CN111914395B - High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR - Google Patents
High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR Download PDFInfo
- Publication number
- CN111914395B CN111914395B CN202010610261.0A CN202010610261A CN111914395B CN 111914395 B CN111914395 B CN 111914395B CN 202010610261 A CN202010610261 A CN 202010610261A CN 111914395 B CN111914395 B CN 111914395B
- Authority
- CN
- China
- Prior art keywords
- matrix
- prediction
- calculation
- data
- valley amplitude
- 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
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a prediction analysis method for valley amplitude deformation of a high arch dam based on ARIMA-GC-SVR, which comprises the steps of firstly obtaining dam area data and preprocessing the data; an ARIMA model and an LASSO model are selected, and an influence factor matrix meeting the calculation requirement is constructed according to the length of a Grain Calculation (GC) window required by the actual engineering calculation requirement; according to the influence factor matrix, combining with the valley amplitude deformation value matrix, selecting grains for calculation to construct a granulation matrix; carrying out least square support vector machine regression (SVR) prediction analysis with the prediction length as a unit window on the granulation matrix; and finally, restoring the scale of the granulation data, if the predicted end point date is less than the target prediction time, updating the influence factor matrix and the granulation matrix, and predicting again until the target prediction time requirement is met. The method has high efficiency and low cost, simultaneously enlarges the predictable time range, improves the long-time prediction precision, and provides reference for the long-term prediction of the valley amplitude deformation.
Description
Technical Field
The invention relates to a prediction analysis method for valley amplitude deformation of a high arch dam, in particular to a prediction analysis method for valley amplitude deformation of a high arch dam based on ARIMA-GC-SVR.
Background
The valley amplitude deformation is a natural phenomenon which mainly occurs during the construction and operation of the high arch dam, and data show that the valley amplitude deformation is an important challenge facing the high arch dam and can influence the working state and long-term safety of the arch dam. The principle is that a plurality of pairs of measuring lines are arranged on two banks of a river valley as required, and the change relation of the valley amplitude along with time is obtained through analysis by recording the change of the length of the measuring lines.
At present, relatively few research results on the deformation of the valley amplitude exist, a unified recognition and research system is not formed, and qualitative research through monitoring data comparative analysis is mainly focused on. In addition, some scholars develop researches related to the deformation of the valley amplitude by establishing a reasonable numerical model and adopting a finite element simulation method, but the researches on the stress-strain state and the safety characteristic of the arch dam by analyzing the deformation of the valley amplitude by utilizing finite element simulation analysis are not developed in detail aiming at the specific phenomenon of the deformation of the valley amplitude.
The ARIMA model is a time series model for predicting variables with long-term change characteristics, and is widely applied to the aspect of forecasting hydrological meteorological elements; GC is a fuzzy calculation method, complex data point information can be fused into information elements with representatives, and the information elements are used as basic units for calculation, so that the calculation efficiency can be greatly improved; the SVR is a least squares support vector machine regression model and is usually used for continuous variable regression prediction, but the SVR combined with particle calculation has the defect of poor long-term prediction accuracy.
For safety supervision of both banks and safety supervision and early warning of an arch dam, a GC (gas chromatography) and SVR (support vector regression) model can be selected to predict the valley amplitude deformation value, but the precision of the conventional prediction method is not high under the long-term prediction condition.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the specific phenomenon of the deformation of the valley amplitude and considering the problems in the prior art, the invention aims to provide a prediction analysis method of the deformation of the valley amplitude of the high arch dam based on the ARIMA-GC-SVR, so as to provide long-term prediction of the deformation of the valley amplitude with higher prediction precision and larger coverage range, solve the problem of shorter prediction time range of the SVR interval, and provide reference for prediction analysis of the deformation value of the valley amplitude.
The technical scheme is as follows: a high arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR comprises the following steps:
(1) Preprocessing dam region related monitoring data, and establishing an influence factor data analysis system; the data analysis system comprises three layers of data matrixes, namely a first-level data matrix, a second-level data matrix and a third-level data matrix;
(2) Performing prediction processing on the dam area related monitoring data, analyzing and comparing the characteristics of the monitoring data, performing target time interval prediction on a primary data matrix by using an ARIMA method, expanding the primary data matrix, and updating a secondary data matrix and a tertiary data matrix;
(3) Determining the length of a calculation window according to engineering calculation requirements, calculating the number of windows, and constructing a single calculation influence factor matrix based on three-level data matrixes;
(4) According to the single calculation influence factor matrix, combining with a valley amplitude deformation value matrix, and selecting Grain Calculation (GC) to construct a granulation matrix;
(5) Constructing an SVR model according to the granulation matrix, and carrying out interval prediction analysis with the prediction length as a unit window on the granulation matrix;
(6) Reducing the scale of the granulation data, and determining a predicted end point date; if the predicted end point date is smaller than the target predicted date, updating the single calculation influence factor matrix in the step (3) and the granulation matrix in the step (4), and performing interval prediction again until the predicted end point date is larger than the target predicted date;
(7) And sorting the calculation results to obtain a valley amplitude deformation value prediction interval.
Further, in the step (1), the dam area related monitoring data includes a reservoir water level elevation, a reservoir water level lifting rate, an accumulated value of valley amplitude deformation of each measurement line, a dam area daily air temperature mean value and a dam area daily rainfall.
Further, in the step (1), the actually monitored influence factor value matrix is used as a primary data matrix, the high-dimensional influence factor matrix representing the potential expression form of the influence factor is used as a secondary data matrix, and the relatively important influence factor matrix screened out through calculation is used as a tertiary data matrix.
Further, in step (1), the preprocessing includes removing abnormal values and deleting missing values.
Furthermore, the determination method of the three-level data matrix is to establish a high-dimensional factor regression model based on LASSO, take the basic influence factors contained in the two-level matrix as independent variables and the deformation values as dependent variables, and screen out relatively important influence factors through a sparse matrix to serve as basic elements of the three-level matrix.
Further, in the step (3), the length of the particle calculation window is a fixed value in the calculation process, the number of the calculation windows is sequentially increased by one according to the circulation condition, and a single calculation influence factor matrix with the dimension (i x n) is obtained by calculation based on the three-level data matrix with the dimension (l x n), the length of the single calculation influence factor matrix is sequentially increased by (w x n) according to the circulation condition, wherein i is the number of the windows, w is the length of the particle calculation window, n is the number of the relatively important influence factors, and l is the length of the prediction time.
Further, in the step (4), a GC is used for selecting a proper kernel function, and a granulation matrix X is constructed based on the single calculation influence factor matrix ij Constructing a granulation matrix Y based on the valley amplitude deformation value matrix i Combining the two calculation results to obtain a granulation matrix (X) ij ,Y i )。
Further, the step (5) specifically includes the following steps:
granulating matrix X ij As an independent variable matrix, Y i Constructing an SVR model as a dependent variable matrix;
respectively training the models in three conditions [ L, M and R ], selecting a grid search method to optimize SVR model parameters, and selecting an optimal model by using a cross-validation method;
and calculating predicted values under three conditions [ L, M and R ], and sorting the predicted results to obtain a valley amplitude deformation value prediction interval of the next window.
Further, the step (6) specifically includes the following contents:
reducing the scale of the granulation data according to the window length, and determining the predicted end point date;
judging whether the end point date is larger than the target prediction date;
if the end point date is less than the target prediction date, returning to the step (3), enabling the window number i = i +1, and simultaneously predicting the end point y i ＝[lo i ,r i ,u i ]Assigning value to Y in step (4) i Recalculating the valley amplitude deformation value prediction interval of the next window;
and if the end date is greater than the target prediction date, jumping out of the loop.
Further, the step (7) specifically includes the following steps:
sorting the calculation results of each cycle, i.e. sorting result matrix Y i ＝[LO i ,R i ,U i ]The interval vector of (1), the interval deformation value corresponding to each time point, and the time point corresponding to the ith window is t i And (5) obtaining a valley amplitude deformation value prediction interval matrix Y = w = i[LO,R,U]。
The invention has the beneficial effects that: compared with the prior art, the method for predicting and analyzing the deformation of the valley amplitude of the high arch dam, disclosed by the invention, has the advantages that an influence factor data analysis system is constructed, the potential expression form of the influence factor is considered, relatively important influence factors are determined, and more reliable and effective basic data are provided for subsequent prediction; hydrological data and reservoir water level data in a target time period are subjected to predictive analysis by combining an ARIMA model, so that the data range is expanded, and basic data are provided for the subsequent rolling prediction of a long-term deformation value; meanwhile, a model is established based on LASSO, so that the characteristic screening of the influence factors is realized, representative important influence factors are selected finally, and more effective influence factors are provided for subsequent prediction; in addition, a GC method is selected, information is subjected to fuzzy granulation, interval prediction is given to the valley amplitude deformation value by combining with an SVR model, the idea of loop iteration is introduced, rolling prediction is carried out, and the problem that the precision of the existing prediction technology is not high under the long-term prediction condition is solved. The method provides a basis for long-term prediction and safety control of the deformation of the valley amplitude of the high arch dam, and has high reliability.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a schematic view of a portion of a granulation matrix in accordance with an embodiment of the present invention;
fig. 3 is a schematic diagram of comparison between the monitoring values and the prediction intervals according to the embodiment of the present invention.
Detailed Description
The invention discloses a prediction analysis method for valley amplitude deformation of a high arch dam based on ARIMA-GC-SVR, which comprises the steps of firstly, obtaining dam area related data, and preprocessing the data; an ARIMA model and an LASSO model are selected, and an influence factor matrix meeting the calculation requirement is constructed according to the Grain Calculation (GC) window length required by the actual engineering calculation requirement; then, adopting GC to fuzzify the data into [ L, M, R ]3 dimensions, and constructing a granulation matrix according to the influence factor matrix and the valley amplitude deformation value matrix; carrying out least square support vector machine (SVR) prediction analysis with prediction length as a unit window on the granulation matrix; and finally, restoring the scale of the granulation data, if the predicted end point date is less than the target prediction time, updating the influence factor matrix and the granulation matrix, and predicting again until the target prediction time requirement is met.
The technical solution of the present invention is further described in detail with reference to the drawings and the embodiments.
As shown in fig. 1, the method for predicting and analyzing the valley amplitude deformation of the high arch dam based on ARIMA-GC-SVR of the present invention comprises:
step one, sorting and collecting dam area related monitoring data, preprocessing the data such as removing abnormal values and deleting missing values, and establishing an influence factor data analysis system, wherein the data analysis system comprises three layers of data matrixes (a first layer, a second layer and a third layer);
in this embodiment, the relevant monitoring data includes a reservoir water level elevation, a reservoir water level lifting rate, an accumulated value of valley amplitude deformation of each measurement line, a single-day temperature average value of the dam area, and a single-day rainfall of the dam area.
The influence factor data analysis system comprises the basic data of reservoir water level elevation, reservoir water level lifting rate, single-day air temperature mean value and single-day rainfall data of a dam area, wherein the data system comprises three layers of data matrixes, and specifically comprises the following steps: the method comprises the following steps of taking an actual monitoring influence factor value matrix as a primary data matrix, taking a high-dimensional influence factor matrix capable of embodying potential expression forms of influence factors as a secondary data matrix, establishing a high-dimensional factor regression model through LASSO, and taking a relatively important influence factor matrix screened out by calculating a sparse matrix as a tertiary data matrix, wherein the method specifically comprises the following steps of:
1. suppose with A l4 Representing a primary data matrix with a visible matrix dimension of [ l 4]Wherein l represents the number of monitoring data groups, and the unit day is the measurement unit of 1 group, and 4 represents that four influence factors are considered, then an actual monitoring influence factor value matrix can be established as a primary data matrix, as shown in table 1:
TABLE 1 schematic table of primary data matrix of influence factor data analysis system
| Date of day | Elevation of reservoir water level | Reservoir water level rise and fall rate | Mean value of temperature of single day | Single day rainfall |
| yy/mm/dd | a 11 | a 12 | a 13 | a 14 |
| … | … | … | … | … |
| yy/mm/dd | a n1 | a n2 | a n3 | a n4 |
* The dates in the table being consecutive dates
2. Suppose with B lk Representing a two-level data matrix with a visible matrix dimension of [ l k]Wherein l represents the number of monitoring data groups, the unit day is a measuring unit of 1 group, k represents the number of potential expression forms of the influence factors, and in the embodiment, the diversity of hydrologic conditioning effects is considered, the rainfall is accumulated in the previous n days, the maximum difference of the rainfall in the previous n days, and the rainfall in the previous n daysAccumulating the gas temperature value and the maximum difference value of the former n weather temperatures, wherein n belongs to { 5.. 85 }, then establishing a two-level data matrix by taking the actual monitoring influence factor value matrix as shown in table 2:
TABLE 2 two-level data matrix schematic table of influence factor data analysis system
* The dates in the table being consecutive dates
3. Suppose with C ln Representing three levels of data matrix, the dimension of the visible matrix is [ l n ]]Wherein l represents the number of monitoring data sets, the unit day is a 1-set metering unit, n represents the number of influence factors screened out based on LASSO dimension reduction, the value of n in this embodiment is 6, the calculation results are rainfall amount on the day, cumulative rainfall amount in the previous 10 days, maximum difference of rainfall amount in the previous 10 days, and maximum difference of temperature in the previous 25 days, and a relatively important influence factor matrix screened out can be constructed by combining with the elevation of the reservoir water level and the variation rate of the reservoir water level to be a three-level data matrix, as shown in Table 3:
TABLE 3 schematic table of three-level data matrix of influence factor data analysis system
* The dates in the table being consecutive dates
Step two, further predicting the monitoring data, and taking the predicted value as a granulation matrix X in step five ij The method comprises the steps of analyzing and comparing the characteristics of monitored data, expanding reservoir water level elevation, the average temperature per day of the dam area and the rainfall per day of the dam area in a primary data matrix by an ARIMA method, updating the primary data matrix, and sequentially updating a secondary matrix and a tertiary matrix.
And step three, determining the length w of a calculation window according to the engineering calculation requirement, calculating the number i of windows, and constructing a single calculation influence factor matrix with the dimension (i x n) based on the three-level data matrix. The particle calculation window length w calculation process is a fixed value, the calculation window number i is sequentially added by one according to the circulation condition, the single calculation influence factor matrix is calculated based on the three-level data matrix with the dimension being (l × n), the dimension is sequentially added by (w × n) according to the circulation condition, wherein n is the number of relatively important influence factors, l is the prediction time length, w =10 in the embodiment, initial l =1300, and l =1580 at the end.
Selecting a proper kernel function, applying GC granulation data, and constructing a granulation X based on the single calculation influence factor matrix ij Constructing a granulation matrix Y based on the valley amplitude deformation value matrix i Combining the two calculation results to obtain a granulation matrix of (X) ij ,Y i )，Y i From a plurality of y i Composition y i Is the result of one operation. Granulation matrix Y constructed based on valley amplitude deformation value matrix in the embodiment i A partial schematic view is shown in fig. 2.
And fifthly, constructing an SVR model according to the granulation matrix, and calculating to obtain a valley amplitude deformation value prediction interval of the next window. The method comprises the following specific steps:
(1) granulating matrix X ij As an independent variable matrix, Y i Constructing an SVR model as a dependent variable matrix;
(2) respectively training the models in three conditions [ L, M and R ], selecting a grid search method to optimize SVR model parameters, and selecting an optimal model by using a cross validation method (five-fold cross validation);
(3) calculate three cases [ L, M, R]And sorting the prediction results to obtain a deformation value prediction interval y i ＝[lo i ,r i ,u i ]，lo i Is a lower boundary, r i Is a median value of u i Is the upper boundary.
And step six, restoring the scale of the granulated data according to the window length, determining the predicted end point date, and meanwhile judging whether the end point date is greater than the target predicted date, wherein in the embodiment, the end point date is required to be greater than 1580, specifically 28 times in a circulating manner. The method comprises the following specific steps:
(1) reducing the scale of the granulation data according to the window length, and determining the predicted end point date;
(2) judging whether the end point date is larger than the target predicted date or not;
(3) if not, returning to the step three, enabling the window number i = i +1, and meanwhile predicting the result y i ＝[lo i ,r i ,u i ]Assign Y in step four i Recalculating;
(4) and if so, jumping out of the loop and performing the step seven.
Step seven, sorting the calculation results of each cycle, and obtaining a valley amplitude deformation value prediction interval according to the obtained matrix [ L, M, R ]; summarizing the calculation results of each time to obtain a predicted valley-amplitude deformation curve; and (4) giving a comparison reference by combining the average change rate of the valley amplitude deformation and the change rate of the reservoir water level scheduling similar period. In this embodiment, the data of the first 1300 days is used as the basic data, the data of the second 280 days is used as the target predicted deformation value, and the prediction result can be represented as a schematic diagram shown in fig. 3, where the red line is the predicted upper boundary value, the blue line is the predicted lower boundary value, and the green line is the original monitoring data.
In conclusion, the influence factor data analysis system is constructed, the range of basic data prediction is expanded based on the ARIMA model, the potential expression form of the influence factors is considered, relatively important influence factors are determined based on the LASSO, and the basic data with larger range, more reliability and effectiveness are provided for deformation prediction; meanwhile, based on the idea of loop iteration, GC and SVR, the interval deformation value of the valley amplitude deformation is predicted in a rolling mode, and the dam area engineering safety and stability supervision is facilitated.
Claims (7)
1. A high arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR is characterized by comprising the following steps:
(1) Preprocessing dam region related monitoring data, and establishing an influence factor data analysis system; the data analysis system comprises three layers of data matrixes, namely a first-level data matrix, a second-level data matrix and a third-level data matrix; taking an actual monitoring influence factor value matrix as a primary data matrix, taking a high-dimensional influence factor matrix representing the potential expression form of the influence factor as a secondary data matrix, and taking a relatively important influence factor matrix screened out through calculation as a tertiary data matrix;
(2) Performing prediction processing on the dam area related monitoring data, analyzing and comparing the characteristics of the monitoring data, performing target time interval prediction on a primary data matrix by using an ARIMA method, expanding the primary data matrix, and updating a secondary data matrix and a tertiary data matrix;
(3) Determining the length of a calculation window according to engineering calculation requirements, calculating the number of windows, and constructing a single calculation influence factor matrix based on three-level data matrixes;
(4) According to the single calculation influence factor matrix, combining with a valley amplitude deformation value matrix, selecting grain calculation to construct a granulation matrix; the specific process is as follows: selecting kernel functions by using GC (gas chromatography), and constructing a granulation matrix X based on the single calculation influence factor matrix ij Constructing a granulation matrix Y based on the valley amplitude deformation value matrix i Combining the two calculation results to obtain a granulation matrix (X) ij ,Y i ) Wherein, Y i ＝[LO i ,R i ,U i ]；
(5) Constructing an SVR model according to the granulation matrix, and carrying out interval prediction analysis with the prediction length as a unit window on the granulation matrix; the method comprises the following steps:
granulating matrix X ij As an independent variable matrix, Y i Constructing an SVR model as a dependent variable matrix;
respectively training the models under three conditions [ L, M and R ], selecting a grid search method to optimize SVR model parameters, and selecting an optimal model by using a cross validation method;
calculating predicted values under three conditions [ L, M and R ], and sorting the predicted results to obtain a valley amplitude deformation value prediction interval of the next window;
(6) Reducing the scale of the granulation data, and determining a predicted end point date; if the predicted end point date is smaller than the target predicted date, updating the single calculation influence factor matrix in the step (3) and the granulation matrix in the step (4), and performing interval prediction again until the predicted end point date is larger than the target predicted date;
(7) And sorting the calculation results to obtain a valley amplitude deformation value prediction interval.
2. The method for predictive analysis of the valley amplitude deformation of a high arch dam as claimed in claim 1, wherein: in the step (1), the dam area related monitoring data comprises a reservoir water level elevation, a reservoir water level lifting rate, an accumulated value of valley amplitude deformation of each measured line, a single-day air temperature mean value of the dam area and a single-day rainfall of the dam area.
3. The high arch dam valley amplitude deformation prediction analysis method according to claim 1 or 2, characterized in that: in the step (1), the preprocessing comprises removing abnormal values and deleting missing values.
4. The method for predictive analysis of the valley amplitude deformation of a high arch dam as claimed in claim 1, wherein: the method for determining the three-level data matrix comprises the steps of establishing a high-dimensional factor regression model based on LASSO, taking basic influence factors contained in a secondary matrix as independent variables and deformation values as dependent variables, and screening out relatively important influence factors through a sparse matrix to serve as basic elements of the three-level matrix.
5. The method for predictive analysis of the valley amplitude deformation of a high arch dam as claimed in claim 1, wherein: in the step (3), the length of the particle calculation window is a fixed value in the calculation process, the number of the calculation windows is sequentially added by one according to the circulation condition, a single calculation influence factor matrix with the dimension (i x n) is obtained through calculation based on the three-level data matrix with the dimension (l x n), the length of the single calculation influence factor matrix is sequentially added by (w x n) according to the circulation condition, wherein i is the number of the windows, w is the length of the particle calculation window, n is the number of the relatively important influence factors, and l is the length of the prediction time.
6. The prediction analysis method for the deformation of the valley amplitude of the high arch dam as claimed in claim 1, wherein the step (6) comprises the following steps:
reducing the scale of the granulation data according to the window length, and determining the predicted end point date;
judging whether the end point date is larger than the target prediction date;
if the end point date is less than the target prediction date, returning to the step (3), enabling the window number i = i +1, and simultaneously predicting the result y i ＝[lo i ,r i ,u i ]Assigning to the matrix Y in the step (4) i Recalculating the valley amplitude deformation value prediction interval of the next window; and if the end date is greater than the target prediction date, jumping out of the loop.
7. The prediction analysis method for the deformation of the valley amplitude of the high arch dam as claimed in claim 6, wherein the step (7) comprises the following steps:
sorting the calculation results of each cycle, i.e. sorting result matrix Y i ＝[LO i ,R i ,U i ]The interval vector of (1), the interval deformation value corresponding to each time point, and the time point corresponding to the ith window is t i And (5) obtaining a valley amplitude deformation value prediction interval matrix Y = [ LO, R, U ] by = w × i]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010610261.0A CN111914395B (en) | 2020-06-30 | 2020-06-30 | High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010610261.0A CN111914395B (en) | 2020-06-30 | 2020-06-30 | High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111914395A CN111914395A (en) | 2020-11-10 |
| CN111914395B true CN111914395B (en) | 2022-11-08 |
Family
ID=73226220
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010610261.0A Active CN111914395B (en) | 2020-06-30 | 2020-06-30 | High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111914395B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106951976A (en) * | 2016-10-12 | 2017-07-14 | 华南理工大学 | A kind of bus passenger flow Forecasting Methodology based on pattern classification |
| CN110374047A (en) * | 2019-05-28 | 2019-10-25 | 中国水利水电科学研究院 | Arch dam runtime real-time security monitoring Threshold based on deformation |
| CN110378070A (en) * | 2019-08-05 | 2019-10-25 | 成都理工大学 | Based on PSO-SVR and the united Prediction of Displacement in Landslide method of DES |
| CN111275341A (en) * | 2020-01-21 | 2020-06-12 | 河海大学 | High arch dam valley amplitude deformation analysis method based on lasso and random forest |
-
2020
- 2020-06-30 CN CN202010610261.0A patent/CN111914395B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106951976A (en) * | 2016-10-12 | 2017-07-14 | 华南理工大学 | A kind of bus passenger flow Forecasting Methodology based on pattern classification |
| CN110374047A (en) * | 2019-05-28 | 2019-10-25 | 中国水利水电科学研究院 | Arch dam runtime real-time security monitoring Threshold based on deformation |
| CN110378070A (en) * | 2019-08-05 | 2019-10-25 | 成都理工大学 | Based on PSO-SVR and the united Prediction of Displacement in Landslide method of DES |
| CN111275341A (en) * | 2020-01-21 | 2020-06-12 | 河海大学 | High arch dam valley amplitude deformation analysis method based on lasso and random forest |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111914395A (en) | 2020-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Farmani et al. | Pipe failure prediction in water distribution systems considering static and dynamic factors | |
| Parhizkar et al. | Evaluation and improvement of energy consumption prediction models using principal component analysis based feature reduction | |
| CN112465243B (en) | Air quality forecasting method and system | |
| WO2020043028A1 (en) | Method for predicting air pollution using historical air quality data features | |
| CN109408848B (en) | Distributed attribution method considering runoff evolution space-time heterogeneity | |
| CN108009674A (en) | Air PM2.5 concentration prediction methods based on CNN and LSTM fused neural networks | |
| CN109116444B (en) | PCA-kNN-based air quality model PM2.5Forecasting method | |
| CN110648014A (en) | Regional wind power prediction method and system based on space-time quantile regression | |
| CN113632101A (en) | Method for predicting atmospheric pollution through vectorization analysis | |
| CN104866923A (en) | Steel enterprise blast furnace by-product gas emergence size prediction method | |
| CN107169289A (en) | It is a kind of based on the Landslide Hazard Assessment method of optimal weights combination method can be opened up | |
| CN106526710A (en) | Haze prediction method and device | |
| CN114169434A (en) | Load prediction method | |
| CN111080039A (en) | Fan cluster fault prediction method and system | |
| CN114997534A (en) | Similar rainfall forecasting method and equipment based on visual features | |
| Sha et al. | Development of a key-variable-based parallel HVAC energy predictive model | |
| Sun et al. | Spatial-temporal prediction of air quality based on recurrent neural networks | |
| CN115630337A (en) | Quantitative evaluation method and system for extreme rainfall attribution based on large-scale climate remote correlation | |
| CN117909888B (en) | Intelligent artificial intelligence climate prediction method | |
| CN116205508A (en) | Distributed photovoltaic power generation abnormality diagnosis method and system | |
| CN115423146A (en) | Self-adaptive runoff forecasting method based on multi-factor nearest neighbor sampling regression and support vector machine | |
| CN114581141A (en) | Short-term load prediction method based on feature selection and LSSVR | |
| CN106897497A (en) | Based on the prediction of precipitation method for sliding gray model and weighted Markov chain | |
| CN111914395B (en) | High arch dam valley amplitude deformation prediction analysis method based on ARIMA-GC-SVR | |
| CN116739642B (en) | Engineering cost analysis method, system and storage medium |
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 |