CN108228988A - A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving - Google Patents
A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving Download PDFInfo
- Publication number
- CN108228988A CN108228988A CN201711415882.8A CN201711415882A CN108228988A CN 108228988 A CN108228988 A CN 108228988A CN 201711415882 A CN201711415882 A CN 201711415882A CN 108228988 A CN108228988 A CN 108228988A
- Authority
- CN
- China
- Prior art keywords
- data
- slope
- displacement prediction
- analysis
- prediction
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
- G01B21/02—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring length, width, or thickness
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Pit Excavations, Shoring, Fill Or Stabilisation Of Slopes (AREA)
Abstract
Invention provides a kind of Slope Displacement Prediction based on big data driving and slides method of discrimination.This method include data acquisition, data prediction, adaptive regression spline analysis and model select and etc..This method is smaller to the dependence of traditional Slope Sliding theory and experience, and basic function is automatically performed according to data in calculating process, without manually setting, can accurately obtain the internal relations between slope displacement and more multivariate datas.This method has a extensive future, and is a kind of hazard prediction early warning technology of accurate quick.
Description
Technical field
The present invention relates to geological hazards prediction early warning technology, more particularly to a kind of slope displacement based on big data driving is pre-
Survey and slide method of discrimination.
Background technology
Landslide is a kind of common geological disaster in mountain area, since landslide side's amount is huge, often in the coverage of landslide
People's lives and properties and engineering construction facility cause greatly to threaten, and still, the generation on overwhelming majority landslide all has tendency,
Therefore it is analyzed for the monitoring data of side slope, predicts the development and generation on landslide, for identification and the prevention meaning on landslide
It is very great.
Existing slope displacement prediction method mainly has multi-variate statistical analysis, neural network (ANN) and support vector machines
(SVM) the methods of algorithm.Multi-variate statistical analysis can play part reference role in Slope Displacement Prediction, but can not be according to big
Data carry out landslide Accurate Model and effectively prediction；Artificial neural network can be by adjusting mutual between internal great deal of nodes
The relationship of connection, side slope monitoring big data, which is handled, carrys out qualitative forecasting Slope Sliding trend, but learning and memory has
Unstability, algorithm the convergence speed is slow, and cannot accurately provide prediction slope displacement and slide the Explicit functions differentiated；Branch
It holds vector machine and is built upon VC (Vapnik-Chervonenkis Dimension) the dimensions theory of Statistical Learning Theory and structure wind
A kind of supervised learning method on the basis of dangerous minimum principle, generalization ability are better than artificial neural network, and algorithm has the overall situation most
Dominance according to Small Sample Database Accurate Curve-fitting slope displacement and can slide differentiation Explicit functions, but its model is not easy to provide
Dominant explanation, and for large-scale training sample, which is difficult to carry out.
Invention content
The object of the present invention is to provide a kind of Slope Displacement Prediction based on big data driving and method of discrimination is slided, with solution
Certainly problems of the prior art.
To realize the present invention purpose and the technical solution adopted is that such, a kind of slope displacement based on big data driving
Prediction and slip method of discrimination, include the following steps：
1) the more multivariate datas of acquisition side slope within a specified time as dependent variable and are summarized.Wherein, it is described
More multivariate datas include geology and geomorphology data, ground Physical and mechanical properties data, crustal stress data, meteorological model number
According to the construction operation data near, displacement monitoring data and side slope.
2) summarize the abnormal data in Various types of data by variance analysis rejecting step 1).Wherein, data that treated are made
For the selected more field datas calculated for analysis.
3) 3/4 data of step 2) selected data are randomly selected as training data, remaining 1/4 data is as check number
According to, and preserved respectively to two texts.
4) prepare Multivariate adaptive regression splines batten tool box.Invocation step 3) in text carry out Multivariate adaptive regression splines batten
Analysis.Change basic function quantity, obtain the corresponding dependent variable model of different basic function quantity.
5) optimal models are selected, and passes through input and explains that the explanation to result of calculation is checked in order, obtain each basic function
Expression formula.
6) the Slope Displacement Prediction explicit function Y of this analysis is obtainedj。
7) the parameter relative importance of more multivariate datas of input is determined using Variance Decomposition Analysis.
8) determine that slope sliding differentiates Explicit functions f (Y using Logic Regression Modelsj)。
Further, after step 8), also have at regular intervals, side slope displacement prediction function differentiates table with sliding
The correlation step being updated up to formula.
Further, Logic Regression Models use sigmoid functions in step 8), and obtained slope sliding differentiates expression formula such as
Formula (1).
Wherein, as f (Yj) ＜ 0.5 when, side slope is in stable state.f(YjDuring) >=0.5, side slope plays pendulum.
The solution have the advantages that unquestionable：
A. dependence theoretical to traditional Slope Sliding and experience is smaller, and basic function is automatically complete according to data in calculating process
Into without manually setting, can accurately obtaining the internal relations between slope displacement and more multivariate datas；
B. it calculates quickly, can be with the larger sample of processing data amount, precision of prediction is high, and the explicit function tool finally obtained
There is interpretation；
C. a large amount of Slope Sliding trend dynamic data bases based on big data can be quickly established, according to more monitorings of side slope
Data differentiate expression formula with slope sliding by the parameter relative importance for analyzing input variable, realize the pre- of side slope sliding
Early warning is surveyed, plays the purpose prevented and reduced natural disasters.
Description of the drawings
Fig. 1 is method flow diagram.
Specific embodiment
With reference to embodiment, the invention will be further described, but should not be construed the above-mentioned subject area of the present invention only
It is limited to following embodiments.Without departing from the idea case in the present invention described above, according to ordinary skill knowledge and used
With means, various replacements and change are made, should all be included within the scope of the present invention.
Embodiment 1：
The present embodiment be based on big data thinking, using more multivariable monitor databases of acquisition as foundation, by it is polynary from
This data-driven method of regression spline is adapted to accurately to determine and establish the Slope Displacement Prediction lain between each measured data
Explicit function and slip differentiate Explicit functions.The present embodiment disclose a kind of Slope Displacement Prediction driven based on big data and
Method of discrimination is slided, referring to Fig. 1, is included the following steps：
1) data acquire：It acquires side slope digital elevation DEM at the appointed time in section (such as several months or several years) of side slope A, bore
The geology and geomorphology data X detected1i, slope ground physics and mechanical performance data X2i, GPS displacement monitoring data X3i, drilling
Tilt X4i, rainfall X5i, crustal stress X6iAnd neighbouring construction operation data XniAs dependent variable.The data of collection are abundanter,
Subsequent analysis result is better.In the present embodiment, GPS displacements are replaced using the component in main glide direction in GPS displacement datas
Data replace deviational survey data using the component in the main glide direction of bore inclining.The average value for taking every class data daily is aggregated into
In table, participate in subsequent analysis and calculate.
2) data prediction：Step 1) is summarized to more field data unified coordinate systems in table, and passes through variance analysis and picks
Except the abnormal data group in every class data.Data after rejecting are as the selected more field datas calculated for analysis.
3) it selects and preceding 3/4 data of selected more field datas is taken in step 2) as training data and are preserved to text document simultaneously
It is named as " Training ", rear 1/4 data are named as " Testing " as verification data and preservation to text document.
4) Matlab tool boxes prepare：Multivariate adaptive regression splines batten tool box ARESLab files are downloaded from network simultaneously
The toolbox files being copied under the installation directory of Matlab softwares, and this document folder path is added in Matlab programs.
5) Multivariate adaptive regression splines Spline analysis is carried out：Write or network download invocation step 3) in text document meter
Code is calculated, Multivariate adaptive regression splines Spline analysis is carried out after inputting Matlab, changes basic function (one group of special base of function space
Element, different functions can be gone out with composite construction) quantity j, obtain different from training data and verification data fitting degree
Dependent variable model.Wherein, the numerical value change section of basic function quantity j is number n to the n of dependent variable2.Basic function quantity j is from head
Item n, tolerance are incremented by for 10, reach n2When stop cycle calculations, obtain the corresponding dependent variable model Y of different basic function quantityj。
6) model selection and explanation：Using basic function quantity j as horizontal axis, the training data goodness of fit r of calculating2For the longitudinal axis,
Draw r2With the change curve of j, r is selected2Maximum model is to get to optimal models.Order is explained by inputting, and can be checked
Explanation of the software to result of calculation.
7) for optimal models, areseq (model, 5) orders is inputted in Matlab order lines, check software to model
As a result explanation can obtain the expression formula of each basic function and the Explicit functions of last slope displacement.Wherein, 5 represent it is small
Several digits remain into after decimal point 5.
8) according to the variation of GCV in model explanation (Generalized Cross Validation), using Variance Decomposition Analysis, the more of input are determined
The parameter relative importance of field multivariate data.
9) determine that slope sliding differentiates Explicit functions using Logic Regression Models.Logic Regression Models use sigmoid
Function, obtained slope sliding differentiate expression formula such as formula (1).
Slope sliding method of discrimination is as f (Yj) ＜ 0.5 when, side slope is in stable state.f(YjDuring) >=0.5, at side slope
In unstable state.This analysis terminates.
10) at regular intervals, side slope A displacement predictions function differentiates that expression formula is updated with sliding, to keep it
Dynamic.
What deserves to be explained is the present embodiment is by the very strong Multivariate adaptive regression splines specifically for high dimensional data of generalization ability
Spline algorithms are introduced into slope and land slide prediction, are not only calculated quickly, can be with the larger sample of processing data amount, precision of prediction
Height, and the explicit function finally obtained has interpretation.The technology can quickly establish a large amount of side slopes based on big data and slide
Shifting trend dynamic data base, according to more monitoring data of side slope, by the parameter relative importance and side slope of analyzing input variable
It slides and differentiates Explicit functions, realize the prediction and warning of side slope sliding, play the purpose prevented and reduced natural disasters.Its application prospect is wide
It is wealthy, it is a kind of hazard prediction early warning technology of accurate quick.Solve in the prior art can not carry out according to big data
Even if Accurate Model predict and can carry out Accurate Curve-fitting cannot but provide model explicitly explain (such as neural network and support to
Amount machine method) the problem of.
Claims (4)
1. a kind of Slope Displacement Prediction and slip method of discrimination based on big data driving, which is characterized in that include the following steps：
1) the more multivariate datas of acquisition side slope within a specified time as dependent variable and are summarized；Wherein, described more
Multivariate data includes geology and geomorphology data, ground Physical and mechanical properties data, crustal stress data, meteorological model data, position
Move the construction operation data near monitoring data and side slope；
2) summarize the abnormal data in Various types of data by variance analysis rejecting step 1)；Data that treated are as selected use
In more field datas that analysis calculates.
3) 3/4 data of step 2) selected data are randomly selected as training data, remaining 1/4 data as verification data, and
It is preserved respectively to two texts；
4) prepare Multivariate adaptive regression splines batten tool box；Invocation step 3) in text carry out Multivariate adaptive regression splines batten point
Analysis；Change basic function quantity j, obtain the corresponding dependent variable model of different basic function quantity；
5) optimal models are selected, and passes through input and explains that the explanation to result of calculation is checked in order, obtain the table of each basic function
Up to formula；
6) the Slope Displacement Prediction explicit function Y of this analysis is obtainedj；
7) the parameter relative importance of more multivariate datas of input is determined using Variance Decomposition Analysis；
8) determine that slope sliding differentiates Explicit functions f (Y using Logic Regression Modelsj)。
2. a kind of Slope Displacement Prediction and slip method of discrimination based on big data driving according to claim 1, special
Sign is：After step 8), also have at regular intervals, side slope displacement prediction function differentiates that expression formula carries out with sliding
Newer correlation step.
3. a kind of Slope Displacement Prediction and slip method of discrimination based on big data driving according to claim 3, special
Sign is：Logic Regression Models use sigmoid functions in step 8), and obtained slope sliding differentiates expression formula such as formula (1)；
Wherein, as f (Yj) ＜ 0.5 when, side slope is in stable state；f(YjDuring) >=0.5, side slope plays pendulum.
4. a kind of Slope Displacement Prediction and slip method of discrimination based on big data driving according to claim 1, special
Sign is that selecting optimal models described in step 5) specifically includes following steps：Using basic function quantity j as horizontal axis, the training of calculating
Data goodness of fit r2For the longitudinal axis, r is drawn2With the change curve of j, r is selected2Maximum model is optimal models.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711415882.8A CN108228988A (en) | 2017-12-25 | 2017-12-25 | A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711415882.8A CN108228988A (en) | 2017-12-25 | 2017-12-25 | A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108228988A true CN108228988A (en) | 2018-06-29 |
Family
ID=62648683
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711415882.8A Pending CN108228988A (en) | 2017-12-25 | 2017-12-25 | A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108228988A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240357A (en) * | 2021-07-12 | 2021-08-10 | 成都理工大学 | Rapid evaluation method for stability of highway side slope in red layer area |
| CN113536659A (en) * | 2021-06-09 | 2021-10-22 | 上海交通大学 | Method, system and storage medium for rapidly predicting post-earthquake road disaster area |
| CN115492175A (en) * | 2022-09-19 | 2022-12-20 | 中交第一公路勘察设计研究院有限公司 | Automatic monitoring system and method for highway side slope |
| CN116561563A (en) * | 2023-07-11 | 2023-08-08 | 电子科技大学 | Slope displacement prediction method and related device based on residual prediction model |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207952A (en) * | 2013-04-23 | 2013-07-17 | 华北科技学院 | Slope displacement prediction method |
| CN105869100A (en) * | 2016-04-29 | 2016-08-17 | 中国地质大学（武汉） | Method for fusion and prediction of multi-field monitoring data of landslides based on big data thinking |
-
2017
- 2017-12-25 CN CN201711415882.8A patent/CN108228988A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207952A (en) * | 2013-04-23 | 2013-07-17 | 华北科技学院 | Slope displacement prediction method |
| CN105869100A (en) * | 2016-04-29 | 2016-08-17 | 中国地质大学（武汉） | Method for fusion and prediction of multi-field monitoring data of landslides based on big data thinking |
Non-Patent Citations (3)
| Title |
|---|
| ÁNGEL FELICÍSIMO等: "Mapping landslide susceptibility with logistic regression,multiple adaptive regression splines, classification and regression trees, and maximum entropy methods:a comparative study", 《LANDSLIDES》 * |
| CHRISTIAN CONOSCENTI等: "Assessment of susceptibility to earth-flow landslide using logistic regression and multivariate adaptive regression splines: A case of the Belice River basin (western Sicily, Italy)", 《GEOMORPHOLOGY》 * |
| 张永磊等: "基于经验模式分解的边坡位移预测模型研究", 《人民长江》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113536659A (en) * | 2021-06-09 | 2021-10-22 | 上海交通大学 | Method, system and storage medium for rapidly predicting post-earthquake road disaster area |
| CN113240357A (en) * | 2021-07-12 | 2021-08-10 | 成都理工大学 | Rapid evaluation method for stability of highway side slope in red layer area |
| CN115492175A (en) * | 2022-09-19 | 2022-12-20 | 中交第一公路勘察设计研究院有限公司 | Automatic monitoring system and method for highway side slope |
| CN116561563A (en) * | 2023-07-11 | 2023-08-08 | 电子科技大学 | Slope displacement prediction method and related device based on residual prediction model |
| CN116561563B (en) * | 2023-07-11 | 2023-09-29 | 电子科技大学 | Slope displacement prediction method and related device based on residual prediction model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hengl et al. | Predictive soil mapping with R | |
| CN108228988A (en) | A kind of Slope Displacement Prediction and slip method of discrimination based on big data driving | |
| RU2723805C1 (en) | Method and computer system for control of drilling of the wells | |
| Qi et al. | Comparative study of hybrid artificial intelligence approaches for predicting hangingwall stability | |
| CA2640725C (en) | Methods, systems, and computer-readable media for fast updating of oil and gas field production models with physical and proxy simulators | |
| US8504341B2 (en) | Methods, systems, and computer readable media for fast updating of oil and gas field production models with physical and proxy simulators | |
| CN108363886A (en) | Deformation prediction method and system based on deep learning | |
| Wambeke et al. | A simulation-based geostatistical approach to real-time reconciliation of the grade control model | |
| CN109492775B (en) | Geological structure interpretation detection method, geological structure interpretation detection device and readable storage medium | |
| WO2023133213A1 (en) | Method for automated ensemble machine learning using hyperparameter optimization | |
| CN117235524A (en) | Learning training platform of automatic valuation model | |
| Seidu et al. | Impact of data partitioning in groundwater level prediction using artificial neural network for multiple wells | |
| Han et al. | Comprehensive analysis for production prediction of hydraulic fractured shale reservoirs using proxy model based on deep neural network | |
| Silva et al. | Prognosis of forest production using machine learning techniques | |
| CN105706105A (en) | Static earth model grid cell scaling and property re-sampling methods and systems | |
| Fattakhov et al. | Using artificial neural networks for analyzing efficiency of advanced recovery methods | |
| Dursun et al. | A workflow for intelligent data-driven analytics software development in oil and gas industry | |
| Belashev | Process “regime” shift recognition algorithm | |
| Caers et al. | Integration of engineering and geological uncertainty for reservoir performance prediction using a distance-based approach | |
| Vodenčarević et al. | Data analytics for manufacturing systems | |
| Li et al. | Reservoir ranking map sketching for selection of infill and replacement drilling locations using machine learning technique | |
| Gerges et al. | Machine-Learning-Assisted Well-Log Data Quality Control and Preprocessing Lab | |
| Saade et al. | LCA—interpretation of results | |
| Yousefzadeh et al. | Uncertainty Management in Reservoir Engineering | |
| Lombardi et al. | The first world catalog of earthquake‐rotated objects (EROs), Part II: Statistical analyses and related interpretations |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180629 |