CN-122022496-A - Municipal building underground drainage pipe network operation risk assessment method and system

CN122022496ACN 122022496 ACN122022496 ACN 122022496ACN-122022496-A

Abstract

The invention discloses a municipal building underground drainage pipe network operation risk assessment method and a system, and relates to the technical field of data processing.A real-time hydraulic situation three-dimensional tensor is constructed by collecting liquid level and flow velocity data of monitoring points, and a hydraulic topology expansion gating network is input to obtain a future hydraulic situation prediction tensor; the method comprises the steps of carrying out space-time grid segmentation on a target analysis pipe section, solving internal physical field characteristics by combining a prediction tensor and physical constraints to obtain a pipe wall mechanical response characteristic matrix, constructing a pipe wall structure fatigue time-varying state transition chain based on the matrix, obtaining transient structure damage risk probability by Markov chain iterative computation, constructing a cross-mode cross attention operator by combining real-time physical field characteristics to obtain a weight square matrix, and finally combining the two to calculate a comprehensive risk score. The invention breaks through the traditional physical monitoring blind area and realizes continuous high-precision analysis of microscopic stress state and dynamic quantification of transient fatigue cracking risk.

Inventors

LIU CHENGLONG
ZHAO JIANAN
Yin Xiaochang
ZHANG CHONG
XU XIAOYU
YE LIYANG
ZHANG WENGANG

Assignees

萨奇水务(厦门)股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The running risk assessment method for the municipal building underground drainage pipe network is characterized by comprising the following steps of: Step S1, monitoring data at monitoring points of a drainage pipe network are collected, a real-time hydraulic situation three-dimensional tensor is constructed, the real-time hydraulic situation three-dimensional tensor is input into a hydraulic topology expansion gating network, and a future hydraulic situation prediction tensor is obtained; S2, segmenting a target analysis tube section among monitoring points, constructing a space-time grid sampling point set, combining a future hydraulic situation prediction tensor, solving physical field characteristics inside the target analysis tube section through physical constraint, and obtaining a tube wall mechanical response characteristic matrix; Step S3, constructing a fatigue time-varying state transition chain of the pipe wall structure based on the pipe wall mechanical response characteristic matrix, and obtaining the risk probability of damage of the transient structure through Markov chain type iterative computation; s4, combining the space-time grid sampling point set with a real-time hydraulic situation three-dimensional tensor, acquiring real-time physical field characteristics inside a target analysis pipe section through a spectral domain physical evolution feedforward network, constructing a cross-modal cross attention operator, and acquiring a cross attention weight matrix; And S5, calculating a comprehensive risk score based on the transient structure damage risk probability and the cross attention weight square matrix, and obtaining an operation risk assessment result of the municipal building underground drainage pipe network.
2. The method for evaluating the running risk of the underground drainage network of the municipal building according to claim 1, wherein in the step S1, the process of obtaining the future hydraulic situation prediction tensor specifically comprises the following steps: collecting monitoring data, wherein the monitoring comprises liquid level and flow rate, and the monitoring points are regarded as nodes in a drainage pipe network; constructing a real-time hydraulic situation three-dimensional tensor, wherein the real-time hydraulic situation three-dimensional tensor represents the liquid level and the flow velocity of a monitoring node in a time step; constructing a directional adjacent matrix, wherein the rows and the columns of the directional adjacent matrix correspond to nodes in a drainage pipe network, and elements in the directional adjacent matrix represent the flow direction relationship between the two corresponding nodes; Inputting the real-time hydraulic situation three-dimensional tensor into a pre-trained hydraulic topology expansion gating network for calculation to obtain a future hydraulic situation prediction tensor; The hydraulic topology expansion gating network comprises a space topology conducting layer, a parallel main information extraction branch, a gating signal extraction branch, a gating fusion layer and a residual error connecting layer.
3. The method for evaluating the running risk of the municipal building underground drainage pipe network according to claim 2, wherein in the step S2, the process of obtaining the pipe wall mechanical response feature matrix specifically comprises the following steps: taking an underground drainage pipe section between two nodes as a target analysis pipe section; Collecting the length of a target analysis pipe section, discretizing and cutting the length of the target analysis pipe section according to a preset length interval, and discretizing and cutting a preset scheduling period according to a preset time interval; combining all length dividing points with scheduling period dividing points in pairs to generate a space-time grid sampling point set; constructing a space-time coordinate matrix based on the space-time grid sampling point set; Inputting the space-time coordinate matrix into each order chebyshev polynomial equation solidified in advance to perform algebraic calculation to obtain a basis function matrix; Extracting hydraulic space-time characteristic slices corresponding to an upstream monitoring node and a downstream monitoring node of a target analysis pipe section from a future hydraulic situation prediction tensor, flattening and splicing data of the hydraulic space-time characteristic slices of the upstream monitoring node and the downstream monitoring node to generate boundary condition characteristic vectors; inputting boundary condition feature vectors into a spectral domain physical evolution feedforward network and outputting an expansion coefficient matrix, wherein the spectral domain physical evolution feedforward network performs data dimension reduction and nonlinear mapping through a plurality of hidden layers, the activation functions in all the hidden layers are hydraulic mutation self-adaptive activation functions, and the mathematical expression of the hydraulic mutation self-adaptive activation functions is as follows: ; Wherein, the For the adaptive scaling parameters to be solved for, Linear mapping characteristic values of the input characteristics of the previous layer received by the hidden layer neuron; And performing matrix multiplication operation on the basis function matrix and the expansion coefficient matrix to obtain a physical field characteristic matrix, wherein the physical field characteristic matrix comprises a predicted water depth value, a predicted flow velocity value, a predicted pipe wall circumferential stress value and a predicted pipe wall longitudinal strain value.
4. The method for evaluating the running risk of the municipal building underground drainage pipe network according to claim 3, wherein in the step S2, the process of obtaining the pipe wall mechanical response characteristic matrix further comprises: solving analysis partial derivatives of each column number value in the physical field feature matrix relative to the input space-time coordinate matrix by utilizing an automatic differential mechanism, and calculating a mass conservation residual error, a momentum conservation residual error and a pipe-soil coupling mechanical residual error by utilizing a san-Vietnam equation and a mole-coulomb model; Constructing a physical control loss function with volume weight, adopting an adaptive moment estimation optimization algorithm, taking the minimized physical control loss function as a target, continuously updating an ownership matrix, a bias vector and an adaptive scaling parameter vector in a spectral domain physical evolution feedforward network through back propagation, and judging convergence when the calculated value of the physical control loss function is continuously iterated to a preset number of times or the variation is smaller than a preset minimum threshold value, and stopping iteration; extracting a column vector representing circumferential stress and a column vector representing longitudinal strain from a physical field feature matrix in the last iteration, and splicing according to the column vector to generate a tube wall mechanical response feature matrix.
5. The method for evaluating the running risk of the municipal building underground drainage network according to claim 4, wherein in the step S3, the process of obtaining the risk probability of the transient structure damage specifically comprises the following steps: Dividing a preset scheduling period into a plurality of time windows according to the pre-divided discrete time steps, extracting dividing points with different lengths and row vectors corresponding to different time steps in each time window in a pipe wall mechanical response characteristic matrix, traversing all the absolute values of hoop stress in the same time window, selecting the row vector with the largest absolute value of the hoop stress in the time window as a stressed row vector under the time step, and splicing the stressed row vectors under each time step according to time sequence to obtain a stressed time sequence matrix; and extracting the circumferential stress and longitudinal strain of the tube wall of each row of the stress time sequence matrix, and executing mechanical equivalent conversion to construct a stress intensity column vector.
6. The method for evaluating the running risk of a municipal building underground drainage network according to claim 5, wherein in the step S3, the process of obtaining the risk probability of damage to the transient structure further comprises: Initializing a blank state transition probability square matrix for each time step in a preset scheduling period, wherein the state transition probability square matrix is used for recording the state transition probability of the pipeline state from a preset first state to a preset second state and from the preset second state to a preset third state in each time step; and calculating the state transition probability through a logic cliff activation function and an equivalent stress intensity factor.
7. The method for evaluating the running risk of a municipal building underground drainage network according to claim 6, wherein in the step S3, the process of obtaining the risk probability of damage to the transient structure further comprises: constructing an initial health state row vector, wherein the initial health state row vector is a preset row vector; Continuously performing matrix continuous multiplication operation on the initial health state row vector and a state transition probability matrix in a preset scheduling period, and finally outputting a termination state row vector at the end of the last time step; And extracting the last element of the termination state row vector as the transient structure damage risk probability.
8. The method for evaluating the running risk of the underground drainage network of the municipal building according to claim 7, wherein in the step S4, the process of obtaining the cross attention weight matrix specifically comprises the following steps: extracting liquid level data of a downstream monitoring node of a target analysis pipe section in the monitoring data, and constructing an actual liquid level column vector; Extracting corresponding real-time hydraulic situation slices of upstream and downstream monitoring nodes of a target analysis pipe section from the real-time hydraulic situation three-dimensional tensor, flattening and splicing data of the real-time hydraulic situation slices of the upstream monitoring nodes and the downstream monitoring nodes, and generating a real-time boundary condition feature vector; When analyzing the target analysis pipe section each time, copying a current spectral domain physical evolution feedforward network parameter as an independent copy, inputting the real-time boundary condition feature vector into the network parameter copy, and continuously executing reverse optimization iterative updating for preset times on the network parameter copy according to a physical control loss function and a self-adaptive moment estimation optimization algorithm; after the network parameter copy converges, acquiring a real-time physical field feature matrix output by the last iteration, and extracting a liquid level feature column corresponding to the actual liquid level column vector on a space topology node as a theoretical liquid level column vector formed by theoretical liquid level values in each time step in a preset time period in the past aiming at the space dimension expansion of the real-time physical field feature matrix caused by grid discretization; Subtracting the theoretical liquid level column vector from the actual liquid level column vector to generate a hydraulic situation residual column vector; Extracting a column representing the hoop stress from the real-time physical field feature matrix, traversing all space dividing points under each time step, selecting the maximum hoop stress value corresponding to each time step to splice into an extremum column vector, calculating the maximum hoop stress difference between adjacent time steps, and generating a pipe-soil coupling stress gradient column vector; performing matrix multiplication operation on the hydraulic situation residual column vector and a preset first attention mapping matrix, and outputting a query matrix; Performing matrix multiplication operation on the pipe-soil coupling stress gradient column vector and a preset second attention mapping matrix, and outputting a key matrix; And calculating the cross-correlation degree of the query matrix and the key matrix by adopting a dot product attention mechanism.
9. The method for evaluating the running risk of the municipal building underground drainage pipe network according to claim 8, wherein in the step S5, the process of obtaining the running risk evaluation result of the municipal building underground drainage pipe network specifically comprises the following steps: Calculating the arithmetic average value of all elements in the hydraulic situation residual column vector, and recording the arithmetic average value as a residual average value; if the residual error mean value is smaller than a preset liquid level abnormal threshold value, judging that the pipe network operates normally; If the residual error mean value is greater than or equal to a preset liquid level abnormal threshold value, judging that the operation of the pipe network is abnormal, and calculating a comprehensive risk score, wherein the comprehensive risk score is a product result of the transient structure damage risk probability and a preset rigid punishment amplification factor; The process for obtaining the hard punishment amplification factor specifically comprises the following steps: Extracting main diagonal elements in the cross attention weight square matrix, calculating an arithmetic average value of the main diagonal elements, and recording the arithmetic average value as a diagonal average value; And if the diagonal average value is greater than or equal to the preset stress mutation threshold value, judging that the target analysis pipe section is at a structural sedimentation and water accumulation risk.
10. The municipal building underground drainage pipe network operation risk assessment system is applied to the municipal building underground drainage pipe network operation risk assessment method according to any one of claims 1-9, and is characterized by comprising an acquisition module, a physical field analysis module, a fatigue state evolution module, a cross-mode attention module and an assessment module; The acquisition module is used for acquiring monitoring data at monitoring points of the drainage pipe network, constructing a real-time hydraulic situation three-dimensional tensor, inputting the real-time hydraulic situation three-dimensional tensor into a hydraulic topology expansion gating network, and acquiring a future hydraulic situation prediction tensor; The physical field analysis module is used for segmenting a target analysis pipe section among monitoring points, constructing a space-time grid sampling point set, combining a future hydraulic situation prediction tensor, solving physical field characteristics inside the target analysis pipe section through physical constraint, and obtaining a pipe wall mechanical response characteristic matrix; The fatigue state evolution module is used for constructing a fatigue time-varying state transition chain of the pipe wall structure based on the pipe wall mechanical response feature matrix, and obtaining the risk probability of damage of the transient structure through Markov chain type iterative computation; The cross-modal attention module is used for combining the space-time grid sampling point set with a real-time hydraulic situation three-dimensional tensor, acquiring real-time physical field characteristics inside a target analysis pipe section through a spectral domain physical evolution feed-forward network, constructing a cross-modal cross attention operator and acquiring a cross attention weight square matrix; The evaluation module is used for calculating a comprehensive risk score based on the transient structure damage risk probability and the cross attention weight square matrix and obtaining an operation risk evaluation result of the municipal building underground drainage pipe network.

Description

Municipal building underground drainage pipe network operation risk assessment method and system Technical Field The invention relates to the technical field of data processing, in particular to a municipal building underground drainage pipe network operation risk assessment method and system. Background The invention discloses an intelligent drainage pipe network management system, which is characterized in that the scale and complexity of an underground drainage system are exponentially increased along with the acceleration of the urban process, the operation state of the underground drainage system is directly related to the flood control and drainage safety of the city and the stability of a road foundation structure, under the condition of strong drainage of a rainy period or a pump station, the pipe network is required to bear severe impact of internal water flow and possibly face the compression caused by uneven settlement of external soil, and at present, the China patent with the application number 202511093820.4 discloses an intelligent drainage pipe network management system which improves the accuracy of data integrity assessment by comprehensively monitoring the water flow speed, the pipeline pressure, the fluid temperature, the pipe wall vibration frequency and the data transmission state and analyzing the time sequence, extracts the key parameters of the local pipeline based on the data integrity index, acquires the health evaluation value of the local pipeline by analyzing the trend change rate of target time interval data and calls the water flow state, the pressure and the vibration characteristics of an abnormal area pipeline to analyze the abnormal type and match the known fault mode data; The prior art mainly relies on simple fluctuation rate and trend change rate analysis on discrete macroscopic data at a sensor node, is limited by traditional analysis dimensionality, is difficult to accurately capture time delay characteristics and space spreading rules of water hammer shock waves transmitted in a long-distance main pipe caused by strong drainage of an upstream pump station and the like in a pipe network in a plurality of time steps, causes insufficient dynamic prediction capability on future hydraulic situations, can only acquire discrete monitoring data of the node position or carries out macroscopic structural stability assessment through external pipe wall vibration frequency, and still belongs to physical monitoring blind areas in a continuous stress state in the pipe, can not accurately analyze deep physical field characteristics such as internal transient hydraulic compression and external soil mass differential settlement co-extrusion, and the traditional pipe network health or abnormal assessment often lacks dynamic mechanical evolution mechanisms, can not quantitatively evaluate transient fatigue damage caused by water hammer shock to an old pipe wall, can detect blockage or abnormal fluctuation, but is difficult to analyze the physical situation of the old pipe wall when the liquid level is abnormally increased, and the operation risk of the pipe network is limited by the comprehensive operation risk assessment of the basic physical situation is limited by the accuracy of operation and the comprehensive assessment. Disclosure of Invention The method solves the technical problems that in the prior art, the time delay characteristic and the space spreading rule of the propagation of the water hammer shock wave in the long-distance main pipe caused by the forced drainage of an upstream pump station and the like in a pipe network are difficult to accurately capture across a plurality of time steps, the dynamic prediction capability of future hydraulic situation is insufficient, the physical detection of the continuous stress state in the pipe still belongs to a blind area in the prior art, the deep physical field characteristics such as the microscopic continuous pipe wall circumferential stress and the longitudinal strain in the pipe section under the co-extrusion of the internal transient hydraulic compression and the uneven settlement of an external soil body cannot be accurately analyzed in the prior art, the instantaneous fatigue damage caused by the water hammer impact to the old pipe wall cannot be quantitatively estimated, and the fundamental physical cause of the hydraulic situation cannot be easily analyzed in the prior art when the liquid level is abnormally increased due to the congestion of the pipe network. In order to solve the technical problems, the invention provides the following technical scheme that the running risk assessment method for the municipal building underground drainage pipe network comprises the following steps: Step S1, monitoring data at monitoring points of a drainage pipe network are collected, a real-time hydraulic situation three-dimensional tensor is constructed, the real-time hydraulic situation three-dimensional tensor is input into a h