Micro-tunnel construction regulation method and system based on multi-source data fusion

Abstract

The application provides a micro pipe jacking construction regulation and control method based on multi-source data fusion, comprising: acquiring pipe jacking equipment operation parameters, geological parameters and environmental data through a multi-source sensor to generate a multi-source data set; performing time-space alignment and data cleaning on the multi-source data set to obtain a standardized data set; extracting features from the standardized data set through a neural network model and performing classification to output a dynamic geological classification result; identifying a jacking resistance level based on the dynamic geological classification result, determining equipment regulation and control parameters through an optimization algorithm; adjusting the equipment regulation and control parameters and matching a mud formula scheme according to real-time monitoring data; triggering model updating and optimizing regulation and control parameters through a prediction error, and adjusting a decision threshold according to a segmented evaluation result.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of information technology, and in particular to a micro-jacking construction regulation method and system based on multi-source data fusion. BACKGROUND

[0002] With the acceleration of urbanization and the complexity of underground space development, this technology has become the key to solving the needs of narrow space and high-precision construction. However, the existing regulation methods mostly rely on manual experience or a single data source, which is difficult to cope with complex and variable stratum conditions and construction environments, resulting in insufficient construction precision, increased equipment wear and tear, and even increased engineering risks. These limitations make traditional methods often show insufficient adaptability when facing diversified geological conditions such as sand layer, clay layer or composite stratum, and cannot achieve dynamic optimization and intelligent decision-making. Under this background, the core challenges of micro-jacking construction regulation mainly focus on how to effectively fuse multi-source data, accurately identify geological conditions, and adaptively adjust regulation parameters. Due to the heterogeneity and real-time nature of multi-source data, it is difficult to unify, and existing systems cannot build accurate working condition identification models, which further affects the dynamic matching of jacking force, rotation speed and mud parameters. In addition, the optimization of regulation strategies lacks continuous learning ability, making it difficult to break away from experience dependence in complex strata, and the decision-making efficiency and adaptability are limited. These technical problems not only increase the uncertainty of construction, but also limit the application potential of micro-jacking technology in a wider range of scenarios. Therefore, how to establish an intelligent working condition identification model based on multi-source data fusion and realize the dynamic optimization of regulation parameters through adaptive algorithms has become a key problem to improve the adaptability and efficiency of micro-jacking construction. Solving this problem requires a breakthrough from data integration to decision optimization to meet the complex needs of different stratum conditions and construction stages. SUMMARY

[0003] The present application provides a micro-jacking construction regulation method based on multi-source data fusion, mainly including:

[0004] Obtain jacking equipment operation parameters, geological parameters and environmental data through multi-source sensors to generate a multi-source data set; perform spatio-temporal alignment and data cleaning on the multi-source data set to obtain a standardized data set; extract features from the standardized data set through a neural network model and perform classification to output a dynamic geological classification result; identify the jacking resistance level based on the dynamic geological classification result and determine the equipment regulation parameters through an optimization algorithm; adjust the equipment regulation parameters and match the mud formula scheme according to real-time monitoring data; update and optimize the regulation parameters through a prediction error triggering model, and adjust the decision threshold according to the segmented evaluation result.

[0005] Further, the acquisition of the pipe jacking equipment operation parameters, geological parameters and environmental data by the multi-source sensor, and the generation of the multi-source data set, comprises: acquiring the equipment operation parameters by the inclination sensor, the earth pressure gauge and the torque instrument installed on the pipe jacking equipment; acquiring the rock and soil mechanics parameters by the geological detector; acquiring the underground water level and surface vibration data by the environmental monitoring device; acquiring the mud viscosity parameters by the rheometer, and generating the formation permeability curve in combination with the data of the permeability tester; and integrating the equipment operation parameters, the rock and soil mechanics parameters and the underground water level and surface vibration data into the multi-source heterogeneous data set.

[0006] Further, the spatio-temporal alignment and data cleaning of the multi-source data set to obtain the standardized data set, comprises: performing spatio-temporal alignment processing on the multi-source data set by using the distributed Kalman filtering algorithm; and removing the abnormal data in the multi-source data set by using the abnormal value elimination algorithm based on the 3σ principle.

[0007] The missing data in the multi-source data set is filled by using the KNN missing value interpolation algorithm; and the standardized data set with a unified time stamp is generated.

[0008] Further, the extraction of features from the standardized data set and the classification by using the neural network model to output the dynamic geological classification result, comprises: inputting the standardized data set into the convolutional recurrent neural network to extract the geological spatial distribution features, the equipment power spectrum features and the environmental time-varying features; establishing a feature correlation matrix for the extracted features by using the fuzzy C-means clustering algorithm; and outputting the dynamic geological classification result containing the sand layer, the clay layer and the composite stratum according to the feature correlation matrix.

[0009] Further, the identification of the jacking resistance level based on the dynamic geological classification result, and the determination of the equipment control parameters by using the optimization algorithm, comprises: inputting the dynamic geological classification result into the pre-trained working condition identification model, wherein the working condition identification model is constructed by using the bidirectional LSTM network containing the attention mechanism; outputting the jacking resistance level label according to the comparison between the rock and soil shear strength characteristic value and the dynamic decision threshold; calling the historical construction database based on the jacking resistance level label, and calculating the Pareto optimal solution set of the jacking force and the rotation speed by using the improved particle swarm algorithm with the equipment power constraint and the jacking speed constraint.

[0010] Further, the adjustment of the equipment control parameters and the matching of the mud formula scheme according to the real-time monitoring data, comprises: real-time monitoring of the jacking force deviation and the rotation speed fluctuation rate; if the ratio of the jacking force deviation and the rotation speed fluctuation rate exceeds the set tolerance, calculating the compensation increment by using the gradient descent optimizer; according to the jacking force after compensation and the mud viscosity-formation permeability curve, searching the formula scheme with the highest matching degree from the pre-established mud formula knowledge base; and updating the knowledge base parameters based on the implementation effect of the mud formula.

[0011] Further, the model updating triggered by the prediction error and the optimization of the regulation parameter include: dividing real-time construction data flow into sample blocks according to time windows; if the prediction error rate of the working condition recognition model on the continuous three sample blocks exceeds a preset model error threshold, updating the network weight by using an elastic weight solidification algorithm; and outputting the updated model as prior knowledge into a feature correlation matrix, and generating an iteratively optimized regulation parameter set through heterogeneous data fusion.

[0012] Further, the adjustment of the stratum disturbance threshold and the equipment decision threshold according to the segmented evaluation result includes: when the jacking distance reaches a segmented threshold, acquiring the soil stress change amount of each monitoring point and a weight coefficient; calculating a stratum disturbance coefficient according to the soil stress change amount, the weight coefficient and an average jacking force; and dynamically adjusting the decision threshold and a tolerance ratio parameter of a subsequent stage according to the stratum disturbance coefficient; and the set tolerance is obtained according to the stratum disturbance coefficient of a previous construction segment.

[0013] The application further provides a micro pipe jacking construction regulation system for implementing any method in claims 1-8, and the system comprises:

[0014] A multi-source data acquisition module is configured to acquire pipe jacking equipment operation parameters, geological parameters and environmental data through multi-source sensors, and generate a multi-source data set;

[0015] A data preprocessing module is configured to perform time-space alignment and data cleaning on the multi-source data set, and obtain a standardized data set;

[0016] A dynamic geological classification module is configured to extract features from the standardized data set and perform classification through a neural network model, and output a dynamic geological classification result;

[0017] A resistance regulation module is configured to identify a jacking resistance grade based on the dynamic geological classification result, and determine equipment regulation parameters through an optimization algorithm;

[0018] A real-time decision module is configured to adjust the equipment regulation parameters and match a mud formula scheme according to real-time monitoring data;

[0019] A model iteration module is configured to trigger model updating and optimize regulation parameters through a prediction error;

[0020] A threshold optimization module is configured to adjust a decision threshold according to a segmented evaluation result.

[0021] The technical scheme provided by the embodiments of the application can have the following beneficial effects:

[0022] The application discloses a micro pipe jacking construction regulation method based on multi-source data fusion. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 A flowchart of the micro pipe jacking construction regulation method based on multi-source data fusion. DETAILED DESCRIPTION

[0024] In order to further understand the content of the application, the application will be described in detail in conjunction with the drawings and examples. It can be understood that the specific examples described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that, in order to facilitate description, only the parts related to the application are shown in the drawings.

[0025] As Figure 1 The micro pipe jacking construction regulation method based on multi-source data fusion can specifically include the following steps.

[0026] S101, the multi-source sensor is used to obtain the pipe jacking equipment operation parameter, geological parameter and environmental data, and a multi-source data set is generated, including: the equipment operation parameter is obtained by the inclination sensor, soil pressure gauge and torque instrument installed on the pipe jacking equipment; the rock and soil mechanics parameter is obtained by the geological detector; the underground water level and ground vibration data are obtained by the environmental monitoring device; the mud viscosity parameter is obtained by the rheometer, and the formation permeability curve is generated in combination with the data of the permeability tester; and the equipment operation parameter, the rock and soil mechanics parameter and the underground water level and ground vibration data are integrated into a multi-source heterogeneous data set.

[0027] The equipment operation parameter is obtained by the inclination sensor, soil pressure gauge and torque instrument installed on the pipe jacking equipment. These sensors can monitor the state of the equipment in real time. The inclination sensor is used to measure the inclination angle of the equipment, so that the pipe jacking machine can keep horizontal or at a preset angle during tunneling, and construction quality problems caused by angle deviation can be avoided. For example, during pipe jacking tunneling, the inclination sensor monitors that the inclination angle of the equipment is 0.5°, which exceeds the allowable range of 0.3°, so the system will immediately issue an alarm and adjust the equipment posture.

[0028] Soil pressure gauges are used to measure the pressure exerted by the soil on the equipment during excavation, helping to determine the stability of the ground. For example, when the soil pressure gauge shows a pressure value of 0.8 MPa, it indicates that the soil layer is relatively dense, and the excavation speed needs to be adjusted to prevent the equipment from being overloaded. Torque meters are used to monitor the rotational torque of the equipment, reflecting the working load of the equipment. When the torque value reaches the set threshold, the system will automatically reduce the excavation speed to avoid equipment damage.

[0029] Geological exploration instruments are used to obtain rock and soil mechanics parameters, including the shear strength, compression modulus, and internal friction angle of the soil layer. For example, the geological exploration instrument measures the shear strength of the soil layer at a certain depth to be 50 kPa, the compression modulus to be 10 MPa, and the internal friction angle to be 25°, which are used to evaluate the stability of the ground and the difficulty of excavation.

[0030] Geological exploration instruments use ultrasonic or electromagnetic wave technology to detect the structure of the ground, generating a ground profile for scientific construction. For example, in the clay layer, the exploration instrument shows that the ground structure is uniform, while in the sand layer, there are local loose areas that need to be reinforced.

[0031] Environmental monitoring devices are used to obtain groundwater level and surface vibration data, which are used to evaluate the impact of construction on the surrounding environment. Groundwater level monitoring devices measure the water level depth through pressure sensors. For example, the water level is measured to be 1.5 m at a depth of 10 m, indicating that the area has abundant groundwater, and dewatering measures need to be taken.

[0032] Surface vibration monitoring devices measure vibration intensity through acceleration sensors. For example, the vibration acceleration is measured to be 0.1g at the surface, indicating that the construction has a small impact on surrounding buildings, but continuous monitoring is needed to prevent vibration from intensifying.

[0033] Rheometers are used to obtain mud viscosity parameters, and combined with the data of the permeability tester, a ground permeability curve is generated. Rheometers are used to measure the rheological properties of mud. For example, the mud viscosity is measured to be 30 cP, indicating that the mud has good fluidity and is suitable for use in sand layers.

[0034] Permeability testers generate permeability curves by measuring the permeation speed of mud in the ground. For example, the permeability is measured to be 10^-6 m / s in the clay layer, indicating that the mud has poor permeability, and the mud formula needs to be adjusted to improve the permeation effect.

[0035] The equipment operation parameters, rock and soil mechanics parameters, and groundwater level and surface vibration data are integrated into a multi-source heterogeneous data set. These data come from different sources and have different formats, and need to be uniformly processed. For example, equipment operation parameters are stored in time series, rock and soil mechanics parameters are stored in tables, and groundwater level data are stored in continuous signals.

[0036] Through data cleaning and format conversion, these data are integrated into a unified dataset, providing a basis for subsequent analysis. The integrated dataset can comprehensively reflect the equipment status, formation characteristics and environmental impact during construction, providing support for dynamic geological classification and parameter optimization.

[0037] In S102, the multi-source dataset is subjected to spatio-temporal alignment and data cleaning to obtain a standardized dataset, including: using a distributed Kalman filtering algorithm to process the multi-source dataset for spatio-temporal alignment; removing abnormal data in the multi-source dataset by an outlier elimination algorithm based on the 3σ principle;

[0038] filling in missing data in the multi-source dataset by a KNN missing value interpolation algorithm; and generating a standardized dataset with a unified timestamp.

[0039] The distributed Kalman filtering algorithm is used for spatio-temporal alignment processing of the multi-source dataset. The core idea is to synchronize the data of multiple sensors in time and space, eliminating the inconsistency of data caused by different sampling frequencies or data transmission delays. For example, in the process of pipe jacking construction, the inclination sensor and the soil pressure meter may collect data at a frequency of one per second and one per two seconds respectively, and there may be a delay of milliseconds in the transmission process. Through distributed Kalman filtering, these data can be unified to a time reference, ensuring that the data for subsequent analysis has a consistent timestamp. This processing method can effectively avoid analysis errors caused by time misalignment and improve the reliability of data.

[0040] The outlier elimination algorithm based on the 3σ principle is used to remove abnormal data in the multi-source dataset. The 3σ principle is a statistical method that assumes data follows a normal distribution and considers data within ±3 standard deviations of the mean to be normal, and data outside this range to be outliers. For example, when collecting torque data of pipe jacking equipment, if a measurement value is significantly higher than the historical mean and exceeds the 3σ range, it can be determined that the data is an outlier and is removed. This method can effectively filter out false data caused by sensor failure or external interference, ensuring the accuracy and consistency of the dataset.

[0041] The KNN missing value imputation algorithm is used to fill in the missing data in the multi-source data set. The KNN algorithm calculates the distance between the missing data points and the neighboring data points, selects the nearest K data points for weighted averaging or directly takes the mean value to fill in the missing values. For example, when collecting groundwater level data, due to sensor failure or data transmission interruption, some time points may be missing. Through the KNN algorithm, the missing values can be reasonably estimated according to the water level data of the adjacent time points, ensuring the integrity of the data set. This method can reduce the analysis bias caused by missing data and improve data utilization. Generating a standardized data set with uniform timestamps is an important goal of data processing.

[0042] Through the above steps, the multi-source data set is aligned in time, the abnormal data is removed, and the missing data is filled in, and finally a standardized data set is generated. For example, in the monitoring system of pipe jacking construction, the inclination, soil pressure, torque and other data are stored with a time stamp of one second after processing, forming a complete and consistent data set. This standardized data set provides a high-quality basis for subsequent feature extraction and analysis, improving the accuracy of geological classification and equipment control.

[0043] S103, the features are extracted from the standardized data set by the neural network model and classified, and the dynamic geological classification result is output, including: inputting the standardized data set into a convolutional recurrent neural network to extract geological spatial distribution features, equipment power spectrum features and environmental time-varying features; a feature correlation matrix is established for the extracted features by fuzzy C-means clustering algorithm; and the dynamic geological classification result containing sand layer, clay layer and composite stratum is output according to the feature correlation matrix.

[0044] The standardized data set is input into a convolutional recurrent neural network to extract geological spatial distribution features, equipment power spectrum features and environmental time-varying features. The convolutional recurrent neural network combines the spatial feature extraction capability of convolutional neural network and the time series processing capability of recurrent neural network, which can effectively capture the spatial distribution law of geological data, the power change trend in the equipment running process and the time-varying characteristics of environmental parameters. For example, the spatial correlation of rock and soil mechanics parameters can be extracted by convolution layer for geological spatial distribution features, the periodic change of torque meter and soil pressure gauge data can be analyzed by recurrent neural network for equipment power spectrum features, and the dynamic change of underground water level and ground vibration can be captured by time series modeling for environmental time-varying features. This multi-dimensional feature extraction method can fully reflect the geological conditions and equipment running state, providing a reliable basis for subsequent classification.

[0045] The convolutional recurrent neural network includes an input layer, a convolutional layer, a recurrent layer, and an output layer. In the input layer, the standardized dataset has a dimension of time steps x number of features, which includes geological parameters (such as shear strength, compression modulus), equipment parameters (jacking force, torque), and environmental parameters (groundwater level, vibration acceleration). The convolutional layer uses a 3-layer one-dimensional convolution kernel (kernel size = 5, step = 1) with a ReLU activation function, and the output feature map has a dimension of time steps x 32. The recurrent layer includes bidirectional LSTM units (hidden layer node number = 64), and the attention mechanism focuses on key parameters such as rock-soil shear strength by calculating feature weights. The output layer uses a Softmax classifier to output the probability distribution of sand layer, clay layer, and composite stratum. During training, the Adam optimizer (learning rate = 0.001) is used, and the loss function is cross-entropy. The training dataset includes 100 groups of historical construction data covering different geological conditions.

[0046] A feature correlation matrix is established for the extracted features using the fuzzy C-means clustering algorithm. Fuzzy C-means clustering is a soft clustering method that allows data points to belong to multiple categories with certain membership degrees, making it suitable for situations with fuzzy boundaries in geological classification. For example, in the extracted geological spatial distribution features, the rock-soil mechanics parameters of certain areas may have both sand layer and clay layer characteristics. Fuzzy C-means clustering can construct a feature correlation matrix by calculating the similarity between features, and clearly define the correlation between features. This clustering method can effectively handle the uncertainty in geological classification and improve the accuracy of the classification results.

[0047] The dynamic geological classification results including sand layer, clay layer, and composite stratum are output based on the feature correlation matrix. The feature correlation matrix reflects the correlation between different features, and combined with the comprehensive analysis of geological spatial distribution features, equipment power spectrum features, and environmental time-varying features, it can achieve dynamic identification of geological categories. For example, when the feature correlation matrix shows that the geological spatial distribution features of a certain area are highly correlated with sand layer, and the equipment power spectrum features and environmental time-varying features also support this conclusion, the system will classify that area as sand layer. For composite stratum, the feature correlation matrix will show that the feature correlation of multiple categories is relatively high, and thus output the classification result of composite stratum. This dynamic classification method can adapt to changes in complex geological conditions and provide accurate geological information support for pipe jacking construction.

[0048] S104, the dynamic geological classification result is input into a pre-trained working condition recognition model, the working condition recognition model is constructed by using a bidirectional LSTM network containing an attention mechanism; according to the comparison between the rock-soil shear strength characteristic value and the dynamic decision threshold, the jacking resistance level label is output; based on the jacking resistance level label, the historical construction database is called, and the improved particle swarm algorithm with device power constraint and jacking speed constraint is used to calculate the Pareto optimal solution set of jacking force and rotating speed.

[0049] When the dynamic geological classification result is input into the pre-trained working condition recognition model, the specific content of the dynamic geological classification result needs to be determined first. Dynamic geological classification usually includes rock-soil type, stratum structure, groundwater state and other multi-dimensional information, which is collected and processed in real time by sensors or geological exploration equipment. The pre-trained working condition recognition model uses a bidirectional LSTM network containing an attention mechanism, which can capture the time dependence and spatial correlation in geological data. The introduction of attention mechanism enables the model to pay more attention to key features, such as rock-soil shear strength characteristic value, thereby improving the accuracy of recognition. For example, when there is a high shear strength rock-soil layer in the stratum, the model will focus on analyzing the data in this area through the attention mechanism, ensuring that the output jacking resistance level label is more accurate.

[0050] According to the comparison between the rock-soil shear strength characteristic value and the dynamic decision threshold, the jacking resistance level label is output. The process involves setting the dynamic decision threshold. The dynamic decision threshold is usually based on historical construction data and geological conditions to adapt to different working conditions. For example, when the rock-soil shear strength characteristic value exceeds the threshold, the model will output a higher jacking resistance level label, indicating that the current stratum has a larger resistance to jacking equipment, and appropriate control measures need to be taken. The key to this step is the reasonable setting of the dynamic decision threshold to ensure that the model can accurately identify the resistance level under different geological conditions.

[0051] When the jacking resistance level label is called based on the historical construction database, the historical construction database contains the construction parameters and effect records under similar geological conditions in the past. By analyzing these data, the best control parameters of the equipment under the current resistance level can be determined. For example, when the jacking resistance level is high, historical data may show that jacking force needs to be increased or rotating speed needs to be reduced to improve construction efficiency. This step makes full use of the experience accumulation of historical data to provide a reliable reference for current construction.

[0052] When calculating the Pareto optimal solution set of jacking force and rotation speed using the improved particle swarm algorithm with device power constraints and jacking speed constraints, the improved particle swarm algorithm introduces constraints on device power and jacking speed to ensure that the parameters obtained meet the construction requirements and do not exceed the safe operating range of the device. The objective function of the improved particle swarm algorithm is minFjacking force + ω⋅Δτtorque, and the constraint conditions are: P < 50 KW, V ≥ 0.2 m / min, where ω is the weight coefficient, which is determined by historical data fitting (ω = 0.3).

[0053] The Pareto optimal solution set refers to finding a balance point between multiple objectives, such as finding the optimal combination between jacking force and rotation speed to maximize construction efficiency. For example, when the jacking resistance level is high, the algorithm may recommend a larger jacking force and a lower rotation speed to ensure that the device can smoothly penetrate the stratum while avoiding device overload. This step achieves precise control of device parameters through optimization algorithms, improving the safety and efficiency of construction.

[0054] S105, adjusting the device control parameters and matching the mud formulation scheme according to the real-time monitoring data, including: real-time monitoring of jacking force deviation and rotation speed fluctuation rate; if the ratio of the jacking force deviation and the rotation speed fluctuation rate exceeds the set tolerance, calculate the compensation increment through the gradient descent optimizer; according to the compensated jacking force, combine the mud viscosity-stratum permeability curve, and retrieve the formulation scheme with the highest matching degree from the pre-established mud formulation knowledge base; update the knowledge base parameters based on the mud formulation implementation effect.

[0055] Real-time monitoring of jacking force deviation and rotation speed fluctuation rate is a key link in device control. Through sensors, real-time data of jacking force F and rotation speed ω are collected to calculate the deviation ΔF and fluctuation rate δ_ω. For example, when the real-time value of jacking force F is 1000 kN and the set value is 950 kN, the deviation ΔF is 50 kN; at the same time, the fluctuation rate of rotation speed ω is 5%, i.e. δ_ω = 0.05. By monitoring these data in real time, abnormal situations in device operation can be discovered in a timely manner, providing a basis for subsequent adjustments.

[0056] If the ratio of jacking force deviation and rotation speed fluctuation rate exceeds the set tolerance, start the gradient descent optimizer to calculate the compensation increment ΔF'. For example, set the tolerance to 1000, when ΔF / δ_ω = 50 / 0.05 = 1000, it just reaches the upper limit of the tolerance. At this time, the gradient descent optimizer determines the compensation increment ΔF' by calculating the partial derivative of the stratum resistance prediction error function E ∂E / ∂F. Assuming that the optimizer calculates the compensation coefficient α = 0.1, then ΔF' = 0.1 × ∂E / ∂F, thus achieving accurate compensation of jacking force.

[0057] According to the compensated jacking force F+ΔF', combined with the mud viscosity-stratum permeability curve, the highest matching formula scheme is retrieved from the pre-established mud formula knowledge base. For example, the compensated jacking force is 1050 kN, combined with the current stratum permeability of 0.02 mD and the mud viscosity of 30 mPa·s, the formula scheme of bentonite content β=8% is retrieved through the fuzzy inference rule of the knowledge base. This step ensures the high matching of the mud formula and the stratum characteristics, and improves the construction efficiency.

[0058] Updating the knowledge base parameters based on the implementation effect of the mud formula is an important link in the optimization process. For example, in actual construction, if it is found that the permeation effect of the current mud formula is better than expected, the relevant data (such as stratum porosity φ=0.15, bentonite content β=8%) can be fed back to the knowledge base to update the parameter relationship β=k·ln(1 / φ), where β is the bentonite content in the mud, φ is the stratum porosity, and k is the action coefficient selected according to the construction experience value.

[0059] By continuously optimizing the knowledge base parameters, the matching accuracy of subsequent formula schemes can be improved, providing more reliable guidance for construction. The above steps form a closed-loop control mechanism through real-time monitoring, optimization calculation, formula matching and knowledge base updating. For example, during construction, if the stratum characteristics change, the equipment parameters and mud formula can be quickly adjusted through real-time monitoring data to ensure the smooth progress of construction. This dynamic adjustment mechanism not only improves the stability and efficiency of equipment operation, but also reduces the construction risk, providing strong support for construction under complex geological conditions.

[0060] S106, the model update triggered by the prediction error and the optimization of the control parameters, including: dividing the real-time construction data stream into sample blocks according to the time window; if the prediction error rate of the working condition recognition model on the continuous 3 sample blocks exceeds the preset model error threshold, update the network weight using the elastic weight hardening algorithm; the updated model output is injected into the feature correlation matrix as prior knowledge, and the iterative optimized control parameter set is generated through heterogeneous data fusion.

[0061] Dividing the real-time construction data stream into sample blocks according to the time window is to convert the continuous data stream into discrete time periods for analysis. For example, in pipe jacking construction, data every 10 minutes can be taken as a sample block, including thrust, torque, speed, etc. This segmentation method can effectively capture the dynamic changes in the construction process, facilitating subsequent model analysis and error calculation. If a too short time window is used, the data fluctuation may be too large; a too long time window may mask the key changes. Therefore, reasonable selection of the time window length is the key to ensuring the accuracy of model analysis.

[0062] If the prediction error rate of the working condition recognition model on the last three sample blocks exceeds the preset model error threshold, for example, the error rate exceeds 5%, the model updating mechanism is triggered. The error rate here is calculated by comparing the model predicted value with the actual value. The error rate of the last three sample blocks exceeds the standard, indicating that the model's performance under the current working condition is consistently poor, which may be due to a significant change in the working condition or the model parameters are no longer applicable. At this time, the network weights are updated using the elastic weight solidification algorithm, which can quickly adapt to the new working condition without destroying the existing knowledge of the model. For example, in pipe jacking construction, when the stratum changes from a sandy layer to a clay layer, the prediction error rate of the model may increase significantly. At this time, updating the network weights using the elastic weight solidification algorithm can make the model better adapt to the new stratum conditions. Updating the network weights of the working condition recognition model using the elastic weight solidification algorithm can protect the important weights of previous tasks while learning the current new task, to overcome the problem of catastrophic forgetting of the working condition recognition model during the updating process. In the elastic weight solidification algorithm, the importance matrix of the old task is calculated (L is the loss function, the network weights); the new task loss function adds a regularization term: where λ = 0.5.

[0063] Injecting the updated model output into the feature correlation matrix as prior knowledge is to use the latest prediction results of the model to optimize the subsequent control parameter generation. For example, in the updated model, if it is found that the sensitivity of the thrust to the stratum disturbance has significantly increased, this information can be injected into the feature correlation matrix as an important basis for subsequent control parameter generation.

[0064] The fuzzy C-means clustering algorithm calculates the feature correlation weight through the membership function to generate the feature correlation matrix where is the membership of feature i to class j.

[0065] Mapping the newly connected sensor data into the feature space through principal component analysis can unify data from different sources into the same feature space, facilitating subsequent analysis. For example, mapping the data of temperature sensors and pressure sensors into the same feature space can better capture the correlation between them.

[0066] Generating an iteratively optimized control parameter set is achieved by combining the updated model output and the heterogeneous data fusion result to dynamically adjust the construction parameters. For example, in pipe jacking construction, if the updated model predicts a high stratum disturbance coefficient, the thrust speed and cutter head speed can be adjusted to reduce the disturbance. This iterative optimization method can ensure that the construction parameters are always matched with the current working condition, improving construction efficiency and safety.

[0067] By dynamically adjusting the decision threshold and tolerance ratio, the performance of the model can be further optimized. For example, when the stratum disturbance coefficient is high, the tolerance ratio can be appropriately reduced to improve the sensitivity of the model to anomalies.

[0068] S107, the stratum disturbance threshold and the equipment decision threshold are adjusted according to the segment evaluation results, including: when the jacking distance reaches the segment threshold, the soil stress change amount and the weight coefficient of each monitoring point are obtained; the stratum disturbance coefficient is calculated according to the soil stress change amount, the weight coefficient and the average jacking force; the decision threshold and the tolerance ratio parameters of the subsequent stage are dynamically adjusted according to the stratum disturbance coefficient; and the tolerance is obtained according to the stratum disturbance coefficient of the previous construction segment.

[0069] When the jacking distance reaches the segment threshold, the soil stress change amount and the weight coefficient of each monitoring point are obtained. During the jacking construction process, the stratum will be disturbed to different degrees, and the soil stress change amount is collected in real time through the monitoring points arranged at different positions. For example, when the jacking distance reaches every 100 meters, the system automatically extracts the soil stress change amounts Δσ_A, Δσ_B and Δσ_C of monitoring points A, B and C, and allocates weight coefficients v_A, v_B and v_C according to the importance of each monitoring point. Monitoring point A is located in the key stratum area, and the weight coefficient v_A is set to 0.5; monitoring points B and C are located in the secondary area, and the weight coefficients v_B and v_C are set to 0.3 and 0.2 respectively. In this way, the stratum disturbance situation can be comprehensively reflected, and accurate data can be provided for subsequent calculation.

[0070] The stratum disturbance coefficient is calculated according to the soil stress change amount and the weight coefficient and the average jacking force. After obtaining the soil stress change amount and the weight coefficient of each monitoring point, the stratum disturbance coefficient K_d is calculated in combination with the average jacking force P_avg of the current construction segment. For example, the soil stress change amount Δσ_A of monitoring point A is 50 kPa, the soil stress change amounts of monitoring points B and C are 30 kPa and 20 kPa respectively, and the average jacking force P_avg is 1000 kN, then K_d=(50×0.5+30×0.3+20×0.2) / 1000=0.036. The stratum disturbance coefficient K_d reflects the degree of stratum disturbance, and the larger the K_d value, the more serious the stratum disturbance, and more strict construction control measures need to be taken.

[0071] The decision threshold and tolerance ratio parameters of the subsequent stage are dynamically adjusted according to the formation disturbance coefficient. Based on the calculated formation disturbance coefficient K_d, the system dynamically adjusts the decision threshold epsilon and the tolerance ratio parameter of the subsequent construction stage. For example, when K_d is 0.036, the system adjusts the decision threshold epsilon from the default value 0.05 to 0.04, and the tolerance ratio parameter from 1.2 to 1.1. This adjustment can better adapt to the changes of the formation conditions, avoid overloading of equipment or decline of construction quality due to excessive formation disturbance. By dynamically adjusting the parameters, the safety and efficiency of construction can be improved, and the construction risk can be reduced.

[0072] The tolerance is set according to the formation disturbance coefficient of the previous construction section. After each construction section, the system will reset the tolerance according to the formation disturbance coefficient K_d of the previous construction section. For example, if the K_d of the previous construction section is 0.036, the system will set the tolerance to ±5kPa. This tolerance setting method based on the formation disturbance coefficient can more accurately control the construction parameters and avoid construction deviations caused by sudden changes in formation conditions. By dynamically adjusting the tolerance, the construction process can be more stable, and the construction quality can be improved.

[0073] Embodiments of the present application also provide a micro-tube construction regulation system for implementing any of the methods of claims 1-8, characterized in that it comprises:

[0074] A multi-source data acquisition module for acquiring pipe jacking equipment operation parameters, geological parameters and environmental data through multi-source sensors to generate a multi-source data set;

[0075] A data preprocessing module for spatio-temporal alignment and data cleaning of the multi-source data set to obtain a standardized data set;

[0076] A dynamic geological classification module for extracting features from the standardized data set and classifying them through a neural network model to output a dynamic geological classification result;

[0077] A resistance regulation module for identifying the jacking resistance level based on the dynamic geological classification result and determining the equipment regulation parameters through an optimization algorithm;

[0078] A real-time decision module for adjusting the equipment regulation parameters and matching the mud formulation scheme according to real-time monitoring data;

[0079] A model iteration module for triggering model updating and optimizing the regulation parameters through prediction error;

[0080] A threshold optimization module for adjusting the decision threshold according to the segmented evaluation result.

[0081] Although the present application has been described in detail with general description and specific embodiments above, it is obvious to those skilled in the art that some modifications or improvements can be made on the basis of the present application. Therefore, these modifications or improvements made on the basis of not deviating from the spirit of the present application, all belong to the scope of protection claimed by the present application.

Claims

1. A micro-tunneling construction regulation method based on multi-source data fusion, characterized in that, The application relates to a dynamic geological classification method for pipe jacking equipment, and belongs to the field of pipe jacking construction. Obtaining pipe jacking equipment operation parameters, geological parameters and environmental data through a multi-source sensor to generate a multi-source data set; Performing time-space alignment and data cleaning on the multi-source data set to obtain a standardized data set; Extracting features from the standardized data set through a neural network model and performing classification to output a dynamic geological classification result; Based on the dynamic geological classification result, the jacking resistance level is identified, and the equipment control parameters are determined through an optimization algorithm; specifically, the dynamic geological classification result is input into a pre-trained working condition identification model, the working condition identification model is constructed by using a bidirectional LSTM network containing an attention mechanism; according to the comparison between the rock-soil shear strength characteristic value and the dynamic decision threshold, the jacking resistance level label is output; based on the jacking resistance level label, the historical construction database is called, and the improved particle swarm algorithm with equipment power constraint and jacking speed constraint is used to calculate the Pareto optimal solution set of the jacking force and the rotating speed; Adjusting the equipment control parameters and matching the mud formula scheme according to the real-time monitoring data; specifically, the jacking force deviation and the rotating speed fluctuation rate are monitored in real time; if the ratio of the jacking force deviation and the rotating speed fluctuation rate exceeds the set tolerance, the compensation increment is calculated through the gradient descent optimizer; according to the compensated jacking force, the mud viscosity-permeability curve is combined to search the formula scheme with the highest matching degree from the pre-established mud formula knowledge base; the knowledge base parameters are updated based on the mud formula implementation effect; Triggering model updating and optimization of the control parameters through prediction error; Adjusting the decision threshold according to the segmented evaluation result.

2. The micro-tunneling construction regulating method according to claim 1, characterized in that, The method comprises the following steps: Obtaining equipment operation parameters through an inclination sensor, a soil pressure gauge and a torque instrument installed on the pipe jacking equipment; Obtaining rock-soil mechanics parameters through a geological detector; Obtaining underground water level and ground vibration data through an environmental monitoring device; Obtaining mud viscosity parameters through a rheometer, and generating a formation permeability curve in combination with the data of a permeability tester; Integrating the equipment operation parameters, the rock-soil mechanics parameters and the underground water level and ground vibration data into a multi-source heterogeneous data set.

3. The micro-tunneling construction regulation method of claim 1, wherein, The method comprises the following steps: Performing time-space alignment processing on the multi-source data set by using a distributed Kalman filtering algorithm; Removing abnormal data in the multi-source data set through an abnormal value elimination algorithm based on the 3sigma principle; Filling in missing data in the multi-source data set through a KNN missing value interpolation algorithm; Generating a standardized data set with a unified time stamp.

4. The micro-tunneling construction regulation method according to claim 3, wherein, The method comprises the following steps: Inputting the standardized data set into a convolutional recurrent neural network to extract geological spatial distribution features, equipment power spectrum features and environmental time-varying features; Establishing a feature correlation matrix for the extracted features through a fuzzy C-means clustering algorithm; Outputting a dynamic geological classification result containing sand layers, clay layers and composite strata according to the feature correlation matrix.

5. The micro-tunneling construction regulation method according to claim 4, wherein, The model update triggered by the prediction error and the optimization of the regulation and control parameters comprise: Divide the real-time construction data stream into sample blocks according to time windows; If the prediction error rate of the working condition recognition model on the continuous three sample blocks exceeds the preset model error threshold, update the network weight by using the elastic weight solidification algorithm; The updated model output is injected into the feature correlation matrix as prior knowledge, and the iterative optimized regulation and control parameter set is generated through heterogeneous data fusion.

6. The micro-tunneling construction regulation method of claim 1, wherein, The adjustment of the stratum disturbance threshold and the equipment decision threshold according to the segmented evaluation results comprises: When the jacking distance reaches the segmented threshold, the soil stress change amount and the weight coefficient of each monitoring point are obtained; According to the soil stress change amount and the weight coefficient and the average jacking force, the stratum disturbance coefficient is calculated; According to the stratum disturbance coefficient, the decision threshold and the tolerance ratio parameter of the subsequent stage are dynamically adjusted; The tolerance is calculated according to the stratum disturbance coefficient of the previous construction section.

7. A micro-tunneling construction regulation system for implementing any one of the methods of claims 1-6, characterized by, Comprise: A multi-source data acquisition module is configured to acquire pipe jacking equipment operation parameters, geological parameters and environmental data through multi-source sensors to generate a multi-source data set; A data preprocessing module is configured to perform time-space alignment and data cleaning on the multi-source data set to obtain a standardized data set; A dynamic geological classification module is configured to extract features from the standardized data set and perform classification through a neural network model to output a dynamic geological classification result; A resistance regulation and control module is configured to identify the jacking resistance level based on the dynamic geological classification result and determine the equipment regulation and control parameters through an optimization algorithm; A real-time decision module is configured to adjust the equipment regulation and control parameters and match the mud formula scheme according to real-time monitoring data; A model iteration module is configured to trigger model update and optimize regulation and control parameters through prediction error; A threshold optimization module is configured to adjust the decision threshold according to the segmented evaluation results.

Images