CN110969370B - Quality risk analysis method for building structural member - Google Patents
Quality risk analysis method for building structural member Download PDFInfo
- Publication number
- CN110969370B CN110969370B CN201911324065.0A CN201911324065A CN110969370B CN 110969370 B CN110969370 B CN 110969370B CN 201911324065 A CN201911324065 A CN 201911324065A CN 110969370 B CN110969370 B CN 110969370B
- Authority
- CN
- China
- Prior art keywords
- data
- ica
- structural member
- dimension
- regression
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000012502 risk assessment Methods 0.000 title claims abstract description 12
- 238000012880 independent component analysis Methods 0.000 claims abstract description 55
- 238000001514 detection method Methods 0.000 claims abstract description 19
- 238000000342 Monte Carlo simulation Methods 0.000 claims abstract description 18
- 238000004364 calculation method Methods 0.000 claims abstract description 17
- 230000009467 reduction Effects 0.000 claims abstract description 17
- 238000012544 monitoring process Methods 0.000 claims abstract description 14
- 238000000611 regression analysis Methods 0.000 claims abstract description 11
- 238000011156 evaluation Methods 0.000 claims abstract description 9
- 230000001419 dependent effect Effects 0.000 claims description 12
- 238000002790 cross-validation Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 11
- 238000006073 displacement reaction Methods 0.000 claims description 9
- 230000003044 adaptive effect Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 6
- 229910000831 Steel Inorganic materials 0.000 claims description 4
- 239000010959 steel Substances 0.000 claims description 4
- 238000011835 investigation Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 238000006386 neutralization reaction Methods 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 238000000926 separation method Methods 0.000 abstract description 3
- 230000007547 defect Effects 0.000 description 11
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000007667 floating Methods 0.000 description 7
- 238000010276 construction Methods 0.000 description 5
- 239000011435 rock Substances 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000003908 quality control method Methods 0.000 description 3
- 238000004445 quantitative analysis Methods 0.000 description 3
- 238000012954 risk control Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 239000002689 soil Substances 0.000 description 3
- 238000013178 mathematical model Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004321 preservation Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000009933 burial Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000007586 pull-out test Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000004062 sedimentation Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2134—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on separation criteria, e.g. independent component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
A method of quality risk analysis of a building structural member, comprising: storing engineering detection and monitoring data of the structural member into a computer readable format; performing dimension reduction trial calculation on the self-variable data by adopting an ICA (independent component analysis) method; performing regression analysis on m-1 data sets by adopting a multi-element self-adaptive regression MARS; selecting an optimal data set subjected to ICA dimension reduction and ICA dimension reduction n corresponding to the data set according to the evaluation result; regression of the multiple self-adaptive regression MARS is used for obtaining a predicted final model of the resistance of the structural member; based on the predicted final model R and load; and simulating the model by adopting a Monte Carlo method, so as to obtain the override probability or the failure probability. The invention gives a risk analysis prediction lower conclusion taking failure probability or overrun probability as output, which is beneficial to the establishment of the separation between the financial insurance industry and the engineering industry.
Description
Technical Field
The invention relates to the field of buildings, in particular to a quality risk analysis method for building structural members.
Background
The existing engineering quality control method generally judges whether an engineering can be accepted or not and meets the standard requirement of a qualified line through engineering detection, but cannot provide a quantitative method which is based on performance and gives probability. The traditional engineering acceptance and qualification method is insufficient for engineering quality defect insurance risk control mechanisms which are vigorously implemented in China in recent years (failure probability or overrun probability cannot be predicted generally), but the engineering quality probability prediction method based on big data is a key capable of achieving insurance risk control, rate definition and even accurate calculation.
Construction engineering quality potential defect insurance (Inherent defects insurance, IDI) is insurance applied by a construction unit, and an insurance company performs compensation obligations for the occurrence of insurance building damage due to engineering quality potential defects within an insurance range and an insurance period according to insurance clause agreements. IDI belongs to the coming world of the european union, and has long been studied and practiced in countries such as finland, france, italy and spain where IDI is developed as a mandatory insurance. The Beijing, shanghai and Shenzhen all put forward the implementation rule time line of engineering quality defect insurance (IDI), such as basic underwriting scope of IDI as foundation and main structural engineering (10 years), heat preservation and waterproof engineering (5 years) specified in Beijing city residential engineering quality potential defect insurance temporary management method put forward in 2019 Beijing:
foundation and major structural engineering defects include: 1. wholly or partially collapsing; 2. the foundation generates uneven settlement beyond the allowable design specification; 3. cracks, deformations, damages and breaks affecting the structural safety appear at the base and main structure parts; 4. the overhanging components such as a balcony, a rain fly, a cornice, an air conditioner plate and the like are cracked, deformed, damaged and broken which influence the use safety; 5. quality defects that the falling off, collapse and the like of the outer wall surface affect the use safety; 6. other foundation foundations and main body structure parts have potential defects of engineering quality which affect structural safety.
The defects of the heat-resistant and waterproof engineering include: 1. the heat preservation layer of the enclosure structure is damaged and falls off; 2. waterproof leakage between the underground, roofing and toilet bath room; 3. leakage of the outer wall (comprising the junction of the outer window and the outer wall); 4. other parts with waterproof requirements leak.
The quality of engineering quality can be evaluated by qualitative or quantitative methods. In the category of the quantitative method, the probability method based on Monte Carlo simulation can intuitively give the failure probability or the overrun probability of the building structural member, and the probability level directly reflects the quality of the building structural member, but the application of the method requires the knowledge of the limit state equation of the structural member, and the mathematical limit state equation Z can be expressed by resistance R and load Q as follows:
Z＝g(R，Q)＝R-Q
when (1) Z is more than 0, the building structural member meets the functional requirement; (2) Z is less than 0, and the building structural member fails or the load level Q exceeds the preset resistance performance index R warning value; (3) z=0, the building structural member is in a critical state. It should be noted that the resistance R and the load Q herein include not only force, stress, and other variables, but also deformation, sedimentation, seepage, and other variables of interest in the construction engineering.
In practice, however, the limit state equations for many building structural members are difficult to obtain, particularly those for underground structural members, because uncertainty in the nature of the ground-based rock-soil mass and uncertainty in the boundary conditions lead to an undetermined mathematical and mechanical model of the structural member; some of the upper structural member limit state equations are available but require large numerical simulation calculation costs (economic costs) due to excessively complex actual boundary conditions, and economic benefits are not taken into consideration in actual engineering.
The invention provides engineering quality process control and acceptance means which are legal for engineering detection or monitoring of building structural members in the field of construction engineering in China, and provides a method for establishing a limit state equation of a corresponding building structural member by using engineering detection or monitoring data through big data and machine learning. Whether the existing building structural member has a known limit state equation or not, the limit state equation based on detection or monitoring data can be established through the method provided by the invention, and the failure probability or the overrun probability of the building structural member can be further obtained through Monte Carlo simulation calculation.
In the prior art, the risk analysis with probability output mainly uses the Monte Carlo method in the construction field and is concentrated in bridge and tunnel engineering, such as the method for establishing the limit state equation by combining the finite element model and the response surface based on polynomial expansion, which is proposed in Chinese patent No. 107844651A, and the main difference between the method and the method is how to establish the limit state equation, because the establishment and the solution of the response surface based on finite element and polynomial expansion need definite boundary conditions, have higher calculation cost and difficult variable value, and are difficult to apply in practical engineering.
Disclosure of Invention
The invention provides a quality risk analysis method for building structural members, which aims to solve at least one technical problem.
To solve the above-described problems, as one aspect of the present invention, there is provided a quality risk analysis method of a building structural member, comprising:
step one, engineering detection and monitoring data of a structural member are stored into a computer-readable format, wherein the data comprise a plurality of variables, one variable is a dependent variable, the other variables are independent variables, and the dependent variable is the resistance of the structural member;
performing dimension reduction trial calculation on the self-variable data by adopting an ICA (independent component analysis) method, wherein the reduced dimension n is less than or equal to the dimension m of the input data, and n is traversed from 1 dimension to m-1 dimension, so that m-1 data sets are obtained;
performing regression analysis on the m-1 data sets by adopting a multi-element adaptive regression MARS, evaluating the quality of regression analysis results by adopting a cross-validation method, and respectively giving out evaluation result indexes of mean square error MSE, GCV and R for the m-1 data sets 2 The following formula is shown:
wherein y is i Is the ith factor changePredicted value of quantity, Y i D=3 for the measured value of the ith dependent variable, m is the number of basis functions in the multivariate adaptive regression MARS, N is the number of data,
average value of the variables;
selecting an optimal data set subjected to ICA dimension reduction and ICA dimension reduction n corresponding to the data set according to the evaluation result, and selecting the MSE with the lowest standard, the GCV with the lowest standard and the R2 with the highest standard or any combination of the three;
step five, obtaining a predicted final model R of the resistance of the structural member based on the regression of the ICA dimension n selected in the step four and the dimension-reduced total data set corresponding to the ICA dimension n by using a multi-element adaptive regression MARS;
step six, based on the final model R and load Q of the prediction of the resistance of the structural member given in step five, a limit state equation z=r-Q is given, wherein Q may be a limit value given in accordance with a national, industry or local standard, or may be a trigger value in insurance clauses, or a value set by a user;
and step seven, simulating the model by adopting a Monte Carlo method, so as to obtain the overrunning probability or failure probability of the structural member.
Preferably, the structural members include beams, columns, plates, piles, anchors, and the like.
Preferably, the engineering detection and monitoring data comprises: load-displacement data or stress-strain data of field test, related data (physical and mechanical parameters of a rock and soil layer) in a geotechnical engineering investigation report, and material test data of concrete, steel bars and the like.
Preferably, the criterion for the neutralization of step four is dependent on the purpose of the person using the data, and it is desirable that these three indicators point to the same ICA-reduced data set, with MSE applied if absolute error priority is not met, and R applied if capture priority of data law is considered 2 GCV should be used if the main goal of reducing overfitting is considered.
The invention has the advantages that:
1) The invention is purely data driven and is suitable for various building structural members.
2) The engineering detection and data based on the invention are usually from quality control projects which are legal to be completed in engineering process and acceptance links, and the method can also be used by independently and additionally collecting detection or monitoring data.
3) The invention gives a risk analysis prediction lower conclusion taking failure probability or overrun probability as output, which is beneficial to the establishment of the separation between the financial insurance industry and the engineering industry.
Drawings
FIG. 1 schematically illustrates a flow chart of the present invention;
FIG. 2 schematically illustrates a schematic view of an anti-floating anchor pullout test apparatus;
FIG. 3 schematically illustrates a modeling and model verification process for an anchor load-displacement predictive model;
FIG. 4 schematically illustrates the behavior of MSE and GCV at different ICA dimensions;
FIG. 5 schematically shows R 2 Performance in different ICA dimensions;
FIG. 6 schematically illustrates a comparison of measured deformation values of an anchor rod with predicted values of an ICA-MARS model;
fig. 7 schematically shows an anchor load-displacement curve: comparison of measured values with predicted values of ICA-MARS model.
Detailed Description
Embodiments of the invention are described in detail below with reference to the attached drawings, but the invention can be implemented in a number of different ways, which are defined and covered by the claims.
The invention provides a quality risk analysis method of a building structural member, which is essentially a mathematical model for evaluating the quality of the building structural member based on building engineering detection and monitoring data, wherein the mathematical model comprises key technologies, specific implementation algorithms and steps of autonomous innovation of the invention. The input of the model is engineering detection and monitoring data, the output of the model is (override or failure) probability, and the model is purely data-driven, i.e. the model is not changed for different building structural member types.
The purpose of the invention is that: the method is directly based on the legal engineering detection or monitoring data in engineering, and the limit state equation and the equation variable with mutually independent statistics of the building structural member are obtained by adopting the autonomous innovative machine learning method. On the basis, a general Monte Carlo simulation method is adopted to obtain the failure probability or the overrun probability of the building structural member, and visual basis is provided for engineering insurance risk control.
The method based on the combination of machine learning and Monte Carlo simulation is ICA-MARS-MC, wherein ICA represents an independent principal component analysis algorithm, MARS refers to a multi-element self-adaptive regression spline, MC refers to Monte Carlo simulation, and ICA-MARS-MC forms three algorithms into an algorithm chain to make up for the defects of each other and improve the prediction accuracy. The specific steps of ICA-MARS-MC can be described as follows:
step one: engineering detection, monitoring data such as load-displacement data or stress-strain data of field tests of structural members such as beams, columns, plates, piles, anchors and the like, related data (physical and mechanical parameters of a rock soil layer), concrete, steel bars and other material test data in a geotechnical engineering investigation report are stored into a computer-readable format in an automatic or manual mode, such as an excel electronic table, a csv format and the like, and usually the data have multiple dimensions (variables), wherein one variable is resistance of the structural member, the variable is taken as a dependent variable (a predicted object) here, and other variables are independent variables;
step two: performing dimension reduction trial calculation on the self-variable data input in the step one by adopting an independent component analysis ICA method, wherein the reduced dimension n is less than or equal to the dimension m of the input data, and n is traversed from 1 dimension to m-1 dimension, so that m-1 data sets are obtained;
step three: regression analysis is carried out on the m-1 data sets obtained in the second step by adopting a multi-element self-adaptive regression MARS, the quality of regression analysis results is evaluated by adopting a cross-validation method, and evaluation result indexes are respectively given to the m-1 data setsFor mean square error MSE, GCV, and R 2 The following formula is shown:
in which y i Is the predicted value of the ith dependent variable, Y i D=3 for the measured value of the ith dependent variable, m is the number of basis functions in the multivariate adaptive regression MARS, N is the number of data,
is the average of the variables.
The mean square error MSE is the mean value of the sum of squares of errors of corresponding points of the predicted data and the original data; the generalized cross validation mean square error GCV is based on MSE as a numerator, and describes the complexity of the model in a denominator, so that indexes of generalization capability and precision of the model are comprehensively considered; determining the coefficient R 2 Is an index of a common evaluation model, and the molecular part represents the sum of the square error of a true value and a predicted value, which is similar to the mean square error MSE; the denominator part represents the sum of the squared differences of the true and mean values, R 2 The value range is [0,1 ]]The closer to 1 the general case is, the better the model effect is
Step four: according to the evaluation result in the third step, selecting an optimal data set subjected to ICA dimension reduction and ICA dimension n corresponding to the data set, and selecting MSE with the lowest standard, GCV with the lowest standard, R2 with the highest standard or any combination of the three, wherein the three indexes point to the same data set subjected to ICA dimension reduction according to the purpose of using a data person, if the three indexes cannot meet the requirement of applying MSE when taking absolute error priority, if taking capture priority of a data rule into consideration, R2 is applied when taking reduction of fitting as a main target, and GCV is used.
Step five: obtaining a predicted final model R of the resistance of the structural member based on regression of the ICA dimension n selected in the fourth step and all data sets of the dimension n corresponding thereto (note that all are not the cross validation method of the third step) used for the multi-element adaptive regression MARS;
step six: based on the final model R and the load Q of the prediction of the resistance of the structural member given in the step five, a limit state equation z=r-Q is given, wherein Q may be a limit value given in accordance with a national, industry or local standard, a trigger value in insurance clauses, or a value self-formulated by a user;
step seven: the model is simulated by using the monte carlo method, so as to obtain the overrun probability or failure probability of the structural member (i.e. statistics Z <0 ratio in monte carlo simulation=overrun probability or failure probability of the structural member).
The method based on the combination of machine learning and Monte Carlo simulation is beneficial in that:
(1) The ICA is used as the data preprocessing, so that the effect of data dimension reduction is realized, and the prediction accuracy can be effectively improved by obtaining the independent hidden dimension ratio through ICA analysis and performing regression analysis independently;
(2) The variables obtained by ICA transformation are mutually independent, which provides basis for random sampling to generate mutually independent samples when the MC method is used, and avoids uncertainty of an original data set on correlation coefficient matrix simulation caused by sample sparseness;
(3) The explicit prediction model provided by the MARS is convenient for engineers to understand more intuitively, and compared with numerical calculation of a finite element method and the like, the calculation cost is greatly reduced, and the possibility of increasing the sample number obviously when MC is used subsequently to improve the calculation accuracy is provided.
The key and innovation of the invention is that:
1) The first step of the invention is to use independent component analysis ICA algorithm to reduce the dimension, and obtain a group of dimension-reduced variables with statistics independent of each other.
2) The invention adopts cross validation after MARS regression analysis to select the dimension of ICA algorithm dimension reduction.
3) The building structure member resistance prediction model R is established through MARS regression analysis after the dimension of ICA dimension reduction is determined through cross verification, namely the input of the MARS regression analysis is the output of ICA dimension reduction.
3) The invention is based on the use of the first ICA, and random sampling in the last MC monte carlo simulation step does not take into account the correlation between the variables, which are statistically uncorrelated after ICA transformation.
The invention has the advantages that:
1) The invention is purely data driven and is suitable for various building structural members.
2) The engineering detection and data based on the invention are usually from quality control projects which are legal to be completed in engineering process and acceptance links, and the method can also be used by independently and additionally collecting detection or monitoring data.
3) The invention gives a risk analysis prediction lower conclusion taking failure probability or overrun probability as output, which is beneficial to the establishment of the separation between the financial insurance industry and the engineering industry.
Examples: case for predicting deformation and evaluating failure probability of engineering anti-floating anchor rod
The embodiment is a sampling drawing test of the anti-floating anchor engineering in the engineering detection acceptance stage of the typical building underground structural member anti-floating anchor. The test adopts a core-penetrating jack and an anchor as drawing devices, and the counterforce is transmitted to the ground through the steel plate. The hydraulic jack is adopted for grading loading and unloading during the test, the loading capacity of each stage is controlled by a pressure gauge on the manual oil pump, and the displacement sensor is used for measuring the displacement of the anchor head when the anchor head is pulled out during the loading.
1. Engineering overview
The project is located in Shenzhen Longhua region with a floor area of about 54607.07m 2 The bottom is provided with 5 layers of business skirt buildings, and the upper part is provided with 6 super high-rise towers (1 office and 5 business buildings); wherein the office building (44F) is 250m, the business building is provided with 4 seats (48+4F) 165m and 1 seat (29+4F) 109.9 mThe underground is 4 layers, and the burial depth is about 18m. Building structural form: the main building adopts a shear wall structure; the skirt house adopts a frame structure and the basement adopts a plate column structure. The engineering foundation design grade is grade A, and the building foundation pile design grade is grade A. The design basic form is as follows: pile foundations, pier foundations, natural rock foundations and anti-floating anchor rods. For the anti-floating anchor rod, the rock anchor rod is required, the rock entering depth is not less than 3.5m, the hole forming diameter is 180mm, 3 reinforcing bars are 32mm HRB400, and the characteristic value of the anti-pulling bearing capacity is 450kN.
TABLE 1 engineering natural foundation mechanical parameter suggestion table
2. ICA-MARS-MC model building
Determination of ICA dimension and model verification
Drawing tests are carried out on 87 anti-floating anchors for engineering detection acceptance to obtain 783 groups of test data, and according to a data dividing method, modeling and model verification processes shown in figure 3 are carried out. Where ICA and MARS are both purely data driven algorithms, but the dimensions of the ICA algorithm output are typically determined by artificially set criteria, how does it determine the dimensions of the ICA output in the project?
The project proposes to select the optimal ICA dimension by adopting a cross-validation method, namely, three model evaluation indexes of MSE, GCV and R2 of the 5-fold cross-validation average under different ICA dimensions are calculated by a 5-fold cross-validation method, so that the ICA dimension is selected. As shown in FIGS. 4 and 5, when ICA dimension is 11, the average value of MSE and GCV given by the 5-fold cross validation model is minimum, R 2 At the highest, comparing the cross-validated average results for such trial-calculations that control only one variable of the ICA dimension can therefore clearly indicate that the model is optimal when the ICA dimension is 11.
While determining ICA dimensions, the 5-fold cross-validation method validated the model, resulting in mse=1.96, gcv=2.89, r 2 =0.92, this precision should be acceptable for practical geotechnical engineering problemsTo accept.
2. Predictive model
Through the above-described cross-validated model procedure, the model of ICA-MARS (ICA dimension=11) can be validated as the model used for final prediction, so all known data will be used for training of the model to obtain the parameters of the model to build the best-precision predictive model. The prediction model whose final prediction model is displayed as ICA (i=0 to 10) given by ICA variables of 11 ICA variables is shown in table 2, and the evaluation index calculation result of the model is as follows: mse=1.8159, gcv=2.6202, r2= 0.9233. Fig. 6 and fig. 7 show a comparison of the measured deformation value of the anchor rod and the load-displacement curve, respectively, and it can be found that the model of ICA-MARS (ICA dimension=11) gives better effects from the prediction of the deformation value and the simulation of the load-displacement curve morphology.
TABLE 2 ICA-MARS prediction model
| Basis function BF | Coefficient beta m | Basis function BF | Coefficient beta m |
| Intercept of (intercept of) | -164.95 | h(-0.11089-ICA3) | 362.45 |
| h(ICA10-0.0105835) | -3600.26 | h(ICA7+0.0742366) | 69.1896 |
| h(0.0105835-ICA10) | 4288.52 | h(ICA7+0.219545) | 162.601 |
| h(ICA2-0.0175126) | -469.376 | h(-0.219545-ICA7) | -319.813 |
| h(0.0175126-V2) | 410.445 | h(0.0379017-ICA10) | -188.226 |
| h(0.068245-ICA10) | -1029.26 | h(0.220921-ICA2) | 182.498 |
| h(ICA10-0.051618) | 486.973 | h(0.0727203-ICA10) | 442.368 |
| h(-0.0289058-ICA4) | -113.691 | h(ICA0-0.075522) | -77.0152 |
| h(0.0603567-ICA4) | 80.2604 | h(ICA4-0.0807966) | 97.4059 |
| h(0.0445167-ICA0) | 174.393 | h(-0.22519-ICA6) | 77.3032 |
| h(ICA5+0.0901701) | 14621.9 | h(ICA5+0.0208572) | 225.696 |
| h(-0.0901701-ICA5) | -1556.77 | h(ICA5-0.0086334) | -131.23 |
| h(ICA5+0.0922799) | -14644.2 | h(ICA5-0.0580034) | 45.5171 |
| h(ICA10+0.0154523) | -513.484 | h(ICA9+0.151809) | -349.425 |
| h(-0.0788279-ICA0) | -155.33 | h(ICA10-0.119567) | -172.238 |
| h(0.0247939-ICA7) | 126.578 |
3. Monte Carlo method for calculating failure probability
The use of the previous ICA algorithm greatly facilitates the sampling process when the monte carlo method is applied, because the 11 variables output by the ICA are all statistically independent from each other, so that the relation between the different variables, i.e., the correlation coefficient, is not considered to be 0 when generating random numbers in the monte carlo simulation, which technically solves an important problem: when the sample size is small and the data distribution is sparse, a large error is possible to exist in the real correlation coefficient or covariance matrix between the fitted variables from the samples, and the error can be avoided when random numbers are generated in the Monte Carlo simulation if the variables are statistically independent, namely, the correlation coefficient is 0.
In the project, a random number is generated by adopting a basic function numpy.random () of Python, and the random sample number adopted in each calculation is tens of millions of times due to the high efficiency of processing an ICA-MARS explicit model by a computer, namely table 4.2, and the obtained calculation result is shown in table 3, and the variation coefficient corresponding to the overrun probability is extremely less than 0.05%. The calculation time of the average 32s also meets the efficiency requirements in actual engineering.
TABLE 3 Monte Carlo simulation calculations
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (2)
1. A method of quality risk analysis of a building structural member, comprising:
step one, engineering detection and monitoring data of a structural member are stored into a computer-readable format, wherein the data comprise a plurality of variables, one variable is a dependent variable, the other variables are independent variables, and the dependent variable is the resistance of the structural member; the structural members include beams, columns, plates, piles and anchors; the engineering detection and monitoring data comprise: load-displacement data or stress-strain data of field test, related data in geotechnical engineering investigation report, concrete and steel bar material test data;
performing dimension reduction trial calculation on the self-variable data by adopting an ICA (independent component analysis) method, wherein the reduced dimension n is less than or equal to the dimension m of the input data, and n is traversed from 1 dimension to m-1 dimension, so that m-1 data sets are obtained;
performing regression analysis on the m-1 data sets by adopting a multi-element adaptive regression MARS, evaluating the quality of regression analysis results by adopting a cross-validation method, and respectively giving out evaluation result indexes of mean square error MSE, GCV and R for the m-1 data sets 2 The following formula is shown:
wherein y is i Is the predicted value of the ith dependent variable, Y i D=3 for the measured value of the ith dependent variable, m is the number of basis functions in the multivariate adaptive regression MARS, N is the number of data,
average value of the variables;
selecting an optimal data set subjected to ICA dimension reduction and ICA dimension reduction n corresponding to the data set according to the evaluation result, and selecting the MSE with the lowest standard, the GCV with the lowest standard and the R2 with the highest standard or any combination of the three;
step five, obtaining a predicted final model R of the resistance of the structural member based on the regression of the ICA dimension n selected in the step four and the dimension-reduced total data set corresponding to the ICA dimension n by using a multi-element adaptive regression MARS;
step six, based on the final model R and load Q of the prediction of the resistance of the structural member given in step five, giving a limit state equation z=r-Q, wherein Q is a limit value given in a national, industry or local standard, or a trigger value in insurance clauses, or a value self-formulated by a user;
and step seven, simulating the model by adopting a Monte Carlo method, so as to obtain the overrunning probability or failure probability of the structural member.
2. The method according to claim 1, wherein the neutralization selection criteria in step four is dependent on the purpose of the person using the data, and the ideal condition is that the three metrics point to the same ICA-reduced data set, the MSE is applied if absolute error priority is not satisfied, and the R is applied if capture priority of data law is considered 2 GCV should be used if the main goal of reducing overfitting is considered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911324065.0A CN110969370B (en) | 2020-01-14 | 2020-01-14 | Quality risk analysis method for building structural member |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911324065.0A CN110969370B (en) | 2020-01-14 | 2020-01-14 | Quality risk analysis method for building structural member |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110969370A CN110969370A (en) | 2020-04-07 |
| CN110969370B true CN110969370B (en) | 2023-05-02 |
Family
ID=70035412
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911324065.0A Active CN110969370B (en) | 2020-01-14 | 2020-01-14 | Quality risk analysis method for building structural member |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110969370B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113408927B (en) * | 2021-06-30 | 2023-01-10 | 四川交达预应力工程检测科技有限公司 | Big data-based prestressed construction quality evaluation method and system |
| CN113592090B (en) * | 2021-09-27 | 2021-12-24 | 中建八局第三建设有限公司 | Building quality prediction method and device based on deep learning and storage medium |
| CN113962407A (en) * | 2021-10-19 | 2022-01-21 | 河南工业大学 | Multidimensional feature analysis method for monitoring data of active civil buildings |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105517019A (en) * | 2015-12-17 | 2016-04-20 | 南京华苏科技股份有限公司 | Method for detecting LTE (Long Term Evolution) network performance by using integrated regression system |
| CN106407498A (en) * | 2016-08-03 | 2017-02-15 | 无锡金世纪国民体质与健康研究有限公司 | Falling risk model establishment method |
| CN107451101A (en) * | 2017-07-21 | 2017-12-08 | 江南大学 | It is a kind of to be layered integrated Gaussian process recurrence soft-measuring modeling method |
| CN110110981A (en) * | 2019-04-26 | 2019-08-09 | 重庆第二师范学院 | A kind of credit rating Default Probability estimates and method for prewarning risk |
| CN110147321A (en) * | 2019-04-19 | 2019-08-20 | 北京航空航天大学 | A kind of recognition methods of the defect high risk module based on software network |
-
2020
- 2020-01-14 CN CN201911324065.0A patent/CN110969370B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105517019A (en) * | 2015-12-17 | 2016-04-20 | 南京华苏科技股份有限公司 | Method for detecting LTE (Long Term Evolution) network performance by using integrated regression system |
| CN106407498A (en) * | 2016-08-03 | 2017-02-15 | 无锡金世纪国民体质与健康研究有限公司 | Falling risk model establishment method |
| CN107451101A (en) * | 2017-07-21 | 2017-12-08 | 江南大学 | It is a kind of to be layered integrated Gaussian process recurrence soft-measuring modeling method |
| CN110147321A (en) * | 2019-04-19 | 2019-08-20 | 北京航空航天大学 | A kind of recognition methods of the defect high risk module based on software network |
| CN110110981A (en) * | 2019-04-26 | 2019-08-09 | 重庆第二师范学院 | A kind of credit rating Default Probability estimates and method for prewarning risk |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110969370A (en) | 2020-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110969370B (en) | Quality risk analysis method for building structural member | |
| Cavaleri et al. | Cyclic response of masonry infilled RC frames: Experimental results and simplified modeling | |
| CN104989456B (en) | A kind of Large Span Underground engineering excavation surrounding rock stability monitoring and pre-alarming method | |
| CN110889588A (en) | Method for evaluating risk level of shield tunnel construction adjacent building by using factor judgment matrix | |
| CN107515025B (en) | Ancient city wall disease grade evaluation method | |
| CN110991918B (en) | Safety risk assessment method in existing building construction project | |
| Uva et al. | Effects in conventional nonlinear static analysis: Evaluation of control node position | |
| Porco et al. | Seismic assessment of irregular existing building: Appraisal of the influence of compressive strength variation by means of nonlinear conventional and multimodal static analysis | |
| Saleemuddin et al. | Seismic damage assessment of reinforced concrete structure using non-linear static analyses | |
| Arrayago et al. | System-based reliability analysis of stainless steel frames under gravity loads | |
| CN116738556B (en) | Modeling method, device and equipment for construction wall and storage medium | |
| Mata et al. | Correlation between compressive strength of concrete and ultrasonic pulse velocity: A case of study and a new correlation method | |
| Liu et al. | Statistical calibration of federal highway administration simplified models for facing tensile forces of soil nail walls | |
| CN115526483A (en) | Rapid evaluation method and system for anti-seismic toughness of shield tunnel structure | |
| Miano et al. | Probabilistic seismic and fire assessment of an existing reinforced concrete building and retrofit design | |
| Angjeliu et al. | Rapid assessment of seismic vulnerability of historic masonry structures through fragility curves approach and national database data | |
| Tanner et al. | Semi-probabilistic models for the assessment of existing concrete structures | |
| Ferlito et al. | Assessment of a vulnerability model against post-earthquake damage data: The case study of the historic city centre of L’Aquila in Italy | |
| Afshari et al. | Simplified time-dependent column shortening analysis in special reinforced concrete moment frames | |
| CN111582634A (en) | Multi-factor safety grading method and system for underground large-space construction | |
| KR100602832B1 (en) | Amentment method of differential shortenint of vertical member in reinforced-concrete structures | |
| CN115795599A (en) | Civil defense engineering structure earthquake vulnerability assessment method | |
| Opabola et al. | Repairability-based design of concrete structures | |
| D’Amato et al. | Risk analysis of existing building heritage through damage assessment after L’Aquila earthquake 2009 | |
| Cao | [Retracted] Life Prediction Model of Wooden Structure Based on Artificial Intelligence Algorithm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |