CN117093919A - Geotechnical engineering geological disaster prediction method and system based on deep learning - Google Patents
Geotechnical engineering geological disaster prediction method and system based on deep learning Download PDFInfo
- Publication number
- CN117093919A CN117093919A CN202311353388.9A CN202311353388A CN117093919A CN 117093919 A CN117093919 A CN 117093919A CN 202311353388 A CN202311353388 A CN 202311353388A CN 117093919 A CN117093919 A CN 117093919A
- Authority
- CN
- China
- Prior art keywords
- geological disaster
- geological
- information
- steps
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 84
- 238000013135 deep learning Methods 0.000 title claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 claims abstract description 29
- 230000004044 response Effects 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 24
- 230000000737 periodic effect Effects 0.000 claims abstract description 19
- 230000004927 fusion Effects 0.000 claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 15
- 230000008859 change Effects 0.000 claims abstract description 10
- 210000002569 neuron Anatomy 0.000 claims description 35
- 238000007781 pre-processing Methods 0.000 claims description 20
- 238000012544 monitoring process Methods 0.000 claims description 17
- 230000005540 biological transmission Effects 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 13
- 230000006399 behavior Effects 0.000 claims description 11
- 238000005516 engineering process Methods 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 9
- 238000013461 design Methods 0.000 claims description 8
- 230000007246 mechanism Effects 0.000 claims description 8
- 238000013468 resource allocation Methods 0.000 claims description 8
- 230000009471 action Effects 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 claims description 3
- 238000011160 research Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000004140 cleaning Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a geotechnical engineering geological disaster prediction method and system based on deep learning, wherein the method comprises the following steps: geological information acquisition and processing, signal processing, geological disaster prediction, geological disaster classification, and geological disaster response and early warning. The invention belongs to the technical field of geological engineering, in particular to a geotechnical engineering geological disaster prediction method and system based on deep learning, wherein the scheme adopts a dynamic system reconstruction method to process historical geological disaster data, reconstruct a weight matrix, and convert a one-dimensional time sequence into a track in a high-dimensional phase space so as to obtain periodic change of geological disasters; the BP neural network is adopted to process multidimensional features, self-adaptive learning is carried out, and after the neural network training is completed, the BP neural network rapidly responds in practical application, so that automatic geological disaster prediction is realized; and combining the evidences by adopting logic or fusion rules, and selecting the category with the highest confidence as a final geological disaster classification result, so that the robustness is high and the fault tolerance is strong.
Description
Technical Field
The invention belongs to the technical field of geological engineering, and particularly relates to a geotechnical engineering geological disaster prediction method and system based on deep learning.
Background
Geotechnical engineering geological disaster prediction is a technical method for predicting and early warning geological disasters by comprehensively investigating, analyzing and evaluating geological environments, and aims to rapidly and accurately identify and analyze the geological disasters, reduce artificial investigation burden, provide important support for geotechnical engineering planning, design and construction and reduce risks and losses of the geological disasters. However, the existing geotechnical engineering geological disaster prediction method has the technical problems that the internal evolution rule of a geological disaster system is complex and nonlinear association is caused, so that the occurrence periodicity rule of the geological disaster is difficult to predict; the technical problems that geological disaster prediction involves a plurality of characteristic variables and a large number of data samples, and the prediction effect and model generalization capability are poor exist; the method has the technical problems that the uncertainty problem is faced in the geological disaster prediction process, the geological disaster prediction process is comprehensively influenced by various factors, the geological environment is varied variously, and the geological disaster classification result has uncertainty.
Disclosure of Invention
Aiming at the situation, in order to overcome the defects of the prior art, the invention provides a geotechnical engineering geological disaster prediction method and a geotechnical engineering geological disaster prediction system based on deep learning, aiming at the technical problems that the internal evolution rule of a geological disaster system is complex and nonlinear association, so that the periodical rule of geological disaster occurrence is difficult to predict, a dynamic system reconstruction method is adopted to process historical geological disaster data, a weight matrix is reconstructed, a one-dimensional time sequence is converted into a track in a high-dimensional phase space, and the dynamic behavior, nonlinear characteristics and periodical change of an analysis system are analyzed by the reconstructed phase space track, so that the periodical change of the geological disaster is obtained; aiming at the technical problems that geological disaster prediction involves a plurality of characteristic variables and a large number of data samples, and the prediction effect and model generalization capability are poor, a BP neural network is adopted to process multidimensional characteristics, self-adaptive learning is carried out, and the BP neural network is trained to respond quickly in practical application, so that automatic geological disaster prediction is realized; aiming at the technical problems that uncertainty is faced in the geological disaster prediction process, the geological disaster prediction process is comprehensively influenced by various factors, the geological environment is varied, the geological disaster classification result has uncertainty, the logic or fusion rule is adopted to merge evidence items, and the category with the highest confidence is selected as the final geological disaster classification result, so that the robustness is high, and the fault tolerance is strong.
The technical scheme adopted by the invention is as follows: the geotechnical engineering geological disaster prediction method based on deep learning provided by the invention comprises the following steps:
step S1: the geological information acquisition and processing, specifically, the geological information is acquired through a remote sensing technology and preprocessed;
step S2: the method comprises the steps of signal processing, namely determining embedding dimension and delay time, processing historical geological disaster data by using a dynamic system reconstruction method, reconstructing a weight matrix, converting a one-dimensional time sequence into a track in a high-dimensional phase space, and analyzing dynamic behaviors, nonlinear characteristics and periodic changes of an analysis system by using the reconstructed phase space track so as to obtain the periodic changes of geological disasters;
step S3: the geological disaster prediction is specifically to collect and preprocess a historical geological disaster data set, design a BP neural network structure, train and optimize a geological disaster prediction model, and then input a real-time geological information data set for practical application;
step S4: the method comprises the steps of geological disaster classification, specifically, carrying out weight adjustment on confidence coefficient of evidence items, fusing all evidence item information by using logic or fusion rules, merging the evidence items, and obtaining a final geological disaster classification result according to the merged confidence coefficient;
step S5: the geological disaster response and early warning are optimized, and particularly the response mechanism is optimized, so that resource allocation and information sharing are accelerated, and geological disaster monitoring equipment, an optimized data transmission network and standard early warning information are continuously improved.
Further, in step S1, the geological information acquisition and processing includes the following steps:
step S11: the remote sensing technology is applied, the remote sensing image is utilized to provide large-scale information of terrains, landforms, vegetation and water bodies, the geological features of the research area are comprehensively known, and a real-time geological information data set is obtained;
step S12: preprocessing data, namely preprocessing a collected real-time geological information data set, including data cleaning, abnormal value removal and data normalization;
step S13: and (3) feature engineering, wherein a statistical method is used for analyzing correlation coefficients between each geological information feature and geological disaster prediction, and geological information features related to the geological disaster prediction are selected according to the correlation coefficients.
Further, in step S2, the signal processing includes the steps of:
step S21: data preparation, namely preparing a historical geological disaster data set, and ensuring that the historical geological disaster data set has enough time span and accuracy;
step S22: determining a delay time and an embedding dimension, wherein the delay time is a time interval for observing the same feature occurrence in a time sequence, and the embedding dimension is a minimum dimension for describing the dynamic feature of the system in a phase space;
step S23: the power system reconstruction method converts a one-dimensional time sequence into a corresponding Gao Weixiang space track, and comprises the following steps of:
step S231: inputting a historical geological disaster dataset: x= { X 1 ，x 2 ，...，x n N is the total number of geological disaster samples;
step S232: determining a neighborhood, determining the neighborhood by using Euclidean distance for each time point, selecting a group of surrounding neighborhood geological disaster samples, defining the number of the neighborhood geological disaster samples as k, and defining the index set of k neighborhood geological disaster samples as N;
step S233: reconstructing a weight matrix, distributing reconstruction weights for each neighborhood geological disaster sample, determining a linear relation between each neighborhood geological disaster sample and the ith geological disaster sample, defining the reconstruction weight matrix as Q, and constructing a weight vector Q ij ，Q ij Is the weight of the ith geological disaster sample and the jth neighborhood geological disaster sample in the reconstruction process, and calculates a weight vector Q ij So that the geological disaster sample x i Can be reconstructed from a linear combination of its neighboring samples using the following formula:
；
wherein x is i Is the ith geological disaster sample, x j Is the jth neighborhood geological disaster sample,is the vector x i －x j Absolute value of (2);
step S234: calculating a high-dimensional track in the phase space according to the reconstruction weight matrix, and decomposing the high-dimensional track in the phase space by using the characteristic values to obtain a low-dimensional representation of the high-dimensional track, so as to obtain a track after phase space reconstruction;
step S24: analyzing dynamic behaviors and nonlinear characteristics, namely analyzing the dynamic characteristics by calculating track indexes after phase space reconstruction, including variances and associated dimensions, and analyzing the nonlinear characteristics of a system by using a nonlinear analysis method;
step S25: and (3) periodically analyzing the change, namely periodically analyzing the phase space track, including spectrum analysis and wavelet analysis, to obtain the periodic change in the geological disaster signal.
Further, in step S3, the geological disaster prediction includes the steps of:
step S31: preprocessing geological information, namely collecting a historical geological disaster data set and preprocessing the data, wherein the preprocessing comprises data cleaning, abnormal value removal and data normalization;
step S32: designing a network structure, constructing a geological disaster prediction model, and designing a BP neural network structure according to the characteristics and the prediction targets of geological disasters, wherein the method comprises the following steps of:
step S321: the method comprises the steps of designing an input layer and an output layer, determining the dimension of input data and the total number of neurons of the input layer, wherein the dimension of the input data is determined by the total number of neurons of the input layer of a BP network, and the total number of neurons of the input layer and the total number of neurons of the output layer are determined according to the classification number of expected transmission signals;
step S322: the hidden layer design, use the network structure that only one hidden layer when designing BP neural network, reduce BP network complexity, reduce training time to reduce the error through increasing the quantity of hidden layer neuron, set up the neuron quantity according to experience, the formula that uses is as follows:
；
wherein p is the total number of neurons of the input layer, M is the total number of neurons of the hidden layer, q is the total number of neurons of the output layer, and b is a constant greater than 0 and less than 10;
step S33: the method comprises the steps of training a geological disaster prediction model, importing a historical geological disaster data set into the geological disaster prediction model, dynamically adjusting connection weights, repeatedly adjusting the connection weights among all levels in a network through forward propagation and reverse propagation, perfecting mapping relations of all levels, and adjusting the connection weights among all levels under the condition of adhering to an adjustment rule, wherein the adjustment rule is that all levels of neurons share output errors and are transmitted back to a hidden layer and an input layer step by step, so that the connection weights among all levels are adjusted;
step S34: training verification and optimization, verifying a geological disaster prediction model obtained through training, evaluating the performance and accuracy of the network according to the error between the prediction result and the actual result of the BP neural network on a verification set, and optimizing the geological disaster prediction model according to the evaluation result, wherein the method comprises the steps of adjusting a network structure, adjusting a learning rate and using a regularization method;
step S35: the real-time geological information data set is input into a geological disaster prediction model to realize automatic geological disaster prediction, and the real-time geological disaster prediction model is used for real-time geological disaster early warning and decision support.
Further, in step S4, the geological disaster classification includes the steps of:
step S41: defining evidence items, regarding the output of each neuron as one evidence item, representing the confidence level of occurrence of the geological disaster, and mapping the output to a probability value between [0,1] by an activation function in the neural network, wherein the probability value is the probability of occurrence of the geological disaster;
step S42: confidence weight adjustment is carried out on the confidence of each evidence item according to different neurons because of the complexity of the network and uncertainty in the learning process in the neural network and the conflict of classification results of different neurons, so that the accuracy of the classification results is improved;
step S43: selecting a fusion rule, carrying out logic or operation on probability values of all evidence items by using the logic or fusion rule to obtain a total probability value, mapping the total probability value between [0,1], and fusing information of all evidence items;
step S44: combining the evidence items, combining the evidence items subjected to weight adjustment into an evidence item set, and combining the evidence item set into a total evidence set to obtain a combined confidence level, wherein the total evidence set comprises all evidence items and corresponding confidence levels;
step S45: and (3) selecting the geological disaster category with the highest confidence as a final geological disaster classification result according to the combined confidence.
Further, in step S5, the geological disaster response and early warning includes the following steps:
step S51: optimizing a response mechanism, namely formulating event priority and a corresponding time response target, and accelerating resource allocation and information sharing;
step S52: improving monitoring equipment, in particular updating and upgrading the monitoring equipment and increasing the distribution density of monitoring points;
step S53: optimizing a data transmission network, specifically constructing a reliable and efficient data transmission network, and transmitting and processing monitoring data in real time by using an advanced communication technology;
step S54: standardizing early warning information, perfecting the following four parts in the early warning information: disaster type, risk level, area of impact, and action advice;
step S55: and (5) updating feedback, and listening to user feedback to continuously update the optimized early warning system and the information transmission channel.
The geotechnical engineering geological disaster prediction system based on deep learning comprises a geological information acquisition and processing module, a signal processing module, a geological disaster prediction module, a geological disaster classification module and a geological disaster response and early warning module;
the geological information acquisition and processing module is used for acquiring geological information through a remote sensing technology and preprocessing the geological information;
the signal processing module processes historical geological disaster data by using a dynamic system reconstruction method, converts a one-dimensional time sequence into a track in a high-dimensional phase space by selecting an embedding dimension and delay time, and analyzes dynamic behaviors, nonlinear characteristics and periodic changes of an analysis system by using the reconstructed phase space track so as to obtain the periodic changes of geological disasters;
the geological disaster prediction module is used for collecting and preprocessing a historical geological disaster data set, designing a BP neural network structure, training and optimizing a geological disaster prediction model, and inputting a real-time geological information data set for practical application;
the geological disaster classification module specifically carries out weight adjustment on the confidence coefficient of the evidence items, fuses all the evidence item information by utilizing logic or fusion rules, merges the evidence items, and obtains a final geological disaster classification result according to the merged confidence coefficient;
the geological disaster response and early warning module is particularly an optimized response mechanism, accelerates resource allocation and information sharing, and continuously improves geological disaster monitoring equipment, optimizes a data transmission network and standardizes early warning information.
The beneficial results obtained by adopting the scheme of the invention are as follows:
(1) Aiming at the technical problems that the internal evolution rule of a geological disaster system is complex and nonlinear associated, and the occurrence periodicity rule of the geological disaster is difficult to predict, a dynamic system reconstruction method is adopted to process historical geological disaster data, a weight matrix is reconstructed, a one-dimensional time sequence is converted into a track in a high-dimensional phase space, and the reconstructed phase space track is used for analyzing the dynamic behavior, nonlinear characteristics and periodicity change of an analysis system, so that the periodicity change of the geological disaster is obtained;
(2) Aiming at the technical problems that geological disaster prediction involves a plurality of characteristic variables and a large number of data samples, and the prediction effect and model generalization capability are poor, a BP neural network is adopted to process multidimensional characteristics, self-adaptive learning is carried out, and the BP neural network is trained to respond quickly in practical application, so that automatic geological disaster prediction is realized;
(3) Aiming at the technical problems that uncertainty is faced in the geological disaster prediction process, the geological disaster is influenced by the combination of a plurality of factors, the geological environment is varied, the geological disaster classification result has uncertainty, the evidence items are combined by adopting logic or fusion rules, and the geological disaster category with the highest confidence is selected as the final geological disaster classification result, so that the robustness is high, and the fault tolerance is strong.
Drawings
FIG. 1 is a schematic flow chart of a geotechnical engineering geological disaster prediction method based on deep learning;
FIG. 2 is a schematic diagram of a geotechnical engineering geological disaster prediction system based on deep learning, which is provided by the invention;
FIG. 3 is a flow chart of step S2;
FIG. 4 is a flow chart of step S3;
fig. 5 is a flow chart of step S4.
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all embodiments of the invention; all other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the description of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "top," "bottom," "inner," "outer," and the like indicate orientation or positional relationships based on those shown in the drawings, merely to facilitate description of the invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the invention.
Referring to fig. 1, the geotechnical engineering geological disaster prediction method based on deep learning provided by the invention comprises the following steps:
step S1: the geological information acquisition and processing, specifically, the geological information is acquired through a remote sensing technology and preprocessed;
step S2: the method comprises the steps of signal processing, namely processing historical geological disaster data by using a dynamic system reconstruction method, converting a one-dimensional time sequence into a track in a high-dimensional phase space by selecting an embedding dimension and delay time, and analyzing dynamic behaviors, nonlinear characteristics and periodic changes of an analysis system by using the reconstructed phase space track so as to obtain the periodic changes of geological disasters;
step S3: the geological disaster prediction is specifically to collect and preprocess a historical geological disaster data set, design a BP neural network structure, train and optimize a geological disaster prediction model, and input a real-time geological information data set for practical application;
step S4: the method comprises the steps of geological disaster classification, specifically, carrying out weight adjustment on confidence coefficient of evidence items, fusing all evidence item information by using logic or fusion rules, merging the evidence items, and obtaining a final geological disaster classification result according to the merged confidence coefficient;
step S5: the geological disaster response and early warning are optimized, and particularly the response mechanism is optimized, so that resource allocation and information sharing are accelerated, and geological disaster monitoring equipment, an optimized data transmission network and standard early warning information are continuously improved.
In a second embodiment, referring to fig. 1, the geological information acquisition and processing in step S1 includes the following steps:
step S11: the remote sensing technology is applied, the remote sensing image is utilized to provide large-scale information of terrains, landforms, vegetation and water bodies, the geological features of the research area are comprehensively known, and a real-time geological information data set is obtained;
step S12: preprocessing data, namely preprocessing a collected real-time geological information data set, including data cleaning, abnormal value removal and data normalization;
step S13: and (3) feature engineering, wherein a statistical method is used for analyzing correlation coefficients between each geological information feature and geological disaster prediction, and geological information features related to the geological disaster prediction are selected according to the correlation coefficients.
Embodiment three, referring to fig. 1 and 3, based on the above embodiment, in step S2, the signal processing includes the following steps:
step S21: data preparation, namely preparing a historical geological disaster data set, and ensuring that the historical geological disaster data set has enough time span and accuracy;
step S22: determining a delay time and an embedding dimension, wherein the delay time is a time interval for observing the same feature occurrence in a time sequence, and the embedding dimension is a minimum dimension for describing the dynamic feature of the system in a phase space;
step S23: the power system reconstruction method converts a one-dimensional time sequence into a corresponding Gao Weixiang space track, and comprises the following steps of:
step S231: inputting a historical geological disaster dataset: x= { X 1 ，x 2 ，...，x n N is the total number of geological disaster samples;
step S232: determining a neighborhood, determining the neighborhood by using Euclidean distance for each time point, selecting a group of surrounding neighborhood geological disaster samples, defining the number of the neighborhood geological disaster samples as k, and defining the index set of k neighborhood geological disaster samples as N;
step S233: reconstructing a weight matrix, distributing reconstruction weights for each neighborhood geological disaster sample, determining a linear relation between each neighborhood geological disaster sample and the ith geological disaster sample, defining the reconstruction weight matrix as Q, and constructing a weight vector Q ij ，Q ij Is the weight of the ith geological disaster sample and the jth neighborhood geological disaster sample in the reconstruction process, so that the geological disaster sample x i Can be reconstructed by linear combination of adjacent samples, and weight vector Q is calculated ij The formula used is as follows:
；
wherein x is i Is the ith geological disaster sample, x j Is the jth neighborhood geological disaster sample,is the vector x i －x j Absolute value of (2);
step S234: calculating a high-dimensional track in the phase space according to the reconstruction weight matrix, and decomposing the high-dimensional track in the phase space by using the characteristic values to obtain a low-dimensional representation of the high-dimensional track, so as to obtain a track after phase space reconstruction;
step S24: analyzing dynamic behaviors and nonlinear characteristics, namely analyzing the dynamic characteristics by calculating track indexes after phase space reconstruction, including variances and associated dimensions, and analyzing the nonlinear characteristics of a system by using a nonlinear analysis method;
step S25: and (3) periodically analyzing the change, namely periodically analyzing the phase space track, including spectrum analysis and wavelet analysis, to obtain the periodic change in the geological disaster signal.
By executing the operation, the dynamic system reconstruction method is adopted to process the historical geological disaster data, the weight matrix is reconstructed, the one-dimensional time sequence is converted into the track in the high-dimensional phase space, and the reconstructed phase space track is used for analyzing the dynamic behavior, the nonlinear characteristics and the periodic variation of the analysis system, so that the periodic variation of the geological disaster is obtained, and the technical problems that the internal evolution rule of the geological disaster system is complex and nonlinear association, and the periodic rule of the occurrence of the geological disaster is difficult to predict are solved.
Embodiment four, referring to fig. 1 and 4, based on the above embodiment, in step S3, the geological disaster prediction includes the following steps:
step S31: preprocessing geological information, namely collecting a historical geological disaster data set and preprocessing the data, wherein the preprocessing comprises data cleaning, abnormal value removal and data normalization;
step S32: designing a network structure, constructing a geological disaster prediction model, and designing a BP neural network structure according to the characteristics and the prediction targets of geological disasters, wherein the method comprises the following steps of:
step S321: the method comprises the steps of designing an input layer and an output layer, determining the dimension of input data and the total number of neurons of the input layer, wherein the dimension of the input data is determined by the total number of neurons of the input layer of a BP network, and the total number of neurons of the input layer and the total number of neurons of the output layer are determined according to the classification number of expected transmission signals;
step S322: the hidden layer design, correctly setting the hidden layer number is a key step of the hidden layer design, in order to reduce the complexity of the BP network and the training time, a network structure with only one hidden layer is used when the BP neural network is designed, errors are reduced by increasing the number of neurons of the hidden layer, and the number of neurons is set according to experience, wherein the following formula is used:
；
wherein p is the total number of neurons of the input layer, M is the total number of neurons of the hidden layer, q is the total number of neurons of the output layer, and b is a constant greater than 0 and less than 10;
step S33: the method comprises the steps of training a geological disaster prediction model, importing a historical geological disaster data set into the geological disaster prediction model, dynamically adjusting connection weights, repeatedly adjusting the connection weights among all levels in a network through forward propagation and reverse propagation, perfecting mapping relations of all levels, and adjusting the connection weights among all levels under the condition of adhering to an adjustment rule, wherein the adjustment rule is that all levels of neurons share output errors and are transmitted back to a hidden layer and an input layer step by step, so that the connection weights among all levels are adjusted;
step S34: training verification and optimization, verifying a geological disaster prediction model obtained through training, evaluating the performance and accuracy of the network according to the error between the prediction result and the actual result of the BP neural network on a verification set, and optimizing the geological disaster prediction model according to the evaluation result, wherein the method comprises the steps of adjusting a network structure, adjusting a learning rate and using a regularization method;
step S35: the real-time geological information data set is input into a geological disaster prediction model to realize automatic geological disaster prediction, and the real-time geological disaster prediction model is used for real-time geological disaster early warning and decision support.
Through executing the operation, the BP neural network is adopted to process multidimensional features, self-adaptive learning is carried out, the BP neural network is trained and then responds quickly in practical application, automatic geological disaster prediction is realized, and the technical problems that the geological disaster prediction relates to a plurality of feature variables and a large number of data samples, and the prediction effect and the model generalization capability are poor are solved.
Embodiment five, referring to fig. 1 and 5, the embodiment is based on the above embodiment, and in step S4, the geological disaster classification includes the following steps:
step S41: defining evidence items, regarding the output of each neuron as one evidence item, representing the confidence level of occurrence of the geological disaster, and mapping the output to a probability value between [0,1] by an activation function in the neural network, wherein the probability value is the probability of occurrence of the geological disaster;
step S42: confidence weight adjustment is carried out on the confidence of each evidence item according to different neurons because of the complexity of the network and uncertainty in the learning process in the neural network and the conflict of classification results of different neurons, so that the accuracy of the classification results is improved;
step S43: selecting a fusion rule, carrying out logic or operation on probability values of all evidence items by using the logic or fusion rule to obtain a total probability value, mapping the total probability value between [0,1], and fusing information of all evidence items;
step S44: combining the evidence items, namely combining the evidence items subjected to weight adjustment into an evidence item set, and combining the evidence item set into a total evidence set to obtain a combined confidence coefficient, wherein the total evidence set comprises all evidence items and the combined confidence coefficient;
step S45: and (3) selecting the geological disaster category with the highest confidence as a final geological disaster classification result according to the combined confidence.
By executing the operations, the evidence items are combined by adopting logic or fusion rules, the category with the highest confidence coefficient is selected as the final geological disaster classification result, the robustness is high, the fault tolerance is strong, the technical problems that uncertainty is faced in the geological disaster prediction process, the geological disaster prediction process is comprehensively influenced by a plurality of factors, the geological environment is varied, and the geological disaster classification result has uncertainty are solved.
Embodiment six, referring to fig. 1, the embodiment is based on the above embodiment, and in step S5, the geological disaster response and early warning includes the following steps:
step S51: optimizing a response mechanism, namely formulating event priority and a corresponding time response target, and accelerating resource allocation and information sharing;
step S52: improving monitoring equipment, in particular updating and upgrading the monitoring equipment and increasing the distribution density of monitoring points;
step S53: optimizing a data transmission network, specifically constructing a reliable and efficient data transmission network, and transmitting and processing monitoring data in real time by using an advanced communication technology;
step S54: standardizing early warning information, perfecting the following four parts in the early warning information: disaster type, risk level, area of impact, and action advice;
step S55: and (5) updating feedback, and listening to user feedback to continuously update the optimized early warning system and the information transmission channel.
An embodiment seven, referring to fig. 2, based on the embodiment, the geotechnical engineering geological disaster prediction system based on deep learning provided by the invention comprises a geological information acquisition and processing module, a signal processing module, a geological disaster prediction module, a geological disaster classification module and a geological disaster response and early warning module;
the geological information acquisition and processing module is used for acquiring geological information through a remote sensing technology and preprocessing the geological information;
the signal processing module processes historical geological disaster data by using a dynamic system reconstruction method, converts a one-dimensional time sequence into a track in a high-dimensional phase space by selecting an embedding dimension and delay time, and analyzes dynamic behaviors, nonlinear characteristics and periodic changes of an analysis system by using the reconstructed phase space track so as to obtain the periodic changes of geological disasters;
the geological disaster prediction module is used for collecting and preprocessing a historical geological disaster data set, designing a BP neural network structure, training and optimizing a geological disaster prediction model, and inputting a real-time geological information data set for practical application;
the geological disaster classification module specifically carries out weight adjustment on the confidence coefficient of the evidence items, fuses all the evidence item information by utilizing logic or fusion rules, merges the evidence items, and obtains a final geological disaster classification result according to the merged confidence coefficient;
the geological disaster response and early warning module is particularly an optimized response mechanism, accelerates resource allocation and information sharing, and continuously improves geological disaster monitoring equipment, optimizes a data transmission network and standardizes early warning information.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
The invention and its embodiments have been described above with no limitation, and the actual construction is not limited to the embodiments of the invention as shown in the drawings. In summary, if one of ordinary skill in the art is informed by this disclosure, a structural manner and an embodiment similar to the technical solution should not be creatively devised without departing from the gist of the present invention.
Claims (8)
1. The geotechnical engineering geological disaster prediction method based on deep learning is characterized by comprising the following steps of: the method comprises the following steps:
step S1: the geological information acquisition and processing, specifically, the geological information acquisition and pretreatment are carried out;
step S2: the method comprises the steps of signal processing, namely determining embedding dimension and delay time, processing historical geological disaster data by using a dynamic system reconstruction method, reconstructing a weight matrix, and converting a one-dimensional time sequence into a track in a high-dimensional phase space so as to obtain periodic change of geological disasters;
step S3: the geological disaster prediction is specifically to design a BP neural network structure, train and optimize a geological disaster prediction model, and input a real-time geological information data set for practical application;
step S4: the method comprises the steps of geological disaster classification, specifically, carrying out weight adjustment on confidence coefficient of evidence items, fusing all evidence item information by using logic or fusion rules, merging the evidence items, and obtaining a final geological disaster classification result according to the merged confidence coefficient;
step S5: the geological disaster response and the early warning are specifically that the geological disaster response is optimized and the early warning is carried out.
2. The geotechnical engineering geological disaster prediction method based on deep learning according to claim 1, wherein the method comprises the following steps: in step S2, the signal processing includes the steps of:
step S21: data preparation, namely preparing a historical geological disaster data set;
step S22: determining a delay time and an embedding dimension;
step S23: the power system reconstruction method converts a one-dimensional time sequence into a corresponding Gao Weixiang space track, and comprises the following steps of:
step S231: inputting a historical geological disaster dataset: x= { X 1 ，x 2 ，...，x n N is the total number of geological disaster samples;
step S232: determining a neighborhood, determining the neighborhood by using Euclidean distance for each time point, selecting a group of surrounding neighborhood geological disaster samples, defining the number of the neighborhood geological disaster samples as k, and defining the index set of k neighborhood geological disaster samples as N;
step S233: reconstructing a weight matrix, distributing reconstruction weights for each neighborhood geological disaster sample, determining a linear relation between each neighborhood geological disaster sample and the ith geological disaster sample, defining the reconstruction weight matrix as Q, and constructing a weight vector Q ij ，Q ij Is the ith geological disaster sample and the jth neighborhood geological disasterWeight of the harmful sample in the reconstruction process, and calculating a weight vector Q ij The formula used is as follows:
；
wherein x is i Is the ith geological disaster sample, x j Is the jth neighborhood geological disaster sample,is the vector x i －x j Absolute value of (2);
step S234: calculating a high-dimensional track in the phase space according to the reconstruction weight matrix, and decomposing the high-dimensional track in the phase space by using the characteristic values to obtain a low-dimensional representation of the high-dimensional track, so as to obtain a track after phase space reconstruction;
step S24: analyzing dynamic behaviors and nonlinear characteristics, analyzing the dynamic characteristics by calculating track indexes after phase space reconstruction, and analyzing the nonlinear characteristics of a system by using a nonlinear analysis method;
step S25: and (3) periodically analyzing the periodic variation to obtain the periodic variation in the geological disaster signal through periodically analyzing the phase space track.
3. The geotechnical engineering geological disaster prediction method based on deep learning according to claim 1, wherein the method comprises the following steps: in step S4, the geological disaster classification includes the following steps:
step S41: defining an evidence item, regarding the output of each neuron as an evidence item, representing the confidence level of geological disaster occurrence, and mapping the output of each neuron to a probability value between [0,1] by an activation function in the BP neural network;
step S42: confidence weight adjustment, namely performing weight adjustment on the confidence of each evidence item according to different neurons;
step S43: selecting a fusion rule, carrying out logic or operation on probability values of all evidence items by using the logic or fusion rule to obtain a total probability value, mapping the total probability value between [0,1], and fusing information of all evidence items;
step S44: combining the evidence items, combining the evidence items subjected to weight adjustment to obtain an evidence item set, and combining the evidence item set into a total evidence set to obtain a combined confidence level;
step S45: and (3) selecting the geological disaster category with the highest confidence as a final geological disaster classification result according to the combined confidence.
4. The geotechnical engineering geological disaster prediction method based on deep learning according to claim 1, wherein the method comprises the following steps: in step S3, the geological disaster prediction includes the steps of:
step S31: preprocessing geological information, collecting a historical geological disaster data set and preprocessing;
step S32: designing a network structure, constructing a geological disaster prediction model, and designing a BP neural network structure according to the characteristics and the prediction targets of geological disasters, wherein the method comprises the following steps of:
step S321: the input layer and the output layer are designed, and the dimension of input data and the total number of neurons of the input layer and the output layer are determined;
step S322: the hidden layer design uses BP neural network structure with only one hidden layer, sets the number of neurons according to experience, and uses the following formula:
；
wherein p is the total number of neurons of the input layer, M is the total number of neurons of the hidden layer, q is the total number of neurons of the output layer, and b is a constant greater than 0 and less than 10;
step S33: training a geological disaster prediction model, importing a historical geological disaster data set into the geological disaster prediction model, dynamically adjusting the connection weight, repeatedly adjusting the connection weight between all levels in a network through forward propagation and reverse propagation under the condition of adhering to an adjustment rule, and perfecting mapping relations of all levels;
step S34: training verification and optimization, verifying a geological disaster prediction model obtained through training, evaluating the performance and accuracy of the network according to the error between the prediction result and the actual result of the BP neural network on a verification set, and optimizing the geological disaster prediction model according to the evaluation result, wherein the method comprises the steps of adjusting a network structure, adjusting a learning rate and using a regularization method;
step S35: the actual prediction application is that a real-time geological information data set is input into a geological disaster prediction model, so that automatic geological disaster prediction is realized.
5. The geotechnical engineering geological disaster prediction method based on deep learning according to claim 1, wherein the method comprises the following steps: in step S5, the geological disaster response and early warning includes the following steps:
step S51: optimizing a response mechanism, namely formulating event priority and a corresponding time response target, and accelerating resource allocation and information sharing;
step S52: improving monitoring equipment, in particular updating and upgrading the monitoring equipment and increasing the distribution density of monitoring points;
step S53: optimizing a data transmission network, specifically constructing a reliable and efficient data transmission network, and transmitting and processing monitoring data in real time by using an advanced communication technology;
step S54: standardizing early warning information, perfecting the following four parts in the early warning information: disaster type, risk level, area of impact, and action advice;
step S55: and (5) updating feedback, and listening to user feedback to continuously update the optimized early warning system and the information transmission channel.
6. The geotechnical engineering geological disaster prediction method based on deep learning according to claim 1, wherein the method comprises the following steps: in step S1, the geological information acquisition and processing includes the following steps:
step S11: the remote sensing technology is applied, the geological features of the research area are comprehensively known by utilizing the remote sensing image, and a real-time geological information data set is obtained;
step S12: preprocessing data, namely preprocessing a collected real-time geological information data set;
step S13: and (3) feature engineering, namely analyzing a correlation coefficient between the geological information feature and geological disaster prediction, and selecting the geological information feature related to the geological disaster prediction according to the correlation coefficient.
7. Geotechnical engineering geological disaster prediction system based on deep learning, for implementing the geotechnical engineering geological disaster prediction method based on deep learning as claimed in any one of claims 1 to 6, characterized in that: the system comprises a geological information acquisition and processing module, a signal processing module, a geological disaster prediction module, a geological disaster classification module and a geological disaster response and early warning module.
8. The deep learning based geotechnical engineering geological disaster prediction system according to claim 7, wherein: the geological information acquisition and processing module is used for acquiring geological information through a remote sensing technology and preprocessing the geological information;
the signal processing module is used for determining embedding dimension and delay time, processing historical geological disaster data by using a dynamic system reconstruction method, reconstructing a weight matrix, converting a one-dimensional time sequence into a track in a high-dimensional phase space, and analyzing dynamic behaviors, nonlinear characteristics and periodic changes of an analysis system by using the reconstructed phase space track so as to obtain the periodic changes of geological disasters;
the geological disaster prediction module is used for collecting and preprocessing a historical geological disaster data set, designing a BP neural network structure, training and optimizing a geological disaster prediction model, and inputting a real-time geological information data set for practical application;
the geological disaster classification module specifically takes the output of each neuron as an evidence item, carries out weight adjustment on the confidence coefficient of the evidence item, fuses all evidence item information by utilizing logic or fusion rules, merges the evidence items, and obtains a final geological disaster classification result according to the merged confidence coefficient;
the geological disaster response and early warning module is particularly an optimized response mechanism, accelerates resource allocation and information sharing, and continuously improves geological disaster monitoring equipment, optimizes a data transmission network and standardizes early warning information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311353388.9A CN117093919B (en) | 2023-10-19 | 2023-10-19 | Geotechnical engineering geological disaster prediction method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311353388.9A CN117093919B (en) | 2023-10-19 | 2023-10-19 | Geotechnical engineering geological disaster prediction method and system based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117093919A true CN117093919A (en) | 2023-11-21 |
| CN117093919B CN117093919B (en) | 2024-02-02 |
Family
ID=88780092
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311353388.9A Active CN117093919B (en) | 2023-10-19 | 2023-10-19 | Geotechnical engineering geological disaster prediction method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117093919B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086935A (en) * | 2018-08-07 | 2018-12-25 | 中国地质环境监测院 | A kind of Prediction of Regional Geo-hazards method |
| CN113486575A (en) * | 2021-06-21 | 2021-10-08 | 武汉科技大学 | Surface deformation damage prediction early warning method and device and storage medium |
| CN115130375A (en) * | 2022-06-17 | 2022-09-30 | 北京科技大学 | Rock burst intensity prediction method |
| US20230144184A1 (en) * | 2021-11-11 | 2023-05-11 | Shandong University | Advanced geological prediction method and system based on perception while drilling |
| CN116863652A (en) * | 2023-07-20 | 2023-10-10 | 中国煤炭地质总局一一九勘探队 | Geological disaster early warning method based on Internet adding technology |
-
2023
- 2023-10-19 CN CN202311353388.9A patent/CN117093919B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086935A (en) * | 2018-08-07 | 2018-12-25 | 中国地质环境监测院 | A kind of Prediction of Regional Geo-hazards method |
| CN113486575A (en) * | 2021-06-21 | 2021-10-08 | 武汉科技大学 | Surface deformation damage prediction early warning method and device and storage medium |
| US20230144184A1 (en) * | 2021-11-11 | 2023-05-11 | Shandong University | Advanced geological prediction method and system based on perception while drilling |
| CN115130375A (en) * | 2022-06-17 | 2022-09-30 | 北京科技大学 | Rock burst intensity prediction method |
| CN116863652A (en) * | 2023-07-20 | 2023-10-10 | 中国煤炭地质总局一一九勘探队 | Geological disaster early warning method based on Internet adding technology |
Non-Patent Citations (2)
| Title |
|---|
| 于锦 等: "基于栈式自编码器的土质边坡失稳风险评估", 安全与环境工程, no. 05 * |
| 董辉 等: "基于Takens理论和SVM的滑坡位移预测", 中国公路学报, no. 05 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117093919B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Energy data generation with wasserstein deep convolutional generative adversarial networks | |
| CN111079999A (en) | Flood disaster susceptibility prediction method based on CNN and SVM | |
| CN116029435B (en) | Environmental comfort early warning system is bred to live pig facility | |
| CN116451553A (en) | Improved variation modal decomposition and BiGRU fusion water quality prediction method | |
| CN116861331A (en) | Expert model decision-fused data identification method and system | |
| CN110516792A (en) | Non-stable time series forecasting method based on wavelet decomposition and shallow-layer neural network | |
| CN117235661B (en) | AI-based direct drinking water quality monitoring method | |
| Terbuch et al. | Hybrid machine learning for anomaly detection in industrial time-series measurement data | |
| CN117093919B (en) | Geotechnical engineering geological disaster prediction method and system based on deep learning | |
| CN109657907B (en) | Quality control method and device for geographical national condition monitoring data and terminal equipment | |
| CN116738192A (en) | Digital twinning-based security data evaluation method and system | |
| Balaji et al. | Plant Infirmity Detection Using Vgg-16 Convolutional Neural Network | |
| Tang et al. | Leveraging socioeconomic information and deep learning for residential load pattern prediction | |
| CN114062812A (en) | Fault diagnosis method and system for metering cabinet | |
| CN113537759A (en) | User experience measurement model based on weight self-adaptation | |
| CN113269380A (en) | Return-to-school scheme estimation method for epidemic situation prevention and control | |
| CN117649209B (en) | Enterprise revenue auditing method, system, equipment and storage medium | |
| CN117473435B (en) | Method for detecting false abnormal information of sudden public health event based on space-time characteristics | |
| CN117593101B (en) | Financial risk data processing and analyzing method and system based on multidimensional data | |
| Siraj et al. | Data mining and neural networks: the impact of data representation | |
| KR102627837B1 (en) | Method and system for risk assessment of alien species using global-scale environmental data and economic loss database | |
| CN117670066B (en) | Questor management method, system, equipment and storage medium based on intelligent decision | |
| Wang et al. | Product Key Reliability Characteristics Identification Method Based on XGBoost in Manufacturing Process | |
| Schiller | A comparative review of forecasting methods for the availability of charging stations for electric vehicles | |
| Liu et al. | A novel student achievement prediction method based on deep learning and attention mechanism |
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 |