CN-122022072-A - Data processing method for risk prediction of urban drainage pipe network

CN122022072ACN 122022072 ACN122022072 ACN 122022072ACN-122022072-A

Abstract

The application is applicable to the technical field of urban infrastructure risk management, and provides a data processing method for urban drainage pipe network risk prediction, which comprises the following steps of S5, multi-disaster coupling risk assessment, namely calculating multi-disaster coupling risk indexes by combining a pre-constructed waterlogging inundation model based on the risk probability of each node obtained in the S4, and generating an urban drainage pipe network risk dynamic distribution map; and S6, intelligent scheduling of emergency knowledge and resources, namely generating an optimal scheduling path from an emergency resource point to a risk point by adopting a path optimization algorithm based on an urban network emergency knowledge graph, and outputting an emergency decision scheme. According to the data processing method for risk prediction of the urban drainage pipe network, in the aspect of data processing, in the step S1, semantic alignment and standardization processing are carried out on multi-source heterogeneous data through the pre-built unified data dictionary, so that the semantic consistency and high quality of the data are ensured.

Inventors

LI SHAOHUA
WANG YAO
ZHANG FENG
ZHANG SHUFENG
ZHENG GUOXING
GUAN CAN
YIN MINGLEI

Assignees

中建海峡建设发展有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (9)

1. The data processing method for risk prediction of the urban drainage pipe network is characterized by comprising the following steps of: Step S1, multi-source heterogeneous data fusion and standardization processing are carried out, namely basic geographic information data, internet of things perception data, meteorological data and historical disaster data of a drainage pipe network are collected, semantic alignment and standardization processing are carried out on the collected multi-source data based on a pre-constructed unified data dictionary, and a standardized fusion data set is generated; S2, space-time grid-based holographic digital twin modeling, namely carrying out three-dimensional space grid segmentation and coding on the urban drainage pipe network, and constructing a pipe network holographic digital twin model supporting dynamic updating; s3, multi-mode risk feature extraction, namely extracting pipe network running state features, environment features and historical disaster features from the fusion data set, constructing a graph representation of a pipe network topological structure, and generating risk feature vectors of each pipe network node through a graph neural network; S4, risk probability prediction based on a dynamic Bayesian network, namely constructing a dynamic Bayesian network model, wherein the model is defined by initial state distribution, state transition probability and observation likelihood function; S5, multi-disaster coupling risk assessment, namely calculating multi-disaster coupling risk indexes based on the risk probability of each node obtained in the step S4 and combining a pre-built waterlogging inundation model to generate a city drainage pipe network risk dynamic distribution map; And S6, intelligent scheduling of emergency knowledge and resources, namely generating an optimal scheduling path from an emergency resource point to a risk point by adopting a path optimization algorithm based on an urban network emergency knowledge graph, and outputting an emergency decision scheme.
2. The method according to claim 1, wherein the step S1 further comprises: r2RML mapping rules are introduced, unstructured historical disaster data are converted into a resource description framework triplet format, and the unstructured historical disaster data are fused with structured data to construct a base layer of the urban drainage pipe network knowledge graph.
3. The method according to claim 1, wherein the step S2 further comprises: And when the perceived data of the Internet of things is updated, positioning the corresponding component according to the grid coding index, and triggering the local update of the digital twin model.
4. The method according to claim 1, wherein the step S3 further comprises: the method comprises the steps of modeling a drainage pipe network topological structure into a graph structure, enabling nodes to represent pipe network nodes and edges to represent pipe section connection relations, aggregating and updating node characteristics by adopting a graph convolution network, and taking the node characteristics output by the network as risk characteristic vectors.
5. The method according to claim 1, wherein the step S4 further comprises: The initial state distribution is the probability that the node is in each risk state at the initial moment; The state transition probability is the probability of the node transitioning from the risk state at the previous moment to the risk state at the current moment, and is positively related to the service life of the pipeline and the soil corrosiveness; The observation likelihood function is the probability density of the current risk feature vector observed by the node in a certain risk state; And inputting the risk feature vector at the current moment, and recursively calculating the posterior probability at the current moment through Bayesian filtering based on the current observation likelihood value, the posterior probability at the last moment and the state transition probability.
6. The method of claim 5, wherein the method of determining the state transition probability further comprises: and for the data sparse area, adopting a physical model based on a pipeline aging rule to preset.
7. The method of claim 5, wherein the method of determining the observation likelihood function further comprises: and collecting historical monitoring data of the nodes in each risk state, and fitting probability distribution of the observation data in each risk state by adopting a kernel density estimation method to obtain an observation likelihood function.
8. The method according to claim 1, wherein calculating the multi-disaster coupling risk index in step S5 further comprises: calculating a coupling risk index of the node i at the current moment, wherein the coupling risk index is obtained by three weighted summation: The first term is the first weight coefficient multiplied by the waterlogging risk probability of the node i; The second term is the probability of gas leakage risk of the area where the node i is located multiplied by the second weight coefficient; The third term is the product of the waterlogging risk probability and the gas leakage risk probability multiplied by a third weight coefficient; the sum of the first weight coefficient, the second weight coefficient and the third weight coefficient is 1, and the third weight coefficient is used for representing a risk amplification effect when waterlogging and leakage occur simultaneously.
9. The method according to claim 1, wherein the step S6 further comprises: The method comprises the steps of taking the shortest emergency response time as an objective function and taking a road traffic state as a constraint condition, dynamically updating the road traffic state according to the real-time ponding depth, and generating an emergency resource scheduling path by adopting a Dijkstra algorithm or a differential evolution algorithm.

Description

Data processing method for risk prediction of urban drainage pipe network Technical Field The invention belongs to the technical field of urban infrastructure risk management, and particularly relates to a data processing method for urban drainage pipe network risk prediction. Background The urban drainage pipe network is used as the core of urban life line engineering, and the safety operation of the urban drainage pipe network is related to public safety and urban toughness. At present, drainage systems in China face multiple challenges such as extreme weather frequency, facility aging, data isomerization, emergency response lag and the like, and the prior art has obvious defects in the following aspects: Firstly, multi-source data fusion is difficult. The pipe network relates to various data such as geographic information, internet of things perception, weather, historical disaster conditions and the like, the sources are dispersed, the format difference is obvious, the semantics are not uniform, the semantic gap cannot be solved by traditional manual input or simple splicing, the data fragmentation is caused, and high-quality data set support risk analysis is difficult to form. And secondly, digital twin modeling hysteresis. Traditional pipe network models are mostly static structures, rely on periodic manual mapping, and cannot be dynamically associated with real-time sensing data of the Internet of things. The modeling efficiency of the ultra-large-scale urban pipe network is low, the updating period is long, and real-time monitoring and dynamic deduction of the risk are difficult to support. Thirdly, extracting the risk characteristics. The traditional method is limited to the isolated analysis of single-point monitoring data, and does not consider the spatial dependency relationship among nodes in a pipe network topological structure. For example, failure to effectively trace liquid level anomalies is caused by upstream water supply or downstream drainage blockage, resulting in a lack of global risk identification. In summary, a comprehensive technical scheme capable of integrating multi-source data, realizing dynamic update of a model, fusing topological features, quantifying multi-disaster coupling effect and supporting intelligent emergency dispatch is needed. Disclosure of Invention The embodiment of the invention aims to provide a data processing method for risk prediction of an urban drainage pipe network, and aims to solve the problems. The invention discloses a data processing method for risk prediction of an urban drainage pipe network, which comprises the following steps: Step S1, multi-source heterogeneous data fusion and standardization processing are carried out, namely basic geographic information data, internet of things perception data, meteorological data and historical disaster data of a drainage pipe network are collected, semantic alignment and standardization processing are carried out on the collected multi-source data based on a pre-constructed unified data dictionary, and a standardized fusion data set is generated; S2, space-time grid-based holographic digital twin modeling, namely carrying out three-dimensional space grid segmentation and coding on the urban drainage pipe network, and constructing a pipe network holographic digital twin model supporting dynamic updating; s3, multi-mode risk feature extraction, namely extracting pipe network running state features, environment features and historical disaster features from the fusion data set, constructing a graph representation of a pipe network topological structure, and generating risk feature vectors of each pipe network node through a graph neural network; S4, risk probability prediction based on a dynamic Bayesian network, namely constructing a dynamic Bayesian network model, wherein the model is defined by initial state distribution, state transition probability and observation likelihood function; S5, multi-disaster coupling risk assessment, namely calculating multi-disaster coupling risk indexes based on the risk probability of each node obtained in the step S4 and combining a pre-built waterlogging inundation model to generate a city drainage pipe network risk dynamic distribution map; And S6, intelligent scheduling of emergency knowledge and resources, namely generating an optimal scheduling path from an emergency resource point to a risk point by adopting a path optimization algorithm based on an urban network emergency knowledge graph, and outputting an emergency decision scheme. According to a further technical scheme, the step S1 further includes: r2RML mapping rules are introduced, unstructured historical disaster data are converted into a resource description framework triplet format, and the unstructured historical disaster data are fused with structured data to construct a base layer of the urban drainage pipe network knowledge graph. According to a further technical scheme, the step S2 further includes: A