CN108615098B - Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method - Google Patents

Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method Download PDF

Info

Publication number
CN108615098B
CN108615098B CN201810448471.7A CN201810448471A CN108615098B CN 108615098 B CN108615098 B CN 108615098B CN 201810448471 A CN201810448471 A CN 201810448471A CN 108615098 B CN108615098 B CN 108615098B
Authority
CN
China
Prior art keywords
pipe
explosion
risk
pipeline
tube
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810448471.7A
Other languages
Chinese (zh)
Other versions
CN108615098A (en
Inventor
信昆仑
陈能
颜合想
陶涛
李树平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tongji University
Original Assignee
Tongji University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tongji University filed Critical Tongji University
Priority to CN201810448471.7A priority Critical patent/CN108615098B/en
Publication of CN108615098A publication Critical patent/CN108615098A/en
Application granted granted Critical
Publication of CN108615098B publication Critical patent/CN108615098B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities

Abstract

The invention relates to a water supply network pipeline pipe burst risk prediction method based on Bayesian survival analysis, which comprises the following steps: (1) establishing a detonator database according to the collected detonator historical data, and extracting key information as covariates; (2) carrying out spatial clustering analysis on the tube explosion point positions, and quantifying spatial distribution information of the tube explosion point positions to be used as a new covariate to be supplemented to a tube explosion database; (3) constructing a pipe explosion risk prediction model by adopting a Bayesian survival analysis method based on a pipe explosion database; (4) and predicting the pipe explosion risk of the pipeline by adopting a pipe explosion risk prediction model. Compared with the prior art, the prediction result is more accurate and reasonable.

Description

Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method
Technical Field
The invention relates to a water supply network pipeline pipe explosion risk prediction method, in particular to a Bayesian survival analysis-based water supply network pipeline pipe explosion risk prediction method.
Background
The water supply network, one of the important public infrastructures in cities, is the main artery in the whole city and plays a role in delivering water to users. In recent ten years, the urbanization development speed of China is continuously accelerated, the water demand of cities is increased dramatically, and the scale of urban water supply networks is increased day by day. However, the water supply pipe network in China still has a series of problems that the laying time of partial pipelines is too long, the aging phenomenon is serious, poor-quality pipes (such as grey cast iron pipes) occupy a large proportion in the pipe network, and the like. Meanwhile, in the aspect of pipe network operation safety guarantee and management technology, aiming at prevention and control of pipe network pipe explosion accidents, scientific research and technical application of the system are lacked in the aspects of informatization management, pipeline material quality, pipe network safety detection, accident reaction and treatment technology and the like, the pipe network operation safety and management technology level is low, accordingly, pipe explosion accidents of urban water supply pipe networks in China occur frequently, precious water resources are wasted, drinking water safety of people is seriously influenced, and a series of chain reactions such as economic loss and social negative effects caused by the pipe explosion accidents cannot be ignored. Therefore, the operation state of the water supply pipe network is evaluated by adopting a scientific method, the pipeline with higher pipe explosion risk in the pipe network is found out, and the maintenance work is strengthened in a targeted manner, so that the occurrence of pipe explosion accidents is reduced.
At home and abroad, the research on the aspect is many, and the pipe explosion risk prediction model can be simply divided into a physical model, a statistical model and a data mining model, and the following are some representative researches:
1) physical model
As in the literature:
[1]:Moglia M,Davis P,Burn S.Strong exploration of a cast iron pipe failure model.Reliability Engineering&System Safety,2008,93(6):885-896.
the method adopts the following main technical measures: based on a fracture mechanics theory, firstly, the residual yield strength of the pipeline is supposed to meet Weibull distribution, and then a pipe explosion risk prediction model of the grey cast iron pipeline is obtained through Monte Carlo simulation and regression analysis by using historical data.
The advantages and disadvantages are as follows: the method has the advantages that the established model is based on some specific pipe bursting physical mechanisms, and the prediction result is accurate. But has the disadvantages that: (1) the physical mechanism of pipe explosion is complex and the pipe explosion inducing factors are various, and the physical prediction model cannot cover all the factors, so that the application difficulty is high; (2) the data required to build the model is difficult or expensive to acquire.
2) Statistical model
As in the literature:
[2]:Park S,Jun H,Agbenowosi N,et al.The proportional hazards modeling of water main failure data incorporating the time-dependent effects of covariates.Water resources management,2011,25(1):1-19.
[3]:Kimutai E,Betrie G,Brander R,et al.Comparison of statistical models for predicting pipe failures:Illustrative example with the City of Calgary water main failure.Journal of Pipeline Systems Engineering and Practice,2015,6(4):04015005.
the method adopts the following main technical measures: based on a large amount of historical data of pipe bursting, a certain relation is established between factors influencing pipe bursting and objective functions such as pipe bursting risk or pipe bursting time through statistical analysis, and therefore a model is established.
The advantages and disadvantages are as follows: the method has the advantages of simple model form, strong operability and small application difficulty. However, such methods have certain requirements on the quality and quantity of the pipe explosion historical data, and the obtained results are not ideal under the condition of less or incomplete data.
3) Data mining model
As in the literature:
[4]:Ahmad A,Mcbean E,Bahram G,et al.Forecasting watermain failure using artificial neural network modelling.Canadian Water Resources Journal,2013,38(1):24-33.
[5]:Harvey R,Mcbean EA,Gharabaghi B.Predicting the Timing of Water Main Failure Using Artificial Neural Networks.Journal of Water Resources Planning&Management,2013,140(4):425-434.
the method adopts the following main technical measures: based on a large amount of historical data of pipe explosion, analyzing the historical data by applying various data mining algorithms, and finally predicting the pipe explosion risk based on a certain principle.
The advantages and disadvantages are as follows: such a method is advantageous in that it is not limited to a specific functional form. But has the disadvantages that: (1) a large amount of pipe burst historical data is needed for training; (2) the predicted results from a purely data-driven model may not be consistent with the actual observed results.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a method for predicting the pipe bursting risk of a water supply network pipeline based on Bayesian survival analysis.
The purpose of the invention can be realized by the following technical scheme:
a water supply network pipeline pipe explosion risk prediction method based on Bayesian survival analysis comprises the following steps:
(1) establishing a detonator database according to the collected detonator historical data, and extracting key information as covariates;
(2) carrying out spatial clustering analysis on the tube explosion point positions, and quantifying spatial distribution information of the tube explosion point positions to be used as a new covariate to be supplemented to a tube explosion database;
(3) constructing a pipe explosion risk prediction model by adopting a Bayesian survival analysis method based on a pipe explosion database;
(4) and predicting the pipe explosion risk of the pipeline by adopting a pipe explosion risk prediction model.
Step (2) taking the quantitative result of whether other pipe explosion exist in the set space distance of the pipe explosion point position as a new covariate, specifically: when other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 1, and when no other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 0.
The step (3) is specifically as follows:
(31) and (3) standardizing tube explosion historical data in a tube explosion database: dividing each pipe bursting pipeline historical record into three parts,
Figure BDA0001657813980000031
wherein D isiRepresents the history of the ith burst pipe, tiIndicating the survival time of the ith burst pipe, ZiRepresenting the covariate vector, delta, corresponding to the ith burst pipeiData type indicating variable, delta, representing the ith burst pipei1 denotes the ith burst pipe history as complete data, δiThe ith burst pipe history record is regarded as deleted data as 0, and i is 1,2, … …, N and N are the total number of burst pipes;
(32) establishing a benchmark risk function: the reference risk function takes the survival time of the pipe explosion pipeline as an independent variable and takes the pipe explosion risk of the pipeline as a dependent variable, wherein the pipe explosion risk of the pipeline is the pipe explosion times of the average unit pipe length per year;
(33) selecting main factors influencing the pipe explosion risk as covariates, and establishing a Bayesian survival analysis model for parameter estimation to determine covariate parameter estimation values;
(34) and determining a pipe explosion risk prediction model according to the reference risk function and the covariate parameter estimation value.
The step (33) is specifically:
(331) dividing the survival time of all pipe explosion pipelines into intervals of 0 & lt s1＜s2···＜sJAnd < ∞, recording the survival time of the ith explosion pipe as tiAnd for all tiAre all less than sJJ time intervals (0, s) are obtained1],(s1,s2],···,(sJ-1,sJ]；
(332) Assuming a fixed reference hazard function h for each time interval0(ti)＝λj，ti∈(sj-1,sj-1) Obtaining a Bayesian survival analysis model function:
Figure BDA0001657813980000041
wherein, h (t)i,Zi) Risk of tube explosion for the ith tube, tiIndicates the survival time of the ith pipeline, ZiRepresents a covariate vector, beta, corresponding to the ith pipelineTFor regression coefficient vectors corresponding to covariates, λjJ is 1,2, … …, J is the total number of divided time intervals, i is 1,2, … …, N represents the total number of pipe bursting pipelines.
The pipe explosion risk model in the step (34) is as follows:
h(t,Z)＝(at2+bt+c)exp(βTZ)；
wherein h (t, Z) is the risk of pipe explosion, t represents the life time of the pipe, at2+ bt + c is the reference risk function, a, b and c are the fitting coefficients of the reference risk function, Z represents the impact on the pipelineCovariate vector, beta, of the main factor of the risk of pipe explosionTThe regression coefficient vector corresponding to the covariate.
And performing spatial clustering analysis on the pipe explosion point positions by adopting a DBSCAN clustering algorithm.
Compared with the prior art, the invention has the following advantages:
(1) when the covariate regression coefficient is calculated, the research result of the predecessor is introduced as prior distribution, so that the obtained covariate regression coefficient is more consistent with the general explosion rule, and the prediction result of the explosion risk prediction model is more accurate and reasonable;
(2) the invention combines the advantages of the Bayesian method and the traditional survival analysis method, and is more suitable for the conditions of less and incomplete pipe burst historical records.
Drawings
FIG. 1 is a general flow chart of the pipeline pipe bursting risk prediction based on Bayesian survival analysis in the invention;
FIG. 2 is a block diagram of the DBSCAN clustering algorithm of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. Note that the following description of the embodiments is merely a substantial example, and the present invention is not intended to be limited to the application or the use thereof, and is not limited to the following embodiments.
Examples
As shown in fig. 1, a method for predicting pipe explosion risk of a water supply network pipeline based on bayesian survival analysis comprises the following steps:
(1) establishing a detonator database according to the collected detonator historical data, and extracting key information as covariates;
(2) carrying out spatial clustering analysis on the tube explosion point positions, and quantifying spatial distribution information of the tube explosion point positions to be used as a new covariate to be supplemented to a tube explosion database;
(3) constructing a pipe explosion risk prediction model by adopting a Bayesian survival analysis method based on a pipe explosion database;
(4) and predicting the pipe explosion risk of the pipeline by adopting a pipe explosion risk prediction model.
The method comprises the following steps that (1) relevant pipe explosion information is recorded in detail aiming at a pipe explosion event, the reason and the rule of the pipe explosion are analyzed, and even a data basis for formulating a prevention and emergency strategy is established, and data collection aiming at the pipe explosion analysis comprises the following data:
1) and (3) maintaining and recording pipe burst: the system mainly refers to maintenance record information of a water supply enterprise on a pipeline with a pipe explosion accident, wherein the maintenance record information comprises a series of related data such as pipeline ID, pipe length, pipe material, pipe explosion time, pipeline laying time, pipe explosion geographic position and maintenance measures, and the more detailed the collected data is, the clearer the reason for pipe explosion and related pipe explosion factors are;
2) pipe network GIS data: the system mainly refers to static data of an urban water supply network, and comprises self structure data such as ID (identity) of each pipeline, pipe length and the like, and data such as positions of each water plant or booster pump station;
3) SCADA monitoring data: the method mainly comprises the steps of monitoring data of each pressure monitoring point in a pipe network in real time, wherein the data of the real-time pressure when a pipeline bursts can be obtained;
4) other data: the method mainly refers to the information of the surrounding environment of the pipeline, such as data of soil corrosion indexes around the laid pipeline, ground load or economic development degree of the pipeline and the like, the data have a certain relation with the pipeline explosion, and a more accurate explosion risk prediction model can be established by collecting the data.
After collecting the tube explosion historical data, removing error information in the data and recording repeated data, and establishing a corresponding tube explosion database after certain pretreatment.
To a certain extent, the pipe explosion risk of the pipeline is higher at the place where the pipe explosion point positions are dense, so the spatial clustering analysis is carried out on the pipe explosion point positions by adopting a DBSCAN clustering algorithm in the step (2), and the quantitative result of whether other pipe explosion exist in the set spatial distance of the pipe explosion point positions is taken as a new covariate, specifically: when other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 1, and when no other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 0. Specifically, the algorithm, as shown in fig. 2, includes the following steps:
1) randomly selecting a point k from a data set as a clustering starting point, and searching an object which can be reached from k density in the neighborhood of the data concentration point k;
2) if k is a core object, dividing all points in the neighborhood of k into a cluster, taking the points as candidate points of the next round, and continuously searching the points with the density reaching from the candidate points to expand the cluster where the candidate points are located until the points with the density reaching cannot be found, wherein the found points are a complete cluster;
3) if k is not a core object, i.e., no points are reachable from k density, then k is temporarily labeled as a noise point. Then, repeating the above process for the next point in the whole data set, and completing the expansion of one cluster after all the objects in the data set are inspected;
4) if unprocessed points still exist in the data set, performing expansion of another cluster; otherwise, points in the data set that do not belong to any cluster are determined as noise points.
And after clustering is finished, quantifying the clustering result to be used as an input variable of the pipe explosion risk model.
The step (3) is specifically as follows:
(31) and (3) standardizing tube explosion historical data in a tube explosion database: dividing each pipe bursting pipeline historical record into three parts,
Figure BDA0001657813980000061
wherein D isiRepresents the history of the ith burst pipe, tiIndicating the survival time of the ith burst pipe, ZiRepresenting the covariate vector, delta, corresponding to the ith burst pipeiData type indicating variable, delta, representing the ith burst pipei1 denotes the ith burst pipe history as complete data, δiThe ith burst pipe history record is regarded as deleted data as 0, and i is 1,2, … …, N and N are the total number of burst pipes;
(32) establishing a benchmark risk function: the benchmark risk function uses the tube burst pipeline survival time as an independent variable, uses the tube burst risk as a dependent variable, the tube burst risk is specifically the tube burst frequency of the average unit tube length in year, and the tube burst risk unit is: sub/(a.km);
(33) selecting main factors influencing the pipe explosion risk as covariates, and establishing a Bayesian survival analysis model for parameter estimation to determine covariate parameter estimation values;
(34) and determining a pipe explosion risk prediction model according to the reference risk function and the covariate parameter estimation value.
The step (33) is specifically:
(331) dividing the survival time of all pipe explosion pipelines into intervals of 0 & lt s1＜s2···＜sJAnd < ∞, recording the survival time of the ith explosion pipe as tiAnd for all tiAre all less than sJJ time intervals (0, s) are obtained1],(s1,s2],···,(sJ-1,sJ]；
(332) Assuming a fixed reference hazard function h for each time interval0(ti)＝λj，ti∈(sj-1,sj-1) Obtaining a Bayesian survival analysis model function:
Figure BDA0001657813980000071
wherein, h (t)i,Zi) Risk of tube explosion for the ith tube, tiIndicates the survival time of the ith pipeline, ZiRepresents a covariate vector, beta, corresponding to the ith pipelineTFor regression coefficient vectors corresponding to covariates, λjJ is 1,2, … …, J is the total number of divided time intervals, i is 1,2, … …, N represents the total number of pipe bursting pipelines.
The pipe explosion risk model in the step (34) is as follows:
h(t,Z)＝(at2+bt+c)exp(βTZ)；
wherein h (t, Z) isRisk of pipe bursting, t represents pipe life time, at2+ bt + c is a reference risk function, a, b and c are fitting coefficients of the reference risk function, Z represents a covariate vector composed of main factors influencing the pipe bursting risk, and betaTThe regression coefficient vector corresponding to the covariate.
After the pipe explosion risk prediction model is built, pipe explosion risks of all pipelines in the water supply network can be calculated by combining GIS pipe network information of the actual water supply network, and a pipeline pipe explosion risk distribution map is drawn according to the calculation result. The risk grade of pipe explosion of the pipeline can be divided into four grades of high risk, danger, low risk and safety according to the calculation result of the risk of pipe explosion of the pipeline. Meanwhile, in order to verify the rationality of the prediction result, the prediction accuracy of the model is verified by comparing the prediction result with the actual pipe explosion point position through spatial analysis.
For the present embodiment, specifically:
(1) collecting and preprocessing pipe bursting historical data
For the example pipe network, 449 original pipe burst history records and the GIS pipe network of the water supply pipe network are collected in total, and a corresponding pipe burst database is established according to the pipe burst history records and mainly comprises 23 fields of pipe diameter, pipe material, pipe length, laying time, receiving and reporting time and the like of pipe burst pipelines. And removing the incomplete field records in the original pipe burst record or the records of pipe bursts caused by artificial damage, and obtaining 275 pipe burst historical records which can be used for modeling.
(2) Spatial clustering analysis of pipe burst point locations
According to the spatial distribution characteristics of the pipe explosion point positions, the distance threshold Eps of the DBSCAN clustering algorithm is 100m, the density threshold MinPts of the DBSCAN clustering algorithm is 2, and the pipe explosion point positions are clustered and analyzed to obtain clustering results. And when the clustering result of the pipe explosion point positions is quantized, assigning values according to the existence of other pipe explosion point positions in the pipe explosion point positions of 100m, and taking the assignment results as covariates of the pipe explosion risk prediction model. And (3) the Cluster assignment of each detonation history record is expressed by a variable Cluster, and then an assignment formula is expressed as the following formula:
Figure BDA0001657813980000081
(3) construction of pipe explosion risk prediction model
And (4) arranging the pretreated pipe burst history records into a standard format, and establishing a corresponding reference risk function. According to the characteristics of the pipe explosion history record, two main factors of the pipe diameter D and the Cluster assignment value are selected as covariates influencing the pipe explosion risk of the pipeline, and meanwhile, the uniform distribution in the (-1, 0) interval is selected as the prior distribution of the pipe diameter D. And for Cluster assignment, assuming that the covariate regression coefficient of the Cluster assignment conforms to normal distribution. The prior distribution of the regression coefficients for each covariate is shown by the following equation:
f(β1)～U(-1,0)，
Figure BDA0001657813980000082
wherein f (. beta.) is1) And f (. beta.)2) A priori distribution of regression coefficients, beta, for covariates D and Cluster, respectively1And beta2And respectively corresponding regression coefficients of the covariates D and Cluster, and after the prior distribution of each covariate is determined, the coefficient of each covariate can be obtained by adopting a Bayesian survival analysis method. In addition, in order to verify the convergence of the Bayesian model, two MCMC chains were used, and the initial values of the covariate regression coefficients in each chain were (β)Cluster＝0，βD0 and (. beta.) ofCluster＝0.5，βDAnd-0.5), performing 2000 times of pre-iteration, performing 10000 times of iteration to obtain posterior distribution of regression coefficients of each covariate, and taking the posterior average value as a final parameter estimation result to finally obtain the pipe explosion risk prediction model.
(4) Predicting pipe bursting risk of pipeline by using pipe bursting risk prediction model
And (4) calculating the pipe explosion risk of each pipeline in the water supply network by combining the GIS pipe network and utilizing the constructed pipe explosion risk prediction model. In this example, a high-risk pipeline with a pipe explosion risk of more than 0.04 times/(a · km), a high-risk pipeline with a pipe explosion risk of more than 0.02 and 0.04 times/(a · km), a low-risk pipeline with a pipe explosion risk of more than 0.01 and 0.02 times/(a · km), and a safe pipeline with a pipe explosion risk of 0.01 times/(a · km) or less.
The above embodiments are merely examples and do not limit the scope of the present invention. These embodiments may be implemented in other various manners, and various omissions, substitutions, and changes may be made without departing from the technical spirit of the present invention.

Claims (5)

1. A water supply network pipeline pipe explosion risk prediction method based on Bayesian survival analysis is characterized by comprising the following steps:
(1) establishing a detonator database according to the collected detonator historical data, and extracting key information as covariates;
(2) carrying out spatial clustering analysis on the tube explosion point positions, and quantifying spatial distribution information of the tube explosion point positions to be used as a new covariate to be supplemented to a tube explosion database;
(3) constructing a pipe explosion risk prediction model by adopting a Bayesian survival analysis method based on a pipe explosion database;
(4) predicting the pipe explosion risk of the pipeline by adopting a pipe explosion risk prediction model;
the step (3) is specifically as follows:
(31) and (3) standardizing tube explosion historical data in a tube explosion database: dividing each pipe bursting pipeline historical record into three parts,
Figure FDA0003169243380000011
wherein D isiRepresents the history of the ith burst pipe, tiIndicating the survival time of the ith burst pipe, ZiRepresenting the covariate vector, delta, corresponding to the ith burst pipeiData type indicating variable, delta, representing the ith burst pipei1 denotes the ith burst pipe history as complete data, δiThe ith burst pipe history record is regarded as deleted data as 0, and i is 1,2, … …, N and N are the total number of burst pipes;
(32) establishing a benchmark risk function: the reference risk function takes the survival time of the pipe explosion pipeline as an independent variable and takes the pipe explosion risk of the pipeline as a dependent variable, wherein the pipe explosion risk of the pipeline is the pipe explosion times of the average unit pipe length per year;
(33) selecting main factors influencing the pipe explosion risk as covariates, and establishing a Bayesian survival analysis model for parameter estimation to determine covariate parameter estimation values;
(34) and determining a pipe explosion risk prediction model according to the reference risk function and the covariate parameter estimation value.
2. The Bayesian survival analysis-based water supply network pipeline pipe bursting risk prediction method according to claim 1, wherein the step (2) takes the quantitative result of whether other pipe bursting exists within the set spatial distance of the pipe bursting point as a new covariate, specifically: when other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 1, and when no other tube explosions exist within the set space distance of the tube explosion point location, the quantization result is 0.
3. The Bayesian survival analysis-based water supply network pipeline pipe explosion risk prediction method according to claim 1, wherein the step (33) specifically comprises:
(331) dividing the survival time of all pipe explosion pipelines into intervals of 0 & lt s1＜s2···＜sJAnd < ∞, recording the survival time of the ith explosion pipe as tiAnd for all tiAre all less than sJJ time intervals (0, s) are obtained1],(s1,s2],···,(sJ-1,sJ]；
(332) Assuming a fixed reference hazard function h for each time interval0(ti)＝λj，ti∈(sj-1,sj-1) Obtaining a Bayesian survival analysis model function:
Figure FDA0003169243380000021
wherein, h (t)i,Zi) Risk of tube explosion for the ith tube, tiIndicates the survival time of the ith pipeline, ZiRepresents a covariate vector, beta, corresponding to the ith pipelineTFor regression coefficient vectors corresponding to covariates, λjJ is 1,2, … …, J is the total number of divided time intervals, i is 1,2, … …, N represents the total number of pipe bursting pipelines.
4. The Bayesian survival analysis-based water supply network pipeline pipe bursting risk prediction method according to claim 1, wherein the pipe bursting risk model in the step (34) is as follows:
h(t,Z)＝(at2+bt+c)exp(βTZ)；
wherein h (t, Z) is the risk of pipe explosion, t represents the life time of the pipe, at2+ bt + c is a reference risk function, a, b and c are fitting coefficients of the reference risk function, Z represents a covariate vector composed of main factors influencing the pipe bursting risk, and betaTThe regression coefficient vector corresponding to the covariate.
5. The water supply network pipeline pipe bursting risk prediction method based on the Bayesian survival analysis as recited in claim 1 or 2, wherein the spatial clustering analysis is performed on pipe bursting point positions by adopting a DBSCAN clustering algorithm.
CN201810448471.7A 2018-05-11 2018-05-11 Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method Active CN108615098B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810448471.7A CN108615098B (en) 2018-05-11 2018-05-11 Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810448471.7A CN108615098B (en) 2018-05-11 2018-05-11 Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method

Publications (2)

Publication Number Publication Date
CN108615098A CN108615098A (en) 2018-10-02
CN108615098B true CN108615098B (en) 2021-11-09

Family

ID=63663052

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810448471.7A Active CN108615098B (en) 2018-05-11 2018-05-11 Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method

Country Status (1)

Country Link
CN (1) CN108615098B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109886506B (en) * 2019-03-14 2023-05-23 重庆大学 Water supply network pipe explosion risk analysis method
CN110705018B (en) * 2019-08-28 2023-03-10 泰华智慧产业集团股份有限公司 Water supply pipeline pipe burst positioning method based on hot line work order and pipeline health assessment
CN115293656B (en) * 2022-10-08 2023-04-07 西南石油大学 Parallel oil and gas pipeline domino effect risk analysis method based on Bayesian network
CN116090605A (en) * 2022-12-07 2023-05-09 南栖仙策(南京)科技有限公司 Pipe network early warning method and device, electronic equipment and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102222169A (en) * 2011-06-21 2011-10-19 天津大学 Method for predicting and analyzing pipe burst of urban water supply network
CN103226741A (en) * 2013-05-10 2013-07-31 天津大学 Urban water supply network tube explosion prediction method
CN107145691A (en) * 2017-06-23 2017-09-08 广东青藤环境科技有限公司 A kind of public supply mains booster prediction analysis method

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7412551B2 (en) * 2004-06-30 2008-08-12 Intel Corporation Methods and apparatus for supporting programmable burst management schemes on pipelined buses
CN107425906B (en) * 2017-07-25 2019-09-27 电子科技大学 Distributing optical fiber sensing signal processing method towards underground pipe network safety monitoring

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102222169A (en) * 2011-06-21 2011-10-19 天津大学 Method for predicting and analyzing pipe burst of urban water supply network
CN103226741A (en) * 2013-05-10 2013-07-31 天津大学 Urban water supply network tube explosion prediction method
CN107145691A (en) * 2017-06-23 2017-09-08 广东青藤环境科技有限公司 A kind of public supply mains booster prediction analysis method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"a pipe network skeleton method based on GIS network analysis technologis";tao tao et al.;《ASCE Library》;20120426;第521-528页 *
"一种通用的生命线工程网络事件空间聚类分析算法";张忠贵 等;《灾害学》;20150120;第30卷(第1期);第29-33页 *
"供水管网系统爆管预测";王祎 等;《天津大学学报》;20100915;第43卷(第9期);第840-843页 *

Also Published As

Publication number Publication date
CN108615098A (en) 2018-10-02

Similar Documents

Publication Publication Date Title
CN108615098B (en) Bayesian survival analysis-based water supply network pipeline pipe burst risk prediction method
Ji et al. Probabilistic physical modelling of corroded cast iron pipes for lifetime prediction
US10082478B2 (en) Methods for evaluation and estimation of external corrosion damage on buried pipelines
CN103455682B (en) A method of prediction hp-ht well corrosion set pipe residue lifetime
CN108894282B (en) Municipal pipe network operation safety dynamic early warning method
CN111178758A (en) Concrete dam monitoring data intelligent management and real-time evaluation system based on BIM
CN102072409B (en) Pipe network leakage monitoring method combining leakage probability calculation and recorder monitoring
CN107133400A (en) A kind of aircraft structure fatigue reliability Bayes&#39;s combination forecasting method
CN110929359B (en) Pipe network siltation risk prediction modeling method based on PNN neural network and SWMM technology
CN112232522B (en) Intelligent recommendation and dynamic optimization method for deep roadway support scheme
Amiri-Ardakani et al. Pipe break rate assessment while considering physical and operational factors: a methodology based on global positioning system and data-driven techniques
CN108460230A (en) Corrosion oil-gas pipeline remaining life based on data fusion and Reliability Prediction Method
CN105653791A (en) Data mining based corrosion failure prediction system for on-service oil pipe column
Shu et al. A multi-task learning-based automatic blind identification procedure for operational modal analysis
Assad et al. Data mining algorithms for water main condition prediction—comparative analysis
Xu et al. A cloud model-based risk assessment methodology for tunneling-induced damage to existing tunnel
Luo et al. A prediction model of structural settlement based on EMD-SVR-WNN
CN116482227B (en) Pipeline corrosion monitoring method, device and system
CN111859779A (en) Early warning method and device for preventing third-party construction damage risk of gas pipe network
Sheng et al. Risk assessment of urban gas pipeline based on different unknown measure functions
CN111125965A (en) Method and system for evaluating reliability of emergency regulation and control equipment under disaster-causing factor condition
Gao Systematic review for water network failure models and cases
CN105678053A (en) Method for predicting corrosion and invalidity of in-service tubing strings based on data mining
Yusof et al. Markov chain model for predicting pitting corrosion damage in offshore pipeline
CN112884170A (en) Predictive intelligent operation and maintenance system and method for comprehensive pipe gallery

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant