CN-122021035-A - Bridge structure fault prediction method and system based on monitoring data

CN122021035ACN 122021035 ACN122021035 ACN 122021035ACN-122021035-A

Abstract

The invention discloses a bridge structure fault prediction method and a bridge structure fault prediction system based on monitoring data, and relates to the field of bridge structure health monitoring, wherein the bridge structure fault prediction method comprises the steps of obtaining an original monitoring data stream, preprocessing, extracting features based on the preprocessed data, and obtaining a standardized feature data set; the method comprises the steps of inputting a standardized feature data set into a causal feature mining framework based on physical information guidance to construct a causal directed acyclic graph among sensor nodes, quantifying contribution degrees of different factors to structural response based on the causal directed acyclic graph by using causal intervention effect, screening out a high-order damage sensitive feature vector, and inputting the high-order damage sensitive feature vector into a digital twin model integrated with a phase field method fracture model. According to the method, the probability decision tree is constructed to quantitatively evaluate the effect of different maintenance strategies on inhibiting crack development, and hierarchical early warning and specific maintenance decision reports are generated based on simulation prediction results and dynamic risk thresholds.

Inventors

Ba Huaiqiang
SONG HUI
ZHANG QIANG
CHEN JIAQI
ZHANG PEIJUN

Assignees

云南陆寻高速公路有限公司

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1. The bridge structure fault prediction method based on the monitoring data is characterized by comprising the steps of obtaining an original monitoring data stream, preprocessing, extracting features based on the preprocessed data, and obtaining a standardized feature data set; Inputting the standardized feature data set into a causal feature mining framework guided based on physical information, and constructing a causal directed acyclic graph among sensor nodes; based on a causal directed acyclic graph, quantifying the contribution degree of different factors to structural response by using causal intervention effect, and screening out a high-order damage sensitive feature vector; Inputting the high-order damage sensitive feature vector into a digital twin model integrated with a phase field method fracture model, and evaluating the intervention effect of different maintenance strategies on crack propagation paths by constructing a decision tree of damage state transition probability; and performing damage evolution physical simulation of a physical mechanism based on the multisource data and the impact load acquired in real time, outputting visual previewing data in the damage evolution process, and starting a multistage early warning mechanism to generate a maintenance decision suggestion report.
2. The bridge structure fault prediction method based on the monitoring data according to claim 1, wherein the method comprises the steps of obtaining an original monitoring data stream, preprocessing the original monitoring data stream, extracting features based on the preprocessed data, and obtaining a standardized feature data set, and comprises the following steps: collecting an original monitoring data stream containing stress, vibration, displacement, temperature and images, and performing preprocessing operations including missing value filling, outlier rejection and low-pass filtering on the original monitoring data stream to obtain a preprocessed data stream; And performing a feature extraction operation based on the preprocessed data stream, extracting a first third-order natural frequency from the vibration part, and extracting a mean value and a variance from the strain part to obtain a standardized feature data set.
3. The bridge structure fault prediction method based on monitoring data as set forth in claim 2, wherein the standardized feature data set is input into a causal feature mining framework based on physical information guidance to construct a causal directed acyclic graph between sensor nodes, comprising the steps of: inputting the standardized feature data set into a causal feature mining framework guided based on physical information, and embedding structural vibration differential equation constraints; In a causal feature mining framework based on physical information guidance, performing causal discovery analysis on a standardized feature dataset by adopting a PC algorithm; and constructing a causal directed acyclic graph among the sensor nodes of the temperature node, the load node, the strain node, the frequency node and the damage state node according to the results of causal discovery analysis.
4. The bridge structure fault prediction method based on the monitoring data as claimed in claim 3, wherein the method is characterized in that based on a causal directed acyclic graph, contribution degrees of response of different factors to the structure are quantified by using causal intervention effects, and a high-order damage sensitive feature vector is screened out, and comprises the following steps: Starting from the causal directed acyclic graph, performing inverse fact intervention on temperature factors, load factors and fatigue accumulation factors through a structure nesting model, and stripping the mixed effect among the factors to quantify the independent contribution degree values of the structural response factors; Performing weighted conversion on the causal directed acyclic graph by using independent contribution values, executing a random walk algorithm of a radix Yu Mengte Carlo simulation, calculating average first hit probability from each input factor node to the damage state node, and constructing a probability weighted causal network; Performing self-adaptive convolution operation on the average first hit probability of each path in the probability weighted causal network and the real-time operation risk level to generate a dynamic causal sensitivity threshold matched with the current risk state; And screening a high-sensitivity causal path in the probability weighted causal network according to the dynamic causal sensitivity threshold, and extracting the corresponding features in the standardized feature dataset mapped by the path to form a high-order damage sensitive feature vector.
5. The bridge structure fault prediction method based on monitoring data as claimed in claim 4, wherein the method is characterized in that a high-order damage sensitive feature vector is input into a digital twin model integrated with a phase field method fracture model, and the intervention effect of different maintenance strategies on crack propagation paths is evaluated by constructing a decision tree of damage state transition probability, and comprises the following steps: according to the average first hit probability, adjusting initial distribution of corresponding physical parameters in the phase field fracture model; defining crack grouting, pasting carbon fiber cloth and applying in-vitro prestress in a digital twin model integrated with a phase field fracture model, and coding each maintenance strategy into intervention operation of a specific high-sensitivity causal path in a probability weighted causal network; executing physical simulation of a digital twin model integrated with a phase field fracture model, counting branching, turning and healing mode changes of cracks in grids by comparing crack evolution results before and after intervention operation of applying a maintenance strategy, and constructing a decision tree reflecting damage state transition probability of state transition under different intervention; traversing a decision tree of the damage state transition probability, and evaluating the expected intervention effect of the crack grouting strategy on the main crack expansion length by analyzing the path transition probability distribution from the initial damage state node to each termination state node under different maintenance strategy intervention operation sequences.
6. The bridge structure fault prediction method based on the monitoring data according to claim 5, wherein the method is characterized by performing physical simulation of damage evolution of a physical mechanism based on multi-source data and impact load acquired in real time, outputting visual preview data of the damage evolution process, and comprises the following steps: Acquiring real-time vehicle weight information, real-time temperature and humidity information and real-time traffic flow information as multi-source data, and identifying and recording impact load information; Inputting real-time multi-source data and impact load into a digital twin model integrated with a phase field method fracture model, and executing damage evolution physical simulation based on a physical mechanism; and extracting data of the change of crack length, expansion angle and stress intensity factor along with time from the damage evolution physical simulation of the physical mechanism, and outputting the data as visual preview data of the damage evolution process.
7. The bridge structure fault prediction method based on monitoring data as claimed in claim 6, wherein the step of starting a multi-stage early warning mechanism to generate a maintenance decision suggestion report comprises the following steps: Comparing the visual previewing data of the damage evolution process with an early warning level threshold preset based on historical safety data, structural design specifications and a real-time operation environment, and judging the deviation degree of crack expansion speed, path and stress intensity factors and the early warning level threshold to obtain a comparison result; Triggering a multi-stage early warning mechanism comprising blue attention, yellow early warning, orange warning and red emergency according to the comparison result; And generating a maintenance decision proposal report according to the visual previewing data and the multi-level early warning mechanism in the damage evolution process.
8. The bridge structure fault prediction system based on the monitoring data is based on the bridge structure fault prediction method based on the monitoring data according to any one of claims 1-7, and is characterized by comprising a feature extraction module, a feature extraction module and a feature extraction module, wherein the feature extraction module acquires an original monitoring data stream and performs preprocessing, and performs feature extraction based on the preprocessed data to obtain a standardized feature data set; The construction module inputs the standardized feature data set into a causal feature mining framework guided by physical information to construct a causal directed acyclic graph among the sensor nodes; The screening module is used for quantifying the contribution degree of the response of different factors to the structure by utilizing the causal intervention effect based on the causal directed acyclic graph and screening out a high-order damage sensitive feature vector; The evaluation module inputs the high-order damage sensitive feature vector into a digital twin model integrated with a phase field method fracture model, and evaluates the intervention effect of different maintenance strategies on the crack propagation path by constructing a decision tree of damage state transition probability; and the advice report module is used for carrying out physical simulation on the damage evolution of the physical mechanism based on the multisource data and the impact load acquired in real time, outputting visual previewing data of the damage evolution process, and starting a multistage early warning mechanism to generate a maintenance decision advice report.
9. The computer equipment comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is characterized in that the processor realizes the steps of the bridge structure fault prediction method based on the monitoring data according to any one of claims 1-7 when executing the computer program.
10. A computer readable storage medium having a computer program stored thereon, wherein the computer program when executed by a processor implements the steps of the bridge structure failure prediction method based on monitoring data as set forth in any one of claims 1 to 7.

Description

Bridge structure fault prediction method and system based on monitoring data Technical Field The invention relates to the field of bridge structure health monitoring, in particular to a bridge structure fault prediction method and system based on monitoring data. Background In the field of bridge structure health monitoring, intelligent fault prediction based on monitoring data is an important research direction for guaranteeing safe operation of infrastructure, and the prior art scheme, for example, a bridge state assessment method integrating a deep neural network and digital twinning, represents a typical practice in the field, collects structural response data by deploying a sensor network, learns complex mapping relations between damage characteristics and states by using the deep neural network, realizes visualization and anomaly detection of structural states by combining a digital twinning technology, and improves automation and intuitiveness of state assessment by combining a data driving model and a physical model. The prior art has limitations in the aspects of the interpretability of a prediction result and the accurate decision support based on prediction, the black box characteristic of a deep neural network model makes it difficult to clarify the specific causal contribution of various environments and operation factors to structural damage, so that the prediction result lacks clear causal logic support, the existing digital twin multi-emphasis is on state reproduction and visualization, the simulation process is often not deeply combined with a causal mechanism mined in data, and therefore, it is difficult to evaluate reliably and quantitatively how different maintenance measures influence future evolution of damage, so that the output of the existing method is generally remained on a risk early warning level, and is difficult to directly convert into a clear decision basis for guiding specific maintenance actions, so that the prediction and maintenance links are disjointed. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the bridge structure fault prediction method based on the monitoring data solves the problems that in the prior art, a prediction result caused by lack of physical mechanism and causal analysis cannot be explained and is difficult to be directly converted into a targeted maintenance decision basis. In order to solve the technical problems, the invention provides the following technical scheme: In a first aspect, the invention provides a bridge structure fault prediction method based on monitoring data, which comprises the steps of obtaining an original monitoring data stream, preprocessing, extracting features based on the preprocessed data, and obtaining a standardized feature data set; Inputting the standardized feature data set into a causal feature mining framework guided based on physical information, and constructing a causal directed acyclic graph among sensor nodes; based on a causal directed acyclic graph, quantifying the contribution degree of different factors to structural response by using causal intervention effect, and screening out a high-order damage sensitive feature vector; Inputting the high-order damage sensitive feature vector into a digital twin model integrated with a phase field method fracture model, and evaluating the intervention effect of different maintenance strategies on crack propagation paths by constructing a decision tree of damage state transition probability; and performing damage evolution physical simulation of a physical mechanism based on the multisource data and the impact load acquired in real time, outputting visual previewing data in the damage evolution process, and starting a multistage early warning mechanism to generate a maintenance decision suggestion report. The bridge structure fault prediction method based on the monitoring data is used for obtaining an original monitoring data stream, preprocessing the original monitoring data stream, extracting features based on the preprocessed data, and obtaining a standardized feature data set, and comprises the following steps: collecting an original monitoring data stream containing stress, vibration, displacement, temperature and images, and performing preprocessing operations including missing value filling, outlier rejection and low-pass filtering on the original monitoring data stream to obtain a preprocessed data stream; And performing a feature extraction operation based on the preprocessed data stream, extracting a first third-order natural frequency from the vibration part, and extracting a mean value and a variance from the strain part to obtain a standardized feature data set. The bridge structure fault prediction method based on the monitoring data is characterized in that a standardized characteristic data set is input into a causal characteristic mining framework based on physi