Tunnel monitoring data analysis method based on long-term monitoring data

Abstract

The invention discloses a tunnel monitoring data analysis method based on long-term monitoring data, and belongs to the field of constructional engineering. The method comprises the following steps: firstly, acquiring historical long-term monitoring data of a target tunnel, and dividing the long-term monitoring data into continuous data and classified data; adopting an interpolation algorithm to generate a continuous data change curve corresponding to the continuous data; the continuous data prediction comprises a time sequence analysis method, a regression analysis method and a machine learning method; the analysis of the classified data is divided into observation of whether the classified data is abnormal or not and prediction of the classified data; verifying the collected classification data through a neural network technology; predicting the classified data through a deep learning technology; according to the method, tunnel monitoring data is divided into continuous data and classified data, and analysis methods of the continuous data and the classified data are provided respectively; a method for processing different tunnel monitoring data by adopting different interpolation methods, machine learning methods and the like is provided.

Description

Technical Field

[0001] The invention belongs to the field of construction engineering, and in particular relates to a tunnel monitoring data analysis method based on long-term monitoring data. Background Art

[0002] The analysis of tunnel monitoring data is an important application in the field of construction engineering. The analysis of tunnel monitoring data includes the following steps: sensor technology and data acquisition, data processing and analysis, structural health monitoring technology, mathematical modeling and simulation analysis, and real-time processing and decision support of monitoring data. Through the above steps, real-time monitoring of tunnels can be achieved to prevent tunnel accidents.

[0003] But at the same time, there are some shortcomings in the analysis of tunnel monitoring data, including: the complexity and variability of data, data quality, data processing methods, real-time data acquisition, model accuracy and complexity, etc.; therefore, when analyzing tunnel monitoring data, it is necessary not only to be able to obtain correct monitoring data in real time, but also to choose the correct analysis method to analyze the monitoring data to ensure that the analyzed data is accurate and effective. Summary of the invention

[0004] In order to overcome the problems existing in the above technologies, the present invention provides a tunnel monitoring data analysis method based on long-term monitoring data, the method comprising:

[0005] Acquire the historical long-term monitoring data of the target tunnel, and divide the long-term monitoring data into continuous data and classified data; the continuous data refers to data that changes continuously, including displacement data, strain data and temperature data; the classified data refers to data that does not change continuously and is used for classification, including safety alarm data, image data and video data;

[0006] The analysis of continuous data includes the establishment of continuous data change curves and the prediction of continuous data; the establishment of continuous data change curves uses corresponding interpolation algorithms to process the data in the form of data to generate continuous data change curves corresponding to the continuous data. The interpolation algorithms used include: linear interpolation, Newton interpolation and spline curve interpolation; the prediction of continuous data uses methods including time series analysis, regression analysis and machine learning to predict the data at the next moment of the continuous data change curve;

[0007] The analysis of classified data includes observing whether there are anomalies in the classified data and predicting the classified data; the observation of whether there are anomalies in the classified data refers to the existence of classification and data correspondence errors in the collected classified monitoring data, including: security alarm errors, image data classification errors and video data classification errors; the method of observing whether there are anomalies in the classified data is: verifying the collected classified data through neural network technology; the prediction of classified data includes: predicting the classified data through deep learning technology, and predicting the state of the carrier of the classified data at the next moment.

[0008] The acquisition of the target tunnel's historical long-term monitoring data includes: the long-term refers to the continuous collection of data, rather than a fixed period of time; factors for determining whether the tunnel monitoring data is long-term include: data collection cycle, change trend and application requirements.

[0009] The establishment of the continuous data change curve adopts the corresponding interpolation algorithm for processing according to the form of the data, including: temperature data and humidity data are processed by linear interpolation algorithm, displacement data and vibration data are processed by Newton interpolation algorithm, and strain data and time series data are processed by spline interpolation algorithm; the corresponding data are processed by interpolation algorithm to obtain the continuous data change curve of the corresponding data.

[0010] The prediction of the continuous data adopts methods including time series analysis, regression analysis and machine learning to predict the data at the later stage of the continuous data change curve, including: what kind of prediction method is selected for a set of continuous data depends on the application object and analysis method of the data; for example: temperature, humidity, strain and vibration data show obvious time correlation and periodic changes, and time series analysis methods are often used to identify and model trends, seasonal changes and periodic changes in data, so it is possible to analyze the temperature, humidity, strain and vibration data through time series analysis to predict the temperature, humidity, strain and vibration changes of the target tunnel at the next moment; the next moment is determined according to the time interval for the equipment to collect the corresponding data information of the target tunnel; for example: considering the tunnel structure The relationship between the displacement and strain of the structure and external factors including temperature and humidity, and regression analysis is suitable for processing the relationship between different variables in tunnel monitoring, especially those cases where one or more independent variables are used to predict dependent variables; therefore, regression analysis can be used to process data with humidity and temperature as independent variables and displacement and strain as dependent variables, and predict the corresponding data at the next moment based on the generated prediction model; machine learning methods, with flexible model architecture, non-parametric and parametric methods, the improvement of big data and computing power, the development of feature engineering and the advancement of integrated learning methods, can be used to process almost all types of data predictions, as long as the appropriate training model is selected according to the data type during the model training process.

[0011] The verification of the collected classification data through the neural network technology includes: training the correct data collected historically to obtain a neural network model; inputting the verified target data into the neural network model to obtain an output result; comparing the output result with the actual classification result to observe whether the output result and the actual classification result are the same. If they are not the same, it means that an error occurs in the device that classifies the target data.

[0012] The state of the carrier of the predicted classified data at the next moment includes: the carrier is the object for the device to collect classified data; the state at the next moment includes: the state data of the carrier and the classification of the state data of the carrier; the state data of the carrier includes: security alarm data, image data and video data; the classification of the state data of the carrier includes: security alarm data classification, target image data classification and target video data classification.

[0013] The beneficial effects of the present invention are:

[0014] The present invention provides a tunnel monitoring data analysis method based on long-term monitoring data, which has the following advantages: the present invention divides tunnel monitoring data into continuous data and classified data, and provides analysis methods for continuous data and classified data respectively; and provides a method for processing different tunnel monitoring data using different interpolation methods and machine learning methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 : A flow chart of a tunnel monitoring data analysis method based on long-term monitoring data of the present invention. DETAILED DESCRIPTION

[0016] The specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. It should be understood that the specific embodiments given here are only used to illustrate and explain the present invention, and cannot be used to limit the present invention.

[0017] like Figure 1As shown in the figure, it is a flow chart of a tunnel monitoring data analysis method based on long-term monitoring data of the present invention; the flow chart includes: step S100, obtaining the long-term monitoring data of the target tunnel in history, and dividing the long-term monitoring data into: continuous data and classified data; the continuous data refers to data that changes continuously, including: displacement data, strain data and temperature data; the classified data refers to data that does not change continuously and is used for classification, including: safety alarm data, image data and video data; step S200, the analysis of continuous data includes the establishment of a continuous data change curve and the prediction of continuous data; the establishment of the continuous data change curve is processed by using a corresponding interpolation algorithm according to the form of the data to generate a continuous data change curve corresponding to the continuous data, and the interpolation algorithm used includes: line interpolation, Newton interpolation and spline interpolation; the prediction of continuous data adopts methods including time series analysis, regression analysis and machine learning to predict the data of the next moment of the continuous data change curve; step S300, the analysis of classified data includes observing whether the classified data has anomalies and predicting the classified data; the observation of whether the classified data has anomalies refers to the problem that the collected classified monitoring data has classification and data correspondence errors, including: security alarm errors, image data classification errors and video data classification errors; the method of observing whether the classified data has anomalies is: verifying the collected classified data through neural network technology; the prediction of classified data includes: predicting the classified data through deep learning technology, and predicting the state of the carrier of the classified data at the next moment.

[0018] Specifically, the present invention first obtains the long-term monitoring data of the target tunnel in history, and divides the long-term monitoring data into continuous data and classified data; the analysis of continuous data includes the establishment of a continuous data change curve and the prediction of continuous data; the interpolation algorithms used include linear interpolation, Newton interpolation and spline curve interpolation to generate a continuous data change curve corresponding to the continuous data; the prediction of continuous data uses methods including time series analysis, regression analysis and machine learning to predict the data at the next moment of the continuous data change curve; the analysis of classified data is divided into observing whether the classified data has anomalies and predicting the classified data; observing whether the classified data has anomalies includes: safety alarm errors, image data classification errors and video data classification errors; the method of observing whether the classified data has anomalies is: verifying the collected classified data through neural network technology; the prediction of classified data includes: predicting the classified data through deep learning technology to predict the state of the carrier of the classified data at the next moment.

[0019] Step S100, obtaining the long-term monitoring data of the target tunnel in history, and dividing the long-term monitoring data into: continuous data and classified data; the continuous data refers to data that changes continuously, including: displacement data, strain data and temperature data; the classified data refers to data that does not change continuously and is used for classification, including: safety alarm data, image data and video data.

[0020] Specifically, long-term refers to the continuous collection of data rather than a fixed period of time; factors that determine whether tunnel monitoring data is long-term include: data collection cycle, change trend and application requirements.

[0021] Step S200, the analysis of continuous data includes the establishment of a continuous data change curve and the prediction of continuous data; the establishment of the continuous data change curve is processed by a corresponding interpolation algorithm according to the form of the data to generate a continuous data change curve corresponding to the continuous data, and the interpolation algorithms used include: linear interpolation, Newton interpolation and spline curve interpolation; the prediction of continuous data uses methods including time series analysis, regression analysis and machine learning to predict the data of the continuous data change curve at the next moment.

[0022] Specifically, the temperature data and humidity data are processed by a linear interpolation algorithm, the displacement data and vibration data are processed by a Newton interpolation algorithm, and the strain data and time series data are processed by a spline interpolation algorithm; the corresponding data are processed by the interpolation algorithm to obtain a continuous data change curve of the corresponding data.

[0023] In the above embodiment, specifically, what kind of prediction method is selected for a set of continuous data depends on the application object and analysis method of the data; for example: temperature, humidity, strain and vibration data show obvious time correlation and periodic changes, and time series analysis methods are often used to identify and model trends, seasonal changes and periodic changes in data, so it is possible to analyze temperature, humidity, strain and vibration data through time series analysis to predict the temperature, humidity, strain and vibration changes of the target tunnel at the next moment; the next moment is determined according to the time interval for the equipment to collect the corresponding data information of the target tunnel; for example: considering the displacement and strain of the tunnel structure and the external factors including temperature and humidity The relationship between variables is a function of the relationship between them, and regression analysis is suitable for processing the relationship between different variables in tunnel monitoring, especially those situations where one or more independent variables are used to predict dependent variables; therefore, regression analysis can be used to process data with humidity and temperature as independent variables and displacement and strain as dependent variables, and predict the corresponding data at the next moment based on the generated prediction model; machine learning methods, with flexible model architectures, non-parametric and parametric methods, the improvement of big data and computing power, the development of feature engineering and the advancement of integrated learning methods, can be used to process almost all types of data predictions, as long as the appropriate training model is selected according to the data type during the model training process.

[0024] Step S300, the analysis of the classified data includes observing whether the classified data has anomalies and predicting the classified data; the observation of whether the classified data has anomalies refers to the existence of classification and data correspondence errors in the collected classified monitoring data, including: security alarm errors, image data classification errors and video data classification errors; the method of observing whether the classified data has anomalies is: verifying the collected classified data through neural network technology; the prediction of the classified data includes: predicting the classified data through deep learning technology, and predicting the state of the carrier of the classified data at the next moment.

[0025] Specifically, verifying the collected classified data through neural network technology includes: training the correct data collected historically to obtain a neural network model; inputting the verified target data into the neural network model to obtain an output result; comparing the output result with the actual classification result to observe whether the output result and the actual classification result are the same. If they are not the same, it means that there is an error in the device that classifies the target data.

[0026] In the above embodiment, specifically, the carrier is an object for the device to collect classified data; the state at the next moment includes: the state data of the carrier and the classification of the state data of the carrier; the state data of the carrier includes: security alarm data, image data and video data; the classification of the state data of the carrier includes: security alarm data classification, target image data classification and target video data classification.

[0027] It should be understood that the above-mentioned embodiments are one or more embodiments of the present invention, and there are many other embodiments and variations thereof based on the present invention; the variations and modifications made by ordinary technicians in this industry through the present invention without making pioneering innovations all fall within the scope of protection of the present invention.

Claims

1. A tunnel monitoring data analysis method based on long-term monitoring data, It is characterized in that The method comprises: Acquire the historical long-term monitoring data of the target tunnel, and divide the long-term monitoring data into continuous data and classified data; the continuous data refers to data that changes continuously, including displacement data, strain data and temperature data; the classified data refers to data that does not change continuously and is used for classification, including safety alarm data, image data and video data; The analysis of continuous data includes the establishment of continuous data change curves and the prediction of continuous data; the establishment of continuous data change curves uses corresponding interpolation algorithms to process the data in the form of data to generate continuous data change curves corresponding to the continuous data. The interpolation algorithms used include: linear interpolation, Newton interpolation and spline curve interpolation; the prediction of continuous data uses methods including time series analysis, regression analysis and machine learning to predict the data at the next moment of the continuous data change curve; The analysis of classified data includes observing whether there are anomalies in the classified data and predicting the classified data; the observation of whether there are anomalies in the classified data refers to the existence of classification and data correspondence errors in the collected classified monitoring data, including: security alarm errors, image data classification errors and video data classification errors; the method of observing whether there are anomalies in the classified data is: verifying the collected classified data through neural network technology; the prediction of classified data includes: predicting the classified data through deep learning technology, and predicting the state of the carrier of the classified data at the next moment.

2. A tunnel monitoring data analysis method based on long-term monitoring data according to claim 1, It is characterized in that The acquisition of the target tunnel's historical long-term monitoring data includes: the long-term refers to the continuous collection of data, rather than a fixed period of time; factors for determining whether the tunnel monitoring data is long-term include: data collection cycle, change trend and application requirements.

3. The tunnel monitoring data analysis method based on long-term monitoring data according to claim 1, It is characterized in that The establishment of the continuous data change curve adopts the corresponding interpolation algorithm for processing according to the form of the data, including: temperature data and humidity data are processed by linear interpolation algorithm, displacement data and vibration data are processed by Newton interpolation algorithm, and strain data and time series data are processed by spline interpolation algorithm; the corresponding data are processed by interpolation algorithm to obtain the continuous data change curve of the corresponding data.

4. The tunnel monitoring data analysis method based on long-term monitoring data according to claim 1, It is characterized in that The prediction of the continuous data adopts methods including time series analysis, regression analysis and machine learning to predict the data at the later stage of the continuous data change curve, including: what kind of prediction method is selected for a set of continuous data depends on the application object and analysis method of the data; for example: temperature, humidity, strain and vibration data show obvious time correlation and periodic changes, and time series analysis methods are often used to identify and model trends, seasonal changes and periodic changes in data, so it is possible to analyze the temperature, humidity, strain and vibration data through time series analysis to predict the temperature, humidity, strain and vibration changes of the target tunnel at the next moment; the next moment is determined according to the time interval for the equipment to collect the corresponding data information of the target tunnel; for example: considering the tunnel structure The relationship between the displacement and strain of the structure and external factors including temperature and humidity, and regression analysis is suitable for processing the relationship between different variables in tunnel monitoring, especially those cases where one or more independent variables are used to predict dependent variables; therefore, regression analysis can be used to process data with humidity and temperature as independent variables and displacement and strain as dependent variables, and predict the corresponding data at the next moment based on the generated prediction model; machine learning methods, with flexible model architecture, non-parametric and parametric methods, the improvement of big data and computing power, the development of feature engineering and the advancement of integrated learning methods, can be used to process almost all types of data predictions, as long as the appropriate training model is selected according to the data type during the model training process.

5. The tunnel monitoring data analysis method based on long-term monitoring data according to claim 1, It is characterized in that The verification of the collected classification data through the neural network technology includes: training the correct data collected historically to obtain a neural network model; inputting the verified target data into the neural network model to obtain an output result; comparing the output result with the actual classification result to observe whether the output result and the actual classification result are the same. If they are not the same, it means that an error occurs in the device that classifies the target data.

6. The tunnel monitoring data analysis method based on long-term monitoring data according to claim 1, It is characterized in that The state of the carrier of the predicted classified data at the next moment includes: the carrier is the object for the device to collect classified data; the state at the next moment includes: the state data of the carrier and the classification of the state data of the carrier; the state data of the carrier includes: security alarm data, image data and video data; the classification of the state data of the carrier includes: security alarm data classification, target image data classification and target video data classification.

Images