Discussion. Optimal Design of Stormwater Drainage Networks for Roads

Abstract

Dr T e m p l e m a n and M r W a i t e r s The Paper covers theoretical aspects of the methods we have developed. Fig. 11 shows the layout of the optimum design of the computer program. At its core is the existing Highway Engineering Computer Branch's program DAPHNE which is used twice. Using BIPS road data and user-specified drainage data, DAPHNE is first used to prepare a conventional minimum cover design assuming manholes at each possible location. This provides design flows, pipe sizes and depths. These sizes and depths are used to define a new range of depth and diameter variables for optimization and an optimum design is then generated which has optimal manhole positions and pipe diameters. This optimum design is again input to DAPHNE which adjusts the pipe slopes keeping the optimal manhole positions and pipe diameters.37. As DAPHNE is a practical design program the optimizing program which relies heavily on DAPHNE also produces fully practical designs. Our program uses only those methods currently used in design offices for drainage design but uses them in an optimal way to produce cheaper designs than could be produced manually. Initial results seem to show that in terms of cost savings size/slope optimization saves a lot, manhole positions rather less and cross-drain positions very little. In its final version the optimizing program omits the variable cross-drain facility because the difficulties of integrating it with the DAPHNE program seemed to outweigh the potential advantages. Fig. 11 shows that the rational design basis could be easily replaced by alternative methods if necessary.The Authors' methods should provide engineers with powerful tools to produce economic and efficient designs. However, their use will depend on a number of factors including their reliability and robustness, the amount of data to be collected and prepared for the programs, and their success.39. Design methods have a tendency to perform well for those who develop them but less well for other users. I think designers should be encouraged to use the methods advocated. Therefore, before they are made generally available, the methods should be clearly and concisely documented, and objectively tested by