Skip to main content

Table 1 A summary of predictive analytics applications for water pipes failure

From: Asset management analytics for urban water mains: a literature review

Cluster of research

Techniques

Applications

Input parameters

Deterministic model

– Multiple Regression models

– Linear Regression

– Poisson Regression

– Multivariate adaptive regression splines

– Water mains annual break rates prediction

– Finding correlation between water main failure rates and input parameters

Physical Factors:

DIA, LEN, AGE, MAT

Probabilistic model

– Cox proportional hazard model

– Weibull proportional hazard survival analysis

– Water main failures prediction

– Time to next break prediction

Physical Factors:

DIA, LEN, AGE, MAT, THK

Operational Factors:

WP, VEL, TRF, RT, WPH

Environmental Factors: SR, SPH, MC, FI, SZN

Artificial intelligence model

– ANN

– Random Forest

– Xgboost

– LR

– SVC

– Evolutionary Polynomial Regression

– Boosted regression tree

– Predict the failure rate of pipeline networks

– Binary classification which shows whether or not the pipe break

Physical Factors:

DIA, LEN, AGE, MAT, YEAR, NC, NT

Operational Factors:

WP, TRF, NB, BD, BY

Environmental Factors:

MC, ST, PP, LU, LO