Intelligent extension method of water level-discharge relation based on similar flood

Abstract

The present application relates to the technical field of hydrological data analysis, and more particularly to a method for intelligently extending the water level-flow relationship based on similar floods. The method comprises the following steps: obtaining measured water level and flow data and performing curve fitting; performing three-property testing on the fitted curve, and if the testing is not satisfied, using multi-point segmented fitting to obtain an output curve; based on a historical flood pattern library, filtering similar curves and performing extension processing to form an extended water level-flow relationship curve; and using the extended curve to push the flow for real-time water level, thereby achieving accurate prediction and dynamic extension of the water level-flow relationship. The present application guarantees the reliability of the curve through measured water level-flow data fitting and segmented optimization, combined with three-property testing, and uses a historical flood similar curve extension and dynamic matching mechanism to achieve intelligent completion and real-time correction in the case of insufficient data, thereby significantly improving the accuracy, continuity and adaptability of the water level-flow relationship curve, and ensuring the accuracy and practicality of the hydrological flow pushing process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrological data analysis, and particularly relates to an intelligent extension method for water level-flow relationship based on similar floods. BACKGROUND

[0002] In recent years, the timeliness requirement of hydrological data for flood control, water resources assessment, water ecological protection and the like is higher and higher, and real-time flow data estimation has become a new requirement for current data compilation. However, the existing real-time flow data estimation mostly relies on manual experience line setting, and excessively relies on manual experience and technical management, so that the water level-flow relationship curve cannot be extended in time and accurately, and the real-time estimated flow has a large error. Therefore, how to extend the water level-flow relationship curve in time and accurately is a problem to be solved in the current hydrological field. SUMMARY

[0003] Therefore, it is necessary to provide an intelligent extension method for water level-flow relationship based on similar floods to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, an intelligent extension method for water level-flow relationship based on similar floods, the method comprises the following steps:

[0005] Step S1: acquiring measured flow and water level process data; fitting the measured flow and water level process data according to the measured water level points and flow points to obtain a fitted water level-flow relationship curve;

[0006] Step S2: judging whether the fitted water level-flow relationship curve passes the three-property test, if yes, marking the fitted water level-flow relationship curve as an output curve; if no, performing multi-point segmented fitting on the fitted water level-flow relationship curve to obtain a secondary fitted output curve;

[0007] Step S3: filtering a similar curve matched with the output curve by using a preset historical flood pattern library, and extending the water level-flow process of the output curve through the similar curve to obtain an extended water level-flow relationship curve;

[0008] Step S4: performing water level flow estimation on the collected real-time water level based on the extended water level-flow relationship curve.

[0009] Preferably, the multi-point segmented fitting on the fitted water level-flow relationship curve in step S2 comprises:

[0010] dividing the fitted water level-flow relationship curve into a plurality of continuous water level sections according to the slope change characteristics of the water level process, and taking the measured water level points and flow points in each section as a feature point set of the section;

[0011] For any one water level section, a first difference feature of the water level section is obtained according to a difference between a flow value of each feature point in the section and an average of flow values of all feature points in the section;

[0012] A second difference feature of the water level section is obtained according to a difference between a water level slope of each feature point in the section and an average of water level slopes of all feature points in the section;

[0013] The segmented fitting curve of the water level section is determined according to the first difference feature and the second difference feature;

[0014] The segmented fitting curves of all water level sections are connected in order of water levels to obtain an output curve after quadratic fitting.

[0015] Preferably, the determination of the segmented fitting curve of the water level section according to the first difference feature and the second difference feature comprises:

[0016] The maximum first difference feature point and the maximum second difference feature point and the remaining feature points in the water level section are screened according to the first difference feature and the second difference feature;

[0017] The segmented curve contour is established based on the maximum first difference feature point and the maximum second difference feature point;

[0018] The curve point fitting of the remaining feature points is performed through the segmented curve contour to obtain the segmented fitting curve of the water level section.

[0019] Preferably, the curve point fitting of the remaining feature points through the segmented curve contour comprises:

[0020] The number of the feature points of the remaining feature points is determined, and when the number of the remaining feature points is greater than a preset fitting standard number, the adjacent remaining feature points are merged so that the number of the feature points is equal to the preset fitting standard number;

[0021] When the number of the remaining feature points is less than the preset fitting standard number, the preset compensation points are inserted based on intervals of the adjacent remaining feature points so that the number of the feature points is equal to the preset fitting standard number.

[0022] Preferably, the determination of whether the fitted water level-flow relationship curve passes the three-property test in step S2 specifically comprises a sign test, a line fitting test, and a deviation value test.

[0023] Preferably, step S3 comprises the following steps:

[0024] Step S31: The water level range of the output curve is taken as an initial retrieval boundary, and similar curves matching the output curve are screened from a preset historical flood pattern library;

[0025] Step S32: determining the extension section of the output curve and performing smoothing processing according to the curve line type of the similar curve, to obtain the smoothed curve extension section;

[0026] Step S33: extending the water level-flow process of the output curve through the curve extension section, to obtain the extended water level-flow relationship curve.

[0027] Preferably, step S31 comprises the following steps:

[0028] Step S311: taking the water level range of the output curve as the initial retrieval boundary;

[0029] Step S312: screening the similar curve matching the output curve from the preset historical flood pattern library according to the initial retrieval boundary, and if the matching is successful, directly outputting the similar curve;

[0030] Step S313: if the matching fails, cyclically expanding the matching range by a preset step until the similar curve is outputted after the matching is successful, and recording the expansion times and the corresponding similarity change data.

[0031] Preferably, step S33 further comprises the following steps:

[0032] Collecting the real-time water level;

[0033] Confirming the water level range based on the real-time water level;

[0034] If the range maximum or the range minimum of the water level range does not belong to the extended water level-flow relationship curve, repeating the execution of steps S31-S33.

[0035] Preferably, the method for obtaining the measured water level points and the flow points comprises the following steps:

[0036] Obtaining the station position data; based on the station position data, arranging the water level observation device, and collecting the water level data at a preset time interval to obtain the continuous measured water level sequence as the measured water level points;

[0037] Obtaining the water level observation section position data; based on the corresponding section position of the water level observation section position data, arranging the flow measurement device, and calculating the flow value through the collected measured flow velocity and the section area to form the corresponding measured flow sequence as the flow points.

[0038] Preferably, the method for constructing the preset historical flood pattern library comprises the following steps:

[0039] Obtaining the historical water level-flow curve data;

[0040] Extracting the curve features of the historical water level-flow curve data to construct the curve feature vector library;

[0041] The real-time data of the curve feature vector library is updated to obtain a preset historical flood pattern library.

[0042] The present application has the following advantages:

[0043] I. By fitting the measured water level and flow points with a curve and combining the three-property test for effectiveness determination, the problem of excessive deviation or insufficient stability of the traditional single fitting method can be avoided. When the fitting result fails to meet the three-property test, the present application uses multi-point segmented fitting, which segments the curve according to the slope change characteristics of the water level process and combines the difference characteristics for fitting optimization, effectively improving the accuracy and rationality of the fitting curve and ensuring the continuity and stability of the output curve, thereby improving the reliability of the water level-flow relationship.

[0044] II. By establishing a historical flood pattern library and extracting curve feature vectors, similar curves can be quickly retrieved from the historical library and extended in the case of limited current fitting curve data, overcoming the defect that the existing method cannot derive a complete relationship curve when the flood data is insufficient or missing. At the same time, through similar curve extension and smoothing processing, the natural transition of the extended curve segment is effectively ensured, avoiding sudden changes or discontinuities in the curve splicing process, and improving the overall fitting effect of the water level-flow relationship.

[0045] III. In the process of extending the curve, dynamic matching and cyclic expansion retrieval strategies are introduced. If the initial retrieval boundary fails to match successfully, similar curves are obtained by gradually expanding the retrieval range, and the similarity change data is recorded to provide a reference for the accuracy optimization of subsequent curve extension. This dynamic retrieval method enhances the adaptability and robustness of the method, which can adapt to the differences in curve characteristics under different hydrological conditions.

[0046] IV. By combining real-time water level data, the extended water level-flow relationship curve is corrected in real time. When the real-time hydrological range exceeds the current curve range, the historical library can be called again for iterative extension to ensure that the derived result always remains consistent with the actual hydrological process, thereby improving the real-time and adaptability of the method. Through this mechanism, the present application can support the dynamic flow derivation demand in the flood process, providing more accurate support for hydrological forecasting and flood control scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 Figure 1 is a schematic diagram of the step flow of the intelligent extension method of the water level-flow relationship based on similar flood;

[0048] Figure 2 Figure 2 is a detailed implementation step flow schematic diagram of step S3 in Figure 1; Figure 1

[0049] Figure 3 ​The water level flow pushing curve chart of the water level flow relationship intelligent extension method based on similar flood;

[0050] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0051] The technical method of the present application will be described clearly and completely below in combination with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0052] In addition, the accompanying drawings are only schematic illustrations of the present application, and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus repeated description thereof will be omitted. Some block diagrams shown in the drawings are functional entities, and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0053] It should be understood that although the terms "first", "second" and the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the exemplary embodiments, a first element can be referred to as a second element, and similarly a second element can be referred to as a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0054] To achieve the above-mentioned purpose, please refer to Figures 1 to 3 The method comprises the following steps:

[0055] Step S1: obtaining the measured flow and water level process data; fitting the measured flow and water level process data according to the measured water level points and flow points to obtain the fitted water level flow relationship curve;

[0056] In an embodiment, the measured water level process data is collected in real time by automatic water level gauges (such as ultrasonic water level gauges or pressure water level gauges) arranged at the river or channel section; and the measured flow data is calculated by flow meters (such as electromagnetic flow meters, acoustic Doppler flow profile meters) combined with the section area. To ensure the reliability of the fitted curve, the sampling time interval is preferably set to 5 minutes to 30 minutes, and the sampling interval is appropriately shortened during the flood process or sudden increase of incoming water to improve the monitoring accuracy.

[0057] In another embodiment, the collection range of the measured data should cover the low water level, normal water level and high water level stages to ensure the applicability and accuracy of the fitted relationship curve. For example, at least 20 groups of water level-flow corresponding points can be collected during the dry season, wet season and flood season respectively to form a data set covering the entire hydrological process. In the fitting method, an empirical formula method (such as a power function form , a least squares regression method or a piecewise fitting method can be selected to obtain the water level-flow relationship curve. Among them, represents the flow, represents the water level, represents the initial water level, is the fitting coefficient. Through fitting, a continuous and smooth water level-flow curve can be obtained.

[0058] For example, the measured data collected at a certain river section shows that when the water level is 1.2 m to 3.5 m, the measured flow range is 10 m³ / s to 320 m³ / s; and the curve obtained by power function regression fitting is , and the fitting determination coefficient is 0.985, indicating that the fitting effect is good.

[0059] Step S2: judging whether the fitted water level-flow relationship curve passes the three-property test, if yes, marking the fitted water level-flow relationship curve as an output curve; if no, performing multi-point piecewise fitting on the fitted water level-flow relationship curve to obtain a secondary fitted output curve;

[0060] In an embodiment, the three-property test specifically includes a sign test, a line fitting test and a deviation value test: in the sign test, the number of positive and negative signs of the measured points deviating from the curve (the value of zero is equally distributed to the positive and negative signs) is counted respectively, and the statistical value is calculated according to , wherein is the statistical value, is the total number of measured points, is the number of positive or negative signs, is the probability of positive or negative sign, each is 0.5, For continuous correction number (discrete type to continuous type), and check the distribution of statistical value is in line with randomness and regularity. When the residual positive and negative distribution, and there is no large interval within the sign of a single or systematic deviation, determine the pass; otherwise, determine the fail.

[0061] Goodness-of-fit test: according to the water level of the measuring point from low to high in the order, from the second point to start statistics deviation positive and negative sign change, change sign is recorded as 1, otherwise 0, and compared with the theoretical expected number. When the number of sign change falls into the reasonable interval under the level of significance, determine the pass; if significantly less or more, determine the fail.

[0062] Deviation value test: by formula And formula Calculate the statistical quantity, the standard deviation of the average relative deviation value, respectively, wherein the formula For statistical quantity, For the average relative deviation value, For the standard deviation of the average relative deviation value, For the standard deviation of the relative deviation value, For the total number of measuring points, For the relative deviation value of the measuring point and the relationship curve, and Compared with the theoretical expected number, when the statistical quantity falls into the reasonable interval under the level of significance, determine the reasonable, that is, accept the test; otherwise, reject the original hypothesis.

[0063] In another embodiment, when the fitting curve fails to pass any one of the above three nature test, the system will automatically perform multi-point segmented fitting: that is, according to the water level interval is divided into several segments (such as low water level interval 1.0m-2.5m, medium water level interval 2.5m-3.8m, high water level interval 3.8m-5.0m), the local function fitting (power function, quadratic polynomial or logarithmic function, etc.) is carried out for each segment, and the smooth transition is used at the junction of the segments, so as to generate the quadratic fitting output curve which meets the sign rule, has good adaptability and deviation within the allowable range.

[0064] For example, in a certain river section monitoring, the sign test of the overall fitting curve fails, and the deviation of the high water level interval exceeds 20%. After using segmented fitting: the low water level segment uses power function , =0.981; the medium water level segment uses quadratic polynomial , =0.989; the high water level segment uses logarithmic function , =0.985. After segmented splicing and smooth transition, the three nature tests are all passed, and the final output curve can be accurately used for subsequent water level and flow conversion.

[0065] Step S3: Screen similar curves matching the output curve from the preset historical flood pattern library, and extend the water level-flow process of the output curve through the similar curves to obtain an extended water level-flow relationship curve;

[0066] In an embodiment, the historical flood pattern library pre-stores water level-flow relationship curve data of a plurality of typical flood processes, covering different river basins and different flood types (such as slow-rising and slow-falling type, rapid-rising and rapid-falling type, double-peak type, etc.). After the output curve in step S2 is obtained, the similar curve screening and extension are performed through the following steps:

[0067] The output curve and each curve in the historical flood pattern library are normalized, and a dynamic time warping (DTW), a correlation coefficient matching or a Euclidean distance minimization method is used to calculate a similarity index. When the similarity is greater than 0.9, it is determined as a similar curve candidate. Further error analysis is performed on a plurality of candidate curves, and the curve with the smallest residual sum of squares and the closest trend is selected as the final similar curve. The end point of the output curve is used as a connection point, the water level-flow change rule of the similar curve after the connection point is used for extension splicing, and a smooth transition algorithm (such as weighted average or spline interpolation) is used at the splicing place to generate an extended water level-flow relationship curve.

[0068] In another embodiment, to ensure the reasonableness of the extended part, a constraint condition is set for the extended curve: the flow rate of change of the extended segment is not more than twice that of the last segment of the original curve, so as to prevent the occurrence of non-physical mutations; the peak flow and the peak water level of the extended segment should be kept within the 95% confidence interval of the historical similar flood curve; the fitting determination coefficient of the entire extended curve is kept above 0.95.

[0069] For example, in the monitoring of a certain river basin in the 2022 flood season, the output curve lacks data at the high water level segment, resulting in an incomplete flood peak interval. By calling the “rapid-rising and slow-falling type” flood process curve in the historical flood pattern library, a similarity of 0.936 is obtained through similarity calculation, and the curve is selected as the optimal candidate curve. Finally, a complete water level-flow relationship curve is obtained through extension splicing, so that the flood peak flow is extended from the original fitting of 2100 m³ / s to 2650 m³ / s, which is in good agreement with the historical similar flood process.

[0070] Step S4: Water level flow is pushed based on the extended water level-flow relationship curve.

[0071] In an embodiment, reference is made to Figure 3 ​, first, the real-time water level data of the target section is obtained through a real-time water level monitoring device (such as a water level gauge or a radar water level sensor), and the real-time water level is taken as an input. The extended water level-flow relationship curve obtained in step S3 is called to match the real-time water level with the curve corresponding relationship, and the corresponding real-time flow value is obtained. The specific processing process includes: positioning the real-time water level value to the coordinate axis of the extended water level-flow relationship curve, and confirming the corresponding interval; when the real-time water level falls between the known sampling points, the piecewise linear interpolation or cubic spline interpolation method is used to calculate the corresponding flow value, so as to avoid the insufficient accuracy caused by discrete points; the real-time water level and the calculated flow value are paired to form a real-time water level-flow pair, which is output as the water level flow pushing result.

[0072] In another embodiment, in order to ensure the stability of the flow pushing result, the real-time calculated flow value is set to have a smoothing processing and a constraint condition: when the flow change rate in the continuous two sampling periods exceeds 15%, an abnormal correction is triggered, and a sliding average method is used to smooth the result; when the real-time water level exceeds the definition interval of the extended water level-flow relationship curve, the system triggers an extrapolation algorithm, and the end trend is used for extension calculation, but the extrapolation interval does not exceed 10% of the water level at the end of the curve, so as to prevent distortion.

[0073] For example, in real-time monitoring of a certain hydrological station, when the measured water level is 12.8 m, the corresponding flow is 1840 m³ / s according to the extended water level-flow relationship curve; when the water level rises to 14.2 m, the corresponding flow is calculated to be 2460 m³ / s. Through the water level flow pushing method, the real-time monitoring and prediction of the flood process can be realized in the case of lacking measured flow.

[0074] Preferably, the multi-point piecewise fitting of the fitted water level-flow relationship curve in step S2 includes:

[0075] The fitted water level-flow relationship curve is divided into a plurality of continuous water level sections according to the slope change characteristics of the water level process, and the measured water level points and flow points in each section are taken as the characteristic point set of the section;

[0076] For any water level section, the first difference characteristic of the water level section is obtained according to the difference between the flow value of each characteristic point in the section and the average flow value of all characteristic points in the section;

[0077] The second difference characteristic of the water level section is obtained according to the difference between the water level slope of each characteristic point in the section and the average water level slope of all characteristic points in the section;

[0078] The piecewise fitting curve of the water level section is determined according to the first difference characteristic and the second difference characteristic;

[0079] The segmented fitting curves of all water level segments are connected in water level order to obtain the output curve after quadratic fitting.

[0080] In an embodiment, when the fitted water level-flow relationship curve fails to pass the three-property test, it is fitted by multi-point segmentation to improve the matching accuracy of the fitted curve and the measured data.

[0081] First, the fitted water level-flow relationship curve is segmented according to the slope change characteristics of the water level process. Specifically, the slope change rate of each point on the curve is calculated, and when the slope change rate of adjacent points exceeds a set threshold (for example, 15%-25%), a demarcation point is automatically generated to divide the overall curve into several continuous water level segments. In each water level segment, the measured water level points and flow points of the segment are extracted as a set of feature points. For any water level segment, the difference between the flow value of each feature point and the average flow value of all feature points in the segment is calculated and recorded as the first difference characteristic of the segment. At the same time, the difference between the water level slope of each feature point in the segment and the average water level slope of all feature points in the segment is calculated and recorded as the second difference characteristic of the segment.

[0082] In another embodiment, the system determines the segmented fitting curve of the segment based on the first difference characteristic and the second difference characteristic. Specifically, when the first difference characteristic is small and the second difference characteristic is large, a quadratic polynomial fitting is preferred to reflect the rapid change of the slope; when the first difference characteristic is large and the second difference characteristic is small, a segmented linear fitting is used to ensure the stability of the flow value in the segment; when both are large, a segmented cubic spline fitting is used to balance the smoothness and accuracy of the fitting. Finally, the segmented fitting curves of all water level segments are spliced and connected in water level order to obtain a continuous and smooth output curve after quadratic fitting.

[0083] For example, in a certain measured data set, the original fitted curve deviates significantly in the interval of water level 8.5m to 10.2m. Through multi-point segmented fitting, the interval is divided into three sub-segments, of which the first segment is fitted by linear fitting, the second segment is fitted by quadratic fitting, and the third segment is fitted by cubic spline fitting. The final spliced quadratic fitting curve significantly improves the overall fitting accuracy and successfully passes the three-property test.

[0084] Preferably, determining the segmented fitting curve of the water level segment according to the first difference characteristic and the second difference characteristic comprises:

[0085] According to the first difference characteristic and the second difference characteristic, the maximum first difference characteristic point and the maximum second difference characteristic point in the water level segment are screened out from the rest of the feature points.

[0086] Based on the maximum first difference characteristic point and the maximum second difference characteristic point, a segmented curve contour is established.

[0087] fitting the remaining feature points by the segmented curve profile to obtain a segmented fitting curve of the water level section.

[0088] In an embodiment, the first difference feature (i.e. the degree of deviation of flow from the mean value) and the second difference feature (i.e. the degree of deviation of water level slope from the mean value) are calculated for all feature points in the water level section. After the calculation, the maximum first difference feature point and the maximum second difference feature point are selected from the feature points, and the remaining feature points are regarded as ordinary points.

[0089] Secondly, the segmented curve profile of the water level section is established based on the spatial position and numerical relationship of the maximum first difference feature point and the maximum second difference feature point. The maximum first difference feature point mainly reflects the extreme change trend of flow, and the maximum second difference feature point mainly reflects the extreme change trend of water level slope, and the two points jointly determine the key inflection point and curve trend of the water level section.

[0090] Then, the remaining feature points in the section are fitted by taking the segmented curve profile as a constraint condition. Specifically, the ordinary points are fitted by the weighted least square method or the spline interpolation method under the premise of ensuring consistency with the segmented curve profile, so as to obtain the segmented fitting curve of the water level section.

[0091] In another embodiment, if the water level interval between the maximum first difference feature point and the maximum second difference feature point exceeds a preset threshold (for example, 0.3m-0.5m), an additional auxiliary point is introduced into the segmented curve profile to avoid over-bending or distortion of the fitting curve.

[0092] For example, in a certain water level section (9.0m-10.0m), the calculation result shows that the flow of a certain feature point deviates from the section mean value by 15%, which is determined as the maximum first difference feature point; another feature point deviates from the section mean value by 20% in slope, which is determined as the maximum second difference feature point. After the segmented curve profile is established based on the two extreme points, the other ordinary points are fitted by weighting, and the segmented curve finally obtained can reflect the mutation characteristics of flow and the change trend of water level slope, and the fitting accuracy is significantly improved.

[0093] Preferably, the fitting the remaining feature points by the segmented curve profile comprises:

[0094] judging the number of the remaining feature points, and when the number of the remaining feature points is greater than a preset fitting standard number, merging adjacent remaining feature points so that the number of the feature points is equal to the preset fitting standard number.

[0095] When the number of the rest feature points is less than the preset fitting standard number, preset compensation points are inserted based on the interval of the adjacent rest feature points, so that the number of the feature points is equal to the preset fitting standard number.

[0096] In an embodiment, a rest feature point set except the maximum first difference feature point and the maximum second difference feature point is obtained, and the number of the rest feature points is counted. When it is determined that the number of the rest feature points is greater than the preset fitting standard number (for example, 6 points), a feature point merging operation is performed. Specifically, two adjacent feature points are merged into one representative point based on the order of the feature points on the water level coordinate axis, and the merging mode adopts arithmetic average or weighted average, so that the number of the finally reserved feature points is exactly equal to the preset fitting standard number. This mode can avoid the problem of excessive curve fitting caused by too many feature points.

[0097] When it is determined that the number of the rest feature points is less than the preset fitting standard number, a feature point compensation operation is performed. Specifically, one or more compensation points are inserted in the water level interval of adjacent feature points according to the linear interpolation or spline interpolation mode, until the number of the rest feature points reaches the fitting standard number. This mode can avoid the problem of too rough curve caused by insufficient feature points. Finally, the processed feature point set (the number of which is equal to the fitting standard number) is combined with the segmented curve contour to generate a smooth segmented fitting curve of the water level section through least square fitting or spline curve fitting.

[0098] In a specific embodiment, the fitting standard number can be preferably set to 5-8 points. For example, in a certain water level section, if the number of the rest feature points is 12 and the fitting standard number is set to 6, 6 representative points are finally obtained by merging adjacent points; if the number of the rest feature points is only 3, 3 compensation points are inserted between adjacent points, so that the total number reaches 6, ensuring the stability and reliability of the fitting curve.

[0099] Preferably, the judgment in step S2 whether the fitted water level-flow relationship curve passes the three-property test specifically includes a sign test, a line fitting test, and a deviation value test.

[0100] In an embodiment, the judgment whether the fitted water level-flow relationship curve passes the three-property test specifically includes the following test mode: the difference between the measured flow and the fitted flow is calculated point by point for the fitted water level-flow relationship curve, and the number of sign changes of the difference is counted. If the number of sign changes is greater than a set reasonable range (for example, no more than 30% of the total number of points), it is indicated that the curve is consistent with the measured data in the overall trend, and it is determined that the sign test passes; otherwise, it is determined that the sign test does not pass.

[0101] The correlation coefficient between the fitted water level-flow relationship curve and the measured points is calculated. When the correlation coefficient reaches a preset threshold (for example, more than 0.95), it is indicated that the fitted curve can better reflect the linear or nonlinear relationship of the measured data, so that it is determined that the linearity test is passed; otherwise, it is determined that it is not passed.

[0102] The absolute value of the deviation between the measured flow and the fitted flow is calculated, and the root mean square error (RMSE) of all absolute deviation values is obtained. If the root mean square error is less than a set threshold (for example, not more than 5%-10% of the average value of the measured flow), it is indicated that the fitting precision of the fitted curve to the measured data meets the requirements, and it is determined that the deviation value test is passed; otherwise, it is determined that it is not passed.

[0103] When all three of the above three tests are passed, the fitted water level-flow relationship curve is determined as an output curve that meets the requirements of precision and stability; if any test is not passed, the subsequent multi-point segmented fitting is triggered to improve the fitting effect and stability of the curve.

[0104] In a specific embodiment, the threshold of the symbol test can be selected to be not more than 25%-30% of the total number of points; the correlation coefficient threshold of the linearity test can be preferably 0.95-0.98; and the root mean square error threshold of the deviation value test can be set to not more than 10% of the average value of the measured flow. Through the comprehensive determination of the above indexes, the overall reliability of the fitted curve in trend consistency, correlation and precision can be effectively guaranteed.

[0105] As an example of the present application, reference is made to Fig. 1, which shows a flowchart of a method for fitting a water level-flow relationship curve according to an embodiment of the present application. In the present example, the step S3 includes: Figure 2

[0106] Step S31: Based on the water level range of the output curve as the initial search boundary, and from the preset historical flood pattern library, similar curves matching the output curve are screened;

[0107] Step S32: According to the curve line type of the similar curves, the extension section of the output curve is determined and smoothed to obtain the smoothed curve extension section;

[0108] Step S33: The water level-flow process of the output curve is extended through the curve extension section to obtain the extended water level-flow relationship curve.

[0109] ​In an embodiment, a water level range of the output curve is determined as an initial retrieval boundary. Then, a curve with a similar line type as the output curve in the water level range is screened from a preset historical flood pattern library. The screening process can include the following operations: normalizing the output curve and the historical library curve, unifying the dimensions of water level and flow rate; calculating a similarity index of the output curve and the historical library curve within the initial retrieval boundary, which can use the Euclidean distance, dynamic time warping (DTW), or correlation coefficient method; and screening a number of similar curves according to the similarity index as a candidate extended curve set.

[0110] According to the line type characteristics of the candidate similar curves, an extension section of the output curve is determined. The specific method includes: analyzing the line type trend of the similar curve outside the water level range of the output curve, such as the slope change, convexity, and other characteristics; performing curve smoothing on the extension section of the similar curve and the end point of the output curve, so that the extension section and the output curve have no obvious jump or mutation at the joint, which can use the spline smoothing or moving average filtering method; obtaining the smoothed extension section curve, and ensuring that the extension section is continuous and consistent with the output curve in the water level change trend.

[0111] The water level and flow rate process of the output curve is extended through the smoothed extension section. Specifically, the smoothed extension section is connected to the end of the output curve, so that the output curve can be extended outside the original water level range; the connection is locally fine-tuned to ensure continuous and non-mutated flow rate change; and finally, the extended water level and flow rate relationship curve is obtained, which can be used for subsequent real-time water level flow pushing or flood forecasting.

[0112] In a specific embodiment, the joint of the end of the output curve and the extension section of the similar curve can be set to a ±0.5% flow rate change tolerance to ensure that the extended curve is smooth and applicable.

[0113] Preferably, the step S31 includes the following steps:

[0114] Step S311: determining a water level range of the output curve as an initial retrieval boundary;

[0115] Step S312: screening a similar curve matching the output curve from a preset historical flood pattern library according to the initial retrieval boundary, and if the matching is successful, directly outputting the similar curve;

[0116] Step S313: if the matching fails, cyclically expanding the matching range by a preset step size until the similar curve is outputted by matching, and recording the expansion times and the corresponding similarity change data.

[0117] In an embodiment, the initial search boundary is determined according to the water level range of the output curve. The initial search boundary is the water level interval between the lowest water level point and the highest water level point in the output curve, which is used to limit the screening range of the historical flood pattern library. Within the initial search boundary, similar curves matching the output curve are screened from the preset historical flood pattern library. The specific method can include: normalizing the output curve and the historical library curve to unify the dimension; using a similarity calculation method, such as Euclidean distance, dynamic time warping (DTW), or correlation coefficient, to compare the matching degree of the output curve and the historical curve within the search boundary; if the similarity of the screened historical curve is higher than a preset threshold, it is determined that the matching is successful, and the historical curve is directly used as the output similar curve.

[0118] If no curve satisfying the similarity threshold is matched within the initial search boundary, the matching range is expanded in a loop: the water level range of the output curve is gradually expanded by a preset step, for example, increasing the water level interval by ±0.01 meters each time; after each expansion, the similar curve screening is re-executed, and the matching similarity is calculated; the above operations are repeatedly executed until the matching is successful; after the matching is successful, the number of loop expansions and the similarity change data corresponding to each expansion are recorded, which are used for subsequent analysis of the stability and reliability of the output curve matching.

[0119] In a specific embodiment, the preset matching similarity threshold is 0.85 to 0.95, and the matching range step is set to 0.1 to 0.2 meters, which can ensure that when the historical library lacks an accurate matching curve, a similar trend curve can still be found and reasonably extended.

[0120] Preferably, step S33 further includes:

[0121] Collecting real-time water level;

[0122] Confirming the water level range based on the real-time water level;

[0123] If the maximum value or the minimum value of the real-time hydrological range does not belong to the extended water level-flow relationship curve, steps S31-S33 are repeatedly executed.

[0124] In an embodiment, the real-time water level of the target hydrological station or monitoring point is collected to obtain the latest water level data sequence. Based on the collected real-time water level data, the real-time hydrological range is confirmed, that is, the minimum value and the maximum value in the current water level data sequence are determined to reflect the fluctuation interval of the current hydrological condition.

[0125] Comparing the real-time hydrological range with the extended water level-discharge relation curve: if the maximum or minimum of the real-time hydrological range is not within the water level range covered by the extended water level-discharge relation curve, it is determined that the extended curve cannot completely cover the current hydrological condition; in this case, steps S31 to S33 are repeatedly executed, i.e., the water level range of the current output curve is taken as the initial search boundary, similar curves in the historical flood pattern library are screened, and a new extended curve is generated, until the extended water level-discharge relation curve covers the real-time hydrological range.

[0126] In a preferred embodiment, the maximum number of times of cyclic execution can be set, for example, no more than 5, to avoid infinite loop; at the same time, the water level range change and the historical curve matching similarity at each time of curve extension can be recorded for analyzing the stability and reliability of curve extension.

[0127] Preferably, the method for obtaining the measured water level points and the measured discharge points comprises:

[0128] Obtaining station position data; based on the station position data, arranging water level observation devices, and collecting water level data at a preset time interval to obtain a continuous measured water level sequence as the measured water level points;

[0129] Obtaining water level observation section position data; based on the corresponding section position of the water level observation section position data, arranging discharge measurement devices, and calculating the discharge value through the collected measured flow velocity and the section area to form a corresponding measured discharge sequence as the discharge points.

[0130] In an embodiment, first, the station position data of the target hydrological station is obtained, including the latitude and longitude coordinates and elevation information of the station. Based on the station position data, water level observation devices such as water level gauges, water gauges or ultrasonic water level sensors are arranged on site, and water level data is automatically collected at a preset time interval (for example, every 5 to 30 minutes) to form a continuous water level time sequence. Each sampling point in the water level time sequence is a measured water level point.

[0131] The position data of the water level observation section is obtained, including the start point, end point and section geometric shape information of the section. Discharge measurement devices such as multi-point flow velocity measurement instruments, electromagnetic flowmeters or radar flow velocity sensors are arranged at the water level observation section, and the discharge value is calculated by combining the measured flow velocity data collected on the section with the section geometric area. The discharge values of the sampling points of the section are integrated to obtain a continuous discharge time sequence corresponding to the measured water level sequence, forming the measured discharge points.

[0132] To ensure the time correspondence of the water level points and the discharge points, a unified time synchronization device such as a GPS time synchronization module can be used to ensure that the sampling time error of water level and discharge measurement is within ±1 second.

[0133] Preferably, the method for constructing the preset historical flood pattern library comprises:

[0134] Obtaining historical water level-flow curve data;

[0135] Extracting curve features of the historical water level-flow curve data to construct a curve feature vector library;

[0136] Performing real-time data updating on the curve feature vector library to obtain the preset historical flood pattern library.

[0137] In an embodiment, the water level-flow curve data of historical hydrological stations, including water level time series and corresponding flow time series, are collected, covering as long a historical time range as possible (such as ten years or more years of continuous flood event records) to ensure the completeness and representativeness of the data in the library.

[0138] The collected historical water level-flow curve data are subjected to feature extraction, including but not limited to: starting and ending water levels of the curve, peak flow, rising / descending rate, number and position of inflection points, curve shape index, etc. The features of each curve are sorted into a curve feature vector to form a curve feature vector library.

[0139] To ensure the real-time performance and scalability of the historical flood pattern library, a real-time data updating mechanism is set for the curve feature vector library. When new water level-flow measurement data or newly occurring flood event data are collected, the corresponding curve feature vectors are calculated and added to the feature vector library; at the same time, if there are repeated or outdated feature vectors, they can be overwritten or marked to maintain the effectiveness and availability of the data in the library. After the above processing, the obtained curve feature vector library is constructed as the preset historical flood pattern library, which can be used for subsequent similar curve matching of output curves and water level-flow process extension operations, providing reliable reference data for hydrological prediction and water level-flow prediction.

[0140] Especially importantly, the flow value calculated from the collected measured flow velocity and cross-sectional area also includes:

[0141] The collected measured flow velocity is sampled and sorted to obtain flow velocity sequence data;

[0142] The cross-sectional area is measured and calibrated to obtain cross-sectional area sequence data;

[0143] Based on the flow velocity sequence data and the cross-sectional area sequence data, dynamic deviation correction multiplication calculation and smoothing processing are performed to generate smoothed and corrected flow data;

[0144] The smoothed and corrected flow data are arranged in time sequence to generate measured flow sequence data.

[0145] In an embodiment, the flow velocity measuring device arranged by the water level observation section collects real-time flow velocity data of the water body at each section position, and data collection is performed according to a preset sampling time interval (such as every 5 seconds or once per minute), so as to obtain continuous flow velocity sequence data. The section is measured and calibrated on site, including measuring the section width, water depth and section contour shape, and calculating the effective cross-sectional area of the section according to the measurement results to form section area sequence data. For irregular sections, the section can be divided into several small unit areas to measure the area and flow velocity respectively, and comprehensive calculation is performed.

[0146] Based on the collected flow velocity sequence data and section area sequence data, the instantaneous flow is calculated by using a dynamic deviation correction method. Specifically, the flow velocity sequence is subjected to outlier rejection and mean value smoothing processing; the section area sequence is subjected to seasonal or water level height change correction; the corrected flow velocity is multiplied by the section area at each time or point to obtain instantaneous flow data. The instantaneous flow data is subjected to data smoothing to obtain smoothed and corrected flow data, and the smoothed and corrected flow data calculated at each time is arranged in time sequence to form complete measured flow sequence data, which is used for subsequent water level-flow relationship curve fitting and flood propagation analysis.

[0147] Especially important is that the dynamic deviation correction multiplication calculation based on the flow velocity sequence data and the section area sequence data also includes:

[0148] Local trend analysis is performed on the flow velocity sequence data to generate flow velocity trend correction factor data;

[0149] The section area sequence data is subjected to differential calibration to generate area correction factor data;

[0150] Based on the flow velocity trend correction factor data and the area correction factor data, dynamic deviation correction multiplication calculation is performed on the flow velocity and the section area to generate instantaneous flow data;

[0151] Short-term smoothing filtering is performed on the instantaneous flow data to obtain smoothed and corrected flow data.

[0152] In an embodiment, local trend analysis is performed on the collected flow velocity sequence data. The specific method is that, in the continuous flow velocity sequence, the average change rate and the standard deviation in each window are calculated in a sliding window (for example, 3-5 sampling points) as a unit, flow velocity trend correction factor data is generated for correcting local mutation or short-term abnormal fluctuation.

[0153] The section area sequence data is subjected to differential calibration, including correcting the small area deviation caused by water level change or measurement error, to generate area correction factor data. The differential calibration can fit a standard function of the section area changing with the water level based on the historical measurement data, and then correct the deviation between the real-time area data and the standard function.

[0154] The flow rate sequence and the cross-sectional area sequence are multiplied point by point by using the flow rate trend correction factor data and the area correction factor data to realize dynamic deviation correction and generate instantaneous flow data. Specifically, the instantaneous flow at each sampling point is equal to the flow rate at the moment multiplied by the cross-sectional area at the moment, and multiplied by the corresponding correction factors.

[0155] To eliminate short-time noise fluctuation, short-term smoothing filtering processing is performed on the instantaneous flow data, and the sliding average, weighted average or exponential smoothing method can be used to obtain the smoothed and corrected flow data, so as to ensure that the generated measured flow sequence not only reflects the real flow change, but also has high stability and reliability.

[0156] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the application file are intended to be included in the present application.

[0157] The above description is only a specific implementation of the present application, enabling those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent extension of water level-discharge relationship based on similar flood, characterized in that, The method comprises the following steps: Step S1: acquiring measured flow and water level process data; fitting the measured flow and water level process data according to measured water level points and flow points to obtain a fitted water level-flow relationship curve; Step S2: determining whether the fitted water level-flow relationship curve passes the three-property test, if yes, marking the fitted water level-flow relationship curve as an output curve; if no, performing multi-point segmented fitting on the fitted water level-flow relationship curve to obtain a secondarily fitted output curve; wherein the multi-point segmented fitting on the fitted water level-flow relationship curve in step S2 comprises: dividing the fitted water level-flow relationship curve into a plurality of continuous water level segments according to the slope change characteristics of the water level process, and taking the measured water level points and flow points in each segment as the characteristic point set of the segment; for any water level segment, obtaining a first difference characteristic of the water level segment according to the difference between the flow value of each characteristic point in the segment and the mean value of the flow values of all characteristic points in the segment; obtaining a second difference characteristic of the water level segment according to the difference between the water level slope of each characteristic point in the segment and the mean value of the water level slopes of all characteristic points in the segment; determining the segmented fitting curve of the water level segment according to the first difference characteristic and the second difference characteristic; connecting the segmented fitting curves of all water level segments in water level order to obtain the secondarily fitted output curve; wherein determining the segmented fitting curve of the water level segment according to the first difference characteristic and the second difference characteristic comprises: screening the maximum first difference characteristic point and the maximum second difference characteristic point in the water level segment and the remaining characteristic points according to the first difference characteristic and the second difference characteristic; establishing a segmented curve contour based on the maximum first difference characteristic point and the maximum second difference characteristic point; performing curve point fitting on the remaining characteristic points through the segmented curve contour to obtain the segmented fitting curve of the water level segment; Step S3: screening a similar curve matching the output curve from a preset historical flood pattern library, and extending the water level-flow process of the output curve through the similar curve to obtain an extended water level-flow relationship curve; Step S4: performing water level flow pushing on the collected real-time water level based on the extended water level-flow relationship curve.

2. The intelligent extension method of stage-discharge relation based on similar flood according to claim 1, characterized in that, The performing curve point fitting on the remaining characteristic points through the segmented curve contour comprises: determining the number of characteristic points of the remaining characteristic points, when the number of the remaining characteristic points is greater than a preset fitting standard number, merging adjacent remaining characteristic points so that the number of characteristic points is equal to the preset fitting standard number; when the number of the remaining characteristic points is less than the preset fitting standard number, inserting a preset compensation point based on the interval of adjacent remaining characteristic points so that the number of characteristic points is equal to the preset fitting standard number.

3. The intelligent extension method of stage-discharge relation based on similar flood according to claim 1, characterized in that, The determining whether the fitted water level-flow relationship curve passes the three-property test in step S2 specifically comprises a sign test, a curve fitting test and a deviation value test.

4. The intelligent extension method of stage-discharge relationship based on similar flood according to claim 1, characterized in that, Step S3 comprises the following steps: Step S31: taking the water level range of the output curve as an initial retrieval boundary, and screening a similar curve matching the output curve from a preset historical flood pattern library; Step S32: determining the extension section of the output curve and performing smoothing processing according to the curve line type of the similar curve, to obtain the smoothed curve extension section; Step S33: extending the water level-flow process of the output curve through the curve extension section, to obtain the extended water level-flow relationship curve.

5. The intelligent extension method of water level-discharge relation based on similar flood according to claim 4, characterized in that, Step S31 includes the following steps: Step S311: taking the water level range of the output curve as the initial search boundary; Step S312: screening the similar curve matching the output curve from the preset historical flood pattern library according to the initial search boundary, and if the matching is successful, directly outputting the similar curve; Step S313: if the matching fails, cyclically expanding the matching range by a preset step size until the similar curve is outputted by matching, and recording the expansion times and the corresponding similarity change data.

6. The intelligent extension method of stage-discharge relationship based on similar flood according to claim 4, characterized in that, After step S33, it further includes: Collecting the real-time water level; Confirming the water level range based on the real-time water level; If the maximum value or the minimum value of the water level range does not belong to the extended water level-flow relationship curve, repeating steps S31-S33.

7. The intelligent extension method of stage-discharge relationship based on similar flood as claimed in claim 1, wherein, The method for obtaining the measured water level points and flow points includes: Obtaining the station position data; based on the station position data, arranging the water level observation device, and collecting the water level data at a preset time interval to obtain the continuous measured water level sequence as the measured water level points; Obtaining the water level observation section position data; based on the corresponding section position of the water level observation section position data, arranging the flow measurement device, and calculating the flow value through the collected measured flow velocity and the section area to form the corresponding measured flow sequence as the flow points.

8. The intelligent extension method of stage-discharge relationship based on similar flood according to claim 1, characterized in that, The construction method of the preset historical flood pattern library includes: Obtaining the historical water level-flow curve data; Extracting the curve features of the historical water level-flow curve data to construct the curve feature vector library; Performing real-time data updating on the curve feature vector library to obtain the preset historical flood pattern library.

Images