation persistence and the rainfall variation concentration degree to form a rainfall variation response feature set Rrf;

[0118] based on the flow rate variation amount sequence, selecting an analysis time period composed of a plurality of continuous time nodes, and performing direction determination, rate statistics and stationarity determination on the flow rate variation amount sequence in the analysis time period;

[0119] determining direction based on a sign distribution of the flow rate variation amount sequence in the analysis time period, obtaining flow rate variation direction, when the proportion of positive values in the flow rate variation amount sequence is higher than the proportion of negative values, determining that the flow rate variation direction is an increasing direction, when the proportion of negative values in the flow rate variation amount sequence is higher than the proportion of positive values, determining that the flow rate variation direction is a decreasing direction, when the proportion of positive values and the proportion of negative values are not dominant, determining that the flow rate variation direction is a shock direction;

[0120] obtaining flow rate variation rate description information based on a mean value and a peak value of the flow rate variation amount sequence in the analysis time period, obtaining flow rate variation rate, the peak value is used to represent the maximum surge degree of the flow rate variation, and the mean value is used to represent the overall change level of the flow rate variation;

[0121] determining stationarity based on a dispersion degree of the flow rate variation amount sequence in the analysis time period, obtaining flow rate variation stationarity, when the dispersion degree is lower than a preset fluctuation threshold, determining that the flow rate variation stationarity is high stationarity, when the dispersion degree is higher than the preset fluctuation threshold, determining that the flow rate variation stationarity is low stationarity;

[0122] combining the flow rate variation direction, the flow rate variation rate and the flow rate variation stationarity to form a flow rate variation response feature set Vrf;

[0123] The continuous segment threshold value, the continuous segment threshold value and the fluctuation threshold value are all reference determination conditions determined according to historical hydrological change data before determining the response characteristics of hydrological change, and the determination process includes the following steps:

[0124] Based on the multi-source hydrological synchronous time series data set Syn formed in the historical period, the water level change amount sequence, the rainfall change amount sequence and the flow velocity change amount sequence corresponding to the historical period are respectively obtained;

[0125] In the historical period, the continuous symbol segment length distribution of the water level change amount sequence which is continuously positive or negative is respectively counted, the continuous positive symbol segment length distribution of the rainfall change amount sequence is counted, and the change dispersion degree distribution of the flow velocity change amount sequence in each historical analysis time period is counted;

[0126] On the basis of statistics, the typical value range of the continuous symbol segment length distribution of the water level change amount sequence is determined as the continuous segment threshold value, which is used to determine the continuity of the water level change trend in the time dimension;

[0127] The typical value range of the continuous positive symbol segment length distribution of the rainfall change amount sequence is determined as the continuous segment threshold value, which is used to determine the continuity of the rainfall change process in the time dimension;

[0128] The typical value range of the change dispersion degree distribution of the flow velocity change amount sequence in the historical analysis time period is determined as the fluctuation threshold value, which is used to determine the stability of the flow velocity change process.

[0129] In this embodiment, by calculating the change amount of the multi-source hydrological synchronous time series data set Syn and segmenting analysis, the originally "numerical and instantaneous" hydrological observation data is converted into the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow velocity change response feature set Vrf which can reflect the hydrological evolution law, thereby effectively solving the problem of "only looking at the current value, not looking at the change process" in the prior art. Specifically, by constructing the water level change amount sequence, the rainfall change amount sequence and the flow velocity change amount sequence, and introducing direction judgment, amplitude or intensity statistics, continuity or persistence judgment and stationarity or concentration degree judgment in the analysis period, each type of hydrological element not only has the judgment result of "whether the change", but also has the multi-dimensional description ability of "whether the change direction is consistent, whether the change is continuous, and whether the change is concentrated or intense". For example, in actual basin operation, there may be a situation that the rainfall change amount sequence presents short-time concentrated enhancement, but the water level change amount sequence has not been significantly lifted. The traditional method often has difficulty in distinguishing whether this is an accidental rainfall or a process with the potential to form a flood peak; and the method can identify the structural characteristics of the rainfall process in advance through the rainfall change intensity, the rainfall change persistence and the rainfall change concentration degree simultaneously embodied in the rainfall change response feature set Rrf, and form a comparable response basis with the water level change response feature set Lrf and the flow velocity change response feature set Vrf. In addition, by the continuous segment threshold, the continuous segment threshold and the fluctuation threshold determined based on the historical data, the judgment of the response feature is not dependent on fixed experience parameters, but is matched with the historical hydrological behavior of the specific basin, thereby significantly improving the recognition accuracy of "abnormal change form" when constructing the hydrological data time consistency offset Tco, which is an actual technical effect that cannot be achieved by simply analyzing the instantaneous water level or a single index.

[0130] Embodiment 4: please refer to Figure 1 and Figure 3 Specifically, the S3 includes S31;

[0131] It should be noted that:

[0132] Figure 3 for showing the change of the time response difference between the multi-source hydrological data under the unified time reference;

[0133] As shown in Figure 3 , the horizontal coordinate in the figure represents the discrete time points in the analysis period, and the discrete time points are specific time positions generated in turn according to the preset time granularity under the unified time reference, and are used to identify the occurrence time corresponding to each hydrological change response difference result;

[0134] Among them, Figure 3The abscissa in the figure explicitly indicates a specific time point, rather than the length of a time interval; the values corresponding to adjacent scales in the abscissa indicate different time point positions, such as the 5th time node, the 6th time node, etc., rather than a time interval corresponding to a range of 5 to 10; the time interval between adjacent time points is determined by the time granularity of the unified time reference, which can be set to a minute level, an hour level or other fixed time scales according to the hydrological data sampling frequency;

[0135] S31, reading the feature information corresponding to the time period in the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow velocity change response feature set Vrf;

[0136] Performing significant change determination on the feature information corresponding to the water level change response feature set Lrf to obtain the water level change significant change time position;

[0137] Performing significant change determination on the feature information corresponding to the rainfall change response feature set Rrf to obtain the rainfall change significant change time position;

[0138] Performing significant change determination on the feature information corresponding to the flow velocity change response feature set Vrf to obtain the flow velocity change significant change time position;

[0139] The significant change determination is performed by comparing the change amplitudes of the feature information corresponding to adjacent time nodes, and when the change amplitudes of the feature information between adjacent time nodes reach a preset significant change proportion condition, the corresponding time node is determined as a significant change time position;

[0140] Aligning the water level change significant change time position, the rainfall change significant change time position and the flow velocity change significant change time position under the time reference to form the time dimension offset relationship of the multi-source hydrological change response at each time point;

[0141] The feature information change proportion is used to represent the proportion relationship of the feature information change amount between adjacent time nodes relative to its historical change range or the current analysis time period change range;

[0142] The determination process of the preset significant change proportion condition includes the following steps:

[0143] Based on the multi-source hydrological synchronous time series data set Syn formed in the historical period, the corresponding feature information in the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow velocity change response feature set Vrf is read respectively in multiple historical analysis time periods; in each historical analysis time period, the change amplitude of the corresponding feature information of adjacent time nodes is calculated respectively, and the change amplitude is calculated by ratio with the overall change range of the corresponding feature information in the historical analysis time period to obtain the historical change proportion distribution; based on the historical change proportion distribution, the proportion threshold for distinguishing general change and significant change is determined, and the proportion threshold is determined as the preset significant change proportion condition for subsequent significant change determination process.

[0144] The S3 further includes S32;

[0145] S32, based on the obtained water level change significant change time position, rainfall change significant change time position and flow velocity change significant change time position, the time difference is calculated, including:

[0146] The time difference between the water level change significant change time position and the rainfall change significant change time position is calculated to form the water level rainfall time difference result Lrt;

[0147] The time difference between the water level change significant change time position and the flow velocity change significant change time position is calculated to form the water level flow velocity time difference result Lvt;

[0148] The time difference between the rainfall change significant change time position and the flow velocity change significant change time position is calculated to form the rainfall flow velocity time difference result Rvt;

[0149] The water level rainfall time difference result Lrt, the water level flow velocity time difference result Lvt and the rainfall flow velocity time difference result Rvt are proportioned according to the time length of the analysis time period to obtain the normalized time difference result: water level rainfall time difference result Lrt, water level flow velocity time difference result Lvt and rainfall flow velocity time difference result Rvt, and are summarized to obtain the hydrological data time consistency offset Tco.

[0150] In this embodiment, by performing significant change determination on the feature information in the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow rate change response feature set Vrf under the same time reference, the water level change significant change time position, the rainfall change significant change time position and the flow rate change significant change time position are extracted respectively, so that the "key change moment" of water level, rainfall and flow rate can be clearly located; on this basis, the water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt and the rainfall flow rate time difference result Rvt are further constructed, and the influence of different analysis time length on the time difference result is eliminated through scaling processing, and finally the hydrological data time consistency offset Tco is formed. For example, in actual basin operation, the common situation is that the rainfall change response feature set Rrf first appears significant change, then the flow rate change response feature set Vrf responds, and the water level change response feature set Lrf obviously lags behind; the traditional method can only determine whether each index is abnormal, and it is difficult to quantify the degree of "sequential mismatch", but the method can unify the three types of time difference results through the hydrological data time consistency offset Tco, so that the time dislocation degree between rainfall-flow rate-water level is intuitively quantified, thereby providing a clear and comparable basis for subsequent abnormal trend determination based on the normal hydrological response consistency reference interval Ref. This analysis method with time response consistency as the core makes the early warning judgment no longer depend on whether a single index is out of limit, but can truly reflect the collaborative destruction process of multi-source hydrological elements in the flood incubation process, which is an important supplement and substantial improvement to the existing hydrological early warning technology.

[0151] In a specific embodiment, the S4 comprises an S41.

[0152] The S41 compares each normalized time difference result in the obtained hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref respectively.

[0153] The interval comparison analysis determines whether the water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt and the rainfall flow rate time difference result Rvt in the hydrological data time consistency offset Tco fall within the normal hydrological response consistency reference interval Ref.

[0154] When at least one normalized time difference result in the hydrological data time consistency offset Tco exceeds the normal hydrological response consistency reference interval Ref, it is determined that the consistency of the multi-source hydrological data in the time response dimension is abnormal.

[0155] When the normalized time difference result in the hydrological data time consistency offset Tco does not exceed the normal hydrological response consistency reference interval Ref, it is determined that the multi-source hydrological data is consistent in the time response dimension without abnormality;

[0156] At the same time, according to the number of normalized time difference results exceeding the normal hydrological response consistency reference interval Ref, an abnormal trend determination result Alm is generated.

[0157] The S4 further comprises S42;

[0158] S42, based on the abnormal trend determination result Alm, confirming the warning level of the current hydrological change process, specifically generating a first-level warning, a second-level warning and a third-level warning according to the number in the abnormal trend determination result Alm, and forming the hydrological abnormal warning information War corresponding to the warning level;

[0159] Synchronously, the obtained hydrological abnormal warning information War is converted into warning instruction information, and is issued through an instruction signaling device to send an abnormal warning instruction to the corresponding hydrological monitoring or management system.

[0160] In this embodiment, by item-by-item interval judgment of each normalized time difference result in the hydrological data time consistency offset Tco, the fine classification identification of the degree of hydrological anomaly is realized. Specifically, by comparing the water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt and the rainfall flow rate time difference result Rvt with the normal hydrological response consistency reference interval Ref respectively, not only whether the multi-source hydrological data exists time response consistency anomaly can be judged, but also according to the number of normalized time difference results exceeding the normal hydrological response consistency reference interval Ref, the abnormal trend judgment result Alm with clear level meaning can be generated. On this basis, the abnormal trend judgment result Alm is directly mapped into the hydrological anomaly warning information War of the first, second and third level warning, so that the warning result can truly reflect the severity of the current hydrological evolution process. For example, in actual flood control scheduling, if only the water level rainfall time difference result Lrt exceeds the normal hydrological response consistency reference interval Ref, the system can generate a first level warning to prompt that the rainfall and water level response start to appear mismatch; and when the water level rainfall time difference result Lrt, the water level flow rate time difference result Lvt and the rainfall flow rate time difference result Rvt exceed the reference interval at the same time, a third level warning can be directly generated to prompt that the multi-source hydrological response is overall unbalanced and the flood risk is significantly increased. By converting the classified hydrological anomaly warning information War into a warning instruction and issuing it to the hydrological monitoring or management system, the method can provide management personnel with decision-making basis matching the actual risk level, avoid the extensive warning problem of "either not reporting or reporting the highest level" in the prior art, and significantly improve the operability and response effectiveness of hydrological warning in real river basin scenarios.

[0161] Embodiment 6: A hydrological data real-time monitoring and warning system based on data analysis, please refer to Figure 2 , specifically: including a hydrological data acquisition module, a time sequence feature extraction module, a consistency offset analysis module and a monitoring and warning module;

[0162] The hydrological data acquisition module acquires water level data, rainfall data and flow rate data in the monitoring river basin, and generates a multi-source hydrological synchronous time sequence data set Syn;

[0163] The time sequence feature extraction module extracts water level change response features, rainfall change response features and flow rate change response features based on the multi-source hydrological synchronous time sequence data set Syn, to form a water level change response feature set Lrf, a rainfall change response feature set Rrf and a flow rate change response feature set Vrf;

[0164] The consistency offset analysis module constructs a hydrological data time consistency offset Tco reflecting the time response consistency of the multi-source hydrological data based on the water level change response feature set Lrf, the rainfall change response feature set Rrf and the flow velocity change response feature set Vrf;

[0165] The monitoring and early warning module compares and analyzes the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref, generates an abnormal trend determination result Alm, and issues the result through an instruction signaling device.

[0166] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and modifications can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended technical solutions and their equivalents.

Claims

1. A real-time monitoring and early warning method for hydrological data based on data analysis, characterized in that: Includes the following steps: S1. Collect and monitor water level data, rainfall data, and flow velocity data within the watershed to generate a multi-source hydrological synchronous time-series data set Syn; S2. Based on the multi-source hydrological synchronous time-series data set Syn, extract water level change response features, rainfall change response features and flow velocity change response features respectively to form water level change response feature set Lrf, rainfall change response feature set Rrf and flow velocity change response feature set Vrf; S3. Based on the water level change response feature set Lrf, the rainfall change response feature set Rrf, and the flow velocity change response feature set Vrf, construct a hydrological data time consistency offset Tco that reflects the time response consistency of multi-source hydrological data. S3 includes S31; S31. By reading the feature information corresponding to the time period in the water level change response feature set Lrf, the rainfall change response feature set Rrf, and the flow velocity change response feature set Vrf; Significant changes are determined by analyzing the feature information corresponding to the water level change response feature set Lrf, and the time and location of significant water level changes are obtained. Significant changes are determined in the feature information corresponding to the rainfall change response feature set Rrf to obtain the time and location of significant changes in rainfall; Significant changes are determined in the feature information corresponding to the flow velocity change response feature set Vrf to obtain the time and location of significant changes in flow velocity. The significant change determination is achieved by comparing the change range of corresponding feature information at adjacent time points. When the change range of feature information between adjacent time points reaches a preset significant change ratio condition, the corresponding time point is determined as the significant change time position. By aligning the time locations of significant changes in water level, rainfall, and flow velocity on a time reference, the offset relationship of the multi-source hydrological change response at each time point in the time dimension is formed. The feature information change ratio is used to characterize the proportion of the feature information change between adjacent time nodes relative to their historical change range or the change range of the current analysis period. S3 further includes S32; S32. Based on the obtained time locations of significant changes in water level, rainfall, and flow velocity, calculate the time difference, including: Calculate the time difference between the locations of significant changes in water level and the locations of significant changes in rainfall to form the water level-rainfall time difference result Lrt; Calculate the time difference between the locations of significant changes in water level and significant changes in flow velocity to form the time difference result Lvt between water level and flow velocity. Calculate the temporal difference between the locations of significant changes in rainfall and the locations of significant changes in flow velocity to form the rainfall-flow velocity temporal difference result Rvt; The time difference results of water level and rainfall (Lrt), water level and flow velocity (Lvt), and rainfall and flow velocity (Rvt) are proportionalized according to the length of the analysis period to obtain normalized time difference results: water level and rainfall time difference results (Lrt), water level and flow velocity (Lvt), and rainfall and flow velocity (Rvt). These results are then summarized to obtain the hydrological data time consistency offset (Tco). S4. Compare and analyze the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref to generate the abnormal trend judgment result Alm, and send it out through the command transmission device.

2. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 1, characterized in that: S1 includes S11; S11. Within the monitoring basin, based on the existing hydrological data acquisition system, acquire water level data, rainfall data, and flow velocity data respectively; Among them, for water level data, the water level observation values ​​at corresponding time points are obtained by periodically sampling the changes in water height at the monitoring section; For rainfall data, the rainfall observation value at the corresponding time point is obtained by continuously recording the precipitation per unit time in the monitoring area; For flow velocity data, the flow velocity observation values ​​at corresponding time points are obtained by periodically sampling the water flow state at the monitoring section. The water level observations are arranged in chronological order of acquisition to form a water level time series data set Lev; The rainfall observations are arranged in chronological order of acquisition to form a rainfall time-series data set Rai; The flow velocity observations are arranged in chronological order of acquisition to form a flow velocity time series data set, Vel.

3. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 2, characterized in that: S1 further includes S12; S12. Based on a unified time reference, the acquisition time of each data in the water level time series data set Lev, the rainfall time series data set Rai, and the flow velocity time series data set Vel is corrected to establish a corresponding relationship between data from different sources at the same time scale. Specifically, when there is missing data or inconsistent sampling at any time point in the water level time series data set Lev, the rainfall time series data set Rai, and the flow velocity time series data set Vel, the missing data is adjusted through time mapping. The processed water level time series data set Lev, rainfall time series data set Rai, and flow velocity time series data set Vel are integrated to generate a multi-source hydrological synchronous time series data set Syn.

4. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 3, characterized in that: S2 includes S21; S21. Based on the Syn set of multi-source hydrological synchronous time series data, calculate the changes in water level data, rainfall data and flow velocity data at the same time point respectively. For water level data, the change in water level between adjacent time points is calculated based on the difference in water level values ​​between adjacent time points, forming a water level change sequence; For rainfall data, the change in rainfall between adjacent time points is calculated based on the difference in rainfall values ​​between adjacent time points, forming a rainfall change sequence; For velocity data, the change in velocity between adjacent time points is calculated based on the difference in velocity values ​​between adjacent time points, forming a velocity change sequence.

5. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 4, characterized in that: S2 further includes S22; S22. Based on the water level change sequence, select an analysis period consisting of multiple consecutive time nodes, and determine the direction, amplitude, and continuity of the water level change sequence within the analysis period. The direction of water level change is determined by analyzing the sign distribution of the water level change sequence within the analysis period. The sign of the change is determined by the change in the water level change sequence. When the change is increasing, the sign is positive; when the change is decreasing, the sign is negative. Meanwhile, the magnitude of water level change is obtained based on the difference between the maximum and minimum values ​​of the water level change sequence within the analysis period; The continuity of water level changes is determined by analyzing the length of the continuous change symbol segment of the water level change sequence within the analysis period. The direction, magnitude, and continuity of water level changes are combined to form the water level change response feature set Lrf; Based on the rainfall change sequence, an analysis period consisting of multiple consecutive time nodes is selected, and the intensity, persistence, and concentration of the rainfall change sequence are statistically analyzed within the analysis period. The intensity of rainfall change is obtained based on the peak value and mean value of the rainfall change sequence within the analysis period. The duration of rainfall changes is determined by the length of consecutive positive segments in the rainfall change sequence within the analysis period. The concentration of positive values ​​in the rainfall change sequence within the analysis period is determined by analyzing the distribution of time nodes, thereby obtaining the persistence of rainfall changes. The intensity, duration, and concentration of rainfall changes are combined to form a rainfall change response feature set Rrf; Based on the velocity change sequence, an analysis period consisting of multiple consecutive time nodes is selected, and the velocity change sequence is subjected to direction determination, rate statistics, and stationarity determination within the analysis period. The direction of the flow velocity change is determined based on the sign distribution of the velocity change sequence within the analysis period. Based on the mean and peak values ​​of the velocity change sequence within the analysis period, flow velocity change rate description information is obtained, and flow velocity change rate is obtained. The stationarity of the flow velocity change sequence within the analysis period is determined based on the degree of dispersion, thereby obtaining the stationarity of the flow velocity change. The direction of velocity change, the rate of velocity change, and the stability of velocity change are combined to form the velocity change response characteristic set Vrf.

6. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 1, characterized in that: S4 includes S41; S41. Compare the normalized time difference results in the obtained hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref. Among them, the interval comparison analysis determines whether the time difference results of water level and rainfall time, Lrt, Lvt, and Rvt of rainfall and flow velocity in the hydrological data time consistency offset Tco fall within the normal hydrological response consistency reference interval Ref. When at least one normalized time difference result in the hydrological data time consistency offset Tco exceeds the normal hydrological response consistency reference interval Ref, the consistency of multi-source hydrological data in the time response dimension is determined to be abnormal. When no normalized time difference result in the hydrological data time consistency offset Tco exceeds the normal hydrological response consistency reference interval Ref, it is determined that the multi-source hydrological data has no abnormality in the time response dimension. Meanwhile, based on the number of normalized time difference results that exceed the normal hydrological response consistency reference interval Ref, an abnormal trend judgment result Alm is generated.

7. The method for real-time monitoring and early warning of hydrological data based on data analysis according to claim 6, characterized in that: S4 further includes S42; S42. Based on the abnormal trend judgment result Alm, the early warning level of the current hydrological change process is confirmed. Specifically, according to the quantity in the abnormal trend judgment result Alm, a first-level early warning, a second-level early warning, and a third-level early warning are generated, and hydrological abnormal early warning information War corresponding to the early warning level is generated. The acquired hydrological anomaly warning information (War) is simultaneously converted into warning instruction information and sent out through the instruction transmitting device to send the anomaly warning instruction to the corresponding hydrological monitoring or management system.

8. A real-time monitoring and early warning system for hydrological data based on data analysis, applied to the real-time monitoring and early warning method for hydrological data based on data analysis as described in any one of claims 1 to 7, characterized in that: It includes a hydrological data acquisition module, a time-series feature extraction module, a consistency offset analysis module, and a monitoring and early warning module; The hydrological data acquisition module collects and monitors water level data, rainfall data, and flow velocity data within the watershed, generating a multi-source synchronous time-series hydrological data set Syn. The time-series feature extraction module extracts water level change response features, rainfall change response features, and flow velocity change response features based on the multi-source hydrological synchronous time-series data set Syn, forming water level change response feature set Lrf, rainfall change response feature set Rrf, and flow velocity change response feature set Vrf. The consistency offset analysis module constructs a hydrological data time consistency offset Tco that reflects the time response consistency of multi-source hydrological data based on the water level change response feature set Lrf, the rainfall change response feature set Rrf, and the flow velocity change response feature set Vrf. The monitoring and early warning module compares and analyzes the hydrological data time consistency offset Tco with the pre-established normal hydrological response consistency reference interval Ref, generates the abnormal trend judgment result Alm, and sends it out through the command sending device.

Images