Knowledge techniques for the analysis of urban runoff using SWMM
Abstract
An expert system was built to facilitate and to automate the calibration of the Runoff block of the Storm Water Management Model (SWMM). The expert system is divided in two parts that can be linked together. The first one calibrates the runoff quantity simulation model. The second, that calibrates the runoff quality simulation module, has two versions. One calibrates the model with the only objective of a correct prediction of the pollutant loads. The second calibrates the model on the basis of the loadographs shapes. The expert systems act as a front end to counsel the user on the choice of the parameters. It interprets the results using backward chaining and suggest some useful changes in the values of the relevant parameters using forward chaining. The calibrations achieved by means of the expert systems has been tested on several watersheds. It has been found that, for runoff quality simulation, the accuracy of the prediction results is limited by the quality of the calibration data that are used and by the model itself that does not take into account all the processes involved in the transport of pollutant by urban runoff. A fuzzy set approach is proposed to try to take into account some of the uncertainties related with runoff quality simulation. A simplified fuzzy set version of SWMM is proposed that has been tested on two watersheds, giving some encouraging results.
Degree
Ph.D.
Advisors
Delleur, Purdue University.
Subject Area
Hydrology
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.