Improvement of Dam Management in Terms of WAM Using Machine Learning
The evolution of a data-rich ecosystem for Water Asset Management (WAM) comes with a rapid changing of the information landscape. Data inventories are created with asset information, which includes information such as maintenance history, piping ID, flow, failure mode, risk assessment history, uncertainty, spillway, weir, slope, and seepage. With more massive datasets across departments involving in dam operation management, data silos bring challenges of integrating and utilising the data sources into a single platform for analysis, visualisation, and decision-making. Therefore, this research intends to introduce a conceptual framework for WAM to meet the needs of practical engineering applications in dam management using machine learning. The idea of engaging machine learning is to determine rules that are embedded within the data by applying concepts like pattern recognition and data analytics. The rules are replicated into algorithms to model prediction, forecast or projection. Collecting and analysing data from the total asset of dam safety monitoring brings valuable insights in terms of asset management strategies. For instance, enhancing optimal maintenance and replacement of the assets, and consequently improving the utilities more efficiently. This paper discusses the research framework on how machine learning is used as a transformative tool for the data-rich world of WAM. The proposed approach offers the potential of a relatively new area of research in Malaysia dam management beneficial to dam managers and stakeholders, which facilitates effective data management and developing strategic benefits to the community, environment, and economy.
Authors
Abstract
The evolution of a data-rich ecosystem for Water Asset Management (WAM) comes with a rapid changing of the information landscape. Data inventories are created with asset information, which includes information such as maintenance history, piping ID, flow, failure mode, risk assessment history, uncertainty, spillway, weir, slope, and seepage. With more massive datasets across departments involving in dam operation management, data silos bring challenges of integrating and utilising the data sources into a single platform for analysis, visualisation, and decision-making. Therefore, this research intends to introduce a conceptual framework for WAM to meet the needs of practical engineering applications in dam management using machine learning. The idea of engaging machine learning is to determine rules that are embedded within the data by applying concepts like pattern recognition and data analytics. The rules are replicated into algorithms to model prediction, forecast or projection. Collecting and analysing data from the total asset of dam safety monitoring brings valuable insights in terms of asset management strategies. For instance, enhancing optimal maintenance and replacement of the assets, and consequently improving the utilities more efficiently. This paper discusses the research framework on how machine learning is used as a transformative tool for the data-rich world of WAM. The proposed approach offers the potential of a relatively new area of research in Malaysia dam management beneficial to dam managers and stakeholders, which facilitates effective data management and developing strategic benefits to the community, environment, and economy.
Citation
Plain Text: Fazli, Bashirah & Abdullah, Mohammad & Zulkifli, Harlisa & Kasim, Mohd Fadhil & Lee, Hin & Noh, Md & Yahya, Fara. (2020). Improvement of Dam Management in Terms of WAM Using Machine Learning. 10.1007/978-981-15-1971-0_23.
BibText: @inbook{inbook, author = {Fazli, Bashirah and Abdullah, Mohammad and Zulkifli, Harlisa and Kasim, Mohd Fadhil and Lee, Hin and Noh, Md and Yahya, Fara}, year = {2020}, month = {01}, pages = {226-236}, title = {Improvement of Dam Management in Terms of WAM Using Machine Learning}, isbn = {978-981-15-1970-3}, doi = {10.1007/978-981-15-1971-0_23} }