Muddy clay pile foundation drilling parameter regulation method and system

Abstract

The application discloses a kind of silt clay pile foundation drilling parameter regulation methods and systems, belong to pile foundation drilling construction technical field, method includes: collection and pretreatment silt clay pile foundation drilling well historical construction data, constructs data set;Based on data set construction and training machine learning model, obtain drilling parameter time series prediction model;Real-time acquisition of actual drilling parameter in the process of drilling, and compared with the predicted drilling parameter of model output, determine drilling parameter error;Based on drilling parameter error, the drilling parameter of rig is dynamically adjusted.The application method realizes the intelligent control method of key parameters such as drilling pressure, rotational speed and mud concentration, to improve the efficiency and hole quality of pile foundation construction in complex soft soil layer, reduce construction risk, ensure engineering safety.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of pile foundation drilling construction, and particularly relates to a method and system for adjusting and controlling parameters of pile foundation drilling in silt clay. BACKGROUND

[0002] With the rapid development of urbanization construction, pile foundation, as an important form of deep foundation, is widely used in various engineering projects such as high-rise buildings, bridges and ports. However, when pile foundation construction is carried out in complex soft soil foundation such as silt clay, due to the characteristics of high water content, low bearing capacity, poor permeability and loose structure of the stratum, a series of problems such as low drilling efficiency, poor hole quality, difficult control of sediment and hole wall collapse are often encountered. These problems not only affect the construction progress, but also directly threaten the bearing capacity and safety of the pile foundation.

[0003] The existing pile foundation drilling construction mainly relies on the experience of construction personnel to adjust the drilling pressure, rotation speed and mud ratio to adapt to different stratum conditions. However, due to the complex and drastic changes in stratum distribution, the limited and lagging information acquisition at the construction site makes the adjustment lag behind the stratum changes, and it is impossible to achieve fine and dynamic control. At the same time, the existing monitoring methods mostly focus on single parameters such as drilling pressure or mud flow rate, and cannot fully utilize multi-source and multi-dimensional data to comprehensively perceive and model the stratum characteristics, resulting in limited control effect.

[0004] In recent years, with the development of sensing technology and data analysis technology, multi-modal sensors can simultaneously acquire mechanical, fluid, vibration, electrical and other information during drilling, and machine learning has shown superior performance in pattern recognition, prediction and optimization control. However, the deep combination of multi-modal sensing and machine learning in pile foundation drilling construction is still in its infancy, and a mature and efficient dynamic control system has not yet been formed.

[0005] Therefore, there is an urgent need for a method that can fully utilize multi-modal sensors to acquire multi-dimensional parameters during drilling in real time, use machine learning methods to accurately model and dynamically predict stratum characteristics and drilling state, and intelligently control key parameters such as drilling pressure, rotation speed, mud ratio and sediment thickness, in order to improve the efficiency and hole quality of pile foundation construction in complex soft soil stratum, reduce construction risks and ensure engineering safety. SUMMARY

[0006] In view of the above deficiencies in the prior art, the method and system for adjusting and controlling parameters of pile foundation drilling in silt clay provided by the present application solve the problems of low utilization rate of collected data, lagging response of parameter adjustment and insufficient intelligent level of adjustment in the prior art, thereby improving the efficiency and hole quality of pile foundation drilling construction.

[0007] In order to achieve the above-mentioned purposes, the technical scheme adopted by the present application is as follows: a method for adjusting drilling parameters of a sludge clay pile foundation, comprising the following steps:

[0008] S100, collecting and preprocessing historical construction data of sludge clay pile foundation drilling to construct a data set;

[0009] S200, constructing and training a machine learning model based on the data set to obtain a drilling parameter time series prediction model;

[0010] S300, collecting actual drilling parameters in real time during the drilling process and comparing them with the predicted drilling parameters output by the model to determine the drilling parameter error;

[0011] S400, dynamically adjusting the drilling parameters of the drilling rig based on the drilling parameter error.

[0012] Further, the step S100 comprises the following sub-steps:

[0013] S101, collecting historical construction data of sludge clay pile foundation drilling, including drilling pressure, drilling speed and mud bentonite concentration of each construction;

[0014] S102, normalizing and standardizing the collected data;

[0015] S103, performing feature fusion on the normalized and standardized data to obtain a principal component data set;

[0016] S104, performing feature selection and transformation on the data in the principal component data set to construct a data set.

[0017] Further, the step S200 comprises:

[0018] The machine learning model is an LSTM model;

[0019] Each LSTM model learns the long and short dependency relationship of drilling pressure, drilling speed and mud bentonite concentration in the time series in the data set, and the time series features learned by each LSTM model are fused to obtain a drilling parameter time series prediction model.

[0020] Further, in the step S300, the drilling parameter error includes drilling speed error, drilling pressure error and mud bentonite concentration error;

[0021] The drilling speed error is:

[0022]

[0023] In the formula, represents the drilling speed predicted by the drilling parameter time series prediction model, represents the actual collected drilling speed, represents the number of data collection groups;

[0024] The drilling pressure error is:

[0025]

[0026] In the formula, represents the model predicted drilling pressure, represents the actual collected drilling pressure, represents the number of data collection groups;

[0027] The mud bentonite concentration error is:

[0028]

[0029] In the formula, represents the model predicted mud bentonite concentration, represents the actual collected mud bentonite concentration, and N represents the number of data collection groups.

[0030] Further, in the step S400, the dynamic adjustment includes short-term adjustment and medium-term adjustment;

[0031] The short-term adjustment is to compare the next time drilling pressure and drilling speed predicted by the drilling parameter time series prediction model with the actual collected current drilling pressure and drilling speed, and then adjust the drilling pressure and drilling speed.

[0032] The medium-term adjustment is to construct a reward function of a reinforcement learning model according to the drilling parameter error, and then adjust the mud bentonite concentration based on the reinforcement learning model.

[0033] Further, the formula for adjusting the drilling pressure is:

[0034]

[0035]

[0036] In the formula, represents the target adjustment drilling pressure at the next time, represents the actual collected current drilling pressure, represents the predicted next time drilling pressure, represents the allowed drilling pressure prediction error threshold;

[0037] The formula for adjusting the drilling speed is:

[0038]

[0039]

[0040] wherein, denotes the target adjusted drilling speed at next time, denotes the drilling speed adjustment value, denotes the actual collected current drilling speed, denotes the predicted output drilling speed at next time, denotes the allowed drilling speed prediction error threshold.

[0041] Further, the reward function is:

[0042]

[0043] wherein, , and denote the weight coefficients of the drilling pressure prediction error, the drilling speed prediction error and the mud bentonite concentration prediction error, respectively; denote the weight coefficients of the drilling pressure prediction error, the drilling speed prediction error and the mud bentonite concentration prediction error, respectively; , and denote the weight coefficients of the drilling pressure prediction error, the drilling speed prediction error and the mud bentonite concentration prediction error, respectively;

[0044] The method for adjusting the mud bentonite concentration is: at each time t, adjusting the mud bentonite concentration according to the action output by the reinforcement learning model based on the reward function The adjustment formula of the bentonite concentration is:

[0045]

[0046]

[0047] =

[0048] wherein, denotes the current error of the mud bentonite concentration, denotes the target adjusted bentonite concentration at next time, denotes the predicted output bentonite concentration at next time, denotes the actual collected current mud bentonite concentration, denotes the allowed bentonite concentration prediction error threshold.

[0049] A mucky clay pile foundation drilling parameter regulation system comprises:

[0050] A model construction module is configured to construct a drilling parameter time sequence prediction model based on historical construction data of mucky clay pile foundation drilling, wherein the drilling parameters include drilling pressure, drilling speed and mud bentonite concentration.

[0051] A data acquisition module is configured to acquire actual drilling parameters.

[0052] a data comparison module configured to compare the actual drilling parameters with the predicted drilling parameters output by the model to determine drilling parameter errors;

[0053] a parameter adjustment module configured to dynamically adjust the drilling parameters of the drilling rig according to the determined drilling parameter errors.

[0054] Further, the model construction module comprises a data processing unit and a model construction unit.

[0055] The data processing unit is configured to sequentially perform standardization / normalization processing, feature fusion processing, and feature selection and transformation processing on the historical construction data to construct a data set.

[0056] The model construction unit is configured to learn long-short dependency relationships of the drilling pressure, drilling speed, and mud bentonite concentration in the time series in the data set by using LSTM models, and fuse the time series features learned by the LSTM models to obtain a drilling parameter time series prediction model.

[0057] Further, the parameter adjustment module comprises a short-term adjustment unit and a medium-term adjustment unit.

[0058] The short-term adjustment unit is configured to compare the drilling pressure and drilling speed predicted by the drilling parameter time series prediction model at the next time with the actual drilling pressure and drilling speed collected at the current time, and then adjust the drilling pressure and drilling speed at the next time.

[0059] The medium-term adjustment unit is configured to construct a reward function of a reinforcement learning model according to the drilling parameter errors, and adjust the mud bentonite concentration based on the reinforcement learning model.

[0060] The present application has the following advantages:

[0061] (1) Intelligent dynamic regulation of pile foundation drilling parameters: by integrating multi-modal sensors, real-time acquisition of drilling pressure, drilling speed and other multi-dimensional key parameters during drilling, construction of a high-dimensional "configuration-spatiotemporal feature matrix", and combination with machine learning models, the drilling pressure, rotation speed, and mud bentonite concentration of the drilling rig can be predicted and closed-loop regulated, thereby significantly improving the drilling efficiency and hole quality.

[0062] (2) Enhanced adaptability to complex soft soil strata: by training deep learning models with historical construction data and multi-physical field simulation data, using graph convolution networks to capture the relationship between strata depths, and combining LSTM models and incremental learning techniques, the model's understanding and real-time adaptability to strata structures can be improved, overcoming the challenges posed by complex strata changes.

[0063] (3) Construct a closed-loop control and continuous optimization mechanism: the application realizes the integrated closed-loop control of prediction, feedback and fine-tuning in the construction process, combines the short-term (second-level) and medium-term (minute-level) dynamic adjustment mechanism, continuously optimizes the control strategy by using reinforcement learning and incremental learning, realizes the self-adaptive evolution of the model and the long-term stable improvement of the construction quality. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 A flowchart of the sludge clay pile foundation drilling parameter regulation method provided by the application. DETAILED DESCRIPTION

[0065] The specific embodiments of the application are described below to facilitate those skilled in the art to understand the application, but it should be clear that the application is not limited to the scope of the specific embodiments, and for those skilled in the art, it is obvious that various changes are within the spirit and scope of the application defined and determined by the appended claims, and all applications utilizing the concept of the application are within the scope of protection.

[0066] Example 1:

[0067] The embodiment of the application provides a sludge clay pile foundation drilling parameter regulation method;

[0068] Referring to Figure 1 , the method comprises the following steps:

[0069] S100, collecting and preprocessing the historical construction data of the sludge clay pile foundation drilling well, and constructing a data set;

[0070] S200, constructing and training a machine learning model based on the data set to obtain a drilling parameter time series prediction model;

[0071] S300, collecting the actual drilling parameters of the drilling process in real time, and comparing them with the predicted drilling parameters output by the model to determine the drilling parameter error;

[0072] S400, dynamically adjusting the drilling parameters of the drilling rig based on the drilling parameter error.

[0073] The step S100 of the embodiment of the application comprises the following sub-steps:

[0074] S101, collecting the historical construction data of the sludge clay pile foundation drilling well, including the drilling pressure, drilling speed and mud bentonite concentration of each construction;

[0075] Specifically, more than 100,000 groups of historical construction data are collected, covering sludge clay (N value 10-25), including drilling pressure (P), drilling speed (V) and bentonite concentration (Q) in mud and other data;

[0076] S102, normalize and standardize the collected data;

[0077] Specifically, for the data with large range of WOB and ROP values, Z-score standardization is performed; for the data with small range of bentonite concentration in mud, Min-Max normalization is performed; wherein, the formula of Z-score standardization is:

[0078]

[0079] In the formula, represents the original data, represents the mean of the data, represents the standard deviation of the data, represents the standardization result;

[0080] The formula of Min-Max normalization is:

[0081]

[0082] In the formula, represents the original data, represents the maximum value of the original data, represents the minimum value of the original data, represents the normalized result;

[0083] S103, feature fusion is performed on the normalized and standardized data to obtain a principal component data set;

[0084] Specifically, through principal component analysis (PCA), the covariance matrix of the data matrix of WOB, ROP and bentonite concentration in mud is calculated, and then eigenvalue decomposition is performed to obtain eigenvectors and eigenvalues, and the first k principal components with the largest eigenvalues are selected to retain most of the variances, and the original data is projected onto these principal components to obtain the reduced data, realizing the conversion of multi-dimensional features into a principal component set, reducing redundant information while retaining main features. Wherein, the covariance matrix of the data matrix is:

[0085]

[0086] In the formula, represents the covariance matrix, represents the total number of samples, represents the eigenvector of the i-th sample;

[0087] S104, feature selection and transformation are performed on the data in the principal component data set to construct a data set;

[0088] Specifically, Pearson correlation coefficient is adopted, and the features with high correlation are selected as the model input, and redundant or noise data are removed; wherein, the Pearson correlation coefficient is:

[0089]

[0090] If Close to 1 or -1, it indicates that the two features have a strong linear relationship.

[0091] The step S200 of the embodiment of the application comprises:

[0092] The machine learning model is an LSTM model;

[0093] The long and short dependency relationship of the WOB, the ROP and the mud shale concentration in the time sequence in the data set is learned by an LSTM model respectively, and the time sequence features learned by the LSTM models are fused to obtain the drilling parameter time sequence prediction model.

[0094] Specifically, in the embodiment, the process that the LSTM model learns the time sequence mode of the WOB, the ROP and the mud shale concentration in the drilling process is as follows:

[0095] The input gate:

[0096]

[0097] The forget gate:

[0098]

[0099] The candidate memory cell:

[0100]

[0101] Memory update:

[0102]

[0103] The output gate:

[0104]

[0105] Final output:

[0106]

[0107] Wherein, is the input gate control signal, is the input of the current time step WOB, mud flow rate and other parameters, is the hidden state of the last time, is the forget gate control signal, Sigmoid activation function range 0~1, is the current candidate memory (mapped by tanh), is the memory cell state at the current moment, is the output gate control signal, is the weight matrix input to the gate unit, is the weight matrix of the previous hidden state to the gate unit, is the bias term of the gate unit, is the hyperbolic tangent function with an output range of-1~.

[0108] In the embodiment, the time sequence features learned by each LSTM model are fused, and the prediction output of the model is:

[0109]

[0110] In the formula, indicates the fused features, indicates the weight matrix, indicates the bias term, indicates the prediction output of the model.

[0111] In step S300 of the embodiment of the application, in the actual drilling process, the drilling pressure, the drilling speed and the bentonite concentration data in the mud collected by the mud pump outlet are compared with the prediction results of the machine learning model, and the drilling parameter error is calculated, including the drilling speed error, the drilling pressure error and the mud bentonite concentration error;

[0112] The drilling speed error is :

[0113]

[0114] In the formula, indicates the drilling speed predicted by the drilling parameter time sequence prediction model, indicates the actually collected drilling speed, indicates the number of data collection groups;

[0115] The drilling pressure error is :

[0116]

[0117] In the formula, indicates the drilling pressure predicted by the model, indicates the actually collected drilling pressure, indicates the number of data collection groups;

[0118] The mud bentonite concentration error is :

[0119]

[0120] In the formula, represents the model predicted bentonite concentration in the mud, represents the actually collected mud bentonite concentration, and N represents the number of data collection groups.

[0121] In step S400 of the embodiment of the present application, the parameters and the model predicted by machine learning are compared with the actually collected data, the drilling parameters of the drilling rig are dynamically adjusted, and real-time adjustment and closed-loop control are realized.

[0122] The dynamic adjustment in the embodiment includes short-term adjustment and medium-term adjustment.

[0123] The short-term adjustment is to compare the predicted output drilling pressure and drilling speed of the next moment according to the drilling parameter time series prediction model with the actually collected current drilling pressure and drilling speed, and then adjust the drilling pressure and drilling speed; the medium-term adjustment is to construct a reward function of a reinforcement learning model according to the drilling parameter error, and then adjust the mud bentonite concentration based on the reinforcement learning model.

[0124] The formula for adjusting the drilling pressure is:

[0125]

[0126]

[0127] In the formula, represents the target adjusted drilling pressure of the next moment, represents the actually collected current drilling pressure, represents the predicted output drilling pressure of the next moment, represents the allowed drilling pressure prediction error threshold value.

[0128] The formula for adjusting the drilling speed is:

[0129]

[0130]

[0131] In the formula, represents the target adjusted drilling speed of the next moment, represents the drilling speed adjustment value, represents the actually collected current drilling speed, represents the predicted output drilling speed of the next moment, represents the allowed drilling speed prediction error threshold value.

[0132] In the embodiment, in the medium-term regulation process, the bentonite concentration of the mud is regulated, the regulation range is ±0.5%, a reinforcement learning (RL) method is used to dynamically regulate the concentration of the bentonite in the mud, and a reward function is designed by comprehensively considering the drilling pressure error, the drilling speed error and the bentonite concentration error in the mud, so as to minimize the loss function and ensure the stability of the drilling process; wherein the reward function is:

[0133]

[0134] In the formula, and respectively represent the drilling pressure prediction error, the drilling speed prediction error and the bentonite concentration prediction error in the mud, are weight coefficients of and, and according to the importance of the drilling pressure, the drilling speed and the bentonite ratio in the mud in the overall optimization target, a is taken as 1, b is taken as 2, and g is taken as 0.5.

[0135] In the embodiment, the method for regulating the bentonite concentration of the mud is that at each time t, the action output by the reinforcement learning model based on the reward function is used to regulate the bentonite concentration The regulation formula is:

[0136]

[0137]

[0138] =

[0139] In the formula, represents the current error of the bentonite concentration in the mud, represents the target regulation bentonite concentration at the next time, represents the predicted output bentonite concentration at the next time, represents the actually collected current bentonite concentration in the mud, represents the allowed bentonite concentration prediction error threshold.

[0140] In the embodiment of the application, based on the above regulation mode, a feedback channel is further provided to compare the prediction and the actual drilling parameters, realize error feedback correction and model fine-tuning, the cuttings bed thickness is fed back in real time by the ultrasonic hole measuring instrument / laser radar, and the hole wall deformation is distinguished by the geological radar system; when the parameter error is greater than the threshold (for example, the drilling pressure prediction deviation is greater than 10%), the model fine-tuning is automatically triggered; and an incremental learning mechanism is adopted to continuously optimize the model performance in the drilling process, and the adaptability to complex formations is enhanced.

[0141] Embodiment 2:​​​​

[0142] The embodiment is further limited on the basis of embodiment 1, and aims to provide a mucky clay pile foundation drilling parameter regulation system, which is realized based on the mucky clay pile foundation drilling parameter regulation method in embodiment 1, and other parts not mentioned are referred to embodiment 1 or prior art.

[0143] The mucky clay pile foundation drilling parameter regulation system in the embodiment comprises:

[0144] The model construction module is configured to construct a drilling parameter time sequence prediction model based on historical construction data of mucky clay pile foundation drilling, and the drilling parameters include drilling pressure, drilling speed and mud bentonite concentration.

[0145] The data acquisition module is configured to acquire actual drilling parameters.

[0146] The data comparison module is configured to compare the actual drilling parameters with the predicted drilling parameters output by the model to determine drilling parameter errors.

[0147] The parameter adjustment module is configured to dynamically adjust the drilling parameters of the drilling rig according to the determined drilling parameter errors.

[0148] In the embodiment of the present application, the model construction module comprises a data processing unit and a model construction unit.

[0149] The data processing unit is configured to sequentially perform standardization / normalization processing, feature fusion processing, and feature selection and transformation processing on the historical construction data to construct a data set.

[0150] The model construction unit is configured to learn the long-short dependency relationship of drilling pressure, drilling speed and mud bentonite concentration in the time sequence in the data set through an LSTM model, and fuse the time sequence features learned by each LSTM model to obtain a drilling parameter time sequence prediction model.

[0151] In the embodiment of the present application, the parameter adjustment module comprises a short-term adjustment unit and a medium-term adjustment unit.

[0152] The short-term adjustment unit is configured to compare the drilling pressure and drilling speed predicted by the drilling parameter time sequence prediction model at the next moment with the current drilling pressure and drilling speed actually acquired, and then adjust the drilling pressure and drilling speed at the next moment.

[0153] The medium-term adjustment unit is configured to construct a reward function of a reinforcement learning model according to the drilling parameter errors, and adjust the mud bentonite concentration based on the reinforcement learning model.

[0154] Specifically, in the short-term adjustment unit, the formula for adjusting the drilling pressure is:

[0155]

[0156]

[0157] In the formula, represents the target adjusted WOB at the next moment, represents the actually collected current WOB, represents the predicted output WOB at the next moment, represents the allowable WOB prediction error threshold;

[0158] The formula of the ROP adjustment is:

[0159]

[0160]

[0161] In the formula, represents the target adjusted ROP at the next moment, represents the ROP adjustment value, represents the actually collected current ROP, represents the predicted output ROP at the next moment, represents the allowable ROP prediction error threshold.

[0162] In this embodiment, in the medium-term adjustment process, the mud bentonite concentration is adjusted, the adjustment range is ±0.5%, the reinforcement learning (RL) method is used to dynamically adjust the mud bentonite concentration, and the reward function is designed by comprehensively considering the WOB error, the ROP error, and the mud bentonite concentration error, the loss function is minimized, and the stability of the drilling process is ensured; wherein the reward function is:

[0163]

[0164] In the formula, , and respectively represent the WOB prediction error, the ROP prediction error, and the mud bentonite concentration prediction error, respectively represent the weight coefficients of , and , according to the importance of the WOB, the ROP, and the mud bentonite ratio in the overall optimization target, α is 1, β is 2, and γ is 0.5.

[0165] In this embodiment, the method for adjusting the mud bentonite concentration is: at each moment t, the action output by the reinforcement learning model based on the reward function is used to adjust the bentonite concentration The adjustment formula is:

[0166]

[0167] .

[0168] =

[0169] In the formula, represents the current error of the mud bentonite concentration, represents the target adjusted bentonite concentration at the next moment, represents the predicted output of the bentonite concentration at the next moment, represents the current mud bentonite concentration actually collected, represents the allowed bentonite concentration prediction error threshold.

[0170] In the embodiment of the present application, based on the above adjustment mode, a feedback channel is further arranged to compare the predicted and measured drilling parameters, so as to realize error feedback correction and model fine-tuning. The cuttings thickness is fed back in real time by the ultrasonic hole measuring instrument / laser radar, and the hole wall deformation is distinguished by the geological radar system. When the parameter error is greater than the threshold (for example, the drilling pressure prediction deviation is greater than 10%), the model fine-tuning is automatically triggered. And the incremental learning mechanism is adopted to continuously optimize the model performance in the drilling process, and to enhance the adaptability to complex formations.

[0171] The principles and implementation manners of the present application are described by specific embodiments in the present application, and the above embodiment descriptions are only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed, and the above description should not be understood as the limitation of the present application.

[0172] Those skilled in the art will understand that the embodiments described herein are for the purpose of helping the reader to understand the principles of the present application, and should be understood as the protection scope of the present application not being limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspirations disclosed in the present application without departing from the essence of the present application, and these modifications and combinations still fall within the protection scope of the present application.

Claims

1. A method for adjusting drilling parameters for silty clay pile foundations, characterized in that, Includes the following steps: S100. Collect and preprocess historical construction data of silty clay pile foundation drilling to construct a dataset; S200. Based on the dataset, a machine learning model is constructed and trained to obtain a drilling parameter time series prediction model; S300: Real-time acquisition of actual drilling parameters during the drilling process, and comparison with the predicted drilling parameters output by the model to determine the drilling parameter error; the drilling parameters include drilling pressure, drilling speed, and mud bentonite concentration; S400: Dynamically adjust the drilling parameters of the drilling rig based on drilling parameter errors; The dynamic adjustment includes both short-term and medium-term adjustments. The short-term adjustment is as follows: the drilling pressure and drilling speed predicted by the drilling parameter time series prediction model for the next moment are compared with the actual current drilling pressure and drilling speed, and then the drilling pressure and drilling speed are adjusted accordingly. The intermediate adjustment is as follows: a reward function for a reinforcement learning model is constructed based on the drilling parameter error, and then the concentration of bentonite in the drilling mud is adjusted based on the reinforcement learning model. The formula for adjusting drilling pressure is as follows: In the formula, This indicates the target adjustment of drilling pressure at the next moment. This indicates the current drilling pressure that was actually collected. This indicates the drilling pressure at the next moment in the predicted output. This indicates the allowable threshold for drilling pressure prediction error; The formula for adjusting the drilling rate is: In the formula, This indicates the target adjustment rate for drilling at the next moment. This indicates the drilling speed adjustment value. This indicates the current drilling speed that was actually collected. This indicates the predicted drilling speed at the next moment. This indicates the allowable drilling speed prediction error threshold; The method for adjusting the bentonite concentration in the mud is as follows: at each time t, based on the action output by the reinforcement learning model based on the reward function... The formula for adjusting the bentonite concentration is as follows: = In the formula, This indicates the current error in the bentonite concentration of the mud. This indicates the target adjustment for bentonite concentration at the next moment. This indicates the predicted bentonite concentration for the next time step. This indicates the actual current bentonite concentration in the collected mud. This indicates the allowable error threshold for bentonite concentration prediction.

2. The method for adjusting drilling parameters for silty clay pile foundations according to claim 1, characterized in that, Step S100 includes the following sub-steps: S101. Collect historical construction data for drilling of silty clay pile foundations, including drilling pressure, drilling speed and bentonite concentration in mud for each operation; S102. Normalize and standardize the collected data; S103. Perform feature fusion on the normalized and standardized data to obtain the principal component dataset; S104. Perform feature selection and transformation on the data in the principal component dataset to construct the dataset.

3. The method for adjusting drilling parameters for silty clay pile foundations according to claim 1, characterized in that, Step S200 includes: The machine learning model is constructed as an LSTM model; The time series dependence of drilling pressure, drilling speed and mud bentonite concentration in the dataset is learned by using an LSTM model, and the time series features learned by each LSTM model are fused to obtain a drilling parameter time series prediction model.

4. The method for adjusting drilling parameters for silty clay pile foundations according to claim 1, characterized in that, In step S300, the drilling parameter errors include drilling speed error, drilling pressure error, and mud bentonite concentration error. The drilling rate error for: In the formula, This represents the drilling rate predicted by the drilling parameter time-series prediction model. This indicates the actual drilling speed. Indicates the number of data collection groups; The drilling pressure error for: In the formula, This represents the drilling pressure predicted by the model. This indicates the actual drilling pressure collected. Indicates the number of data collection groups; The error in the concentration of bentonite in the mud for: In the formula, This represents the bentonite concentration in the mud predicted by the model. This indicates the actual concentration of bentonite in the collected mud, and N represents the number of data collection groups.

5. The method for adjusting drilling parameters for silty clay pile foundations according to claim 1, characterized in that, The reward function is: In the formula, , and These represent the errors in drill pressure prediction, drill speed prediction, and mud bentonite concentration prediction, respectively. They represent , and The weighting coefficients.

6. A drilling parameter control system for silty clay pile foundations, implemented based on the drilling parameter control method for silty clay pile foundations according to any one of claims 1 to 5, characterized in that, include: Model building module: used to build a time series prediction model for drilling parameters based on historical construction data of silty clay pile foundation drilling; The drilling parameters include drilling pressure, drilling speed, and bentonite concentration in the drilling mud. Data acquisition module: used to collect actual drilling parameters; Data comparison module: used to compare actual drilling parameters with predicted drilling parameters output by the model to determine drilling parameter errors; Parameter adjustment module: Used to dynamically adjust the drilling parameters of the drilling rig based on the determined drilling parameter error.

7. The drilling parameter control system for silty clay pile foundations according to claim 6, characterized in that, The model building module includes a data processing unit and a model building unit; The data processing unit is used to perform standardization / normalization, feature fusion, and feature selection and transformation on historical construction data in sequence to construct a dataset; The model building unit is used to learn the time-series dependence of drilling pressure, drilling speed and mud bentonite concentration in the dataset through LSTM models, and to fuse the time-series features learned by each LSTM model to obtain a drilling parameter time-series prediction model.

8. The drilling parameter control system for silty clay pile foundations according to claim 6, characterized in that, The parameter adjustment module includes a short-term adjustment unit and a medium-term adjustment unit; The short-term adjustment unit is used to compare the drilling pressure and drilling speed at the next moment predicted by the drilling parameter time-series prediction model with the actual current drilling pressure and drilling speed, and then adjust the drilling pressure and drilling speed at the next moment. The intermediate adjustment unit is used to construct a reward function for a reinforcement learning model based on drilling parameter errors, and to adjust the bentonite concentration in the drilling mud based on the reinforcement learning model.

Images