A low-strain method-based pile rapid detection method

Abstract

This invention provides a rapid pile detection method based on the low-strain method, belonging to the field of pile detection technology, and includes the following steps: S1, acquiring spatiotemporal metadata containing timestamps and real-time physical coordinates; S2, processing the low-strain signal to generate quality indicators; S3, using a preset detection process entropy model, calculating the process entropy value based on the spatiotemporal metadata and quality indicators; S4, generating a risk control instruction when the process entropy value exceeds a preset threshold; S5, executing control actions based on the risk control instruction. This invention improves the identification and utilization rate of sub-health state data, avoids unnecessary retesting, and improves detection efficiency; real-time and powerful intervention for serious operational errors ensures the basic quality of data from the source; and quantitative feedback guidance reduces excessive reliance on operator experience, improving the overall consistency and reliability of detection results.

Description

Technical Field

[0001] This invention relates to the field of pile testing technology, specifically a rapid pile testing method based on the low-strain method. Background Technology

[0002] With the deep integration of digital information technology into the field of civil engineering, engineering quality inspection is evolving from a traditional model relying on manual experience to a data-driven, intelligent, and automated approach. Especially in large-scale infrastructure construction, how to utilize electronic digital data processing technology to efficiently and reliably manage and analyze massive amounts of on-site inspection data, and ensure the traceability of the data throughout its entire lifecycle, has become a key technical issue for improving engineering quality control and ensuring structural safety.

[0003] However, in the specific application scenario of low-strain pile foundation testing, existing technologies still face several challenges. Firstly, data quality assessment mechanisms have limitations. Traditional signal quality assessments largely rely on post-processing analysis of single hammer impact waveforms, lacking real-time, quantitative monitoring of the stability of the testing process itself. This makes it difficult to detect latent data quality problems introduced by inconsistent operator behavior or dynamic environmental changes, creating hidden dangers for subsequent analysis. Secondly, data tracing and signal analysis are disconnected. Tracing information such as spatiotemporal location and signal waveform data are usually recorded and managed as two independent, static datasets. The lack of a real-time, dynamic correlation analysis mechanism between the two prevents the system from using spatiotemporal information changes during the process to assist in judging signal quality, and vice versa. Summary of the Invention

[0004] The purpose of this invention is to provide a rapid detection method for foundation piles based on the low strain method, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A rapid detection method for foundation piles based on the low-strain method includes the following steps: S1. Acquire spatiotemporal metadata containing timestamps and real-time physical coordinates, which is acquired synchronously with the low-strain signal; S2. Process the low-strain signals to generate corresponding quality indices that characterize the signal quality. S3. Using a preset detection process entropy model, based on the spatiotemporal metadata and the sequence of quality indicators within a preset time window, calculate the process entropy value used to characterize the deviation of the current detection process from the preset baseline behavior paradigm. S4. When the process entropy value exceeds the preset threshold, generate a risk control instruction to adjust a set of control parameters used when processing subsequent low strain signals. S5. Execute control actions based on risk control instructions.

[0006] Furthermore, the steps for S1 to obtain spatiotemporal metadata are as follows: compare the real-time physical coordinates with the design coordinates corresponding to the same station number obtained from the server; and only when the distance between the real-time physical coordinates and the design coordinates is less than a preset distance threshold will the real-time physical coordinates and timestamp be used as valid spatiotemporal metadata.

[0007] Furthermore, the steps for generating quality indicators in S2 are as follows: For multiple low-strain signals associated with the same station number, calculate the signal similarity coefficient between each pair of the multiple low-strain signals; and generate quality indicators based on the statistical distribution of the signal similarity coefficients.

[0008] Furthermore, the steps for obtaining spatiotemporal metadata are as follows: The input consists of a data frame from the sensor containing raw low-strain signals, timestamps, and real-time physical coordinates, as well as a design coordinate dataset obtained from the server; From the design coordinate dataset, retrieve the corresponding design coordinates based on the station number selected by the operator. Call the coordinate transformation subroutine to convert the real-time physical coordinates into the same local construction coordinate system as the design coordinates; Calculate the three-dimensional Euclidean distance between the transformed real-time physical coordinates and the design coordinates; Obtain the design pile diameter of the current pile and calculate the preset distance threshold; Determine whether the calculated 3D Euclidean distance is less than a preset distance threshold; If it is less than, the positioning verification is successful; mark the timestamp and real-time physical coordinates in the data frame as valid, and cache the low strain signal in the data frame, which is the spatiotemporal metadata; If the deviation is not less than the specified value, the verification fails, an alarm "Positioning deviation is too large" is output to the user interface, and the data frame is discarded, waiting for the next acquisition.

[0009] Furthermore, the steps for obtaining quality indicators are as follows: Continue executing the spatiotemporal metadata acquisition steps until N verified low-strain signals have been successfully cached for the same station number; The signal processing subroutine is invoked to preprocess the N verified low-strain signals and calculate the normalized cross-correlation between each pair to generate a set of signal similarity coefficients. Based on the set of signal similarity coefficients, the final weighted similarity index of the discreteness penalty is calculated; The final output is a floating-point value, which is the weighted similarity index with dispersion penalty, i.e., the quality index; it will be used as the direct input for subsequent quality entropy calculation.

[0010] Furthermore, the steps for calculating the process entropy value using the detection process entropy model include: S3.1 Calculate spatial entropy based on the spatiotemporal metadata sequence within a preset time window; S3.2 Calculate the quality entropy based on the quality index sequence within a preset time window; S3.3. Normalize and weightedly fuse the spatial entropy and mass entropy in sequence to generate the process entropy value.

[0011] Furthermore, the specific steps for calculating the process entropy are as follows: The input consists of two synchronized data sequences, both stored in a first-in-first-out queue of length equal to the sliding time window size: one is a spatiotemporal metadata sequence generated by S1, and the other is a weighted similarity index sequence with discrete penalty generated by S2. Process the spatiotemporal metadata sequence; sequentially perform coordinate transformation, grid generation, point probability statistics, Shannon entropy calculation, and normalization; output the normalized spatial entropy; Process the weighted similarity index sequence with dispersion penalty; sequentially perform interval partitioning, quality probability statistics, Shannon entropy calculation and normalization; output normalized quality entropy; Input the normalized spatial entropy and the normalized mass entropy, and calculate the dynamic weights of the spatial entropy and the mass entropy. Input the normalized spatial entropy, normalized mass entropy, dynamic weight of spatial entropy, and dynamic weight of mass entropy; calculate the final process entropy value; The final output is a floating-point value, namely the process entropy value; the process entropy value will be used as the basis for decision-making in subsequent steps to determine whether to trigger control.

[0012] Furthermore, the step of generating risk control instructions in S4 includes: mapping the process entropy value to a process risk level; and selecting the corresponding risk control instruction from a preset instruction library according to the process risk level.

[0013] Furthermore, the steps of S5 executing the linkage control action include: based on the risk control instruction, synchronously adjusting the judgment threshold in a set of control parameters used to determine whether the signal similarity meets the standard, and outputting a process abnormality alarm corresponding to the process risk level to the user interface.

[0014] Furthermore, the steps for generating and executing risk control instructions are as follows: Input the process entropy value and compare it with the pre-set medium-risk activation threshold and high-risk activation threshold to determine the discrete process risk level; At the same time, a continuous adjustment scaling factor is calculated based on the process entropy value; The process risk level and adjustment ratio factor are encapsulated into structured risk control instructions; Risk control instructions are distributed in parallel to enable algorithm parameter adjustment and user interface interaction; Algorithm parameter adjustment: Parse the adjustment ratio factor in the risk control instruction; calculate the adjusted similarity judgment threshold, and immediately overwrite the original judgment threshold in memory with the new value; thereafter, all new signal quality judgments will use the new judgment threshold; User interface interaction: After receiving risk control instructions, the user interface interaction parses out the process risk level; Select the corresponding alarm scheme based on the process risk level value; The process risk level is low: no alarm actions are executed, and the interface remains normal. The process risk level is medium: A yellow text alert is displayed in a specific status bar of the user interface, which reads "Operational stability has decreased, please be careful"; The process risk level is high: A red, flashing text alert is displayed in the center area of ​​the user interface, which reads "Operational stability has seriously deteriorated. Please check and adjust immediately!" and is optionally accompanied by a short beep. The final output consists of adjustments to internal algorithm parameters and changes to the external environment, and alarms are displayed through the user interface, forming a complete closed-loop human-machine collaborative control circuit.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention aims to fundamentally solve the technical problems in existing technologies where data traceability and signal quality analysis are disconnected and cannot effectively cope with the dynamic uncertainties of the detection process by constructing a closed-loop intelligent detection system that integrates automatic data matching, real-time process monitoring, and risk control.

[0016] The system downloads and pre-loads spreadsheet data, including project information, pile location design coordinates, and pile type parameters, into the testing instrument via a cloud interface. A high-precision positioning device is integrated into the testing sensor to acquire the precise geographic coordinates of the operation point in real time. When the operator selects a specific project on-site, the system enters an automatic matching mode. The real-time acquired coordinates are compared with the pre-set pile location design coordinates. If the real-time coordinates fall within the preset error threshold range for a particular design pile location, the system automatically locks and retrieves the pile number, seamlessly switching the interface to the corresponding signal acquisition state. This enables rapid and automated identification of the target pile, avoiding errors from manual selection. Each subsequent low-strain signal is forcibly bound to its verified geographic coordinates from the outset, and this high-precision spatiotemporal metadata is embedded into the original signal curve, providing an immutable and highly reliable foundation for subsequent data analysis and full lifecycle traceability. This invention introduces a detection process entropy model to achieve real-time, quantitative evaluation of the quality of the detection behavior itself. It transcends the traditional model of judging only a single signal waveform, instead providing a comprehensive understanding of the operational sequence within a time window. It receives and processes two types of heterogeneous data streams in parallel: the first is a quality index sequence reflecting the consistency of signal waveforms, derived by calculating the similarity between multiple consecutive signals; the second is a position coordinate sequence reflecting the spatial stability of the acquisition point, directly derived from a high-precision positioning instrument. By fusing and calculating the inherent uncertainty of these two types of sequences, the process entropy value is finally output. The process entropy value objectively quantifies the orderliness and reliability of the entire detection process; a stable and standardized operation corresponds to a low process entropy value, while a chaotic and arbitrary operation will produce a high process entropy value. This invention achieves its goals through a closed-loop control mechanism that integrates human and machine collaboration, with process entropy as the decision-making hub. It exerts influence on both the algorithm and personnel ends in parallel, using process entropy to dynamically adjust the similarity threshold of the backend signal analysis algorithm. When the process entropy increases, indicating increased process risk, the review standards are automatically raised to filter signals with greater caution, achieving adaptive management of analysis risks. The process entropy is mapped to the front-end operation interface in real time, providing on-site operators with immediate and objective feedback on their operational stability through multi-level alarms including safety, warnings, and interventions, guiding them to proactively correct non-standard behaviors. Only signals and their analysis results confirmed as valid after passing through intelligent process quality control are uploaded to the cloud for automatic generation of test reports. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the detection system of the present invention; Figure 2 This is a schematic diagram of the overall method steps of the present invention; Figure 3 This is a schematic diagram illustrating the steps for acquiring spatiotemporal metadata according to the present invention; Figure 4 This is a schematic diagram illustrating the steps for obtaining the quality indicators of the present invention; Figure 5 This is a schematic diagram of the process entropy step of the present invention; Figure 6 This is a schematic diagram illustrating the steps for generating and executing risk control instructions according to the present invention; Figure 7 This is a schematic diagram of the instrument's working interface according to the present invention; Figure 8 This is a schematic diagram of the on-site operation of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example 1: Please see Figures 1 to 8 This invention provides a technical solution: a rapid detection method for foundation piles based on the low-strain method, comprising the following steps: S1. Acquire spatiotemporal metadata containing timestamps and real-time physical coordinates, which is acquired synchronously with the low-strain signal; The steps for S1 to obtain spatiotemporal metadata are as follows: compare the real-time physical coordinates with the design coordinates corresponding to the same station number obtained from the server; and only when the distance between the real-time physical coordinates and the design coordinates is less than a preset distance threshold will the real-time physical coordinates and timestamps be used as valid spatiotemporal metadata. The process is executed by the detector, which downloads engineering data containing the design coordinates of each pile from the cloud server; the real-time physical coordinates collected by the sensors are used to calculate the three-dimensional Euclidean distance between the converted real-time physical coordinates and the design coordinates; the three-dimensional Euclidean distance is compared with a preset distance threshold set based on the pile diameter (in this embodiment, the preset distance threshold is set to 0.8D, where D is the pile diameter); only when the calculated Euclidean distance is less than the preset distance threshold is the current positioning considered valid, and the collected {timestamp, real-time physical coordinates} are marked as valid and sent to the subsequent processing stage.

[0021] S2. Process the low-strain signals to generate corresponding quality indices that characterize the signal quality. The steps for generating quality indicators using S2 are as follows: For multiple low-strain signals associated with the same station number, calculate the signal similarity coefficient between each pair of the multiple low-strain signals; and generate quality indicators based on the statistical distribution of the signal similarity coefficients.

[0022] The N low-strain signals generated by the detector after N consecutive hammer blows (N=3 in this embodiment) to the same pile are calculated using a normalized cross-correlation algorithm to calculate the signal similarity coefficient between the C(N,2) pairs of signals. These coefficient values ​​({0.95, 0.92, 0.75} in this embodiment) are used as a set to generate a quality index that can characterize the overall quality. In this embodiment, the quality index is the average or median of the set.

[0023] The execution of this embodiment relies on a detector configured to perform digital data processing, which has the ability to communicate with external sensors and servers; the required key parameters are defined as follows: The parameter notation for the design coordinate dataset is set to... It is a structured dataset, which is a set of three-dimensional coordinates of the design center points of all foundation piles in the engineering project; it is obtained by the detector from a cloud server storing building information model or computer-aided design engineering drawings through a wireless communication interface; before the data is acquired, the preprocessing module on the server side has extracted the pile position information in the drawings and converted it into tabular data containing unique pile number identifiers and corresponding design coordinates.

[0024] The parameter sign of the real-time physical coordinates is set to ; is a three-dimensional coordinate point, representing the precise position of the low-strain sensor in the Earth coordinate system at the moment of impact; provided by a high-precision real-time dynamic positioning module integrated with the low-strain sensor; the real-time dynamic positioning module has the ability to receive signals from satellite navigation systems and differential correction signals from ground base stations, with a positioning accuracy better than 5 centimeters.

[0025] The parameter symbol for the preset distance threshold is set to ; is a scalar value used to determine real-time physical coordinates. With design coordinate dataset Whether the design coordinates of the corresponding pile to be tested match the maximum allowed spatial distance; The calculation logic is as follows: It is necessary to obtain the design pile diameter parameters of the pile to be measured, with the parameter symbols as follows: The calculation process for the preset distance threshold is as follows: obtain the design pile diameter. ; Obtain the pile diameter multiple as a safety factor, with parameter symbol as ; Design pile diameter Multiples of pile diameter Perform a multiplication operation to obtain the final preset distance threshold, which is set to 0.8D, i.e., a multiple of the pile diameter. The value is 0.8.

[0026] The parameter notation of the similarity coefficient set is set to ; is a one-dimensional array containing multiple floating-point numbers, which is a set of signal similarity coefficients calculated pairwise for low-strain signals generated by multiple consecutive hammer blows on the same pile; The calculation logic is as follows: N low-strain signals for the same station number are acquired to form a signal sequence; for each signal in the signal sequence, the DC bias is removed, i.e., the arithmetic mean of all sampling points is calculated, and the arithmetic mean is subtracted from the value of each sampling point; for any two signals in the signal sequence, normalized cross-correlation is performed to obtain a signal similarity coefficient ranging from -1 to 1; the calculation logic for the signal similarity coefficient is as follows: the products of corresponding points of the two signals are accumulated, and normalized using the energy of each signal; all calculated signal similarity coefficients are stored in a set, which is the similarity coefficient set. .

[0027] The parameter sign of the dispersion penalty factor is set to ; is a positive real number, used to control the penalty strength for the degree of dispersion of signal consistency when calculating composite quality indicators; Dispersion penalty factor For hyperparameters that require offline calibration; the calibration process is as follows: construct a large-scale low-strain signal sample library containing various working conditions; organize three or more senior testing engineers to score the consistency of each group (N channels) of signals in the sample library, with a score range of 1 (excellent) to 5 (very poor), and take the average score as the expert label of the signal; set a dispersion penalty factor. The candidate value range is from 0.1 to 10 in this embodiment, with a step size of 0.1. For each candidate value, a multi-parameter fusion mechanism is used to calculate the weighted similarity index of the dispersion penalty for all signal groups in the sample library. The Pearson correlation coefficient between the weighted similarity index and the expert label is calculated. The candidate value that maximizes the absolute value of the Pearson correlation coefficient is selected as the final dispersion penalty factor. The final dispersion penalty factor The value is 2.5.

[0028] The parameter sign of the weighted similarity index for dispersion penalty is set to... It is a floating-point number ranging from zero to one, and is a quality indicator that combines the average quality of the signal and the stability of the acquisition process. The calculation logic is as follows: weighted similarity index with dispersion penalty. The calculation process is as follows: Obtain the set of similarity coefficients. ; Calculate the set of similarity coefficients The arithmetic mean of all elements in the matrix is ​​used to obtain the mean similarity. ; Calculate the set of similarity coefficients The standard deviation of all elements is used to obtain the similarity standard deviation. Using the standard deviation of similarity Divide by mean similarity The dimensionless similarity coefficient of variation was obtained. Obtain the dispersion penalty factor The coefficient of variation of similarity With dispersion penalty factor Perform a multiplication operation, then take the negative value of the product; use the natural constant e as the base, and perform an exponential operation on the negative product to obtain the consistency penalty coefficient. ; Mean similarity Consistency penalty coefficient Performing a multiplication operation yields the final weighted similarity index with dispersion penalty. Weighted similarity index with dispersion penalty The range of its value is limited to between zero and one.

[0029] Multi-parameter fusion mechanism methods are based on exponentially penalized nonlinear fusion methods: The degree of dispersion of signal quality (by the similarity coefficient of variation) The contribution of the similarity coefficient to the final metric is not linear, but rather punitive, especially when the data collection process is highly stable (e.g., coefficient of variation). (approaching zero), consistency penalty coefficient Approaching 1, the final index is approximately equal to the mean similarity. Once the process experiences fluctuations (similarity coefficient of variation) Increase), consistency penalty coefficient It decays exponentially, thus affecting the mean similarity. Imposing a large penalty leads to a weighted similarity index with a final dispersion penalty. A sharp decline.

[0030] Dispersion penalty factor The calculation is as follows: Obtain multichannel signals and calculate the set of similarity coefficients between signals. And based on the set of similarity coefficients Generate quality metrics; preprocess the multichannel signals by removing DC bias before calculating the signal similarity coefficients; calculate the signal similarity coefficients using a normalized cross-correlation algorithm; calculate the set of similarity coefficients. mean similarity and similarity coefficient of variation Based on similarity coefficient of variation The consistency penalty coefficient is calculated using an exponential function. ; and mean similarity Consistency penalty coefficient Multiplying yields the quality index; the exponential function includes a dispersion penalty factor determined through offline optimization calibration of a low-strain signal sample library. .

[0031] The steps for obtaining spatiotemporal metadata are as follows: The input consists of raw low-strain signals, timestamps, and real-time physical coordinates from the sensor. The data frame, and the design coordinate dataset obtained from the server. ; From the design coordinate dataset In the process, the corresponding design coordinates are retrieved based on the station number selected by the operator. Call the coordinate transformation subroutine to convert real-time physical coordinates in WGS-84 format. Convert to the same local construction coordinate system as the design coordinates; Calculate the transformed real-time physical coordinates The three-dimensional Euclidean distance between the coordinates and the design coordinates; Get the design pile diameter of the current pile And calculate the preset distance threshold. ; Determine whether the calculated 3D Euclidean distance is less than a preset distance threshold. ; If the value is less than the specified value, the location verification is successful; the timestamp and real-time physical coordinates in the data frame will be used. Mark it as valid and cache the low strain signal in the data frame, which is the spatiotemporal metadata; If the deviation is not less than the specified value, the verification fails, an alarm "Positioning deviation is too large" is output to the user interface, and the data frame is discarded, waiting for the next acquisition.

[0032] The local construction coordinate system transformation is as follows: Input real-time physical coordinates, specifically a triplets containing longitude, latitude, and geodetic height; The coordinate transformation subroutine calls the transformation function; the transformation function converts the input latitude, longitude and altitude coordinates into three-dimensional coordinates in a geocentric rectangular coordinate system with the Earth's center of mass as the origin (0,0,0); the three components (X,Y,Z) of the output coordinates represent the distances relative to the Earth's center along the three orthogonal axes. Output the three-dimensional coordinates of the point in geocentric rectangular coordinates; Input the 3D coordinates of the point in geocentric Cartesian coordinates, and the preset localization transformation parameters; the localization transformation parameters include: three translation parameters (translation amount in the X, Y, Z axes), three rotation parameters (rotation angle around the X, Y, Z axes) and scale parameters; The coordinate transformation subroutine performs the Bursa-Wolf model or a similar seven-parameter Helmert transformation; logically, it includes: The X, Y, and Z components of the input coordinate point are added to the three translation parameters respectively to compensate for the positional difference between the origins of the two coordinate systems. The translated coordinate points are rotated in three-dimensional space around the three coordinate axes in sequence according to the angles specified by the three rotation parameters, so as to align the coordinate axis directions of the two coordinate systems. The X, Y, and Z components of the rotated coordinate points are multiplied by the scale parameter to eliminate the slight differences in scale between the two coordinate systems. Outputs the final three-dimensional Cartesian coordinates in the local construction coordinate system, in meters; that is, the converted real-time physical coordinates.

[0033] The steps for calculating three-dimensional Euclidean distance are as follows: Input the three-dimensional components of the converted real-time physical coordinates and the three-dimensional components of the design coordinates; The coordinate transformation subroutine performs three subtraction operations respectively: The difference in the X-axis is obtained by subtracting the X-component of the design coordinates from the X-component of the converted real-time physical coordinates. The difference in the Y-axis is obtained by subtracting the Y-component of the design coordinates from the Y-component of the converted real-time physical coordinates. The Z-axis difference is obtained by subtracting the Z-component of the design coordinates from the Z-component of the converted real-time physical coordinates. Input the X-axis difference, Y-axis difference, and Z-axis difference; The coordinate transformation subroutine performs three multiplication operations and two addition operations respectively: Calculate the product of the X-axis difference and itself to obtain the square of the X-axis difference; Calculate the product of the Y-axis difference and itself to obtain the square of the Y-axis difference; Calculate the product of the Z-axis difference and itself to obtain the square of the Z-axis difference; Add the squares of the differences along the X-axis, Y-axis, and Z-axis to get the sum of squares; The input is the sum of squares; the coordinate transformation subroutine calls the standard mathematical function library to calculate the arithmetic square root of the sum of squares; the output is a non-negative floating-point number, i.e., the three-dimensional Euclidean distance between the two three-dimensional coordinate points, in meters, used to compare with a preset distance threshold. Compare them.

[0034] The steps for obtaining quality indicators are as follows: Continue executing the spatiotemporal metadata acquisition steps until N verified low-strain signals have been successfully cached for the same station number; The signal processing subroutine is invoked to preprocess the N verified low-strain signals, calculate the normalized cross-correlation between each pair, and generate a set of similarity coefficients. ; Based on similarity coefficient set The final weighted similarity index with dispersion penalty is calculated. ; The final output is a floating-point value, which is the weighted similarity index with dispersion penalty. This refers to the quality index; it will serve as the direct input for subsequent quality entropy calculations.

[0035] S3. Using a preset detection process entropy model, based on the spatiotemporal metadata and the sequence of quality indicators within a preset time window, calculate the process entropy value used to characterize the deviation of the current detection process from the preset baseline behavior paradigm. The steps used by the detection process entropy model to calculate the process entropy value include: S3.1 Calculate spatial entropy based on the spatiotemporal metadata sequence within a preset time window; S3.2 Calculate the quality entropy based on the quality index sequence within a preset time window; S3.3. Normalize and weightedly fuse the spatial entropy and mass entropy in sequence to generate the process entropy value.

[0036] The key parameters required for this embodiment are defined as follows: The parameter sign for the sliding time window size is set to... ; is an integer representing the number of the nearest data points captured when calculating process entropy; sliding time window size. The value is set to 10.

[0037] The sign of the parameter for the number of spatial grids is set to ; is an integer representing the total number of cells used to discretize the two-dimensional region at the top of the pile when calculating spatial entropy; the number of spatial grids. The value is set to 9.

[0038] The parameter sign of the normalized spatial entropy is set as follows: ; is a floating-point number ranging from zero to one, a normalized quantitative indicator that characterizes the dispersion of sensor placement positions; The calculation logic is as follows: Obtain the latest sliding time window size within the sliding time window. Real-time physical coordinates; uniformly transform coordinate points to a two-dimensional local coordinate system with the current pile design center as the origin; based on the current pile design diameter. Define a square area that exactly covers the top of the pile, and divide the top of the pile into a spatial grid of [number] grids. There are 1 grid; iterate through all coordinate points, count the number of points falling within the i-th grid, and then divide by the sliding time window size. To obtain the point probability of the i-th grid. The probability of each non-zero point. Calculate the probability of the location Logarithm to base 2, multiplied by the probability of the point The process involves: obtaining the calculation results; summing the results of all grids and taking the negative of the sum to obtain the original spatial entropy value; and calculating the number of grids in the spatial division. The maximum possible entropy is obtained by taking the logarithm to base 2; the original spatial entropy is then divided by the maximum possible entropy to obtain the final normalized spatial entropy. .

[0039] The parameter sign for the number of quality index intervals is set to ; is an integer representing the number of bins used to discretize continuous quality indicators when calculating quality entropy; in this embodiment, it refers to the number of quality indicator intervals. The value is set to 5; Weighted similarity index based on dispersion penalty within the current sliding time window The sequence; perform the following steps to determine the number of quality index intervals. : Obtain a weighted similarity index with dispersion penalty The sequence length, i.e., the sliding time window size. ; Calculate the sliding time window size The logarithm is calculated to base 10; for a value of 10, the logarithm to base 10 is 1. The resulting logarithmic value is multiplied by a fixed logarithmic conversion factor (3.322 in this embodiment); for a value of 1, the product is 3.322. The product is added to 1 to obtain the floating-point baseline bin number; for a value of 3.322, the sum is 4.322. Quantitative fine-tuning based on engineering experience is performed; the obtained baseline bin number is rounded up to obtain... Integer value; for a value of 4.322, rounding up results in 5; to avoid unreasonable bin counts due to excessively small or large window sizes, a minimum and maximum bin count are preset; in this embodiment, the minimum bin count is set to 5 to ensure that there are at least enough bins to capture the basic shape of the quality distribution; the maximum bin count is set to 20 to avoid generating too many empty bins with limited data points; the obtained integer value is compared with the minimum and maximum bin counts; if the integer value is less than the minimum bin count, the final quality index interval number is... The value is the minimum number of boxes; if the integer value is greater than the maximum number of boxes, then the final quality index interval number is... The value is the maximum number of bins; otherwise, the number of quality index intervals. This is the integer value obtained; in this embodiment, since the result of rounding up is 5, which is exactly equal to the minimum number of bins, the final output quality index interval number is... It is 5.

[0040] The parameter sign of the normalized mass entropy is set to ; is a floating-point number ranging from zero to one, and is a normalized quantitative indicator that characterizes the stability of signal acquisition quality; The calculation logic is as follows: Obtain the latest sliding time window size within the sliding time window. Weighted similarity index with discreteness penalty The value; dividing the range of zero to one into equal intervals for quality indicators. Each interval; traversing all weighted similarity indices with dispersion penalties. The value is calculated, and the number of values ​​falling within the j-th interval is counted. Then, this number is divided by the sliding time window size. The quality probability of the j-th interval is obtained. ; The normalized spatial entropy used for calculation The exact same logic applies to the probability of this location. Replace with quality probability and the number of grids Replace with quality index intervals This will give us the final normalized mass entropy. .

[0041] The parameter sign of the spatial entropy dynamic weight is set to The parameter sign of the dynamic weight of mass entropy is set to It is a floating-point number in the range of zero to one. The calculation logic is as follows: Obtain the currently calculated normalized spatial entropy. and normalized mass entropy Normalized spatial entropy and normalized mass entropy Summing yields the total entropy value; to avoid the total entropy being zero and rendering subsequent calculations meaningless, a very small positive number is introduced, with the parameter sign . As a smoothing term; in this embodiment, The value is 0.000001; the total entropy value is compared with... Add them together to get the smoothed total entropy value; use the normalized spatial entropy. Dividing by the smoothed total entropy value yields the spatial entropy dynamic weight. ; Using normalized spatial entropy Dividing by the smoothed total entropy value yields the dynamic weight of the mass entropy. If the normalized spatial entropy is calculated at a certain moment... The normalized mass entropy is 0.6. The value is 0.2; the total entropy value is 0.8; therefore, the spatial entropy dynamic weight is... The calculation is 0.6 divided by 0.8, resulting in 0.75; dynamic weighting of mass entropy. The calculation is 0.2 divided by 0.8, resulting in 0.25; this indicates that the inconsistency in spatial location is the main source of process risk at this moment, and the weight is dynamically increased. In another embodiment, the currently calculated normalized spatial entropy is obtained. and normalized mass entropy Normalized spatial entropy and normalized mass entropy Summing yields the total entropy value; to avoid the total entropy being zero and rendering subsequent calculations meaningless, a very small positive number is introduced, with the parameter sign . As a smoothing term; in this embodiment, The value is 0.000001; the total entropy value is compared with... Add them together to get the smoothed total entropy value; use the normalized spatial entropy. Dividing by the smoothed total entropy value yields the spatial entropy dynamic weight. After calculating the dynamic weights of spatial entropy Then, the dynamic weight of mass entropy. The calculation is 1 minus the spatial entropy and dynamic weight. .

[0042] The parameter symbol for the process entropy value is set to ; is a floating-point number between zero and one, which is a comprehensive indicator that ultimately quantifies the disorder or non-standardization of the current detection process; The calculation logic is as follows: Obtain the normalized spatial entropy. and spatial entropy dynamic weight Perform a multiplication operation on the two to obtain the spatial entropy contribution term; obtain the normalized mass entropy. and dynamic weights of mass entropy Multiply the two to obtain the mass entropy contribution term; sum the spatial entropy contribution term and the mass entropy contribution term to obtain the final process entropy value. .

[0043] Process entropy The calculation steps are as follows: The input consists of two synchronized data sequences, both stored in a space of length equal to the sliding time window size. In the first-in-first-out queue: one is the spatiotemporal metadata sequence generated by S1, and the other is the weighted similarity index with discreteness penalty generated by S2. sequence; Process the spatiotemporal metadata sequence; sequentially perform coordinate transformation, grid generation, point probability statistics, Shannon entropy calculation, and normalization; output the normalized spatial entropy. ; Weighted similarity index for handling dispersion penalty Sequence; sequentially perform interval partitioning, quality probability statistics, Shannon entropy calculation, and normalization; output normalized quality entropy. ; Input normalized spatial entropy and normalized mass entropy Calculate the dynamic weight of spatial entropy and dynamic weights of mass entropy ; Input normalized spatial entropy Normalized mass entropy Spatial entropy dynamic weight and dynamic weights of mass entropy ; Calculate the final process entropy value ; The final output is a floating-point value, which is the process entropy value. Process entropy This will serve as the basis for determining whether to trigger control in subsequent steps.

[0044] S4. When the process entropy value exceeds the preset threshold, generate a risk control instruction to adjust a set of control parameters used when processing subsequent low strain signals. The steps of generating risk control instructions in S4 include: mapping process entropy values ​​to a process risk level; and selecting the corresponding risk control instruction from a preset instruction library based on the process risk level.

[0045] S5. Execute control actions based on risk control instructions; The steps of S5 in executing the linkage control action include: based on the risk control instruction, synchronously adjusting the judgment threshold in a set of control parameters used to determine whether the signal similarity meets the standard, and outputting a process abnormality alarm corresponding to the process risk level to the user interface.

[0046] The key parameters involved in this embodiment are defined as follows: The parameter sign for the medium-risk activation threshold is set to The parameter sign for the high-risk activation threshold is set to... ; is a floating-point number between zero and one, and is a boundary value used to discretize continuous process entropy values ​​into three risk levels: "low", "medium" and "high". The calculation logic is as follows: Collect a large number of historical detection sessions from multiple different projects and operators, and calculate the entropy values ​​of all processes generated in each session. This generates a massive historical entropy dataset; all entropy values ​​in the historical entropy dataset are sorted in ascending order; and a medium-risk activation threshold is determined. In this embodiment, the 70th percentile of the historical entropy dataset is selected as the medium-risk activation threshold. That is, 70% of the entropy values ​​in the dataset are less than or equal to the medium-risk activation threshold. Determine the high-risk activation threshold. In this embodiment, the 90th percentile of the historical entropy dataset is selected as the high-risk activation threshold. In this embodiment, the medium-risk activation threshold Set to 0.4, a high-risk activation threshold. It is set to 0.7.

[0047] The parameter sign of the benchmark similarity threshold is set to ; is a floating-point number between zero and one, and is the default or basic threshold used to determine whether signal similarity meets the standard under low-risk conditions; in this embodiment, according to industry practice, the benchmark similarity judgment threshold It was set to 0.8.

[0048] The parameter sign for the maximum threshold adjustment range is set to... ; is a positive floating-point number, representing the threshold for determining the baseline similarity when the process entropy reaches its maximum value (in this embodiment, the maximum value is set to 1). The maximum amount that can be increased; The calculation logic is as follows: In this embodiment, the maximum threshold adjustment range The threshold for determining the baseline similarity is set to 0.15. 0.8 plus the maximum threshold adjustment range The resulting threshold is 0.95; this represents the worst-case scenario where the signal must achieve near-perfect repeatability to be accepted. It is a strict but still achievable upper limit, avoiding the need for a baseline similarity threshold. The value was adjusted to a value that could not be achieved; in this embodiment, it is set to be greater than 1.

[0049] Adjust the sign of the scaling factor parameter to: ; is a floating-point number between zero and one, used to represent the entropy value of a continuous process. The core proportionality factor is converted into the intensity of continuous control actions; The calculation logic is as follows: Obtain the current process entropy value. Activation threshold for medium risk ; Determine the entropy value of the process Is it less than the medium-risk activation threshold? If it is less than, adjust the scaling factor. It is directly set to 0; if it is not less than, it is determined from the process entropy value. Subtract the medium-risk activation threshold This yields the excess entropy value; subtract the medium-risk activation threshold from 1. The effective entropy interval width is obtained; the excess entropy value is divided by the effective entropy interval width, and the result is the adjustment scaling factor. To ensure the adjustment of the scaling factor The result should not exceed 1; if the calculated result is greater than 1, the scaling factor will be adjusted. Limited to 1.

[0050] In this embodiment: Medium-risk activation threshold The entropy value is 0.4; if the current process entropy value is... If the value is 0.6, then the excess entropy value is 0.2, and the effective entropy interval width is 0.6; adjust the scaling factor. The calculation is 0.2 divided by 0.6, resulting in 0.333; if the current process entropy value... If the value is 0.9, then the excess entropy value is 0.5, the effective entropy interval width is 0.6, and the adjustment factor is... The calculation is 0.5 divided by 0.6, and the result is 0.833.

[0051] The parameter symbol of the risk control instruction is set to It is a structured data object, generated by S4, containing complete control information, and used to drive the execution of S5 internal instructions; The calculation logic is as follows: Obtain the process entropy value. Based on process entropy value With medium-risk activation threshold and high-risk activation threshold The comparison results determine the discrete process risk level. The values ​​are "low", "medium", and "high"; based on the process entropy value. Calculate the continuous adjustment ratio factor ; classify process risk levels and adjustment scaling factor Encapsulated into a data structure, forming the final risk control instructions. .

[0052] The parameter sign of the adjusted similarity threshold is set to It is a floating-point number between zero and one, which is the final threshold used to determine signal quality after dynamic adjustment based on real-time risk. The calculation logic is as follows: Obtain the baseline similarity threshold. Maximum threshold adjustment range and by risk control instructions Input adjustment factor Adjust the maximum threshold range With adjustment scaling factor Perform a multiplication operation to obtain the dynamic adjustment amount; set the baseline similarity threshold. The summation with the obtained dynamic adjustment amount yields the final adjusted similarity threshold. .

[0053] In this embodiment: the benchmark similarity determination threshold The maximum threshold adjustment range is 0.8. The scaling factor is adjusted to 0.15. If the value is 0.333, then the dynamic adjustment amount is 0.05; the final adjusted similarity threshold. Adding 0.05 to 0.8 equals 0.85.

[0054] The steps for generating and executing risk control instructions are as follows: Input process entropy and process entropy value With the pre-set medium-risk activation threshold and high-risk activation threshold Compare and determine the discrete process risk level. ; At the same time, according to the process entropy value Calculate the continuous adjustment ratio factor ; Process risk level and adjustment scaling factor Encapsulated into structured risk control instructions ; Risk control instructions Parallel distribution enables algorithm parameter adjustment and user interface interaction; Algorithm parameter adjustment: Parsing risk control instructions Adjustment scaling factor ; Calculate the adjusted similarity threshold The new value will immediately overwrite the existing threshold value in memory; thereafter, all new signal quality determinations will use the new threshold value. User interface interaction: The user interface receives risk control instructions. Then, the process risk level was analyzed. ; Based on process risk level Select the corresponding alarm scheme based on the value; Process risk level Low: No alarm action is executed, and the interface remains normal; Process risk level For the middle: A yellow text alert is displayed in a specific status bar of the user interface, which reads "Operational stability has decreased, please be careful"; Process risk level High: A red, flashing text alert is displayed in the center of the user interface, stating "Operational stability has severely deteriorated. Please check and adjust immediately!", optionally accompanied by a short beep. The final output consists of adjustments to internal algorithm parameters and changes to the external environment, and alarms are displayed through the user interface, forming a complete closed-loop human-machine collaborative control circuit.

[0055] Figure 1 and Figure 8 The automatic vibration device with adjustable distance and impact height is used to hammer the foundation piles; An automatic vibration device includes at least the following structure: Vibration hammer: A standard mechanical hammer made of special steel that meets the requirements of the "Technical Specification for Dynamic Testing of Foundation Piles". Its weight and hammerhead curvature radius are standard values; a permanent magnet is embedded in the upper part of the hammer body. Holding and releasing mechanism: A DC electromagnet is installed directly above the hammer; when it is necessary to hold the hammer head at a preset height, the electromagnet is energized, generating a strong magnetic force to hold the hammer head in place; Reset mechanism: A small cylinder or linkage mechanism is provided on the side of the falling path of the hammer to push the hammer head back or lift it directly under the electromagnet after one strike, so that it can be attracted and lifted again next time.

[0056] The operating steps of the automatic vibration device are as follows: Based on the automatically collected coordinate information, select the matching pile number and verify the pile information and parameters; the operator places the automatic vibration device on the pile head and levels it; turn on the device power, and select to connect the automatic vibration device on the low strain detector's operating interface. The operator installs the wireless sensor on the pile head; on the detector software interface, the sensor position is usually defined as a specific point in the coordinate system. In this embodiment, it is the point (0,R) on the Y-axis, where R is the pile radius. The operator clicks on one or more vibration points required by the specifications on the pile head diagram in the detector software; the detector sends these clicked locations to the automatic vibration device via Bluetooth. After receiving the command, the automatic vibration device simultaneously drives the rotary motor and the radial telescopic motor to precisely move the vibration hammer directly above the target point; The detector sends a height command according to the specifications or user settings; it drives the vertical lifting motor to move the electromagnet and the attracted hammer to the target height. The operator clicks "Start Testing" on the detector; the detector sends a tapping command; the automatic vibration device immediately cuts off the electromagnet power, and the vibration hammer falls freely, completing one standard tapping; the sensor collects the signal, and the detector records it synchronously. After the tapping is completed, the reset mechanism resets the hammer head, ready to receive the positioning and tapping command for the next point, thereby realizing automated, rapid and continuous detection of multiple measuring points; The automatic vibration device hammers the pile head, generating stress waves that propagate through the pile and simultaneously cause severe vibration at the top of the pile. The piezoelectric crystal inside the sensor, which is rigidly connected to the top of the pile, generates an electric charge signal due to the vibration, thus enabling the acquisition of low-strain signals.

[0057] It should be noted that all calculation formulas in this application employ regression analysis, including but not limited to machine learning algorithms, to deeply analyze the collected parameters and identify their natural trends and interrelationships. Specialized software, such as Python's Scikit-learn library or the R language, is used to automatically generate mathematical models that match the data. Then, cross-validation and other methods are used to objectively evaluate the model performance, and continuous feedback and optimization are combined to ensure that the created formulas truly reflect the inherent laws of the data, thereby guaranteeing their effectiveness and accuracy. In all calculation formulas in this application, the parameters in each formula undergo dimensionless processing within a consistent range to ensure that different physical quantities are compared on the same scale; dimensionless processing techniques include, but are not limited to, Min-Max Normalization and Z-Score standardization. The technical solution of this invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of this invention.

[0058] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0059] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A low-strain method-based pile rapid detection method, characterized in that, Includes the following steps: S1. Acquire spatiotemporal metadata containing timestamps and real-time physical coordinates, which is acquired synchronously with the low-strain signal; S2. Process the low-strain signals to generate corresponding quality indices that characterize the signal quality. S3. Using a preset detection process entropy model, based on the spatiotemporal metadata and the sequence of quality indicators within a preset time window, calculate the process entropy value used to characterize the deviation of the current detection process from the preset baseline behavior paradigm. S4. When the process entropy value exceeds the preset threshold, generate a risk control instruction to adjust a set of control parameters used when processing subsequent low strain signals. S5. Execute control actions based on risk control instructions.

2. The method according to claim 1, wherein the method is a low-strain method. The steps for S1 to obtain spatiotemporal metadata are as follows: compare the real-time physical coordinates with the design coordinates corresponding to the same station number obtained from the server; and only when the distance between the real-time physical coordinates and the design coordinates is less than a preset distance threshold will the real-time physical coordinates and timestamps be used as valid spatiotemporal metadata.

3. The rapid detection method for foundation piles based on the low-strain method according to claim 2, characterized in that: The steps for generating quality indicators using S2 are as follows: For multiple low-strain signals associated with the same station number, calculate the signal similarity coefficient between each pair of the multiple low-strain signals; and generate quality indicators based on the statistical distribution of the signal similarity coefficients.

4. The rapid pile detection method based on low strain method according to claim 3, characterized in that: The steps for obtaining spatiotemporal metadata are as follows: The input consists of a data frame from the sensor containing raw low-strain signals, timestamps, and real-time physical coordinates, as well as a design coordinate dataset obtained from the server; From the design coordinate dataset, retrieve the corresponding design coordinates based on the station number selected by the operator. Call the coordinate transformation subroutine to convert the real-time physical coordinates into the same local construction coordinate system as the design coordinates; Calculate the three-dimensional Euclidean distance between the transformed real-time physical coordinates and the design coordinates; Obtain the design pile diameter of the current pile and calculate the preset distance threshold; Determine whether the calculated 3D Euclidean distance is less than a preset distance threshold; If it is less than, the positioning verification is successful; mark the timestamp and real-time physical coordinates in the data frame as valid, and cache the low strain signal in the data frame, which is the spatiotemporal metadata; If the deviation is not less than the specified value, the verification fails, an alarm "Positioning deviation is too large" is output to the user interface, and the data frame is discarded, waiting for the next acquisition.

5. The rapid detection method for foundation piles based on the low-strain method according to claim 4, characterized in that: The steps for obtaining quality indicators are as follows: Continue executing the spatiotemporal metadata acquisition steps until N verified low-strain signals have been successfully cached for the same station number; The signal processing subroutine is invoked to preprocess the N verified low-strain signals and calculate the normalized cross-correlation between each pair to generate a set of signal similarity coefficients. Based on the set of signal similarity coefficients, the final weighted similarity index of the discreteness penalty is calculated; The final output is a floating-point value, which is the weighted similarity index with dispersion penalty, i.e., the quality index; it will be used as the direct input for subsequent quality entropy calculation.

6. The rapid detection method for foundation piles based on the low-strain method according to claim 5, characterized in that: The steps used by the detection process entropy model to calculate the process entropy value include: S3.1 Calculate spatial entropy based on the spatiotemporal metadata sequence within a preset time window; S3.2 Calculate the quality entropy based on the quality index sequence within a preset time window; S3.

3. Normalize and weightedly fuse the spatial entropy and mass entropy in sequence to generate the process entropy value.

7. The rapid pile detection method based on low strain method according to claim 6, characterized in that: The specific steps for calculating the process entropy are as follows: The input consists of two synchronized data sequences, both stored in a first-in-first-out queue of length equal to the sliding time window size: one is a spatiotemporal metadata sequence generated by S1, and the other is a weighted similarity index sequence with discrete penalty generated by S2. Process the spatiotemporal metadata sequence; sequentially perform coordinate transformation, grid generation, point probability statistics, Shannon entropy calculation, and normalization. Output the normalized spatial entropy; A weighted similarity index sequence for handling dispersion penalty; The process involves sequentially performing interval partitioning, quality probability statistics, Shannon entropy calculation, and normalization. Output normalized mass entropy; Input the normalized spatial entropy and the normalized mass entropy, and calculate the dynamic weights of the spatial entropy and the mass entropy. Input normalized spatial entropy, normalized mass entropy, dynamic weight of spatial entropy, and dynamic weight of mass entropy; Calculate the final process entropy value; The final output is a floating-point value, namely the process entropy value; the process entropy value will be used as the basis for decision-making in subsequent steps to determine whether to trigger control.

8. The rapid detection method for foundation piles based on the low-strain method according to claim 7, characterized in that: The steps of generating risk control instructions in S4 include: mapping process entropy values ​​to a process risk level; and selecting the corresponding risk control instruction from a preset instruction library based on the process risk level.

9. A rapid pile detection method based on low strain method according to claim 8, characterized in that: The steps of S5 in executing the linkage control action include: based on the risk control instruction, synchronously adjusting the judgment threshold in a set of control parameters used to determine whether the signal similarity meets the standard, and outputting a process abnormality alarm corresponding to the process risk level to the user interface.

10. The rapid detection method for foundation piles based on the low-strain method according to claim 9, characterized in that: The steps for generating and executing risk control instructions are as follows: Input the process entropy value and compare it with the pre-set medium-risk activation threshold and high-risk activation threshold to determine the discrete process risk level; At the same time, a continuous adjustment scaling factor is calculated based on the process entropy value; The process risk level and adjustment ratio factor are encapsulated into structured risk control instructions; Risk control instructions are distributed in parallel to enable algorithm parameter adjustment and user interface interaction; Algorithm parameter adjustment: Parse the adjustment ratio factor in the risk control instruction; Calculate the adjusted similarity threshold and immediately overwrite the original threshold in memory with the new value; From now on, all new signal quality determinations will use the new determination threshold; User interface interaction: After receiving risk control instructions, the user interface interaction parses out the process risk level; Select the corresponding alarm scheme based on the process risk level value; The process risk level is low: no alarm actions are executed, and the interface remains normal. The process risk level is medium: A yellow text alert is displayed in a specific status bar of the user interface, which reads "Operational stability has decreased, please be careful"; The process risk level is high: A red, flashing text alert is displayed in the center area of ​​the user interface, which reads "Operational stability has seriously deteriorated. Please check and adjust immediately!" and is optionally accompanied by a short beep. The final output consists of adjustments to internal algorithm parameters and changes to the external environment, and alarms are displayed through the user interface, forming a complete closed-loop human-machine collaborative control circuit.

Images