A prediction model and monitoring system for settlement of deep reclamation areas of heliports

Abstract

The application discloses a kind of for the prediction model and monitoring system of heliport deep backfill area settlement, including the following steps: hierarchical settlement monitoring is carried out to deep backfill area and obtains real-time settlement data, according to the real-time settlement data screening of field situation, eliminate abnormal data;Sensitivity analysis parameter set is obtained and finite element analysis model is constructed, the analysis model is selected to the sensitivity analysis parameter set is analyzed and outputs sensitivity analysis simulation data, and the sensitivity analysis simulation data includes the creep deformation data of deep backfill area;According to the sensitivity analysis simulation data and the real-time settlement data are simulated and analyzed, and settlement law data set is obtained;According to grey system theory and neural network, the settlement law data set is trained and learned to form settlement combined prediction model. Through prediction model, consolidation settlement and creep settlement can be simultaneously predicted. The problem that the prediction data precision of prediction model to settlement area is low is solved.

Description

Technical Field

[0001] This invention relates to the field of transportation construction engineering technology, and in particular to a predictive model and monitoring system for settlement in deep backfill areas of helipads. Background Technology

[0002] With my country's economic development, helicopters have become a major means of transportation. This development necessitates the construction of helipads to support their operation, enabling helicopters to provide onshore and offshore oil services, marine monitoring, aerial photography, and pilotage operations to domestic and international clients. Due to the large area required for helipads, they are typically located in suburban areas. In cities like Shenzhen, where land is scarce, helipads are often situated in mountainous or forested areas. These locations often feature significant elevation differences and complex terrain, requiring extensive earthwork, leveling, and compaction, including deep backfilling. Furthermore, the predominantly clayey soil in Shenzhen demands high-quality construction to prevent large-scale subsidence later on.

[0003] Due to the construction site and geographical environment of helipads, large areas of deep backfill are often encountered during construction. Since the soil in mountainous areas is mostly cohesive, this poses a potential risk of settlement in the deep backfill areas later on. In order to reduce the impact of large-scale settlement in the deep backfill areas on the normal operation of the helipad, settlement prediction is usually carried out in the deep backfill areas of the helipad before and after construction to control the settlement as much as possible. However, most existing settlement prediction models only predict one of the two: consolidation settlement or creep settlement. In the case of deep backfill areas in the mountainous areas of Shenzhen, not only consolidation settlement but also creep settlement will occur. Therefore, the accuracy of the prediction data obtained by using a single settlement prediction model needs to be improved.

[0004] Therefore, it is necessary to propose a prediction model and monitoring system for settlement in deep backfill areas of helicopter airports. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a prediction model and monitoring system for settlement in deep backfill areas of helicopter airports.

[0006] On the one hand, the technical solution of the present invention is as follows:

[0007] A predictive model for settlement in deep backfill areas of helicopter airports includes the following steps: performing stratified settlement monitoring of the deep backfill area and obtaining real-time settlement data; filtering the real-time settlement data according to the site conditions and removing abnormal data.

[0008] A set of sensitivity analysis parameters is obtained and a finite element analysis model is constructed. The finite element analysis model is selected to analyze the set of sensitivity analysis parameters and output sensitivity analysis simulation data, which includes creep deformation data of the deep backfill area.

[0009] Among them, a dataset of settlement patterns is obtained by performing simulation analysis based on the sensitivity analysis simulation data and the real-time settlement data.

[0010] The settlement pattern dataset is trained and learned using grey system theory and neural networks to form a joint settlement prediction model.

[0011] In one possible implementation, constructing the finite element analysis model includes: performing finite element mesh generation on the finite element analysis model to discretize the continuous structure into a finite number of elements.

[0012] In one possible implementation, the set of sensitivity analysis parameters includes at least the filling height parameter, the soil-to-rock ratio parameter, the foundation treatment method parameter, the structural analysis parameter, and the thermal analysis parameter in the deep backfill area, and a finite element analysis model is constructed based at least on the soil-to-rock ratio parameter and the foundation treatment method parameter.

[0013] In one possible implementation, the step of selecting the finite element analysis model to analyze the time set of sensitivity analysis parameters and outputting sensitivity analysis simulation data further includes: performing stratification on the deep backfill area, with each layer undergoing creep according to a creep formula, wherein the creep formula is: In the formula The degree of creep deformation, This represents the degree of creep deformation corresponding to the creep time t. The unit time is the same as the creep deformation time t, and n is the viscosity coefficient of the soil in the deep backfill area.

[0014] In one possible implementation, the method further includes stratifying the deep backfill area, with each layer undergoing creep according to a creep formula, and deriving a creep settlement formula for the deep backfill area using a stratified summation method. The creep settlement formula is as follows: In the formula The creep settlement is defined as the unit creep settlement with the same dimensions as the creep settlement s, with a value of 1, and n is the creep rate parameter of the deep backfill zone. This represents the final creep settlement under the current load, which is in a highly consolidated state. This represents the final creep settlement under the current load and in a normal consolidation state.

[0015] In one possible implementation, the step of training and learning the settlement law dataset based on grey system theory and neural networks to form a joint settlement prediction model includes at least: the joint settlement prediction model predicts the deep backfill area and outputs prediction data, and inputs real-time settlement data into the joint settlement prediction model and compares and verifies it with the prediction data.

[0016] In one possible implementation, before training and learning the settlement pattern dataset to form a joint settlement prediction model, the method further includes performing isochronous processing on the settlement pattern dataset using cubic spline interpolation.

[0017] In one possible implementation, the settling pattern dataset is further included as input to the model of the grey system theory. After accumulating the generated sequence And calculate the predicted value of the grey system theory. .

[0018] In one possible implementation, the input values ​​of the grey system theory are also included. and predicted value The real-time settlement data serves as the input value to the model of the neural network, and the model outputs the predicted value after setting parameters and training the neural network.

[0019] On the other hand, the present invention also provides a monitoring system applicable to the above-mentioned prediction model for settlement of deep backfill areas in helicopter airports, comprising at least a settlement monitoring module for monitoring the settlement amount of deep backfill areas and a horizontal monitoring module for monitoring the horizontal displacement of deep backfill areas.

[0020] According to the above-described scheme, the beneficial effects of this invention are as follows: It provides a prediction model and monitoring system for settlement in deep backfill areas of helicopter airports. The invention offers a prediction model for settlement in deep backfill areas of helicopter airports. By collecting a set of sensitivity analysis parameters with creep deformation data through sensitivity analysis, it is possible to better understand the model's behavior and performance, as well as the degree of influence of input parameters on the model's output, thereby better optimizing and improving the model. Analyzing the sensitivity analysis parameter set through a finite element analysis model helps to better understand the degree of influence of various parameters on the final result of the prediction model in practical problems. By using a finite element analysis model to simulate and analyze sensitivity analysis data and real-time settlement data to form a settlement law dataset, it provides data support for the prediction model, including consolidation settlement laws and creep settlement laws. Furthermore, by learning from the settlement law dataset containing consolidation and creep settlement laws through grey system theory and neural networks, this invention can simultaneously predict both consolidation and creep settlement in deep backfill areas of helicopter airports, thereby improving the accuracy of predicting settlement trend data in deep backfill areas of helicopter airports. Attached Figure Description

[0021] Figure 1 This is a flowchart of a prediction model for settlement in deep backfill areas of helicopter airports according to the present invention. Detailed Implementation

[0022] The present invention will now be further described with reference to the accompanying drawings and embodiments:

[0023] It should be noted that the mountainous soil in Shenzhen is highly compressible clay with a high proportion of organic soil. Deep backfill areas formed by this type of soil are prone to settlement in the later stages. Existing settlement prediction models usually only predict consolidation settlement or creep settlement. When dealing with deep backfill areas formed by mountainous soil in Shenzhen, the data predicted by existing settlement prediction models are not accurate enough, resulting in a significant risk of settlement in the later stages of the construction of the heliport. Therefore, it is necessary to propose a prediction model for settlement of deep backfill areas of heliports to improve the accuracy of settlement prediction and thus predict the settlement after the construction of heliports in Shenzhen.

[0024] like Figure 1 As shown, a prediction model for settlement in deep backfill areas of helicopter airports includes the following steps:

[0025] S100. Conduct stratified settlement monitoring of the deep backfill area and obtain real-time settlement data. Filter the real-time settlement data according to the site conditions and remove abnormal data.

[0026] In this embodiment, by dividing the deep backfill area into layers and establishing a settlement monitoring system within the deep backfill area, real-time settlement data of each layer is collected. The collected settlement data is then filtered according to the site conditions to remove abnormal data, thereby avoiding errors in the prediction results caused by abnormal data and improving the authenticity and accuracy of the settlement data. This, to a certain extent, improves the prediction results of the present invention.

[0027] S200. Obtain the sensitivity analysis parameter set and construct a finite element analysis model. Select the finite element analysis model to analyze the sensitivity analysis parameter set and output the sensitivity analysis simulation data. The sensitivity analysis simulation data includes creep deformation data of the deep backfill area.

[0028] In this embodiment, relevant parameters are set to form a sensitivity analysis parameter set, and a finite element analysis model is constructed. By determining the sensitivity analysis parameter set, the analysis direction of the model can be initially determined. The sensitivity analysis parameter set is analyzed by the finite element analysis model, and sensitivity analysis simulation data is output. The sensitivity analysis simulation data reflects the analysis accuracy of the finite element analysis model. Based on this, the finite element analysis model is optimized and modified to improve the analysis accuracy of the finite element analysis model, so as to output sensitivity analysis simulation data that is closer to the actual situation. This provides more reliable data support for the subsequent settlement joint prediction model, thereby improving the prediction accuracy of the prediction model of this invention to a certain extent.

[0029] It should be explained that the sensitivity analysis parameters include creep deformation data for deep backfill areas. This creep deformation data primarily includes parameters affecting creep in backfill soil, such as backfill depth and volume, soil properties and filling materials, geological environment and foundation conditions, applied loads, and climate and hydrological conditions. Specifically, the filling height and volume of high embankments will experience long-term creep settlement under their own weight stress. The physical and mechanical properties of the filling materials in high embankments are complex, such as uneven moisture content distribution and difficulty in unifying particle size distribution, which also affect the creep properties and settlement deformation of the soil. Due to the complex terrain of the mountainous areas in Shenzhen, the bottom of deep backfill locations is covered with soft soil layers of varying thickness and unevenness. Combined with the self-weight stress of the high embankment foundation, this easily causes creep in the soft soil layers, leading to unstable settlement. Furthermore, practical load factors and climatic and hydrological conditions also affect the settlement and deformation of the foundation at deep backfill sites. This is especially true when geological conditions and backfill materials are difficult to unify, where the impact of load factors and climatic and hydrological conditions on deep backfill is even greater. Specifically, by modifying and setting different sensitivity analysis parameter sets and inputting different parameter values, several parameter data sets are generated. These parameter data sets are then analyzed using a finite element model to output sensitivity analysis simulation data. Based on the sensitivity analysis simulation data, it can be determined which key parameter values ​​have a more significant impact on the finite element analysis model. This data serves as a basis for necessary corrections and optimizations to the finite element analysis model. Simultaneously, by analyzing the sensitivity analysis simulation data from different runs, it is possible to determine which parameter data has the greatest and least impact on creep in the deep backfill area.

[0030] S210. The construction of the finite element analysis model includes: performing finite element meshing on the finite element analysis model to discretize the continuous structure into a finite number of elements.

[0031] In this embodiment, when constructing the finite element analysis model, the deep backfill area model of the helicopter airport is discretized into a finite number of elements by using a finite mesh. This helps to simulate and consider the non-homogeneity and anisotropy among the backfill soil in more detail, thus providing more accurate data for the settlement joint prediction model. This allows for a more realistic simulation of the creep movement of the backfill soil, improving the creep characteristics of the model. Consequently, when the finite element analysis model simulates creep settlement, it can improve the accuracy of the creep settlement law in the sensitivity analysis simulation data, thereby improving the prediction accuracy of the settlement joint prediction model to a certain extent.

[0032] S211. The set of sensitivity analysis parameters includes at least the filling height parameter, the soil-to-rock ratio parameter, the foundation treatment method parameter, the structural analysis parameter, and the thermal analysis parameter in the deep backfill area, and a finite element analysis model is constructed based at least on the soil-to-rock ratio parameter and the foundation treatment method parameter.

[0033] In this embodiment, the sensitivity analysis parameter set includes at least the filling height parameter, the soil-to-rock ratio parameter, the foundation treatment method parameter, structural analysis parameters, and thermal analysis parameters for the deep backfill area of ​​the helipad. The thicker the filling height of the deep backfill area, the greater the pressure on the backfill soil, and the greater the settlement of the backfill soil. The soil-to-rock ratio of the deep backfill area of ​​the helipad can be at least 100:0, 70:30, 50:50, or 30:70. Different soil-to-rock ratios will lead to different settlement of the embankment. Specifically, as the percentage of gravel in the backfill soil increases, both the settlement of the backfill soil and the total settlement decrease significantly. However, when the soil-to-rock mixing ratio of the backfill soil is 50:50, the decrease in settlement is no longer linear as the specific gravity of the gravel increases. Foundation treatment methods can increase the bearing capacity and stability of the deep backfill area by replacing the bottom soft soil layer with strong and stable materials such as gravel, sand, and concrete. Alternatively, dynamic compaction of the bottom of the deep backfill area can densify the foundation, improving its bearing capacity and reducing the probability of settlement. Structural analysis parameters mainly include the structural parameters of the deep backfill area, such as the shape, displacement, and stress of the structure, as well as parameters under static and dynamic loads. Thermal analysis parameters include the structure's temperature distribution, heat flow, and thermal stress. Sensitivity analysis parameters comprehensively cover a variety of parameters, including those related to the environment the deep backfill area may encounter. Under the analysis of the finite element analysis model, input data that is more accurate than actual conditions is provided, thus offering more precise data for the settlement prediction model.

[0034] S220. The step of selecting the finite element analysis model to analyze the sensitivity analysis parameter set and outputting sensitivity analysis simulation data further includes: performing stratification on the deep backfill area, with each layer undergoing creep according to the creep formula, wherein the creep formula is: In the formula The degree of creep deformation, This represents the degree of creep deformation corresponding to the creep time t. The unit time is the same as the creep deformation time t, and n is the viscosity coefficient of the soil in the deep backfill area.

[0035] In this embodiment, the deep backfill area is layered to form several layers. Each layer undergoes creep deformation using a creep formula. The degree of creep deformation for each layer can be determined based on the creep formula, and the expression for the degree of creep deformation is as follows: Wherein, when the creep time t=0, =0; when creep time , = That is, creep deformation increases over time and eventually reaches a steady value. .

[0036] S221, further comprising performing stratification of the deep backfill area as described above, wherein each layer undergoes creep according to the creep formula, and the creep settlement formula of the deep backfill area is derived by combining the stratification summation method, wherein the creep settlement formula is: In the formula The creep settlement is defined as the unit creep settlement with the same dimensions as the creep settlement s, with a value of 1, and n is the creep rate parameter of the deep backfill zone. This represents the final creep settlement under the current load, which is in a highly consolidated state. This represents the final creep settlement under the current load and in a normal consolidation state.

[0037] In this embodiment, the creep settlement formula of the deep backfill area is derived by combining the creep deformation degree of each layer of the deep backfill area with the sum of the layers, thereby calculating the creep settlement of the deep backfill area. The data obtained from the creep settlement calculation is then analyzed by the finite element analysis model to obtain the creep settlement law of the deep backfill area of ​​the helicopter airport.

[0038] S300, wherein, based on the sensitivity analysis simulation data and the real-time settlement data, a settlement law dataset is obtained through simulation analysis;

[0039] In this embodiment, by inputting real-time settlement data into the finite element analysis model, the model performs simulation analysis based on the real-time settlement data and sensitivity analysis simulation data to obtain the settlement law of the deep backfill area of ​​the helicopter airport.

[0040] It needs to be explained that sensitivity analysis simulation data is obtained by analyzing the sensitivity analysis parameter set through a finite element analysis model. This data is the sum of deep backfill area settlement data obtained from different sensitivity analysis parameter sets analyzed by the finite element analysis model. Specifically, by analyzing the sensitivity analysis simulation data, the settlement simulation situation of the deep backfill area under different set parameters can be determined. Inputting the sensitivity analysis simulation data and real-time settlement data into the finite element analysis model for simulation analysis yields the settlement patterns that may occur in the deep backfill of the helipad under different parameter conditions, forming a settlement pattern dataset. This dataset includes at least consolidation settlement and creep settlement patterns, allowing for the analysis of the settlement patterns and deformation trends of the deep backfill area.

[0041] S400. Based on grey system theory and neural networks, the settlement law dataset is trained and learned to form a joint settlement prediction model.

[0042] In this embodiment, the settlement pattern dataset is trained and learned using grey system theory and neural networks to form a joint settlement prediction model for the deep backfill area of ​​the helipad. The joint settlement prediction model can predict the settlement pattern and deformation trend of the deep backfill area of ​​the helipad, thereby allowing for early adjustment of the construction plan for the deep backfill area and avoiding large-scale settlement problems at the helipad during or after construction.

[0043] It should be explained that, in this embodiment, a joint settlement prediction model is formed by training and learning the settlement law dataset based on grey system theory and neural networks. The grey system theory GM(1,1) model has the advantage of weakening data randomness and generating regular data sequences. Therefore, the grey system theory GM(1,1) model is used to learn the consolidation settlement law in the settlement law dataset, enabling the model to better learn the characteristics of the consolidation settlement law, thus training the grey system theory GM(1,1) model. Furthermore, the BP neural network model has strong fault tolerance and strong adaptive performance, and can self-learn to adapt to changing external environments. Therefore, the BP neural network model is used to learn the creep settlement law in the settlement law dataset, enabling the model to better learn the characteristics of the creep settlement law, thus training the BP neural network model. By using a settlement pattern dataset as input variables, the grey system theory model and the neural network model learn and train on the features in the settlement pattern dataset. This enables the joint settlement prediction model to simultaneously predict both consolidation settlement and creep settlement in the deep backfill area of ​​a helicopter airport. Compared to existing prediction models that can only predict either consolidation settlement or creep settlement in the deep backfill area of ​​a helicopter airport, the settlement prediction model in this invention has both types of prediction data. Therefore, this invention can improve the accuracy of settlement prediction data for the deep backfill area of ​​a helicopter airport.

[0044] S410. The step of training and learning the settlement law dataset based on grey system theory and neural network to form a joint settlement prediction model includes at least the following: the joint settlement prediction model predicts the deep backfill area and outputs prediction data, and inputs real-time settlement data into the joint settlement prediction model and compares and verifies it with the prediction data.

[0045] In this embodiment, the settlement joint prediction model is applied to the deep backfill area of ​​a helicopter airport to obtain prediction data. This predicted data is then compared and verified with real-time settlement data to validate the accuracy and reliability of the settlement joint prediction model. Based on the comparison results, the settlement joint prediction model is adjusted and optimized to improve its prediction accuracy. Furthermore, the comparison results reveal whether the settlement joint prediction model is adaptable and practical under different times and conditions, allowing for further improvement and expansion of the model based on the data.

[0046] S420. Before training and learning the settlement pattern dataset to form a joint settlement prediction model, the method further includes performing isochronous processing on the settlement pattern dataset using cubic spline interpolation.

[0047] In this embodiment, the settlement pattern dataset contains data with non-equidistant time intervals. Cubic spline interpolation can transform the original non-equidistant settlement data sequence into an equidistant data sequence, making the data sequence in the settlement pattern dataset more regular and consistent, thereby improving the prediction accuracy of the grey system theory GM(1,1) model. Therefore, using cubic spline interpolation to perform equidistant processing on the settlement pattern dataset can better adapt to the modeling requirements of the grey system theory GM(1,1) model and BP neural network, improving the model's prediction accuracy and reliability.

[0048] S430, further includes: the settlement law dataset as input values ​​for the gray system theory model. After accumulating the generated sequence And calculate the predicted value of the grey system theory. .

[0049] In this embodiment, the settlement pattern dataset after cubic spline interpolation is used as the input value of the grey system theory GM(1,1) model. And by accumulating the generated sequence The GM(1,1) model is based on the input value and sequence Calculations are performed to obtain the predicted values ​​from the grey system theory. It needs to be explained that grey system theory defines all random variables as grey quantities that vary within a certain range. It uses data generation methods to integrate these grey quantities into a highly regularized generated sequence, and finally builds a model for prediction. This method is particularly suitable for solving system problems with limited or uncertain information. Specifically, by performing cubic spline interpolation on the settlement law dataset input to grey system theory, the non-equidistant time interval data in the dataset can be processed, transforming it into equidistant data with equal time intervals. This facilitates the processing of the settlement law dataset by grey system theory, leading to a better understanding and prediction of the dynamic development trend of settlement.

[0050] S440, also includes the input values ​​of the gray system theory. and predicted value The real-time settlement data serves as the input value to the model of the neural network, and the model outputs the predicted value after setting parameters and training the neural network model.

[0051] In this embodiment, after processing the sedimentation law dataset using the grey system theory GM(1,1) model, the input values ​​obtained can be used as input values ​​for the BP neural network model. and predicted value This study utilizes real-time settlement data as the expected output value of a BP neural network model. By setting parameters, training the BP neural network model, and outputting predicted values, settlement prediction for deep backfill areas of helicopter airports can be achieved. It's important to note that BP neural network models possess strong fault tolerance and adaptability, enabling them to learn and adapt to changing external environments. Using real-time settlement data as the expected output value of the BP neural network model leverages its adaptability and fault tolerance to achieve more accurate settlement predictions. Furthermore, training the BP neural network model allows for finding the optimal combination point between the grey system theory GM(1,1) model and the neural network model, resulting in a joint prediction model with higher accuracy and reliability. Finally, by comprehensively utilizing the advantages of grey system theory and neural networks, efficient and accurate prediction of complex, multi-factorial unknown systems can be achieved, providing a more reliable basis and support for construction schemes in deep backfill areas of helicopter airports.

[0052] According to another aspect of the present invention, a monitoring system is provided, applicable to the aforementioned prediction model for settlement of deep backfill areas in helicopter airports, comprising at least a settlement monitoring module for monitoring the settlement amount of the deep backfill area and a horizontal monitoring module for monitoring the horizontal displacement of the deep backfill area.

[0053] In this embodiment, the settlement monitoring module includes at least a Global Positioning System (GPS). GPS technology can be used to monitor the settlement of the helipad, including the amount and rate of settlement after filling, thus providing accurate real-time settlement data. Additionally, the horizontal detection module includes at least Real-Time Kinematic (RTK) monitoring. RTK monitoring can detect the horizontal displacement of the helipad, including lateral and longitudinal displacement. By monitoring horizontal displacement, the stability of the embankment can be determined, and accurate horizontal displacement data can be provided for real-time settlement data. By setting up the monitoring system, real-time data from the helipad can be collected, providing more accurate data references for the joint settlement prediction model.

[0054] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

[0055] The present invention has been described above with reference to the accompanying drawings. Obviously, the implementation of the present invention is not limited to the above-described manner. Any improvements made using the inventive concept and technical solution of the present invention, or the direct application of the inventive concept and technical solution of the present invention to other situations without modification, are all within the protection scope of the present invention.

Claims

1. A predictive model for settlement in deep backfill areas of helicopter landing pads, characterized in that, The steps include: conducting stratified settlement monitoring of the deep backfill area and obtaining real-time settlement data; filtering the real-time settlement data according to the site conditions and removing abnormal data. A sensitivity analysis parameter set is obtained and a finite element analysis model is constructed. The finite element analysis model is selected to analyze the sensitivity analysis parameter set and output sensitivity analysis simulation data. The sensitivity analysis parameter set includes at least the filling height parameter, the soil-rock ratio parameter, the foundation treatment method parameter, the structural analysis parameter, and the thermal analysis parameter in the deep backfill area. The finite element analysis model is constructed based at least on the soil-rock ratio parameter and the foundation treatment method parameter. The sensitivity analysis simulation data includes creep deformation data of the deep backfill area. The process of selecting the finite element analysis model to analyze the sensitivity analysis parameter set and output sensitivity analysis simulation data also includes: performing stratification on the deep backfill area, with each layer undergoing creep according to the creep formula, wherein the creep formula is: In the formula The degree of creep deformation, This represents the degree of creep deformation corresponding to the creep time t. Let n be the unit time with the same dimension as the creep time t, and n be the viscosity coefficient of the soil in the deep backfill area. Based on the layered treatment of the deep backfill area, each layer undergoes creep according to the creep formula. Combining the layered summation method, the creep settlement formula for the deep backfill area is derived. The creep settlement formula is: In the formula The creep settlement is defined as the unit creep settlement with the same dimensions as the creep settlement s, with a value of 1, and n is the creep rate parameter of the deep backfill zone. This represents the final creep settlement under the current load, which is in a highly consolidated state. This represents the final creep settlement under the current load and in a normal consolidation state. Among them, a dataset of settlement patterns is obtained by performing simulation analysis based on the sensitivity analysis simulation data and the real-time settlement data. The settlement pattern dataset is trained and learned using grey system theory and neural networks to form a joint settlement prediction model, which includes at least: the joint settlement prediction model predicts the deep backfill area and outputs prediction data; real-time settlement data is input into the joint settlement prediction model and compared and verified with the prediction data; before training and learning the settlement pattern dataset to form the joint settlement prediction model, cubic spline interpolation is used to perform equal time interval processing on the settlement pattern dataset.

2. The prediction model for settlement in deep backfill areas of helicopter landing pads according to claim 1, characterized in that, The construction of the finite element analysis model includes: performing finite element meshing on the finite element analysis model to discretize the continuous structure into a finite number of elements.

3. The prediction model for settlement in deep backfill areas of helicopter landing pads according to claim 1, characterized in that, Also includes: The sedimentation law dataset serves as the input value for the grey system theory model. After accumulating the generated sequence And calculate the predicted value of the grey system theory. .

4. The prediction model for settlement in deep backfill areas of helicopter landing pads according to claim 3, characterized in that, It also includes the input values ​​of the grey system theory. and predicted value The real-time settlement data serves as the input value to the model of the neural network, and the model outputs the predicted value after setting parameters and training the neural network.

5. A monitoring system applicable to the prediction model for settlement in deep backfill areas of helicopter landing pads as described in any one of claims 1-4, characterized in that, It includes at least a settlement monitoring module for monitoring the settlement of deep backfill areas and a horizontal monitoring module for monitoring the horizontal displacement of deep backfill areas.

Images