A small watershed mountain flood disaster risk dynamic early warning method and system

Abstract

The application discloses a small watershed mountain flood disaster risk dynamic early warning method and system, relates to the technical field of hydrological monitoring and disaster early warning, and comprises the following steps: collecting structural acoustic vibration signals and water flow pulsation signals at the moment when the basic structure is impacted by flood, forming a spatial distribution atlas of energy disturbance in a short time, and constructing the spatial distribution atlas into an impact echo fingerprint strip for providing a reference basis of abnormal jump signals; based on the energy changes between the energy segments at the end of the impact echo fingerprint strip, identifying the synchronous mutation points and the continuous time length of adjacent segments, extracting the energy jump area, constructing a pseudo-peak suspicious point index, and marking the spatial positioning information of the abnormal jump signals. The application constructs an energy atlas by synchronously collecting acoustic vibration signals and water flow signals, identifies pseudo-peak abnormalities, reconstructs an energy turning path, combines time sequence anchor points to drive multi-stage intervention devices, realizes energy guidance and dissipation, effectively suppresses misjudgment, and improves the accuracy, stability and response reliability of mountain flood early warning.

Description

Technical Field

[0001] This invention relates to the field of hydrological monitoring and disaster early warning technology, specifically to a method and system for dynamic early warning of flash flood risks in small watersheds. Background Technology

[0002] Dynamic early warning for flash flood risks in small watersheds refers to the real-time analysis of risk trends during the incubation, development, and outbreak of flash floods in small watersheds characterized by large topographic relief, short runoff times, concentrated rainfall, and extremely rapid flood response. Driven by continuously changing monitoring information such as rainfall, water level, slope runoff, and soil moisture content, a warning judgment mechanism is established that automatically adjusts as monitoring data is updated. Its core meaning lies in the fact that traditional static early warning thresholds are insufficient to adapt to the suddenness and time-varying nature of flash floods in small watersheds. By introducing dynamic risk indicators, watershed response models, multi-source monitoring data fusion, and adaptive threshold correction, the risk levels at each stage before, during, and after flood formation are frequently updated. This shifts from reactive alerts to proactive predictions, enabling the early warning system to respond sensitively to disaster signs in a very short time, improving the timeliness, accuracy, and reliability of flash flood prediction in small watersheds, and ensuring the safety of people and engineering facilities in downstream areas.

[0003] The existing technology has the following shortcomings:

[0004] In existing technologies, flash flood early warning systems for small watersheds generally rely on real-time hydrological monitoring data as the basis for risk assessment. Sensors are typically fixedly installed on foundation structures such as bridge piers, support piles, and bank slope bases, continuously collecting parameters such as water level, flow velocity, and wave intensity to dynamically record the rise and fall of floodwaters. However, during the transient phase of high-speed flood impact on structures, the fluid impact on the structural surface, the reflection of internal cavities, and local eddy current disturbances may superimpose to form abnormal energy backflow zones. This causes short-term backflow interference in the signal link of the sensor equipment, resulting in a false peak jump in the output signal. This spurious peak signal appears as an instantaneous extreme point in the monitoring curve and returns to normal within a very short time, causing the dynamic early warning system to misjudge the flood trend at critical judgment points. Because these transient spurious peaks are sudden and highly uncertain, the early warning system may mistake them for a sharp increase in flood flow or a rapid rise in water level when reading these abnormal values. This can trigger an incorrect jump in the early warning level, causing emergency evacuation procedures, response mechanisms, and command chains to be activated prematurely. This can lead to disordered resource allocation and passive amplification of emergency response links, seriously affecting the robustness and operational rhythm of the overall early warning system.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for dynamic early warning of flash flood risks in small watersheds, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic early warning of flash flood risks in small watersheds, comprising the following steps:

[0008] The acoustic vibration signal and water flow pulsation signal of the basic structure are collected at the moment of the flood impact, and the spatial distribution spectrum of energy disturbance is formed in a short time. It is then constructed into an impact echo fingerprint strip containing high-energy segments and low-energy segments to provide a reference basis for abnormal jump signals.

[0009] Based on the energy changes between energy segments at the end of the impact echo fingerprint band, the synchronous abrupt change points and duration of adjacent segments are identified, energy jump regions are extracted, a pseudo-peak suspicious point index is constructed, and the spatial location information of abnormal jump signals is marked.

[0010] Based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, the spatial return path of the interference energy is reconstructed, and an energy return manifold trajectory diagram is generated to construct a continuous model of the interference energy propagation path;

[0011] By pulsating along the path at the end of the energy return manifold trajectory, the impact leader point and the attenuation tail are identified, the key time location segment where false peaks occur is determined, and a list of time-series anchor points is generated to calibrate the time reference of jump events.

[0012] Based on the key node information in the time-series anchor list, the breathing impedance window, the reverse adsorption curtain, and the return energy bypass are controlled to operate in a coordinated manner according to the time rhythm. By opening and closing the nodes in a time-division manner, the interference energy is guided to gradually pull, weaken, and dissipate along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of spurious peak signals.

[0013] Preferably, the steps for generating the impact echo fingerprint strip are as follows:

[0014] During the initial impact phase of the flood, the concrete foundation of the bank slope, the steel pile support node and the load-bearing platform under the bridge were selected as the installation locations of the basic structure. An integrated multi-channel induction array was deployed to simultaneously collect structural acoustic and vibration signals and water flow pulsation signals.

[0015] After acquiring the signal, the spatial differences between the multi-point acquired signals are used to calculate the local energy distribution of the structural surface during the impact process, forming an energy disturbance distribution map that includes the horizontal, vertical and time dimensions.

[0016] Based on the energy disturbance distribution map, the continuous region is divided into high-energy segments and low-energy segments according to the energy change cycle, and the impact response chain is constructed in time sequence.

[0017] By reconstructing the impact response chain with time as the main axis, an impact echo fingerprint band containing energy feature levels is formed, providing a reference basis for the identification of abnormal jump signals.

[0018] The preferred steps for constructing the pseudo-peak suspicious point index are as follows:

[0019] The terminal region of the impact echo fingerprint was selected as the detection focus. The amplitude difference, energy density change rate and duration between adjacent energy segments were analyzed to identify synchronous abrupt change points and determine energy jump characteristics.

[0020] The identified synchronous mutation points are clustered and merged, adjacent mutation segments are combined into energy mutation regions, and the time boundary, the position of the central segment and the energy peak point are recorded.

[0021] The energy jump region is mapped to the spatial positioning number of the sensing probe, and combined with the spatial distribution tag in the impact echo fingerprint strip, a spatial positioning structure of the jump signal is generated.

[0022] The spatial positioning structure of all energy jump regions is organized to form pseudo-peak suspicious point index data, so as to realize spatiotemporal marking and rapid retrieval of abnormal jump signals.

[0023] Preferably, the spatial positioning structure of the energy jump region is generated by multi-point joint positioning. The central node of each jump region is marked in coordinate form and a spatiotemporal composite index format is established in accordance with the time label to ensure the accurate correspondence and traceability of the pseudo-peak suspicious point index data in the time and spatial dimensions.

[0024] Preferably, the steps for generating the energy return manifold trajectory diagram are as follows:

[0025] Extract the endpoint fragment of the jump region from the pseudo-peak suspicious point index, identify the spatial coordinates and time labels of the energy jump, and extract the boundary parameters of the equipment installation body and cavity to establish an environmental model;

[0026] In the environmental model, the return path of the interference energy is constructed segment by segment according to the location of the jump point, and the reflection angle, path length and remaining energy density of each jump point are recorded to form a discrete jump point sequence;

[0027] In the sequence of jump points, determine the continuity between each jump point and select path segments that conform to the propagation law to form a directed energy return path;

[0028] All energy return paths are integrated and arranged in chronological order to generate an energy return manifold trajectory diagram to represent the continuous propagation model of disturbance energy.

[0029] Preferably, the steps for generating the time-series anchor list are as follows:

[0030] The end path segment of the energy return manifold trajectory is selected as the analysis area. The energy intensity, duration and propagation direction in the path segment are analyzed to identify short-term sudden energy pulsation phenomena and determine the pulsation time range.

[0031] Using the pulsation point as a clue, trace back along the direction of energy propagation to find the starting position where energy is first excited and continuously forms multi-hop propagation, and define it as the impact leader point;

[0032] The tail region where energy gradually decays and loses its propagation ability is defined as the decay tail section when the trailing part of the path continues to be tracked.

[0033] The time information and spatial coordinates of the impact leader, path pulsation segment and decay tail segment are integrated to form a list of time-series anchor points to calibrate the time reference of the jump event.

[0034] Preferably, the time information and spatial coordinates of the impact leader point, path pulsation segment and attenuation tail segment recorded in the time sequence anchor point list are arranged in time sequence to form a continuous time sequence chain. The time sequence chain is used to indicate the starting point, peak and ending point of energy propagation, and to provide accurate time parameters for the rhythmic control of subsequent interference energy.

[0035] Preferably, based on the key node information in the time-series anchor point list, the breathing impedance window, the anti-phase adsorption curtain, and the return energy bypass are controlled to open and close in a time-division manner according to the time rhythm, guiding the interference energy to be drawn, weakened, and dissipated segment by segment along the preset path. The steps are as follows:

[0036] Based on the key node information in the time-series anchor point list, high-probability jump points in the return energy propagation path are extracted and a time rhythm sequence is constructed.

[0037] A breathing impedance window is set at the beginning of the path and opened and closed in a high-frequency beat sequence according to a rhythmic sequence to reduce energy incidence.

[0038] An anti-phase adsorption curtain is installed in the middle of the energy propagation section, and energy wave cancellation is achieved through electronically controlled preheating and phase anti-adsorption.

[0039] A reversing energy bypass is set at the end of the energy propagation process, and the energy is dissipated step by step through a spiral guide groove and a compression chamber.

[0040] Based on the time rhythm sequence, the breathing impedance window, the reverse adsorption curtain, and the return energy bypass are operated in sequence to achieve the attraction, weakening, and dissipation of interference energy.

[0041] A dynamic early warning system for flash flood risks in small watersheds includes an energy disturbance fingerprint construction module, a pseudo-peak and suspicious point indexing module, a return energy path reconstruction module, a time-series anchor point generation module, and a rhythmic energy regulation module.

[0042] The energy disturbance fingerprint construction module collects structural acoustic vibration signals and water flow pulsation signals at the moment the basic structure is impacted by flood, and forms a spatial distribution map of energy disturbance in a short time. It then constructs an impact echo fingerprint band containing high-energy segments and low-energy segment sequences to provide a reference basis for abnormal jump signals.

[0043] The pseudo-peak suspicious point indexing module identifies synchronous abrupt change points and durations of adjacent segments based on energy changes between energy segments at the end of the impact echo fingerprint band, extracts energy jump regions, constructs a pseudo-peak suspicious point index, and marks the spatial location information of abnormal jump signals.

[0044] The energy return path reconstruction module reconstructs the spatial return path of interference energy based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, and generates an energy return manifold trajectory map, which is used to construct a continuous model of the interference energy propagation path.

[0045] The time-series anchor point generation module identifies the impact leader point and the attenuation tail segment along the path pulsation at the end of the energy return manifold trajectory diagram, determines the key time location segment where false peaks occur, and generates a time-series anchor point list to calibrate the time reference of jump events.

[0046] The rhythmic energy regulation module, based on the key node information in the time-series anchor point list, controls the breathing impedance window, the anti-phase adsorption curtain, and the return energy bypass to operate in a time rhythm. By opening and closing the nodes in a time-division manner, it guides the interference energy to be gradually pulled, weakened, and dissipated along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of spurious peak signals.

[0047] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0048] This invention achieves high-precision characterization of flood dynamic coupling behavior by simultaneously acquiring structural acoustic vibration and water flow pulsation signals at the moment of flood impact and establishing a spatial distribution map of energy disturbance. Through the construction of impact echo fingerprints and a pseudo-peak index, it can quickly identify and locate the formation area of ​​signal anomalies, avoiding misjudging interference peaks as measured flood extremes. Furthermore, by reconstructing the energy return manifold trajectory map and generating a time-series anchor point list, it achieves full-process tracking and time-layered control of the interference energy propagation process. Finally, through the rhythmic linkage of a breathing impedance window, an anti-phase adsorption curtain, and a return energy bypass, it guides energy jumps to a safe dissipation path, effectively weakening the conditions for pseudo-peak formation. This method enables time-division control of energy flow within the structure during the dynamic changes of floods, allowing interference signals to be gradually pulled, weakened, and dissipated, fundamentally avoiding false triggering of warning thresholds. Compared with traditional static judgment mechanisms, this technology significantly improves the anti-interference capability, data reliability, and response robustness of small watershed flash flood early warning systems, and achieves an integrated closed-loop prevention and control effect from signal identification, energy tracking to dynamic regulation, providing high-timeliness, high-accuracy, and high-reliability technical support for intelligent early warning of flash flood disasters. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0050] Figure 1 This is a flowchart of a method for dynamic early warning of flash flood disaster risk in small watersheds according to the present invention.

[0051] Figure 2 This is a schematic diagram of a module for a dynamic early warning system for flash flood disaster risks in small watersheds according to the present invention. Detailed Implementation

[0052] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0053] This invention provides, for example Figure 1 The method for dynamic early warning of flash flood risk in small watersheds, as shown, includes the following steps:

[0054] The acoustic vibration signal and water flow pulsation signal of the basic structure are collected at the moment of the flood impact, and the spatial distribution spectrum of energy disturbance is formed in a short time. It is then constructed into an impact echo fingerprint strip containing high-energy segments and low-energy segments to provide a reference basis for abnormal jump signals.

[0055] To effectively identify and intervene in anomalous signal jumps during flood impacts in small watersheds, it is necessary to accurately collect and analyze structural acoustic vibration and water flow pulsation information at the instant the flood impacts the foundation structure, and convert it into an energy spectrum with spatiotemporal directionality. This provides a clear reference for subsequent pseudo-peak identification and control. The specific implementation steps are as follows:

[0056] During the initial impact phase of a flood, typical foundation structure installation locations were selected, including the concrete foundation of the bank slope, steel pile support nodes, and the load-bearing platform under the bridge. Integrated multi-channel induction arrays were deployed at each location. Each array was equipped with at least a structural acoustic vibration response sensing unit and a multi-frequency water flow pulsation sampling probe. The signal sampling frequency was no less than 5000 Hz, and the response time was less than 0.01 seconds, ensuring the ability to capture sudden disturbances under the influence of floodwaters. The sensing units needed to be deployed perpendicular to the impact surface to reproduce the vertical hydrodynamic impact characteristics as closely as possible. Each induction array had to be started synchronously to ensure that multiple locations were collected on the same time axis, thus obtaining a complete set of instantaneous impact signals. Taking a flash flood caused by a localized rainstorm as an example, the floodwaters originated upstream and impacted the bank slope structure within approximately 10 seconds. At this time, the system continuously collected a steel bending wave signal with an upper amplitude limit of approximately 2.3 mm and a surge wave period with a frequency of 240 Hz, forming an initial disturbance signal packet within 2.6 seconds.

[0057] After acquiring the structural acoustic and vibration signals and water flow pulsation signals, the spatial differences between the multi-point acquired signals were used to calculate the local energy distribution of the structural surface during the impact process. By synchronously comparing the signal intensity changes and frequency expansion trends of each measuring point within the same time period, the temporal and spatial characteristics of energy disturbances at each location were extracted, restoring the overall hydrodynamic impact behavior to a spatiotemporal energy mapping. For example, the induction probe below the bank slope foundation recorded three wave peaks within 6 consecutive seconds, each lasting 0.4 seconds, with an average amplitude 1.8 times higher than that of the adjacent support pile, indicating that the impact generated a high energy concentration area at that location. Based on such signal differences, spatial interpolation was used to generate a disturbance distribution map containing lateral, longitudinal, and temporal dimensions. The map clearly shows the strength and duration of energy excitation in different regions, laying the foundation for subsequent segmentation.

[0058] Based on the perturbation distribution map, continuous energy change regions are divided into different segments according to temporal sequence and amplitude stability. The segment length is determined based on the energy change period, with a length range of 0.2 to 2.5 seconds. Continuous energy segments with energy values ​​more than 150% above the regional average are classified as high-energy segments, and segments with energy values ​​less than 50% below the average are classified as low-energy segments. Taking a single acquisition result as an example, 22 energy segments were extracted within a 10-second sampling time, including 6 high-energy segments, 10 low-energy segments, and the rest being neutral segments. After segmentation, all segments are concatenated in chronological order to construct a complete impact response chain. To further improve structure and readability, transition points between segments are retained for subsequent identification of anomalous jump points, avoiding misidentification or omission of energy change regions due to segmentation.

[0059] Based on the segment sequence, a time-based reconstruction is performed to form an impact echo fingerprint band containing energy characteristic layers. This fingerprint band not only reflects the actual energy response of the structure under flood impact but also provides a spatial correspondence between concentrated and sparse energy distribution, possessing high identifiability and directionality. By comprehensively labeling the energy intensity, duration, and differences between adjacent segments for each segment, the fingerprint band can be used as a reference for locating suspected false peaks. For example, in a certain event, segments 14 to 16 of the fingerprint band showed a high-intensity jump, with amplitude fluctuations exceeding three times that of the preceding segments. Retrospective analysis confirmed that this segment was the starting point of the abnormal signal surge. The constructed impact echo fingerprint band is ultimately presented in a time-series structure and compared with the original data to ensure the integrity of the fingerprint sequence, the absence of overlap or omissions, and the traceability and stability characteristics required for subsequent interference identification.

[0060] Based on the energy changes between energy segments at the end of the impact echo fingerprint band, the synchronous abrupt change points and duration of adjacent segments are identified, energy jump regions are extracted, a pseudo-peak suspicious point index is constructed, and the spatial location information of abnormal jump signals is marked.

[0061] After constructing the impact echo fingerprint, it is necessary to further explore the subtle changes in the energy segments at the ends of the fingerprint to identify the starting points and regions of energy jumps, and extract sensitive areas that may generate spurious peak signals. This lays the foundation for subsequent interference path reconstruction and rhythmic control. The specific implementation steps are as follows:

[0062] The terminal region of the impact echo fingerprint strip is selected as the focus of detection. The terminal region is defined as the set of the most recently formed continuous energy segments in the fingerprint strip. The number of segments is proportional to the total sampling volume of the previous stage, typically set to the last 20% of the segment segments. This part focuses on analyzing the amplitude difference, energy density change rate, and duration extension between each pair of adjacent energy segments. The signal difference curve between each pair of segments is used to determine whether abrupt changes occur. For example, in a flood impact signal, the terminal region contains a total of 8 pairs of segments. Among them, 3 pairs of segments exhibit energy transitions within 0.4 seconds, with amplitude differences exceeding 2.2 times and density changes exceeding a slope of 8 joules per second. These can be preliminarily considered as candidate regions for synchronous abrupt changes. Simultaneously, to avoid misjudgments caused by occasional jumps, duration is introduced as an evaluation factor. If a jump segment continuously maintains abnormal amplitude for 0.1 seconds, the jump is confirmed to have true abrupt change characteristics.

[0063] Identified synchronization abrupt change points are clustered and merged to group multiple consecutive or adjacent abrupt change segments into a single energy abrupt change region, ensuring the region's integrity, clear boundaries, and temporal consistency. During the merging process, energy segments extending 0.05 seconds before and after the abrupt change point are preserved, with the time axis as the core, to avoid mistakenly deleting nearby high-frequency disturbance sources. For example, in a 10-second flood impact sampling, four pairs of abrupt change pairs were found in the terminal region, two of which had an interval of less than 0.15 seconds. These were merged into a 0.7-second abrupt change region with an average amplitude 1.9 times that of the preceding segment, and an energy release density reaching 9.6 joules / second in a local area, exhibiting significant anomalous characteristics. Based on this, the temporal boundaries, central segment location, and energy peak point of the abrupt change region are accurately recorded, preparing for subsequent spatial mapping.

[0064] The sudden jump regions are mapped back to the corresponding sensor probe's location number in physical space. Combined with the spatial distribution labels of each segment in the impact echo fingerprint strip constructed in the previous stage, multi-point joint localization of the jump regions is performed to generate a spatial localization structure for the jump signal. This structure not only marks the time information of the abnormal signal occurrence but also indicates its specific propagation path and distribution nodes in three-dimensional space. For example, segments 25 to 27 in the fingerprint strip are grouped into one jump region. Mapping reveals that this region involves three probe points on the northwest side of the bank slope, located at an elevation of 3.4 meters, a lateral distance of 9.2 meters, and a longitudinal distance of 6.7 meters, respectively. This spatial region is located directly below the water flow reversal channel, posing a high risk of disturbance. In the spatial localization structure, the central node of each jump region is marked in coordinate form and corresponds one-to-one with the time label of the jump region, forming a spatiotemporal composite index format.

[0065] All spatial location structures of sudden jump areas were centrally organized to generate a set of pseudo-peak suspicious point index data. This index uses event number as the primary key and includes detailed indicators such as time label, spatial coordinates, segment number, jump intensity, amplitude slope, duration, and the difference between the mean of preceding and following segments, all packaged and stored in tabular format. This index possesses high temporal resolution and spatial tracking capabilities, enabling rapid retrieval of pseudo-peak generation areas, repeated verification of abnormal segments, and targeted control in subsequent intervention stages. In a typical flash flood monitoring case, this index generated five sets of pseudo-peak suspicious point records, concentrated between the 7th and 8th seconds in time and spatially concentrated at structural boundary locations, which were prioritized for extraction during subsequent energy return trajectory reconstruction. Through the precise indexing of this index, the entire anomaly detection process possesses a clear logical chain and high executability, providing accurate input data for dynamic control.

[0066] Based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, the spatial return path of the interference energy is reconstructed, and an energy return manifold trajectory diagram is generated to construct a continuous model of the interference energy propagation path;

[0067] To accurately trace and reconstruct the return path of interference energy within the structure, starting from identified spurious peaks, it is necessary to combine the location of the sudden signal, the structural details of the equipment installation, and the boundary construction of the cavity it is attached to. This requires segment-by-segment path analysis at the physical level, ultimately forming a continuous and complete energy propagation trajectory diagram. The specific implementation steps are as follows:

[0068] The endpoint segments of each abrupt jump region are extracted from the pseudo-peak suspicious point index to identify the spatial coordinates and time labels of the energy jump. This jump point represents the end node of the abnormal energy release and also the possible starting source of energy back propagation. To accurately define the location of this point within the structure, the specific installation parameters of the corresponding sensor probe need to be retrieved, including the geometry, material properties, positioning method, and positional relationship relative to the structural boundary of the base to which it is attached. Taking the embedded installation at the bottom of the support pile as an example, if the probe is installed on the outside of a reinforced concrete column, and the column is a hollow cylinder with a diameter of 120 mm and a wall thickness of 20 mm, then its boundary conditions must include the changes in the angle between the internal cavity space, the wall thickness, and the outer annular structure. Furthermore, the incident angle information of the water flow impact direction needs to be obtained. For example, when a jump signal occurs, the flood impacts the structure at an average speed of 5.2 m / s and an angle of 38 degrees, and the resulting energy propagation direction has a clear vector directionality. The extraction of this series of physical parameters constitutes the basic environmental model required for path tracing.

[0069] In the environmental model, possible return paths of interference energy are established around the jump point. This path is not a simple linear propagation, but rather a multi-fold, variable, and bend-characteristic reflection route formed by the complex internal morphology of the structure. Taking a concrete base with a trapezoidal cross-section and flow-guiding grooves at the edges of a cavity as an example, the energy return may bounce off the lower edge slope to the middle and then be reflected by the top bend, propagating to the release port at the top of the structure after three bounces. During each path construction process, the propagation direction, point of action, reflection angle, and loss rate of the energy wave need to be calculated. For example, after the water flow impacts the bottom slope of the base at a 45-degree angle, it will form the first jump point, with a return path direction of 16 degrees southeast and a path length of 0.7 meters. Subsequently, the second jump point occurs at the corner of the groove, with a reflection angle of 31 degrees and a path length of 1.2 meters, and the energy density is reduced to 68% of the original value. The third jump point is located in the left bend area at the top of the cavity; after analysis, the reflection direction deviates from the initial direction by 26 degrees, and the path length is 0.9 meters. During this process, all path segments are recorded in three-dimensional coordinates to form a discrete sequence of jump points. The location information, reflection angle, path length and remaining energy density of each jump point are fully marked to ensure the integrity of subsequent connections.

[0070] Based on the sequence of jump points, the continuity between each jump point is determined, and jump segments with inconsistent directions, excessively long path distances, or excessively rapid energy attenuation are eliminated, retaining only path combinations that conform to the laws of energy propagation. The continuity judgment is based on the following criteria: the distance between jump points does not exceed 1.5 meters, the angle between the incident direction and the reflection direction of the previous jump segment is less than 20 degrees, and the remaining energy density is not less than 40% of the previous jump segment. If a jump segment does not meet the above conditions with the previous jump segment, it is considered a path interruption point and will not be extended further. Connections are established between jump points that meet the conditions, and energy propagation chains are generated sequentially. Taking an energy jump event as an example, starting from the jump point at 8.6 seconds, propagation occurs through three consecutive jump points, with a total path length of 3.2 meters, a rebound angle variation range of 13 degrees, and an average energy density attenuation rate of 12% per meter. These jump segments are combined to form a directed energy return path, each segment being associated with a direction vector, physical length, structural medium type, and reflective surface material properties, constituting a complete spatial propagation chain.

[0071] All completed energy propagation chains are integrated and arranged chronologically to generate an energy return manifold trajectory diagram. This diagram is based on a three-dimensional spatial structure, with time as the horizontal axis and spatial coordinates as the vertical axis. Energy levels are distinguished by color intensity, and the directionality, speed changes, and concentration areas of energy propagation are fully represented through different colors, line thicknesses, and node labels. Each jump segment in the trajectory diagram clearly marks its start and end points, reflection angle, and path length. Jump points are represented by dots, and the color filled within the dots uses a gradient from red to blue according to different energy intensities. Taking a typical impact process as an example, the flood impacts the bottom of the structure at 8.3 seconds, generating energy. The return path propagates continuously along the inner wall of the structure to the top drainage hole, forming five jump structures involving six energy nodes, which are all dissipated at 9.1 seconds. After the trajectory map is completed, it is compared one-to-one with the previously identified abrupt jump signals in time and space to verify whether the starting point of each energy propagation path coincides with the suspected spurious peak. This determines the contribution of the return path to the formation of the spurious peak and provides high-precision spatial-temporal input for the next stage of key node identification and dynamic energy control. As a visual representation of the entire energy propagation process, the trajectory map provides detailed and actionable technical support for subsequent construction of intervention channels, regulation of device operation rhythm, and location of potential spurious peak high-incidence points.

[0072] By pulsating along the path at the end of the energy return manifold trajectory, the impact leader point and the attenuation tail are identified, the key time location segment where false peaks occur is determined, and a list of time-series anchor points is generated to calibrate the time reference of jump events.

[0073] After constructing the energy return manifold trajectory map, it is necessary to finely identify the propagation pulsations of its terminal path segment. By clearly defining the energy excitation starting point and the final decay region, time segments with sudden jump characteristics are extracted, and these key moments are compiled into a time-series anchor point list, serving as the basis for time control of subsequent dynamic interventions. The specific implementation steps are as follows:

[0074] The analysis region is selected from the final path segment of the energy return manifold trajectory. This segment typically concentrates in the last 0.5 to 1 second of energy propagation, a region where energy density gradually decreases but may still generate spurious peak signals. Within this segment, the energy intensity, duration, and propagation direction of each jump segment are analyzed individually to identify short-term, sudden energy pulsations. These pulsations manifest as a sudden increase in the energy amplitude of a particular jump segment or combination of jump segments within an overall energy decay trend, forming a fluctuation pattern significantly different from the preceding and following segments. For example, between 9.45 and 9.61 seconds, the seventh and eighth jumps in the trajectory diagram show rapid changes in energy peak values. The amplitude of the seventh jump jumps abruptly from 0.8 mm to 2.6 mm in just 0.07 seconds, while the eighth jump maintains an amplitude of 2.4 mm for 0.11 seconds, far exceeding the previous average of 1.1 mm. These short-period, high-intensity fluctuations indicate a certain energy concentration trend within the path, potentially forming sudden jump signals, and should be used as a reference for subsequent critical time periods. By locating these sudden pulsations, their initial, peak, and end times are marked on the timeline, providing input information for subsequent path structure analysis.

[0075] Using the pulsation point as a clue, the path segment is traced backward along the energy propagation direction to find the starting position where energy is first excited and continuously forms multiple pulsations, which is defined as the impact leader point. This point is the first spur segment in the path with continuous excitation capability, which is usually characterized by: the energy density first breaking through the previous stable level, followed by at least two spur segments forming continuously within 0.3 to 0.8 seconds, each with a consistent propagation direction and no rapid decrease in energy density. Taking one analysis result as an example, between 8.48 seconds and 8.76 seconds, the third spur segment experienced an increase in energy amplitude, reaching 2.9 times the previous average value, and three energy spurs with consistent directions and reflection angles between 32 and 38 degrees were formed continuously within 0.28 seconds. Its path direction is propagated upward along the inclined surface of the inner wall of the structure, which is consistent with the characteristics of continuous excitation. Therefore, 8.48 seconds is taken as the impact leader point of this path segment. Subsequently, using this starting point as a reference, it is extended backward to the terminal pulsation segment identified in the previous step to form a complete potential pseudo-peak excitation time segment.

[0076] At the rear of the potential spurious peak excitation time range, the attenuation of energy propagation in the path is further tracked. The region where energy density gradually dissipates and the jump amplitude shrinks to a point where propagation is impossible is defined as the attenuation tail segment. The attenuation tail segment typically appears at the end of structural cavities, energy-absorbing nodes, or the edge of the water flow re-entering the mainstream. The attenuation characteristics are: the last jump segment lasts less than 0.08 seconds, the amplitude is no higher than 30% of the average amplitude of the preceding segment, and no effective rebound jump is formed within 0.1 seconds afterward. Between 9.86 and 9.94 seconds, the tenth jump segment in the energy propagation path has an amplitude of 0.3 mm, lower than 25% of the average amplitude of the previous five segments, a duration of only 0.06 seconds, a propagation direction deviating 37 degrees from the original path, and fails to excite the next jump segment; this is identified as the attenuation tail segment. This, together with the aforementioned impact leader point, forms a complete time loop used to calibrate the time range of the sudden jump event within the path structure.

[0077] Information on the identified impact leader, path pulsation segment, and decay tail segment in each path segment was integrated to create a high-resolution temporal anchor point list. The list is arranged chronologically, with each row containing an anchor point event. Records include a timestamp, jump segment number, maximum amplitude, duration, propagation direction, energy density change rate, start and end node coordinates, and the corresponding path number. In one analysis task, a total of 12 anchor point locations were identified, including 4 impact leader points, 5 pulsation center points, and 3 energy decay tail segments. Taking a pulsation point at 9.55 seconds as an example, its timestamp is 9.55 seconds, corresponding to the seventh jump segment, with a peak amplitude of 2.4 mm, a duration of 0.11 seconds, a propagation direction of 21 degrees north of northeast, an energy change rate increasing by 0.33 mm every 0.01 seconds, and the coordinates of the previous jump point being 3.4 meters horizontally, 7.6 meters vertically, and 2.1 meters above sea level. This list can directly provide key time points for time-sharing intervention in subsequent control processes, and can be used to deduce the temporal evolution trend of spurious peak events, improving the accuracy of jump identification and intervention timing. Through the comprehensive construction of this anchor point list, jump events are no longer random abnormal signals, but rather ordered evolutionary processes with three-dimensional characteristics of time, space, and energy, providing a predictable, intervention-friendly, and controllable technological foundation.

[0078] Based on the key node information in the time-series anchor list, the breathing impedance window, the reverse adsorption curtain, and the return energy bypass are controlled to operate in a coordinated manner according to the time rhythm. By opening and closing the nodes in a time-division manner, the interference energy is guided to be gradually pulled, weakened, and dissipated along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of the spurious peak signal.

[0079] After constructing the time-series anchor point list, it is necessary to sequentially drive the breathing impedance window, the anti-phase adsorption curtain, and the foldback energy bypass according to the distribution characteristics of each key time node listed in the list. This coordinated operation must be completed according to a strict rhythmic sequence to achieve dynamic intervention of interference energy from initial guidance, mid-stage attenuation, to final dissipation, effectively suppressing the formation of spurious peak signals. The specific implementation steps are as follows:

[0080] Based on the time stamps, spatial coordinates, energy peaks, and propagation directions recorded in the time-series anchor point list, key jump points with high trigger probability are extracted from each return energy propagation path, constructing multiple time rhythm sequences. Each rhythm sequence corresponds to an energy propagation path and includes three elements: initial control time, rhythm triggering period, and spatial propagation direction. For example, in the path from the northeast wall to the top bend angle, the jump times of the eighth to tenth segments are 9.44 seconds, 9.56 seconds, and 9.69 seconds, respectively. The average propagation period between each segment is 0.12 seconds, the propagation direction remains between 22 degrees east of north and 26 degrees east of north, and the energy amplitude increases from 2.3 mm to 2.8 mm and then gradually decreases to 1.7 mm. This path is identified as a high-energy jump chain, requiring full-process rhythmic intervention. These time rhythms are converted into standardized intervention parameters to provide precise time and space references for subsequent component opening and closing control.

[0081] At the structural entrance corresponding to the initial segment of the path, breathing impedance windows are pre-installed, mainly distributed at the side gaps of the support piles, the front end of the guide ramp in the middle of the structure, the bottom of the downstream cavity inner wall, and both sides of the inner cavity entrance platform. The breathing impedance window consists of three parts: a waveform elastic diaphragm, a servo hydraulic limit rail, and a fast-opening and closing solenoid valve assembly. During execution, it can complete high-frequency beat-like opening and closing according to a preset rhythm sequence. For example, 50 milliseconds before the northeast wall impact node identified at 9.43 seconds, the impedance window is pre-activated, reaching a fully open state at 9.425 seconds, remaining open for 0.08 seconds before automatically closing, and reopening at 9.505 seconds to complete the second round of blocking. Each breathing impedance window's action sequence consists of no less than three rounds, with each rhythmic cycle strictly synchronized with the corresponding rhythmic sequence to ensure precise weakening of energy in the initial stage of incidence. Through this process, the kinetic energy wave along the inner wall of the structure is cut off or slowed down in the early propagation stage, avoiding continuous backlash.

[0082] After entering the mid-path of energy propagation, anti-phase adsorption curtains are deployed at the mid-corner of the structure, the side cavity area of ​​the top cavity, the refractive inclined surface of the inner wall, and the narrow transition slit of the cavity. The anti-phase adsorption curtains consist of a three-layer composite structure, including a surface conductive glass fiber energy-absorbing film, a middle thermally responsive deformation layer, and a bottom fixed micro-wave enhancement unit. 100 milliseconds before the target node is activated, the adsorption curtains are preheated electronically, putting the middle deformation layer in a semi-activated state. When the anchor point is triggered, the energy-absorbing film locally forms an adsorption interface opposite to the direction of the energy wave. Taking the bending angle at 9.56 seconds as an example, this point is the area with the highest energy concentration in the mid-path. After the anti-phase adsorption curtain is activated, it completes the phase cancellation process with the energy wave within 0.05 seconds, reducing its effective propagation length by approximately 1.6 meters, with an equivalent energy attenuation of approximately 3.2 joules. Each adsorption curtain is deployed in layers with spatially offset angles to ensure that the reflected energy from different paths can effectively respond under different reflection angles, avoiding interference omissions caused by path migration.

[0083] A reversing energy bypass is deployed at the end of the energy propagation path. Its structure consists of a bent spiral guide channel and a wide-body slow-release channel. The installation location covers the top drainage seam, the confluence point at the cavity corner, and the finishing groove area at the edge of the structure. The energy bypass channel internally comprises three continuously variable diameter compression chambers, two sets of spiral baffle guides, and an arc-shaped energy release plate at the end. It has the ability to guide high-energy fluid to flow in a non-mainstream direction and dissipate it step by step. When the energy peak at 9.69 seconds is detected at the end of the path and its propagation direction is concentrated at a 45-degree angle to the right of the top of the structure, the reversing energy bypass opens 0.07 seconds in advance, forming a guide inlet. Energy is received within the next 0.15 seconds and guided into the compression chamber along the spiral guide. Inside the compression chamber, the high-energy fluid forms a rotating vortex structure within 0.08 seconds, with a kinetic energy loss rate of 65%. It then diffuses into the slow-release zone through the arc-shaped release plate, preventing the energy flow from completing a closed path and thus blocking the subsequent propagation chain.

[0084] Based on the complete rhythm sequence, the breathing impedance window, the reverse adsorption curtain, and the reversing energy bypass are executed sequentially at their respective action periods and spatial nodes, ensuring temporal overlap between the three components to form a complete intervention loop from the initial stage of energy excitation to the final stage of dissipation. For example, during the interference cycle from 9.43 seconds to 9.88 seconds, the impedance window completes three rounds of opening and closing cycles at the beginning of the path, with an average opening time of 0.08 seconds and a closing time of 0.06 seconds per round; the adsorption curtain responds concentratedly to the seventh and eighth jump segments of the path in the middle section, with its activation time covering from 9.53 seconds to 9.58 seconds; the energy bypass remains open for 0.14 seconds after 9.69 seconds and completes the high-kinetic energy discharge at the end. The total time span of the three working together is 0.45 seconds, effectively suppressing more than 90% of the energy accumulation in the path and disrupting the energy connectivity of multiple key excitation points in the reversing chain. Ultimately, the interference energy was gradually drawn in, weakened, and completely dissipated, preventing the formation of false peak signals at the source and ensuring that flash flood warnings in small watersheds still have high reliability and stability under complex hydrodynamic conditions.

[0085] This invention achieves high-precision characterization of flood dynamic coupling behavior by simultaneously acquiring structural acoustic vibration and water flow pulsation signals at the moment of flood impact and establishing a spatial distribution map of energy disturbance. Through the construction of impact echo fingerprints and a pseudo-peak index, it can quickly identify and locate the formation area of ​​signal anomalies, avoiding misjudging interference peaks as measured flood extremes. Furthermore, by reconstructing the energy return manifold trajectory map and generating a time-series anchor point list, it achieves full-process tracking and time-layered control of the interference energy propagation process. Finally, through the rhythmic linkage of a breathing impedance window, an anti-phase adsorption curtain, and a return energy bypass, it guides energy jumps to a safe dissipation path, effectively weakening the conditions for pseudo-peak formation. This method enables time-division control of energy flow within the structure during the dynamic changes of floods, allowing interference signals to be gradually pulled, weakened, and dissipated, fundamentally avoiding false triggering of warning thresholds. Compared with traditional static judgment mechanisms, this technology significantly improves the anti-interference capability, data reliability, and response robustness of small watershed flash flood early warning systems, and achieves an integrated closed-loop prevention and control effect from signal identification, energy tracking to dynamic regulation, providing high-timeliness, high-accuracy, and high-reliability technical support for intelligent early warning of flash flood disasters.

[0086] This invention provides, for example Figure 2 The system shown is a dynamic early warning system for flash flood risks in small watersheds, comprising an energy disturbance fingerprint construction module, a pseudo-peak and suspicious point indexing module, a return energy path reconstruction module, a time-series anchor point generation module, and a rhythmic energy regulation module.

[0087] The energy disturbance fingerprint construction module collects structural acoustic vibration signals and water flow pulsation signals at the moment the basic structure is impacted by flood, and forms a spatial distribution map of energy disturbance in a short time. It then constructs an impact echo fingerprint band containing high-energy segments and low-energy segment sequences to provide a reference basis for abnormal jump signals.

[0088] The pseudo-peak suspicious point indexing module identifies synchronous abrupt change points and durations of adjacent segments based on energy changes between energy segments at the end of the impact echo fingerprint band, extracts energy jump regions, constructs a pseudo-peak suspicious point index, and marks the spatial location information of abnormal jump signals.

[0089] The energy return path reconstruction module reconstructs the spatial return path of interference energy based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, and generates an energy return manifold trajectory map, which is used to construct a continuous model of the interference energy propagation path.

[0090] The time-series anchor point generation module identifies the impact leader point and the attenuation tail segment along the path pulsation at the end of the energy return manifold trajectory diagram, determines the key time location segment where false peaks occur, and generates a time-series anchor point list to calibrate the time reference of jump events.

[0091] The rhythmic energy regulation module, based on the key node information in the time-series anchor point list, controls the breathing impedance window, the anti-phase adsorption curtain, and the return energy bypass to operate in a time rhythm. By opening and closing the nodes in a time-division manner, it guides the interference energy to be gradually pulled, weakened, and dissipated along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of spurious peak signals.

[0092] The present invention provides a method for dynamic early warning of flash flood risk in small watersheds, which is implemented through the aforementioned dynamic early warning system for flash flood risk in small watersheds. For details of the specific method and process of the dynamic early warning system for flash flood risk in small watersheds, please refer to the aforementioned embodiment of the method for dynamic early warning of flash flood risk in small watersheds, which will not be repeated here.

[0093] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for dynamic early warning of flash flood disaster risk in small watersheds, characterized in that, Includes the following steps: The structural acoustic vibration signal and water flow pulsation signal were collected at the moment when the basic structure was hit by the flood, and the spatial distribution map of energy disturbance was formed in a short time and constructed as an impact echo fingerprint. Based on the energy changes between energy segments at the end of the impact echo fingerprint band, the synchronous abrupt change points and duration of adjacent segments are identified, energy jump regions are extracted, a pseudo-peak suspicious point index is constructed, and the spatial location information of abnormal jump signals is marked. Based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, the spatial return path of the interference energy is reconstructed, and an energy return manifold trajectory diagram is generated. By pulsating along the path at the end of the energy return manifold trajectory, the impact leader point and the attenuation tail segment are identified, the key time location segment where false peaks occur is determined, and a list of time-series anchor points is generated. Based on the key node information in the time-series anchor list, the breathing impedance window, the reverse adsorption curtain, and the return energy bypass are controlled to operate in a coordinated manner according to the time rhythm. By opening and closing the nodes in a time-division manner, the interference energy is guided to gradually pull, weaken, and dissipate along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of spurious peak signals.

2. The method for dynamic early warning of flash flood risk in small watersheds according to claim 1, characterized in that, The steps for generating an impact echo fingerprint strip are as follows: During the initial impact phase of the flood, the concrete foundation of the bank slope, the steel pile support node and the load-bearing platform under the bridge were selected as the installation locations of the basic structure. An integrated multi-channel induction array was deployed to simultaneously collect structural acoustic and vibration signals and water flow pulsation signals. After acquiring the signal, the spatial differences between the multi-point acquired signals are used to calculate the local energy distribution of the structural surface during the impact process, forming an energy disturbance distribution map; Based on the energy disturbance distribution map, the continuous region is divided into high-energy segments and low-energy segments according to the energy change cycle, and the impact response chain is constructed in time sequence. The impact response chain is reconstructed with time as the main axis to form an impact echo fingerprint band, providing a reference basis for the identification of abnormal jump signals.

3. The method for dynamic early warning of flash flood risk in small watersheds according to claim 2, characterized in that, The steps for constructing the pseudo-peak suspicious point index are as follows: The terminal region of the impact echo fingerprint was selected as the detection focus. The amplitude difference, energy density change rate and duration between adjacent energy segments were analyzed to identify synchronous abrupt change points and determine energy jump characteristics. The identified synchronous mutation points are clustered and merged, adjacent jump segments are combined into energy jump regions, and the time boundary, the position of the central segment and the energy peak point are recorded. The energy jump region is mapped to the spatial positioning number of the sensing probe, and combined with the spatial distribution tag in the impact echo fingerprint strip, a spatial positioning structure of the jump signal is generated. The spatial positioning structure of all energy jump regions is organized to form a pseudo-peak suspicious point index data.

4. The method for dynamic early warning of flash flood risk in small watersheds according to claim 3, characterized in that, The spatial positioning structure of the energy jump region is generated through multi-point joint positioning. The central node of each jump region is marked in coordinate form and a spatiotemporal composite index format is established in accordance with the time label to ensure the accurate correspondence and traceability of the pseudo-peak suspicious point index data in the time and spatial dimensions.

5. A method for dynamic early warning of flash flood risk in small watersheds according to claim 3, characterized in that, The steps for generating the energy return manifold trajectory diagram are as follows: Extract the endpoint fragment of the jump region from the pseudo-peak suspicious point index, identify the spatial coordinates and time labels of the energy jump, and extract the boundary parameters of the equipment installation body and cavity to establish an environmental model; In the environmental model, the return path of the interference energy is constructed segment by segment according to the location of the jump point, and the reflection angle, path length and remaining energy density of each jump point are recorded to form a discrete jump point sequence; In the sequence of jump points, determine the continuity between each jump point and select path segments that conform to the propagation law to form a directed energy return path; All energy return paths are integrated and arranged in chronological order to generate an energy return manifold trajectory diagram to represent the continuous propagation model of disturbance energy.

6. A method for dynamic early warning of flash flood disaster risk in small watersheds according to claim 5, characterized in that, The steps for generating the time-series anchor list are as follows: The end path segment of the energy return manifold trajectory is selected as the analysis area. The energy intensity, duration and propagation direction in the path segment are analyzed to identify short-term sudden energy pulsation phenomena and determine the pulsation time range. Using the pulsation point as a clue, trace back along the direction of energy propagation to find the starting position where energy is first excited and continuously forms multi-hop propagation, and define it as the impact leader point; The tail region where energy gradually decays and loses its propagation ability is defined as the decay tail section when the trailing part of the path continues to be tracked. The time information and spatial coordinates of the impact leader, path pulsation segment and decay tail segment are integrated to form a list of time-series anchor points to calibrate the time reference of the jump event.

7. A method for dynamic early warning of flash flood risk in small watersheds according to claim 6, characterized in that, The time information and spatial coordinates of the impact leader point, path pulsation segment and attenuation tail segment recorded in the time sequence anchor point list are arranged in time sequence to form a continuous time sequence chain. The time sequence chain is used to indicate the starting point, peak and ending point of energy propagation, and to provide precise time parameters for the rhythm control of subsequent disturbance energy.

8. A method for dynamic early warning of flash flood risk in small watersheds according to claim 6, characterized in that, Based on the key node information in the time-series anchor point list, the breathing impedance window, the anti-phase adsorption curtain, and the return energy bypass are controlled to open and close in a time-division manner according to the time rhythm, guiding the interference energy to be drawn, weakened, and dissipated segment by segment along the preset path. The steps are as follows: Based on the key node information in the time-series anchor point list, high-probability jump points in the return energy propagation path are extracted and a time rhythm sequence is constructed. A breathing impedance window is set at the beginning of the path and opened and closed in a high-frequency beat sequence according to a rhythmic sequence to reduce energy incidence. An anti-phase adsorption curtain is installed in the middle of the energy propagation section, and energy wave cancellation is achieved through electronically controlled preheating and phase anti-adsorption. A reversing energy bypass is set at the end of the energy propagation process, and the energy is dissipated step by step through a spiral guide groove and a compression chamber. Based on the time rhythm sequence, the breathing impedance window, the reverse adsorption curtain, and the return energy bypass are operated in sequence to achieve the attraction, weakening, and dissipation of interference energy.

9. A dynamic early warning system for flash flood disaster risks in small watersheds, used to implement the dynamic early warning method for flash flood disaster risks in small watersheds as described in any one of claims 1-8, characterized in that, It includes an energy perturbation fingerprint construction module, a pseudo-peak indexing module, a return energy path reconstruction module, a time-series anchor point generation module, and a rhythmic energy regulation module: The energy disturbance fingerprint construction module collects structural acoustic vibration signals and water flow pulsation signals at the moment when the basic structure is impacted by flood, and forms a spatial distribution map of energy disturbance in a short time, and constructs it as an impact echo fingerprint band. The pseudo-peak suspicious point indexing module identifies synchronous abrupt change points and durations of adjacent segments based on energy changes between energy segments at the end of the impact echo fingerprint band, extracts energy jump regions, constructs a pseudo-peak suspicious point index, and marks the spatial location information of abnormal jump signals. The energy return path reconstruction module reconstructs the spatial return path of the interference energy based on the sudden jump position at the end of the pseudo-peak suspicious point index, combined with the boundary parameters of the equipment installation body and cavity, and generates an energy return manifold trajectory diagram. The time-series anchor point generation module identifies the impact leader point and the attenuation tail segment along the path pulsation at the end of the energy return manifold trajectory diagram, determines the key time location segment where false peaks occur, and generates a time-series anchor point list. The rhythmic energy regulation module, based on the key node information in the time-series anchor point list, controls the breathing impedance window, the anti-phase adsorption curtain, and the return energy bypass to operate in a time rhythm. By opening and closing the nodes in a time-division manner, it guides the interference energy to be gradually pulled, weakened, and dissipated along a predetermined rhythm, thereby completing the dynamic intervention and risk suppression of spurious peak signals.

Images