CN115217152B - Method and device for predicting open-close deformation of immersed tunnel pipe joint - Google Patents
Method and device for predicting open-close deformation of immersed tunnel pipe joint Download PDFInfo
- Publication number
- CN115217152B CN115217152B CN202210904682.3A CN202210904682A CN115217152B CN 115217152 B CN115217152 B CN 115217152B CN 202210904682 A CN202210904682 A CN 202210904682A CN 115217152 B CN115217152 B CN 115217152B
- Authority
- CN
- China
- Prior art keywords
- ssa
- component
- data
- value
- svr model
- 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
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02D—FOUNDATIONS; EXCAVATIONS; EMBANKMENTS; UNDERGROUND OR UNDERWATER STRUCTURES
- E02D29/00—Independent underground or underwater structures; Retaining walls
- E02D29/063—Tunnels submerged into, or built in, open water
- E02D29/073—Tunnels or shuttering therefor assembled from sections individually sunk onto, or laid on, the water-bed, e.g. in a preformed trench
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02D—FOUNDATIONS; EXCAVATIONS; EMBANKMENTS; UNDERGROUND OR UNDERWATER STRUCTURES
- E02D33/00—Testing foundations or foundation structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- 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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Structural Engineering (AREA)
- Computer Hardware Design (AREA)
- Civil Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Paleontology (AREA)
- Mining & Mineral Resources (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Mathematical Optimization (AREA)
- Architecture (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Environmental & Geological Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a device for predicting the opening and closing deformation of a pipe joint of a immersed tunnel, which are characterized in that the opening and closing deformation data are divided into a trend component, a period component and a residual component for parallel processing, so that the complexity of original opening and closing data is effectively reduced, the prediction efficiency is improved, and the condition of each component in the opening and closing amount of the pipe joint can be clearly reflected; compared with the traditional SVR model, the SVR model has more excellent prediction performance, small model prediction error and strong generalization capability.
Description
Technical Field
The invention relates to the technical field of tunnel deformation prediction, in particular to a method and a device for predicting open-close deformation of a immersed tunnel pipe joint connector.
Background
With a series of immersed tunnel engineering in China being put into operation successively, the service safety performance of the immersed tunnel engineering is widely paid attention to by people. The immersed tube tunnel structure in the operation period is in a severe and changeable environment and complex stress characteristics and is in a complex bending, twisting, pulling and pressing state for a long time, wherein the tube joint is used as the weakest and most stressed part in the whole tunnel system, deformation such as differential settlement, opening and closing, torsion and the like is very easy to generate in the state, and the excessive stretching amount of the tube joint can cause the occurrence of actions such as water stop belt breakage, shear key damage, water leakage and the like, so that the service safety of the immersed tube tunnel is seriously threatened. Therefore, the future situation of the opening and closing deformation of the pipe joint connector is timely and effectively forecasted, and the method has important practical significance for timely finding potential safety hazards and guaranteeing operation safety.
At present, methods such as model test, numerical simulation and the like are mainly adopted for researching the deformation of the pipe joint. However, the model test has the defects of undetermined similarity ratio, time and labor consumption, simplified model in numerical simulation and the like, and the development of the traditional method is limited to a certain extent. Along with the continuous integration of the intersection disciplines, deformation prediction research based on an artificial intelligence algorithm has become a hotspot, is mainly concentrated on surrounding rock deformation, surface subsidence and the like in the field of tunnel engineering, and has not been disclosed in the report of prediction research on open-close deformation of immersed tube tunnel joints.
The Support Vector Regression (SVR) model has perfect theoretical foundation and strong nonlinear fitting capability, and is a commonly used model in deformation prediction research, but the existing research shows that the model still lacks an efficient method in the selection of the nuclear parameter g and the penalty factor C.
Accordingly, in order to overcome the above-described drawbacks of the prior art, a solution is needed.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a method and a device for predicting the opening and closing deformation of a immersed tube tunnel joint, which are used for solving the technical problem that the opening and closing deformation prediction is difficult to be performed efficiently in the prior art.
A prediction method for open-close deformation of a immersed tunnel pipe joint is characterized by comprising the following steps: acquiring opening and closing deformation data and environment data of a pipe joint to be predicted, wherein the environment data at least comprises temperature data and water depth data; decomposing the opening and closing deformation data by adopting a singular spectrum analysis method, and reconstructing according to the components obtained by decomposition to respectively obtain a trend component, a periodic component and a residual component; dividing the trend component, the period component and the residual component into a training set and a testing set which are respectively corresponding to each other; fitting and predicting the trend component by adopting a least square method to obtain a trend component prediction result; training the constructed first SSA-SVR model by adopting a training set of periodic components to obtain a first target SSA-SVR model, and carrying out periodic component prediction by adopting the first target SSA-SVR model to obtain a periodic component prediction result; training the constructed second SSA-SVR model by adopting a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by adopting the second target SSA-SVR model to obtain a residual component prediction result; and accumulating and calculating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on a time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation.
In one embodiment, the step of obtaining the opening and closing deformation data and the environmental data at the joint of the pipe joint to be predicted, wherein the environmental data at least includes temperature data and water depth data further includes: judging whether the opening and closing deformation data set is equidistant interval data or not; if yes, carrying out a subsequent decomposition step of the opening and closing deformation sequence by adopting a singular spectrum analysis method; if not, performing equidistant processing by adopting a cubic spline interpolation method to obtain an equidistant opening and closing deformation data set, and then performing a decomposition step on the opening and closing deformation data by adopting a singular spectrum analysis method.
In one embodiment, the step of fitting and predicting the trend component by using a least square method to obtain a trend component prediction result includes: fitting a least square function based on the training set of the trend component to obtain a target least square function; and carrying out trend component prediction according to the target least square function to obtain a trend component prediction result.
In one embodiment, training the constructed first SSA-SVR model by using a training set of periodic components to obtain a first target SSA-SVR model, and performing periodic component prediction by using the first target SSA-SVR model to obtain a periodic component prediction result, including: respectively carrying out normalization processing on the training set, the prediction set and the corresponding environmental data of the periodic component; establishing a first SVR model, taking the opening and closing deformation data and the environmental data of 1 day before the predicted day as input, and taking the opening and closing deformation data of the predicted day as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the first SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; and using the obtained optimal combination value of the penalty factors and the nuclear parameters in the first SVR model, predicting a test set of the periodic components, and performing inverse normalization operation to obtain a prediction result of the periodic components.
In one embodiment, the step of training the constructed second SSA-SVR model with the training set of residual components to obtain a second target SSA-SVR model, and predicting residual components with the second target SSA-SVR model to obtain a residual component prediction result includes: respectively carrying out normalization processing on the training set and the prediction set of the residual components; establishing a second SVR model, wherein the residual component value of 3-6 days before the predicted day in the residual component is taken as input, and the residual component value of the predicted day is taken as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the second SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; and using the obtained optimal combination value of the penalty factor and the kernel parameter in the second SVR model to predict a test set of the residual component, and performing inverse normalization operation to obtain a prediction result of the residual component.
In one embodiment, after the step of accumulating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on the time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation, the method comprises the following steps: and quantitatively evaluating the prediction result by adopting the determination coefficient, the average absolute error and the root mean square error.
The utility model provides a immersed tube tunnel tube coupling connects prediction unit who opens and shuts deformation which characterized in that includes: the system comprises a data acquisition module, a data decomposition module and a data prediction module, wherein the data acquisition module is used for acquiring opening and closing deformation data and environment data of a pipe joint to be predicted, and the environment data at least comprises temperature data and water depth data; the data decomposition module is used for decomposing the opening and closing deformation data by adopting a singular spectrum analysis method, and reconstructing according to the components obtained by decomposition to respectively obtain a trend component, a periodic component and a residual component; dividing the trend component, the period component and the residual component into a training set and a testing set which are respectively corresponding to each other; the data prediction module is used for fitting and predicting the trend component by adopting a least square method to obtain a trend component prediction result; training the constructed first SSA-SVR model by adopting a training set of periodic components to obtain a first target SSA-SVR model, and carrying out periodic component prediction by adopting the first target SSA-SVR model to obtain a periodic component prediction result; training the constructed second SSA-SVR model by adopting a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by adopting the second target SSA-SVR model to obtain a residual component prediction result; and accumulating and calculating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on a time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation.
A computer device comprising a memory, a processor and a computer program stored in the memory and operable on the processor, wherein the processor executes the program to implement the steps of a method for predicting opening and closing deformation of a immersed tube tunnel pipe joint according to the above embodiments.
A storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of a method of predicting open-close deformation of a immersed tube tunnel joint as described in the above embodiments.
According to the technical scheme, the beneficial technical effects of the invention are as follows:
1. according to the scheme, the opening and closing deformation data are divided into the trend component, the period component and the residual component for parallel processing, so that the complexity of the original opening and closing data is effectively reduced, and the situation of each component in the opening and closing amount of the pipe joint can be clearly reflected; compared with the traditional SVR model, the SVR model has more excellent prediction performance, small model prediction error and strong generalization capability.
2. According to the scheme, the SSA is combined with SVR to construct a model, and the optimal nuclear parameters and penalty factors can be efficiently selected, so that the prediction efficiency is greatly improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. Like elements or portions are generally identified by like reference numerals throughout the several figures. In the drawings, elements or portions thereof are not necessarily drawn to scale.
FIG. 1 is a flow chart of a method for predicting the opening and closing deformation of a pipe joint of a immersed tunnel in one embodiment;
FIG. 2 is a simplified flow chart of a method for predicting the opening and closing deformation of a pipe joint of a immersed tunnel in one embodiment;
FIG. 3 is a schematic diagram of arch top settlement equidistant data in one embodiment;
FIG. 4 is an exploded view of a tubing joint tensioning amount sequence in one embodiment;
FIG. 5 is a schematic representation of trend component fitting and prediction results in one embodiment;
FIG. 6 is a schematic diagram of a periodic component prediction result in one embodiment;
FIG. 7 is a diagram of residual component prediction results in one embodiment;
FIG. 8 is a schematic diagram of overall prediction results of opening and closing deformation in one embodiment;
FIG. 9 is a schematic structural diagram of a device for predicting the opening and closing deformation of a pipe joint of a immersed tunnel according to an embodiment;
fig. 10 is an internal structural view of a computer device in one embodiment.
Detailed Description
Embodiments of the technical scheme of the present invention will be described in detail below with reference to the accompanying drawings. The following examples are only for more clearly illustrating the technical aspects of the present invention, and thus are merely examples, and are not intended to limit the scope of the present invention.
It is noted that unless otherwise indicated, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.
In one embodiment, as shown in fig. 1, a method for predicting open-close deformation of a pipe joint of a immersed tunnel is provided, which includes the following steps:
s110, opening and closing deformation data and environment data of the joint of the pipe joint to be predicted are obtained, wherein the environment data at least comprise temperature data and water depth data.
Specifically, opening and closing deformation data of a pipe joint to be predicted and temperature and water level information of the opening and closing deformation data are obtained.
In one embodiment, after step S110, further includes: judging whether the opening and closing deformation data set is equidistant interval data or not; if yes, carrying out a subsequent decomposition step of the opening and closing deformation sequence by adopting a singular spectrum analysis method; if not, performing equidistant processing by adopting a cubic spline interpolation method to obtain an equidistant opening and closing deformation data set, and then performing a decomposition step on the opening and closing deformation data by adopting a singular spectrum analysis method. In particular, in practical engineering, however, there are many uncertain factors in the data acquisition process, and the monitored data often has a missing phenomenon, and if the monitored data is not reasonable to directly analyze, it should be firstly determined whether the monitored data is equidistant data after the relevant data is taken. If yes, go to step S120 directly; if not, performing equidistant processing by adopting a cubic spline difference method, and then performing step S120.
S120, decomposing the opening-closing deformation data by adopting a singular spectrum analysis method, and reconstructing according to the components obtained by decomposition to respectively obtain a trend component, a periodic component and a residual component.
Specifically, a singular spectrum analysis method is adopted to decompose the opening-closing deformation sequence, and a trend component, a periodic component and a residual component are respectively obtained by reconstruction according to the component obtained by decomposition. Can be expressed as:
y(t)＝T(t)+S(t)+I(t) (1)
wherein: y (T) is the joint opening and closing deformation value, T (T) is the trend component, S (T) is the periodic component, and I (T) is the residual component. The trend component is controlled by the factors of aging, permanent load and the like of the structure, the period component is controlled by the factors of temperature, tidal load and the like, and the residual component is controlled by the factors of sudden disasters, monitoring conditions and the like.
S130, dividing the trend component, the periodic component and the residual component into a training set and a testing set which are respectively corresponding.
Specifically, the trend component, the periodic component and the residual component are respectively divided into two groups, namely a training set and a testing set. The training set is used for fitting functions or training models, and the testing set is used for verifying the accuracy of the fitted functions or the prediction accuracy and generalization capability of the models.
And S140, fitting and predicting the trend component by adopting a least square method to obtain a trend component prediction result.
In one embodiment, step S140 includes: fitting a least square function on the basis of a training set of trend components to obtain a target least square function; and carrying out trend component prediction according to the target least square function to obtain a trend component prediction result.
Specifically, a least square method is adopted to fit and predict trend components. And then predicting the trend component by using the fitted function to obtain a trend component prediction result.
S150, training the constructed first SSA-SVR model by using a training set of periodic components to obtain a first target SSA-SVR model, and predicting the periodic components by using the first target SSA-SVR model to obtain a periodic component prediction result.
Specifically, the training set in the periodic component is used for training an SSA-SVR model, wherein a Sparrow Search Algorithm (SSA), SVR and SVR regression algorithm are called support vector regression or SVR, and the support vector regression is a supervised learning algorithm used for predicting discrete values. And then predicting the periodic component by using the trained SSA-SVR model to obtain a periodic component prediction result. Wherein, the SVR model adopts RBF kernel function. RBF (Radial Basis Function) kernel function: the function of the kernel function is a function of converting from a low-dimensional space to a high-dimensional space because the problem of being unable to be classified linearly in the low-dimensional space to the high-dimensional space can be classified linearly. The solution is that after the point of the low-dimensional space is mapped to the point of the high-dimensional space, the inner product of the two points is a kernel function.
In one embodiment, step S150 includes: respectively carrying out normalization processing on the training set, the prediction set and the corresponding environmental data of the periodic component; establishing a first SVR model, taking the opening and closing deformation data and the environmental data of 1 day before the predicted day as input, and taking the opening and closing deformation data of the predicted day as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of nuclear parameters, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the first SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of a finder, a joiner and an early warning person in an SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; and using the obtained optimal combination value of the penalty factors and the nuclear parameters in a first SVR model, predicting a test set of the periodic components, and performing inverse normalization operation to obtain a prediction result of the periodic components.
Specifically, a Sparrow Search Algorithm (SSA) and an SVR model are combined, so that the sparrow search algorithm is called an SSA-SVR model, (1) tidal load is simulated through water depth change, and a training set and a prediction set of periodic components in open-close deformation data and temperature data and water depth data corresponding to each part are respectively normalized, so that the training set and the prediction set are positioned between [ -1 and 1] to eliminate the influence of data dimension, and the algorithm is easier to converge. (2) The temperature, water depth and opening and closing deformation data of the first 1 day of the prediction day are taken as input, and the opening and closing deformation data of the prediction day are taken as output. (3) Initializing sparrow search algorithm parameters, wherein the parameters mainly comprise maximum iteration times, population scale, penalty factor C, a value range of kernel parameter g, finder proportion, safety value and the like. (4) Each sparrow individual represents two parameters g and C of the SVR model, the mean square error of the predicted output value and the actual output value is used as the fitness function of the sparrow search algorithm, the fitness values of all sparrows are calculated, the current global optimal solution is found, and the position of the current global optimal solution is determined. (5) Updating the positions of a finder, a jointer and an early warning person, checking the individual fitness value after the position updating, and determining the position of the optimal sparrow individual, wherein the finder, the jointer and the early warning person are parameters of a sparrow search algorithm; (6) And judging whether the suspension criterion is met, namely obtaining a global optimal solution or meeting the maximum iteration times. If so, outputting the optimal sparrow position as the optimal value of the parameters g and C, otherwise, continuing to execute the step (5). (7) And using the obtained g and C optimal combination value in an SVR model, predicting a test set of the periodic component, and performing inverse normalization operation to obtain a prediction result of the periodic component.
S160, training the constructed second SSA-SVR model by using a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by using the second target SSA-SVR model to obtain a residual component prediction result.
Specifically, training the SSA-SVR model by adopting a training set in the residual components, and then predicting the residual components by utilizing the trained SSA-SVR model to obtain a residual component prediction result. Wherein, the SVR model adopts RBF kernel function. RBF (Radial Basis Function) kernel function: the function of the kernel function is a function of converting from a low-dimensional space to a high-dimensional space because the problem of being unable to be classified linearly in the low-dimensional space to the high-dimensional space can be classified linearly. The solution is that after the point of the low-dimensional space is mapped to the point of the high-dimensional space, the inner product of the two points is a kernel function.
In one embodiment, step S160 includes: respectively carrying out normalization processing on the training set and the prediction set of the residual components; establishing a second SVR model, wherein the historical residual component value of 3-6 days before the predicted day in the residual component is taken as input, and the residual component value of the predicted day is taken as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of nuclear parameters, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the second SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of a finder, a joiner and an early warning person in an SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; and using the obtained optimal combination value of the penalty factors and the kernel parameters in a second SVR model, predicting a test set of the residual components, and performing inverse normalization operation to obtain a prediction result of the residual components.
Specifically, (1) in actual engineering, it is difficult to obtain the quantized value of the influence factor of the residual component, so, starting from the data of the residual component, the response relationship between the history information and the future information is constructed, and at the same time, the training set and the prediction set of the residual component in the tensor deformation data are respectively normalized, so that the training set and the prediction set are positioned between [ -1,1] to eliminate the influence of the data dimension, and the algorithm is easier to converge. (2) The residual components of 3-6 days before the predicted day are taken as input, the residual components of the predicted day are taken as output, and in actual use, the residual components of 3 days before the predicted day are taken as input. (3) Initializing sparrow search algorithm parameters, wherein the parameters mainly comprise maximum iteration times, population scale, penalty factor C, a value range of kernel parameter g, finder proportion, safety value and the like. (4) Each sparrow individual represents two parameters g and C of the SVR model, the mean square error of the predicted output value and the actual output value is used as the fitness function of the sparrow search algorithm, the fitness values of all sparrows are calculated, the current global optimal solution is found, and the position of the current global optimal solution is determined. (5) Updating the positions of a finder, a jointer and an early warning person, checking the individual fitness value after the position updating, and determining the position of the optimal sparrow individual; (6) And judging whether the suspension criterion is met, namely obtaining a global optimal solution or meeting the maximum iteration times. If so, outputting the optimal sparrow position as the optimal value of the parameters g and C, otherwise, continuing to execute the step (5). (7) And using the obtained g and C optimal combination value in an SVR model, predicting a test set of the residual component, and performing inverse normalization operation to obtain a prediction result of the residual component.
And S170, accumulating and calculating a trend component prediction result, a periodic component prediction result and a residual component prediction result based on a time sequence superposition principle to obtain an integral prediction value of the pipe joint opening and closing deformation.
Specifically, based on a time sequence superposition principle, a trend component prediction result, a periodic component prediction result and a residual component prediction result are accumulated and calculated to obtain a pipe joint opening and closing deformation integral prediction value.
In one embodiment, after step S170, further includes: and quantitatively evaluating the prediction result by adopting the determination coefficient, the average absolute error and the root mean square error. In particular, the method comprises the steps of,by determining the coefficient R 2 The prediction results are quantitatively evaluated by mean absolute error MAE and root mean square error RMSE. R is R 2 The value is (0, 1), and the larger the value is, the higher the degree that the predicted value is consistent with the actual value is; MAE can reflect the actual situation of the predicted value error, RMSE is extremely sensitive to larger error, and can reflect the precision of prediction, and the smaller the two indexes are, the better. The calculation formula is as follows:
wherein:
predicted value of joint opening and closing deformation, y i For the actual value of the opening and closing deformation of the joint +.>
And N is the group number of the data, which is the average value of the actual values of the opening and closing deformation of the joint.
In one embodiment, a specific example is performed according to the method described above, as shown in fig. 3, to provide data of opening and closing deformation, temperature and water depth after equidistant treatment. And decomposing the open-close data at the joint by adopting a singular spectrum analysis method, wherein the window length is set to 365, the 1 st component is a trend component, the 2 nd to 40 th components are periodic components, and the 41 st to 365 th components are residual components. As shown in fig. 4, the trend component T, the periodic component S, and the residual component I obtained by decomposing the pipe joint opening amount monitoring data by the singular spectrum analysis method are shown. And fitting the training set of the trend components by adopting a least square method, and predicting the test set of the trend components by utilizing a fitting function. The linear equation and the precision of the fitting are shown in table 1, and the fitting and predicting results are shown in fig. 5.
TABLE 1 trend component fitting equation and accuracy
And (5) predicting a periodic component. The 22 th day of 4 months in 2020 is taken as a node, the training set is taken before the node, and the training set is divided into the test set after the node. Based on a Python programming language building model, the maximum iteration number of the SSA algorithm is set to 300, the population size is set to 50, the upper and lower boundaries of the optimizing parameter range are respectively 1000 and 0.001, the ratio of discoverers is 0.7, the ratio of early warning persons is 0.2, and the safety value is set to 0.6. The SVR model selects RBF kernel function. To further verify the validity and accuracy of the SSA-SVR model, the conventional SVR model was used as a comparison, the prediction results are shown in fig. 6, and the prediction accuracy and error are shown in table 2.
TABLE 2 precision and error of periodic component prediction
As can be seen in fig. 6: it can be obviously seen that the SSA-SVR optimization model has better prediction effect than the traditional SVR model, because the sparrow search algorithm has stronger global optimizing capability, the optimal penalty factor and kernel function parameters of the model can be obtained, and therefore a better prediction result is obtained. By further analyzing the quantitative evaluation indexes in the table 2, the SSA-SVR model obtains an optimal result, compared with the SVR models, the RMSE and the MAE are respectively reduced to 0.1946 and 0.1581 from 0.2904 and 0.2440, the RMSE is reduced by 28.19 percent and 27.38 percent, and the actual conditions that the RMSE is sensitive to larger errors and the M AE can reflect the predicted value errors can be seen from the calculation formulas of the two indexes, so that the SSA-SVR model has higher stability and prediction precision and is suitable for predicting the periodic components.
And (5) predicting residual components. With 22 days of 4 months 2020 as nodes, the nodes were preceded by a training set and then divided into test sets, the remaining components 3 days before the predicted day were selected as inputs in this example. Based on a Python programming language building model, the maximum iteration number of the SSA algorithm is set to 300, the population size is set to 50, the upper and lower boundaries of the optimizing parameter range are respectively 1000 and 0.001, the ratio of discoverers is 0.7, the ratio of early warning persons is 0.2, and the safety value is set to 0.6. The SVR model selects RBF kernel function. To further verify the validity and accuracy of the SSA-SVR model, the conventional SVR model was used as a comparison, the prediction results are shown in fig. 7, and the prediction accuracy and error are shown in table 3.
TABLE 3 residual component prediction precision and error
As can be seen from fig. 7: the SSA-SVR optimization model has better prediction effect than the traditional SVR model as a whole, but obviously deviates from an actual value at a curve turning position, and the traditional SVR model also has the phenomenon, and compared with the SSA-SVR model, the error is larger, which shows that the SVR model has stronger sensitivity to data mutation and poorer anti-interference capability than the SSA-SVR model. Thus, the SSA-SVR model presented herein is still reliable to apply in the prediction of the residual component. Further analysis in combination with quantitative evaluation index in Table 2, SSA-SVR model predicts R 2 = 0.9308, rmse= 0.0401mm, mae=0.0305 mm, svr model predictive R 2 The 3 evaluation indexes of the model are optimized by an SSA-SVR optimization model, namely 0.8919, RMSE= 0.0501mm and M AE=0.0368 mm, which further illustrates that the SSA-SVR model has higher prediction precision.
And (3) superposing the trend component, the periodic component and the residual component to obtain an optimal prediction result of the two-joint stretching quantity, as shown in fig. 8. As can be seen from FIG. 8, the combination model provided herein is used for obtaining a good prediction effect in the open-close deformation of the immersed tube tunnel joint, and the prediction result can approximately represent the actual open-close deformation trend of the joint. R is R 2 0.9792; RMSE is 0.1900mm; the MAE is 0.1587mm, and has higher precision and stability. The main reasons are the following 3 points: 1) The trend component, the period component and the residual component decomposed by adopting the singular spectrum analysis method are characterized by taking into considerationThe deformation characteristics of the components select a proper prediction method, so that a better prediction result is obtained; 2) The sparrow search algorithm has stronger global optimizing capability, and can effectively perform SVR model parameter optimizing; 3) The RFB kernel function has strong nonlinear excavation capability, and can more accurately predict the opening and closing deformation condition of the joint.
In one embodiment, as shown in fig. 9, there is provided a device for predicting the opening and closing deformation of a pipe joint of a immersed tunnel, comprising: a data acquisition module 210, a data decomposition module 220, and a data prediction module 230, wherein,
the data acquisition module 210 is configured to acquire opening and closing deformation data and environmental data at a joint of a pipe joint to be predicted, where the environmental data at least includes temperature data and water depth data;
the data decomposition module 220 is configured to decompose the opening-closing deformation data by using a singular spectrum analysis method, and reconstruct according to the component obtained by the decomposition to obtain a trend component, a periodic component and a residual component respectively; dividing the trend component, the period component and the residual component into a training set and a testing set which are respectively corresponding to each other;
the data prediction module 230 is configured to fit and predict the trend component by using a least square method, so as to obtain a trend component prediction result; training the constructed first SSA-SVR model by adopting a training set of the periodic component to obtain a first target SSA-SVR model, and carrying out periodic component prediction by adopting the first target SSA-SVR model to obtain a periodic component prediction result; training the constructed second SSA-SVR model by adopting a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by adopting the second target SSA-SVR model to obtain a residual component prediction result; and accumulating and calculating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on a time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing configuration templates and can also be used for storing target webpage data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to realize a prediction method of open-close deformation of the immersed tube tunnel joint.
It will be appreciated by those skilled in the art that the structure shown in fig. 10 is merely a block diagram of some of the structures associated with the present application and is not limiting of the computer device to which the present application may be applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, there is also provided a storage medium storing a computer program comprising program instructions which, when executed by a computer, cause the computer to perform a method as described in the preceding embodiment, the computer being part of a device for predicting a pipe joint opening and closing deformation of a immersed tube tunnel as mentioned above.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented in a general purpose computing device, they may be centralized on a single computing device, or distributed across a network of computing devices, or they may alternatively be implemented in program code executable by computing devices, such that they may be stored on a computer storage medium (ROM/RAM, magnetic or optical disk) for execution by computing devices, and in some cases, the steps shown or described may be performed in a different order than what is shown or described herein, or they may be individually manufactured as individual integrated circuit modules, or a plurality of modules or steps in them may be manufactured as a single integrated circuit module. Therefore, the present invention is not limited to any specific combination of hardware and software.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention, and are intended to be included within the scope of the appended claims and description.
Claims (7)
1. A prediction method for open-close deformation of a immersed tunnel pipe joint is characterized by comprising the following steps:
acquiring opening and closing deformation data and environment data of a pipe joint to be predicted, wherein the environment data at least comprises temperature data and water depth data;
decomposing the opening and closing deformation data by adopting a singular spectrum analysis method, and reconstructing according to the components obtained by decomposition to respectively obtain a trend component, a periodic component and a residual component;
dividing the trend component, the period component and the residual component into a training set and a testing set which are respectively corresponding to each other;
fitting and predicting the trend component by adopting a least square method to obtain a trend component prediction result;
training the constructed first SSA-SVR model by adopting a training set of periodic components to obtain a first target SSA-SVR model, and carrying out periodic component prediction by adopting the first target SSA-SVR model to obtain a periodic component prediction result;
respectively carrying out normalization processing on the training set, the prediction set and the corresponding environmental data of the periodic component;
establishing a first SVR model, taking the opening and closing deformation data and the environmental data of 1 day before the predicted day as input, and taking the opening and closing deformation data of the predicted day as output;
initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value;
taking the mean square error between the predicted value and the actual value pre-output by the first SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution;
updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm;
judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution;
the obtained optimal combination value of the penalty factors and the nuclear parameters is used in the first SVR model, a test set of the periodic components is predicted, and inverse normalization operation is carried out to obtain a prediction result of the periodic components;
training the constructed second SSA-SVR model by adopting a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by adopting the second target SSA-SVR model to obtain a residual component prediction result;
respectively carrying out normalization processing on the training set and the prediction set of the residual components;
establishing a second SVR model, wherein the residual component value of 3-6 days before the predicted day in the residual component is taken as input, and the residual component value of the predicted day is taken as output;
initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value;
taking the mean square error between the predicted value and the actual value pre-output by the second SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution;
updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm;
judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution;
the obtained optimal combination value of the penalty factors and the kernel parameters is used in the second SVR model, a test set of the residual components is predicted, and inverse normalization operation is carried out to obtain a prediction result of the residual components;
and accumulating and calculating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on a time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation.
2. The method according to claim 1, wherein the step of obtaining the opening and closing deformation data and the environmental data at the joint of the pipe section to be predicted, the environmental data including at least temperature data and water depth data further includes:
judging whether the opening and closing deformation data set is equidistant interval data or not;
if yes, carrying out a subsequent decomposition step of the opening and closing deformation sequence by adopting a singular spectrum analysis method;
if not, performing equidistant processing by adopting a cubic spline interpolation method to obtain an equidistant opening and closing deformation data set, and then performing a decomposition step on the opening and closing deformation data by adopting a singular spectrum analysis method.
3. The method of claim 1, wherein the step of fitting and predicting the trend component using a least squares method to obtain a trend component prediction result comprises:
fitting a least square function based on the training set of the trend component to obtain a target least square function;
and carrying out trend component prediction according to the target least square function to obtain a trend component prediction result.
4. The method according to claim 1, wherein after the step of accumulating the trend component prediction result, the periodic component prediction result, and the residual component prediction result based on a time sequence superposition principle to obtain the overall prediction value of the pipe joint opening and closing deformation, the method comprises:
and quantitatively evaluating the prediction result by adopting the determination coefficient, the average absolute error and the root mean square error.
5. The utility model provides a immersed tube tunnel tube coupling connects prediction unit who opens and shuts deformation which characterized in that includes: a data acquisition module, a data decomposition module and a data prediction module, wherein,
the data acquisition module is used for acquiring opening and closing deformation data and environment data of a pipe joint to be predicted, wherein the environment data at least comprises temperature data and water depth data;
the data decomposition module is used for decomposing the opening and closing deformation data by adopting a singular spectrum analysis method, and reconstructing according to the components obtained by decomposition to respectively obtain a trend component, a periodic component and a residual component; dividing the trend component, the period component and the residual component into a training set and a testing set which are respectively corresponding to each other;
the data prediction module is used for fitting and predicting the trend component by adopting a least square method to obtain a trend component prediction result; training the constructed first SSA-SVR model by adopting a training set of periodic components to obtain a first target SSA-SVR model, and carrying out periodic component prediction by adopting the first target SSA-SVR model to obtain a periodic component prediction result; respectively carrying out normalization processing on the training set, the prediction set and the corresponding environmental data of the periodic component; establishing a first SVR model, taking the opening and closing deformation data and the environmental data of 1 day before the predicted day as input, and taking the opening and closing deformation data of the predicted day as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the first SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; the obtained optimal combination value of the penalty factors and the nuclear parameters is used in the first SVR model, a test set of the periodic components is predicted, and inverse normalization operation is carried out to obtain a prediction result of the periodic components; training the constructed second SSA-SVR model by adopting a training set of the residual components to obtain a second target SSA-SVR model, and predicting the residual components by adopting the second target SSA-SVR model to obtain a residual component prediction result; respectively carrying out normalization processing on the training set and the prediction set of the residual components; establishing a second SVR model, wherein the residual component value of 3-6 days before the predicted day in the residual component is taken as input, and the residual component value of the predicted day is taken as output; initializing parameters of an SSA algorithm, wherein the parameters at least comprise maximum iteration times, population scale, penalty factors, a value range of a kernel parameter, a finder proportion and a safety value; taking the mean square error between the predicted value and the actual value pre-output by the second SVR model as the fitness function of the SSA algorithm, calculating the fitness values of all SSAs, determining the optimal solution in the current SSA algorithm and determining the position of the optimal solution; updating the positions of discoverers, joiners and early warning persons in the SSA algorithm, checking individual fitness values after the positions are updated, and determining the position of an optimal solution in the SSA algorithm; judging whether the SSA algorithm obtains a global optimal solution or meets the maximum iteration times, if so, outputting the position of the optimal solution in the SSA algorithm as an optimal value of a penalty factor and a kernel parameter, and if not, repeatedly executing the step of updating the position to determine the optimal solution; the obtained optimal combination value of the penalty factors and the kernel parameters is used in the second SVR model, a test set of the residual components is predicted, and inverse normalization operation is carried out to obtain a prediction result of the residual components; and accumulating and calculating the trend component prediction result, the periodic component prediction result and the residual component prediction result based on a time sequence superposition principle to obtain the integral prediction value of the pipe joint opening and closing deformation.
6. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 4 when the computer program is executed.
7. A storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the method of any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210904682.3A CN115217152B (en) | 2022-07-29 | 2022-07-29 | Method and device for predicting open-close deformation of immersed tunnel pipe joint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210904682.3A CN115217152B (en) | 2022-07-29 | 2022-07-29 | Method and device for predicting open-close deformation of immersed tunnel pipe joint |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115217152A CN115217152A (en) | 2022-10-21 |
| CN115217152B true CN115217152B (en) | 2023-05-16 |
Family
ID=83614765
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210904682.3A Active CN115217152B (en) | 2022-07-29 | 2022-07-29 | Method and device for predicting open-close deformation of immersed tunnel pipe joint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115217152B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117349602A (en) * | 2023-12-06 | 2024-01-05 | 江西省水投江河信息技术有限公司 | Water conservancy facility operation state prediction method, system and computer |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112903091A (en) * | 2021-01-19 | 2021-06-04 | 昆明理工大学 | Elevator guide shoe fault diagnosis method based on VMD-SVD and SSA-SVM |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106407581A (en) * | 2016-09-28 | 2017-02-15 | 华中科技大学 | Intelligent prediction method for ground surface settlement induced by subway tunnel construction |
| CN110442979B (en) * | 2019-08-08 | 2021-04-13 | 山东大学 | BP neural network-based shield construction tunnel total deformation prediction method and system |
| CN110969096A (en) * | 2019-11-07 | 2020-04-07 | 国电南京自动化股份有限公司 | Motor fault mode diagnosis method based on particle swarm optimization support vector machine |
| CN111270713A (en) * | 2020-02-10 | 2020-06-12 | 大连科迈尔防腐科技有限公司 | Ultrasonic thickness measurement probe mounting structure, tunnel immersed tube and mounting method |
| AU2020100405A4 (en) * | 2020-03-17 | 2020-04-30 | Qingdao university of technology | A slop risk comprehensive assessment method based on slope failures forms |
| CN111709569B (en) * | 2020-06-11 | 2023-09-22 | 国网冀北电力有限公司电力科学研究院 | Wind power plant output power prediction correction method and device |
| CN113472851A (en) * | 2021-05-27 | 2021-10-01 | 中交天津港湾工程研究院有限公司 | Automatic monitoring management system for immersed tube tunnel cloud |
| CN113484785A (en) * | 2021-07-07 | 2021-10-08 | 李忠 | Battery health state estimation and residual life prediction method based on sparrow search and least square support vector machine |
| CN113591700A (en) * | 2021-07-30 | 2021-11-02 | 兰州大学 | Prediction method for potential landslide and river blockage |
| CN113569324A (en) * | 2021-08-03 | 2021-10-29 | 招商局重庆交通科研设计院有限公司 | Slope deformation monitoring abnormal data analysis and optimization method |
-
2022
- 2022-07-29 CN CN202210904682.3A patent/CN115217152B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112903091A (en) * | 2021-01-19 | 2021-06-04 | 昆明理工大学 | Elevator guide shoe fault diagnosis method based on VMD-SVD and SSA-SVM |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115217152A (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | A compound framework for wind speed forecasting based on comprehensive feature selection, quantile regression incorporated into convolutional simplified long short-term memory network and residual error correction | |
| Galelli et al. | An evaluation framework for input variable selection algorithms for environmental data-driven models | |
| CN112990556A (en) | User power consumption prediction method based on Prophet-LSTM model | |
| CN110119854A (en) | Voltage-stablizer water level prediction method based on cost-sensitive LSTM Recognition with Recurrent Neural Network | |
| Song et al. | Value of information analysis in non-stationary stochastic decision environments: A reliability-assisted POMDP approach | |
| CN111784061B (en) | Training method, device and equipment for power grid engineering cost prediction model | |
| CN117175588B (en) | Space-time correlation-based electricity load prediction method and device | |
| CN111325403A (en) | Method for predicting remaining life of electromechanical equipment of highway tunnel | |
| CN115217152B (en) | Method and device for predicting open-close deformation of immersed tunnel pipe joint | |
| CN114358435A (en) | Pollution source-water quality prediction model weight influence calculation method of two-stage space-time attention mechanism | |
| CN115510748A (en) | Landslide displacement prediction method based on variational modal decomposition and CNN-GRU | |
| CN115687322A (en) | Water quality time series missing data completion method based on encoder-decoder and autoregressive generated countermeasure network | |
| CN113988210A (en) | Method and device for restoring distorted data of structure monitoring sensor network and storage medium | |
| CN113151842B (en) | Method and device for determining conversion efficiency of wind-solar complementary water electrolysis hydrogen production | |
| Zou et al. | Reliability analysis of Digital Instrumentation and Control software system | |
| CN112767692A (en) | Short-term traffic flow prediction system based on SARIMA-GA-Elman combined model | |
| CN112911533A (en) | Temperature detection system based on remove end App | |
| CN115330085A (en) | Wind speed prediction method based on deep neural network and without future information leakage | |
| CN111859798A (en) | Flow industrial fault diagnosis method based on bidirectional long-time and short-time neural network | |
| Cai et al. | A self-attention-LSTM method for dam deformation prediction based on CEEMDAN optimization | |
| Zhao et al. | An interpretable ultra-short-term wind power prediction model based on the feature matrix reconstruction through regression trees | |
| CN117829822B (en) | Power transformer fault early warning method and system | |
| CN117977584B (en) | Power load probability prediction method, system, medium, device and program product | |
| CN114741948B (en) | Aero-engine degradation trend prediction method based on residual stacking convolution network of sequence reconstruction | |
| CN116108361B (en) | Intelligent oil quality detection method and monitoring system |
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 |