Engineering construction slag discharge data abnormal early warning analysis method and system

Abstract

The application discloses an abnormal early warning analysis method and system for slagging data of engineering construction, and relates to the field of data processing, wherein the method comprises the following steps: collecting multi-source operation data associated with the slagging process to construct a slagging behavior data set during the equipment tunneling process; dividing the slagging process into multiple slagging behavior units according to a preset time window, and constructing a multi-dimensional behavior feature vector for each unit; based on the pressure balance mechanism of the earth pressure shield excavation face, a coupling constraint relationship between the slagging behavior and the earth chamber pressure and the tunneling parameters is constructed to form a dynamic benchmark model; the multi-dimensional deviation degree of the actual behavior relative to the benchmark is calculated by inputting the multi-dimensional behavior feature vector into the model, and the comprehensive deviation index is obtained by fusing each deviation degree; and the slagging abnormal early warning is output. The application solves the problems of low monitoring accuracy, poor real-time performance, false alarm and missed alarm of the existing slagging monitoring, and achieves the effects of improving the abnormal monitoring accuracy and real-time performance, and reducing the false alarm and missed alarm.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to a method and system for early warning analysis of abnormal construction waste discharge data. Background Technology

[0002] During earth pressure balance (EPB) tunnel boring machine (TBM) construction, the stability of the muck removal process directly affects the safety of the excavation face, construction efficiency, and project quality. If abnormalities occur during muck removal and are not detected and addressed promptly, they can easily lead to safety accidents such as ground subsidence and TBM jamming. Therefore, monitoring and early warning of abnormalities in the muck removal process is one of the core aspects of EPB TBM construction. Currently, the mainstream method for monitoring abnormalities in muck removal in the industry mainly relies on manual monitoring by on-site personnel. This involves real-time observation of the muck removal status of the screw conveyor, manual recording of key data such as muck removal volume and soil chamber pressure, and combining this with their own construction experience to determine whether there are any abnormalities in the muck removal process. A simple threshold alarm device is also used as a supplement, with preset fixed thresholds for parameters such as muck removal volume and soil chamber pressure. When the monitored data exceeds the threshold, an alarm is issued. Existing methods have obvious limitations. On the one hand, manual monitoring is greatly affected by the experience, sense of responsibility and fatigue of the staff, which can easily lead to problems such as monitoring omissions and judgment biases. Moreover, manual data recording is inefficient and prone to errors, and cannot achieve real-time and continuous monitoring of slag discharge data. On the other hand, simple threshold alarm devices can only alarm for a single parameter and do not consider the correlation between parameters, which can easily lead to false alarms and missed alarms. They are difficult to accurately reflect the overall abnormal state of the slag discharge process and cannot meet the accuracy and real-time requirements of modern earth pressure shield tunneling for abnormal slag discharge early warning.

[0003] At present, the monitoring of abnormal muck discharge in earth pressure shield tunneling has technical problems such as low monitoring accuracy, poor real-time performance, and susceptibility to false alarms and missed alarms, and it cannot accurately reflect the overall abnormal state of the muck discharge process. Summary of the Invention

[0004] This application provides an anomaly early warning analysis method and system for muck discharge data during engineering construction. It collects multi-source operational data in real time during earth pressure shield tunneling, including muck discharge volume of the screw conveyor, time series, earth chamber pressure, cutterhead torque and speed, and propulsion speed, and integrates this data into a unified muck discharge behavior dataset. The continuous muck discharge process is divided into multiple muck discharge behavior units according to a preset time window. A multi-dimensional behavior feature vector is extracted and constructed for each unit. Based on the pressure balance mechanism of the earth pressure shield excavation face, a coupling constraint relationship is established between muck discharge behavior, earth chamber pressure, and tunneling parameters, forming a dynamic benchmark model adapted to the construction state. This model integrates the data from each muck discharge behavior unit... The model inputs feature vectors to calculate four types of multi-dimensional deviations: slag volume, rhythm, pressure response, and energy consumption correlation. These deviations are then fused to obtain a comprehensive deviation index. Based on this comprehensive deviation index, the model judges and outputs corresponding slag discharge anomaly warnings. This approach solves the technical problems of low monitoring accuracy and poor real-time performance in existing earth pressure shield tunneling slag discharge anomaly monitoring, which are prone to false alarms and missed alarms, and cannot accurately reflect the overall abnormal state of the slag discharge process. This approach achieves the technical effect of improving the accuracy and real-time performance of slag discharge anomaly monitoring, effectively reducing false alarms and missed alarms, accurately reflecting the overall abnormal state of the slag discharge process, and providing reliable protection for construction safety.

[0005] This application provides a method for abnormal early warning analysis of muck discharge data in engineering construction, including: during equipment tunneling, collecting multi-source operational data associated with the muck discharge process, such as muck discharge volume data of the screw conveyor, muck discharge time series data, earth chamber pressure data, cutterhead torque and speed data, and propulsion speed data, to construct a unified muck discharge behavior dataset; based on the muck discharge behavior dataset, dividing the muck discharge process according to a preset time window, discretizing the continuous muck discharge process into multiple muck discharge behavior units, and constructing a multi-dimensional behavior feature vector for each muck discharge behavior unit; based on the earth pressure shield tunneling face pressure balance mechanism, constructing a coupling constraint relationship between muck discharge behavior, earth chamber pressure, and tunneling parameters to form a dynamic benchmark model of muck discharge behavior; inputting the multi-dimensional behavior feature vector of the muck discharge behavior unit into the dynamic benchmark model, calculating the multi-dimensional deviation of the actual behavior relative to the dynamic benchmark, such as muck discharge volume deviation, rhythm deviation, pressure response deviation, and energy consumption correlation deviation, and fusing the deviations to obtain a comprehensive deviation index; and outputting an abnormal muck discharge warning based on the comprehensive deviation index.

[0006] In a possible implementation, the following processing is performed: the coupling constraint relationship includes the balance relationship between the slag discharge volume and the earth chamber pressure, the synchronization relationship between the slag discharge rhythm and the propulsion speed, and the correlation relationship between the cutterhead energy consumption and the slag discharge volume.

[0007] In a possible implementation, a dynamic benchmark model of muck removal behavior is formed, and the following processing is performed: Based on the muck removal behavior dataset, historical muck removal behavior units with soil chamber pressure fluctuations within a preset stable range and continuous advance speed changes are selected to construct a benchmark sample set; intervariate correlation analysis is performed on the multidimensional behavioral features in the benchmark sample set to determine the balance correlation function between muck removal volume and soil chamber pressure, the synchronization response function between muck removal rhythm and advance speed, and the coupling mapping function between cutterhead energy consumption and muck removal volume; multivariate coupling constraint relationships are established based on the intervariate correlation analysis results; based on the coupling constraint relationships, the benchmark sample sets of different tunneling sections are partitioned and fitted to form multiple local benchmark sub-models corresponding to different construction states; based on the matching degree between the characteristics of the muck removal behavior units and the local benchmark sub-models, the local benchmark sub-models are called and the coupling constraint relationships are dynamically corrected to form a dynamic benchmark model.

[0008] In a possible implementation, a dynamic benchmark model is formed, and the following processes are performed: obtaining the feature distribution center corresponding to each local benchmark sub-model; configuring the matching degree based on the distance value between the feature of the slag discharge behavior unit and the feature distribution center; filtering and calling local benchmark sub-models based on the matching degree, and performing dynamic correction of coupling constraint relationships.

[0009] In a possible implementation, the multidimensional behavioral feature vector of the slag discharge behavior unit is input into the dynamic benchmark model, and the following processing is performed: the comparison processing layer within the dynamic benchmark model is invoked to perform residual comparison of the multidimensional behavioral feature vectors, and a standardized residual dataset is established; the standardized residual dataset is aligned within the same time window to form a deviation combination vector; it is determined whether there are any vectors in the deviation combination vector that exceed their respective benchmark thresholds; when at least two deviation combination vectors exceed their corresponding benchmark thresholds simultaneously, a coupling enhancement term is constructed based on the corresponding deviation combination vectors; when only a single deviation combination vector exceeds its corresponding benchmark threshold, the original value is retained to construct a univariate deviation term; based on the superposition of the coupling enhancement term and the univariate deviation term, the superposition result is used for sliding accumulation processing within the time window to construct a comprehensive deviation index.

[0010] In possible implementations, the following processing is performed: the multidimensional behavioral feature vector includes the rate of change of slag volume per unit time calculated based on slag discharge time series data, the slag discharge stability index calculated based on the fluctuation amplitude of slag volume within a continuous time window, the pressure fluctuation amplitude and pressure recovery time calculated based on earth chamber pressure data, the energy consumption ratio corresponding to the unit slag volume calculated based on cutterhead torque and speed data, and the propulsion-slag discharge matching index constructed based on the ratio relationship between propulsion speed data and slag volume data.

[0011] In a possible implementation, based on the comprehensive deviation index, an abnormal slag discharge warning is output, and the following processing is also performed: an abnormal warning signal is configured according to the comprehensive deviation index, and a warning instruction is established; the warning instruction is sent to the warning device, and a visual warning output management is performed.

[0012] In a possible implementation, the multidimensional deviation of the actual behavior relative to the dynamic benchmark is calculated, and the following processing is performed: time series consistency analysis is performed on the multidimensional behavior feature vectors of multiple consecutive slag discharge behavior units to construct a change trend sequence between adjacent time windows; the change trend sequence is compared with the trend change of the corresponding variable in the dynamic benchmark model to generate a trend deviation coefficient; and multidimensional deviation is calculated and compensated based on the trend deviation coefficient.

[0013] In a possible implementation, the multi-source operational data associated with the slag discharge process is collected, and the following processing is performed: performing time window self-verification analysis of the multi-source operational data to establish a self-verification signal; configuring an additional acquisition strategy based on the self-verification signal, and performing a reset acquisition update of the multi-source operational data.

[0014] This application also provides an anomaly early warning analysis system for engineering construction muck discharge data, including: a muck discharge behavior dataset construction module, used to collect multi-source operational data related to the muck discharge process during equipment tunneling, the multi-source operational data including screw conveyor muck discharge volume data, muck discharge time series data, earth chamber pressure data, cutterhead torque and speed data, and propulsion speed data, to construct a unified muck discharge behavior dataset; a muck discharge behavior unit division module, used to divide the muck discharge process according to a preset time window based on the muck discharge behavior dataset, discretize the continuous muck discharge process into multiple muck discharge behavior units, and construct a multi-dimensional behavior feature vector for each muck discharge behavior unit; coupling The module for constructing a constraint relationship is used to construct a coupled constraint relationship between muck discharge behavior and earth chamber pressure and tunneling parameters based on the pressure balance mechanism of the earth pressure shield excavation face, forming a dynamic benchmark model for muck discharge behavior; the module for calculating multidimensional deviation is used to input the multidimensional behavioral feature vector of the muck discharge behavior unit into the dynamic benchmark model, calculate the multidimensional deviation of the actual behavior relative to the dynamic benchmark, the multidimensional deviation includes muck discharge volume deviation, rhythm deviation, pressure response deviation, and energy consumption correlation deviation, and fuse the deviations to obtain a comprehensive deviation index; the module for outputting anomaly warnings for muck discharge is used to output anomaly warnings for muck discharge based on the comprehensive deviation index.

[0015] The proposed method and system for abnormal early warning analysis of muck discharge data in engineering construction, as described in this application, firstly collects multi-source operational data related to the muck discharge process during equipment tunneling. This multi-source operational data includes muck discharge volume data from the screw conveyor, muck discharge time series data, earth chamber pressure data, cutterhead torque and speed data, and propulsion speed data, constructing a unified muck discharge behavior dataset. Then, based on this muck discharge behavior dataset, the muck discharge process is divided according to a preset time window, discretizing the continuous muck discharge process into multiple muck discharge behavior units. For each muck discharge behavior unit, a multi-dimensional behavior feature vector is constructed. Then, based on the earth pressure shield tunneling face pressure balance mechanism, a coupling constraint relationship is constructed between muck discharge behavior, earth chamber pressure, and tunneling parameters, forming a dynamic benchmark model of muck discharge behavior. Next, the multi-dimensional behavior feature vector of the muck discharge behavior unit is input into the dynamic benchmark model to calculate the multi-dimensional deviation of the actual behavior relative to the dynamic benchmark. This multi-dimensional deviation includes muck discharge volume deviation, rhythm deviation, pressure response deviation, and energy consumption correlation deviation. These deviations are then fused to obtain a comprehensive deviation index. Finally, an abnormal muck discharge warning is output based on the comprehensive deviation index. Through the above process, the method and system proposed in this application achieve the technical effect of improving the accuracy and real-time performance of slag discharge anomaly monitoring, effectively reducing false alarms and missed alarms, accurately reflecting the overall abnormal state of the slag process, and providing reliable protection for construction safety. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 A flowchart illustrating the abnormal early warning analysis method for construction slag discharge data provided in this application embodiment.

[0018] Figure 2 A schematic diagram of the structure of the abnormal early warning analysis system for construction slag discharge data provided in this application embodiment.

[0019] Figure labeling: 10 for slag discharge behavior dataset construction module, 20 for slag discharge behavior unit partitioning module, 30 for coupling constraint relationship construction module, 40 for multidimensional deviation calculation module, and 50 for slag discharge anomaly early warning output module. Detailed Implementation

[0020] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following describes in detail the specific implementation manners, structures, features, and effects of the present invention in combination with the accompanying drawings and preferred embodiments.

[0021] The embodiment of the present application provides an abnormal warning analysis method for engineering construction mucking data, as Figure 1 shown, the method includes: Step S100, during the tunneling process of the equipment, collect multi-source operation data associated with the mucking process. The multi-source operation data includes the mucking volume data of the screw conveyor, the mucking time series data, the earth pressure data in the soil bin, the cutter head torque and rotation speed data, and the propulsion speed data, and construct a unified mucking behavior data set.

[0022] Specifically, during the tunneling process of the equipment, a multi-source data acquisition system supporting an earth pressure balance shield machine is used to collect multi-source operation data associated with the mucking process. Among them, the mucking volume data of the screw conveyor is collected by a weight sensor自带 by the screw conveyor, and the sensor is installed below the discharge port of the screw conveyor. The collected data is the real-time mucking weight; the mucking time series data is collected by a timestamp module and is collected synchronously with various types of operation data, recording the collection time corresponding to each piece of mucking-related data to form a continuous time series; the earth pressure data in the soil bin is collected by pressure sensors installed inside the soil bin, and the number of sensors is 3, which are evenly distributed on the inner wall of the soil bin. After collection, the average value of the 3 sensors is taken as the current earth pressure data in the soil bin; the cutter head torque and rotation speed data are collected by torque sensors and rotation speed sensors of the cutter head drive system; the propulsion speed data is collected by displacement sensors of the propulsion cylinders. All the above-mentioned collected multi-source operation data is imported into the data acquisition terminal, and a data cleaning algorithm is used to process abnormal data. Specifically, null values and jump values that appear during the collection process are剔除. The determination criterion for jump values is that the difference between two adjacent data points exceeds 30% of the previous data point, and then the format of the cleaned data is standardized to统一 the timestamp format and unit of the data, and finally a unified mucking behavior data set is constructed. The data set is indexed by a time series and stores various types of operation data in an associated manner.

[0023] In one possible implementation, multi-source operational data associated with the slag discharge process is collected. Step S100 further includes step S110, which involves performing a time window self-verification analysis of the multi-source operational data and establishing a self-verification signal. Specifically, a sliding time window algorithm is used to perform the time window self-verification analysis of the multi-source operational data. The time window length is set to 5 minutes, and the sliding step is 1 minute. Consistency verification is performed on each type of multi-source operational data within each time window. The verification process is as follows: First, calculate the average and standard deviation of a certain type of data within each time window. Then, determine whether the deviation of each data point from the average within the time window exceeds 3 times the standard deviation. If the deviation exceeds 3 times the standard deviation, the data point is determined to be suspicious data, and the time window is marked as an abnormal time window. If two or more types of data show suspicious data within the same time window, it is determined that the data collection within the time window is abnormal, and a self-verification abnormal signal is generated. If only a single type of data shows suspicious data or there is no suspicious data, the data collection is determined to be normal, and a self-verification normal signal is generated. Finally, a self-verification signal covering the verification results of each time window is established. The self-verification signal includes the time window identifier, the verification result (normal / abnormal), and suspicious data information.

[0024] Step S120: Configure an additional acquisition strategy based on the self-verification signal to perform a reset acquisition update of the multi-source operating data. Specifically, configure the additional acquisition strategy based on the self-verification signal. If the self-verification signal is normal, maintain the original acquisition strategy and continue to acquire various types of data at the predetermined sampling frequency. If the self-verification signal is abnormal, trigger the additional acquisition strategy, specifically by increasing the sampling frequency corresponding to the abnormal time window to twice the original frequency, and simultaneously adding the acquisition of screw conveyor motor current data as auxiliary verification data through the motor current sensor. After completing the additional acquisition, replace and update the newly acquired data with the original abnormal time window data to achieve a reset acquisition update of the multi-source operating data and ensure the accuracy of the slag discharge behavior dataset.

[0025] Step S200: Based on the slag discharge behavior dataset, the slag discharge process is divided into multiple slag discharge behavior units according to a preset time window. For each slag discharge behavior unit, a multi-dimensional behavior feature vector is constructed. The multi-dimensional behavior feature vector includes the slag discharge rate per unit time calculated based on the slag discharge time series data, the slag discharge stability index calculated based on the slag discharge fluctuation amplitude within the continuous time window, the pressure fluctuation amplitude and pressure recovery time calculated based on the earth chamber pressure data, the energy consumption ratio corresponding to the unit slag discharge calculated based on the cutterhead torque and speed data, and the propulsion-slag discharge matching index constructed based on the ratio relationship between the propulsion speed data and the slag discharge data.

[0026] Specifically, based on the constructed slag discharge behavior dataset, a sliding time window method is used to divide the slag discharge process. For example, the preset time window length can be set to 10 minutes, and the sliding step size is 5 minutes. The continuous slag discharge process is discretized into multiple interconnected and non-overlapping slag discharge behavior units. Each slag discharge behavior unit corresponds to all slag discharge-related operational data within a time window. For each slag discharge behavior unit, feature parameters of each type of operational data are extracted to construct a multi-dimensional behavioral feature vector. The specific process is as follows: the rate of change of slag discharge per unit time is obtained by dividing the difference in slag discharge between two adjacent 1-minute time intervals within the slag discharge behavior unit by the time interval (1 minute); the slag discharge stability index is obtained by calculating the fluctuation amplitude of slag discharge in all 1-minute time intervals within the slag discharge behavior unit and taking the average of the fluctuation amplitudes. The fluctuation amplitude is the absolute difference between the slag discharge in each 1-minute time interval and the average slag discharge in the unit. The average of the fluctuation amplitudes in all time intervals is the slag discharge stability index; The force fluctuation amplitude is the difference between the maximum and minimum values ​​of the pressure in the earth chamber within the slag discharge unit. The pressure recovery time is the time required for the pressure in the earth chamber to recover from a value 10% below the average value to the average value. The energy consumption ratio corresponding to the unit slag discharge is obtained by dividing the total energy consumption of the cutterhead in the slag discharge unit by the total slag discharge. The total energy consumption of the cutterhead is obtained by multiplying the cutterhead torque and rotational speed by time. The propulsion-slag discharge matching index is obtained by the ratio of the average propulsion speed to the average slag discharge in the slag discharge unit. All the characteristic parameters calculated above are integrated to form a multidimensional behavioral characteristic vector of the slag discharge unit.

[0027] Step S300: Based on the earth pressure shield tunneling face pressure balance mechanism, construct the coupling constraint relationship between muck discharge behavior and earth chamber pressure and tunneling parameters to form a dynamic benchmark model of muck discharge behavior. The coupling constraint relationship includes the balance relationship between muck discharge amount and earth chamber pressure, the synchronization relationship between muck discharge rhythm and advance speed, and the correlation relationship between cutterhead energy consumption and muck discharge amount.

[0028] Specifically, based on the pressure balance mechanism of the earth pressure shield tunneling excavation face, and combined with the muck discharge behavior dataset, a coupled constraint relationship between muck discharge behavior, earth chamber pressure, and tunneling parameters is constructed to form a dynamic benchmark model of muck discharge behavior. The specific construction process of the coupled constraint relationship is as follows: Based on the mechanism that earth chamber pressure equals the sum of the earth pressure at the excavation face and the muck discharge resistance of the screw conveyor, and combined with actual construction data, the balance relationship between muck discharge volume and earth chamber pressure is determined. That is, when the muck discharge volume increases, the earth chamber pressure decreases accordingly, and when the muck discharge volume decreases, the earth chamber pressure increases accordingly; the two are negatively correlated. The muck discharge rhythm is as follows: The synchronization relationship between the number of muck removals per unit time, the muck removal rhythm, and the advance speed is as follows: the advance speed and the muck removal rhythm are positively correlated. When the advance speed increases, the muck removal rhythm accelerates synchronously to ensure that the excavated muck is discharged in a timely manner. The cutterhead energy consumption increases with the increase of the muck removal volume. The relationship between the cutterhead energy consumption and the muck removal volume is as follows: the two are positively correlated. The larger the muck removal volume, the greater the muck resistance that the cutterhead needs to overcome, and the higher the energy consumption. By integrating the above three relationships, a coupled constraint relationship between the muck removal behavior, the soil chamber pressure, and the tunneling parameters is formed, thereby constructing a dynamic benchmark model of the muck removal behavior.

[0029] In one possible implementation, a dynamic benchmark model of slag discharge behavior is formed. Step S300 further includes step S310, which, based on the slag discharge behavior dataset, filters historical slag discharge behavior units where the pressure fluctuation of the earth chamber is within a preset stable range and the advance speed changes continuously, thus constructing a benchmark sample set. Specifically, based on the slag discharge behavior dataset, a data filtering algorithm is used to filter historical slag discharge behavior units where the pressure fluctuation of the earth chamber is within a preset stable range and the advance speed changes continuously, thus constructing a benchmark sample set. The preset stable range of the earth chamber pressure is determined according to the construction geological conditions. For example, for silty clay soil, the preset stable range is 0.6 to 0.9 MPa, and the pressure fluctuation amplitude of the earth chamber does not exceed 0.1 MPa, indicating it is within the stable range. The criterion for continuous change in advance speed is that the change amplitude of the advance speed within the slag discharge behavior unit does not exceed 5 mm / min, and there are no sudden stops or sudden increases or decreases. During the screening process, all historical slag discharge behavior units in the slag discharge behavior dataset are traversed first. The soil chamber pressure fluctuation data and propulsion speed data of each unit are extracted. The data are then screened against the above judgment criteria. All slag discharge behavior units that meet the criteria are extracted and integrated to form a benchmark sample set. The benchmark sample set contains the multi-dimensional behavior feature vector of each slag discharge behavior unit that meets the criteria and the corresponding operation data.

[0030] Step S320: Perform intervariate correlation analysis on the multidimensional behavioral features in the benchmark sample set to determine the balance correlation function between slag discharge volume and earth chamber pressure, the synchronization response function between slag discharge rhythm and propulsion speed, and the coupling mapping function between cutterhead energy consumption and slag discharge volume. Establish multivariate coupling constraint relationships based on the intervariate correlation analysis results. Specifically, perform intervariate correlation analysis on the multidimensional behavioral features in the benchmark sample set, using Pearson correlation coefficient analysis combined with multiple linear regression algorithm to determine the correlation relationships between variables. The specific execution process is as follows: First, calculate the Pearson correlation coefficient between every two multidimensional behavioral features in the benchmark sample set to determine the degree of correlation between each feature. A correlation coefficient absolute value above 0.8 indicates a strong correlation, 0.5 to 0.8 indicates a moderate correlation, and below 0.5 indicates a weak correlation. Based on the correlation coefficient analysis results, select variable pairs with strong correlation, and fit them using multiple linear regression algorithm. For slag discharge volume and earth chamber pressure, use... With slag discharge volume as the independent variable and earth chamber pressure as the dependent variable, a multiple linear regression algorithm is used to fit the data and obtain the equilibrium correlation function between slag discharge volume and earth chamber pressure. For slag discharge rhythm and propulsion speed, with propulsion speed as the independent variable and slag discharge rhythm as the dependent variable, a multiple linear regression algorithm is used to fit the synchronous response function. For cutterhead energy consumption and slag discharge volume, with slag discharge volume as the independent variable and cutterhead energy consumption as the dependent variable, a multiple linear regression algorithm is used to fit the coupling mapping function. Based on the above three functions, the quantitative relationships between slag discharge volume and earth chamber pressure, slag discharge rhythm and propulsion speed, and cutterhead energy consumption and slag discharge volume are clarified, and multivariate coupling constraint relationships are established.

[0031] Step S330: Based on the aforementioned coupling constraint relationship, the reference sample sets for different tunneling sections are partitioned and fitted to form multiple local reference sub-models corresponding to different construction states. Specifically, according to the geological conditions, burial depth, and other parameters of the construction section, the entire tunneling process is divided into different tunneling sections, such as a silty clay layer section, a sandy soil layer section, and a gravel layer section. For each tunneling section, all muck removal behavior units belonging to that section are extracted from the reference sample set, serving as the partitioned sample set for that section. Using the coupling constraint relationship function form obtained in S320 as a template, the parameters in the function are refitted for the partitioned sample set of each tunneling section. Each tunneling section corresponds to a local reference sub-model, ultimately resulting in multiple local reference sub-models corresponding to the construction states of each tunneling section.

[0032] Step S340: Based on the matching degree between the characteristics of the slag discharge behavior unit and the local benchmark sub-model, the local benchmark sub-model is invoked and the coupling constraint relationship is dynamically corrected to form a dynamic benchmark model. Specifically, the feature distribution center corresponding to each local benchmark sub-model is obtained. The feature distribution center is the average value of the multidimensional behavior feature vector of the sample set of the corresponding partition of the local benchmark sub-model. For example, the feature distribution center of the local benchmark sub-model of the silty clay layer section is the average value of the feature parameters such as the unit time slag discharge rate change rate and slag discharge stability index of all slag discharge behavior units in that section. Then, the Euclidean distance between the multidimensional behavior feature vector of the current slag discharge behavior unit and the feature distribution center of each local benchmark sub-model is calculated. The reciprocal of the Euclidean distance is used as the matching degree. The smaller the Euclidean distance, the higher the matching degree. Based on the matching degree, local benchmark sub-models are selected and invoked. The local benchmark sub-model with the highest matching degree, i.e., the smallest Euclidean distance, is selected as the reference benchmark model for the current slag removal behavior unit. Then, based on the actual operating data of the current slag removal behavior unit, the coupling constraint relationship of this local benchmark sub-model is dynamically corrected. The correction coefficient is adjusted according to the real-time geological parameters of the current construction section to make the coupling constraint relationship more consistent with the current construction reality. After the correction is completed, this local benchmark sub-model becomes the dynamic benchmark model corresponding to the current slag removal behavior unit. As the slag removal process progresses, the above matching and correction process is performed for each slag removal behavior unit to achieve real-time updates of the dynamic benchmark model.

[0033] In one possible implementation, a dynamic benchmark model is formed. Step S340 further includes step S341, obtaining the feature distribution center corresponding to each local benchmark sub-model. Specifically, the feature distribution center corresponding to each local benchmark sub-model is obtained using the mean calculation method. For each local benchmark sub-model's corresponding partition sample set, the multi-dimensional behavioral feature vector of all slag discharge behavior units is extracted. The average value is calculated for each feature parameter. For example, for the feature of the change rate of slag discharge per unit time, the arithmetic mean of the change rate of slag discharge per unit time of all slag discharge behavior units in the partition sample set is calculated as the distribution center component of this feature. The average values ​​of all feature parameters are integrated to form the feature distribution center of the local benchmark sub-model. The dimension of the feature distribution center is consistent with the dimension of the multi-dimensional behavioral feature vector, and each dimension corresponds to the average value of a feature parameter.

[0034] Step S342: Configure the matching degree based on the distance between the features of the slag discharge behavior unit and the feature distribution center. Specifically, the Euclidean distance calculation method is used to configure the matching degree based on the distance between the features of the slag discharge behavior unit and the feature distribution center. The Euclidean distance between the multi-dimensional behavior feature vector of the current slag discharge behavior unit and the feature distribution center of each local benchmark sub-model is calculated. The calculation process is as follows: first, calculate the difference between the corresponding dimension feature parameters of the two vectors, square each difference, then add all the squared differences, and finally take the square root of the sum to obtain the Euclidean distance. The reciprocal of the Euclidean distance is used as the matching degree. The formula for calculating the matching degree is matching degree = 1 / (1 + Euclidean distance). It is ensured that the value of the matching degree is between 0 and 1. The smaller the Euclidean distance, the closer the matching degree is to 1, indicating that the current slag discharge behavior unit matches the local benchmark sub-model more closely.

[0035] Step S343: Based on the matching degree, select and call local reference sub-models, and perform dynamic correction of coupling constraint relationships. Specifically, sort the matching degrees of all local reference sub-models, and select the local reference sub-model with the highest matching degree as the initial reference model for the current slag removal behavior unit. Then, collect real-time geological parameters and equipment operating parameters of the current construction section, and determine correction coefficients based on these parameters. Multiply each parameter in the coupling constraint relationship of the initial reference model by the corresponding correction coefficient to complete the dynamic correction of the coupling constraint relationship. The corrected coupling constraint relationship is more in line with the current construction conditions. After the correction is completed, this model is the dynamic reference model corresponding to the current slag removal behavior unit.

[0036] Step S400: Input the multidimensional behavior feature vector of the slag discharge behavior unit into the dynamic benchmark model, calculate the multidimensional deviation of the actual behavior relative to the dynamic benchmark, the multidimensional deviation includes slag discharge deviation, rhythm deviation, pressure response deviation and energy consumption correlation deviation, and fuse the deviations to obtain a comprehensive deviation index.

[0037] Specifically, based on the coupling constraints of the dynamic benchmark model, the benchmark characteristic parameters corresponding to the current slag discharge behavior unit are calculated, namely, benchmark slag discharge volume, benchmark slag discharge rhythm, benchmark earth chamber pressure response value, and benchmark cutterhead energy consumption. Then, the difference between the actual characteristic parameters and the benchmark characteristic parameters is calculated, and divided by the benchmark characteristic parameters to obtain the deviation degree of each dimension. For example, if the actual slag discharge volume is 12m³... 3 / min, with a baseline slag discharge rate of 10m³ 3If the actual slag discharge rate is 1 / min, then the deviation is (12-10) / 10 = 0.2, or 20%. The rhythm deviation is the difference between the actual slag discharge rhythm and the benchmark slag discharge rhythm divided by the benchmark slag discharge rhythm. The pressure response deviation is the difference between the actual soil chamber pressure fluctuation amplitude and the benchmark soil chamber pressure fluctuation amplitude divided by the benchmark soil chamber pressure fluctuation amplitude. The energy consumption correlation deviation is the difference between the energy consumption ratio per unit slag discharge and the energy consumption ratio per unit slag discharge, divided by the benchmark energy consumption ratio per unit slag discharge. Positive values ​​for each deviation indicate that the actual value is higher than the benchmark value, while negative values ​​indicate that the actual value is lower than the benchmark value. Then, a weighted summation method is used to integrate each deviation. The weight of each deviation is determined according to the construction safety priority. The larger the absolute value of the comprehensive deviation index, the greater the deviation between the actual slag discharge behavior and the benchmark behavior.

[0038] In one possible implementation, the multidimensional behavioral feature vector of the slag discharge unit is input into the dynamic benchmark model. Step S400 further includes step S410, which calls the comparison processing layer within the dynamic benchmark model to perform residual comparison of the multidimensional behavioral feature vector and establish a standardized residual dataset. Specifically, the comparison processing layer uses a residual calculation algorithm to calculate the difference between each feature parameter in the multidimensional behavioral feature vector of the current slag discharge unit and the corresponding benchmark feature parameter output by the dynamic benchmark model, obtaining the residual for each feature parameter: Residual = Actual Feature Parameter - Benchmark Feature Parameter. Then, all residuals are standardized using the Z-score standardization algorithm to calculate the standardized residual for each residual: Standardized Residual = (Residual - Mean Residual) / Standard Residual, where the mean residual is the arithmetic mean of the residuals of all feature parameters of the current slag discharge unit, and the standard residual is the standard deviation of all residuals. The standardized residuals of all feature parameters are integrated to form a standardized residual dataset.

[0039] Step S420: Align the standardized residual dataset within the same time window to form a deviation combination vector. Specifically, using the time window corresponding to the current slag discharge unit as a benchmark, determine the acquisition time corresponding to each standardized residual in the standardized residual dataset. Sort all standardized residuals according to the order of acquisition time to ensure that the timestamp of each standardized residual is completely aligned with the time series within the time window, avoiding time misalignment. Then, arrange the aligned standardized residuals according to the order of feature parameters to form a vector, which is the deviation combination vector. Each component of the vector corresponds to the standardized residual of a feature parameter.

[0040] Step S430: Determine whether any of the deviation combination vectors exceeds their respective benchmark thresholds. Specifically, a benchmark threshold is set for the standardized residual of each feature parameter. The benchmark threshold is determined based on the statistical data of the benchmark sample set, using the 3-standard-deviation rule, i.e., the benchmark threshold is three times the standard deviation of the standardized residual of the corresponding feature parameter in the benchmark sample set. Each component in the deviation combination vector, i.e., the standardized residual of each feature parameter, is then checked to see if it exceeds the corresponding benchmark threshold. If the absolute value of a component is greater than the corresponding benchmark threshold, it is determined that the deviation corresponding to that component exceeds the benchmark threshold; otherwise, it is determined that it does not exceed it.

[0041] Step S440: When at least two deviation combination vectors simultaneously exceed the corresponding benchmark threshold, a coupling enhancement term is constructed based on the corresponding deviation combination vectors. Specifically, all deviation combination vector components that exceed the corresponding benchmark threshold are selected, the sum of the absolute values ​​of these components is calculated, and then multiplied by the coupling coefficient to obtain the coupling enhancement term. The coupling coefficient is determined based on the number of components exceeding the threshold; when two components exceed the threshold, the coupling coefficient is 1.2, and when three or more components exceed the threshold, the coupling coefficient is 1.5. The coupling enhancement term is used to reflect the synergistic effect of multiple deviation dimensions and enhance the sensitivity of the comprehensive deviation index to multi-dimensional anomalies.

[0042] Step S450: When only a single deviation vector exceeds the corresponding baseline threshold, the original values ​​are retained to construct a univariate deviation term. Specifically, when only a single deviation vector exceeds the corresponding baseline threshold, only the original standardized residual value of the component exceeding the baseline threshold is extracted as the univariate deviation term. Components that do not exceed the threshold are not included in the construction of the univariate deviation term. The univariate deviation term is used to reflect anomalies in a single dimension, avoiding the neglect of anomalies in a single dimension.

[0043] Step S460: Based on the superposition of the coupling enhancement term and the univariate deviation term, a sliding accumulation process is performed within a time window using the superposition result to construct a comprehensive deviation index. Specifically, if a coupling enhancement term exists, it is superimposed with all univariate deviation terms to obtain the superposition result; if no coupling enhancement term exists, and only univariate deviation terms exist, the superposition result is the sum of the univariate deviation terms; if there are neither coupling enhancement terms nor univariate deviation terms, the superposition result is 0. Then, a sliding accumulation algorithm is used, starting from the current time window, sliding forward three time windows, and accumulating and summing the superposition results of these four time windows to obtain the comprehensive deviation index. The sliding accumulation process is used to avoid false alarms caused by accidental anomalies in a single time window, thereby improving the accuracy of the warning.

[0044] In one possible implementation, calculating the multidimensional deviation of the actual behavior relative to the dynamic benchmark, step S400 further includes step S470, performing time series consistency analysis on the multidimensional behavior feature vectors of multiple consecutive slag discharge behavior units to construct a change trend sequence between adjacent time windows. Specifically, using a trend analysis algorithm, five consecutive slag discharge behavior units are selected, and each feature parameter in the multidimensional behavior feature vector of each slag discharge behavior unit is extracted. Then, the difference between the corresponding feature parameters of two adjacent slag discharge behavior units is calculated, and the change trend of the feature parameter is determined according to the sign of the difference: a positive difference indicates an upward trend, a negative difference indicates a downward trend, and a difference of 0 indicates a stable trend. The change trend of each feature parameter between consecutive adjacent time windows is arranged in chronological order to form a change trend sequence of the feature parameter. The change trend sequences of all feature parameters are integrated to obtain a change trend sequence between adjacent time windows. This sequence contains continuous change trend information of each feature parameter and is used to determine whether the change of slag discharge behavior conforms to normal patterns.

[0045] Step S480: The trend sequence is compared with the trend changes of the corresponding variables in the dynamic benchmark model to generate a trend deviation coefficient. Specifically, the benchmark trend changes corresponding to each feature parameter are extracted from the dynamic benchmark model. The benchmark trend changes are determined based on the feature parameter change patterns of continuous slag discharge behavior units in the benchmark sample set. The actual trend changes of each feature parameter in the trend sequence are compared with the corresponding benchmark trend changes, and the ratio of the actual change difference to the benchmark change difference is calculated as the trend deviation coefficient of that feature parameter. If the actual change trend is opposite to the benchmark trend, the trend deviation coefficient is negative. The trend deviation coefficients of all feature parameters are integrated to obtain the overall trend deviation coefficient, which is used to compensate for multidimensional deviations.

[0046] Step S490: Perform multi-dimensional deviation calculation and compensation based on the trend deviation coefficient. Specifically, multiply the trend deviation coefficient of each characteristic parameter by the corresponding deviation to obtain the compensated deviation. For example, if the slag discharge deviation is 0.2 and the corresponding trend deviation coefficient is 0.6, then the compensated slag discharge deviation = 0.2 × 0.6 = 0.12. All compensated deviations are weighted and summed according to predetermined weights to obtain the compensated comprehensive deviation index. By compensating for the trend deviation coefficient, the deviation calculation deviation caused by changes in the slag discharge behavior trend can be corrected, improving the accuracy of the comprehensive deviation index.

[0047] Step S500: Output an abnormal slag discharge warning based on the comprehensive deviation index.

[0048] Specifically, a pre-set warning threshold for the comprehensive deviation index is established. This threshold is determined based on the statistical data of the comprehensive deviation index in the benchmark sample set, using the three-standard-deviation rule, where three times the standard deviation of the comprehensive deviation index in the benchmark sample set is taken as the warning threshold. Then, it is determined whether the currently calculated comprehensive deviation index exceeds the warning threshold. If the absolute value of the comprehensive deviation index is greater than the warning threshold, the current slag discharge behavior is deemed abnormal, and an abnormality warning signal is output. If the absolute value of the comprehensive deviation index is less than or equal to the warning threshold, the current slag discharge behavior is deemed normal, and no warning signal is output. The abnormality warning signal includes the time of the abnormality, the corresponding slag discharge behavior unit, the deviation degree of each dimension, and the value of the comprehensive deviation index, used to prompt staff to handle the situation promptly.

[0049] In one possible implementation, based on the comprehensive deviation index, an abnormal slag discharge warning is output. Step S500 further includes step S510, configuring an abnormal warning signal according to the comprehensive deviation index and establishing a warning command. Specifically, when the comprehensive deviation index exceeds the warning threshold, the warning signal is divided into different levels according to the absolute value of the comprehensive deviation index. The higher the warning level, the more serious the slag discharge abnormality. The corresponding warning signal content is configured according to the warning level. For example, a level 1 warning signal includes anomaly indication and abnormal data details; a level 2 warning signal adds preliminary handling suggestions on top of the level 1 warning; and a level 3 warning signal adds an emergency shutdown indication on top of the level 2 warning. The warning signal content, warning level, and abnormal related data are integrated to form a warning command. The warning command adopts a standardized format to ensure that the warning equipment can accurately identify and execute it.

[0050] Step S520: The warning command is sent to the warning equipment to execute visual warning dispatch management. Specifically, the warning command is sent to the warning equipment at the construction site via industrial Ethernet. The warning equipment includes a site monitoring terminal, an audible and visual alarm, and a handheld terminal for workers. After receiving the warning command, the site monitoring terminal displays information such as the warning level, abnormal data, and location of the abnormality on the screen, using different colors to distinguish the warning levels, such as yellow for level 1, orange for level 2, and red for level 3. The audible and visual alarm issues corresponding audible and visual prompts according to the warning level, such as intermittent yellow light and sound for level 1, intermittent orange light and rapid sound for level 2, and continuous red light and piercing alarm sound for level 3. After receiving the warning command, the handheld terminal for workers displays a warning message and pushes abnormality details and handling suggestions to ensure that workers can be aware of the abnormal slag discharge situation in a timely manner and take appropriate handling measures, thus completing the visual warning dispatch management.

[0051] This application embodiment collects multi-source operational data in real time during earth pressure shield tunneling, including muck discharge volume of the screw conveyor, time series, earth chamber pressure, cutterhead torque and speed, and propulsion speed, and integrates them into a unified muck discharge behavior dataset. The continuous muck discharge process is divided into multiple muck discharge behavior units according to a preset time window. A multi-dimensional behavior feature vector is extracted and constructed for each unit. Based on the pressure balance mechanism of the earth pressure shield excavation face, a coupling constraint relationship is established between muck discharge behavior, earth chamber pressure, and tunneling parameters, forming a dynamic benchmark model adapted to the construction state. The feature vectors of each muck discharge behavior unit are input into this model, and the calculation... This technology calculates and integrates four multi-dimensional deviations—slag volume, rhythm, pressure response, and energy consumption—to obtain a comprehensive deviation index. Based on this comprehensive deviation index, it judges and outputs corresponding slag discharge anomaly warnings. This solves the technical problems of low monitoring accuracy, poor real-time performance, and the tendency to generate false alarms and missed alarms in existing earth pressure shield tunneling slag discharge anomaly monitoring, which cannot accurately reflect the overall abnormal state of the slag discharge process. It achieves the technical effect of improving the accuracy and real-time performance of slag discharge anomaly monitoring, effectively reducing false alarms and missed alarms, accurately reflecting the overall abnormal state of the slag discharge process, and providing reliable protection for construction safety.

[0052] In the above text, refer to Figure 1 This paper describes in detail a method for abnormal early warning analysis of slag discharge data during engineering construction according to an embodiment of the present invention. Next, we will refer to... Figure 2 An abnormal early warning analysis system for slag discharge data in engineering construction is described according to an embodiment of the present invention.

[0053] The abnormal early warning analysis system for construction muck discharge data according to an embodiment of the present invention addresses the technical problems of low monitoring accuracy, poor real-time performance, and susceptibility to false alarms and missed alarms in existing earth pressure shield tunneling muck discharge anomaly monitoring, which fails to accurately reflect the overall abnormal state of the muck discharge process. The system aims to improve the accuracy and real-time performance of muck discharge anomaly monitoring, effectively reduce false alarms and missed alarms, accurately reflect the overall abnormal state of the muck discharge process, and provide reliable assurance for construction safety. The abnormal early warning analysis system for construction muck discharge data includes: a muck discharge behavior dataset construction module 10, a muck discharge behavior unit division module 20, a coupling constraint relationship construction module 30, a multi-dimensional deviation calculation module 40, and a muck discharge anomaly early warning output module 50.

[0054] The slag discharge behavior dataset construction module 10 is used to collect multi-source operational data related to the slag discharge process during equipment excavation. This multi-source operational data includes slag discharge volume data from the screw conveyor, slag discharge time series data, soil chamber pressure data, cutterhead torque and speed data, and propulsion speed data, to construct a unified slag discharge behavior dataset. The slag discharge behavior unit partitioning module 20 is used to partition the slag discharge process according to a preset time window based on the slag discharge behavior dataset, discretizing the continuous slag discharge process into multiple slag discharge behavior units, and constructing a multi-dimensional behavior feature vector for each slag discharge behavior unit. The coupling constraint relationship construction module 30 is used to... Based on the pressure balance mechanism of the earth pressure shield tunneling face, a coupling constraint relationship between muck discharge behavior and earth chamber pressure and tunneling parameters is constructed to form a dynamic benchmark model of muck discharge behavior. A multidimensional deviation calculation module 40 is used to input the multidimensional behavior feature vector of the muck discharge behavior unit into the dynamic benchmark model to calculate the multidimensional deviation of the actual behavior relative to the dynamic benchmark. The multidimensional deviation includes muck discharge volume deviation, rhythm deviation, pressure response deviation, and energy consumption correlation deviation. The deviations are then fused to obtain a comprehensive deviation index. A muck discharge anomaly early warning output module 50 is used to output a muck discharge anomaly early warning based on the comprehensive deviation index.

[0055] The detailed description of the specific configuration of the coupling constraint relationship construction module 30 is explained as follows: As mentioned above, the coupling constraint relationship construction module 30 may further include: the coupling constraint relationship includes the balance relationship between the slag discharge volume and the earth chamber pressure, the synchronization relationship between the slag discharge rhythm and the propulsion speed, and the correlation relationship between the cutterhead energy consumption and the slag discharge volume.

[0056] The dynamic benchmark model for muck discharge behavior, specifically the coupling constraint relationship construction module 30, may further include: a benchmark sample set construction unit, which, based on the muck discharge behavior dataset, selects historical muck discharge behavior units whose soil chamber pressure fluctuations are within a preset stable range and whose advance speed changes continuously, and constructs a benchmark sample set; a coupling constraint relationship establishment unit, which performs intervariate correlation analysis on the multidimensional behavioral features in the benchmark sample set, determines the balance correlation function between muck discharge volume and soil chamber pressure, the synchronization response function between muck discharge rhythm and advance speed, and the coupling mapping function between cutterhead energy consumption and muck discharge volume, and establishes multivariate coupling constraint relationships based on the intervariate correlation analysis results; a partition fitting processing unit, which performs partition fitting processing on the benchmark sample sets of different tunneling sections based on the coupling constraint relationships, forming multiple local benchmark sub-models corresponding to different construction states; and a dynamic benchmark model generation unit, which, based on the matching degree between the characteristics of the muck discharge behavior units and the local benchmark sub-models, calls the local benchmark sub-models and performs dynamic correction of the coupling constraint relationships to form a dynamic benchmark model.

[0057] The dynamic benchmark model generation unit may further include: a feature distribution center acquisition subunit for acquiring the feature distribution center corresponding to each local benchmark submodel; a matching degree configuration subunit for configuring the matching degree based on the distance between the feature of the slag discharge behavior unit and the feature distribution center; and a coupling constraint relationship dynamic correction subunit for filtering and calling local benchmark submodels based on the matching degree and performing dynamic correction of coupling constraint relationships.

[0058] The detailed description of the specific configuration of the multidimensional deviation calculation module 40 is explained as follows: As mentioned above, the multidimensional behavioral feature vector of the slag discharge behavior unit is input into the dynamic benchmark model. The multidimensional deviation calculation module 40 may further include: a residual comparison unit for calling the comparison processing layer in the dynamic benchmark model to perform residual comparison of the multidimensional behavioral feature vector and establish a standardized residual dataset; an alignment processing unit for aligning the standardized residual dataset within the same time window to form a deviation combination vector; a judgment unit for judging whether there are vectors in the deviation combination vector that exceed their respective benchmark thresholds; a coupling enhancement term construction unit for constructing a coupling enhancement term based on the corresponding deviation combination vector when at least two deviation combination vectors exceed the corresponding benchmark thresholds simultaneously; a univariate deviation term construction unit for retaining the original value and constructing a univariate deviation term when only a single deviation combination vector exceeds the corresponding benchmark threshold; and a sliding accumulation processing unit for performing sliding accumulation processing within the time window based on the superposition of the coupling enhancement term and the univariate deviation term to construct a comprehensive deviation index.

[0059] The detailed description of the specific configuration of the slag discharge behavior unit division module 20 is as follows: As mentioned above, the slag discharge behavior unit division module 20 may further include: a multi-dimensional behavior feature vector including the slag discharge rate per unit time calculated based on slag discharge time series data, the slag discharge stability index calculated based on the slag discharge fluctuation amplitude within a continuous time window, the pressure fluctuation amplitude and pressure recovery time calculated based on the earth chamber pressure data, the energy consumption ratio corresponding to the unit slag discharge calculated based on the cutterhead torque and speed data, and the propulsion-slag discharge matching index constructed based on the ratio relationship between propulsion speed data and slag discharge data.

[0060] The detailed description of the specific configuration of the slag discharge anomaly early warning output module 50 is explained as follows: As mentioned above, the slag discharge anomaly early warning output module 50 may further include: an anomaly early warning signal configuration unit for configuring an anomaly early warning signal according to the comprehensive deviation index and establishing an early warning instruction; and a visual early warning dispatch management unit for sending the early warning instruction to the early warning device and performing visual early warning dispatch management.

[0061] The multidimensional deviation calculation module 40, which calculates the multidimensional deviation of the actual behavior relative to the dynamic benchmark, may further include: a time series consistency analysis unit for performing time series consistency analysis on the multidimensional behavior feature vectors of multiple consecutive slag discharge behavior units to construct a trend sequence between adjacent time windows; a trend deviation coefficient generation unit for comparing the trend sequence with the trend changes of corresponding variables in the dynamic benchmark model to generate a trend deviation coefficient; and a multidimensional deviation calculation compensation unit for performing multidimensional deviation calculation compensation based on the trend deviation coefficient.

[0062] The detailed description of the specific configuration of the slag discharge behavior dataset construction module 10 is as follows: As mentioned above, the slag discharge behavior dataset construction module 10 collects multi-source operating data associated with the slag discharge process. The slag discharge behavior dataset construction module 10 may further include: a time window self-verification analysis unit for performing time window self-verification analysis of multi-source operating data and establishing a self-verification signal; and a reset acquisition update unit for configuring additional acquisition strategies according to the self-verification signal and performing reset acquisition update of multi-source operating data.

[0063] The abnormal early warning analysis system for engineering construction slag discharge data provided in this embodiment of the invention can execute the abnormal early warning analysis method for engineering construction slag discharge data provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0064] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for early warning and analysis of abnormal construction waste discharge data, characterized in that, The method includes: During the equipment excavation process, multi-source operational data related to the slag removal process are collected. The multi-source operational data includes slag removal data of the screw conveyor, slag removal time series data, soil chamber pressure data, cutterhead torque and speed data, and propulsion speed data, to construct a unified slag removal behavior dataset. Based on the slag discharge behavior dataset, the slag discharge process is divided according to a preset time window, the continuous slag discharge process is discretized into multiple slag discharge behavior units, and a multi-dimensional behavior feature vector is constructed for each slag discharge behavior unit. Based on the pressure balance mechanism of earth pressure shield tunneling, a coupled constraint relationship between muck removal behavior and earth chamber pressure and tunneling parameters is constructed to form a dynamic benchmark model of muck removal behavior. The multidimensional behavioral feature vector of the slag discharge behavior unit is input into the dynamic benchmark model to calculate the multidimensional deviation of the actual behavior relative to the dynamic benchmark. The multidimensional deviation includes the slag discharge deviation, rhythm deviation, pressure response deviation, and energy consumption correlation deviation. The deviations are then fused to obtain a comprehensive deviation index. An abnormal slag discharge warning is output based on the comprehensive deviation index.

2. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, The coupling constraints include the balance between slag discharge volume and earth chamber pressure, the synchronization between slag discharge rhythm and propulsion speed, and the correlation between cutterhead energy consumption and slag discharge volume.

3. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 2, characterized in that, A dynamic baseline model for slag discharge behavior is established, including: Based on the slag discharge behavior dataset, historical slag discharge behavior units with soil chamber pressure fluctuations within a preset stable range and continuous changes in propulsion speed are selected to construct a benchmark sample set. Perform intervariate correlation analysis on the multidimensional behavioral features in the benchmark sample set to determine the balance correlation function between slag discharge volume and earth chamber pressure, the synchronization response function between slag discharge rhythm and propulsion speed, and the coupling mapping function between cutterhead energy consumption and slag discharge volume. Establish multivariate coupling constraint relationships based on the results of the intervariate correlation analysis. Based on the aforementioned coupling constraint relationship, the benchmark sample sets of different tunneling sections are partitioned and fitted to form multiple local benchmark sub-models corresponding to different construction states. Based on the matching degree between the characteristics of the slag discharge behavior unit and the local benchmark sub-model, the local benchmark sub-model is called and the coupling constraint relationship is dynamically corrected to form a dynamic benchmark model.

4. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 3, characterized in that, The formation of a dynamic benchmark model also includes: Obtain the feature distribution center corresponding to each local benchmark sub-model; The matching degree is configured based on the distance value between the characteristics of the slag discharge behavior unit and the feature distribution center. Based on the matching degree, the local baseline sub-model is selected and invoked, and the coupling constraint relationship is dynamically corrected.

5. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, The multi-dimensional behavioral feature vector of the slag discharge behavior unit is input into the dynamic benchmark model, including: The comparison processing layer within the dynamic benchmark model is invoked to perform residual comparison of multidimensional behavioral feature vectors, thereby establishing a standardized residual dataset. The standardized residual datasets are aligned within the same time window to form a deviation combination vector. Determine whether any of the deviation combination vectors exceed their respective benchmark thresholds; When at least two deviation combination vectors exceed the corresponding baseline thresholds simultaneously, a coupling enhancement term is constructed based on the corresponding deviation combination vectors; When only a single deviation vector exceeds the corresponding baseline threshold, the original value is retained to construct a single-variable deviation term. Based on the superposition of the coupling enhancement term and the univariate deviation term, the superposition result is used to perform sliding accumulation processing within a time window to construct a comprehensive deviation index.

6. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, The multidimensional behavioral feature vector includes the rate of change of slag discharge per unit time calculated based on slag discharge time series data, the slag discharge stability index calculated based on the fluctuation amplitude of slag discharge within a continuous time window, the pressure fluctuation amplitude and pressure recovery time calculated based on earth chamber pressure data, the energy consumption ratio corresponding to the unit slag discharge calculated based on cutterhead torque and speed data, and the propulsion-slag discharge matching index constructed based on the ratio relationship between propulsion speed data and slag discharge data.

7. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, Based on the aforementioned comprehensive deviation index, an abnormal slag discharge warning is output, which also includes: Configure abnormal early warning signals based on the comprehensive deviation index and establish early warning instructions; The warning command is sent to the warning device to perform visual warning issuance management.

8. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, Calculate the multidimensional deviation of actual behavior from the dynamic baseline, including: Time series consistency analysis is performed on the multidimensional behavioral feature vectors of multiple consecutive slag discharge behavioral units to construct a sequence of changing trends between adjacent time windows; A trend deviation coefficient is generated by comparing the trend changes of the aforementioned trend sequence with the trend changes of the corresponding variables in the dynamic benchmark model. Multidimensional deviation calculation and compensation are performed based on the trend deviation coefficient.

9. The method for abnormal early warning analysis of engineering construction slag discharge data as described in claim 1, characterized in that, The collection of multi-source operational data related to the slag removal process also includes: Perform time-window self-verification analysis on multi-source runtime data and establish a self-verification signal; Configure additional acquisition strategies based on the self-verification signal, and perform reset acquisition and update of multi-source running data.

10. An abnormal early warning and analysis system for construction slag discharge data, characterized in that, The system is used to implement the abnormal early warning analysis method for engineering construction slag discharge data according to any one of claims 1-9, and the system includes: The slag discharge behavior dataset construction module is used to collect multi-source operational data related to the slag discharge process during equipment tunneling. The multi-source operational data includes slag discharge data of screw conveyor, slag discharge time series data, soil chamber pressure data, cutterhead torque and speed data, and propulsion speed data to construct a unified slag discharge behavior dataset. The slag discharge behavior unit division module is used to divide the slag discharge process according to a preset time window based on the slag discharge behavior dataset, discretize the continuous slag discharge process into multiple slag discharge behavior units, and construct a multi-dimensional behavior feature vector for each slag discharge behavior unit. The coupling constraint relationship construction module is used to construct the coupling constraint relationship between muck discharge behavior and earth chamber pressure and tunneling parameters based on the earth pressure shield excavation face pressure balance mechanism, forming a dynamic benchmark model of muck discharge behavior; The multidimensional deviation calculation module is used to input the multidimensional behavior feature vector of the slag discharge behavior unit into the dynamic benchmark model, calculate the multidimensional deviation of the actual behavior relative to the dynamic benchmark, the multidimensional deviation includes slag discharge deviation, rhythm deviation, pressure response deviation and energy consumption correlation deviation, and fuse each deviation to obtain a comprehensive deviation index. The slag discharge anomaly early warning output module is used to output an abnormal slag discharge early warning based on the comprehensive deviation index.

Images