CN-122020017-A - TBM cutter disc rotating speed prediction method based on vibration characteristic driving

CN122020017ACN 122020017 ACN122020017 ACN 122020017ACN-122020017-A

Abstract

The invention relates to the field of shield cutter head related test equipment, in particular to a TBM cutter head rotating speed prediction method based on vibration characteristic driving. The TBM cutterhead rotating speed prediction method based on vibration characteristic driving is characterized in that TBM vibration and tunneling data are collected, the collected data are preprocessed, characteristic engineering is conducted on the vibration data, the first 15 key characteristics are selected as model input parameters through correlation calculation, and data training is conducted through random forests and XGBoost models.

Inventors

WAN XUEYU
Gu Tianxin
CHEN QIAO
LI SHUYUAN
WANG KAI
CHEN RUIXIANG
WANG YANBO
JIN WENJIE
GAO SHICHEN
ZHANG HAO
LI YANGYANG
LIU YONGSHENG
YU KAIXUAN
LI MENGYU
ZHANG HEPEI
ZHANG BIN
LI SHUAO
LI ZHANGUO
WANG CHUNLIANG
GAO PAN
BAO PENG

Assignees

盾构及掘进技术国家重点实验室
中铁隧道局集团有限公司

Dates

Publication Date: 20260512
Application Date: 20251219

Claims (5)

1. The TBM cutterhead rotating speed prediction method based on vibration characteristic driving is characterized by comprising the following steps of: S1, collecting vibration data and tunneling data of a TBM in a certain time period; s2, performing time alignment and pretreatment on the collected vibration and tunneling data; S3, performing characteristic engineering on the vibration parameters and performing correlation calculation; S4, determining input parameters of a model, and performing model training; S5, evaluating and comparing the model; In the step S1, taking a main bearing section metaphor as a conventional dial plate as a reference, the vibration parameters of the TBM mainly select 8 key parameters of main bearing 1 o 'clock-speed, main bearing 5 o' clock-speed, main bearing 7 o 'clock-speed, main bearing 11 o' clock-speed, main bearing 1 o 'clock-displacement, main bearing 5 o' clock-displacement, main bearing 7 o 'clock-displacement and main bearing 11 o' clock-displacement, wherein the tunneling parameters select 5 key parameters of cutter head rotating speed, propelling speed, penetration, cutter head torque and total propelling force; S2, carrying out data time alignment, data preprocessing, data complement and data standardization and normalization processing due to different acquisition frequencies of vibration parameters and tunneling parameters; s3, carrying out feature engineering on vibration parameters, namely counting features and interaction features, and carrying out correlation calculation among the features by using Pelson correlation coefficients; S4, selecting the first 15 key features as model inputs according to correlation calculation; s5, evaluating the performance of the model by adopting MSE, RMSE, MAE and R2.
2. The TBM cutterhead speed prediction method based on vibration feature driving as claimed in claim 1, wherein: in S2, the data complement and data normalization and normalization processes include: (1) Data cleaning is carried out on the collected tunneling and vibration data, missing data and zero values are removed, and abnormal data are modified; (2) Alignment of multi-source data, namely alignment of vibration and tunneling parameters is realized by adopting a sliding window algorithm; (3) Data complementation, namely supplementing missing vibration data and vibration data by using an average value of missing value adjacent points; (4) Data normalization and normalization.
3. The TBM cutterhead rotating speed prediction method based on vibration characteristic driving of claim 1 is characterized by comprising the steps of carrying out characteristic engineering on vibration parameters, carrying out correlation calculation, extracting time domain characteristic statistics (mean value, standard deviation) and time domain characteristic (difference and sliding average) of the speed/displacement of each vibration point, constructing interaction characteristics with definite engineering significance, including dynamic indexes such as vibration energy (speed multiplied by displacement), speed-shift ratio (speed/displacement) and the like, and calculating Pearson correlation calculation screening key characteristics; the energy index reflects the instantaneous vibration energy of each measuring point ; Transfer efficiency of vibration energy ; Timing characteristics First order difference: ; sliding window statistics (window size=3) ; Pearson correlation coefficient: 。
4. The TBM cutterhead speed prediction method based on vibration feature driving of claim 3, wherein S4 comprises determining model input parameters for model training, retaining the features of |ρ| >0.3 and p-value <0.05, finally selecting ' main bearing 7 o ' clock_speed ', ' main bearing 7 o ' clock_speed_rolling_mean ', ' v7_energy ', ' main bearing 7 o ' clock_displacement ', ' v11_ratio ', ' main bearing 7 o ' clock_displacement_rolling_mean ', ' main bearing 11 o ' clock_rolling_mean ', ' main bearing 11 o ' clock_speed_rolling_mean ', ' v7_ratio ', ' main bearing 1 o ' displacement ', ' main bearing 1 o ' clock_bearing 1 o ' displacement ', ' main bearing 1 o ' clock_rolling_mean as the model input.
5. The TBM cutterhead rotating speed prediction method based on vibration characteristic driving as claimed in claim 4 is characterized in that data after data cleaning and characteristic engineering are used as input data of a model, namely 15 characteristics are used as input data sets of the model, the data sets are divided according to a ratio of 8:2, model training is conducted through a random forest and XGBoost algorithm, and then model performance is evaluated through MSE, RMSE, MAE and R2.

Description

TBM cutter disc rotating speed prediction method based on vibration characteristic driving Technical Field The invention relates to the field of shield cutter head related test equipment, in particular to a TBM cutter head rotating speed prediction method based on vibration characteristic driving. Background TBM (Tunnel Boring Machine, full face tunnel boring machine) is a large-scale tunnel construction equipment that collects technologies such as machinery, electron, hydraulic pressure, sensing in an organic whole, can realize the continuous operation of processes such as excavation, support, slag tap, mainly is applicable to the tunnel excavation of hard rock and extremely hard rock. TBM construction has the characteristics of high excavation efficiency, good excavation effect, good safety performance, low economic cost and the like, and TBM construction is adopted as the primary choice. A plurality of researches show that the cutter head vibration has a strong correlation with the TBM rock breaking efficiency. The vibration of the cutterhead is the result of the interaction of the rock-machine and is jointly influenced by the structure of the heading machine, the rock mass condition of the face and the heading parameters. The TBM is faced with various high thrust, high torque and the like in the tunneling process, and various parts of the TBM can generate larger vibration, so that the efficiency and the safety of tunnel construction are affected. Due to the problems of short tunnel construction period, high TBM manufacturing cost, difficult replacement of TBM parts and the like, the tunneling of the TBM under a safe load is ensured to be necessary. At present, students at home and abroad do a great deal of research on TBM tunneling parameters, but the analysis of the relation between cutter head vibration and tunneling parameters is still less. The cutter head vibration is an important measurement index capable of reflecting the TBM tunneling state, so that the research on the relation between the cutter head vibration parameter and the tunneling parameter is an effective way for optimizing the TBM tunneling parameter. And carrying out relation analysis and research between vibration parameters and tunneling parameters by depending on vibration data and tunneling data of the plateau railway TBM tunnel engineering. In the prior art, research on the prediction of the rotating speed of the shield cutterhead is lacking. Disclosure of Invention The invention aims to solve the problems, and provides a TBM cutter rotating speed prediction method based on vibration characteristic driving, which is characterized in that TBM vibration and tunneling data are collected, the collected data are preprocessed, characteristic engineering is carried out on the vibration data, the first 15 key characteristics are selected as model input parameters through correlation calculation, data training is carried out by utilizing random forest and XGBoost models, the models are evaluated, compared and selected to obtain a model with better prediction effect, and cutter rotating speed setting recommendation in the TBM construction process is realized. The specific scheme of the invention is as follows: A TBM cutterhead rotating speed prediction method based on vibration characteristic driving is designed, and the method comprises the following steps: S1, collecting vibration data and tunneling data of a TBM in a certain time period; s2, performing time alignment and pretreatment on the collected vibration and tunneling data; S3, performing characteristic engineering on the vibration parameters and performing correlation calculation; S4, determining input parameters of a model, and performing model training; S5, evaluating and comparing the model; In the step S1, taking a main bearing section metaphor as a conventional dial as a vibration parameter of a reference TBM, mainly selecting 8 key parameters of main bearing 1 o 'clock-speed, main bearing 5 o' clock-speed, main bearing 7 o 'clock-speed, main bearing 11 o' clock-speed, main bearing 1 o 'clock-displacement, main bearing 5 o' clock-displacement, main bearing 7 o 'clock-displacement and main bearing 11 o' clock-displacement, wherein the tunneling parameters are 5 key parameters of cutter head rotating speed, propelling speed, penetration, cutter head torque and total propelling force; S2, carrying out data time alignment, data preprocessing, data complement and data standardization and normalization processing due to different acquisition frequencies of vibration parameters and tunneling parameters; s3, carrying out feature engineering on vibration parameters, namely counting features and interaction features, and carrying out correlation calculation among the features by using Pelson correlation coefficients; S4, selecting the first 15 key features as model inputs according to correlation calculation; S5 the model performance was evaluated using MSE, RMSE, MAE and R 2. In a sp