Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model

Jeroen T. Verheugd; Paulo R. de O. da Costa; Reza Refaei Afshar; Yingqian Zhang; Sjoerd Boersma

doi:10.1109/SMC42975.2020.9282941

Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model

Jeroen T. Verheugd
, Paulo R. de O. da Costa
, Reza Refaei Afshar
, Yingqian Zhang
, Sjoerd Boersma

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

9 Citations (Scopus)

638 Downloads (Pure)

Abstract

Water distribution networks have shown an increased rate of failure due to material deterioration. In this paper, we apply a Recurrent Neural Hawkes Process model to learn the failure intensity function of water pipes. The failure intensity function is learned based on two components: the base failure rate that is determined by the unique pipe profile attributes, and the effect of past failures. Compared to the existing solutions, our model is able to predict the time to next failure on an individual water pipe level. The learned failure intensity function is used to identify value points in the deterioration process of water pipes that represent their economical end-of-life. We use data from a Dutch water distribution network that consists of 49,600 km of pipelines to test the performance of the proposed model. We have made this dataset available online.

Original language	English
Title of host publication	2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020
Publisher	Institute of Electrical and Electronics Engineers
Pages	2628-2633
Number of pages	6
ISBN (Electronic)	978-1-7281-8526-2
DOIs	https://doi.org/10.1109/SMC42975.2020.9282941
Publication status	Published - 14 Dec 2020
Event	2020 IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC 2020 - Virtual, Toronto, Canada Duration: 11 Oct 2020 → 14 Oct 2020

Conference

Conference	2020 IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC 2020
Abbreviated title	IEEE SMC 2020
Country/Territory	Canada
City	Toronto
Period	11/10/20 → 14/10/20

Keywords

Intensity function
Point Process
Predictive maintenance
Recurrent Neural Networks
Water pipe failure

Access to Document

10.1109/SMC42975.2020.9282941

IEEESMC2020_VerheugdFinal author version , 427 KB

Cite this

Verheugd, J. T., de O. da Costa, P. R., Refaei Afshar, R., Zhang, Y., & Boersma, S. (2020). Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model. In 2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020 (pp. 2628-2633). Article 9282941 Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SMC42975.2020.9282941

@inproceedings{e2545554cd2048c797afe72f752d2e6c,

title = "Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model",

abstract = "Water distribution networks have shown an increased rate of failure due to material deterioration. In this paper, we apply a Recurrent Neural Hawkes Process model to learn the failure intensity function of water pipes. The failure intensity function is learned based on two components: the base failure rate that is determined by the unique pipe profile attributes, and the effect of past failures. Compared to the existing solutions, our model is able to predict the time to next failure on an individual water pipe level. The learned failure intensity function is used to identify value points in the deterioration process of water pipes that represent their economical end-of-life. We use data from a Dutch water distribution network that consists of 49,600 km of pipelines to test the performance of the proposed model. We have made this dataset available online.",

keywords = "Intensity function, Point Process, Predictive maintenance, Recurrent Neural Networks, Water pipe failure",

author = "Verheugd, \{Jeroen T.\} and \{de O. da Costa\}, \{Paulo R.\} and \{Refaei Afshar\}, Reza and Yingqian Zhang and Sjoerd Boersma",

year = "2020",

month = dec,

day = "14",

doi = "10.1109/SMC42975.2020.9282941",

language = "English",

pages = "2628--2633",

booktitle = "2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2020 IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC 2020, IEEE SMC 2020 ; Conference date: 11-10-2020 Through 14-10-2020",

}

Verheugd, JT, de O. da Costa, PR, Refaei Afshar, R, Zhang, Y & Boersma, S 2020, Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model. in 2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020., 9282941, Institute of Electrical and Electronics Engineers, pp. 2628-2633, 2020 IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC 2020, Toronto, Ontario, Canada, 11/10/20. https://doi.org/10.1109/SMC42975.2020.9282941

Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model. / Verheugd, Jeroen T.; de O. da Costa, Paulo R.; Refaei Afshar, Reza et al.
2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020. Institute of Electrical and Electronics Engineers, 2020. p. 2628-2633 9282941.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model

AU - Verheugd, Jeroen T.

AU - de O. da Costa, Paulo R.

AU - Refaei Afshar, Reza

AU - Zhang, Yingqian

AU - Boersma, Sjoerd

PY - 2020/12/14

Y1 - 2020/12/14

N2 - Water distribution networks have shown an increased rate of failure due to material deterioration. In this paper, we apply a Recurrent Neural Hawkes Process model to learn the failure intensity function of water pipes. The failure intensity function is learned based on two components: the base failure rate that is determined by the unique pipe profile attributes, and the effect of past failures. Compared to the existing solutions, our model is able to predict the time to next failure on an individual water pipe level. The learned failure intensity function is used to identify value points in the deterioration process of water pipes that represent their economical end-of-life. We use data from a Dutch water distribution network that consists of 49,600 km of pipelines to test the performance of the proposed model. We have made this dataset available online.

AB - Water distribution networks have shown an increased rate of failure due to material deterioration. In this paper, we apply a Recurrent Neural Hawkes Process model to learn the failure intensity function of water pipes. The failure intensity function is learned based on two components: the base failure rate that is determined by the unique pipe profile attributes, and the effect of past failures. Compared to the existing solutions, our model is able to predict the time to next failure on an individual water pipe level. The learned failure intensity function is used to identify value points in the deterioration process of water pipes that represent their economical end-of-life. We use data from a Dutch water distribution network that consists of 49,600 km of pipelines to test the performance of the proposed model. We have made this dataset available online.

KW - Intensity function

KW - Point Process

KW - Predictive maintenance

KW - Recurrent Neural Networks

KW - Water pipe failure

UR - https://www.scopus.com/pages/publications/85098869096

U2 - 10.1109/SMC42975.2020.9282941

DO - 10.1109/SMC42975.2020.9282941

M3 - Conference contribution

SP - 2628

EP - 2633

BT - 2020 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2020

PB - Institute of Electrical and Electronics Engineers

T2 - 2020 IEEE International Conference on Systems, Man, and Cybernetics, IEEE SMC 2020

Y2 - 11 October 2020 through 14 October 2020

ER -

Predicting Water Pipe Failures with a Recurrent Neural Hawkes Process Model

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this