Abstract. The difficulty in knowledge representation of a water distribution network (WDN) problem has contributed to the limited use of artificial intelligence (AI) based expert systems (ES) in the management of these networks. This paper presents a design of a Decision Support System (DSS) that facilitates "on-demand'' knowledge generation by utilizing results of simulation runs of a suitably calibrated and validated hydraulic model of an existing aged WDN corresponding to emergent or even hypothetical but likely scenarios. The DSS augments the capability of a conventional expert system by integrating together the hydraulic modelling features with heuristics based knowledge of experts under a common, rules based, expert shell named CLIPS (C Language Integrated Production System). In contrast to previous ES, the knowledge base of the DSS has been designed to be dynamic by superimposing CLIPS on Structured Query Language (SQL). The proposed ES has an inbuilt calibration module that enables calibration of an existing (aged) WDN for the unknown, and unobservable, Hazen-Williams C-values. In addition, the daily run and simulation modules of the proposed ES further enable the CLIPS inference engine to evaluate the network performance for any emergent or suggested test scenarios. An additional feature of the proposed design is that the DSS integrates computational platforms such as MATLAB, open source Geographical Information System (GIS), and a relational database management system (RDBMS) working under the umbrella of the Microsoft Visual Studio based common user interface. The paper also discusses implementation of the proposed framework on a case study and clearly demonstrates the utility of the application as an able aide for effective management of the study network.
Development of a decision support system for the maintenance of water distribution network