CN103912026B - A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method - Google Patents

A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method Download PDF

Info

Publication number
CN103912026B
CN103912026B CN201410104939.2A CN201410104939A CN103912026B CN 103912026 B CN103912026 B CN 103912026B CN 201410104939 A CN201410104939 A CN 201410104939A CN 103912026 B CN103912026 B CN 103912026B
Authority
CN
China
Prior art keywords
mechanical index
double lumen
neural network
lumen intubation
soil
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
Application number
CN201410104939.2A
Other languages
Chinese (zh)
Other versions
CN103912026A (en
Inventor
蔡清
程江涛
万凯军
于沉香
陈定安
黄静
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.)
Wuhan Surveying Geotechnical Research Institute Co Ltd of MCC
Original Assignee
Wuhan Surveying Geotechnical Research Institute Co Ltd of MCC
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 Wuhan Surveying Geotechnical Research Institute Co Ltd of MCC filed Critical Wuhan Surveying Geotechnical Research Institute Co Ltd of MCC
Priority to CN201410104939.2A priority Critical patent/CN103912026B/en
Publication of CN103912026A publication Critical patent/CN103912026A/en
Application granted granted Critical
Publication of CN103912026B publication Critical patent/CN103912026B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method, and its step is：1., the collection of the mechanical index of Double lumen intubation data and soil and arrangement；2., set up Double lumen intubation mechanical index and determine BP neural network forecast model；3. BP neural network forecast model is trained, to be determined to Double lumen intubation mechanical index；4., determine that BP neural network forecast model is predicted to soil layer mechanical index with the Double lumen intubation mechanical index for having completed training.Its advantage is：To determine that soil layer mechanical index provides reliable theoretical method by static sounding data；It is that mechanical index determines that the reliability of BP neural network forecast model provides strong support, it is ensured that the accuracy for predicting the outcome using a large amount of Double lumen intubation data and great soil group mechanical index as training sample；Probing hole count greatly reduces during survey can be made, so as to shorten quality, saves prospecting cost, and can improve the quality of engineering investigation.

Description

A kind of mechanical index of Double lumen intubation data based on BP neural network determines Method
Technical field
It is specifically a kind of based on BP neural network the present invention relates to rock-soil engineering in-situ test Applied research fields The mechanical index of Double lumen intubation data determines method.
Background technology
Traditional prospecting mode is arranged based on drilling using static sounding as an auxiliary for improving investigation accuracy Apply, and new prospecting mode then emphasize prospecting should be drilled based on static sounding and targetedly, laboratory test and Other in-situ tests, it attaches equal importance to probing, the effect of laboratory test, and the transformation of this prospecting mode makes probing hole count subtract significantly It is few, quality is shortened, prospecting cost is saved, and improve the quality of prospecting.Replaced using static sounding measuring technology Traditional prospecting mode is, it is necessary to carry out deep excavation and sufficiently utilization to static sounding data.
At present, domestic and foreign scholars have been established largely by the statistical analysis between static sounding parameter and soil layer mechanical index Regression equation or empirical equation, but these regression equations by soil nature Classification and Identification, formed employed in environment and regression analysis The reliability effect of static sounding parameter, is difficult to extensive use in engineering practice；Existing some local regulations and engineering geology Handbook mainly lists the experience value relation of single bridge probe data and foundation bearing capacity, has no specific theoretical and side Method, the scope of application is not also extensive.Therefore, how a kind of mechanical index based on Double lumen intubation data is provided and determines method, Set up the forecast model of mechanical index, it has also become this area technical problem urgently to be resolved hurrily.
The content of the invention
The purpose of the present invention is using existing Double lumen intubation data and foundation bearing capacity, shearing strength, compression mould Three mechanical index of amount, using BP neural network algorithm, develop a kind of Double lumen intubation data based on BP neural network Mechanical index determines method.
A kind of mechanical index of the Double lumen intubation data based on BP neural network of the present invention determines method, its specific skill Art method is implemented according to the following steps：
1., the collection of the mechanical index of Double lumen intubation data and soil and arrangement：By compiling sitework ground Matter exploration report, soil test achievement data etc., collect the mechanical index of statistics Double lumen intubation data and soil；
Described Double lumen intubation data include static point resistance qcWith side friction power fs；Described native mechanical index Including foundation bearing capacity f0, shearing strength（Cohesive strength c, internal friction angle）, Modulus of pressure Es。
2., set up Double lumen intubation mechanical index and determine BP neural network forecast model：Based on BP neural network algorithm, Forecast model is set up using step data message 1.；
Described Double lumen intubation mechanical index determine BP neural network model using 1 input layer, 1 output layer and 1 network structure of hidden layer, specifically sets up process and implements in the following manner：
By along the n-th -2, n-1 of depth, n, n+1, n+2 static point resistance qcValue and n-2, n-1, n, n+1, n+2 side are rubbed Resistance fsUsed as input layer, network model input layer is made up of value 10 n dimensional vector ns：
Using the mechanical index of soil at n-th input value correspondence depth point as output layer, network model output layer is 4 dimension arrows Amount：
Y=[Dn,Cn,Jn,Mn,]
DnRepresent n-th foundation bearing capacity of input value correspondence depth, CnRepresent that n-th input value corresponds to depth Cohesive strength, JnRepresent that n-th input value corresponds to depthMnRepresent n-th compression mould of input value correspondence depth Amount.
3. BP neural network forecast model is trained, to be determined to Double lumen intubation mechanical index：It is soft with matlab Part, setting network training type function, output layer activation primitive type, maximum iteration epochs, anticipation error minimum value The learning efficiency lr of setting value goal and amendment weights, is trained to BP neural network model；
Described network training type function uses traincgf functions；
Described output layer activation primitive type uses purelin functions；
Described maximum iteration epochs is set as 1000 times；
Described anticipation error minimum value setting value goal is set as 0.01；
The learning efficiency lr of described amendment weights is set as 0.05.
4., with step 3. in completed training Double lumen intubation mechanical index determine BP neural network predict mould Type, is predicted by soil layer Double lumen intubation statistical average to soil layer mechanical index；
Described soil layer Double lumen intubation statistical average is all static point resistance q in each layer soilcWith side friction power fs Arithmetic average.
A kind of mechanical index of the Double lumen intubation data based on BP neural network of the present invention determines that the advantage of method is： To determine that soil layer mechanical index provides reliable theoretical method by static sounding data；With a large amount of Double lumen intubation data With great soil group mechanical index as training sample, for mechanical index determines that the reliability of BP neural network forecast model provides strong branch Support, it is ensured that the accuracy for predicting the outcome；Probing hole count greatly reduces during survey can be made, so as to shorten quality, saves Cost is about reconnoitred, and the quality of engineering investigation can be improved.
Brief description of the drawings
Fig. 1 is the flow chart that a kind of mechanical index of the Double lumen intubation data based on BP neural network determines method；
Fig. 2 determines BP neural network structure chart for Double lumen intubation data mechanical index.
Specific embodiment
Embodiments of the present invention are illustrated below by way of specific instantiation, Fig. 1 is referred to, one kind is based on BP nerve nets The mechanical index of the Double lumen intubation data of network determines method：
1., the collection of the mechanical index of Double lumen intubation data and soil and arrangement：By compiling Yangtze River Delta Area Each typical geology regional cnginering geology exploration report, soil test achievement data etc., collect statistics Double lumen intubation data and The mechanical index of soil, 89 groups altogether, wherein comprising Q4 al28 groups of clay, Q4 al22 groups of silty clay, Q4 al27 groups of silt, Q4 alFlour sand 12 groups；
2., set up Double lumen intubation mechanical index and determine BP neural network forecast model：Based on BP neural network algorithm, Forecast model is set up using step data message 1., training sample example is shown in Table 1；
The Double lumen intubation mechanical index of table 1 determines BP neural network training sample example
3. BP neural network forecast model is trained, to be determined to Double lumen intubation mechanical index：It is soft with matlab Part, setting network model training function uses traincgf functions, hidden layer activation primitive to use tansing functions, output layer Activation primitive uses purelin functions, maximum iteration epochs=1000, anticipation error minimum value setting value goal= 0.01, the learning efficiency lr=0.05 of weights is corrected, BP neural network model is trained using the training sample of table 1；
4., with step 3. in completed training Double lumen intubation mechanical index determine BP neural network predict mould Type, by the Double lumen intubation of each layer soil in Jiangdu Tai Fu harbour affairs Co., Ltd stock yard job number SJT1 static soundings hole Statistical average is predicted to soil layer mechanical index, and SJT1 static soundings hole uses Double lumen intubation in-situ test, test Depth 13.0m, the quiet spy statistical average of doube bridge and mechanical index of each layer soil predict the outcome and are shown in Table 2.
The quiet spy statistical average of doube bridge and mechanical index of each layer soil of Jiangdu stock yard SJT1 of table 2 predict the outcome

Claims (4)

1. a kind of mechanical index of the Double lumen intubation data based on BP neural network determines method, it is characterised in that its side Method is comprised the following steps：
1., the collection of the mechanical index of Double lumen intubation data and soil and arrangement：Surveyed by compiling sitework geology Report, soil test achievement data are examined, the mechanical index of statistics Double lumen intubation data and soil is collected；
2., set up Double lumen intubation mechanical index and determine BP neural network forecast model：Based on BP neural network algorithm, utilize Step data message 1. sets up forecast model；
3. BP neural network forecast model is trained, to be determined to Double lumen intubation mechanical index：With matlab softwares, if Determine network training type function, output layer activation primitive type, maximum iteration epochs, anticipation error minimum value setting value The learning efficiency lr of goal and amendment weights, is trained to BP neural network model；
4., with step 3. in completed training Double lumen intubation mechanical index determine BP neural network forecast model, lead to Soil layer Double lumen intubation statistical average is crossed to be predicted soil layer mechanical index；Described Double lumen intubation mechanical index Determine BP neural network model using 1 input layer, 1 output layer and 1 network structure of hidden layer, specifically set up process by Following manner is implemented：
By along the n-th -2, n-1 of depth, n, n+1, n+2 static point resistance qcValue and n-2, n-1, n, n+1, n+2 side friction fs Used as input layer, network model input layer is made up of value 10 n dimensional vector ns：
X = [ q c n - 2 , q c n - 1 , q c n , q c n + 1 , q c n + 2 , f s n - 2 , f s n - 1 , f s n , f s n + 1 , f s n + 2 ]
Using the mechanical index of soil at n-th input value correspondence depth point as output layer, network model output layer is 4 n dimensional vector ns：
Y=[Dn,Cn,Jn,Mn]
DnRepresent n-th foundation bearing capacity of input value correspondence depth, CnRepresent n-th adhesive aggregation of input value correspondence depth Power, JnRepresent that n-th input value corresponds to depthMnRepresent n-th modulus of compressibility of input value correspondence depth.
2. the mechanical index determination side of a kind of Double lumen intubation data based on BP neural network according to claim 1 Method, it is characterised in that：Described Double lumen intubation data, including static point resistance qcWith side friction power fs；Described native power Index is learned, including：Foundation bearing capacity f0, shearing strength, compression modulus Es.
3. the mechanical index determination side of a kind of Double lumen intubation data based on BP neural network according to claim 1 Method, it is characterised in that：Described network training type function uses traincgf functions；Described output layer activation primitive type Using purelin functions；Described maximum iteration epochs is set as 1000 times；Described anticipation error minimum value setting Value goal is set as 0.01；The learning efficiency lr of described amendment weights is set as 0.05.
4. the mechanical index determination side of a kind of Double lumen intubation data based on BP neural network according to claim 1 Method, it is characterised in that：Described soil layer Double lumen intubation statistical average is all static point resistance q in each layer soilcWith side wall Frictional resistance fsArithmetic average.
CN201410104939.2A 2014-03-20 2014-03-20 A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method Active CN103912026B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410104939.2A CN103912026B (en) 2014-03-20 2014-03-20 A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410104939.2A CN103912026B (en) 2014-03-20 2014-03-20 A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method

Publications (2)

Publication Number Publication Date
CN103912026A CN103912026A (en) 2014-07-09
CN103912026B true CN103912026B (en) 2017-07-11

Family

ID=51037997

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410104939.2A Active CN103912026B (en) 2014-03-20 2014-03-20 A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method

Country Status (1)

Country Link
CN (1) CN103912026B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110147835A (en) * 2019-05-10 2019-08-20 东南大学 Resisting shear strength of reinforced concrete beam-column joints prediction technique based on grad enhancement regression algorithm
CN116244813B (en) * 2023-05-11 2023-07-18 中国铁路设计集团有限公司 Soil body strength parameter inversion method based on sounding technology

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5497091A (en) * 1994-09-01 1996-03-05 Applied Research Associates, Inc. Surface mounted pH sensor for cone penetration testing
CN101504736A (en) * 2009-02-27 2009-08-12 江汉大学 Method for implementing neural network algorithm based on Delphi software
CN102087677A (en) * 2010-12-24 2011-06-08 东南大学 Method for solving solution of elastic foundation beams and for back analysis of value m
KR20120011090A (en) * 2012-01-16 2012-02-06 김제성 Borehole geophysical logging system used safety pipe
CN102425148A (en) * 2011-09-02 2012-04-25 铁道第三勘察设计院集团有限公司 Rapid sub-grade settlement predicting method based on static sounding and BP (Back Propagation) neural network
CN102619209A (en) * 2012-04-25 2012-08-01 重庆大学 Geotechnical engineering cataclysm dynamic real-time intelligent early warning method based on displacement monitoring
EP2535460A1 (en) * 2011-06-16 2012-12-19 Hydro Geotechnique Device for measuring the resistance to ground penetration
CA2851874A1 (en) * 2011-10-11 2013-04-18 Ian Gray Formation pressure sensing system

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5497091A (en) * 1994-09-01 1996-03-05 Applied Research Associates, Inc. Surface mounted pH sensor for cone penetration testing
CN101504736A (en) * 2009-02-27 2009-08-12 江汉大学 Method for implementing neural network algorithm based on Delphi software
CN102087677A (en) * 2010-12-24 2011-06-08 东南大学 Method for solving solution of elastic foundation beams and for back analysis of value m
EP2535460A1 (en) * 2011-06-16 2012-12-19 Hydro Geotechnique Device for measuring the resistance to ground penetration
CN102425148A (en) * 2011-09-02 2012-04-25 铁道第三勘察设计院集团有限公司 Rapid sub-grade settlement predicting method based on static sounding and BP (Back Propagation) neural network
CA2851874A1 (en) * 2011-10-11 2013-04-18 Ian Gray Formation pressure sensing system
KR20120011090A (en) * 2012-01-16 2012-02-06 김제성 Borehole geophysical logging system used safety pipe
CN102619209A (en) * 2012-04-25 2012-08-01 重庆大学 Geotechnical engineering cataclysm dynamic real-time intelligent early warning method based on displacement monitoring

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
"淮北平原地基分层与承载力的BP网络预测方法";戴张俊，余飞，陈善雄，罗红明;《土木建筑与环境工程》;20130615;第35卷(第3期);第18-24页 *

Also Published As

Publication number Publication date
CN103912026A (en) 2014-07-09

Similar Documents

Publication Publication Date Title
Zhou et al. Numerical investigations into cycling of full-flow penetrometers in soft clay
Sarkar et al. Estimation of strength parameters of rock using artificial neural networks
Byrne et al. PISA design methods for offshore wind turbine monopiles
Li et al. A GMDH predictive model to predict rock material strength using three non-destructive tests
Yeter et al. Uncertainty analysis of soil-pile interactions of monopile offshore wind turbine support structures
Zhou et al. Effect of shaft on resistance of a ball penetrometer
Zambrano-Cruzatty et al. Numerical simulation of a free fall penetrometer deployment using the material point method
CN106546525A (en) The method and apparatus for setting up three-dimensional penetration rate model
Erzin et al. The use of neural networks for the prediction of cone penetration resistance of silty sands
Mamat et al. Slope stability prediction of road embankment on soft ground treated with prefabricated vertical drains using artificial neural network
CN103912026B (en) A kind of mechanical index of the Double lumen intubation data based on BP neural network determines method
CN103898890B (en) Soil layer quantization layering method based on double-bridge static sounding data of BP neural network
Robertson Soil behavior type using the DMT
Moayedi et al. A new real-time monitoring technique in calculation of the py curve of single thin steel piles considering the influence of driven energy and using strain gauge sensors
Das et al. Artificial neural network and regression models for prediction of free-field ground vibration parameters induced from vibroflotation
Yang et al. Effects of thixotropy and reconsolidation on the undrained shear characteristics of remoulded marine clays
Al-Taie et al. Compression index and compression ratio prediction by artificial neural networks
Rezaiee-Pajand et al. Crack detection in concrete gravity dams using a genetic algorithm
Lim Numerical study of cone penetration test in clays using press-replace method
CN115855686A (en) Method and device for measuring and calculating shear strength of cement mixing pile composite foundation
Lim et al. Shearing resistance during pile installation in sand
Aghda et al. Estimation of uniaxial compressive strength and modulus of deformability of the Asmari limestone, using neuro-fuzzy system
CN106844861A (en) Orthotropic Steel Bridge Deck welding node measured stress width dynamic probability modeling method
Sarker et al. Applicability of Standard Penetration Test in Bangladesh and Graphical Representation of SPT-N Value
CN113722962A (en) Side slope drainage pipe clogging simulation method based on particle flow analysis

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant