CN-120911278-B - Frame house protection structure optimization method under landslide induction effect of slope cutting building

CN120911278BCN 120911278 BCN120911278 BCN 120911278BCN-120911278-B

Abstract

The invention discloses a frame house protection structure optimization method under the effect of landslide induction by building a cut slope, which comprises the steps of collecting and preprocessing multisource space-time data of a house back slope and a back wall, dividing and extracting sliding windows and features to obtain partition state parameters, establishing a Bayesian network model, reasoning risk probability of each partition structure, optimizing the distribution of partitions, windows and reinforcing ribs according to the risk probability distribution, forming a construction scheme and implementing, continuously monitoring in an operation stage, and dynamically optimizing the structure parameters by periodically inputting data. According to the invention, by fusing multisource space-time data analysis and Bayesian network reasoning, dynamic optimization and toughness self-adaptive regulation and control of the framework house roof rear wall protection structure under the influence of landslide are realized.

Inventors

WU SHUANGSHUANG
ZOU GUANGJUN
HAO SHUAI
ZHOU SHENGTAO
ZHOU BOYU

Assignees

河海大学

Dates

Publication Date: 20260508
Application Date: 20250728

Claims (5)

1. The method for optimizing the protection structure of the frame house under the action of landslide induction during slope cutting and building is characterized by comprising the following steps of: collecting multisource space-time data sets of a house back slope and a back wall area, wherein the multisource space-time data sets comprise topography geological parameters, rainfall duration, soil moisture content, wall stress parameters and structural layout information of the house back slope and the back wall area, and preprocessing the multisource space-time data sets to generate a standardized multisource space-time data set; carrying out sliding window division on the standardized multi-source space-time data set according to a preset space step length and a time step length, and carrying out feature extraction on data in each window by utilizing a sliding window converter to obtain state feature parameters of each space partition; the state characteristic parameters of each space partition are used as node variables, a Bayesian network model is established, the state characteristic parameters are input to perform conditional probability reasoning, and the structural risk probability distribution of each space partition of the rear wall is obtained; According to the structural risk probability distribution, optimally designing structural partitions, window sizes and reinforcing rib arrangements of a rear wall protection structure, setting up a field-shaped embedded column beam structure on the rear wall, realizing wall space partition, arranging small-size lighting windows in two partitions at the upper part of the field-shaped structure, arranging closed walls or reinforcing ribs in other partitions, performing structural construction according to the optimal design, and realizing the structural partitions, the window sizes and the reinforcing rib arrangements on a construction drawing and field implementation; continuously monitoring state characteristic parameters of the rear wall partition at the house operation stage, periodically inputting new monitoring data into a sliding window converter and a Bayesian network model, and dynamically updating parameters to realize the toughness self-adaptive optimization of a house protection structure; The obtaining the structural risk probability distribution of each space partition of the rear wall comprises the following steps: Taking the obtained state characteristic parameters of each space partition as candidate node variables, dynamically dividing the space partition node structure according to the space distribution characteristics and the change trend of the state characteristic parameters, and determining a space partition node set; for the determined space partition node set, according to the abnormal change of the state characteristic parameters and the risk distribution difference, subdividing the space partition nodes with state mutation into a plurality of space partition sub-nodes, merging the space partition nodes with state high correlation or risk balance into a single space partition node, and obtaining the space partition node set; Taking each space partition node in the space partition node set as an input node, taking four monitoring parameters of ground surface deformation, strain, stress and soil moisture content as external disaster causing factor nodes, presetting an integral structure influence node in the multi-layer Bayesian network model structure as an output node of the failure probability of the integral structure, and constructing the multi-layer Bayesian network model structure comprising the space partition nodes, the external disaster causing factor nodes and the integral structure influence node; For each node in the multi-layer Bayesian network model structure, based on historical monitoring data, real-time acquisition data and external disaster-causing environment data, periodically updating the conditional probability distribution of each node to form a multi-source data-driven conditional probability table set; Inputting state characteristic parameters of each space partition node in a condition probability table set and a space partition node set, searching a corresponding state parameter combination in the condition probability table by using a Bayesian inference method, reading the conditional probability of local destruction of the space partition node under the state parameter combination, and if a plurality of father nodes or multiple observation information exist, sequentially carrying out probability weighting and normalization on the states of all the father nodes according to a probability propagation mechanism to obtain the local destruction probability of each space partition node under the current observation information; the method comprises the steps of calculating the failure probability of an overall structure influence node by adopting a probability joint reasoning function for the local failure probability of each space partition node in a space partition node set; Based on the multi-layer Bayesian network model structure and the local destruction probability of each space partition node in the space partition node set, analyzing the correlation among the space partition nodes, establishing network edge connection among the space partition nodes with obvious correlation, deleting the network edge connection among the space partition nodes with low correlation, and obtaining a Bayesian network model structure after dynamic adjustment; outputting the local damage probability of each space partition node in the space partition node set, the overall structure failure probability of the overall structure influence node, and the structure risk probability distribution corresponding to the dynamically adjusted Bayesian network model structure, and obtaining the structure risk probability distribution of each space partition of the back wall.
2. The method for optimizing the protection structure of the frame house under the action of landslide induction by building a cut slope according to claim 1, wherein the preprocessing of the multi-source space-time data set specifically comprises the steps of format unification, missing value filling, outlier rejection and data standardization processing of the multi-source space-time data set.
3. The method for optimizing the protection structure of the frame house under the action of landslide induction by building a cut slope according to claim 1, wherein the step of obtaining the state characteristic parameter of each space partition comprises the following steps: Calculating the change rate of four monitoring parameters of the earth surface deformation, strain, stress and soil moisture content in a continuous time interval respectively for each space partition and each time interval in a standardized multisource space-time data set, weighting and fusing the change rate of the four monitoring parameters according to the earth surface deformation importance weight, the strain importance weight, the stress importance weight and the soil moisture content importance weight to obtain joint change perception indexes corresponding to each space partition and each time interval, and judging whether sudden extreme changes exist in the four monitoring parameters of the earth surface deformation, the strain, the stress and the soil moisture content to obtain sudden change monitoring results corresponding to each space partition and each time interval; According to the joint change sensing index and the sudden change monitoring result, carrying out sliding window self-adaptive division on each space partition and each time interval, when the joint change sensing index is higher than a preset threshold or the sudden change monitoring result is positive, adopting a first space window size and a first time window step length to carry out sliding window division on the space partition and the time interval to obtain a first type of sliding window, when the joint change sensing index is lower than or equal to the preset threshold and the sudden change monitoring result is negative, adopting a second space window size and a second time window step length to carry out sliding window division on the space partition and the time interval to obtain a second type of sliding window, wherein the first space window size and the first time window step length are preset fine granularity parameters, the second space window size and the second time window step length are preset wide scale parameters, and according to the principles of space position continuity and time sequence continuity, identifying directly adjacent sliding window pairs in all sliding windows in space or time, and establishing adjacent sliding window sets; Extracting global statistical features, local extremum features, change rate features and segmentation interval features in the sliding windows according to the spatial scale and the time scale respectively for the first type sliding window and the second type sliding window to generate a multi-scale feature subset; performing feature stitching and weighted fusion on the extracted global statistical features, local extremum features, change rate features and segmented interval features to generate multi-scale feature vectors of each sliding window; Executing a sliding window overlapping strategy on each pair of the sliding windows which are directly adjacent in space or time in the adjacent sliding window sets, and inwards extending the boundary between the adjacent sliding windows according to a set overlapping length on the basis of maintaining the original window boundary sequence to form a sliding window pair set with a data overlapping area in a space position or time sequence; Carrying out trend extrapolation on the multi-scale feature vector of the sliding window boundary by utilizing the change trend of the overlapping region data, fusing the multi-scale feature vector generated by the trend extrapolation with the multi-scale feature vector of the original overlapping region in a Bayesian smooth weight mode to obtain a multi-scale feature vector after boundary continuity processing and anomaly correction, and correcting the multi-scale feature vector after boundary continuity processing and anomaly correction and the multi-scale feature vector of the corresponding adjacent sliding window in a residual error correction mode if feature mutation or anomaly residual error occurs in the overlapping region data; Summarizing the multi-scale feature vector and the multi-scale feature vector subjected to boundary continuity processing and anomaly correction to generate multi-dimensional state feature parameters corresponding to each space partition and each time interval, and obtaining the state feature parameters of each space partition.
4. The method for optimizing the protection structure of a frame house under the action of landslide induction of a cut-slope building as defined in claim 1, wherein the optimizing design is performed on the structural partition, window size and reinforcing rib arrangement of the rear wall protection structure according to the structural risk probability distribution, a field-shaped embedded column beam structure is established on the rear wall to realize the spatial partition of the wall, small-size lighting windows are arranged in two upper partitions of the field-shaped structure, the other partitions are provided with closed walls or reinforcing ribs, the structural construction is performed according to the optimizing design, and the structural partition, window size and reinforcing rib arrangement are implemented in a construction diagram and field implementation, and the method comprises the following steps: According to the structural risk probability distribution of each space partition, a structural protection optimization objective function is established by combining the space partition parameters, the window size parameters and the reinforcing rib arrangement parameters of the rear wall structure; According to the structural protection optimization objective function, constraint optimization is carried out on space partition parameters, window size parameters and reinforcing rib arrangement parameters, the number and the size of space partitions, the window size and the position, the allowable range of a reinforcing rib arrangement mode and design constraint are limited in the optimization process according to structural design specifications and safety standards, and on the premise that all constraint conditions are met, an optimal space partition scheme, an optimal window size scheme and an optimal reinforcing rib arrangement scheme which are the lowest in structural risk probability and reasonable in structural protection cost are determined; The method comprises the steps of applying an optimal space partitioning scheme to a rear wall structure, setting up a field-shaped embedded column beam structure in the rear wall structure, partitioning a rear wall space, and determining the number and the size of space partitions according to the optimal space partitioning scheme; The optimal window size scheme is applied to space partitions of the rear wall structure, lighting windows are arranged in the upper space partitions of the field-shaped embedded column beam structure, the size of the lighting windows is determined according to the optimal window size scheme, and other space partitions are arranged by adopting closed walls or reinforcing ribs; Applying the optimal reinforcing rib arrangement scheme to each space partition of the rear wall structure, and arranging reinforcing ribs in the appointed space partition according to the reinforcing rib arrangement parameters; the design result is integrated to form a complete optimization scheme of the rear wall protection structure, wherein the complete optimization scheme comprises optimal configuration of space partition, window size and reinforcing rib arrangement; according to the complete optimization scheme, a structural construction diagram is compiled, the size of each space partition, the size and the position of each lighting window and the arrangement requirement of each reinforcing rib are defined, and a design file for engineering construction is generated; and applying the compiled structure construction diagram and design file to site structure construction, and implementing space partition, window size and reinforcing rib arrangement optimization design to complete engineering implementation of the rear wall protection structure.
5. The method for optimizing the protection structure of the frame house under the action of landslide induction by building a cut slope according to claim 1, wherein the continuously monitoring the state characteristic parameters of the rear wall partition in the house operation stage periodically inputs new monitoring data into the sliding window converter and the Bayesian network model, dynamically updates the parameters, and realizes the adaptive optimization of the toughness of the protection structure of the house, and comprises the following steps: Continuously collecting monitoring data of each space partition of the rear wall in the running stage of the house to form a new round of space partition state characteristic parameter data set; Carrying out data standardization and pretreatment on the collected back wall space partition state characteristic parameter data set; Periodically inputting the newly acquired and preprocessed space partition state characteristic parameter data set into a sliding window converter, and carrying out characteristic extraction and multi-scale characteristic update; Periodically inputting the space partition state characteristic parameters output by the sliding window converter into a Bayesian network model, and dynamically updating network node parameters and conditional probability distribution; based on the Bayesian network model after dynamic updating, the structural risk probability distribution of each space partition of the rear wall and the overall structural toughness state are inferred and evaluated in real time; According to reasoning and evaluation results, self-adaptive adjustment optimization is carried out on the space partition design of the house protection structure, the window size and the protection scheme of the reinforcing rib arrangement, the toughness and the adaptability of the house protection structure are improved, and the optimized protection scheme is implemented in structure maintenance and operation management.

Description

Frame house protection structure optimization method under landslide induction effect of slope cutting building Technical Field The invention relates to the technical field of civil engineering and structural safety protection, in particular to a method for optimizing a frame house protection structure under the action of landslide induction of a cut-slope building. Background Along with the promotion of mountain development construction, cut slope building is widely adopted in hilly and mountain areas, and the slope cutting action obviously increases the risk of inducing geological disasters such as landslide. The existing house protection structure is mostly based on traditional static design and experience reinforcing measures, and unified partition and conventional reinforcing rib arrangement are generally adopted. In engineering practice, the rear wall protection structure often fails to dynamically analyze multi-source space-time monitoring data such as space partition state parameters, soil deformation characteristics, rainfall, geological factors and the like, and the structural design is difficult to consider the difference of partition risks and the complex stress change under landslide induction. Most of traditional landslide risk assessment methods depend on single physical quantity or experience threshold values, cannot cover the synergistic effect of multiple parameters such as earth surface deformation, strain, stress, soil moisture content and the like, and are difficult to describe the probability distribution relation of expansion of local damage and overall failure between space partitions. The existing structural protection optimization means lack of closed loops for real-time monitoring and partition risk dynamic reasoning, and the protection design cannot timely adjust key parameters such as space partition division, window size, reinforcing rib arrangement and the like according to monitoring data in a house operation stage. Most protection structures are cured for a long time once being constructed, the follow-up partition self-adaptive adjustment and structure toughness optimization cannot be realized, and the house is easy to locally damage and spread or even totally fail under the induction of extreme rainfall or geological disasters. Because the multisource monitoring information cannot be effectively fused, the structural safety and disaster prevention capability are limited. The prior art has obvious defects in the aspects of comprehensive identification of partition risks, multi-parameter dynamic fusion and self-adaptive optimization of structures. Therefore, how to provide a method for optimizing the protection structure of the frame house under the action of landslide induction of building a cut-slope is a problem to be solved by the person skilled in the art. Disclosure of Invention The invention aims to provide a frame house protection structure optimization method under landslide induction effect of slope cutting and house building, which is characterized in that state characteristic parameters of each space partition of a rear wall are extracted by utilizing a sliding window converter through acquisition and processing of multisource space-time monitoring data, and intelligent reasoning and joint evaluation of risk probability of the space partition structure are realized by combining a Bayesian network model, so that structural parameters such as the space partition, window size, reinforcing rib arrangement and the like are optimally designed, and toughness self-adaptive optimization of the structural protection parameters is realized through periodic updating of the monitoring data in a house operation stage. The invention has the advantages of multisource information fusion, accurate regional risk assessment and structure self-adaptive optimization, and improves the safety, durability and disaster prevention capability of the house structure in a slope-cutting house-building scene. According to the embodiment of the invention, the method for optimizing the protection structure of the frame house under the effect of landslide induction by slope cutting and house building comprises the following steps: Collecting multisource space-time data sets of a house back slope and a back wall area, preprocessing the multisource space-time data sets, and generating a standardized multisource space-time data set; carrying out sliding window division on the standardized multi-source space-time data set according to a preset space step length and a time step length, and carrying out feature extraction on data in each window by utilizing a sliding window converter to obtain state feature parameters of each space partition; the state characteristic parameters of each space partition are used as node variables, a Bayesian network model is established, the state characteristic parameters are input to perform conditional probability reasoning, and the structural risk probability distribut