CN113074649A - Method for monitoring foundation pile of high-pile wharf - Google Patents

Method for monitoring foundation pile of high-pile wharf Download PDF

Info

Publication number
CN113074649A
CN113074649A CN202110302961.8A CN202110302961A CN113074649A CN 113074649 A CN113074649 A CN 113074649A CN 202110302961 A CN202110302961 A CN 202110302961A CN 113074649 A CN113074649 A CN 113074649A
Authority
CN
China
Prior art keywords
unfavorable
pile
optical fiber
output
strain sensor
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110302961.8A
Other languages
Chinese (zh)
Other versions
CN113074649B (en
Inventor
周世良
吴俊�
舒岳阶
徐瑛
孙世泉
谢雷
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing Shipping Construction Development Group Co ltd
Chongqing Jiaotong University
Original Assignee
Chongqing Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing Jiaotong University filed Critical Chongqing Jiaotong University
Priority to CN202110302961.8A priority Critical patent/CN113074649B/en
Priority to CN202210679285.0A priority patent/CN114877820B/en
Publication of CN113074649A publication Critical patent/CN113074649A/en
Application granted granted Critical
Publication of CN113074649B publication Critical patent/CN113074649B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/16—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge
    • G01B11/18—Measuring arrangements characterised by the use of optical techniques for measuring the deformation in a solid, e.g. optical strain gauge using photoelastic elements
    • E—FIXED CONSTRUCTIONS
    • E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
    • E02D—FOUNDATIONS; EXCAVATIONS; EMBANKMENTS; UNDERGROUND OR UNDERWATER STRUCTURES
    • E02D33/00—Testing foundations or foundation structures
    • E—FIXED CONSTRUCTIONS
    • E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
    • E02D—FOUNDATIONS; EXCAVATIONS; EMBANKMENTS; UNDERGROUND OR UNDERWATER STRUCTURES
    • E02D5/00—Bulkheads, piles, or other structural elements specially adapted to foundation engineering
    • E02D5/22—Piles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01L—MEASURING FORCE, STRESS, TORQUE, WORK, MECHANICAL POWER, MECHANICAL EFFICIENCY, OR FLUID PRESSURE
    • G01L1/00—Measuring force or stress, in general
    • G01L1/24—Measuring force or stress, in general by measuring variations of optical properties of material when it is stressed, e.g. by photoelastic stress analysis using infrared, visible light, ultraviolet
    • G01L1/242—Measuring force or stress, in general by measuring variations of optical properties of material when it is stressed, e.g. by photoelastic stress analysis using infrared, visible light, ultraviolet the material being an optical fibre
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/727—Offshore wind turbines

Abstract

The invention discloses a method for monitoring a foundation pile of a high-pile wharf, which is characterized in that an optical fiber strain sensor is arranged on the foundation pile, one end of the optical fiber strain sensor is connected with a control center to realize monitoring of the foundation pile, the optical fiber strain sensor is provided with a part arranged on water and a part arranged under water, and the optical fiber strain sensor is vertically arranged along the surface of the foundation pile in a fitting manner and is fixed and protected by virtue of underwater epoxy resin covered on the surface. The method is convenient and quick to implement, has good fixing and protecting effects on the optical fiber sensor, enables monitoring to be more accurate and reliable, has a longer service life, can analyze the cause of damage of the high pile, and can better realize safety monitoring on the wharf.

Description

Method for monitoring foundation pile of high-pile wharf
Technical Field
The invention relates to the technical field of wharf safety monitoring, in particular to a method for monitoring foundation piles of a high-pile wharf.
Background
The high-pile wharf is a wharf mainly composed of an upper structure, a pile foundation, wharf equipment and the like. The superstructure forms a quay surface and is integrally connected with the pile foundation, directly bears the vertical and horizontal loads acting on the quay surface, and transmits the loads to the pile foundation. The pile foundation is used to support the superstructure and transfer the load of the superstructure and quay surface into the foundation through pile-soil (rock) interaction. The high-pile wharf has the advantages of adapting to large water level amplitude variation, good mooring condition, high loading and unloading efficiency and the like, and is a main structural type for constructing wharfs of inland rivers and seaports at present.
The high-pile wharf can gradually generate accumulative damage under the common unfavorable inducement effects of ship nonstandard berthing, local over-limit stacking load, storm flow impact and the like, and the reliability of the structure is seriously influenced. Some investigation results of 29 seaport high-piled wharf structures in east China and south China show that more than 90% of wharfs have structural damage to different degrees within 20 years of wharf operation. The detection statistical data of hundreds of high-pile wharfs such as Ningbo mountain wharfs, northriver wharfs, north warehouse wharfs and the like show that: after the wharf is in service for many years, the structure is generally damaged to a certain degree; the serious damage of the upper structure accounts for 18 percent, the serious damage of the pile foundation accounts for 72 percent, and the damage of other structures accounts for 10 percent. The high piles need to be monitored to ensure the safety of use.
The existing patents and documents mainly aiming at the condition monitoring requirement of the high-pile wharf only introduce a simple monitoring function. As the utility model (CN205879247U) proposes a health monitoring system for high-piled wharf based on optical fiber sensing technology and BIM technology, various sensors are arranged on the high-piled wharf to monitor the stress state of the high-piled wharf; the invention patent (CN104567794A) combines AIS and wireless communication to provide a foundation pile deformation monitoring system. However, the above-mentioned patent methods simply use the sensing technology to monitor the stress condition of the high-pile wharf, and cannot further analyze and identify the unfavorable inducement causing the structural damage according to the monitoring data, so that the unfavorable inducement currently acting on the high-pile wharf cannot be specifically excluded. Because the unfavorable inducement acting on the wharf cannot be known, the action of the unfavorable inducement cannot be eliminated or weakened by taking corresponding measures in time, and the wharf is always in a sub-health state. When damage of the wharf structure is accumulated to a certain degree, the wharf structure may be seriously damaged until the wharf structure is completely failed, and long-term safe and reliable operation of the high-pile wharf is seriously threatened.
In addition, the above patent technologies do not disclose the installation method of the optical fiber sensor, and the optical fiber sensor is only installed by adopting a conventional installation method. When the optical fiber sensor is installed, the optical fiber sensor is fixed on the water part of the foundation pile in a welding, bundling or surface bonding mode, so that the strain state detection of the foundation pile on the water part can be realized only. The pile foundation of the high-pile wharf is dozens of meters long, most of the high-pile wharf is located underwater, and the stress state of the pile foundation cannot be completely reflected only through the strain monitoring values of the above-water partial nodes. When the optical fiber sensor is installed underwater, if a common installation mode is adopted, the fixing effect and the protection effect of the optical fiber sensor are poor, the optical fiber is corroded by seawater for a long time, sea wave impact, invasion of marine organisms and other factors act, the factors directly act on the optical fiber to cause damage, and confusion can be caused to a detection result. Therefore, in the prior art, detection errors are easily caused due to improper installation of the optical fibers, so that the detection precision is quickly reduced, the service life is short, and the long-term safe and reliable operation of the high pile wharf is threatened.
Disclosure of Invention
Aiming at the defects of the prior art, the technical problems to be solved by the invention are as follows: how to provide a convenient quick of implementation, fixed with the protection effectual to fiber sensor for monitoring is more accurate reliable, the long high stake pier foundation pile monitoring method of life to further can analyze the high stake damage and induce the reason, can realize the pier safety monitoring better.
In order to solve the technical problems, the invention adopts the following technical scheme:
a method for monitoring foundation piles of a high-pile wharf is characterized in that an optical fiber strain sensor is provided with a part for installing the foundation piles on water and a part for installing the foundation piles under water, and the optical fiber strain sensor is vertically arranged along the surfaces of the foundation piles in a fitting mode and fixed and protected by means of underwater epoxy resin covering the surfaces.
Like this, optic fibre strain transducer installs on the foundation pile and links to each other with control center in this scheme, and the sensor can real-time detection foundation pile surface receives the stress variation condition, turns into detected signal and then realizes the control to the foundation pile. The epoxy resin in water is a curing material which can be used under water. In the scheme of the invention, the optical fiber strain sensor is attached and fixed on the foundation pile by means of underwater epoxy resin, namely, the optical fiber strain sensor is provided with an overwater part and an underwater part. Therefore, the stress state of the pile foundation can be monitored and reflected more completely and accurately. The optical fiber is fixed and protected by means of underwater epoxy resin, the detection precision is prevented from being influenced by damage caused by seawater erosion, sea wave impact, invasion of marine organisms and the like, and the monitoring reliability is guaranteed. Meanwhile, the method is convenient for the installation and fixation of the optical fiber strain sensor, cannot damage or influence the self quality of the foundation pile, and is suitable for newly built or built high-pile wharfs.
Further, the optical fiber strain sensor is an optical fiber grating distributed strain sensor or an optical fiber Brillouin distributed strain sensor.
Has the advantages of mature product, reliable detection, low cost and the like.
Furthermore, each foundation pile is symmetrically provided with two measuring lines, one measuring line is positioned on one side facing a water area, the other measuring line is positioned on one side facing a bank slope, and each measuring line is correspondingly provided with one optical fiber strain sensor.
Therefore, when the foundation pile is influenced by ship berthing, sea wave impact, wind current and the like, the foundation pile is stressed along the direction of the water surface facing the bank slope under most conditions, two measuring lines are arranged in the direction, the stress and strain conditions of the foundation pile can be detected to the greatest extent by using the least strain sensors, and the monitoring accuracy is improved.
Further, the optical fiber strain sensor is installed according to the following steps:
the first step is as follows: removing impurities on the surface of the pile foundation and keeping the surface smooth;
the second step is that: straightening the optical fiber strain sensor along a preset measuring line position of the foundation pile, and fixing the optical fiber strain sensor on the surface of the foundation pile through a temporary fixing point;
the third step: the glue injection template is adopted for assisting installation, the cross section of the glue injection template is in an arc shape, the length of the glue injection template is equivalent to that of the distributed optical fiber strain sensor to be installed, and the glue injection template covers the optical fiber strain sensor and realizes fixation;
the fourth step: injecting underwater epoxy resin into the glue injection template from bottom to top by using a glue injection machine, filling gaps in the glue injection template with the underwater epoxy glue from bottom to top, and adhering the sensor to the surface of the foundation pile;
the fifth step: and (5) removing the glue injection template after the underwater epoxy resin is hardened, and finishing the installation of the optical fiber strain sensor.
In this way, the glue injection template is adopted for assisting installation, so that the installation is convenient and rapid, the installation and fixation are reliable, and the monitoring precision is ensured; meanwhile, the foundation pile is not damaged, and the method is suitable for newly built or built high-pile wharfs.
Furthermore, in the second step, the optical fiber strain sensor is fixed on the surface of the foundation pile through three temporary fixing points, wherein one point is located at the position, close to the upper end, of the foundation pile, one point is located at the position, close to the lower end, of the foundation pile, and the other point is located at the middle position of the foundation pile.
The operation is simple, convenient and fast, and the fixation is reliable.
Furthermore, in the second step, the fixation of the three temporary fixing points of the optical fiber strain sensor is realized by adopting an underwater epoxy resin glue dispensing mode through a glue injection machine.
Like this, fixed convenient and fast is reliable, and the fixed underwater epoxy of gluing combines as an organic whole with the underwater epoxy of follow-up injecting glue, and the monolithic stationary is more reliable stable.
Further, the both sides of injecting glue template width direction have one section be used for with the laminating portion of foundation pile surface laminating, the middle part of width direction has bellied optic fibre and holds the chamber, optic fibre holds the protruding height in chamber and is greater than the optic fibre diameter, the whole arc that is laminating foundation pile surface of injecting glue template lower extreme tip position, the upper end tip leaves optic fibre and holds the chamber export, the injecting glue template side is close to the lower extreme position and opens there is the injecting glue entry, injecting glue entry and optic fibre hold the chamber and communicate with each other.
Like this, can make things convenient for the injecting glue template to cover optic fibre strain transducer more, and optic fibre holds the die plate of chamber both sides and lower extreme and can laminate with the reference surface, epoxy flows from the gap under water when avoiding the injecting glue, and through the injecting glue machine with epoxy from injecting glue entry position under water pour into during the injecting glue, can realize the injecting glue better.
Further, the glue injection templates are arranged in pairs in a bilateral symmetry mode, a hoop mechanism for compressing is connected and arranged between each pair of glue injection templates, the hoop mechanism comprises a left hoop and a right hoop, one ends of the left hoop and the right hoop are hinged, the other ends of the left hoop and the right hoop are provided with quick lock catches for opening and closing ends, and the glue injection templates are vertically fixed in the middle positions of the left hoop and the right hoop.
Like this, the injecting glue template uses in pairs when using, and a pair of injecting glue template lock respectively on two optic fibre strain sensor of foundation pile, then relies on quick hasp to realize the closed locking of switching end for the injecting glue template is tightly pasted on the foundation pile, and the laminating is fixed very reliably stably. The glue injection template is convenient to use, the glue injection template is not required to be fixed by hands, and the glue injection operation efficiency is greatly improved.
Furthermore, the upper end and the lower end of each pair of glue injection templates are respectively provided with a hoop mechanism close to the end part. Therefore, the glue injection template can be better pressed and fixed.
And further, when the high-pile wharf is an inland river overhead vertical wharf, construction and installation are carried out by utilizing a low water level period.
And furthermore, when the high-pile wharf is a harbor high-pile wharf, arranging divers for construction and installation.
Furthermore, after the optical fiber strain sensor is installed, an unfavorable incentive recognition classifier model is established in the control center, unfavorable incentive recognition training is carried out, and automatic recognition of unfavorable incentives is realized according to detection signals of the optical fiber strain sensor during monitoring.
Therefore, automatic identification training of the unfavorable inducement can realize automatic identification of the detection signal during monitoring, and the unfavorable inducement information of the monitoring personnel is fed back, so that the monitoring personnel can better and conveniently make a targeted response in time. The safety guarantee of wharf monitoring is improved better. The classifier model can adopt a parallel support vector machine classifier model, so that the specific situation of dock unfavorable incentive recognition is better met, and the recognition is more accurate and reliable.
Preferably, the establishment, training and recognition of the unfavorable cause recognition classifier model are realized in the following way:
a. determining unfavorable causes and acquiring test data; determining unfavorable inducement factors of 4 high-pile wharfs, wherein 1 is ship unconventional berthing, 2 is local over-limit stacking, 3 is wind action, and 4 is water flow action; acquiring distributed strain data of pile groups under the independent action of four unfavorable inducers by using the installed optical fiber strain sensor as test data of an unfavorable inducer inversion model;
b. extracting a characteristic vector; carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; carrying out frequency analysis on the test data, and extracting the first third harmonic frequency x of the signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then each distributed optical fiber test data characteristic parameter forms a characteristic vector x ═ (x)1,x2,...,x9) Since each feature value in x has a certain correlation and a too high dimension, the principal component y of x is determined by the PCA algorithm as (y ═ y)1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 4 unfavorable causes, four parallel support vector machine classifiers are established:
the SVM1 distinguishes the abnormal berthing of the ship from other 3 unfavorable inducers, when the ship is in normal berthing, the output of the SVM1 takes +1, otherwise takes-1; f. of1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000051
SVM 2: the classifier distinguishes local over-limit stowage from other 3 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM2 is +1, otherwise, the output is-1; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000052
SVM 3: the classifier distinguishes wind effects from other 3 unfavorable causes, with the SVM3 output taking +1 when wind effects, otherwise-1. f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000053
SVM 4: the classifier distinguishes the water flow effect from other 3 unfavorable inducers, when the water flow effect is achieved, the output of the SVM4 is taken as +1, otherwise, the output is taken as-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000054
d. training model parameters; inputting the sample data obtained in the step b into a parallel support vector machine classifier model, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3 and SVM4 by using an SVM training algorithm, so as to parameterize the parallel support vector machine classifier model;
Figure BDA0002987042560000055
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; in the specific process, if the output of the SVM1 is +1, the situation that the ship does not normally berth exists is indicated, and otherwise, the situation that the ship does not exist; the output of the SVM2 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of the SVM3 is +1, which indicates that wind action exists, otherwise, the wind action does not exist; the SVM4 output is +1 indicating that water flow effects are present, otherwise they are not.
Therefore, by performing unfavorable cause recognition in the above manner, and enumerating all classifiers with an output of +1 during recognition, the type of unfavorable cause acting on the high-piled wharf can be determined. For example, if the outputs of the SVM1 and SVM3 are +1 and the outputs of the SVM2 and SVM4 are-1, the adverse inducement currently acting on the high-piled wharf is that the ship does not regulate the berthing and wind action. In the step a, the acquired data can be acquired through a specific detection or test mode, specifically, the data of the wind action factor and the water flow action factor can be acquired through specific actual detection of a sensor for detecting wind power arranged on the wharf water surface and a sensor for detecting water flow impact force arranged under the wharf water surface. The data of the local overrun stacking action factors can be obtained through a specific test mode, namely the data is obtained after overrun weights are stacked at specific positions on the surface of the wharf. The data of the ship nonstandard berthing factors can be acquired through a specific test mode or can be acquired according to the actual ship berthing condition detection. Thus, the method can better ensure the accuracy and reliability of the recognition of the adverse factors.
As another preferred mode, the establishment, training and recognition of the unfavorable cause recognition classifier model are implemented as follows:
a. determining unfavorable causes and acquiring test data;
determining unfavorable inducement factors of 6 high-pile wharfs, wherein 1 is ship nonstandard berthing, 2 is bank slope uneven settlement, 3 is local ultra-limited stacking load, 4 is material shape deterioration, 5 is wind wave flow impact, and 6 is earthquake action; the method comprises the following steps of obtaining 1 pile group distributed strain data under the independent action of three unfavorable inducements of ship nonstandard berthing, 3 local over-limit stacking load and 5 storm flow impact by using an installed optical fiber strain sensor; meanwhile, for the other three unfavorable inducers, establishing an ABAQUS numerical simulation analysis model of the high-pile wharf, and simulating the pile group distributed strain data under the independent action of the unfavorable inducers in a software model; the actual test data and the simulation data are jointly used as test data of the unfavorable incentive inversion model;
b. extracting a characteristic vector;
carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; carrying out frequency analysis on the test data, and extracting the first third harmonic frequency x of the signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then, a high-dimensional characteristic vector x (x) is extracted from the test data obtained from each distributed optical fiber1,x2,...,x9) Wherein x is1,x2,…,x99 characteristic values of pile group strain distribution data are obtained; since each feature quantity in the high-dimensional feature vector x has a certain correlation and has an excessively high dimension, the principal component y of x is obtained by the PCA algorithm (y is equal to1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 6 unfavorable causes, six parallel support vector machine classifier models are established, and the 6 unfavorable causes are classified and identified:
SVM1: the classifier distinguishes the ship nonstandard berthing from other 5 unfavorable inducers, when the ship is standard berthed, the output of the SVM1 takes +1, otherwise takes-1, f1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000071
SVM 2: the classifier distinguishes bank slope uneven settlement from other 5 unfavorable inducers, when bank slope uneven settlement is adopted, the output of the SVM2 is +1, otherwise-1 is adopted; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000072
SVM 3: the classifier distinguishes local over-limit stowage from other 5 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM3 is +1, otherwise, the output is-1; f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000073
SVM 4: the classifier distinguishes the material property deterioration from other 5 unfavorable causes, when the material property deterioration is detected, the output of the SVM4 is +1, otherwise, the output is-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000074
SVM 5: the classifier distinguishes the storm surge from other 5 unfavorable inducers, when the storm surge is adopted, the output of the SVM5 takes +1, otherwise takes-1; f. of5(y) is a classification function of SVM5, as shown in equation (5), where a5i(i＝1,2,…n)， b5As a classification function f5(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000075
SVM 6: the classifier distinguishes the earthquake action from other 5 unfavorable inducers, when the earthquake is impacted by wind, wave and current, the output of the SVM6 takes +1, otherwise takes-1; f. of6(y) is a classification function of SVM6, as shown in equation (6), where a6i(i＝1,2,…n)，b6As a classification function f6(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000076
d. training model parameters; inputting the sample data obtained in the step b into six parallel support vector machine classifier models, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3, SVM4, SVM5 and SVM6 by using an SVM training algorithm, so as to parameterize the 6 parallel support vector machine classifier models;
Figure BDA0002987042560000081
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; the specific process is as follows: the output of the SVM1 is +1, which indicates that the ship does not standardize berthing, otherwise, the output does not exist; the output of the SVM2 is +1, which indicates that bank slope non-uniformity settlement exists, otherwise, the bank slope non-uniformity settlement does not exist; the output of the SVM3 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of SVM4 is +1, indicating that there is material property degradation, otherwise it is not present; the output of the SVM5 is +1, which indicates that the storm surge exists, otherwise, the storm surge does not exist; the output of SVM6 is +1, indicating the presence of seismic action, otherwise it is not.
The method is adopted to identify unfavorable inducers. Compared with the former mode, the wind flow and wave flow impact often occur simultaneously and have consistency, so the wind flow effect and the wave flow effect are regarded as the same factor to simplify the model. The data acquisition mode of the three unfavorable causes 1, 3 and 5 is the same as that of the first mode, and can be actually acquired through specific detection or test. Meanwhile, three unfavorable inducers which are 2, 4 and 6 and difficult to detect on site are added in the method, a high-pile code head ABAQUS numerical simulation analysis model can be established, and pile group distributed strain data under the independent action of the unfavorable inducers can be simulated in a software model. The ABAQUS software is a set of existing mature finite element software with powerful engineering simulation functions, and the software can be used for simulating the performance of typical engineering materials and solving the problem of simulation analysis of structural stress and displacement of constructional engineering. Is mature prior art. For the application, the data of three unfavorable causes which are difficult to detect on site can be increased, and the comprehensiveness of model identification is greatly improved.
The method has the advantages that the method for inverting the unfavorable inducement of the high-pile wharf in real time according to the group pile distributed strain data is provided for the first time, the input distributed strain data is analyzed through the parallel support vector machine classifier model, the type of the unfavorable inducement currently acting on the wharf is judged in real time, and the method has great significance for timely eliminating dangerous cases, preventing major safety accidents and guaranteeing safe and reliable operation of the high-pile wharf.
In conclusion, the method is convenient and quick to implement, has good fixing and protecting effects on the optical fiber sensor, enables monitoring to be accurate and reliable, has long service life, can analyze the cause of damage of the high pile, and can better realize safety monitoring on the wharf.
Drawings
Fig. 1 is a schematic view of a pier stud in example 1 when an optical fiber strain sensor is installed.
Fig. 2 is an enlarged view a-a of fig. 1.
Fig. 3 is a schematic diagram of the classifier model principle of the parallel support vector machine in embodiment 1.
FIG. 4 is a schematic diagram of the classifier model of the parallel support vector machine in embodiment 2.
Detailed Description
The present invention will be described in further detail with reference to specific embodiments.
Specific example 1: a method for monitoring foundation piles of a high-pile wharf is disclosed, and referring to fig. 1-2, in the method, an optical fiber strain sensor 2 is installed on a foundation pile 1, one end of the optical fiber strain sensor 2 is connected with a control center (not shown in the figure) to realize monitoring of the foundation pile, wherein the optical fiber strain sensor 2 is provided with a part for installing the foundation pile on water and a part for installing the foundation pile under water, and the optical fiber strain sensor 2 is vertically arranged along the surface of the foundation pile 1 in a fitting mode and depends on underwater epoxy resin covered on the surface to realize fixing and protection.
Like this, optic fibre strain transducer installs on the foundation pile and links to each other with control center in this scheme, and the sensor can real-time detection foundation pile surface receives the stress variation condition, turns into detected signal and then realizes the control to the foundation pile. The epoxy resin in water is a curing material which can be used under water. In the scheme of the invention, the optical fiber strain sensor is attached and fixed on the foundation pile by means of underwater epoxy resin, namely, the optical fiber strain sensor is provided with an overwater part and an underwater part. Therefore, the stress state of the pile foundation can be monitored and reflected more completely and accurately. The optical fiber is fixed and protected by means of underwater epoxy resin, the detection precision is prevented from being influenced by damage caused by seawater erosion, sea wave impact, invasion of marine organisms and the like, and the monitoring reliability is guaranteed. Meanwhile, the method is convenient for the installation and fixation of the optical fiber strain sensor, cannot damage or influence the self quality of the foundation pile, and is suitable for newly built or built high-pile wharfs.
The optical fiber strain sensor 2 is an optical fiber grating distributed strain sensor or an optical fiber Brillouin distributed strain sensor.
Has the advantages of mature product, reliable detection, low cost and the like.
Wherein, each foundation pile 1 is gone up the symmetry and is set up two survey lines, one is located and faces waters one side, and another is located and faces bank slope one side, and every survey line is arranged correspondingly and is installed an optic fibre strain sensor 2.
Therefore, when the foundation pile is influenced by ship berthing, sea wave impact, wind current and the like, the foundation pile is stressed along the direction of the water surface facing the bank slope under most conditions, two measuring lines are arranged in the direction, the stress and strain conditions of the foundation pile can be detected to the greatest extent by using the least strain sensors, and the monitoring accuracy is improved.
The optical fiber strain sensor is installed according to the following steps:
the first step is as follows: removing impurities on the surface of the pile foundation and keeping the surface smooth;
the second step is that: straightening the optical fiber strain sensor along a preset measuring line position of the foundation pile, and fixing the optical fiber strain sensor on the surface of the foundation pile through a temporary fixing point;
the third step: the glue injection template 3 is adopted for assisting installation, the cross section of the glue injection template 3 is arc-shaped, the length of the glue injection template 3 is equivalent to that of a distributed optical fiber strain sensor to be installed, and the glue injection template covers the optical fiber strain sensor and realizes fixation;
the fourth step: injecting underwater epoxy resin into the glue injection template from bottom to top by using a glue injection machine, filling gaps in the glue injection template with the underwater epoxy glue from bottom to top, and adhering the sensor to the surface of the foundation pile;
the fifth step: and (5) removing the glue injection template after the underwater epoxy resin is hardened, and finishing the installation of the optical fiber strain sensor.
In this way, the glue injection template is adopted for assisting installation, so that the installation is convenient and rapid, the installation and fixation are reliable, and the monitoring precision is ensured; meanwhile, the foundation pile is not damaged, and the method is suitable for newly built or built high-pile wharfs.
In the second step, the optical fiber strain sensor is fixed on the surface of the foundation pile through three temporary fixing points, wherein one point is located at the position, close to the upper end, of the foundation pile, one point is located at the position, close to the lower end, of the foundation pile, and the other point is located at the middle position of the foundation pile.
The operation is simple, convenient and fast, and the fixation is reliable.
In the second step, the fixation of three temporary fixed points of the optical fiber strain sensor is realized by adopting a glue dispensing mode of underwater epoxy resin by a glue injection machine.
Like this, fixed convenient and fast is reliable, and the fixed underwater epoxy of gluing combines as an organic whole with the underwater epoxy of follow-up injecting glue, and the monolithic stationary is more reliable stable.
Wherein, 3 width direction's of injecting glue template both sides have one section be used for with the laminating portion 4 of foundation pile surface laminating, and width direction's middle part has bellied optic fibre and holds the chamber, and optic fibre holds the protruding height in chamber and is greater than the optic fibre diameter, and 3 lower extreme tip positions of injecting glue template wholly are the arc on laminating foundation pile surface, and the upper end tip leaves optic fibre and holds the chamber export, and the injecting glue template side is close to lower extreme position and opens there is injecting glue entry 5, and injecting glue entry 5 and optic fibre hold the chamber and communicate with each other.
Like this, can make things convenient for the injecting glue template to cover optic fibre strain transducer more, and optic fibre holds the die plate of chamber both sides and lower extreme and can laminate with the reference surface, epoxy flows from the gap under water when avoiding the injecting glue, and through the injecting glue machine with epoxy from injecting glue entry position under water pour into during the injecting glue, can realize the injecting glue better.
Wherein, injecting glue template 3 bilateral symmetry sets up in pairs, and every pair connects between the injecting glue template 3 to be provided with the staple bolt mechanism that compresses tightly the usefulness, and staple bolt mechanism includes left strap 6 and right strap 7, and the articulated setting of one end of left strap 6 and right strap 7, the end of just opening and shutting for the end that opens and shuts of other one end is provided with quick hasp 8, the vertical middle part position of fixing at left strap and right strap of 3 boards of injecting glue mould.
Like this, the injecting glue template uses in pairs when using, and a pair of injecting glue template lock respectively on two optic fibre strain sensor of foundation pile, then relies on quick hasp to realize the closed locking of switching end for the injecting glue template is tightly pasted on the foundation pile, and the laminating is fixed very reliably stably. The glue injection template is convenient to use, the glue injection template is not required to be fixed by hands, and the glue injection operation efficiency is greatly improved.
Wherein, the upper and lower ends of each pair of glue injection templates are respectively provided with a hoop mechanism near the end part. Therefore, the compressing and fixing of the glue injection template can be better realized.
When the high-pile wharf is implemented, when the high-pile wharf is an inland river overhead vertical wharf, construction and installation are carried out by utilizing a low water level period. When the high-pile wharf is a harbor high-pile wharf, arranging divers for construction and installation.
After the optical fiber strain sensor is installed, an unfavorable incentive recognition classifier model is established in a control center, unfavorable incentive recognition training is carried out, and automatic recognition of unfavorable incentives is achieved according to detection signals of the optical fiber strain sensor during monitoring.
Therefore, automatic identification training of the unfavorable inducement can realize automatic identification of the detection signal during monitoring, and the unfavorable inducement information of the monitoring personnel is fed back, so that the monitoring personnel can better and conveniently make a targeted response in time. The safety guarantee of wharf monitoring is improved better. The classifier model can adopt a parallel support vector machine classifier model, so that the specific situation of dock unfavorable incentive recognition is better met, and the recognition is more accurate and reliable.
In this embodiment, the establishment, training and recognition of the unfavorable cause recognition classifier model are implemented as follows, and is shown in fig. 3:
a. determining unfavorable causes and acquiring test data; determining unfavorable inducement factors of 4 high-pile wharfs, wherein 1 is ship unconventional berthing, 2 is local over-limit stacking, 3 is wind action, and 4 is water flow action; acquiring distributed strain data of pile groups under the independent action of four unfavorable inducers by using the installed optical fiber strain sensor as test data of an unfavorable inducer inversion model;
b. extracting a characteristic vector; carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; carrying out frequency analysis on the test data, and extracting the first third harmonic frequency x of the signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then each distributed optical fiber test data characteristic parameter forms a characteristic vector x ═ (x)1,x2,...,x9) Since each feature value in x has a certain correlation and a too high dimension, the principal component y of x is determined by the PCA algorithm as (y ═ y)1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 4 unfavorable causes, four parallel support vector machine classifiers are established:
the SVM1 distinguishes the abnormal berthing of the ship from other 3 unfavorable inducers, when the ship is in normal berthing, the output of the SVM1 takes +1, otherwise takes-1; f. of1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000121
SVM 2: the classifier distinguishes local over-limit stowage from other 3 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM2 is +1, otherwise, the output is-1; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000122
SVM 3: the classifier distinguishes wind effects from other 3 unfavorable causes, with the SVM3 output taking +1 when wind effects, otherwise-1. f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000123
SVM 4: the classifier distinguishes the water flow effect from other 3 unfavorable inducers, when the water flow effect is achieved, the output of the SVM4 is taken as +1, otherwise, the output is taken as-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000124
d. training model parameters; inputting the sample data obtained in the step b into a parallel support vector machine classifier model, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3 and SVM4 by using an SVM training algorithm, so as to parameterize the parallel support vector machine classifier model;
Figure BDA0002987042560000125
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; in the specific process, if the output of the SVM1 is +1, the situation that the ship does not normally berth exists is indicated, and otherwise, the situation that the ship does not exist; the output of the SVM2 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of the SVM3 is +1, which indicates that wind action exists, otherwise, the wind action does not exist; the SVM4 output is +1 indicating that water flow effects are present, otherwise they are not.
Therefore, the present embodiment is a method for monitoring foundation piles of a high-pile wharf based on four-factor recognition, wherein unfavorable cause recognition is performed by using the four-factor recognition method, and the unfavorable cause type acting on the high-pile wharf can be determined by enumerating all classifiers with an output of +1 during recognition. For example, if the outputs of SVM1 and SVM3 are +1 and the outputs of SVM2 and SVM4 are-1, it is indicated that the current unfavorable inducement on the high-piled wharf is the ship irregular berthing and wind action. In the step a, the acquired data can be acquired through a specific detection or test mode, and specifically, the data of the wind action factor and the water flow action factor can be acquired through specific actual detection of a sensor for detecting wind power arranged on the wharf water surface and a sensor for detecting water flow impact force arranged under the wharf water surface. The data of the local overrun stacking action factors can be obtained through a specific test mode, namely the data is obtained after overrun weights are stacked at specific positions on the surface of the wharf. The data of the ship nonstandard berthing factors can be acquired in a specific test mode or detected according to the actual ship berthing condition. Thus, the method can better ensure the accuracy and reliability of the recognition of the adverse factors.
Example 2. The rest of this embodiment is the same as embodiment 1, except that the establishment, training and recognition of the unfavorable cause recognition classifier model are implemented as follows, see fig. 4:
a. determining unfavorable causes and acquiring test data;
determining unfavorable inducement factors of 6 high-pile wharfs, wherein 1 is ship nonstandard berthing, 2 is bank slope uneven settlement, 3 is local ultra-limited stacking load, 4 is material shape deterioration, 5 is wind wave flow impact, and 6 is earthquake action; the method comprises the following steps of obtaining 1 pile group distributed strain data under the independent action of three unfavorable inducements of ship nonstandard berthing, 3 local over-limit stacking load and 5 storm flow impact by using an installed optical fiber strain sensor; meanwhile, for the other three unfavorable inducers, establishing an ABAQUS numerical simulation analysis model of the high-pile wharf, and simulating the pile group distributed strain data under the independent action of the unfavorable inducers in a software model; the actual test data and the simulation data are jointly used as test data of the unfavorable incentive inversion model;
b. extracting a characteristic vector;
carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; the frequency analysis is performed on the test data,extracting first third harmonic frequency x of signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then, a high-dimensional characteristic vector x (x) is extracted from the test data obtained from each distributed optical fiber1,x2,...,x9) Wherein x is1,x2,…,x99 characteristic values of pile group strain distribution data are obtained; since each feature quantity in the high-dimensional feature vector x has a certain correlation and has an excessively high dimension, the principal component y of x is obtained by the PCA algorithm (y is equal to1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 6 unfavorable causes, six parallel support vector machine classifier models are established, and the 6 unfavorable causes are classified and identified:
SVM1: the classifier distinguishes the ship nonstandard berthing from other 5 unfavorable inducers, when the ship is standard berthed, the output of the SVM1 takes +1, otherwise takes-1, f1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000141
SVM 2: the classifier distinguishes bank slope uneven settlement from other 5 unfavorable inducers, when bank slope uneven settlement is adopted, the output of the SVM2 is +1, otherwise-1 is adopted; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000142
SVM 3: the classifier distinguishes local over-limit stowage from other 5 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM3 is +1, otherwise, the output is-1; f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000143
SVM 4: the classifier distinguishes the material property deterioration from other 5 unfavorable causes, when the material property deterioration is detected, the output of the SVM4 is +1, otherwise, the output is-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure BDA0002987042560000144
SVM 5: the classifier distinguishes the storm surge from other 5 unfavorable inducers, when the storm surge is adopted, the output of the SVM5 takes +1, otherwise takes-1; f. of5(y) is a classification function of SVM5, as shown in equation (5), where a5i(i＝1,2,…n)， b5As a classification function f5(y) a coefficient, Φ being a gaussian kernel function;
Figure RE-GDA0003034433930000145
SVM 6: the classifier distinguishes the earthquake action from other 5 unfavorable inducers, when the earthquake is impacted by wind, wave and current, the output of the SVM6 takes +1, otherwise takes-1; f. of6(y) is a classification function of SVM6, as shown in equation (6), where a6i(i＝1,2,…n)，b6As a classification function f6(y) a coefficient, Φ being a gaussian kernel function;
Figure RE-GDA0003034433930000146
d. training model parameters; inputting the sample data obtained in the step b into six parallel support vector machine classifier models, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3, SVM4, SVM5 and SVM6 by using an SVM training algorithm, so as to parameterize the 6 parallel support vector machine classifier models;
Figure BDA0002987042560000151
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; the specific process is as follows: the output of the SVM1 is +1, which indicates that the ship does not standardize berthing, otherwise, the output does not exist; the output of the SVM2 is +1, which indicates that bank slope non-uniformity settlement exists, otherwise, the bank slope non-uniformity settlement does not exist; the output of the SVM3 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of SVM4 is +1, indicating that there is material property degradation, otherwise it is not present; the output of the SVM5 is +1, which indicates that the storm surge exists, otherwise, the storm surge does not exist; the output of SVM6 is +1, indicating the presence of seismic action, otherwise it is not.
Therefore, this embodiment 2 is a method for monitoring foundation piles of a high-pile wharf based on six-factor recognition. Compared with the method in the embodiment 1, the method for identifying the six factors has the advantages that the wind flow and wave flow impact are always generated at the same time and have consistency, so that the wind flow effect and the wave flow effect are regarded as the same factor to simplify the model. The data acquisition mode of the three unfavorable causes 1, 3 and 5 is the same as that of the first mode, and can be actually acquired through specific detection or test. Meanwhile, three unfavorable inducers which are difficult to detect on site are added in the method, namely 2, 4 and 6, and a high-pile wharf ABAQUS numerical simulation analysis model can be established, so that pile group distributed strain data under the independent action of the unfavorable inducers can be simulated in a software model. The ABAQUS software is a set of existing mature finite element software with powerful engineering simulation functions, and can be used for simulating the performance of typical engineering materials and solving the problem of simulation analysis of structural stress and displacement of constructional engineering. Is mature prior art. The data of three unfavorable causes which are difficult to detect on site can be increased, and the comprehensiveness of model identification is greatly improved.

Claims (10)

1. A method for monitoring foundation piles of a high-pile wharf is characterized in that an optical fiber strain sensor is provided with a part for installing the foundation piles on water and a part for installing the foundation piles under water, and the optical fiber strain sensor is vertically arranged along the surfaces of the foundation piles in a fitting mode and fixed and protected by means of underwater epoxy resin covering the surfaces.
2. The method for monitoring foundation piles of high-pile wharf according to claim 1, wherein the optical fiber strain sensor is a fiber grating distributed strain sensor or a fiber Brillouin distributed strain sensor.
3. The method for monitoring foundation piles of a high-pile wharf according to claim 1, wherein two measuring lines are symmetrically arranged on each foundation pile, one measuring line is positioned on the side facing the water area, the other measuring line is positioned on the side facing the bank slope, and each measuring line is correspondingly provided with an optical fiber strain sensor.
4. The method of monitoring a piled wharf foundation pile of claim 3, wherein the optical fiber strain sensor is installed according to the following steps:
the first step is as follows: removing impurities on the surface of the pile foundation and keeping the surface smooth;
the second step is that: straightening the optical fiber strain sensor along a preset measuring line position of the foundation pile, and fixing the optical fiber strain sensor on the surface of the foundation pile through a temporary fixing point;
the third step: the glue injection template is adopted for assisting installation, the cross section of the glue injection template is in an arc shape, the length of the glue injection template is equivalent to that of the distributed optical fiber strain sensor to be installed, and the glue injection template covers the optical fiber strain sensor and realizes fixation;
the fourth step: injecting underwater epoxy resin into the glue injection template from bottom to top by using a glue injection machine, filling gaps in the glue injection template with the underwater epoxy glue from bottom to top, and adhering the sensor to the surface of the foundation pile;
the fifth step: and (5) removing the glue injection template after the underwater epoxy resin is hardened, and finishing the installation of the optical fiber strain sensor.
5. The method for monitoring foundation piles of a high-pile wharf according to claim 4, wherein in the second step, the optical fiber strain sensor is fixed to the surface of the foundation pile through three temporary fixing points, one of which is located at a position near the upper end of the foundation pile, one of which is located at a position near the lower end of the foundation pile, and the other of which is located at a position in the middle of the foundation pile;
in the second step, the fixation of the three temporary fixing points of the optical fiber strain sensor is realized by adopting a glue dispensing mode of underwater epoxy resin by a glue injection machine.
6. The method for monitoring the foundation pile of the high-pile wharf according to claim 4, wherein two sides in the width direction of the glue injection template are provided with a section of attaching portion for attaching to the surface of the foundation pile, the middle portion in the width direction is provided with a convex optical fiber accommodating cavity, the convex height of the optical fiber accommodating cavity is larger than the diameter of the optical fiber, the end portion of the lower end of the glue injection template is integrally in an arc shape for attaching to the surface of the foundation pile, an outlet of the optical fiber accommodating cavity is reserved at the end portion of the upper end, a glue injection inlet is formed in the side surface of the glue injection template, close to the lower end, and the glue.
7. The method for monitoring foundation piles of the high-pile wharf according to claim 5, wherein the glue injection formworks are arranged in left-right symmetrical pairs, a clamp mechanism for pressing is connected between each pair of glue injection formworks, the clamp mechanism comprises a left clamp belt and a right clamp belt, one end of the left clamp belt and one end of the right clamp belt are hinged, the other end of the left clamp belt and the other end of the right clamp belt are open-close ends, the open-close ends are provided with quick latches, and the glue injection formworks are vertically fixed in the middle of the left clamp belt and the right clamp belt.
8. The method for monitoring foundation piles of the high-pile wharf according to claim 4, wherein after the optical fiber strain sensor is installed, an unfavorable cause recognition classifier model is established in a control center to perform unfavorable cause recognition training, and automatic recognition of the unfavorable causes is achieved according to detection signals of the optical fiber strain sensor during monitoring.
9. The method of claim 8, wherein the building, training and recognition of the unfavorable cause recognition classifier model is performed as follows:
a. determining unfavorable causes and acquiring test data; determining unfavorable inducement factors of 4 high-pile wharfs, wherein 1 is ship nonstandard berthing, 2 is local over-limit stacking, 3 is wind action, and 4 is water flow action; acquiring distributed strain data of pile groups under the independent action of four unfavorable inducers by using the installed optical fiber strain sensor as test data of an unfavorable inducer inversion model;
b. extracting a characteristic vector; carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; carrying out frequency analysis on the test data, and extracting the first third harmonic frequency x of the signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then each distributed optical fiber test data characteristic parameter forms a characteristic vector x ═ (x)1,x2,...,x9) Since each feature value in x has a certain correlation and a too high dimension, the principal component y of x is determined by the PCA algorithm as (y ═ y)1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 4 unfavorable causes, four parallel support vector machine classifiers are established:
SVM1 the classifier distinguishes the ship nonstandard berth from other 3 unfavorable inducers when the classification is a ship specificationWhen the model is berthed, the output of the SVM1 is +1, otherwise, -1 is selected; f. of1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000031
SVM 2: the classifier distinguishes local over-limit stowage from other 3 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM2 is +1, otherwise, the output is-1; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000032
SVM 3: the classifier distinguishes wind effects from other 3 unfavorable causes, with the SVM3 output taking +1 when wind effects, otherwise-1. f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000033
SVM 4: the classifier distinguishes the water flow effect from other 3 unfavorable inducers, when the water flow effect is achieved, the output of the SVM4 is taken as +1, otherwise, the output is taken as-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000034
d. training model parameters; inputting the sample data obtained in the step b into a parallel support vector machine classifier model, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3 and SVM4 by using an SVM training algorithm, so as to realize parameterization of the parallel support vector machine classifier model;
Figure FDA0002987042550000035
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; in the specific process, if the output of the SVM1 is +1, the situation that the ship does not normally berth exists is indicated, and otherwise, the situation that the ship does not exist; the output of the SVM2 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of the SVM3 is +1, which indicates that wind action exists, otherwise, the wind action does not exist; the SVM4 output is +1 indicating that water flow effects are present, otherwise they are not.
10. The method of claim 8, wherein the building, training and recognition of the unfavorable cause recognition classifier model is performed as follows:
a. determining unfavorable causes and acquiring test data;
determining unfavorable inducement factors of 6 high-pile wharfs, wherein 1 is ship nonstandard berthing, 2 is bank slope uneven settlement, 3 is local over-limit stacking load, 4 is material shape deterioration, 5 is wind wave flow impact, and 6 is earthquake action; the method comprises the following steps of obtaining 1 pile group distributed strain data under the independent action of three unfavorable inducements of ship nonstandard berthing, 3 local over-limit stacking load and 5 storm flow impact by using an installed optical fiber strain sensor; meanwhile, for the other three unfavorable inducers, establishing an ABAQUS numerical simulation analysis model of the high-pile wharf, and simulating the pile group distributed strain data under the independent action of the unfavorable inducers in a software model; the measured data and the simulation data are used as test data of the unfavorable incentive inversion model;
b. extracting a characteristic vector;
carrying out statistical analysis on the test data and extracting the maximum strain x1Average strain x2As a time domain characteristic parameter; carrying out frequency analysis on the test data, and extracting the first third harmonic frequency x of the signal3、x4、x5As frequency domain characteristic parameters; HHT analysis is carried out on the test data, and the first three-order inherent modal frequency x is extracted6、x7、x8Transient energy x9As a time-frequency domain characteristic parameter; then, a high-dimensional characteristic vector x (x) is extracted from the test data obtained from each distributed optical fiber1,x2,...,x9) Wherein x is1,x2,…,x99 characteristic values of pile group strain distribution data are obtained; since each feature quantity in the high-dimensional feature vector x has a certain correlation and has an excessively high dimension, the principal component y of x is obtained by the PCA algorithm (y is equal to1,y2,...ym)，m<9, using the data as sample data of the later classifier training;
c. establishing a parallel support vector machine classifier model; according to 6 unfavorable causes, six parallel support vector machine classifier models are established, and the 6 unfavorable causes are classified and identified:
SVM1: the classifier distinguishes the ship nonstandard berthing from other 5 unfavorable inducers, when the ship is standard berthed, the output of the SVM1 takes +1, otherwise takes-1, f1(y) is a classification function of SVM1, as shown in equation (1), where a1i(i＝1,2,…n)，b1As a classification function f1(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000041
SVM 2: the classifier distinguishes bank slope non-uniformity settlement from other 5 unfavorable inducers, when bank slope non-uniformity settlement is adopted, the output of the SVM2 is +1, otherwise, the output is-1; f. of2(y) is a classification function of SVM2, as shown in equation (2), where a2i(i＝1,2,…n)，b2As a classification function f2(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000042
SVM 3: the classifier distinguishes local over-limit stowage from other 5 unfavorable inducers, when the local over-limit stowage is adopted, the output of the SVM3 is +1, otherwise, the output is-1; f. of3(y) is a classification function of SVM3, as shown in equation (3), where a3i(i＝1,2,…n)，b3As a classification function f3(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000051
SVM 4: the classifier distinguishes the material property deterioration from other 5 unfavorable causes, when the material property deterioration is detected, the output of the SVM4 is +1, otherwise, the output is-1; f. of4(y) is a classification function of SVM4, as shown in equation (4), where a4i(i＝1,2,…n)，b4As a classification function f4(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000052
SVM 5: the classifier distinguishes the storm surge from other 5 unfavorable inducers, when the storm surge is adopted, the output of the SVM5 takes +1, otherwise takes-1; f. of5(y) is a classification function of SVM5, as shown in equation (5), where a5i(i＝1,2,…n)，b5As a classification function f5(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000053
SVM 6: the classifier distinguishes the earthquake action from other 5 unfavorable inducers, when the earthquake is impacted by wind, wave and current, the output of the SVM6 takes +1, otherwise takes-1; f. of6(y) is a classification function of SVM6, as shown in equation (6), where a6i(i＝1,2,…n)，b6As a classification function f6(y) a coefficient, Φ being a gaussian kernel function;
Figure FDA0002987042550000054
d. training model parameters; inputting the sample data obtained in the step b into 6 parallel support vector machine classifier models, and determining coefficient matrixes a and b of SVM1, SVM2, SVM3, SVM4, SVM5 and SVM6 by using an SVM training algorithm, so as to parameterize the 6 parallel support vector machine classifier models;
Figure FDA0002987042550000055
e. identifying adverse cause action; after the model parameter training is finished, the model can be used for identifying unfavorable incentive effects; the specific process is as follows: the output of the SVM1 is +1, which indicates that the ship does not standardize berthing, otherwise, the output does not exist; the output of the SVM2 is +1, which indicates that bank slope non-uniformity settlement exists, otherwise, the bank slope non-uniformity settlement does not exist; the output of the SVM3 is +1, which indicates that the local overrun heap loading exists, otherwise, the local overrun heap loading does not exist; the output of SVM4 is +1, indicating that there is material property degradation, otherwise it is not present; the output of the SVM5 is +1, which indicates that the storm flow impact exists, otherwise, the storm flow impact does not exist; the output of SVM6 is +1, indicating the presence of seismic action, otherwise it is not.
CN202110302961.8A 2021-03-22 2021-03-22 Method for monitoring foundation pile of high-pile wharf Active CN113074649B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202110302961.8A CN113074649B (en) 2021-03-22 2021-03-22 Method for monitoring foundation pile of high-pile wharf
CN202210679285.0A CN114877820B (en) 2021-03-22 2021-03-22 High pile wharf foundation pile monitoring method based on unfavorable incentive recognition

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110302961.8A CN113074649B (en) 2021-03-22 2021-03-22 Method for monitoring foundation pile of high-pile wharf

Related Child Applications (1)

Application Number Title Priority Date Filing Date
CN202210679285.0A Division CN114877820B (en) 2021-03-22 2021-03-22 High pile wharf foundation pile monitoring method based on unfavorable incentive recognition

Publications (2)

Publication Number Publication Date
CN113074649A true CN113074649A (en) 2021-07-06
CN113074649B CN113074649B (en) 2022-08-23

Family

ID=76613296

Family Applications (2)

Application Number Title Priority Date Filing Date
CN202210679285.0A Active CN114877820B (en) 2021-03-22 2021-03-22 High pile wharf foundation pile monitoring method based on unfavorable incentive recognition
CN202110302961.8A Active CN113074649B (en) 2021-03-22 2021-03-22 Method for monitoring foundation pile of high-pile wharf

Family Applications Before (1)

Application Number Title Priority Date Filing Date
CN202210679285.0A Active CN114877820B (en) 2021-03-22 2021-03-22 High pile wharf foundation pile monitoring method based on unfavorable incentive recognition

Country Status (1)

Country Link
CN (2) CN114877820B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114046743A (en) * 2021-09-24 2022-02-15 浙江大学 Intelligent monitoring system for wharf pile foundation
CN115576243A (en) * 2022-10-08 2023-01-06 中国能源建设集团广东省电力设计研究院有限公司 Ocean erosion dynamic environment three-dimensional monitoring multi-data fusion management system and method

Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005256519A (en) * 2004-03-15 2005-09-22 Takenaka Komuten Co Ltd Axial force measurement method of ground improvement pile
CN1888330A (en) * 2006-07-25 2007-01-03 南京大学 Bored concrete pile foundation distributing optical fiber sensing detecting method and system
CN1900434A (en) * 2006-07-25 2007-01-24 南京大学 Distributive optical fiber detecting method and system for prefabricated pile damage
CN103208007A (en) * 2013-03-19 2013-07-17 湖北微驾技术有限公司 Face recognition method based on support vector machine and genetic algorithm
CN103215975A (en) * 2013-04-19 2013-07-24 浙江华东工程安全技术有限公司 Inbuilt method for distributed type sensing cable in foundation pile
CN203337111U (en) * 2013-07-25 2013-12-11 无锡市政设计研究院有限公司 Hoop type optical fiber grating reinforcement meter
CN103759666A (en) * 2014-01-13 2014-04-30 河海大学 Device and method for monitoring pile body strain of round solid pile
JP2014084598A (en) * 2012-10-22 2014-05-12 Taisei Corp Pile state detection system
CN104367324A (en) * 2013-08-16 2015-02-25 上海微创电生理医疗科技有限公司 Pressure sensor, manufacturing method of pressure sensor, mold and medical catheter
CN105369836A (en) * 2014-08-22 2016-03-02 天津科鉴基础工程检测有限公司 Novel pile foundation detection method
CN105444848A (en) * 2016-01-26 2016-03-30 济南大学 Skip weighing and detecting device and method based on optical fiber sensing
CN106013276A (en) * 2016-07-04 2016-10-12 中国电建集团华东勘测设计研究院有限公司 Stress-strain testing system for large-diameter steel pipe pile of offshore wind turbine and construction method
CN106295692A (en) * 2016-08-05 2017-01-04 北京航空航天大学 Product initial failure root primordium recognition methods based on dimensionality reduction Yu support vector machine
CN107704838A (en) * 2017-10-19 2018-02-16 北京旷视科技有限公司 The attribute recognition approach and device of destination object
CN109338859A (en) * 2018-08-18 2019-02-15 重庆交通大学 The stake of screw-in hanging wire and its mounting device suitable for pan soil construction
CN109837930A (en) * 2018-12-14 2019-06-04 重庆交通大学 Long piled wharf pile foundation based on optical fiber distributed type strain monitoring damages online recognition method
US20190354582A1 (en) * 2018-05-21 2019-11-21 LEVERTON GmbH Post-filtering of named entities with machine learning
CN111561974A (en) * 2020-06-29 2020-08-21 浙江工业大学 Bridge scouring multi-source monitoring system and monitoring method and punching depth evaluation method thereof
CN212296724U (en) * 2020-05-17 2021-01-05 北京华信科创科技有限公司 Auxiliary control integrated acquisition module and auxiliary control platform for offshore wind generating set
CN112195984A (en) * 2020-09-25 2021-01-08 中交投资南京有限公司 Anti-floating anchor rod pile test device and test method

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB201212701D0 (en) * 2012-07-17 2012-08-29 Silixa Ltd Structure monitoring
CN105928646B (en) * 2016-07-15 2018-07-24 重庆交通大学 Suspension cable anchor head performance degradation state monitoring method based on distributed fiber optic sensing
CN106482792A (en) * 2016-11-21 2017-03-08 深圳市道桥维修中心桥梁检测站 Bridge health monitoring system based on Brillouin distributed optical fiber sensing technology
DE102018105905B4 (en) * 2018-03-14 2020-12-31 Bundesrepublik Deutschland, vertreten durch die Bundesministerin für Wirtschaft und Energie, diese vertreten durch den Präsidenten der Bundesanstalt für Materialforschung und-prüfung (BAM) Method for determining a change of a physical parameter with the correct sign and device with an optical fiber
CN109000876B (en) * 2018-04-28 2020-10-20 南京航空航天大学 SNS optical fiber impact identification method based on automatic encoder deep learning
CN111678630B (en) * 2020-06-18 2021-08-06 哈尔滨工业大学（深圳） Steel strand uniaxial stress detection method based on ultrasonic guided wave stress sensitivity analysis
CN112129265A (en) * 2020-09-24 2020-12-25 唐山市神州科贸有限公司 Tank body settlement monitoring device and monitoring method

Patent Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005256519A (en) * 2004-03-15 2005-09-22 Takenaka Komuten Co Ltd Axial force measurement method of ground improvement pile
CN1888330A (en) * 2006-07-25 2007-01-03 南京大学 Bored concrete pile foundation distributing optical fiber sensing detecting method and system
CN1900434A (en) * 2006-07-25 2007-01-24 南京大学 Distributive optical fiber detecting method and system for prefabricated pile damage
JP2014084598A (en) * 2012-10-22 2014-05-12 Taisei Corp Pile state detection system
CN103208007A (en) * 2013-03-19 2013-07-17 湖北微驾技术有限公司 Face recognition method based on support vector machine and genetic algorithm
CN103215975A (en) * 2013-04-19 2013-07-24 浙江华东工程安全技术有限公司 Inbuilt method for distributed type sensing cable in foundation pile
CN203337111U (en) * 2013-07-25 2013-12-11 无锡市政设计研究院有限公司 Hoop type optical fiber grating reinforcement meter
CN104367324A (en) * 2013-08-16 2015-02-25 上海微创电生理医疗科技有限公司 Pressure sensor, manufacturing method of pressure sensor, mold and medical catheter
CN103759666A (en) * 2014-01-13 2014-04-30 河海大学 Device and method for monitoring pile body strain of round solid pile
CN105369836A (en) * 2014-08-22 2016-03-02 天津科鉴基础工程检测有限公司 Novel pile foundation detection method
CN105444848A (en) * 2016-01-26 2016-03-30 济南大学 Skip weighing and detecting device and method based on optical fiber sensing
CN106013276A (en) * 2016-07-04 2016-10-12 中国电建集团华东勘测设计研究院有限公司 Stress-strain testing system for large-diameter steel pipe pile of offshore wind turbine and construction method
CN106295692A (en) * 2016-08-05 2017-01-04 北京航空航天大学 Product initial failure root primordium recognition methods based on dimensionality reduction Yu support vector machine
CN107704838A (en) * 2017-10-19 2018-02-16 北京旷视科技有限公司 The attribute recognition approach and device of destination object
US20190354582A1 (en) * 2018-05-21 2019-11-21 LEVERTON GmbH Post-filtering of named entities with machine learning
CN109338859A (en) * 2018-08-18 2019-02-15 重庆交通大学 The stake of screw-in hanging wire and its mounting device suitable for pan soil construction
CN109837930A (en) * 2018-12-14 2019-06-04 重庆交通大学 Long piled wharf pile foundation based on optical fiber distributed type strain monitoring damages online recognition method
CN212296724U (en) * 2020-05-17 2021-01-05 北京华信科创科技有限公司 Auxiliary control integrated acquisition module and auxiliary control platform for offshore wind generating set
CN111561974A (en) * 2020-06-29 2020-08-21 浙江工业大学 Bridge scouring multi-source monitoring system and monitoring method and punching depth evaluation method thereof
CN112195984A (en) * 2020-09-25 2021-01-08 中交投资南京有限公司 Anti-floating anchor rod pile test device and test method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
NICKY DE BATTISTA 等: "《Distributed fiber optic sensors for monitoring reinforced concrete piles using Brillouin scattering》", 《SIXTH EUROPEAN WORKSHOP ON OPTICAL FIBER SENSORS》, 31 May 2016 (2016-05-31) *
沈国根: "《光纤传感技术在桩基内力监测中的应用研究》", 《山西建筑》, 27 May 2020 (2020-05-27), pages 72 - 73 *
舒岳阶 等: "《FBG应变传感器应力疲劳极限传感寿命评估方法》", 《光子学报》, 3 November 2017 (2017-11-03), pages 106 - 111 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114046743A (en) * 2021-09-24 2022-02-15 浙江大学 Intelligent monitoring system for wharf pile foundation
CN115576243A (en) * 2022-10-08 2023-01-06 中国能源建设集团广东省电力设计研究院有限公司 Ocean erosion dynamic environment three-dimensional monitoring multi-data fusion management system and method
CN115576243B (en) * 2022-10-08 2024-01-09 中国能源建设集团广东省电力设计研究院有限公司 Multi-data fusion management system and method for stereoscopic monitoring of marine dredging power environment

Also Published As

Publication number Publication date
CN114877820B (en) 2023-06-02
CN113074649B (en) 2022-08-23
CN114877820A (en) 2022-08-09

Similar Documents

Publication Publication Date Title
CN113074649B (en) Method for monitoring foundation pile of high-pile wharf
JP5148589B2 (en) A method for evaluating the safety of bridge structures by vibration measurements.
Sha et al. Laboratory tests and numerical simulations of barge impact on circular reinforced concrete piers
CN207244680U (en) A kind of sunk bridge pile foundation washes away real-time monitoring system
Ghorbani et al. Effect of submerged vanes on the scour occurring at a cylindrical pier
CN104614020B (en) The original position whole detection method of long piled wharf horizontal bearing capacity and force model proterties
CN109837930B (en) High-pile wharf pile foundation damage online identification method based on optical fiber distributed strain monitoring
Seo et al. A study on accumulated damage of steel wedges with dead-rise 10 due to slamming loads
CN103810380A (en) Submarine pipeline suspended span security level grading evaluation method and device
Vardanega et al. Assessing the suitability of bridge-scour-monitoring devices
KR101874378B1 (en) Method for Safety Assessment Real-Time of Landing Pier-Type Wharf Structure Based Internet of Things and System thereof
CN110197015B (en) Dam foundation pre-stressed anchor cable effective tensile stress measuring method
Eick Structural health monitoring of Inland navigation infrastructure
CN109799132B (en) Strain test-based method for identifying damage to foundation pile of high-pile wharf
Komatsu et al. Development of offshore wind turbine floater that blends into Japanese waters
CN103981527A (en) Multichannel ocean platform cathode protection parameter monitoring device
CN108256204A (en) A kind of high pile pier structure overall security appraisal procedure based on energy method
Farooq et al. Bridge scour monitoring: challenges and opportunities
CN103632038A (en) Automatic batch checking calculation method for safety of submarine pipeline suspended span sections
Ge The application of the wireless sensor network on port wharf structure health monitoring
Henrique et al. Fast Tool for Structural Monitoring of a Pier After Impact of a Very Large Vessel Using Ambient Vibration Analysis
Hutchinson et al. Measurement of Structure Response to Ferry Berthing Loads
Xu et al. Strain Measurement on Water Intake Coarse Grid of Nuclear Power Plant in Ice-covered Region of China: FBG Sensor
MAKKI et al. REVIEW AND OVERVIEW OF STRUCTURAL HEALTH MONITORING TECHNOLOGY IN BRIDGES
WO2024052259A1 (en) Sensor for detecting scour

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
TA01 Transfer of patent application right

Effective date of registration: 20220803

Address after: 400074 No. 66, Xuefu Avenue, Nan'an District, Chongqing

Applicant after: CHONGQING JIAOTONG University

Applicant after: Chongqing shipping construction development (Group) Co.,Ltd.

Address before: No.66 Xuefu Avenue, Nan'an District, Chongqing 400074

Applicant before: CHONGQING JIAOTONG University

TA01 Transfer of patent application right
GR01 Patent grant
GR01 Patent grant