Multi-source data fusion water conservancy project construction monitoring data supervision system and method

Abstract

This invention discloses a multi-source data fusion system and method for monitoring and supervising water conservancy engineering construction data, relating to the field of engineering construction technology. The system and method include the following steps: S1, acquiring water conservancy construction monitoring data and preprocessing the data; S2, constructing a sliding window, extracting key structural and environmental variables, comprehensively analyzing the degree of structural disturbance, and dividing different disturbance state segments according to the degree of structural disturbance; S3, extracting unsteady time periods, constructing disturbance feature sequences, identifying highly suspected disturbance segments and analyzing the response deviation between the structure and the environment, screening highly reliable anomaly segments, and generating anomaly data structures; S4, jointly analyzing the coupling relationship between structural response and environmental disturbance, and constructing a dynamic imbalance index for measuring the intensity of unsteady states. This solves the problem of high false alarm rates in anomaly identification caused by frequent disturbances to monitoring data in complex construction scenarios.

Description

Technical Field

[0001] This invention relates to the field of engineering construction technology, specifically to a monitoring system and method for water conservancy engineering construction monitoring data fusion based on multi-source data. Background Technology

[0002] As water conservancy projects continue to expand in scale and increase in complexity, intelligent and refined monitoring methods are gradually replacing traditional experience-based manual management. The widespread deployment of multi-source sensing devices has enriched on-site monitoring data, and continuous in-depth research and development of engineering technologies have provided algorithmic and system support for data-driven structural health analysis models and monitoring decision-making mechanisms.

[0003] For example, the invention with publication number CN118761637A discloses a method for supervising the construction safety of water conservancy and hydropower projects, which relates to the field of engineering construction technology and includes the following method: inputting the environmental risk level value, the complexity value of construction technology and the number of defects in construction management into the supervision module, and the supervision module outputs a comprehensive risk index, a response capability index and a safety performance index.

[0004] For example, the invention with publication number CN116090822A discloses a water conservancy construction safety protection and supervision system based on data analysis, including a main control module, a data acquisition module, a safety monitoring module, a pre-arrangement module, and a risk warning module; by assigning weights to temporary power supply areas at various locations in water conservancy construction, key monitoring targets are identified, and targeted monitoring and analysis are conducted on these key monitoring targets. At the same time, based on the water conservancy construction nodes where each temporary power supply area is located, potential temporary power supply areas that may appear in the future are determined.

[0005] However, in actual water conservancy construction, due to the complex construction environment and diverse disturbance factors, existing monitoring systems often suffer from problems such as delayed anomaly identification, high false alarm rates, and fragmented data processing chains. Related engineering technology research and development remains weak, making it difficult to effectively support the needs of refined safety management under dynamic operating conditions. Especially in the early stages of disturbance event evolution, the coupling relationship between structural response and environmental changes has not been fully characterized, making it difficult to capture key risk signals in a timely manner.

[0006] Therefore, in order to address the above problems, there is an urgent need for a multi-source data fusion system and method for monitoring and supervising water conservancy project construction data. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides a multi-source data fusion system and method for monitoring water conservancy engineering construction data, which solves the problem of high false alarm rate in anomaly identification caused by frequent disturbances to monitoring data in complex construction scenarios.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: a multi-source data fusion system and method for monitoring water conservancy engineering construction data, comprising the following steps: S1, acquiring water conservancy construction monitoring data by real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system, and preprocessing the water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data; S2, constructing a sliding window based on the preprocessed water conservancy construction monitoring data, extracting key structural and environmental variables, comprehensively analyzing the degree of structural disturbance, dividing different disturbance state segments according to the degree of structural disturbance, and marking the segment state labels for each time period; S3, extracting unsteady time periods based on the segment state labels, constructing disturbance feature sequences, identifying highly suspected disturbance segments and analyzing the response deviation between the structure and the environment, performing a sliding comparison with the three-dimensional structural response sequence and historical steady-state samples, screening highly reliable abnormal segments, and generating abnormal data structures; S4, based on the identified abnormal data structures, jointly analyzing the coupling relationship between structural response and environmental disturbance, constructing a dynamic imbalance index for measuring the intensity of unsteady state, classifying abnormal levels according to the dynamic imbalance index, and updating the corresponding scheduling strategy and acquisition configuration.

[0011] Furthermore, the specific steps for obtaining water conservancy construction monitoring data through real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system are as follows: Water conservancy construction monitoring data is obtained through real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system. This data includes pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress values ​​of load-bearing components, steel reinforcement stress, and structural vibration acceleration. Specifically, pore water pressure is obtained by reading the hydrostatic pressure changes collected by pore pressure gauges embedded in the dam foundation and slope; seepage rate is obtained by reading the flow velocity conversion results output by the differential pressure sensor; and seepage rate is obtained by reading the data installed on-site... The atmospheric pressure is measured in real time by the air pressure module in the field meteorological unit; the construction temperature and humidity are obtained through the temperature and humidity channel data output by the environmental sensor node; the ground settlement is obtained by reading the coordinate change information of the global navigation satellite system measuring points and combining it with the elevation difference extracted from leveling; the concrete crack width is obtained by collecting the displacement measurement between two points of the crack sensor; the stress value of the stressed component is obtained by reading the electrical signal collected by the stress sensor embedded in the load-bearing structure and calculating it after calibration; the steel stress is obtained by collecting the data of the strain gauge on the surface of the steel bar and combining it with the elastic modulus of the steel bar; and the structural vibration acceleration is obtained by reading the instantaneous acceleration value output by the triaxial accelerometer.

[0012] Further, the water conservancy construction monitoring data is preprocessed, and the specific steps to obtain the preprocessed water conservancy construction monitoring data are as follows: Through collaborative detection using the sliding window-based standard deviation outlier detection method and the isolated forest algorithm, abrupt changes and abnormal data links in pore water pressure, seepage rate, and structural vibration acceleration are identified and removed; through a combined completion mechanism using temporal linear interpolation and spline interpolation algorithms, missing fields in surface settlement, construction humidity, and concrete crack width are completed temporally; through a joint modeling strategy combining exponentially weighted moving average and local regression algorithms, trend fitting and high-frequency noise suppression are performed on construction temperature, steel reinforcement stress, and stress values ​​of stressed components; through a cascaded transformation using power function transformation and maximum-minimum normalization algorithms, the distribution of the water conservancy construction monitoring data is adjusted and numerically compressed, completing the unified interval reconstruction and normalization processing of the water conservancy construction monitoring data.

[0013] Furthermore, based on the pre-processed hydraulic construction monitoring data, a sliding window is constructed to extract key structural and environmental variables, and the specific steps for comprehensively analyzing the degree of structural disturbance are as follows: The time series of pre-processed pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress value of stressed components, steel stress, and structural vibration acceleration are aligned according to a unified sampling time and merged according to the monitoring point number to form a multi-dimensional data structure with a unified temporal and spatial index; a fixed-length sliding window is defined in the multi-dimensional data structure, and the concrete crack width and stress value of stressed components within the sliding window are extracted. The disturbance degree of hydraulic engineering structures under complex construction and environmental conditions is assessed by taking into account factors such as concrete crack width, steel reinforcement stress, pore water pressure, atmospheric pressure, construction temperature, and construction humidity, combined with a crack width threshold. The first part is obtained by calculating the square of the concrete crack width divided by the crack width threshold, calculating the absolute value of the difference between the stress values ​​of the load-bearing members and the steel reinforcement stress, adding this ratio to the sum of absolute values, and multiplying by the structural weight. The second part is obtained by calculating the absolute value of the difference between pore water pressure and atmospheric pressure, calculating the sum of construction temperature, construction humidity, and the minimum value, dividing this absolute value by the ratio of the sum, and multiplying by the environmental weight. The first and second parts are then added together to obtain the structural disturbance assessment value.

[0014] Furthermore, the specific steps for dividing different disturbance state segments according to the degree of structural disturbance and labeling the segment status of each time period are as follows: The structural disturbance assessment value within each sliding window is calculated in real time, and the structural disturbance assessment value is compared with the structural disturbance threshold, which includes the primary structural disturbance threshold and the secondary structural disturbance threshold. When the structural disturbance assessment value is less than or equal to the primary structural disturbance threshold, the current time period is marked as a steady-state segment, and the baseline update strategy is automatically triggered. The original hydraulic construction monitoring data within the current window is incrementally updated in the steady-state sample set, and statistical feature classification is performed. When the structural disturbance assessment value is greater than the primary structural disturbance threshold but less than the secondary structural disturbance threshold, the current time period is marked as a transitional segment, the delayed confirmation mechanism is activated, all water conservancy construction monitoring data within the current window are retained, and a time buffer is set to await subsequent state evolution trend judgment; when the structural disturbance assessment value is greater than or equal to the secondary structural disturbance threshold, the current time period is marked as an unstable segment, the high-frequency resampling logic is invoked, the enhanced sampling mode is enabled for the covered monitoring points in the subsequent time period, and the water conservancy construction monitoring data within the current sliding window is pushed to the anomaly identification and processing flow.

[0015] Each time period is marked as a segment status label and attached to the metadata of the current sliding window. This label contains the status category, time segment range and monitoring point number, and is retained along with the time index for use as an input index and scheduling basis for subsequent identification processes.

[0016] Further, based on the segment state labels, the following specific steps are taken to extract unsteady time periods, construct disturbance feature sequences, identify highly suspected disturbance segments, and analyze the response deviation between the structure and the environment: Extract the time periods marked as unsteady segments by the segment state labels, obtain the corresponding pore water pressure, seepage rate, and structural vibration acceleration within the time periods, and construct a pseudo-anomaly candidate sample set by combining the time index and monitoring point number contained in the state labels; in the pseudo-anomaly candidate sample set, construct a sliding time series around pore water pressure, seepage rate, and structural vibration acceleration, and extract three types of statistical features: standard deviation, jump amplitude, and number of fluctuations; after normalizing each statistical feature, form a feature vector, which is then input... The isolated forest algorithm is used to calculate the corresponding isolation degree value; the isolation degree value and the isolation judgment threshold are compared in real time, and samples with isolation degree values ​​exceeding the isolation judgment threshold are screened out and recorded as high-probability perturbation segments; for high-probability perturbation segments, the deviation between their internal structural response and the external environmental background is further quantified; the absolute value of the difference between the stress value of the stressed component and the stress of the steel reinforcement is divided by the sum of the concrete crack width and the minimum term, and the ratio is recorded as the structural compatibility term; the absolute value of the difference between pore water pressure and atmospheric pressure is calculated, 1 is added, and the logarithm is taken; the structural vibration acceleration is divided by this logarithm value and 1 is added; this ratio is recorded as the perturbation amplification term; the structural compatibility term is multiplied by the perturbation amplification term to obtain the structural response deviation value.

[0017] Furthermore, the specific steps for screening high-confidence anomaly fragments and generating anomaly data structures by combining the three-dimensional structural response sequence with historical steady-state samples through sliding comparison are as follows: The calculated structural response deviation value is compared with the response deviation threshold. When the structural response deviation value is less than the response deviation threshold, the highly suspected disturbance fragment is marked as a pseudo-anomaly candidate and retained for subsequent pseudo-anomaly statistical analysis. When the structural response deviation value is greater than or equal to the response deviation threshold, the highly suspected disturbance fragment is transferred to the next step of feature pattern comparison, performing graph similarity analysis: extracting the concrete crack width, stress value of the stressed member, and steel reinforcement stress of the current highly suspected disturbance fragment to form... The three-dimensional structural response sequence is matched with the stable three-dimensional structural response sequence composed of concrete crack width, stress value of stressed member and steel reinforcement stress in the historical steady-state sample set. If the structural response threshold error range constraint is met within a continuous comparison segment, the current high-suspected disturbance segment is marked as a steady-state deviation; otherwise, the current high-suspected disturbance segment is marked as a high-confidence anomaly segment. The water conservancy construction monitoring data corresponding to the time segment identified as a high-confidence anomaly segment is recorded together with the time index, monitoring point number, structural disturbance assessment value and structural response deviation value to form a complete anomaly data structure, which serves as the input basis for the subsequent anomaly analysis and judgment process.

[0018] Furthermore, based on the identified anomalous data structure, the specific steps for constructing a dynamic imbalance index to measure the intensity of unsteady state are as follows: Based on the anomalous data structure, extract all water conservancy construction monitoring data for the time period corresponding to the high-confidence anomalous segment; combine the structural disturbance assessment value and the structural response deviation value to comprehensively analyze the coupling mode between structural response and environmental disturbance, and quantify the unsteady state driving intensity of the current time period; calculate the square of the sum of the structural disturbance assessment value and the structural response deviation value, and record it as the structural response amplification term; calculate the product of surface subsidence and seepage rate, divide this product by the construction temperature plus a minimum term, and record the ratio as the geological environment coupling term; calculate the absolute value of the difference between pore water pressure and atmospheric pressure, add 1, and take the logarithm, and record this logarithm as the external pressure disturbance adjustment term; add the three terms—structural response amplification term, geological environment coupling term, and external pressure disturbance adjustment term—to obtain the dynamic imbalance value.

[0019] Furthermore, based on the dynamic imbalance index, the anomaly levels are classified, and the specific steps for updating the corresponding scheduling strategy and data acquisition configuration are as follows: After calculating the dynamic imbalance value, the dynamic imbalance value and the imbalance threshold are compared in real time to classify high-confidence anomaly segments into levels; among which, the imbalance threshold includes a first-level imbalance threshold and a second-level imbalance threshold; when the dynamic imbalance value is less than or equal to the first-level imbalance threshold, the current high-confidence anomaly segment is marked as a slight disturbance, triggering the local resampling logic to perform encrypted sampling of the current monitoring point; when the dynamic imbalance value is greater than the first-level imbalance threshold but less than the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current monitoring point is subjected to encrypted sampling, and the water conservancy monitoring data corresponding to the current high-confidence anomaly segment is pushed to the supervision platform with an additional manual review prompt; when the dynamic imbalance value is greater than the first-level imbalance threshold but less than the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current monitoring point is subjected to encrypted sampling, and the water conservancy monitoring data corresponding to the current high-confidence anomaly segment is pushed to the supervision platform with an additional manual review prompt; when the dynamic imbalance value is less than or equal to the first-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current high-confidence anomaly segment ... When the dynamic imbalance value is greater than or equal to the secondary imbalance threshold, the current high-confidence anomaly segment is marked as a severe disturbance, triggering a station-wide alarm process, suspending related construction tasks, and transferring the segment data to the anomaly tracing and processing process. The dynamic imbalance value, anomaly level label, scheduling processing record, and corresponding time index are archived to generate an anomaly event log, which is bound to the original water conservancy construction monitoring data to achieve full-process index tracing. The anomaly level, dynamic imbalance value, structural disturbance assessment value, and structural response deviation value marked in the anomaly event log are synchronously written into the acquisition scheduling configuration table to update the sampling frequency setting, monitoring point priority ranking, and comparison threshold adjustment scheme, realizing adaptive scheduling optimization based on actual anomaly evolution characteristics, and completing a closed-loop process integrating monitoring identification, response control, and acquisition strategy.

[0020] The second aspect of this invention provides a multi-source data fusion system for monitoring and supervising water conservancy engineering construction data, comprising: a water conservancy construction monitoring data acquisition and preprocessing module, a structural disturbance assessment and status marking module, a disturbance identification and anomaly sample extraction module, and an anomaly analysis, judgment, and scheduling linkage module. The water conservancy construction monitoring data acquisition and preprocessing module is used to acquire water conservancy construction monitoring data by real-time data acquisition and synchronous processing from multiple types of sensing devices in the construction monitoring system, and to preprocess the water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data. The structural disturbance assessment and status marking module is used to construct a sliding window based on the preprocessed water conservancy construction monitoring data, extract key structural and environmental variables, and comprehensively analyze the results. The system is divided into several modules: a disturbance degree module, a disturbance identification and anomaly sample extraction module, and an anomaly sample extraction module. The former is used to extract unsteady time periods based on the segment state labels, construct disturbance feature sequences, identify highly suspected disturbance segments, analyze the response deviation between the structure and the environment, and perform sliding comparisons with historical steady-state samples to screen highly reliable anomaly segments and generate anomaly data structures. The latter is used to analyze the coupling relationship between structural response and environmental disturbances based on the identified anomaly data structures, construct a dynamic imbalance index to measure the intensity of unsteady states, classify anomaly levels based on the dynamic imbalance index, and update corresponding scheduling strategies and data acquisition configurations.

[0021] Beneficial effects

[0022] The present invention has the following beneficial effects:

[0023] (1) This multi-source data fusion water conservancy engineering construction monitoring data supervision system and method, by constructing a structural disturbance assessment value and integrating water conservancy construction monitoring data, analyzes the disturbance trend in real time within a sliding window. This indicator can distinguish between three types of segment states: steady state, transitional state, and unstable state, effectively supporting the early identification and state classification of disturbances. Its introduction realizes the quantitative expression of disturbance intensity, avoiding the problem of traditional methods that highly rely on empirical thresholds and single-point monitoring data for anomaly judgment, and providing a precise entry point for subsequent high-frequency sampling scheduling and anomaly screening.

[0024] (2) The water conservancy project construction monitoring data supervision system and method based on multi-source data fusion, by relying on the combination relationship of steel reinforcement stress, stress of stressed components, concrete crack width, structural vibration acceleration, pore water pressure and atmospheric pressure to perform deviation modeling, obtains structural response deviation value, which not only reflects the degree of reaction of disturbance influence inside the structure, but also reflects the amplification effect of external disturbance factors on the structural state. It is a further precise screening after disturbance assessment, effectively improves the identification accuracy of high-confidence abnormal segments, and provides key input for subsequent imbalance level classification.

[0025] (3) The water conservancy project construction monitoring data supervision system and method, which integrates dynamic imbalance values, comprehensive structural disturbance assessment values, structural response deviation values, surface settlement, seepage rate, pore water pressure, atmospheric pressure and other key variables, further characterizes the coupling relationship between disturbance sources and response mechanisms on the basis of abnormal state identification. It breaks through the limitation of the traditional monitoring system's modeling of structural behavior and environmental changes separately, and can dynamically reflect the linkage between different monitoring data, providing a quantitative basis for state perception and trend judgment under complex construction conditions.

[0026] (4) The water conservancy project construction monitoring data supervision system and method, which integrates multi-source data fusion, automatically integrates highly reliable anomaly segments with multi-source information such as time index, monitoring point number, structural disturbance assessment value, and structural response deviation value to construct a structured anomaly data object. This data structure not only retains the original monitoring data but also links the analysis logic and calculation results in the entire identification chain, providing a reliable foundation for subsequent cause investigation, event reconstruction, manual verification, and model optimization, and realizing closed-loop support for the discovery and interpretation of anomaly information. Attached Figure Description

[0027] Figure 1 Flowchart of a method for supervising construction monitoring data of water conservancy projects based on source data fusion;

[0028] Figure 2 Structure diagram of a water conservancy project construction monitoring data supervision system for source data fusion;

[0029] Figure 3 Distribution of structural disturbance assessment values ​​for each time period;

[0030] Figure 4 This is a distribution map of dynamic imbalance values ​​and disturbance levels for highly reliable anomaly segments. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figures 1-4This invention provides a technical solution: a multi-source data fusion system and method for monitoring water conservancy engineering construction data, comprising the following steps: S1, acquiring water conservancy construction monitoring data by real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system, and preprocessing the water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data; S2, constructing a sliding window based on the preprocessed water conservancy construction monitoring data, extracting key structures and environmental variables, comprehensively analyzing the degree of structural disturbance, dividing different disturbance state segments according to the degree of structural disturbance, and marking the segment state labels of each time period; S3, extracting unsteady time periods based on the segment state labels, constructing disturbance feature sequences, identifying highly suspected disturbance segments and analyzing the response deviation between the structure and the environment, combining the three-dimensional structural response sequence with historical steady-state samples for sliding comparison, screening highly reliable abnormal segments and generating abnormal data structures; S4, based on the identified abnormal data structures, jointly analyzing the coupling relationship between structural response and environmental disturbance, constructing a dynamic imbalance index for measuring the intensity of unsteady state, classifying abnormal levels according to the dynamic imbalance index, and updating the corresponding scheduling strategy and acquisition configuration.

[0033] Specifically, the steps for obtaining water conservancy construction monitoring data through real-time data acquisition and synchronous processing of multiple types of sensors in the construction monitoring system are as follows: Water conservancy construction monitoring data is obtained through real-time data acquisition and synchronous processing of multiple types of sensors in the construction monitoring system. This data includes pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress values ​​of load-bearing components, steel reinforcement stress, and structural vibration acceleration. Specifically, pore water pressure is obtained by reading the hydrostatic pressure changes collected by pore pressure gauges embedded in the dam foundation and slope; seepage rate is obtained by reading the flow velocity conversion results output by the differential pressure sensor; and seepage rate is obtained by reading data installed on-site... The meteorological unit measures atmospheric pressure in real time using the barometric pressure module; it obtains construction temperature and humidity data from the temperature and humidity channels output by the environmental sensor nodes; it acquires ground settlement by reading coordinate change information from GPS measurement points and combining it with elevation differences extracted from leveling measurements; it obtains concrete crack width by collecting displacement measurements between two points from crack sensors; it obtains stress values ​​of load-bearing components by reading and calibrating electrical signals collected by stress sensors embedded in the load-bearing structure; it obtains steel stress by collecting data from strain gauges on the steel reinforcement surface and converting it with the steel reinforcement's elastic modulus; and it obtains structural vibration acceleration by reading instantaneous acceleration values ​​output by triaxial accelerometers.

[0034] In this implementation plan, by acquiring and processing real-time data from multiple types of sensing devices in the construction monitoring system, comprehensive acquisition of key hydraulic construction monitoring data, such as pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress value of stressed components, steel reinforcement stress, and structural vibration acceleration, is achieved. This ensures multi-dimensional information coverage of structural response, geological changes, and environmental conditions. By reading data from different types of devices, including pore pressure gauges, differential pressure sensors, air pressure modules, environmental sensing nodes, GPS measuring points, leveling equipment, crack sensors, stress sensors, strain gauges, and triaxial accelerometers, the source of each monitoring data item is separated, indicators are clearly defined, and data consistency is controlled. This provides a stable and accurate data foundation for subsequent anomaly identification, disturbance assessment, and coordinated scheduling.

[0035] Specifically, the preprocessing steps for water conservancy construction monitoring data are as follows: First, a collaborative detection method combining standard deviation outlier detection based on a sliding window and an isolated forest algorithm is used to identify and remove abrupt changes and abnormal data in pore water pressure, seepage rate, and structural vibration acceleration, ensuring the reliability of the source data in terms of temporal continuity and transmission stability. Second, a combined completion mechanism using temporal linear interpolation and spline interpolation algorithms is employed to complete missing fields in surface settlement, construction humidity, and concrete crack width, retaining the source identifiers to support the traceability of subsequent processing nodes. Third, a joint modeling strategy combining exponentially weighted moving average and local regression algorithms is used to perform trend fitting and high-frequency noise suppression on construction temperature, steel reinforcement stress, and stress values ​​of stressed components, reducing false triggering interference of monitoring data while preserving the abnormal structural response characteristics. Fourth, a cascaded transformation using power function transformation and max-min normalization algorithms is employed to adjust the distribution and compress the numerical values ​​of the water conservancy construction monitoring data, completing the scaling and normalization of the data without losing the original data's structural features.

[0036] In this implementation plan, a highly robust preprocessing mechanism for addressing data stability and integrity issues in complex scenarios is constructed. This mechanism involves outlier removal of pore water pressure, seepage rate, and structural vibration acceleration; completion of missing data on surface settlement, construction humidity, and concrete crack width; noise reduction through trend fitting of construction temperature, steel reinforcement stress, and stress values ​​of load-bearing components; and normalization reconstruction of all hydraulic construction monitoring data. Supported by engineering technology research and development, this mechanism not only improves the usability and consistency of hydraulic construction monitoring data but also provides a reliable data foundation for subsequent structural disturbance assessment and anomaly identification, ensuring the monitoring system has effective data support capabilities under unsteady conditions.

[0037] Specifically, the steps for constructing a sliding window based on preprocessed hydraulic construction monitoring data, extracting key structural and environmental variables, and comprehensively analyzing the degree of structural disturbance are as follows: The time series of preprocessed pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress values ​​of stressed components, steel reinforcement stress, and structural vibration acceleration are aligned according to a unified sampling time and merged based on monitoring point numbers to form a multidimensional data structure with unified temporal and spatial indexes, providing consistent input conditions for dynamic comparative analysis between different monitoring points. Within this multidimensional data structure, a fixed-length sliding window is defined to ensure the continuity and comparability of structural disturbance information within the time domain. Subsequently, the concrete crack width, stress value of the load-bearing member, steel reinforcement stress, pore water pressure, atmospheric pressure, construction temperature, and construction humidity within the sliding window are extracted as key structural and environmental variables. Combined with a crack width threshold as a basic reference for structural response anomalies, the degree of disturbance to the hydraulic engineering structure under complex construction and environmental conditions is assessed: the square of the concrete crack width is calculated and divided by the crack width threshold to measure the degree of crack propagation anomalies; the absolute value of the difference between the stress value of the load-bearing member and the stress of the steel reinforcement is calculated to reflect the deviation in internal mechanical transmission; this ratio is added to the sum of absolute values ​​and multiplied by the structural weight to obtain the first part; the absolute value of the difference between pore water pressure and atmospheric pressure is calculated to characterize the influence of external hydrological disturbances; the sum of construction temperature, construction humidity, and a minimum term is calculated to construct an environmental factor adjustment term; the minimum term is a set decimal constant used to avoid numerical anomalies with a denominator of zero, ensuring calculation stability; this absolute value is divided by the ratio of the sum and multiplied by the environmental weight to obtain the second part; the first and second parts are added together to obtain the structural disturbance assessment value, which serves as the basic indicator for subsequent state classification and anomaly identification. Among them, the structural weight and environmental weight are obtained by fitting the joint changes of concrete crack width, stress value of load-bearing member, steel stress, pore water pressure, atmospheric pressure, construction temperature and construction humidity in multiple time periods using the least squares regression algorithm. The values ​​of structural weight and environmental weight are both in the range of [1,0].

[0038] The specific formula for calculating the structural disturbance assessment value is as follows:

[0039]

[0040] In the formula, D represents the structural disturbance assessment value, α represents the structural weight, β represents the environmental weight, and W... c W represents the width of a concrete crack. r σ represents the crack width threshold. s σ represents the stress value of a stressed component. r P represents the stress in the reinforcing steel. w P represents the pore water pressure. adenoted by atmospheric pressure, T represents construction temperature, RH represents construction humidity, and ∈ represents a minterm.

[0041] In this embodiment, Table 1 is a data table of structural disturbance assessment values, which records in detail the key monitoring variables and structural disturbance assessment value calculation results in the process of structural disturbance modeling and analysis at five different time periods. It is used to measure the stability of the structural response under complex construction and environmental conditions. Specifically: In time period 1, the concrete crack width was 0.35, the structural weight was 0.5, the stress value of the load-bearing member was 18.5, the steel reinforcement stress was 17.9, the pore water pressure was 105.4, the atmospheric pressure was 101.3, the construction temperature was 28.5°C, the construction humidity was 74.3%, and the structural disturbance assessment value was 0.52; In time period 2, the concrete crack width was 0.28, the structural weight was 0.5, the stress value of the load-bearing member was 16.2, the steel reinforcement stress was 15.7, the pore water pressure was 97.2, the atmospheric pressure was 101.3, the construction temperature was 27.2°C, the construction humidity was 70.5%, and the structural disturbance assessment value was 0.41; In time period 3, the concrete crack width was 0.42, the structural weight was 0.5, the stress value of the load-bearing member was 19.8, and the steel reinforcement stress was 19.1. In time period 4, the pore water pressure was 112.6, atmospheric pressure was 101.3, construction temperature was 29.6, construction humidity was 76.8, and the structural disturbance assessment value was 0.67. In time period 5, the concrete crack width was 0.33, structural weight was 0.5, stress value of load-bearing members was 17.3, steel reinforcement stress was 16.9, pore water pressure was 101.8, atmospheric pressure was 101.3, construction temperature was 26.8, construction humidity was 72.4, and the structural disturbance assessment value was 0.37. In time period 6, the concrete crack width was 0.31, structural weight was 0.5, stress value of load-bearing members was 16.7, steel reinforcement stress was 16.2, pore water pressure was 99.5, atmospheric pressure was 101.3, construction temperature was 27.9, construction humidity was 71.6, and the structural disturbance assessment value was 0.42.

[0042] Table 1. Structural Disturbance Assessment Values

[0043] Time period <![CDATA[W c ]]> <![CDATA[W r ]]> <![CDATA[σ s ]]> <![CDATA[σ r ]]> <![CDATA[P w ]]> <![CDATA[P a ]]> T RH D Time period 1 0.35 0.5 18.5 17.9 105.4 101.3 28.5 74.3 0.52 Time period 2 0.28 0.5 16.2 15.7 97.2 101.3 27.2 70.5 0.41 Time period 3 0.42 0.5 19.8 19.1 112.6 101.3 29.6 76.8 0.67 Time period 4 0.33 0.5 17.3 16.9 101.8 101.3 26.8 72.4 0.37 Time period 5 0.31 0.5 16.7 16.2 99.5 101.3 27.9 71.6 0.42

[0044] like Figure 3 The figure shows the distribution of structural disturbance assessment values ​​across different time periods, providing a visual overview of the distribution. The figure displays structural disturbance assessment values ​​for five different time periods, fluctuating between 0.37 and 0.67. Time period 3 shows the highest disturbance value at 0.67, reflecting a significant deviation between the structural stress and environmental coupling state during this period. Time period 4 shows the lowest disturbance value at only 0.37, indicating a relatively stable state. This distribution of structural disturbance assessment values ​​across time periods helps to clearly identify disturbance trends, providing a quantitative basis for subsequent structural state identification and dynamic response strategy development.

[0045] This implementation scheme constructs a multidimensional data structure with unified temporal and spatial indexes, defines a sliding window, and extracts key structural and environmental variables such as concrete crack width, stress values ​​of stressed components, steel reinforcement stress, pore water pressure, atmospheric pressure, construction temperature, and construction humidity. Combined with a crack width threshold, a structural disturbance assessment value is constructed, effectively reflecting the degree of disturbance in hydraulic engineering projects under complex construction conditions. This method not only ensures the continuity and comparability of structural disturbance characteristics in the time domain but also enhances the engineering applicability of the calculation results by introducing structural and environmental weights and setting minima to ensure numerical stability. With the support of in-depth engineering technology research and development, it provides operable and quantifiable basic indicators for subsequent state classification and anomaly identification, improving the integrated analysis level of monitoring data and the sensitivity of on-site disturbance response.

[0046] Specifically, the steps for dividing different disturbance state segments according to the degree of structural disturbance and labeling the segment status of each time period are as follows: The structural disturbance assessment value within each sliding window is calculated in real time, and the structural disturbance assessment value is compared with the structural disturbance threshold, which includes the primary structural disturbance threshold and the secondary structural disturbance threshold. When the structural disturbance assessment value is less than or equal to the primary structural disturbance threshold, the current time period is marked as a steady-state segment, and the baseline update strategy is automatically triggered. The original hydraulic construction monitoring data within the current window is incrementally updated in the steady-state sample set, and statistical features are archived. When the structural disturbance assessment value is greater than the primary structural disturbance threshold but less than the secondary structural disturbance threshold, the current time period is marked as a transitional segment, the delayed confirmation mechanism is activated, all water conservancy construction monitoring data within the current window are retained, and a time buffer is set to await subsequent state evolution trend judgment; when the structural disturbance assessment value is greater than or equal to the secondary structural disturbance threshold, the current time period is marked as an unstable segment, the high-frequency resampling logic is invoked, the enhanced sampling mode is enabled for the covered monitoring points in the subsequent time period, and the water conservancy construction monitoring data within the current sliding window is pushed to the anomaly identification and processing flow.

[0047] Each time period is marked as a segment status label and attached to the metadata of the current sliding window. This label contains the status category, time segment range and monitoring point number, and is retained along with the time index for use as an input index and scheduling basis for subsequent identification processes.

[0048] This implementation plan dynamically divides the system into steady-state, transitional, and unstable sections by comparing the assessed structural disturbance values ​​with the set primary and secondary structural disturbance thresholds in real time. Differentiated processing strategies are then developed for each section, such as baseline updates, delayed confirmation, and high-frequency resampling, enhancing adaptability to complex construction conditions and improving the targeted nature of data management. Simultaneously, by binding section status tags to sliding window metadata, recording status categories, time ranges, and monitoring point numbers, traceable labeling and scheduling of structural disturbance states are achieved. This provides accurate and efficient input indexes for subsequent anomaly identification processes, improving the data closure capability and engineering response efficiency of the entire monitoring chain.

[0049] Specifically, the steps for extracting unsteady time periods based on segment state labels, constructing disturbance feature sequences, identifying highly suspected disturbance segments, and analyzing the response deviation between the structure and the environment are as follows: Extracting time periods marked as unsteady segments by segment state labels, obtaining the corresponding pore water pressure, seepage rate, and structural vibration acceleration within the time period, and combining the time index and monitoring point number contained in the state labels to construct a pseudo-anomaly candidate sample set, providing a data source guarantee for subsequent anomaly detection; Constructing a sliding time series around pore water pressure, seepage rate, and structural vibration acceleration in the pseudo-anomaly candidate sample set, extracting three types of statistical features: standard deviation, jump amplitude, and number of fluctuations, to comprehensively reflect the severity and duration of disturbance changes; Normalizing each statistical feature to form a feature vector, inputting it into the isolated forest algorithm to calculate the corresponding isolation degree value, fully utilizing the unsupervised model to automatically characterize and score unsteady disturbance behavior; Real-time comparison of the isolation degree value and the isolation judgment threshold, and... Samples with isolation values ​​exceeding the isolation threshold are selected and recorded as high-probability disturbance segments, serving as the basis for subsequent deviation quantification analysis. For high-probability disturbance segments, the degree of deviation between their internal structural response and the external environmental background is further quantified to improve the judgment criteria for disturbance identification. The absolute value of the difference between the stress value of the stressed component and the stress of the steel reinforcement is divided by the sum of the concrete crack width and the minimum term, and the ratio is recorded as the structural compatibility term, reflecting the internal consistency of the structural mechanical behavior. The minimum term is used to prevent the denominator from being zero, which would cause abnormal calculation results. The absolute value of the difference between pore water pressure and atmospheric pressure is calculated, and then the logarithm is taken to construct a hydrological disturbance amplification adjustment term. The structural vibration acceleration is then divided by this logarithm and added by 1, and the ratio is used as the disturbance amplification term to quantify the amplification effect of external disturbances on structural behavior. Finally, the structural compatibility term is multiplied by the disturbance amplification term to obtain the structural response deviation value, providing a key basic indicator for subsequent identification of high-confidence abnormal segments and dynamic imbalance strength assessment.

[0050] The specific formula for calculating the structural response deviation value is as follows:

[0051]

[0052] In the formula, S represents the structural response deviation value, σ s σ represents the stress value of a stressed component. r W represents the stress in the reinforcing steel. c The value represents the width of the concrete crack, ∈ represents the minimum term, a represents the structural vibration acceleration, and P represents the minimum value of the crack. w P represents the pore water pressure. a It represents atmospheric pressure.

[0053] In this implementation scheme, a pseudo-anomaly candidate sample set is constructed, key statistical features of pore water pressure, seepage rate, and structural vibration acceleration are extracted, and the isolated forest algorithm is used to effectively screen highly suspected disturbance segments, thereby improving the sensitivity and robustness of non-steady-state behavior identification. Simultaneously, a structural compatibility term and a disturbance amplification term are introduced to quantify the amplified impact of the internal mechanical consistency of the structure and the external hydrological environment, ultimately generating a structural response deviation value. This provides a clear analytical basis for further classification of anomaly segments and dynamic imbalance assessment, strengthening the diagnostic capability for structure-environment coupled disturbances. This method possesses strong data adaptability and analytical scalability, providing precise support for the design of subsequent anomaly management strategies.

[0054] Specifically, the steps for screening high-confidence anomalous segments and generating anomalous data structures by combining the three-dimensional structural response sequence with historical steady-state samples through sliding comparison are as follows: The calculated structural response deviation value is compared with a response deviation threshold. When the structural response deviation value is less than the response deviation threshold, the highly suspected disturbance segment is marked as a pseudo-anomaly candidate and retained for subsequent pseudo-anomaly statistical analysis. When the structural response deviation value is greater than or equal to the response deviation threshold, the highly suspected disturbance segment is transferred to the next step of feature pattern comparison, performing graph similarity analysis: extracting the concrete crack width, stress value of the stressed member, and steel reinforcement stress of the current highly suspected disturbance segment to form... The three-dimensional structural response sequence is matched with the stable three-dimensional structural response sequence composed of concrete crack width, stress value of stressed member and steel reinforcement stress in the historical steady-state sample set. If the structural response threshold error range constraint is met within a continuous comparison segment, the current high-suspected disturbance segment is marked as a steady-state deviation; otherwise, the current high-suspected disturbance segment is marked as a high-confidence anomaly segment. The water conservancy construction monitoring data corresponding to the time segment identified as a high-confidence anomaly segment is recorded together with the time index, monitoring point number, structural disturbance assessment value and structural response deviation value to form a complete anomaly data structure, which serves as the input basis for the subsequent anomaly analysis and judgment process.

[0055] In this implementation plan, a three-dimensional structural response sequence comparison mechanism is introduced. Combining the joint characteristics of concrete crack width, stress values ​​of stressed components, and steel reinforcement stress, a sliding comparison strategy is constructed, significantly improving the accuracy and efficiency of screening highly suspected disturbance segments. With the support of engineering technology research and development, by comparing the numerical range of stable structural response sequences with historical steady-state samples, false anomaly segments are systematically eliminated, and highly reliable anomaly segments are clearly marked. Simultaneously, the time index, monitoring point number, structural disturbance assessment value, and structural response deviation value of the highly reliable anomaly segments are recorded together to generate a standardized anomaly data structure. This provides a structured input and decision-making basis for subsequent anomaly analysis and judgment processes, ensuring the traceability and data closure of the anomaly identification process.

[0056] Specifically, based on the identified anomalous data structure, the following steps are taken to construct a dynamic imbalance index for measuring unsteady-state intensity by jointly analyzing the coupling relationship between structural response and environmental disturbances: Based on the anomalous data structure, extract all hydraulic construction monitoring data for the time period corresponding to the high-confidence anomalous segments. This data includes pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress values ​​of stressed components, steel reinforcement stress, and structural vibration acceleration. Combine the structural disturbance assessment value with the structural response deviation value to construct a two-dimensional perspective of structure and environment. Perform a comprehensive analysis of the coupling mode between structural response and environmental disturbances to quantify the unsteady-state driving intensity for the current time period. Calculate the square of the sum of the structural disturbance assessment value and the structural response deviation value to reflect the structural... The anomalous intensity of the cumulative response at the structural level is denoted as the structural response amplification term. The product of surface subsidence and seepage rate is calculated, and this product is divided by the sum of construction temperature and a minimum term, where the minimum term is a set small positive constant used for numerical stability control and to prevent calculation errors caused by the denominator approaching zero. This ratio can characterize the coupling effect between local geological disturbance and groundwater activity, and is denoted as the geological environment coupling term. The absolute value of the difference between pore water pressure and atmospheric pressure is calculated, and 1 is added to it before taking the natural logarithm, which is used to measure the moderating effect of anomalous external pressure changes on structural disturbance. This logarithmic value is denoted as the external pressure disturbance moderating term. The structural response amplification term, the geological environment coupling term, and the external pressure disturbance moderating term are added together to form the dynamic imbalance value corresponding to the current high-confidence anomaly segment, which serves as a key input indicator for subsequent anomaly level determination.

[0057] The specific formula for calculating the dynamic imbalance value is as follows:

[0058]

[0059] In the formula, represents the dynamic imbalance value, D represents the structural disturbance assessment value, S represents the structural response deviation value, h represents the surface settlement, v represents the seepage rate, T represents the construction temperature, represents the pore water pressure, represents the atmospheric pressure, and ε represents the minimum term.

[0060] In this embodiment, Table 2 is a dynamic imbalance value data table, which records in detail the key monitoring data and index calculation results of five high-confidence anomaly segments in the comprehensive analysis of structural disturbance and environmental disturbance, and is used to quantify the degree of structural imbalance under unstable conditions. Specifically: the structural disturbance assessment value corresponding to high-confidence anomaly segment 1 is 1.82, the structural response deviation value is 1.65, the surface settlement is 9.80, the seepage rate is 3.10, the construction temperature is 26.50, the pore water pressure is 89.40, the atmospheric pressure is 101.80, and the dynamic imbalance value is 11.48; in high-confidence anomaly segment 2, the structural disturbance assessment value is 2.10, the structural response deviation value is 1.78, the surface settlement is 12.30, the seepage rate is 3.60, the construction temperature is 27.80, and the pore water pressure is 87. In high-confidence anomaly segment 3, the structural disturbance assessment value is 1.45, the structural response deviation value is 1.52, the surface settlement is 7.20, the seepage rate is 2.50, the construction temperature is 25.60, the pore water pressure is 90.50, the atmospheric pressure is 101.70, and the dynamic imbalance value is 8.52. In high-confidence anomaly segment 4, the structural disturbance assessment value is 2.45, the structural response deviation value is 2.30, the surface settlement is 15.10, and the seepage rate is 4.20.

[0061] The construction temperature was 29.10°C, the pore water pressure was 86.00°C, the atmospheric pressure was 101.20°C, and the dynamic imbalance value was 18.83. In the high-confidence anomaly segment 5, the structural disturbance assessment value was 1.65, the structural response deviation value was 1.70, the surface settlement was 8.40, the seepage rate was 2.80, the construction temperature was 26.90°C, the pore water pressure was 88.60°C, the atmospheric pressure was 101.60°C, and the dynamic imbalance value was 9.71.

[0062] Table 2 Dynamic Imbalance Value Data Table

[0063] Excerpt D S h v T <![CDATA[P w ]]> <![CDATA[P w ]]> Q 1 1.06 0.79 10.41 0.17 27.24 128.52 101.3 6.85 2 1.47 0.73 29.40 0.22 17.79 69.97 101.3 8.68 3 1.31 1.22 26.65 0.31 20.84 101.42 101.3 6.92 4 1.22 1.06 14.25 0.27 22.33 109.24 101.3 7.56 5 0.91 1.12 13.64 0.22 24.12 54.65 101.3 8.12

[0064] like Figure 4The figure shows the distribution of dynamic imbalance values ​​and disturbance levels for five high-confidence anomaly segments. It illustrates the dynamic imbalance values ​​and corresponding disturbance level classifications for these segments, reflecting the system's automatic identification and classification capabilities under different anomaly intensities. The horizontal axis represents the sample number, and the vertical axis represents the dynamic imbalance value. The bar colors are categorized according to the disturbance level, corresponding to mild, critical, and severe disturbances, respectively. The figure shows that the dynamic imbalance values ​​for samples 1 and 3 are 6.85 and 6.92, respectively, indicating mild disturbances; while the dynamic imbalance values ​​for samples 2, 4, and 5 are 8.68, 7.56, and 8.12, respectively, indicating critical disturbances. The specific imbalance value and disturbance level for each sample are clearly labeled above the bars, and the legend provides a clear correspondence between color and level, intuitively reflecting the distribution of dynamic imbalance values ​​in the high-confidence anomaly segments. This provides a quantitative basis for subsequent scheduling response strategy formulation and sampling configuration adjustment.

[0065] In this implementation plan, by introducing structural disturbance assessment values ​​and structural response deviation values, and combining them with key hydraulic construction monitoring data such as surface settlement, seepage rate, construction temperature, pore water pressure, and atmospheric pressure, a comprehensive index, dynamic imbalance value, is constructed. This provides a refined means of expressing the intensity of unsteady-state driving forces. This index integrates three parts: structural response increase term, geological environment coupling term, and external pressure disturbance adjustment term. It systematically reflects the complex coupling relationship between structural response and environmental disturbance, effectively improving the accuracy and continuity of abnormal state characterization, and providing a solid data foundation for subsequent graded diagnosis and scheduling strategy formulation.

[0066] Specifically, based on the dynamic imbalance index, the anomaly levels are classified, and the corresponding scheduling strategies and data acquisition configuration updates are as follows: After calculating the dynamic imbalance value, the dynamic imbalance value and the imbalance threshold are compared in real time to classify the high-confidence anomaly segments; the imbalance threshold includes a first-level imbalance threshold and a second-level imbalance threshold; when the dynamic imbalance value is less than or equal to the first-level imbalance threshold, the current high-confidence anomaly segment is marked as a slight disturbance, triggering local resampling logic and performing encrypted sampling on the current monitoring point; when the dynamic imbalance value is greater than the first-level imbalance threshold but less than the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current monitoring point is subjected to encrypted sampling, and the water conservancy monitoring data corresponding to the current high-confidence anomaly segment is pushed to the supervision platform with an additional manual review prompt; when the dynamic imbalance value is greater than the first-level imbalance threshold but less than the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current monitoring point is subjected to encrypted sampling, and the water conservancy monitoring data corresponding to the current high-confidence anomaly segment is pushed to the supervision platform with an additional manual review prompt; when the dynamic imbalance value is less than or equal to the first-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current high-confidence anomaly segment ... When the dynamic imbalance value is greater than or equal to the secondary imbalance threshold, the current high-confidence anomaly segment is marked as a severe disturbance, triggering a station-wide alarm process, suspending related construction tasks, and transferring the segment data to the anomaly tracing and processing process. The dynamic imbalance value, anomaly level label, scheduling processing record, and corresponding time index are archived to generate an anomaly event log, which is bound to the original water conservancy construction monitoring data to achieve full-process index tracing. The anomaly level, dynamic imbalance value, structural disturbance assessment value, and structural response deviation value marked in the anomaly event log are synchronously written into the acquisition scheduling configuration table to update the sampling frequency setting, monitoring point priority ranking, and comparison threshold adjustment scheme, realizing adaptive scheduling optimization based on actual anomaly evolution characteristics, and completing a closed-loop process integrating monitoring identification, response control, and acquisition strategy.

[0067] This implementation plan introduces a dynamic imbalance value and imbalance threshold classification mechanism to clearly define three anomaly levels: mild disturbance, critical disturbance, and severe disturbance, effectively enhancing the hierarchical response capability of high-confidence anomaly segments. Driven by the dynamic imbalance value, a coordinated scheduling mechanism for local resampling, manual review, station-wide alarms, and anomaly tracing is implemented. Key parameters such as dynamic imbalance values, anomaly level labels, structural disturbance assessment values, and structural response deviation values ​​are archived in the anomaly event log to ensure the integrity of the entire process indexing and tracing. By synchronously writing log results into the acquisition scheduling configuration table, the sampling frequency, monitoring point priority, and comparison threshold settings are dynamically adjusted, constructing a closed-loop scheduling optimization system that integrates monitoring identification, response control, and acquisition strategies, significantly improving the adaptive capability and data decision-making accuracy of water conservancy construction monitoring.

[0068] like Figure 2As shown, the second aspect of this invention provides a multi-source data fusion system for monitoring water conservancy engineering construction data, including: a water conservancy construction monitoring data acquisition and preprocessing module, a structural disturbance assessment and status marking module, a disturbance identification and abnormal sample extraction module, and an abnormal analysis judgment and scheduling linkage module. The water conservancy construction monitoring data acquisition and preprocessing module is used to acquire water conservancy construction monitoring data by real-time data acquisition and synchronous processing from multiple types of sensing devices in the construction monitoring system, and to preprocess the water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data. The structural disturbance assessment and status marking module is used to construct a sliding window based on the preprocessed water conservancy construction monitoring data, extract key structural and environmental variables, and conduct comprehensive analysis. The system is divided into several modules: 1) Structural Disturbance Degree: Different disturbance state segments are defined based on the degree of structural disturbance, and segment state labels are assigned to each time period. 2) Disturbance Identification and Anomaly Sample Extraction: This module extracts unsteady time periods based on segment state labels, constructs disturbance feature sequences, identifies highly probable disturbance segments, analyzes the response deviation between the structure and the environment, performs a sliding comparison between the three-dimensional structural response sequence and historical steady-state samples, filters highly reliable anomaly segments, and generates anomaly data structures. 3) Anomaly Analysis, Judgment, and Scheduling Linkage: Based on the identified anomaly data structures, this module jointly analyzes the coupling relationship between structural response and environmental disturbances, constructs a dynamic imbalance index to measure the intensity of unsteady states, classifies anomaly levels based on the dynamic imbalance index, and updates corresponding scheduling strategies and data acquisition configurations.

[0069] This implementation plan integrates modules for water conservancy construction monitoring data acquisition and preprocessing, structural disturbance assessment and status marking, disturbance identification and anomaly sample extraction, and anomaly analysis, judgment, and scheduling linkage to form a complete data supervision process. Supported by engineering technology research and development, the system can achieve real-time acquisition and unified preprocessing of water conservancy construction monitoring data, ensuring data consistency and availability; it clarifies the evolution characteristics of construction status through quantitative assessment of structural disturbance levels and status segment marking; it accurately extracts highly reliable anomaly segments based on disturbance feature sequences and structural response deviation analysis; and it combines dynamic imbalance index construction to achieve graded response and scheduling strategy updates for anomaly levels, effectively enhancing the monitoring system's adaptive supervision capabilities under complex working conditions and improving the level of data-driven risk identification and control.

[0070] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0071] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for monitoring and supervising construction monitoring data of water conservancy projects through multi-source data fusion, characterized in that, Includes the following steps: S1. By collecting and synchronously processing real-time data from multiple types of sensing devices in the construction monitoring system, water conservancy construction monitoring data is obtained, and the water conservancy construction monitoring data is preprocessed to obtain preprocessed water conservancy construction monitoring data. S2. Based on the preprocessed water conservancy construction monitoring data, a sliding window is constructed to extract key structural and environmental variables, comprehensively analyze the degree of structural disturbance, divide different disturbance state segments according to the degree of structural disturbance, and mark the segment state labels for each time period. S3. Based on the segment state label, extract the unstable time period, construct the disturbance feature sequence, identify high-suspect disturbance segments and analyze the response deviation between the structure and the environment. Combine the three-dimensional structural response sequence with historical steady-state samples for sliding comparison, screen high-confidence abnormal segments and generate abnormal data structures. S4. Based on the identified abnormal data structure, the coupling relationship between structural response and environmental disturbance is jointly analyzed to construct a dynamic imbalance index for measuring the intensity of non-steady state. Anomaly levels are classified according to the dynamic imbalance index, and corresponding scheduling strategies and data acquisition configurations are updated.

2. The method for monitoring and supervising construction monitoring data of water conservancy projects by multi-source data fusion according to claim 1, characterized in that: The specific steps for obtaining water conservancy construction monitoring data through real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system are as follows: By acquiring and synchronously processing real-time data from various types of sensors in the construction monitoring system, hydraulic construction monitoring data is obtained. This data includes pore water pressure, seepage rate, atmospheric pressure, construction temperature, construction humidity, surface settlement, concrete crack width, stress values ​​of load-bearing components, steel reinforcement stress, and structural vibration acceleration. Specifically, pore water pressure is obtained by reading the hydrostatic pressure changes collected by pore pressure gauges embedded in the dam foundation and slopes; seepage rate is obtained by reading the velocity conversion results output by the differential pressure sensor; atmospheric pressure is obtained by reading the real-time measurements from the pressure module installed in the on-site meteorological unit; and environmental pressure is obtained through... The temperature and humidity channels output by the sensing nodes are used to obtain the construction temperature and humidity; the coordinate change information of the measuring points of the Global Navigation Satellite System is read and combined with the elevation difference extracted from the leveling survey to obtain the ground settlement; the displacement measurement between two points of the crack sensor is collected to obtain the width of the concrete crack; the stress value of the stressed component is obtained by reading the electrical signal collected by the stress sensor embedded in the load-bearing structure and calculating it after calibration; the stress of the steel reinforcement is obtained by collecting the data of the strain gauge on the surface of the steel reinforcement and combining it with the elastic modulus of the steel reinforcement; and the structural vibration acceleration is obtained by reading the instantaneous acceleration value output by the triaxial accelerometer.

3. The method for monitoring and supervising water conservancy project construction data through multi-source data fusion according to claim 1, characterized in that: The specific steps for preprocessing water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data are as follows: By employing a collaborative detection method based on sliding window standard deviation outlier detection and isolated forest algorithm, abrupt changes and abnormal data links in pore water pressure, seepage rate, and structural vibration acceleration are identified and removed. A combined completion mechanism using temporal linear interpolation and spline interpolation algorithms is used to complete missing fields in surface settlement, construction humidity, and concrete crack width. A joint modeling strategy combining exponentially weighted moving average and local regression algorithms is used to perform trend fitting and high-frequency noise suppression on construction temperature, steel reinforcement stress, and stress values ​​of stressed components. Finally, a cascaded transformation using power function transformation and max-min normalization algorithms is employed to adjust the distribution and compress the numerical values ​​of water conservancy construction monitoring data, achieving unified interval reconstruction and normalization of the data.

4. The method for monitoring and supervising construction monitoring data of water conservancy projects by multi-source data fusion according to claim 1, characterized in that: The specific steps for dividing the structure into different disturbance state segments based on the degree of structural disturbance and marking the segment state labels for each time period are as follows: The system calculates the structural disturbance assessment value within each sliding window in real time and compares it with the structural disturbance threshold, which includes a primary structural disturbance threshold and a secondary structural disturbance threshold. When the structural disturbance assessment value is less than or equal to the primary structural disturbance threshold, the current time period is marked as a steady-state segment, and the baseline update strategy is automatically triggered. The original water conservancy construction monitoring data within the current window is incrementally updated in the steady-state sample set, and statistical features are archived. When the structural disturbance assessment value is greater than the primary structural disturbance threshold but less than the secondary structural disturbance threshold, the current time period is marked as a transition segment. The delayed confirmation mechanism is activated, all water conservancy construction monitoring data within the current window is retained, and a time buffer is set for subsequent state evolution trend judgment. When the structural disturbance assessment value is greater than or equal to the secondary structural disturbance threshold, the current time period is marked as an unstable segment. The high-frequency resampling logic is invoked, and the enhanced sampling mode is enabled for the covered monitoring points in subsequent time periods. At the same time, the water conservancy construction monitoring data within the current sliding window is pushed to the anomaly identification and processing flow. Each time period is marked as a segment status label and attached to the metadata of the current sliding window. This label contains the status category, time segment range and monitoring point number, and is retained along with the time index for use as an input index and scheduling basis for subsequent identification processes.

5. The method for monitoring and supervising construction monitoring data of water conservancy projects by multi-source data fusion according to claim 1, characterized in that: The specific steps for classifying anomalies based on dynamic imbalance indicators, and updating corresponding scheduling strategies and data collection configurations are as follows: After calculating the dynamic imbalance value, the dynamic imbalance value and the imbalance threshold are compared in real time to classify high-confidence anomaly segments into levels. The imbalance threshold includes a first-level imbalance threshold and a second-level imbalance threshold. When the dynamic imbalance value is less than or equal to the first-level imbalance threshold, the current high-confidence anomaly segment is marked as a slight disturbance, triggering local resampling logic and performing encrypted sampling on the current monitoring point. When the dynamic imbalance value is greater than the first-level imbalance threshold but less than the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a critical disturbance, the current monitoring point is subjected to encrypted sampling, and the water conservancy monitoring data corresponding to the current high-confidence anomaly segment is pushed to the supervision platform with an additional manual review prompt. When the dynamic imbalance value is greater than or equal to the second-level imbalance threshold, the current high-confidence anomaly segment is marked as a severe disturbance, triggering a station-wide alarm process, suspending related construction tasks, and transferring the segment data to the anomaly tracing and processing process. The dynamic imbalance value, anomaly level label, scheduling and processing record and corresponding time index are archived to generate an anomaly event log, which is bound to the original water conservancy construction monitoring data to achieve full-process index traceability. The anomaly level, dynamic imbalance value, structural disturbance assessment value and structural response deviation value marked in the anomaly event log are synchronously written into the acquisition scheduling configuration table to update the sampling frequency setting, monitoring point priority sorting and comparison threshold adjustment scheme, realize adaptive scheduling optimization based on actual anomaly evolution characteristics, and complete the closed-loop process of monitoring identification, response control and acquisition strategy.

6. A multi-source data fusion system for monitoring and supervising construction data in water conservancy projects, characterized in that: include: The system includes a water conservancy construction monitoring data acquisition and preprocessing module, a structural disturbance assessment and status marking module, a disturbance identification and anomaly sample extraction module, and an anomaly analysis, judgment, and scheduling linkage module, among which: The water conservancy construction monitoring data acquisition and preprocessing module is used to acquire water conservancy construction monitoring data by real-time data acquisition and synchronous processing of multiple types of sensing devices in the construction monitoring system, and to preprocess the water conservancy construction monitoring data to obtain preprocessed water conservancy construction monitoring data. The structural disturbance assessment and status labeling module is used to construct a sliding window based on preprocessed hydraulic construction monitoring data, extract key structural and environmental variables, comprehensively analyze the degree of structural disturbance, divide different disturbance state segments according to the degree of structural disturbance, and label the segment status labels for each time period. The disturbance identification and abnormal sample extraction module is used to extract unstable time periods based on segment state labels, construct disturbance feature sequences, identify highly suspected disturbance segments and analyze the response deviation between the structure and the environment, combine the three-dimensional structural response sequence with historical steady-state samples for sliding comparison, screen highly reliable abnormal segments and generate abnormal data structures. The anomaly analysis and scheduling linkage module is used to jointly analyze the coupling relationship between structural response and environmental disturbance based on the identified abnormal data structure, construct a dynamic imbalance index to measure the intensity of non-steady state, classify the anomaly level according to the dynamic imbalance index, and update the corresponding scheduling strategy and data acquisition configuration.

Images