A construction engineering supervision site data intelligent acquisition and abnormal analysis platform

Abstract

This invention discloses an intelligent data acquisition and anomaly analysis platform for construction engineering supervision, relating to the field of construction engineering supervision technology. The platform includes a site perception layer, a data transmission layer, and a data processing layer. The perception layer acquires electrical and attitude data of the vibrating equipment, original spatial coordinates, ambient temperature, and concrete material parameters. The data processing layer performs spatial drift compensation based on density clustering on the original coordinates to obtain calibration coordinates and calculates dynamic operation duration thresholds based on material parameters and ambient temperature. The system extracts equivalent mechanical impedance and attitude validity features, and performs cascaded anomaly judgments including attitude compliance, physical duration, and process form based on a dynamic time warping algorithm through a logic verification engine. If the verification passes, the data is associated with BIM components using the calibration coordinates. This invention eliminates interference from environmental and material differences, achieving accurate quantitative assessment of the quality of engineering operations.

Description

Technical Field

[0001] This invention relates to the field of construction engineering supervision technology, specifically to an intelligent data acquisition and anomaly analysis platform for construction engineering supervision. Background Technology

[0002] In the construction process, concrete pouring is a critical hidden project. Its vibration quality directly determines the density and final strength of structural components. Traditional supervision mainly relies on on-site visual observation and experience judgment by on-site personnel. This method is limited by human resources and subjective factors, making it difficult to achieve all-weather, all-coverage objective supervision. With the application of Internet of Things (IoT) technology, some quality management systems based on intelligent monitoring terminals have emerged in the industry. These systems mainly collect the operating current or vibration status of the vibration equipment to calculate the operation time, attempting to replace manual recording in a digital way.

[0003] However, existing monitoring technologies still have limitations in practical engineering applications. In construction sites with dense steel reinforcement and complex electromagnetic environments, the coordinate data output by spatial positioning devices often exhibit random drift and jumps due to the influence of signal multipath effects and non-line-of-sight propagation. This makes it difficult for the system to accurately map discrete operation data to specific building information model components, resulting in misaligned quality data archiving and the inability to form an accurate spatial traceability chain.

[0004] Furthermore, existing automatic judgment logic typically uses a single, fixed time threshold to determine whether a work operation is qualified, failing to consider the impact of significant changes in ambient temperature and batch slump fluctuations on the rheological properties of the material. Under low-temperature or low-slump conditions, applying the conventional time standard leads to actual under-vibration; while under high-temperature or high-flow-rate conditions, it causes over-vibration and segregation. This static judgment standard lacks dynamic adaptability. Simultaneously, shallow monitoring relying solely on equipment start-stop status cannot deeply identify specific work process forms, making it difficult to effectively distinguish between compliant vertical insertion vibration and ineffective horizontal dragging, using equipment to pry up reinforcing bars, or idling in the air—violations that result in a large amount of invalid or harmful work data being misjudged as qualified by the system, making it difficult to truly guarantee the physical quality of concealed works. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent data acquisition and anomaly analysis platform for construction engineering supervision, which solves the problems of being unable to dynamically adjust work standards according to environmental and material differences and the difficulty in identifying violations.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent data acquisition and anomaly analysis platform for construction engineering supervision, comprising: This invention provides an intelligent data acquisition and anomaly analysis platform for construction engineering supervision, which includes a site perception layer, a data transmission layer, and a data processing layer.

[0007] The site perception layer is configured to acquire multi-source heterogeneous physical data of the construction site. The multi-source heterogeneous physical data includes electrical data and motion posture data of the handheld vibrator, original three-dimensional spatial coordinate data of the target object, site ambient temperature data, and concrete material conveying parameters.

[0008] The data transmission layer is configured to perform time synchronization and frequency unification processing on the multi-source heterogeneous physical data, generate a full-dimensional state vector aligned on the time axis, and upload it to the data processing layer.

[0009] The data processing layer is configured to retrieve a pre-stored building information model and receive the full-dimensional state vector, and then perform the following processing: Spatial drift compensation is performed on the original three-dimensional spatial coordinate data to obtain calibration coordinates; Calculate the dynamic operation duration threshold under the current working conditions based on the material conveying parameters and the ambient temperature data; Equivalent mechanical impedance features and attitude effectiveness features are extracted from the electrical data and motion attitude data; The system determines the current operation status and generates anomaly determination results based on the calibration coordinates, the dynamic operation duration threshold, the equivalent mechanical impedance characteristics, and the attitude validity characteristics.

[0010] The data transmission layer includes an edge computing gateway, which sets a uniform system resampling frequency and uses the discrete time step of this frequency as the reference time axis. For raw three-dimensional spatial coordinate data with an original sampling frequency lower than the system resampling frequency, the edge computing gateway performs linear interpolation processing. For electrical data whose original sampling frequency is higher than the system resampling frequency, the edge computing gateway performs downsampling processing by calculating the effective value or average value, thereby generating a full-dimensional state vector under a unified timestamp.

[0011] The data processing layer utilizes a data preprocessing module to perform density-based clustering spatial drift compensation. The data preprocessing module selects a set of valid work points within a time window, uses a density-based noise-based spatial clustering algorithm to extract the observation centroid of the largest cluster, and searches for the geometric center of the component closest to this observation centroid in the building information model.

[0012] When the Euclidean distance between the two is less than the preset drift radius, the data preprocessing module calculates the drift compensation vector pointing from the observed centroid to the geometric center of the component, and uses this vector to perform translation correction on the original three-dimensional spatial coordinate data to obtain the calibration coordinates.

[0013] The data processing layer uses a physical parameter calculation module to calculate the dynamic operation duration threshold. This module retrieves pre-stored baseline operation parameters, including the baseline liquefaction time, standard slump constant, and standard temperature constant under standard operating conditions. During the calculation, the module calculates a slump correction factor based on the difference between the standard slump constant and the current batch's designed slump value, using a pre-stored slump correction coefficient. The calculation logic for this factor is to multiply the difference by the slump correction coefficient and then sum it with the unit value. In addition, the physical parameter calculation module calculates the temperature correction factor based on the difference between the standard temperature constant and the current ambient temperature data, using a pre-stored temperature correction coefficient. The calculation logic is to multiply the difference by the temperature correction coefficient and then sum it with the unit value.

[0014] Finally, the physical parameter calculation module multiplies the baseline liquefaction time by the slump correction factor and the temperature correction factor in sequence to obtain the dynamic operation time threshold.

[0015] The data processing layer uses a feature extraction module to calculate the original impedance value at each time step by dividing the input voltage by the load current, and then performs a sliding window low-pass filtering on the value to obtain the equivalent mechanical impedance characteristics.

[0016] Meanwhile, the feature extraction module extracts attitude validity features from the motion attitude data. These features are represented by a binary mask: when the angle between the device's main axis and the opposite direction of the gravity vector is within the maximum allowable tilt angle range, it is marked as a valid state; otherwise, it is marked as an invalid state.

[0017] In addition, the feature extraction module calculates the weighted impedance change rate feature, that is, it obtains the second derivative of the equivalent machine impedance feature with respect to time, and uses the attitude validity mask to perform gating filtering on the second derivative. The peak value of the weighted impedance change rate feature is used to characterize the liquefaction phase transition point of concrete.

[0018] The data processing layer utilizes a logic verification engine to perform compliance determinations based on a dynamic time warping algorithm.

[0019] The logic verification engine extracts the measured impedance sequence of a single operation interval and retrieves the pre-stored standard process template sequence. Then, it constructs a cumulative distance matrix of data points between the measured sequence and the standard template sequence, searches for an optimal curved path from the starting point to the ending point in the matrix, and determines the minimum cumulative Euclidean distance corresponding to this path as the DTW distance.

[0020] The logic verification engine executes cascading exception determination logic with priority.

[0021] The first level of posture compliance is verified by calculating the percentage of posture validity characteristics within the work area that indicate a valid state to determine posture violations and anomalies. The second level of verification is physical duration, which determines under-vibration anomalies by comparing the actual operation duration with the dynamic operation duration threshold. The third level of process verification determines process non-compliance anomalies by comparing the DTW distance with the upper limit of waveform similarity tolerance.

[0022] If all level verifications pass, the logical verification engine determines it as a normal operation, uses the calibration coordinates to locate the component ID to which the current operation belongs in the building information model, associates the operation data with the component ID, and stores it in the database.

[0023] Acquire multi-source heterogeneous physical data from the construction site, including electrical and motion posture data of handheld vibratory equipment, original three-dimensional spatial coordinate data of the target object, ambient temperature data, and concrete material conveying parameters. The multi-source heterogeneous physical data is synchronized in time and frequency to generate an aligned full-dimensional state vector. Retrieve the pre-stored building information model and perform spatial drift compensation on the original three-dimensional spatial coordinate data to obtain the calibration coordinates; Calculate the dynamic operation duration threshold based on material conveying parameters and ambient temperature data; Extract equivalent electromechanical impedance features and attitude effectiveness features from electrical data and motion attitude data; Based on the calibration coordinates, dynamic operation duration threshold, equivalent mechanical impedance characteristics, and attitude effectiveness characteristics, multidimensional anomaly analysis is performed to generate anomaly determination results.

[0024] The method can eliminate the interference of environmental differences, power grid fluctuations and human operating habits on quality judgment through physical parameter calculation and multi-dimensional feature extraction, and realize the quantitative evaluation of engineering operation quality.

[0025] This invention provides an intelligent data acquisition and anomaly analysis platform for construction engineering supervision. It has the following beneficial effects: 1. This invention uses a density-based noise spatial clustering algorithm to process the original three-dimensional coordinates, extracts the centroid of the observation and matches it with the geometric center of the component in the building information model, and then calculates the drift compensation vector. This processing method can correct the random drift error of the on-site positioning equipment, ensure that the operation data can be accurately mapped to the specific digital model component, solve the problem of deviation between the spatial position and the actual component when the engineering supervision data is archived, and realize the accurate traceability of construction data.

[0026] 2. This invention introduces a dynamic threshold calculation mechanism based on ambient temperature and material conveying parameters, and uses slump correction factors and temperature correction factors to adjust the benchmark liquefaction time in real time. This allows the quality judgment standard to adapt to the changing environmental conditions and batch differences of concrete at the construction site, avoiding misjudgments caused by using a single fixed time standard in extreme weather or material changes, and ensuring the objectivity and accuracy of quality assessment under different temperature and concrete fluidity conditions.

[0027] 3. This invention employs a cascaded anomaly judgment logic based on posture compliance, physical duration, and process form, and combines a dynamic time warping algorithm to compare the measured impedance sequence with a standard process template. By eliminating invalid work intervals with excessive equipment tilt angles and performing nonlinear similarity matching on the work waveforms, this solution can effectively identify abnormal operations such as under-vibration, illegal horizontal dragging, or using equipment to pry up steel bars, thus realizing a shift from simple work duration assessment to full-process process standardization assessment. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the overall architecture of a smart data acquisition and anomaly analysis platform for construction engineering supervision according to the present invention. Figure 2 This is a timing logic diagram illustrating the synchronous acquisition and timing alignment of multi-source heterogeneous data in an embodiment of the present invention; Figure 3 This is a schematic diagram of the spatial drift compensation process based on density clustering in this invention; Figure 4 This is a schematic diagram of the logical flow of the theoretical liquefaction time threshold calculation model of this invention; Figure 5 This is a schematic diagram of the equivalent impedance calculation and attitude validity gating logic of the present invention; Figure 6 This is a schematic diagram of the multidimensional logical anomaly detection process based on dynamic time warping of the present invention; Figure 7 This is a real-time impedance monitoring diagram for field operations in an application embodiment of the present invention; Figure 8 This is a diagram illustrating the compliance analysis of the posture of intelligent machinery in an application embodiment of the present invention. Figure 9 This is a logic verification engine decision diagram in an application embodiment of the present invention.

[0029] The system comprises: 110, Field Perception Layer; 111, Intelligent Machinery Monitoring Terminal; 1111, Current Sensor; 1112, Voltage Sensor; 1113, Inertial Measurement Unit; 112, Spatial Positioning Tag; 113, Environmental Monitoring Station; 114, Material Conveying Monitoring Unit; 120, Data Transmission Layer; 121, Edge Computing Gateway; 130, Data Processing Layer; 131, Data Preprocessing Module; 132, Physical Parameter Calculation Module; 133, Feature Extraction Module; and 134, Logic Verification Engine. Detailed Implementation

[0030] The technical solutions in 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.

[0031] See attached document Figure 1 The present invention provides a construction engineering supervision on-site data anomaly analysis platform, comprising: an on-site perception layer 110, a data transmission layer 120, and a data processing layer 130.

[0032] The on-site perception layer 110 is used to acquire multi-source heterogeneous physical data of the construction site, including: intelligent machinery monitoring terminal 111, spatial positioning tag 112, environmental monitoring station 113, and material conveying monitoring unit 114.

[0033] The intelligent machinery monitoring terminal 111 is installed on the handheld vibrating equipment. The terminal integrates a current sensor 1111, configured to collect the load current of the vibrating equipment. Voltage sensor 1112 is configured to collect the input voltage of the vibrating equipment. ; and an inertial measurement unit 1113, configured to collect the triaxial vibration acceleration of the vibrating equipment. and posture Euler angle .

[0034] Spatial positioning tag 112 is fixed to the safety helmet of construction workers or the body of equipment, and is configured to output the target object in discrete time steps via ultra-wideband UWB or real-time dynamic differential RTK technology. The original three-dimensional spatial coordinates below .

[0035] Environmental monitoring station 113 is configured to collect real-time ambient temperature data at the construction site. And relative humidity data. The material conveying monitoring unit 114 is configured to acquire the conveying status of concrete and batch design parameters.

[0036] The data transmission layer 120 includes an edge computing gateway 121. The edge computing gateway 121 is communicatively connected to each device in the field sensing layer 110 and configured to synchronize the time of multiple acquired data streams and unify the sampling frequency. The synchronized data is then uploaded to the data processing layer 130.

[0037] The data processing layer 130 includes: a data preprocessing module 131, a physical parameter calculation module 132, a feature extraction module 133, and a logic verification engine 134. The data processing layer 130 performs the following processing flow to achieve logical verification of the project quality.

[0038] Data preprocessing module 131 is configured to receive a full-dimensional state vector containing current, voltage, acceleration, attitude angle, and spatial coordinates. And perform spatial drift compensation based on density clustering. Data preprocessing module 131 selects a time window. Effective work point set within The DBSCAN algorithm is used to extract the centroid of the largest cluster. .

[0039] Data preprocessing module 131 searches for distances in Building Information Modeling (BIM). Nearest component geometric center When the Euclidean distance between the two is less than the preset drift radius At that time, the data preprocessing module 131 calculates the drift compensation vector. : ; The data preprocessing module 131 uses this vector to correct the original coordinates and obtain calibrated coordinates. : ; Data preprocessing module 131 utilizes calibration coordinates Determine the BIM component ID to which the current task belongs. The physical parameter calculation module 132 is configured to calculate the dynamic task duration threshold based on the rheological properties of concrete. The physical parameter calculation module 132 reads the design slump of the current batch of concrete. and ambient temperature provided by environmental monitoring station 113 Calculated based on the following formula : ; in, The reference liquefaction time under standard operating conditions. The standard slump constant, The standard temperature constant, This is the slump correction factor. This is the temperature correction factor.

[0040] Feature extraction module 133 is configured as the equivalent mechanical impedance of the computing device. and pose validity mask Equivalent mechanical impedance The calculation is as follows: ; Pose validity mask Used to remove violations of tilting operations, defined as follows: ; in, The angle between the main axis of the equipment and the opposite direction of the gravity vector. The maximum permissible tilt angle. Feature extraction module 133 further calculates the weighted impedance change rate feature. Used to characterize the liquefaction phase transition point of concrete: ; in, For partial differential operators, This indicates a time variable. This represents a second-order operation on the time dimension. The logic verification engine 134 is configured to perform compliance checks on the work process based on the Dynamic Time Warping (DTW) algorithm. The logic verification engine 134 extracts the impedance sequence for a single work interval. And retrieve the pre-stored standard process template sequence. Calculate the DTW distance between the two. .

[0041] The logic verification engine 134 generates anomaly determination results based on the following logic. : ; in, For sequence length, The effective pose percentage threshold, It is the starting point of the event. It is the end point of the event. This represents the upper limit of waveform similarity tolerance. The type corresponds to a pose violation. Corresponding to under-vibration, type The corresponding process is not compatible.

[0042] The data processing layer 130 will generate the anomaly determination results. It is associated with the corresponding BIM component ID and stored in the database.

[0043] See attached document Figure 2In this embodiment of the invention, the synchronous acquisition mechanism of the four-dimensional data stream is specifically executed collaboratively by sensing devices distributed at the construction site and edge computing gateways.

[0044] The intelligent machinery monitoring terminal 111 conditions sensor signals through its internally integrated analog front-end (AFE) circuit. The current sensor 1111 is a closed-loop Hall effect sensor, mounted on the single-phase power line of the vibrating equipment, configured to convert the large current signal of the main circuit into an analog voltage signal of 0 to 5V. The voltage sensor 1112 is a precision voltage transformer connected in parallel to the power input, configured to proportionally step down the input voltage waveform. The microcontroller unit (MCU) uses an analog-to-digital converter (ADC) at a preset sampling frequency. (e.g., 1000Hz) Synchronous sampling of analog current and voltage signals ensures that subtle waveform changes and transient impacts of the power grid frequency cycle can be captured.

[0045] Meanwhile, the inertial measurement unit 1113 inside the intelligent machine monitoring terminal 111 uses frequency (e.g., 100Hz) Continuously outputs raw triaxial acceleration data and quaternions calculated by the internal digital motion processor (DMP). The microcontroller unit (MCU) converts the quaternions into Euler angles, namely roll, pitch, and yaw, where the pitch angle is denoted as the device's tilt relative to the vertical line of gravity. The MCU packages the electrical data (current, voltage) with the motion data (acceleration, attitude) and sends them to the edge computing gateway 121 via LoRa or Wi-Fi industrial protocols.

[0046] Spatial positioning tag 112 is configured to use frequency (Typically 1Hz to 10Hz, lower than the sampling rate of the equipment data) Outputs positioning data packets. For UWB positioning systems, the tag communicates bidirectionally with the base station deployed on-site via TWR (Time-to-Wide) ranging to calculate the tag's three-dimensional coordinates in the local coordinate system. For RTK-GPS systems, the tag receives satellite signals and differential data from the base station, outputs latitude, longitude, and elevation in the global geodetic coordinate system, and is converted into local engineering coordinates by the gateway. The edge computing gateway 121 is configured as a data stream aggregation and timing alignment center. The gateway runs a time synchronization service based on Network Time Protocol (NTP) or Precision Time Protocol (PTP), periodically sending time synchronization commands to the intelligent equipment monitoring terminal 111 and the spatial positioning tag 112 to ensure that the timestamp deviation of all front-end devices is maintained within the millisecond error range.

[0047] Edge computing gateway 121 is configured with several discrete time steps A circular buffer is formed. Due to the different original sampling frequencies of the sensors ( If there is a discrepancy, the gateway will perform multi-rate signal processing and resampling logic. The gateway will set a uniform system resampling frequency. (This frequency is preferably the same as the frequency of the inertial measurement unit) Consistent, for example, 100Hz).

[0048] The gateway uses the system resampling frequency The discrete time step is used as the reference time axis: For spatial positioning data with a low sampling frequency, a linear interpolation algorithm is used to calculate the interpolated coordinates corresponding to the reference time. For electrical waveform data with extremely high sampling frequencies, a sliding window is used to calculate its time period corresponding to the current time step (i.e., The effective RMS or average value within the specified time period is downsampled to generate a unified timestamp. The alignment state vector below This mechanism ensures that each frame of current characteristics strictly corresponds to a unique spatial location coordinate and machine attitude in subsequent processing, eliminating spatiotemporal misalignment caused by data transmission delay or clock drift, while also avoiding excessive redundancy in data dimensions.

[0049] See attached document Figure 3 In this embodiment of the invention, the data preprocessing module 131 eliminates the system error of the positioning system and the registration deviation between the BIM model and the physical site by executing the compensation algorithm.

[0050] Data preprocessing module 131 first constructs a time sliding window. And filter the set of valid job points within this window. The filtering logic is based on the machine's load current. and vibration acceleration The original spatial coordinates at that moment are only determined when the current intensity exceeds the preset no-load threshold and the acceleration amplitude indicates that the machine is in the open state. Only then was it added to the set of valid work points. This screening step is used to eliminate discrete trajectory data generated by construction workers during the movement, rest, or preparation phases, ensuring that subsequent cluster analysis focuses only on the actual work location.

[0051] For the selected set of valid work points The data preprocessing module 131 executes the density-based noise-applied spatial clustering algorithm DBSCAN. This algorithm sets two hyperparameters: neighborhood radius and neighborhood radius. and minimum number of contained points The algorithm iterates through each point in the set and calculates its... Point density within a neighborhood. If the number of points in the neighborhood of a given point is greater than... If a point is identified as a core point, it is used as the basis for expanding outwards to form a cluster. Through this clustering process, the algorithm automatically filters out drifting isolated points (noise points) caused by multipath effects and aggregates high-density coordinate points into several clusters. The data preprocessing module 131 selects the cluster containing the most points as the main processing cluster. The arithmetic mean of all coordinate points in the cluster is calculated to obtain the observed centroid within the current time window. .

[0052] Subsequently, the data preprocessing module 131 performs a spatial proximity search in the Building Information Modeling (BIM) database. The module iterates through all components (such as columns, walls, and beams) within the current floor, obtaining the theoretical geometric center coordinates of each component. The module then calculates the observed centroid. With respect to the theoretical geometric center of each component The Euclidean distance between them is used to identify the nearest target component.

[0053] In this process, the system introduces a drift radius constraint. (For example, set to 50 cm). Only if the calculated minimum Euclidean distance is less than this drift radius constraint. Only then does the data preprocessing module 131 determine that the match is valid and consider the current coordinate deviation to be a calibrable systematic drift, rather than an erroneous position identification. If the calculated minimum Euclidean distance is greater than the drift radius constraint... If the current work position deviates from all known components (possibly due to testing in a non-work area), the data preprocessing module 131 will stop subsequent drift compensation and mark the current data frame as an unassociated component, which will not participate in subsequent quality archiving.

[0054] After obtaining the drift compensation vector, the data preprocessing module 131 applies the vector to the current time window. All original coordinate data within. The module performs a coordinate translation transformation, converting the original coordinates... Add drift compensation vector Generate calibrated original coordinates This step allows discrete work points that might have previously fallen outside the electronic fence boundary of a BIM component to be shifted and snapped into the correct BIM component space, thus ensuring that the subsequent logic verification engine can accurately associate the work data with the correct building component ID.

[0055] In this embodiment of the invention, the physical parameter calculation module 132 is responsible for establishing a digital mapping relationship between the operating parameters of the physical world and the input variables of the algorithm model, providing a quantitative benchmark for subsequent dynamic threshold calculation.

[0056] The physical parameter calculation module 132 first establishes a data connection with the material conveying monitoring unit 114 via a communication interface. The material conveying monitoring unit 114 uses RFID technology or QR code optical recognition technology to read the electronic tag information on each concrete transport document. This unit analyzes the key rheological influencing factors of the current pouring batch, specifically including the concrete strength grade (e.g., C30, C40) and the design slump value. Slump value, expressed in millimeters (mm), characterizes the fluidity and yield stress of concrete mixture under its own weight.

[0057] To achieve standardized parameter mapping, the physical parameter calculation module 132 has a pre-built material property database. This database stores the rheological property benchmarks of concrete of different strength grades under standard laboratory conditions. The module uses the read strength grade as an index key to retrieve the corresponding benchmark slump constant from the database. (For example, set to 160mm) and the base viscosity coefficient at this mixing ratio. If the actual design slump is read... It includes an allowed fluctuation range, and the module uses the median or lower limit of this range as a conservative calculation basis to convert it into a scalar input that the algorithm can recognize.

[0058] Meanwhile, the physical parameter calculation module 132 continuously receives real-time meteorological data streams from the environmental monitoring station 113. The environmental monitoring station 113 is deployed in the near-field area of ​​the construction work surface and is configured to output ambient temperature data at a minute-level update frequency. And relative humidity. The module performs a moving average filter on the collected temperature series to eliminate data fluctuations caused by gusts or temporary obstructions.

[0059] After completing the acquisition and cleaning of the raw data, the physical parameter calculation module 132 performs parameter normalization mapping. The module sets a standard temperature constant. (For example, 20 degrees Celsius) is used as the thermodynamic zero point. The module calculates the ambient temperature. With standard temperature constant The deviation value is used to establish a temperature influence factor mapping; at the same time, the actual design slump is calculated. With standard slump constant The difference is used to establish a flowability influencing factor mapping. Through this mapping mechanism, the physical parameter calculation module 132 transforms the originally discrete, qualitative working condition descriptions (such as low temperature environment, high-grade dry-hard concrete) into continuous, dimensionless numerical deviation vectors, and inputs these vectors into subsequent physical model calculation units in real time, ensuring that the quality verification standard at each moment can accurately match the current specific working condition. In addition, the module also binds these parameters with the current BIM construction task ID to form a construction log record with environmental information association.

[0060] See attached document Figure 4 In this embodiment of the invention, the physical parameter calculation module 132 not only performs parameter mapping, but its core function is to run the calculation model to determine the lower limit of the physical time required for concrete to transform from solid accumulation to liquid compaction under the current specific working conditions.

[0061] The physical parameter calculation module 132 first retrieves the reference liquefaction time that matches the model of the currently used intelligent vibrating equipment. This reference time is based on standard laboratory conditions (i.e., standard temperature). and standard slump Under certain conditions, this refers to the empirical time required to bring a unit volume of concrete to a fully liquefied state by testing a vibratory compactor with a specific power and frequency. This value is stored in the equipment's firmware or a cloud configuration library as the scalar basis for calculations.

[0062] Next, the physical parameter calculation module 132 calculates the slump correction factor. This factor aims to compensate for the impact of the initial fluidity of concrete on the difficulty of vibration. The physical parameter calculation module 132 is based on the formula... Perform the calculation. Among them, This is the slump correction factor, and its unit is... This coefficient is a constant obtained through linear regression analysis of historical experimental data, used to quantify the impact of slump variation on vibration time. When the actual design slump... Less than the standard slump constant When the concrete is drier and harder with higher viscosity, the difference is positive, and the slump correction factor is calculated. It is greater than 1, thus extending the theoretical threshold based on the reference time.

[0063] Meanwhile, the physical parameter calculation module calculates the temperature correction factor. This factor aims to compensate for the influence of environmental thermodynamic conditions on the viscosity coefficient of cement paste. The physical parameter calculation module 132 is based on the formula... Perform the calculation. Among them, the coefficients... This is the temperature correction factor, and its unit is... This coefficient is also a constant obtained through calibration analysis of vibration test data at different temperatures. When the ambient temperature... Below the standard temperature constant When the difference is positive, the temperature correction factor... Greater than 1. This step reflects the physical characteristic of increased internal friction in materials at low temperatures, and the system determines that the vibration time needs to be increased accordingly.

[0064] The final physical parameter calculation module 132 uses a multiplicative model to synthesize the above two factors and the reference time to generate the final dynamic operation duration threshold. The calculation logic is based on the baseline liquefaction time. Multiply by slump correction factor in sequence Right now The computational model is also configured with boundary constraint logic, that is, when the calculated boundary constraint logic is applied... When the time drops below a preset engineering safety threshold (e.g., 5 seconds), the physical parameter calculation module 132 forcibly... This value is assigned as the safety baseline to prevent the threshold from becoming too low due to abnormal parameters. The calculation is complete. It is transmitted to the logic verification engine as a dynamic standard for determining under-vibration anomalies.

[0065] See attached document Figure 5 In this embodiment of the invention, the feature extraction module 133 performs this step in order to decouple the real dielectric load features from the original electrical signal and use spatial geometric constraints to eliminate invalid and non-compliant operation data.

[0066] Feature extraction module 133 first processes the synchronously acquired voltage. and current Time series analysis. Given that construction sites typically use temporary power networks, the start-up and shutdown of large equipment (such as tower cranes and elevators) can cause significant transient fluctuations in the grid voltage. To eliminate the interference of these power supply fluctuations on load determination, the feature extraction module 133 does not directly use the current value as a feature, but instead performs a division operation to calculate the equivalent mechanical impedance value at each time step in real time. The calculation logic uses the input voltage as the dividend and the load current as the divisor. Through this calculation, the system obtains a physical quantity that is independent of the supply voltage amplitude, which directly reflects the magnitude of the mechanical damping experienced by the rotor of the vibratory tamper motor in the concrete medium. To further improve the signal-to-noise ratio, the module performs a sliding window low-pass filter on the calculated original impedance sequence to filter out high-frequency electrical noise caused by motor commutation or poor contact, generating a smooth impedance envelope.

[0067] Meanwhile, the feature extraction module 133 processes the attitude data from the inertial measurement unit in parallel. The module reads the device's current attitude Euler angles. The system analyzes the pitch component and defines a virtual effective working cone space centered on the gravity vector. The semi-apex angle of this cone is the maximum allowable tilt angle. (For example, set to 15 degrees). The feature extraction module 133 calculates in real time the angle between the main axis direction of the device and the opposite direction of gravity. and compare it with the maximum permissible tilt angle. Compare them.

[0068] Based on the above comparison results, the feature extraction module 133 generates a time-varying binary pose validity mask. When the calculated included angle Less than or equal to the maximum permissible tilt angle When the vibrator is inserted vertically or nearly vertically, it indicates that the vibrator is in a vertical or near-vertical insertion state, which conforms to the concrete vibration process specification. At this point, the binarized attitude validity mask is used. It is assigned a value of 1. Conversely, when the angle is... Greater than the maximum permissible tilt angle This indicates that the equipment is in a significantly tilted state. This situation typically corresponds to construction workers using vibrators to pry up the steel mesh, dragging the equipment horizontally to move materials on the surface, or the equipment being placed horizontally on scaffolding and running idle. In these scenarios, although the motor may generate high load current (such as the stall current when vibrating steel bars), it is not an effective concrete compaction operation. At this time, the binarized attitude validity mask... It was forcibly assigned the value 0.

[0069] Finally, the feature extraction module 133 calculates the equivalent mechanical impedance. With binarized pose validity mask A point-by-point multiplication operation is performed, outputting a gated and filtered effective impedance sequence. This mechanism ensures that subsequent feature analysis is performed only on geometrically compliant vertical insertion actions, eliminating interference from data generated by non-compliant operations on the quality assessment results.

[0070] In this embodiment of the invention, the feature extraction module 133 further performs in-depth analysis on the effective impedance sequence after attitude gating filtering, aiming to identify the precise physical moment when the internal structure of the concrete medium transforms from a solid flocculated state to a liquid leveled state.

[0071] Feature extraction module 133 first processes the discrete equivalent impedance sequence Perform time-domain differentiation. Given that high-frequency quantization noise that may remain in the original signal will be amplified during the differentiation process, the module uses a five-point or seven-point Savitzky-Golay smoothing differentiation filter to suppress noise while calculating the first derivative (i.e., the rate of change of impedance) and the second derivative (i.e., the acceleration of the impedance change) of the impedance with respect to time.

[0072] Module based on formula Generate weighted impedance change rate characteristics In this calculation process, the pose validity mask... To control the effectiveness of the calculation: the system calculates the second derivative feature only when the device maintains a vertical orientation; once the device tilts beyond a threshold, It is immediately assigned a value of zero. This mechanism prevents mechanical disturbances generated by mobile devices from being misinterpreted as media rheological characteristics. Feature extraction module 133 in weighted electrical impedance change rate feature Search for extreme points in the sequence. From a physical rheological perspective, when the vibrator is first inserted into the concrete, the medium is in a high-viscosity state, and the second derivative is close to zero. As vibration energy continues to be input, the yield stress inside the concrete decreases, and the medium liquefies. During this process, mechanical damping decreases, leading to an equivalent impedance. It shows an upward trend.

[0073] Weighted impedance change rate characteristics The characteristic peak corresponds to the moment when the slope of the impedance curve changes the most, i.e., the critical point where the internal structure of the concrete undergoes significant changes. The feature extraction module 133 detects the occurrence time and peak intensity of this characteristic peak. If a characteristic peak meeting the preset strength threshold is detected within a single operation cycle, the module marks this moment as the liquefaction initiation point. The presence of a significant characteristic peak indicates that the concrete has entered a fluid state; if no significant characteristic peak is detected throughout the process, it indicates that the expected physical phase transition has not occurred. Through second-order derivative analysis, the system achieves indirect monitoring of the internal physical state of the engineering structure.

[0074] See attached document Figure 6 In this embodiment of the invention, the logic verification engine 134 performs this step by flexibly matching the measured operation data with the standard process model to achieve a comprehensive judgment on the operation quality.

[0075] The logic verification engine 134 first executes the job event slice. The engine continuously monitors the effective impedance sequence after attitude gating. When the impedance value exceeds the preset air impedance threshold and the machine acceleration indicates that it is in the open state, it marks the event start point. The event is marked as ending when the impedance value falls below the baseline or the equipment is shut down. The engine extracts this time interval. Construct the detection sequence using all available impedance data. .

[0076] To perform waveform comparison, the logic verification engine 134 maintains a standard process template library. This library stores typical impedance change curves recorded by experienced technicians performing standard vibration operations (i.e., vertical rapid insertion, slow withdrawal, and achieving a compacted state) at different concrete slump levels. Based on the physical parameters of the current batch of concrete (provided by the physical parameter calculation module 132), the engine indexes a matching standard reference sequence from the library. .

[0077] The logic verification engine 134 calculates the sequence to be detected. With standard reference sequence Dynamic time-warped distance between Given the inherent randomness in the speed of manual operations (e.g., varying insertion and withdrawal speeds), direct Euclidean distance calculation would lead to waveform misalignment on the time axis. The DTW algorithm constructs a cumulative distance matrix by non-linearly stretching or compressing the sequence on the time axis. The algorithm then searches the matrix for an optimal curved path from the start point to the end point that minimizes the cumulative Euclidean distance along that path. This minimum cumulative distance is... Its numerical value quantifies the degree of deviation of the measured operation from the standard process in terms of morphological trend, while ignoring the time dimension difference caused by the speed of the action.

[0078] After obtaining the above characteristic parameters, the logic verification engine 134 executes priority-based serial cascaded exception judgment logic to ensure the uniqueness and accuracy of the judgment result: The first level is attitude compliance verification. The engine calculates the attitude validity mask within the event interval. The proportion of time steps with a value of 1 to the total number of steps. If this proportion is lower than the preset effective pose percentage threshold... (For example, 90%), the engine directly determines the operation as Type I anomaly: posture violation and terminates subsequent judgments. This usually corresponds to the operator tilting the machine for an extended period of time to rush materials or pry steel bars.

[0079] The second level is physical duration verification. If the attitude verification passes, the engine calculates its actual duration. And compare it with the dynamic operation duration threshold generated by the physical parameter calculation module 132. Compare. If The engine was flagged as a Type II anomaly: under-vibration, and subsequent checks were terminated. This logic ensures that each operating point receives the minimum mechanical energy injection required to overcome the yield stress of the current medium.

[0080] The third level is process morphology verification. If the duration also meets the standard, the engine compares and calculates the DTW distance. Upper limit of waveform similarity tolerance .like The engine was identified as having a Type III anomaly: process non-compliance. This type covers atypical conditions such as the vibrator touching the reinforcing steel frame (causing a sudden spike in impedance), getting stuck in a void, or spinning freely in the air (causing the impedance curve to be flat and unchanged).

[0081] Only when all three levels of verification pass will the logic verification engine 134 determine the operation as normal, and, in conjunction with the coordinate information after spatial drift compensation, include the qualified data in the quality statistics file of the corresponding BIM component.

[0082] In this embodiment of the invention, after the logic verification engine 134 completes the judgment of a job event, the system generates a standardized structured data packet, namely a quality audit log, and persists it in the database for subsequent traceability.

[0083] The structured data packet consists of four core parts: metadata header, spatial context segment, physical environment segment, and diagnostic results segment.

[0084] The metadata header contains a globally unique identifier (UUID) for the event, used to uniquely identify the job event in the distributed system and prevent data duplication or conflicts. The metadata header also records the precise start and end timestamps of the event, with the time format conforming to the ISO 8601 standard and the precision retained to the millisecond level, ensuring the absolute traceability of the job sequence.

[0085] The spatial context stores the spatial information corrected by the data preprocessing module 131. This segment does not record the original noisy positioning coordinates, but rather the final coordinates calibrated by the density clustering drift compensation algorithm. More importantly, this section contains the Target_BIM_GUID field, which stores the globally unique IFC (Industry Base Class) ID of the BIM component locked by the geometric mapping algorithm. Through this field, a unique indexed association is established between the physical world's working data and the virtual model in the digital world.

[0086] The physical environment section records the external boundary conditions at the moment the operation occurs, i.e., the conditions collected by the physical parameter calculation module 132. and concrete design slump The purpose of recording these parameters is to preserve the calculation basis, so that the dynamic operation duration threshold can be verified during subsequent quality reviews. The rationality of it.

[0087] The diagnostic results section is the core of the data packet, containing the final judgment output by the logic verification engine. This section has enumerated status code fields, corresponding to normal, posture violation, under-vibration, and process non-compliance, respectively. In addition to qualitative status codes, this section also records quantitative process parameters in detail, including: actual effective operation time. The dynamic job duration threshold calculated by the system DTW distance between the measured waveform and the standard template And the attitude compliance rate. In addition, to support manual auditing, this section also includes a binary object (BLOB) field that stores the downsampled equivalent impedance corresponding to this event. Curves and attitude angles Curve data serves as the underlying electronic evidence for the judgment result.

[0088] The system uses lightweight data exchange formats (such as JSON or Protocol Buffers) to serialize and encode the above fields, and transmits them to the cloud server or local monitoring center through an encrypted channel. This structured definition ensures the uniformity of heterogeneous data at the storage level, supports fast SQL-based query retrieval (e.g., querying all historical vibration records of a certain column) and multi-dimensional quality analysis based on big data.

[0089] In this embodiment of the invention, the data processing layer 130 transmits the generated structured abnormal event data to the visualization display terminal, and realizes the transparent presentation of the internal quality status of the project through the graphics rendering engine.

[0090] The visualization terminal first loads the Building Information Model (BIM) file of the current construction area. This file conforms to the Industry Basic Class IFC standard and contains geometric, topological, and attribute information of the building components. The terminal then uses WebGL or a similar 3D graphics library to build a rendering pipeline, projecting the virtual building model onto the user interface.

[0091] The system establishes a separate quality data layer within the 3D scene, which is overlaid on the building model. The system reads the calibration space coordinates from each quality audit log entry. And exception status codes. For each discrete job event, the system generates a visual voxel point or sphere marker at the corresponding coordinate position in three-dimensional space.

[0092] The system executes a status code-based color mapping logic, modifying the RGBA channel values ​​of the rendered material to represent different quality states. For example, the system renders work points with normal results as green (RGBA:0,255,0,1.0); under-vibration points as red, indicating insufficient energy injection; attitude violation points as yellow, indicating non-standard operation; and process non-compliance points as purple. Through this spatially dense point cloud color distribution, the interface intuitively presents the vibration density and defect distribution inside the concrete component, forming a heat map reflecting the actual construction quality.

[0093] To achieve macro-level quality control at the component level, the system performs aggregation analysis based on BIM component IDs. The system counts the number of all valid normal operation points belonging to the same BIMGUID component and, combined with the component's volume attributes, calculates the actual vibration point density. If this density is lower than a preset process standard (e.g., it must contain [a certain amount of something] per cubic meter), [the system will then calculate the density at each vibration point]. When there are 1 effective vibration point, the system renders the model appearance of the component as a semi-transparent, highlighted warning state to alert the supervisor that there is an overall risk of vibration leakage in the area.

[0094] Furthermore, this visualization interface supports deep interactive functionality based on ray-casting. When the user's cursor hovers over or clicks on any specific work point marker in the 3D scene, the system initiates a query request to the backend database using the point's unique index (UUID). The system retrieves the binary large object (BLOB) data corresponding to this event, i.e., the impedance process curve. and attitude change curve The interface then displays the two waveforms in a floating window and marks the positions of the characteristic peaks automatically identified by the system. And the calculated DTW distance. This interactive mechanism allows supervisors to view microscopic single waveform records from macroscopic component color blocks, thereby enabling manual verification of disputed anomalies.

[0095] Specific application example: Monitoring the vibration quality of C40 concrete columns in cold environments. Project background and preconditions Project Location: Construction site of a commercial center project in northern China, during the main structure construction phase.

[0096] Target of the work: Frame columns in Zone C of the 3rd floor, with the unique BIM component identification code being COL-C3-05.

[0097] Monitoring period: November 15 (early winter).

[0098] Material parameters: Concrete strength grade: C40.

[0099] Design slump ( ): 140mm (belongs to low fluidity, dry-hard concrete, difficult to vibrate).

[0100] Standard slump constant ( ): 160mm.

[0101] Environmental parameters: The ambient temperature was measured by the on-site IoT monitoring station. : 5℃ (In low-temperature environments, the viscosity coefficient of concrete increases significantly).

[0102] Standard temperature constant ( ): 20℃.

[0103] System monitoring threshold settings: Reference liquefaction time ( ): 8.0 seconds.

[0104] Maximum permissible tilt angle ( ): 15°.

[0105] Slump correction factor ( ): 0.02s / mm.

[0106] Temperature correction factor ( ): 0.05s / ℃.

[0107] Detailed explanation of system operation process: Step 1: Calculation of physical parameters (dynamic threshold generation) Before the operation begins, the edge computing terminal calculates the dynamic operation duration threshold for that condition based on the current environment and material data. The calculation logic is as follows: Slump correction factor: ; (Note: The concrete is relatively dry and has high flow resistance, requiring a 40% extension of the working time.) Temperature correction factor: ; (Note: Low temperature increases the viscosity of the slurry, requiring a 75% extension of the operation time.) Dynamic job duration threshold: ; Conclusion: Under these specific working conditions of cold weather and relatively dry concrete, the attached... Figure 9 The baseline for judgment (vertical dotted line) is dynamically set to 19.6 seconds.

[0108] Step Two: Data Acquisition and Spatial Drift Compensation The monitoring system detected three consecutive vibration operations performed by the worker in the col-c3-05 column area. The system automatically completed spatial drift compensation by verifying the Euclidean distance between the UWB positioning coordinates and the centroid of the BIM model (d≈0.47m<0.5m), locking the data stream to the component.

[0109] Step 3: Logic Verification Engine Judgment and Attachment Analysis The system recorded three operation events (A, B, C), and the logic verification engine combined this with the timing waveform diagram (attached). Figure 7 ), attitude monitoring diagram (attached) Figure 8 ) and comprehensive judgment chart (attached) Figure 9 The following analysis will be conducted: Event A (Normal Operation) Waveform characteristics: Appendix Figure 7 (Section A): The impedance curve Z is smooth, gradually decreasing from the air impedance (approximately 900Ω) to the concrete working impedance (approximately 400Ω), and remains horizontal and stable during operation without significant noise.

[0110] Appendix Figure 8(Section A): The equipment tilt angle θ curve is flat throughout, with a stable value of around 5°, which is far below the threshold dashed line of 15°.

[0111] Decision logic: a. Actual measurement duration .

[0112] b. Maximum tilt angle .

[0113] Final result: as Figure 9 As shown at the top, the length of the dark gray progress bar exceeds the vertical threshold line, which is considered normal.

[0114] Event B (Anomaly Type I: Posture Violation) Waveform characteristics: Appendix Figure 8 (Section B): The curve shows significant periodic large oscillations, with the peaks repeatedly breaking through the maximum permissible tilt angle (15°) indicated by the horizontal dashed line, reaching a peak of about 40°, and the waveform is marked with arrows indicating illegal large tilts.

[0115] Appendix Figure 7 (Section B): Affected by the drastic change in posture, the impedance curve Z also showed associated sinusoidal fluctuations in the bottom working area, indicating that the contact state between the vibrator and the concrete was unstable.

[0116] Scene reconstruction: In order to quickly spread the concrete, the workers illegally used vibrators to pry the steel bars and aggregates.

[0117] Judgment logic: Although the duration reaches 20.0s (exceeding the threshold), the valid operation judgment fails because the attitude data is severely out of control.

[0118] Final result: as Figure 9 As shown in the middle, although the progress bar is long, it is filled with diagonal lines as a warning, which is judged as an anomaly I: posture violation.

[0119] Event C (Anomaly Type II: Under-vibration) Waveform characteristics: Appendix Figure 7 (Segment C): The cathode reactance curve is normal in shape, but the duration of the low level is significantly shorter.

[0120] Appendix Figure 8 (Segment C): The attitude curve is stable and the operating angle is compliant.

[0121] On-site reconstruction: The workers operated according to normal temperature conditions based on their experience, ignoring the fact that the low temperature environment hindered the fluidity of the concrete.

[0122] Judgment logic: Actual test duration Compare the dynamic threshold calculated in step one. There was a severe shortage of time.

[0123] Final result: as Figure 9 As shown at the bottom, the light gray progress bar is significantly shorter than the vertical line. There is an under-vibration region between the threshold and the actual measurement conditions, which intuitively shows that the working time is insufficient. The system judges it as Abnormal II: Under-vibration.

Claims

1. A smart data acquisition and anomaly analysis platform for construction engineering supervision, characterized in that, It includes a field perception layer (110), a data transmission layer (120), and a data processing layer (130). The field perception layer (110) is used to acquire multi-source heterogeneous physical data of the field, including electrical data and motion posture data of the handheld vibrating device, original three-dimensional spatial coordinate data, ambient temperature data and concrete material conveying parameters; The data transmission layer (120) is used to perform time synchronization and frequency unification of the multi-source heterogeneous physical data to generate an aligned full-dimensional state vector and upload it to the data processing layer (130). The data processing layer (130) is used to retrieve the pre-stored building information model and receive the full-dimensional state vector and perform the following processing: Spatial drift compensation is performed on the original three-dimensional spatial coordinate data of the target to obtain calibration coordinates; The dynamic operation time threshold is calculated based on the concrete material conveying parameters and the ambient temperature data, and the equivalent electromechanical impedance characteristics and attitude effectiveness characteristics are extracted from the electrical data and motion attitude data. Anomaly determination results are generated based on the calibration coordinates, the dynamic operation duration threshold, the equivalent mechanical impedance characteristics, and the attitude validity characteristics.

2. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 1, characterized in that, The on-site perception layer (110) includes: The intelligent machinery monitoring terminal (111) is installed on the handheld vibrating device and integrates a current sensor (1111) for collecting load current, a voltage sensor (1112) configured to collect input voltage, and an inertial measurement unit (1113) configured to collect triaxial vibration acceleration and attitude Euler angle. The spatial positioning label (112) is configured to output the original three-dimensional spatial coordinate data of the target object at the discrete time step; An environmental monitoring station (113) is configured to collect the ambient temperature data at the construction site; The material conveying monitoring unit (114) is configured to acquire the material conveying parameters of the concrete, including the design slump value.

3. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 2, characterized in that, The data transmission layer (120) includes an edge computing gateway (121); The edge computing gateway (121) is configured to set a unified system resampling frequency, and with the discrete time step of the system resampling frequency as the reference time axis, perform linear interpolation processing on the original three-dimensional spatial coordinate data whose original sampling frequency is lower than the system resampling frequency, and perform downsampling processing on the electrical data whose original sampling frequency is higher than the system resampling frequency by calculating the effective value or average value, thereby generating the full-dimensional state vector under a unified timestamp and uploading it to the data processing layer (130).

4. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 1, characterized in that, The data processing layer (130) includes a data preprocessing module (131), which is configured to perform the spatial drift compensation in the following manner: Select the set of valid job points within the time window, and use a density-based noise spatial clustering algorithm to extract the observation centroid of the largest cluster; Search the building information model for the geometric center of the component that is closest to the observed centroid of the largest cluster; When the Euclidean distance (a straight line between two points in three-dimensional space) between the observed centroid of the largest cluster and the geometric center of the component is less than the preset drift radius, a drift compensation vector pointing from the observed centroid to the geometric center of the component is calculated, and the original three-dimensional spatial coordinate data within the time window is translated and corrected using the drift compensation vector to obtain the calibration coordinates.

5. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 2, characterized in that, The data processing layer (130) includes a physical parameter calculation module (132), which is configured to calculate the dynamic job duration threshold in the following manner: Retrieve pre-stored benchmark operating parameters, which include benchmark liquefaction time, standard slump constant, and standard temperature constant under standard operating conditions; Based on the difference between the standard slump constant and the design slump value of the current batch, the slump correction factor is calculated using the pre-stored slump correction coefficient. The calculation logic of the slump correction factor is to multiply the difference by the slump correction coefficient and then sum it with the unit value. Based on the difference between the standard temperature constant and the current ambient temperature data, a temperature correction factor is calculated using a pre-stored temperature correction coefficient. The calculation logic of the temperature correction factor is to multiply the difference by the temperature correction coefficient and then sum it with the unit value. The dynamic operation duration threshold is obtained by multiplying the baseline liquefaction time by the slump correction factor and the temperature correction factor in sequence.

6. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 2, characterized in that, The data processing layer (130) includes a feature extraction module (133); The feature extraction module (133) is configured to calculate the original impedance value at each time step by dividing the input voltage by the load current, and to perform sliding window low-pass filtering on the original impedance value to obtain the equivalent mechanical impedance characteristics, so as to eliminate the interference of grid voltage fluctuations on load determination.

7. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 6, characterized in that, The feature extraction module (133) is further configured to extract the posture validity features from the motion posture data; The attitude validity feature is represented by a binary attitude validity mask: when the angle between the main axis of the handheld vibrator and the opposite direction of the gravity vector is less than or equal to the maximum allowable tilt angle, the attitude validity mask is marked as valid; when the angle is greater than the maximum allowable tilt angle, the attitude validity mask is marked as invalid, thus forming an attitude validity feature sequence corresponding to the discrete time step.

8. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 7, characterized in that, The feature extraction module (133) is also configured to calculate the weighted impedance change rate feature; the calculation process of the weighted impedance change rate feature specifically includes: Perform a second-order differential or second-order difference operation with respect to the equivalent machine impedance characteristic sequence to obtain an impedance acceleration value that reflects the degree of impedance change. The impedance acceleration value at each time step is multiplied point by point with the corresponding attitude validity mask value. When the attitude validity mask is a value representing an invalid state, the corresponding impedance acceleration value is set to zero. The sequence obtained after multiplication is determined as the weighted impedance change rate feature, and significant extreme peaks are identified in the feature sequence. The time points corresponding to the extreme peaks are characterized as the liquefaction phase transition points of concrete.

9. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 1, characterized in that, The data processing layer (130) includes a logic verification engine (134). The logic verification engine (134) is configured to perform compliance judgment on the operation process based on the dynamic time warping algorithm (DTW): extract the measured impedance sequence composed of the equivalent mechanical impedance characteristics of a single operation interval, retrieve the pre-stored standard process template sequence, and calculate the DTW distance between the measured impedance sequence and the standard process template sequence by performing nonlinear stretching or compression on the time axis. The calculation of the DTW distance between the measured impedance sequence and the standard process template sequence specifically includes: Construct the cumulative distance matrix of data points between the measured impedance sequence and the standard process template sequence; Search the cumulative distance matrix for an optimal curved path from the starting point to the ending point; The minimum cumulative Euclidean distance corresponding to the optimal curved path is determined as the DTW distance.

10. The intelligent data acquisition and anomaly analysis platform for construction engineering supervision as described in claim 9, characterized in that, The logic verification engine (134) is configured to execute cascading exception determination logic with priority: The first level verifies the compliance of the posture. If the proportion of valid postures in the working area is lower than the preset threshold, it is judged as a posture violation and the subsequent judgment is terminated. The second level verifies the physical duration. If the first level of verification is passed and the actual operation duration is less than the dynamic operation duration threshold, it is determined to be an under-vibration anomaly and the subsequent judgment is terminated. The third level of verification process form. If the second level of verification is passed and the DTW distance is greater than the upper limit of waveform similarity tolerance, it is determined to be a process non-compliance anomaly. If all the above levels of verification are passed, it is determined to be a normal operation. The calibration coordinates are used to locate the building information model component ID to which the current operation belongs, and the data of the normal operation is associated with the building information model component ID and stored in the database.

Images