CN118114845A - Method, device and medium for predicting displacement trend in bridge operation and maintenance period - Google Patents
Method, device and medium for predicting displacement trend in bridge operation and maintenance period Download PDFInfo
- Publication number
- CN118114845A CN118114845A CN202410532632.6A CN202410532632A CN118114845A CN 118114845 A CN118114845 A CN 118114845A CN 202410532632 A CN202410532632 A CN 202410532632A CN 118114845 A CN118114845 A CN 118114845A
- Authority
- CN
- China
- Prior art keywords
- displacement
- prediction
- data
- time period
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000006073 displacement reaction Methods 0.000 title claims abstract description 199
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000012423 maintenance Methods 0.000 title claims abstract description 26
- 238000012544 monitoring process Methods 0.000 claims abstract description 65
- 238000012937 correction Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 8
- 230000009467 reduction Effects 0.000 claims description 6
- 230000007704 transition Effects 0.000 claims description 5
- 230000001502 supplementing effect Effects 0.000 claims description 2
- 230000008859 change Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
- G06F18/295—Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Primary Health Care (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Bridges Or Land Bridges (AREA)
Abstract
The application relates to a method, a device and a medium for predicting displacement trend in bridge operation and maintenance period, wherein the method for predicting displacement trend comprises the following steps: training the LSTM basic model by using the original data set to obtain a trained LSTM model; selecting input data of a second time period, inputting the input data into the trained LSTM model for prediction, and obtaining displacement prediction data of each prediction time period; calculating relative errors corresponding to all the predicted time periods based on displacement actual data corresponding to all the predicted time periods in the original data set; inputting the relative errors corresponding to each prediction time period into a Markov chain model to obtain a correction coefficient; predicting the displacement trend of the target time period by utilizing the LSTM model to obtain target displacement prediction data; and correcting the target displacement prediction data by using the correction coefficient to obtain corrected target displacement prediction data, so that an accurate displacement prediction result can be obtained based on the displacement monitoring data in a shorter period of time.
Description
Technical Field
The application relates to the technical field of bridge displacement trend prediction, in particular to a method, a device and a medium for predicting the displacement trend of a bridge in the operation and maintenance period.
Background
Displacement trend prediction for the future 24 hours and equivalent length of the bridge in the operation and maintenance period is performed by using a deep learning LSTM (Long-short term memory) model, and displacement monitoring data of the bridge in the operation and maintenance period, for example, displacement monitoring data of 3-5 years, is required to accurately perform displacement trend prediction for the bridge in the future period, because the displacement of the bridge is influenced by temperature and other factors, because the longer time data can comprise the displacement condition of the bridge in various different periods. If there is less displacement monitoring data about the bridge, for example, displacement monitoring data having no year period in the early period of the operation and maintenance period of about 3-10 months, the displacement trend of the bridge may not be accurately predicted. If the displacement monitoring data in a short time possibly cannot include the periodic change condition and the condition of large fluctuation of the displacement of the bridge, the displacement trend prediction cannot be performed in the later period of the bridge, and the accuracy is low.
Disclosure of Invention
The present application has been made to solve the above-mentioned drawbacks of the prior art. The method, the device and the medium for predicting the displacement trend in the bridge operation and maintenance period are needed, the displacement trend prediction result can be obtained by utilizing the LSTM model and the correction coefficient of the Markov chain based on displacement monitoring data in a short period of time, and the accuracy of the displacement prediction result can be improved.
According to a first aspect of the present application, there is provided a method for predicting a displacement trend during operation and maintenance of a bridge, the method comprising the steps of. And acquiring displacement monitoring data of a first time period in the bridge operation and maintenance period. Based on the displacement monitoring data in the first time period, selecting part of continuous displacement monitoring data as an original data set, and training an LSTM basic model by utilizing the original data set to obtain a trained LSTM model. Based on the original data set, selecting a plurality of input data of a second time period, inputting the input data into the trained LSTM model, and predicting to obtain displacement prediction data of each prediction time period, wherein the prediction time period is positioned behind the second time period according to a time sequence within a time range corresponding to the first time period. Based on displacement actual data corresponding to each prediction time period in the original data set, calculating a relative error of displacement between the displacement prediction data corresponding to each prediction time period and the displacement actual data. And inputting the relative errors corresponding to the prediction time periods into a Markov chain model to obtain correction coefficients. And predicting the displacement trend of the target time period by using the trained LSTM model based on the original data set to obtain target displacement prediction data, wherein the target time period is positioned after the first time period according to the time sequence. And correcting the target displacement prediction data by using the correction coefficient to obtain corrected target displacement prediction data.
According to a second aspect of the present application, there is provided a device for predicting a displacement trend during operation and maintenance of a bridge, the device comprising an interface and a processor. The interface is configured to obtain displacement monitoring data for a first period of time during operation and maintenance of the bridge. The processor is configured to perform the prediction method according to any of the embodiments of the present application.
According to a third aspect of the present application there is provided a non-transitory computer readable medium having instructions stored thereon which, when executed by a processor, perform the steps of the prediction method according to any of the embodiments of the present application.
According to the method, the device and the medium for predicting the displacement trend in the bridge operation and maintenance period, the displacement monitoring data in the first time period can be displacement monitoring data in a short time, for example, about 3-10 months, the LSTM basic model is firstly trained based on the displacement monitoring data in the first time period, then the trained LSTM basic model is utilized for prediction, and the correction coefficient is obtained by utilizing the Markov chain model based on the predicted value and the actual value, so that the target displacement prediction data can be corrected in the displacement trend prediction in the target time period to obtain an accurate displacement trend prediction result, and therefore the accurate displacement prediction result can be obtained based on the displacement monitoring data in the short time period.
Drawings
In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. The same reference numerals with letter suffixes or different letter suffixes may represent different instances of similar components. The accompanying drawings illustrate various embodiments by way of example in general and not by way of limitation, and together with the description and claims serve to explain the claimed embodiments. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts. Such embodiments are illustrative and not intended to be exhaustive or exclusive of the present apparatus or method.
FIG. 1 shows a flow chart of a method for predicting displacement trends during bridge operation according to an embodiment of the present application; and
Fig. 2 shows a schematic structural diagram of a device for predicting displacement trend during bridge operation according to an embodiment of the present application.
Detailed Description
The present application will be described in detail below with reference to the drawings and detailed description to enable those skilled in the art to better understand the technical scheme of the present application. Embodiments of the present application will be described in further detail below with reference to the drawings and specific examples, but not by way of limitation.
The terms "first," "second," and the like, as used herein do not denote any order, quantity, or importance, but rather are used to distinguish one element from another. The word "comprising" or "comprises" and the like means that elements preceding the word encompass the elements recited after the word, and not exclude the possibility of also encompassing other elements.
The embodiment of the application provides a method for predicting the displacement trend of a bridge in the operation and maintenance period. Fig. 1 shows a flowchart of a method for predicting a displacement trend during operation and maintenance of a bridge according to an embodiment of the present application. The prediction method comprises the following steps. In step 101, displacement monitoring data of a first period of time during operation and maintenance of a bridge is obtained. During the operation and maintenance period of the bridge, the displacement monitoring data can be obtained through the displacement sensor on the bridge, and the displacement monitoring data can be the displacement data transmitted by the displacement sensor on the bridge at a certain frequency (for example, 5 minutes). The bridge operation and maintenance period is the operation period after the completion of the bridge. The first period of time may be 3 months or more, such as 3 months, 5 months, 7 months, 12 months, etc., where the first period of time may be of different lengths. The displacement monitoring data of the first period may be only displacement monitoring data of the bridge at an early or middle operation and maintenance period compared to the displacement monitoring data of the longer period of 3-5 years.
In step 102, based on the displacement monitoring data in the first period, a part of continuous displacement monitoring data is selected as an original data set, and the LSTM basic model is trained by using the original data set, so as to obtain a trained LSTM model. The displacement monitoring data that is partially continuous during the first period of time may be selected based on the data amount of the displacement monitoring data during the first period of time. Specifically, the time span may be the same as the first time period, for example, the displacement monitoring data are all 3 months long, only part of time points are not selected, so as to reduce the data amount, and the time length corresponding to the displacement monitoring data can be ensured, so that the accuracy is improved. The original data set can be selected to be a training set with a 0.8 ratio and a test set with a 0.2 ratio, wherein the ratio can be adjusted, and the original data set can be disturbed when the LSTM basic model is trained. The deep-learning LSTM model is beneficial to more accurately predicting displacement trend of a certain period in the future.
In step 103, based on the original data set, selecting a plurality of input data in a second time period, inputting the input data into the trained LSTM model, and predicting the input data to obtain displacement prediction data in each prediction time period, where the prediction time period is in a time range corresponding to the first time period. The predicted time period is located after the second time period in time order. The second period may be a period of 3-7 days, taking one data point every 15 minutes in 7 consecutive days as an example, the input data of the second period includes displacement monitoring data of 7×24×4 time points, and the displacement prediction data of the predicted period may be displacement prediction data of 24 hours on 8 th day, for example, displacement prediction data of 24 time points from 1 to 24 early morning on 8 th day, displacement prediction data of 3 to 2 early morning on 9 th day on 8 th day, and the like. Here, the second period and the prediction period are both variable, and are selected as needed.
In step 104, based on the displacement actual data corresponding to each of the predicted time periods in the original data set, a relative error of the displacement between the displacement predicted data and the displacement actual data corresponding to each of the predicted time periods is calculated. The displacement prediction data predicted by LSTM may be compared with the displacement actual data in the original data set, for example, the first time period is 90 days, so that the displacement prediction data of 24 time points on day 8 may be compared with the displacement actual data of 24 time points on day 8 corresponding to the original data set, so that the relative error of each data point can be obtained, and each of the relative errors includes the relative error of 24 time points, for example, the relative error of 1 point, the relative error of 2 points, and the like, where the relative error may be a displacement difference. This allows the difference between the displacement predicted value and the displacement actual value to be obtained.
In step 105, the relative error corresponding to each of the predicted time periods is input into a Markov chain model (Markov Chain Model) to obtain a correction coefficient. The relative error of each prediction period is input to a markov chain model, for example, displacement prediction data at 24 time points on day 8, displacement prediction data at 24 time points on day 9, and the like, and the output correction coefficient is obtained by inputting the relative error to the markov chain model. The displacement monitoring data in some time periods can have larger displacement fluctuation under the influence of temperature and other factors, and the frequency of LSTM model iteration can be reduced by using a Markov chain model, so that the prediction accuracy is improved.
In step 106, based on the original data set, the displacement trend of the target time period is predicted by using the trained LSTM model, so as to obtain target displacement prediction data, where the target time period is located after the first time period according to the time sequence. For example, in the case of data of the first 3 months of the bridge operation period, target displacement prediction data of a target period of the 4 th month in the future can be predicted, for example, displacement of 1 to 24 am on the 91 st day can be predicted, so as to obtain a bridge displacement trend of the target period in the future of the first period.
In step 107, the target displacement prediction data is corrected by the correction coefficient, and corrected target displacement prediction data is obtained. The first time period only corresponds to a shorter time in the bridge operation period, and the accuracy of the displacement trend prediction result taking the displacement monitoring data in the first time period as input data can be improved through modification of the correction coefficient.
The displacement monitoring data within one year does not comprise displacement monitoring data of bridges with a period of years, and possibly does not comprise displacement change conditions of the bridges under the influence of temperature and other factors within a long time, so that training prediction of data of years and whole years is not available. The method can better cope with the situation that the existing LSTM model does not have sufficient displacement monitoring data, overcomes the defect that the existing LSTM model needs longer time of displacement monitoring data to obtain accurate results of bridge displacement, and can consume less calculation resources.
In some embodiments, inputting the relative error corresponding to each of the predicted time periods into the markov chain model specifically includes: arranging the relative errors of the displacement in each prediction time period according to the time sequence of the prediction time period; the relative error data arranged in time order is input to a Markov chain model. Taking the original data set of 3 months as an example, the relative errors of all the prediction time periods are arranged according to a time sequence, including the displacement prediction relative errors of 24 data points on the 8 th day, the displacement prediction relative errors of 24 data points on the 9 th day, the displacement prediction relative errors of 24 data points on the 10 th day, and the displacement prediction relative errors of 24 data points up to the 90 th day, and the displacement prediction relative errors of 24 data points are input into a Markov chain model. Since LSTM model predicts displacement values with time order for future time periods, markov chain models are processed in time order to obtain relative errors corresponding to respective data points in time order.
In some embodiments, inputting the relative errors corresponding to the prediction time periods into a markov chain model, and obtaining the correction coefficient specifically includes the following steps. And carrying out state classification on the relative error sequence by using an ordered clustering method. And determining error intervals of the state classifications based on the error data corresponding to the state classifications. And calculating a transition probability matrix based on the error interval and the relative error sequence of each state classification, and obtaining a state matrix based on the transition probability matrix. A target error interval is determined based on the state matrix and the error intervals for each state classification. And obtaining a correction coefficient based on the target error interval.
The relative errors in the relative error sequences herein are arranged in time order, and the relative error sequences may be in the form of a set or list, etc.
Since the LSTM model predicts and outputs displacement values with time sequence in future time periods, the relative error sequence is processed by using the ordered clustering method, which is helpful for predicting displacement values with time sequence in future. Taking the above-mentioned relative error sequences from day 8 to day 90 as an example, the input relative error sequences are classified into four state classifications by an ordered clustering method. Then determining the error interval of each state classification, wherein the error interval is respectively: [ z0, z1], [ z1, z2], [ z2, z3], [ z3, z4]. And calculating a transition probability matrix P0 according to the error interval of each state classification and all the relative error values. And taking the last relative error as an initial state, and obtaining a state matrix m0 according to the state interval (1*4). Probability distribution of first predictor: m0 x P0, and thus a state matrix m1, and a second predicted value m2=m1 x p0.. And determining the target error interval as [ z2, z3] according to the state matrix (0, 1, 0) and the state interval, and then calculating a correction coefficient according to the target error interval. And the correction coefficient is convenient to accurately obtain.
In some embodiments, based on the target error interval, deriving the correction system specifically includes: determining a minimum relative error value and a maximum relative error value within a target error interval; correction coefficient= (minimum relative error value+maximum relative error value)/2. Taking the median value of the target error interval as the final correction coefficient to correct the displacement prediction value, taking the above as an example, the correction coefficient xz= (z2+z3)/2.
In some embodiments, the correcting the target displacement prediction data by using the correction coefficient, where the obtaining corrected target displacement prediction data specifically includes: corrected target displacement prediction data=target displacement prediction data/(1-correction coefficient). The target displacement prediction data after correction coefficient processing is made to be closer to the actual displacement value, and as an example, the corrected target displacement prediction data=target displacement prediction data/(1-xz) is taken to obtain the displacement value of each time point in the corrected target time period.
In some embodiments, training the LSTM base model using the raw dataset, the obtaining the trained LSTM model specifically includes: a first input data set and a first output data set are selected based on the original data set. And training the LSTM basic model by using the first input data set and the first output data set to obtain a trained LSTM model. Since the LSTM model predicts a displacement trend for a future time period, the time period corresponding to the first output data set should be after the time period corresponding to the first input data set in time sequence, for example, the first output data set includes a displacement from 1 early morning on day 10 to 24 early on day 11, a displacement from 2 early morning on day 10 to 1 early morning on day 11, etc., and the first input data set includes data before 1 early morning on day 10 and data before 2 early morning on day 10. The LSTM basic model can be trained better, so that future displacement trend can be predicted.
In some embodiments, selecting the first input data set and the first output data set based on the original data set specifically comprises: each predicted point in time is determined. And selecting displacement monitoring data of a third time period before each prediction time point based on the original data set as a first input data set, and selecting displacement monitoring data of a prediction time period after each prediction time point as a first output data set according to time sequence. The first input data set and the first output data set are adjacent in time and are bounded by predicted points, because the displacement trend of the next adjacent point in time may have a larger influence on the previous point in time, for example, the previous point in time is influenced by an external force, the displacement may have a larger change, the next point in time may follow a larger displacement fluctuation, for example, the predicted point in time is a displacement after 2 points, and data before 1 point is needed. Taking the example of predicting the displacement from 2 to 24 a.m., the predicted time point may be 2 a.m. The third period of time may be displacement monitoring data of 7 days or 6 days before the predicted time point, or the like, and is continuous displacement monitoring data. So as to input more fitting data to the LSTM model and accurately obtain future displacement trend. In addition, the input data in the first input data set are arranged in time sequence, and the time sequence displacement monitoring data is not disturbed.
In some embodiments, the portion of the displacement monitoring data based on the first time period specifically comprises, as an original data set: and performing frequency reduction processing on the displacement monitoring data in the first time period, and supplementing the displacement monitoring data of the missing data time points in the frequency reduction processed displacement monitoring data to obtain supplemented displacement monitoring data which are used as an original data set. Since the frequency of uploading data by the displacement sensor may be relatively high, the displacement monitoring data in the first period is subjected to a process such as frequency reduction, for example, a data interval of 5 minutes, the time interval of adjacent data points may be 15 minutes through the frequency reduction process, and the frequency reduction amplitude may be selected according to the data amount and the frequency of the data of the displacement sensor.
In the original data set, data may be lost due to the sensor, so that data is lost at some time points, and the data can be supplemented by displacement monitoring data at the time points adjacent to each other, for example, the average value of the displacement monitoring data at the time points before and after the time point at which the data is lost.
The embodiment of the application also provides a device for predicting the displacement trend of the bridge in the operation and maintenance period. Fig. 2 shows a schematic structural diagram of a device for predicting displacement trend during bridge operation according to an embodiment of the present application. The prediction apparatus 200 comprises an interface 201 and a processor 202. The interface 201 is configured to obtain displacement monitoring data for a first period of time during operation and maintenance of the bridge. Processor 202 is configured to perform the prediction method described in any of the embodiments of the present application. According to the prediction device, through training of the LSTM model and inputting of the correction coefficient obtained by the Markov chain through the relative error, the bridge displacement result can be accurately obtained under the condition of displacement monitoring data in a short period of time, and even a target period with larger displacement fluctuation still has higher accuracy, so that the prediction device can be used for predicting the displacement trend in early, middle and late stages of bridge operation and maintenance. The method can better cope with the situation that the existing LSTM model does not have sufficient displacement monitoring data, and solves the problem that the existing LSTM model needs long-time displacement monitoring data to obtain accurate bridge displacement results. But also consume less computing resources.
The processor 202 in the present application may be a processing device including more than one general purpose processing device, such as a microprocessor, central Processing Unit (CPU), graphics Processing Unit (GPU), or the like. More specifically, the processor may be a Complex Instruction Set Computing (CISC) microprocessor, a Reduced Instruction Set Computing (RISC) microprocessor, a Very Long Instruction Word (VLIW) microprocessor, a processor running other instruction sets, or a processor running a combination of instruction sets. The processor may also be one or more special purpose processing devices such as an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), a system on a chip (SoC), or the like. The processor may be communicatively coupled to the memory and configured to execute computer-executable instructions stored thereon.
There is also provided, in accordance with an embodiment of the present application, a non-transitory computer-readable medium having instructions stored thereon, which when executed by a processor, perform the steps of the prediction method according to any of the embodiments of the present application. According to the prediction method, the bridge displacement result can be accurately obtained under the condition of displacement monitoring data in a short period of time through training of an LSTM model and inputting of the correction coefficient obtained by a Markov chain by relative errors, and the bridge displacement result still has high accuracy even in a target period of time with large displacement fluctuation, so that the prediction method can be used for predicting the displacement trend in early, middle and late states of bridge operation and maintenance. The method can better cope with the situation that the existing LSTM model does not have sufficient displacement monitoring data, and solves the problem that the existing LSTM model needs long-time displacement monitoring data to obtain accurate bridge displacement results. But also consume less computing resources.
Furthermore, although exemplary embodiments have been described herein, the scope thereof includes any and all embodiments having equivalent elements, modifications, omissions, combinations (e.g., of the various embodiments across), adaptations or alterations as pertains to the present application. The elements in the claims are to be construed broadly based on the language employed in the claims and are not limited to examples described in the present specification or during the practice of the application, which examples are to be construed as non-exclusive. It is intended, therefore, that the specification and examples be considered as exemplary only, with a true scope and spirit being indicated by the following claims and their full scope of equivalents.
The above description is intended to be illustrative and not restrictive. For example, the above-described examples (or one or more aspects thereof) may be used in combination with each other. For example, other embodiments may be used by those of ordinary skill in the art upon reading the above description. In addition, in the above detailed description, various features may be grouped together to streamline the application. This is not to be interpreted as an intention that the features of the non-claimed application are essential to any claim. Rather, the inventive subject matter may lie in less than all features of a particular disclosed embodiment. Thus, the following claims are hereby incorporated into the detailed description as examples or embodiments, with each claim standing on its own as a separate embodiment, and it is contemplated that these embodiments may be combined with one another in various combinations or permutations. The scope of the application should be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements of this application will occur to those skilled in the art, and are intended to be within the spirit and scope of the application.
Claims (10)
1. The method for predicting the displacement trend of the bridge during operation and maintenance is characterized by comprising the following steps of:
Acquiring displacement monitoring data of a first time period in the bridge operation and maintenance period;
Based on the displacement monitoring data in the first time period, selecting part of continuous displacement monitoring data as an original data set, and training an LSTM basic model by utilizing the original data set to obtain a trained LSTM model;
Based on the original data set, selecting a plurality of input data of a second time period, inputting the input data into a trained LSTM model, and predicting to obtain displacement prediction data of each prediction time period, wherein the prediction time period is in a time range corresponding to the first time period, and the prediction time period is positioned behind the second time period according to a time sequence;
calculating the relative error of the displacement between the displacement prediction data corresponding to each prediction time period and the displacement actual data based on the displacement actual data corresponding to each prediction time period in the original data set;
Inputting the relative errors corresponding to the prediction time periods into a Markov chain model to obtain correction coefficients;
Based on the original data set, predicting the displacement trend of a target time period by using the trained LSTM model to obtain target displacement prediction data, wherein the target time period is positioned after the first time period according to time sequence;
and correcting the target displacement prediction data by using the correction coefficient to obtain corrected target displacement prediction data.
2. The prediction method according to claim 1, wherein inputting the relative error corresponding to each of the prediction time periods into a markov chain model specifically comprises: arranging the relative errors of the displacement in each prediction time period according to the time sequence of the prediction time period; the relative error data arranged in time order is input to a Markov chain model.
3. The prediction method according to claim 2, wherein inputting the relative error corresponding to each of the prediction time periods into a markov chain model to obtain the correction coefficient specifically includes:
Carrying out state classification on the relative error sequence by using an ordered clustering method;
determining error intervals of each state classification based on error data corresponding to each state classification;
Calculating a transition probability matrix based on the error interval and the relative error sequence of each state classification, and obtaining a state matrix based on the transition probability matrix;
determining a target error interval based on the state matrix and the error interval of each state classification;
and obtaining a correction coefficient based on the target error interval.
4. The prediction method according to claim 3, wherein obtaining the correction system based on the target error interval specifically includes:
Determining a minimum relative error value and a maximum relative error value within a target error interval;
correction coefficient= (minimum relative error value+maximum relative error value)/2.
5. The prediction method according to claim 1, wherein correcting the target displacement prediction data by using a correction coefficient, the obtaining corrected target displacement prediction data specifically includes: corrected target displacement prediction data=target displacement prediction data/(1-correction coefficient).
6. The method of predicting as claimed in claim 1, wherein training the LSTM base model using the raw dataset, the obtaining the trained LSTM model specifically comprises:
selecting a first input data set and a first output data set based on the original data set;
And training the LSTM basic model by using the first input data set and the first output data set to obtain a trained LSTM model.
7. The method of predicting according to claim 6, wherein selecting the first input data set and the first output data set based on the original data set comprises:
determining each predicted time point;
And selecting displacement monitoring data of a third time period before each prediction time point based on the original data set as a first input data set, and selecting displacement monitoring data of a prediction time period after each prediction time point as a first output data set according to time sequence.
8. The prediction method according to claim 1, wherein the portion of the data in the displacement monitoring data based on the first period of time as the original data set specifically includes: and performing frequency reduction processing on the displacement monitoring data in the first time period, and supplementing the displacement monitoring data of the missing data time points in the frequency reduction processed displacement monitoring data to obtain supplemented displacement monitoring data which are used as an original data set.
9. A device for predicting displacement trend during operation and maintenance of a bridge, the device comprising:
the interface is configured to acquire displacement monitoring data of a first time period in the bridge operation period;
A processor configured to perform the prediction method of any of claims 1-8.
10. A non-transitory computer readable medium having instructions stored thereon, which when executed by a processor, perform the steps of the prediction method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410532632.6A CN118114845A (en) | 2024-04-30 | 2024-04-30 | Method, device and medium for predicting displacement trend in bridge operation and maintenance period |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410532632.6A CN118114845A (en) | 2024-04-30 | 2024-04-30 | Method, device and medium for predicting displacement trend in bridge operation and maintenance period |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118114845A true CN118114845A (en) | 2024-05-31 |
Family
ID=91216347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410532632.6A Pending CN118114845A (en) | 2024-04-30 | 2024-04-30 | Method, device and medium for predicting displacement trend in bridge operation and maintenance period |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118114845A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020253038A1 (en) * | 2019-06-18 | 2020-12-24 | 平安普惠企业管理有限公司 | Model construction method and apparatus |
| CN112434359A (en) * | 2020-11-11 | 2021-03-02 | 东华理工大学 | High-speed railway pier settlement curve prediction method and system |
| CN113449473A (en) * | 2021-06-29 | 2021-09-28 | 南京工业大学 | Wind power bearing degradation trend prediction method, system and device based on DFCAE-BilSTM |
| CN114741976A (en) * | 2022-06-13 | 2022-07-12 | 西南交通大学 | Displacement prediction method, device and equipment and readable storage medium |
| CN115270239A (en) * | 2022-06-16 | 2022-11-01 | 浙江工业大学 | Bridge reliability prediction method based on dynamic characteristics and intelligent algorithm response surface method |
-
2024
- 2024-04-30 CN CN202410532632.6A patent/CN118114845A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020253038A1 (en) * | 2019-06-18 | 2020-12-24 | 平安普惠企业管理有限公司 | Model construction method and apparatus |
| CN112434359A (en) * | 2020-11-11 | 2021-03-02 | 东华理工大学 | High-speed railway pier settlement curve prediction method and system |
| CN113449473A (en) * | 2021-06-29 | 2021-09-28 | 南京工业大学 | Wind power bearing degradation trend prediction method, system and device based on DFCAE-BilSTM |
| CN114741976A (en) * | 2022-06-13 | 2022-07-12 | 西南交通大学 | Displacement prediction method, device and equipment and readable storage medium |
| CN115270239A (en) * | 2022-06-16 | 2022-11-01 | 浙江工业大学 | Bridge reliability prediction method based on dynamic characteristics and intelligent algorithm response surface method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5962190B2 (en) | Method and apparatus for predicting short-term power load | |
| CN113221271B (en) | Digital twin-driven quantitative recognition method for cracks of rotating blades of aircraft engine | |
| EP3859455A1 (en) | Learning apparatus, learning method, learning program, determination apparatus, determination method, determination program, and computer readable medium | |
| CN110633859B (en) | Hydrologic sequence prediction method integrated by two-stage decomposition | |
| CN113657668A (en) | Power load prediction method and system based on LSTM network | |
| CN112633461B (en) | Application assistance system and method, and computer-readable recording medium | |
| CN112100713A (en) | Cable force automatic identification method based on variable-pitch grid | |
| CN112949932B (en) | Ship traffic flow prediction method, device, computer equipment and storage medium | |
| CN114911788B (en) | Data interpolation method and device and storage medium | |
| CN110991761B (en) | Heat supply load prediction method and device | |
| CN118114845A (en) | Method, device and medium for predicting displacement trend in bridge operation and maintenance period | |
| CN112990372A (en) | Data processing method, model training device and electronic equipment | |
| US12031848B2 (en) | Method and computing device for detecting anomalous sensor data | |
| CN110472788B (en) | Personnel information prediction method and device | |
| CN117574835A (en) | Method, equipment and storage medium for simulating Nand false LLR distribution | |
| CN117010723A (en) | Carbon emission prediction method, system, apparatus, and computer-readable storage medium | |
| CN111563078B (en) | Data quality detection method and device based on time sequence data and storage device | |
| CN116701875A (en) | Audible noise probability prediction method and system for extra-high voltage alternating current transmission line | |
| CN115730717A (en) | Power load prediction method and system based on combination of transfer learning strategy and multiple channels | |
| CN114661665A (en) | Determination method of execution engine, model training method and device | |
| CN116086497B (en) | Output correction method, device and medium for optical fiber gyroscope under low angular rate based on neural network | |
| CN117553864B (en) | Sensor acquisition method and system based on big data | |
| EP3518153A1 (en) | Information processing method and information processing system | |
| CN112102942A (en) | Skeletal development grade detection method and terminal equipment | |
| CN117996863B (en) | Sagging control method and device for photovoltaic power supply, computer equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |