CN113916183B - A prediction system for PBA structural deformation risk and its use method - Google Patents

Abstract

The invention discloses a PBA structural deformation risk prediction system and its use method, which includes: a data acquisition unit, a data transmission unit, and a data processing unit; the data acquisition unit is used to obtain the deformation of tunnel vaults, spandrels, and waist Monitoring data of changes in earth pressure and steel stress at the vault and waist; data processing unit used to predict earth pressure and steel stress during construction. This invention can make up for the shortcomings of manual measurement with high risk, low frequency and large errors, thereby realizing real-time automated and accurate monitoring of tunnel soil pressure and steel bar stress, which actually provides guarantee for safe tunnel construction; at the same time, it takes into account the fixation of the collection box and the transmission box. , the instrument is waterproof and anti-collision protected, extending the service life of the collection box and transmission box; using a short-distance wireless transmission system and GPRS information transmission components to replace the use of long-distance signal data lines, effectively reducing monitoring costs and improving equipment efficiency applicability.

Description

A prediction system for PBA structural deformation risk and its use method

Technical field

The present invention relates to the technical field of tunnel data monitoring and parameter identification, and in particular to a prediction system for PBA structural deformation risk and its use method.

Background technique

The stability of the surrounding rock is crucial during the subway construction process. Currently, the stability of the surrounding rock of the tunnel is mainly determined by manually measuring the displacement of the surrounding rock. The manual measurement method is risky, low-frequency, and contains large human errors. Compared with manual measurement, automated monitoring has higher accuracy and more intensive monitoring frequency. However, the automated monitoring currently used in tunnel projects mostly uses one monitoring section as the monitoring target, which belongs to the category of two-dimensional monitoring. As a result, the monitoring data cannot reflect the project. The overall stable state of space has certain limitations.

The current detection equipment in the subway tunnel project is greatly affected by the environmental factors of the subway project, and there are mainly the following problems: (1) The measurement target is in a two-dimensional plane, and the distance between the current monitoring position and the tunnel face is ignored, making it impossible to determine The spatial effect brought about by the advancement of the tunnel face; (2) Monitoring equipment requires workers to measure on site and cannot be monitored automatically. (3) During the subway construction process, the environment was complex and harsh (such as gravel splashing and water seepage caused by blasting), and no corresponding instrument protection devices and corresponding fixing devices were designed. (4) Due to signal problems in the tunnel, a small number of automated collection facilities can only use data lines of considerable length to connect to the outside of the tunnel, which is costly and the data lines in the tunnel are easily damaged.

Contents of the invention

The present invention provides a prediction system and method for PBA structural deformation risk to overcome the above problems.

The system of the present invention includes: a data acquisition unit, a data transmission unit, and a data processing unit;

The data acquisition unit is used to obtain the monitoring data of changes in the earth pressure at the tunnel vault, spandrels and waists and the stress of the steel bars at the vaults and waists;

A data transmission unit, used to integrate the earth pressure and steel stress data collected by the data acquisition unit to form time series monitoring data, and send the time series monitoring data to the data processing unit;

The data processing unit is used to establish a nonlinear mapping relationship between monitoring data and tunnel soil pressure and steel stress through a fitting algorithm based on the least squares support vector machine, and uses a time series-based bacterial foraging algorithm to analyze the conditions during the construction process. Earth pressure and steel stress are predicted.

Further, the data collection unit includes: data monitoring part, data collection part, protection structure part, and support structure part;

The data monitoring part is used to monitor and obtain the change monitoring data of the earth pressure at the vault, spandrel and waist and the stress of the steel bars at the vault and waist;

The data acquisition unit is used to collect and send the data monitoring unit to the data transmission unit;

The protective structure part is used to accommodate each of the data collection parts;

The supporting structure part is used to fix the protective structure part to the tunnel side wall.

Further, the data monitoring part includes: earth pressure vault measurement module, earth pressure spandrel measurement module, earth pressure waist measurement module, steel stress vault measurement module, steel stress waist measurement module;

The earth pressure vault measurement module is used to obtain sampling data corresponding to the vault monitoring direction;

The earth pressure spandrel measurement module is used to obtain sampling data corresponding to the spandrel monitoring direction;

The earth pressure waist measurement module is used to obtain sampling data corresponding to the waist monitoring direction;

The steel stress vault measurement module is used to obtain sampling data corresponding to the vault monitoring direction;

The steel stress waist measurement module is used to obtain sampling data corresponding to the waist monitoring direction;

Further, the data collection part includes: automatic data collection module and automatic data transmission module;

The automatic data collection module is used for real-time monitoring of the data monitoring part.

The automatic data transmission module is connected to the automatic data collection module through a data line to send the collected data to the data transmission unit in real time.

Further, the data collection unit includes: a data receiving part, a data transmission part, a protection structure part, and a supporting structure part;

The data receiving part is used to receive data wirelessly transmitted by the data collection unit;

The data transmission part is used to send the data received by the data receiving part to the data processing unit;

The protection structure part contains each of the data collection parts;

The supporting structure part is capable of fixing the protective structure part to the tunnel side wall.

Further, the data receiving part includes: a data wireless receiver and a transmission box signal receiving antenna, and the data wireless receiver is connected to the transmission box signal receiving antenna through a data line;

The transmission box signal receiving antenna can realize short-distance wireless data transmission.

The method of the present invention includes:

S1. The data processing unit receives the monitoring data transmitted by the data transmission unit and establishes sample data based on time series. The sample data includes: vault earth pressure, spandrel earth pressure, arch waist earth pressure, and vault steel stress on the i-th day. , the value of the stress of the arch waist steel bar;

S2. The data processing unit uses the least squares support vector machine theory to learn the time series data and obtain the nonlinear relationship between the soil pressure value and the steel stress value change sequence;

S3. The data processing unit determines the mapping relationship corresponding to the training set formed by relying on time series monitoring data to obtain a nonlinear prediction model, and based on the nonlinear prediction model, uses the tropism operation, copy operation, and migration operation of the bacterial foraging algorithm. Optimize the nonlinear mapping model to achieve rolling prediction of structural deformation risk.

Further, S2 includes:

S21. Based on the determined distance data between the monitoring position and the tunnel face, obtain the earth pressure and steel stress data of the monitoring section of the tunnel face;

S22. Predict the nonlinear change sequence, that is, obtain the relationship between the earth pressure and steel stress at time p+1 and the earth pressure and steel stress at the previous p moments;

S23. Use the least squares support vector machine theory to train the relationship between earth pressure and steel bar stress changes through time series data, and obtain the nonlinear relationship between earth pressure and steel bar stress change sequences:

Among them, y(x _p+1 ) is the earth pressure and steel stress value at the p+1 moment; b is the offset; is the kernel function, K(x,x _k )=exp{||xx _i || ² /σ ² }, σ ² is the squared bandwidth pass in the Gaussian RBF kernel.

Further, S3 includes:

S31. Use the nonlinear relationship between the soil pressure value and the steel stress value change sequence to establish a nonlinear prediction model. The nonlinear prediction model is used to predict the data of the current position p+1 through the first P historical data;

S32. Add groundwater level influencing parameters to the input time series to correct each step of the prediction process:

S321. Trend operation: For a set of time series monitoring data, perform any possible increase or decrease in the data value within the set range to obtain a new set of time series monitoring data, and repeat the operation;

S322. Copy operation: When the trend operation reaches the set critical number of times, all time series monitoring data are arranged in descending order based on the absolute value of the distance between the predicted value and the monitoring value as the evaluation index, and the last 50% of the data is deleted. Copy the first 50% of the data.

S323. Migration operation: The migration operation occurs when the next data needs to be predicted. The execution reference point is after the copy operation reaches the set limit step. If the data meets the requirements, the data is retained; if the data does not meet the requirements, the data is deleted.

S33. Based on the nonlinear prediction model optimized by the bacterial foraging algorithm, make corresponding rolling predictions for the sampling data obtained by the data collection unit.

This invention can make up for the shortcomings of manual measurement with high risk, low frequency and large errors, thereby realizing real-time automated and accurate monitoring of tunnel soil pressure and steel bar stress, which actually provides guarantee for safe tunnel construction; at the same time, it takes into account the fixation of the collection box and the transmission box. , the instrument is waterproof and anti-collision protected, extending the service life of the collection box and transmission box; using a short-distance wireless transmission system and GPRS information transmission components to replace the use of long-distance signal data lines, effectively reducing monitoring costs and improving equipment efficiency applicability. The data transmission unit of the present invention can upload data to the cloud, enabling real-time acquisition by multiple terminals; again, the data processing unit can train a nonlinear prediction model based on the monitoring data in the cloud to achieve rolling prediction of structural deformation risks.

Description of the drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting any creative effort.

Figure 1 is an overall flow chart of the present invention;

Figure 2 is a schematic structural diagram of the earth pressure box according to the present invention;

Figure 3 is a schematic structural diagram of the steel bar meter according to the present invention;

Figure 4 is a schematic structural diagram of the collection box according to the present invention;

Figure 5 is a schematic structural diagram of the transmission box according to the present invention;

Figure 6 is a schematic structural diagram of the horizontal support frame according to the present invention;

Figure 7 is a schematic diagram of monitoring section settings in one embodiment;

Figure 8 is a schematic diagram of monitoring section monitoring points in one embodiment;

Figure 9 is an overall schematic diagram of the layout of monitoring section equipment in one embodiment;

Figure 10 is an implementation flow chart of the PBA structural deformation risk prediction method based on the intelligent response surface method according to the present invention;

Figure 11 is a flow chart for realizing the work of the collection box according to the present invention;

Figure 12 is a flow chart for realizing the work of the transmission box according to the present invention;

Figure 13 is a stratigraphic map in one embodiment;

Figure 14 is a line chart of earth pressure time series data of the data processing unit in one embodiment.

Explanation of reference numbers:

1. Earth pressure box; 2. Earth pressure box bracket; 3. Earth pressure box signal output line; 4. Steel bars; 5. Steel bar meter; 6. Steel bar meter signal output line; 7. Collection box signal transmitting antenna; 8. Collection box; 9. Data wireless transmitter; 10. Bracket expansion screw; 11. Bracket; 12. Power cord; 13. Earth pressure box signal input line; 14. Reinforcement meter signal input line; 15. Collection box power supply; 16. Automation Collector; 17. Waterproof outside of the collection box; 18. Signal receiving antenna of the transmission box; 19. Transmission box; 20. Waterproof outside of the transmission box; 21. GPRS signal transmitter; 22. GPRS signal transmitting antenna; 23. Power cord of the transmission box ; 24. Transmission box power supply; 25. Data wireless receiver; 26. 40m monitoring section; 27. 75m monitoring section; 28. 110m monitoring section; 29. Vault earth pressure monitoring point; 30. Spandrel earth pressure monitoring point; 31. The soil pressure monitoring point of the arch waist; 32. The stress monitoring point of the steel bars on the top of the vault; 33. The stress monitoring point of the steel bars of the arch waist.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

Example 1

In view of the current problems in the monitoring process of tunnel earth pressure and steel stress, manual measurement has high risks, low frequency, large errors, and the inability to provide advance warning of the risk of surrounding rock deformation. In this embodiment, a method based on the intelligent response surface method was developed. The PBA structural deformation risk prediction method has the following characteristics: (1) The collection box and launch box can be installed on the primary lining wall of the tunnel through brackets. The collection box and launch box are closed structures and have the functions of waterproofing and releasing concrete shotcrete and flying rocks. , able to adapt to complex and harsh engineering environments; (2) Monitoring data on changes in soil pressure at the tunnel vault, spandrels and waist and steel stress at the vault and waist can be obtained at the same time, which can make up for the tunnel excavation space Errors caused by the effect; (3) The collection box and transmission box formed by the information collection unit and the data transmission unit use wireless data transmission instead of traditional wired transmission, improving the practicality of the device; (4) The data processing unit can use vector Machines and bacterial foraging algorithms are used to build prediction models, and real-time predictions are made based on monitored soil pressure and steel stress data, providing engineers with a more intuitive data reference. At the same time, the device is easy to install and disassemble, can be recycled, and has great cost advantages.

Based on the above design points, it can be seen that the PBA structural deformation risk system based on the intelligent response surface method described in this embodiment includes:

(1) The data acquisition unit can simultaneously obtain the monitoring data of the soil pressure at the tunnel vault, spandrel and waist and the change monitoring data of the steel stress at the vault and waist, to more comprehensively reflect the conditions of the tunnel during the tunnel excavation process. The change process of earth pressure and steel stress under the influence of spatial effect; at the same time, the data acquisition unit has dealt with tunnel shotcrete construction, blasting construction and water-rich environment, and the protective structure can realize the functions of waterproofing and flying stone splash prevention ; The data acquisition unit must also have an independent support and fixed structure. For the longitudinal extension characteristics of the tunnel project in space, a support structure that can provide stable support and is easy to disassemble. After the construction of the current area is completed, the device can be disassembled and moved. And used for the next target position, which improves the utilization efficiency of the structure of the present invention;

(2) Data transmission unit, which can collect and remotely transmit the earth pressure and steel stress monitoring information output by the data acquisition unit through the earth pressure box and steel bar meter. The data transmission unit sends data to the data processing unit through the GPRS module. The early warning information is synchronously released through multiple terminals such as web pages, mobile phones, and software platforms through the server to meet the data recognition needs of different construction participating departments and provide project managers with more intuitive and comprehensive surrounding rock status data;

(3) Data processing unit, which can receive data transmitted by the data transmission unit; establish a nonlinear mapping relationship between monitoring data and tunnel earth pressure and steel stress through a fitting algorithm based on the least squares support vector machine , a bacterial foraging algorithm based on time series is used to predict the earth pressure and steel stress during the construction process. If the predicted value exceeds the warning value three times in a day, a monitoring and early warning will be issued; the early warning information is realized through the server on web pages, mobile phones, software platforms, etc. The simultaneous release of data from multiple terminals meets the data recognition needs of different construction participating departments and provides project managers with more intuitive and comprehensive soil pressure and steel stress state data.

The PBA structural deformation risk prediction method based on the intelligent response surface method, composed of the above three units, can realize the release of soil pressure and steel stress monitoring data and risk warning information during the forward construction process of the tunnel face. The combined working form and specific data flow are shown in Figure 10, Figure 11, and Figure 12.

As shown in Figures 1 to 7, 13, and 14, based on the above solution, the present invention first develops a PBA structural deformation risk prediction system based on the intelligent response surface method to apply it to the tunnel section vault and arch of this device. The monitoring process of the earth pressure at the shoulder and waist and the steel stress at the vault and waist. However, in the traditional structural deformation risk prediction and early warning process, no attention is paid to the distance between the monitoring section and the tunnel face. However, changes in the distance between the monitoring section and the tunnel face will affect the data of earth pressure and steel stress, which represents the spatial effect of tunnel excavation. With the excavation of the tunnel, the original support for the upper stratum will be affected. The rock is excavated, destroying the supporting balance of the surrounding rock. The earth pressure and the stress of the initial lining steel bars will establish a new balance. This balance will bring about the displacement of the surrounding rock; as the excavation section advances, the old balance will continue to be broken. , new balances are constantly established, and displacements continue to produce new displacements. In the data processing unit, vector machines and bacterial foraging algorithms are used to establish earth pressure and steel stress prediction models to achieve rolling predictions of the project. The prediction and warning data are uploaded to the cloud, enabling real-time acquisition of data from multiple terminals.

The data collection unit mainly consists of four parts: 1. Data monitoring part; 2. Data collection part; 3. Protection structure part; 4. Support structure part. The data monitoring part 1 can monitor the earth pressure at the vault, spandrel and waist of the tunnel section and the steel stress at the vault and waist, and collect the corresponding tunnel spatial characteristic information, preferably using an earth pressure box and a steel bar meter; The data collection part 2 is connected to the monitoring instrument through a data line, collects corresponding monitoring data, and transmits the monitoring data wirelessly through the antenna.

The protective structure part 3 contains the data monitoring part 1 and the data acquisition part 2. The protective structure part 2 can realize the protection of the instrument. The main protection targets are the water environment in the tunnel, the high dust particle environment, and the collision of engineering equipment. etc.; the support structure part 4 can fix the data monitoring part 1, the data acquisition part 2 and the protection structure part 3 to the side wall of the tunnel. The working flow chart of the collection box is shown in Figure 12.

In a specific embodiment, as shown in Figure 2 and Figure 3, the data monitoring part includes: earth pressure box 1, pressure box bracket 2, earth pressure box signal output line 3, steel bar 4, steel bar meter 5, steel bar meter signal output Line 6, vault earth pressure monitoring point 29, spandrel earth pressure monitoring point 30, arch waist earth pressure monitoring point 31, vault steel bar stress monitoring point 32, arch waist steel stress monitoring point 33; wherein, the earth pressure box 1. The pressure box bracket 2 welded to the primary lining steel bars is fixed on the vault earth pressure monitoring point 29, the spandrel earth pressure monitoring point 30, and the arch waist earth pressure monitoring point 31 to obtain the corresponding vault, spandrel, and Earth pressure monitoring data at the arch waist; the steel bar meter 5 is used to obtain the corresponding vault, Monitoring data of steel stress at the waist of the arch;

In a more specific embodiment, as shown in Figure 8, the earth pressure box 1 and the steel bar gauge 5 are fixed at the vault earth pressure monitoring point 29, the spandrel earth pressure monitoring point 30, the arch waist earth pressure monitoring point 31, and the vault steel bars. After the stress monitoring point 32 and the arch waist steel stress monitoring point 33, the positions of the earth pressure box signal output line 3 and the steel bar meter signal output line 6 need to be adjusted to ensure that the signal output lines are not damaged on site, and to ensure that there is sufficient length. The tunnel is lined outside.

In a more specific embodiment, this case can limit the specific types and attributes of the monitoring instruments, that is, the earth pressure box 1 and the steel bar gauge 5 adopt vibrating wire instruments to facilitate the collection of data acquisition equipment.

In a specific embodiment, as shown in Figure 4, the data collection part includes: collection box signal transmitting antenna 7, collection box 8, data wireless transmitter 9, power cord 12, earth pressure box signal input line 13, steel bar gauge signal Input line 14, collection box power supply (220v to 24v) 15, automated collector 16; wherein, the data wireless transmitter 9, power cord 12, collection box power supply (220v to 24v) 15, and automated collector 16 are all placed in Inside the collection box 8, the collection box power supply (220v to 24v) 15 is responsible for powering the data wireless transmitter 9 and the automated collector 16; the automated collector 16 is connected through the earth pressure box signal input line 13 and the steel bar meter signal input line 14. Thus, the earth pressure at the vault, spandrel and waist and the steel stress at the vault and waist are collected; the data wireless transmitter 9 is connected to the automatic collector 16 through the data line, and the data of the automatic collector 16 is passed through the collection box The external collection box signal transmitting antenna 7 realizes wireless transmission.

In a more specific embodiment, the earth pressure box signal input line 13 and the steel bar meter signal input line 14 are respectively connected to the earth pressure box signal output line 3 and the steel bar meter signal output line 6. Only one pair of signal lines can be used. 1 connection.

In a specific embodiment, as shown in Figure 4, the protective structure part includes a box structure composed of a collection box 8 and a waterproof outer edge 17 of the collection box; the entire collection box 8 is welded with 3mm thick 304 stainless steel, and the cover One side of the plate is fixed on the collection box 8 through a rotating shaft, so that the cover plate can rotate around the rotating shaft; there is a header waterproof outer edge 17 on the top of the collecting box 8 to prevent dripping water in the hole. The 3mm thick 304 stainless steel has a certain resistance to impact and prevents flying stones generated by tunnel spraying from damaging the instrument.

In a more specific embodiment, as shown in Figure 4 and Figure 6, the support structure part includes a water support expansion screw 10, wherein the support expansion screw 10 is fixed in the lining of the tunnel, and the support 11 is passed through the support expansion screw. 10. Fix it on the side wall of the tunnel. Make sure that the collection box is at a certain height from the bottom of the tunnel.

The earth pressure and steel stress data measured by the data acquisition unit are wirelessly transmitted to the data transmission unit through the signal transmitting antenna of the acquisition box.

The data transmission unit mainly consists of four parts: 1. Data receiving part; 2. Data transmission part; 3. Protection structure part; 4. Support structure part. The data receiving part 1 can receive the data wirelessly transmitted by the collection box, and transmit the data to the data transmission part through the data line; the data transmission part 2 mainly consists of a GPRS signal transmitter and transmits the monitoring data wirelessly through the antenna. transmitted to the remote data processing unit. The protective structure part 3 contains the data receiving part 1 and the data transmitting part 2. The protective structure part 3 can realize the protection of the instrument. The main protection targets are the water environment, high dust particle environment and engineering equipment in the tunnel. Collision, etc.; the support structure part 4 can fix the data receiving part 1, the data transmission part 2 and the protective structure part 3 to the side wall of the tunnel. The transmission workflow diagram is shown in Figure 12.

In a specific embodiment, as shown in Figure 5, the data receiving part includes: transmission box signal receiving antenna 18, transmission box 19, transmission box power cord 23, transmission box power supply (220v to 24v) 24, and data wireless receiver 25 ; Among them, the transmission box power supply (220v to 24v) 24 and the data wireless receiver 25 are placed inside the transmission box 19; the transmission box signal receiving antenna 18 receives the signal wirelessly transmitted by the collection box signal transmitting antenna 7 Earth pressure and steel stress monitoring data, and transmit the monitoring data to the data wireless receiver 25; the data wireless receiver 25 is powered by the transmission box power supply (220v to 24v) 24;

In a specific embodiment, as shown in Figure 5, the data transmission part includes: transmission box 19, GPRS signal transmitter 21, GPRS signal transmitting antenna 22, transmission box power cord 23, transmission box power supply (220v to 24v) 24 ; Among them, the GPRS signal transmitter 21 and the transmission box power supply (220v to 24v) 24 are placed inside the transmission box 19, and the transmission box power supply (220v to 24v) 24 is responsible for powering the GPRS signal transmitter 21; GPRS signal transmitter 21 receives the signal of the data wireless receiver 25 through the data line, and wirelessly transmits the received monitoring data to the remote data processing unit through the GPRS signal transmitting antenna 22.

In a specific embodiment, as shown in the figure, the protective structure part includes a box structure composed of a transmission box 19 and a waterproof outer edge 20 of the transmission box; the entire transmission box 19 is welded with 3mm thick 304 stainless steel, and the cover plate One side is fixed on the transmission box 19 through a rotating shaft, so that the cover plate can rotate around the rotating shaft; there is a transmission box waterproof outer edge 20 on the top of the transmission box 19 to prevent dripping water in the hole. The 3mm thick 304 stainless steel has a certain resistance to impact and prevents the flying stone sputtering caused by tunnel spraying from damaging the instrument.

In a more specific embodiment, as shown in Figures 5 and 6, the support structure part includes: water support expansion screws 10, wherein the support expansion screws 10 are fixed in the lining of the tunnel, and the support 11 is expanded by the support. The screws 10 are fixed on the side wall of the tunnel, and it is necessary to ensure that the transmission box 19 is at a certain height from the bottom of the tunnel.

In specific embodiments, the method of the present invention includes:

S1. Establish sample data { _xi }={x ₁ , x ₂ ,..., x _n } based on time series. The sample data includes vault earth pressure, spandrel earth pressure, arch waist earth pressure, and vault steel bars. Stress, arch waist steel stress;

S2. According to the least square support vector machine theory, the nonlinear deformation relationship can be obtained by learning the obtained measured earth pressure value and steel bar stress value through the support vector machine to obtain the nonlinear relationship between the soil pressure value and the steel bar stress value change sequence. ;

S3. Determine the mapping relationship corresponding to the training set to obtain a nonlinear mapping model, and based on the nonlinear mapping model, optimize the nonlinear mapping model through the tendency operation, copy operation, and migration operation of the bacterial foraging algorithm. Realize rolling prediction of structural deformation risks.

Optionally, in one embodiment, the S2 includes:

S21. Based on the determined distance data between the monitoring position and the tunnel face, obtain the earth pressure and steel stress data of the monitoring section corresponding to each step forward of the tunnel face;

S22. To predict this nonlinear change sequence, we need to find the relationship between the earth pressure and steel stress at time p+1 and the earth pressure and steel stress at the previous p moments x ₁ , x ₂ ,..., x _p , that is, x _p+1 =f(x ₁ ,x ₂ ,…,x _p ) is a learning function, which represents the nonlinear relationship between earth pressure and steel stress change sequence.

S23. Use the least squares support vector machine theory to learn time series data: for np deformed sequences x _i , x _i+1 ,..., x _i+p , (i=1, 2,..., np) Learn and obtain the nonlinear relationship between earth pressure and steel stress change sequence,

In the formula: y(x _p+1 ) is the earth pressure and steel stress value at time p+1 x _p+1 is the first p earth pressure and steel stress value at time p+1, x _p+1 =f( x ₁ ,x ₂ ,…,x _p+1 ) are the first p earth pressure and steel stress values at the position at time p+k, x _k = (x _k ,x _k+1 ,…,x _{k+p+ 1} ).

Optionally, in one embodiment, the S3 includes:

S31. Under the above conditions, the mapping form of the learning sample is set to {x ₁ ,x ₂ ,…,x _p }→{x _p+1 }, {x ₂ ,x ₃ ,…,x _p+1 }→ {x _p+2 }…{x _np ,x _n-p+1 ,…,x _n-1 }→{x _n }. The result of vector machine learning is that the data of the current position can be predicted through the P historical data before the prediction point. For example, if you need to predict x _n+1 , you only need to input, {x _n-p+1 ,x _n-p+2 ,…, x _n } to get the prediction result; then use the predicted x _n+1 as a known quantity, and {x _n-p+2 , x _n-p+3 ,...,x _n+1 } as a new Time series predicts x _n+2 .

S32. However, there are still some inevitable errors in each prediction step in the prediction process. As the number of prediction steps increases, such errors continue to accumulate, which may eventually cause the prediction results to fail to accurately express the real working conditions. In order to reduce the impact of this error, Q influence parameters that can be accurately determined are added to the input time series to effectively correct each step of the prediction process. The bacterial foraging optimization algorithm BFOA is used to find the optimal parameters through trend operations, copy operations, and migration operations, and optimally solve the least squares support vector machine nonlinear prediction algorithm for the best historical step number p and influencing factor Q. High dependency problem.

S33. Based on the optimized nonlinear prediction model, perform corresponding rolling predictions on the sampled data obtained by the data acquisition unit.

Example 2

The first step is to install the monitoring device and obtain measurement data:

First, during tunnel construction, install the earth pressure box and steel bar meter to the corresponding measuring point (as shown in Figure 8) and protect the corresponding earth pressure box signal output line and steel bar meter signal output line; at the set monitoring section (as shown in Figure 8) 7) Install the collection box bracket as shown in Figure 6 on the side wall of the tunnel at an appropriate location nearby. Install the transmission box bracket at the tunnel entrance. Use expansion bolts to fix the bracket to the side wall, and place the collection box and transmission box at After the bracket is installed, pass the steel wire through the hole at the bottom of the box to ensure that the bracket is firmly installed; secondly, connect the earth pressure box signal input line and the steel bar meter signal input line of the collection box to the earth pressure box signal output line exposed outside the lining and the Rebar meter signal output line, debug the instrument to ensure that the automated collector can collect data; again, adjust the positions of the collection box signal transmitting antenna, transmission box signal receiving antenna, and GPRS signal transmitting antenna to ensure that the transmission box can receive the signal from the collection box , the remote computer can receive the signal from the transmission box, and finally the connectivity of the data acquisition unit, data transmission unit, and data processing unit.

The second step is monitoring data transmission and data integration:

First, collect the vibrating wire values of the earth pressure box and steel bar gauge, and use the formula Calculate the soil pressure and steel stress; secondly, the data wireless transmitter sends the soil pressure and steel stress measured by the automatic collector to the transmission box through the collection box signal transmitting antenna; secondly, the data wireless receiver of the transmission box receives the signal through the transmission box The antenna receives the monitoring data, and then uses a GPRS signal transmitter to transmit the data to the data processing unit through the GPRS signal transmitting antenna. The data processing unit first processes the data into event sequence monitoring data.

The third step is to calculate the time series monitoring data based on the vector machine and bacterial foraging algorithm, obtain an optimized prediction model, and conduct rolling predictions of earth pressure and steel stress: the creation process of the rolling prediction includes:

S1. Establish sample data { _xi }={x ₁ , x ₂ ,..., x _n } based on time series. The sample data includes vault earth pressure, spandrel earth pressure, arch waist earth pressure, and vault steel bars. Stress, arch waist steel stress;

S2. According to the least square support vector machine theory, the nonlinear deformation relationship can be obtained by learning the obtained measured earth pressure value and steel bar stress value through the support vector machine to obtain the nonlinear relationship between the soil pressure value and the steel bar stress value change sequence. ;The S2 includes:

S21. Based on the determined distance data between the monitoring position and the tunnel face, obtain the earth pressure and steel stress data of the monitoring section corresponding to each step forward of the tunnel face;

S22. To predict this nonlinear change sequence, we need to find the relationship between the earth pressure and steel stress at time p+1 and the earth pressure and steel stress at the previous p moments x ₁ , x ₂ ,..., x _p , that is, x _p+1 =f(x ₁ ,x ₂ ,…,x _p ) is a learning function, which represents the nonlinear relationship between earth pressure and steel stress change sequence.

S23. Using the least squares support vector machine theory, for the given N training samples {x _i , y _i } _i=1...N (where x _i ∈R ⁿ is the n-dimensional training input sample y _i ∈R ⁿ is the training output sample), and the objective optimization function is Optimize the target parameters and learn the time series data: learn np deformation sequences x _i , x _i+1 ,..., x _i+p , (i=1, 2,..., np), and get the earth pressure and Nonlinear relationship between steel stress change sequences (LSSVM regression function): /> In the formula: K(x,x _k )=exp{||xx _i || ² /σ ² } (the kernel function adopts the radial basis kernel function).

In the formula: y(x _p+1 ) is the earth pressure and steel stress value at time p+1 x _p+1 is the first p earth pressure and steel stress value at time p+1, x _p+1 =f ( x ₁ ,x ₂ ,…,x _p+1 ) are the first p earth pressure and steel stress values at the position at time p+k, x _k = (x _k ,x _k+1 ,…,x _{k+p+ 1} ).

S3. Determine the mapping relationship corresponding to the training set to obtain a nonlinear mapping model, and based on the nonlinear mapping model, optimize the nonlinear mapping model through the tendency operation, copy operation, and migration operation of the bacterial foraging algorithm. Realize rolling prediction of structural deformation risks.

S31. Under the above conditions, the mapping form of the learning sample is set to {x ₁ ,x ₂ ,…,x _p }→{x _p+1 }, {x ₂ ,x ₃ ,…,x _p+1 }→ {x _p+2 }…{x _np ,x _n-p+1 ,…,x _n-1 }→{x _n }. The result of vector machine learning is that the data of the current position can be predicted through the P historical data before the prediction point. For example, if you need to predict x _n+1 , you only need to input, {x _n-p+1 ,x _n-p+2 ,…, x _n } to get the prediction result; then use the predicted x _n+1 as a known quantity, and {x _n-p+2 , x _n-p+3 ,...,x _n+1 } as a new Time series predicts x _n+2 .

S32. However, there are still some inevitable errors in each prediction step in the prediction process. As the number of prediction steps increases, such errors continue to accumulate, which may eventually cause the prediction results to fail to accurately express the real working conditions. In order to reduce the impact of this error, Q influence parameters that can be accurately determined are added to the input time series to effectively correct each step of the prediction process. The bacterial foraging optimization algorithm BFOA is used to find the optimal parameters through trend operations, copy operations, and migration operations, and optimally solve the least squares support vector machine nonlinear prediction algorithm for the best historical step number p and influencing factor Q. High dependency problem.

S321. Tendency operation: For a community with a population size of s, use θ(i,g,n,m) to represent the position of individual i after g chemotaxis operations, n replication operations, and m migration operations. Information, C(i) represents the step size, Δ represents any unit random vector on [-1,1], represents the direction after random adjustment, then the position formula of the chemotactic cycle can be expressed as:/>

S322. Copy operation: The copy operation follows the law of natural selection of survival of the fittest. When the trend operation reaches the critical number, all data will be arranged from large to small using the fitness value as the evaluation index. The total number of data will be recorded as 2S _r ; the data will be ranked in the second half. Partial S _r data with smaller fitness values are eliminated, and the first S _r data with larger fitness values are retained; the retained excellent individuals are copied to obtain a data with exactly the same foraging ability. Complete a copy operation.

S323. Migration operation: The migration operation occurs when the resource environment changes, and the execution reference point is after N _re steps of the copy operation. The migration operation means two different results: the entire data is moved to another area or this set of data is cancelled.

S33. Based on the optimized nonlinear prediction model, perform corresponding rolling predictions on the sampled data obtained by the data acquisition unit.

The fourth step is to release prediction and early warning results: officially use the optimized prediction and early warning model pair obtained above, install it in the data collection and processing module, integrate and calculate the real-time collected data, and obtain the prediction and early warning results and send them to the cloud to realize multi-terminal pair calculation. Real-time access to query results.

To implement the embodiments of the present invention, firstly, a PBA structural deformation risk prediction method based on the intelligent response surface method was developed based on the automated collection of the Internet of Things to compensate for the shortcomings of manual measurement with high risk, low frequency, and large errors, and then Real-time automated and accurate monitoring of tunnel soil pressure and steel stress ensures safe tunnel construction; at the same time, the fixation of the collection box and transmission box, instrument waterproofing, and anti-collision protection are also considered... to improve the collection box and transmission The service life of the box is extended; the use of short-distance wireless transmission systems and GPRS information transmission components replaces the use of long-distance signal data lines, effectively reducing monitoring costs and improving the applicability of the equipment. The data transmission unit of the present invention can upload data to the cloud, enabling real-time acquisition by multiple terminals; again, the data processing unit can train a nonlinear prediction model based on the monitoring data in the cloud to achieve rolling prediction of structural deformation risks. It is suitable for the dynamic design process of tunnel construction to improve the intelligent process of tunnel engineering.

Example 3

The data acquisition unit can simultaneously obtain the monitoring data of the soil pressure at the tunnel vault, spandrel and waist and the change monitoring data of the steel stress at the vault and waist;

The data transmission unit can integrate the earth pressure and steel stress data collected by the data acquisition unit to form time series monitoring data and send it to the data processing unit;

The data processing unit uses a fitting algorithm based on the least squares support vector machine to establish a nonlinear mapping relationship between monitoring data and tunnel soil pressure and steel stress, and uses a time series-based bacterial foraging algorithm to calculate the soil pressure during the construction process. Predict the stress of steel bars. If the predicted value exceeds the warning value three times in a day, a monitoring warning will be issued.

Optionally, in one embodiment, the data collection unit includes:

A measurement structure, which is used to monitor and obtain displacement change monitoring data in three directions: vault settlement, tunnel circumferential convergence, and distance between the monitoring position and the tunnel face;

A protective structure, each of the measurement structures is accommodated in the protective structure;

and a support structure capable of fixing the protective structure to the tunnel side wall.

Optionally, in one embodiment, the data collection unit includes:

Data monitoring part, the data monitoring part is used to monitor and obtain the change monitoring data of the earth pressure at the vault, spandrel and waist and the stress of the steel bars at the vault and waist;

Data collection part, the data collection unit is used to collect and send the data monitoring unit to the data transmission unit;

A protective structure part containing each of the data collection parts;

A support structure part capable of fixing the protective structure part to the tunnel side wall.

Optionally, in one embodiment, the data monitoring part includes: earth pressure vault measurement module, earth pressure spandrel measurement module, earth pressure spandrel measurement module, steel stress vault measurement module, steel stress spandrel measurement module Measuring module; wherein, the earth pressure vault measurement module is fixed on the vertical support frame through a bracket to obtain sampling data corresponding to the vault monitoring direction; the earth pressure spandrel measurement module is fixed on the vertical support frame through a bracket. The vertical support frame is used to obtain sampling data corresponding to the spandrel monitoring direction; the earth pressure apex measurement module is fixed on the vertical support frame through a bracket to obtain sampling data corresponding to the spandrel monitoring direction. ; The steel stress vault measurement module is fixed on the vertical support frame by welding to obtain sampling data corresponding to the vault monitoring direction; the steel stress vault measurement module is fixed on the vertical support by welding On the rack, it is used to obtain sampling data corresponding to the waist monitoring direction;

Optionally, in one embodiment, the data collection part

It includes: an automatic data acquisition module and an automatic data transmission module; wherein, the earth pressure box signal input line of the automatic data acquisition module is connected to the earth pressure box signal output line, and is connected to the steel bar meter signal output line through the steel bar meter signal input line. Real-time monitoring of the data monitoring part. The automatic data transmission module is connected to the automatic data collection module through a data line to realize real-time transmission of collected data to the data transmission unit.

Optionally, in one embodiment, the automatic data transmission module includes an automated collector and a collection box signal transmitting antenna. The automated collector is connected to the collection box signal transmitting antenna through a data line; the collection box signal transmitting antenna can realize short-term transmission. Wireless data transmission over long distances.

Optionally, in one embodiment, the protective structure part includes a box structure composed of a top plate, a side plate and a bottom plate, and a waterproof outer edge of the collection box; wherein the front side plate is fixed to the side side plate through a rotating shaft , so that the front side plate can rotate around the rotating axis, and then open the box structure for installation and adjustment; and the waterproof outer edge of the collection box is located on the outer edge of the top plate, which can prevent water dripping from the hole from penetrating into the collection box.

Optionally, in one embodiment, the support structure part includes bracket bolts, wherein the bracket is fixed on the tunnel side wall through bracket bolts, and there are two brackets under one collection box.

Optionally, in one embodiment, the data collection unit includes:

Data receiving part, the data receiving part is used to receive data wirelessly transmitted by the data acquisition unit described in claim 2;

Data transmission part, the data transmission part is used to send the data received by the data receiving part to the data processing unit;

A protective structure part containing each of the data collection parts;

A support structure part capable of fixing the protective structure part to the tunnel side wall.

Optionally, in one embodiment, the data receiving part includes a data wireless receiver and a transmission box signal receiving antenna. The data wireless receiver is connected to the transmission box signal receiving antenna through a data line; the transmission box signal receiving antenna can realize short Wireless data transmission over long distances.

Optionally, in one embodiment, the data transmission part includes a GPRS signal transmitter and a GPRS signal transmitting antenna. The GPRS signal transmitter is connected to the data wireless receiver through a data line for receiving data; GPRS signal transmission The device is connected to the GPRS signal transmitting antenna through a data line, and the data is transmitted to the data processing unit through the GPRS signal transmitting antenna.

Optionally, in one embodiment, the structural deformation risk prediction and warning based on vector machines and bacterial foraging algorithms corresponding to the data processing unit includes:

S1. Establish sample data { _xi }={x ₁ , x ₂ ,..., x _n } based on time series. The sample data includes vault earth pressure, spandrel earth pressure, arch waist earth pressure, and vault steel bars. Stress, arch waist steel stress;

S2. According to the least square support vector machine theory, the nonlinear deformation relationship can be obtained by learning the obtained measured earth pressure value and steel bar stress value through the support vector machine to obtain the nonlinear relationship between the soil pressure value and the steel bar stress value change sequence. ;

S3. Determine the mapping relationship corresponding to the training set to obtain a nonlinear mapping model, and based on the nonlinear mapping model, optimize the nonlinear mapping model through the tendency operation, copy operation, and migration operation of the bacterial foraging algorithm. Realize rolling prediction of structural deformation risks.

Optionally, in one embodiment, the S2 includes:

S21. Based on the determined distance data between the monitoring position and the tunnel face, obtain the earth pressure and steel stress data of the monitoring section corresponding to each step forward of the tunnel face;

S22. To predict this nonlinear change sequence, we need to find the relationship between the earth pressure and steel stress at time p+1 and the earth pressure and steel stress at the previous p moments x ₁ , x ₂ ,..., x _p , that is, x _p+1 =f(x ₁ ,x ₂ ,…,x _p ) is a learning function, which represents the nonlinear relationship between earth pressure and steel stress change sequence.

S23. Use the least squares support vector machine theory to learn time series data: for np deformed sequences x _i , x _i+1 ,..., x _i+p , (i=1, 2,..., np) Learn and obtain the nonlinear relationship between earth pressure and steel stress change sequence,

In the formula: y(x _p+1 ) is the earth pressure and steel stress value at time p+1 x _p+1 is the first p earth pressure and steel stress value at time p+1, x _p+1 =f ( x ₁ ,x ₂ ,…,x _p+1 ) are the first p earth pressure and steel stress values at the position at time p+k, x _k = (x _k ,x _k+1 ,…,x _{k+p+ 1} ).

Optionally, in one embodiment, the S3 includes:

S31. Under the above conditions, the mapping form of the learning sample is set to {x ₁ ,x ₂ ,…,x _p }→{x _p+1 }, {x ₂ ,x ₃ ,…,x _p+1 }→ {x _p+2 }…{x _np ,x _n-p+1 ,…,x _n-1 }→{x _n }. The result of vector machine learning is that the data of the current position can be predicted through the P historical data before the prediction point. For example, if you need to predict x _n+1 , you only need to input, {x _n-p+1 ,x _n-p+2 ,…, x _n } to get the prediction result; then use the predicted x _n+1 as a known quantity, and {x _n-p+2 , x _n-p+3 ,...,x _n+1 } as a new Time series predicts x _n+2 .

S32. However, there are still some inevitable errors in each prediction step in the prediction process. As the number of prediction steps increases, such errors continue to accumulate, which may eventually cause the prediction results to fail to accurately express the real working conditions. In order to reduce the impact of this error, Q influence parameters that can be accurately determined are added to the input time series to effectively correct each step of the prediction process. The bacterial foraging optimization algorithm BFOA is used to find the optimal parameters through trend operations, copy operations, and migration operations, and optimally solve the least squares support vector machine nonlinear prediction algorithm for the best historical step number p and influencing factor Q. High dependency problem.

S33. Based on the optimized nonlinear prediction model, perform corresponding rolling predictions on the sampled data obtained by the data acquisition unit.

Beneficial effects:

1. This invention makes up for the shortcomings of manual measurement such as high risk, low frequency and large errors, thereby realizing real-time automated and accurate monitoring of tunnel soil pressure and steel bar stress, which actually provides guarantee for safe tunnel construction; while taking into account the collection box and transmission box The fixed, instrument waterproof and anti-collision protection improves the service life of the collection box and transmission box; the use of short-distance wireless transmission system and GPRS information transmission components replaces the use of long-distance signal data lines, effectively reducing monitoring costs and improving the suitability of the equipment. The data transmission unit of the present invention can upload data to the cloud, enabling real-time acquisition by multiple terminals; again, the data processing unit can train a nonlinear prediction model based on the monitoring data in the cloud to achieve rolling prediction of structural deformation risks.

2. Constructed an IoT monitoring system suitable for tunnel-pile method (PBA) underground excavation stations, realizing real-time monitoring of soil pressure and steel stress and automatic uploading of monitoring data, making up for the high risk, low frequency and error of manual measurement. Big drawback.

3. The use of short-distance wireless transmission systems and GPRS information transmission components replaces the use of long-distance signal data lines, effectively reducing monitoring costs and improving the applicability of the equipment.

4. The detection box and transmission box are designed with corresponding waterproof and dustproof protective shells and placement brackets, which improves the practicability of the equipment in complex limited space construction environments.

5. The remote data processing unit uses a vector machine optimized by the bacterial foraging algorithm to form a nonlinear prediction model based on the cloud monitoring data. It can predict earth pressure and steel stress in advance, and perform structural deformation on data that exceeds the warning value. Risk Warning.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, but not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features can be equivalently replaced; and these modifications or substitutions do not deviate from the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention. scope.

Claims (6)

1. A system for predicting risk of deformation of a PBA structure, comprising: the device comprises a data acquisition unit, a data transmission unit and a data processing unit;

the data acquisition unit is used for acquiring the soil pressure at the arch crown, the arch shoulder and the arch waist of the tunnel and the change monitoring data of the steel bar stress at the arch crown and the arch waist;

the data transmission unit is used for integrating the soil pressure and the steel bar stress data acquired by the data acquisition unit to form time sequence monitoring data and transmitting the time sequence monitoring data to the data processing unit;

the data processing unit is used for establishing a nonlinear mapping relation between monitoring data and tunnel soil pressure and reinforcing steel bar stress through a fitting algorithm based on a least square support vector machine, and predicting the soil pressure and the reinforcing steel bar stress in the construction process by adopting a bacterial foraging algorithm based on a time sequence; the data acquisition unit includes: the device comprises a data monitoring part, a data acquisition part, a protection structure part and a support structure part;

the data monitoring part is used for monitoring and acquiring change monitoring data of soil pressure at the arch crown, the arch shoulder and the arch waist and the stress of the reinforcing steel bars at the arch crown and the arch waist;

the data acquisition part is used for collecting and sending the data monitoring unit to the data transmission unit; the data transmission unit includes a data receiving section; the data receiving section includes: the data wireless receiver is connected with the transmission box signal receiving antenna through a data line; the transmission box signal receiving antenna is used for wireless transmission of data;

The protection structure part is used for accommodating each data acquisition part; so as to realize water resistance, flying stone sputtering resistance, dust particle resistance and engineering equipment collision resistance;

the protection structure part comprises a box body structure formed by a top plate, side plates and a bottom plate and a waterproof outer edge of the collection box; the side plates comprise a front side plate and a side plate; the front side plate is fixed on the side plate through a rotating shaft, so that the front side plate can rotate around the rotating shaft, and the box body structure is opened for installation and adjustment; the waterproof outer edge of the collecting box is positioned at the outer edge of the top plate, so that dripping water in a hole can be prevented from penetrating into the collecting box;

the support structure portion is used to secure the protection structure portion to a tunnel sidewall.

2. A PBA structural deformation risk prediction system according to claim 1, wherein the data monitoring section comprises: a soil pressure vault measuring module, a soil pressure arch waist measuring module, a steel bar stress vault measuring module and a steel bar stress arch waist measuring module;

the soil pressure vault measuring module is used for acquiring sampling data corresponding to the vault monitoring direction;

the soil pressure arch shoulder measuring module is used for acquiring sampling data corresponding to the arch shoulder monitoring direction;

The soil pressure arch measuring module is used for acquiring sampling data corresponding to the arch monitoring direction;

the steel bar stress vault measuring module is used for acquiring sampling data corresponding to the vault monitoring direction;

the steel bar stress arch measuring module is used for acquiring sampling data corresponding to the arch monitoring direction.

3. A PBA structural deformation risk prediction system according to claim 1, wherein the data acquisition section comprises: the data automatic acquisition module and the data automatic transmission module;

the data automatic acquisition module is used for monitoring the data monitoring part in real time;

the data automatic transmission module is connected with the data automatic acquisition module through a data line, so that the acquired data can be sent to the data transmission unit in real time.

4. A method of using a PBA structural deformation risk prediction system according to any of claims 1-3, comprising:

s1, the data processing unit receives monitoring data transmitted by the data transmission unit and establishes sample data based on a time sequence, wherein the sample data comprises: values for the arch crown soil pressure, arch shoulder soil pressure, arch crown reinforcement stress on day i;

S2, the data processing unit learns time sequence data by utilizing a least square support vector machine theory to obtain a nonlinear relation between a soil pressure value and a reinforcement stress value change sequence;

s3, the data processing unit determines a mapping relation corresponding to a training set formed by means of time sequence monitoring data to obtain a nonlinear prediction model, and optimizes the nonlinear mapping model through trend operation, copy operation and migration operation of a bacterial foraging algorithm based on the nonlinear prediction model to realize rolling prediction of structural deformation risks.

5. The method of claim 4, wherein S2 comprises:

s21, acquiring soil pressure and steel bar stress data of a monitoring section of the tunnel face based on the determined distance data of the monitoring position from the tunnel face;

s22, predicting the nonlinear change sequence, namely obtaining the relation between the soil pressure and the steel bar stress at the time p+1 and the soil pressure and the steel bar stress at the time p before;

s23, training the relation between the soil pressure and the steel bar stress change through time sequence data by utilizing a least square support vector machine theory to obtain a nonlinear relation between the soil pressure and the steel bar stress change sequence:

Wherein y (x) _p+1 ) The soil pressure and the steel bar stress value at the p+1 time are obtained; b is the offset;as a kernel function, K (x, x _k )＝exp{||x-x _i || ² /σ ² }，σ ₂ Is the square bandwidth of the gaussian RBF kernel.

6. A method of predicting risk of deformation of a PBA structure according to claim 4, wherein said data processing unit, said S3 comprises:

s31, establishing a nonlinear prediction model by utilizing a nonlinear relation between a soil pressure value and a reinforcement stress value change sequence, wherein the nonlinear prediction model is used for predicting data of a current position p+1 through the previous P pieces of historical data;

s32, adding groundwater level influence parameters into an input time sequence, and correcting each prediction process:

s321, trending operation: any possible increase or decrease change in the set range is carried out on the data value for a set of time series monitoring data, so as to obtain a set of new time series monitoring data, and the operation is repeated;

s322, copy operation: sequentially arranging all time series monitoring data from large to small according to the absolute value of the predicted value distance monitoring value as an evaluation index when the trend operation reaches the set critical frequency, deleting 50% of data, and copying the data of the previous 50%;

s323, migration operation: when the migration operation occurs in the process of predicting the next data, the execution reference point is the copying operation, and after the copying operation is carried out to a set limit step, if the data meets the requirements, the data is reserved; if the data does not meet the requirements, deleting the data;

S33, performing corresponding rolling prediction on the sampling data obtained by the data acquisition unit based on a nonlinear prediction model optimized by a bacterial foraging algorithm.