Wadiso 6 User Guide > Optimization > Optimization technique and general concepts

Optimization of water distribution systems with Wadiso consists of finding the least cost combination of pipes subject to a set of constraints (e.g. minimum pressures, ranges of pipe sizes) provided by the user. The intent of the Wadiso optimization routine is to use these constraints to size specific pipes in the system. The optimization technique employed by Wadiso is a very simple one. First, the user specifies a range of candidate diameters for each pipe or group of pipes that needs to be sized. One or more load cases are then specified, each with a minimum pressure requirement at the nodes in the network. One such load case may be the peak water demand condition, another a fire-flow condition at a critical node. The program will enumerate all possible pipe size combinations (exhaustive trial-and-error), test each combination to see whether the specified minimum pressures are met, and select the workable solution with the lowest total cost (including energy cost for the pumping and/or tank cost if desired). Where other elegant mathematical optimization techniques normally fail due to the complexity of water distribution systems and the multitude of scenarios that need to be considered in the design of such systems, the enumeration procedure guarantees the global minimum within the specified constraints.

The user may specify a size of 0 (zero) as a candidate size for a pipe to be sized. Selection of size 0 implies elimination of the pipe, at no cost. The optimization routine also allows the user to consider cleaning and re-lining of existing pipes. The optimization can be carried out for one or several load cases for water demand, each with its own set of specified minimum pressures at the nodes.

The optimization routine can also be instructed to determine Pareto Optimal solutions that are close to the optimal solution in terms of cost and minimum pressure. A combination of pipe sizes is said to be a Pareto Optimal (non inferior) solution if there is no other solution that has better pressure at a lower cost. By relaxing the minimum pressure constraint and allowing solutions which result in costs within a certain percentage of the minimum cost, a series of sub-optimal solutions, which give either higher pressures at an increased cost, or which slightly violate the pressure requirements at reduced cost, can be generated. The degree to which the pressure constraint is relaxed and the allowable percentage variation in cost are values which the user can either specify or rely on default values.

The following key aspects of the optimization algorithm warrant further discussion:

•Reasons for an enumeration algorithm

•Reducing the number of candidate solutions

•Enumeration algorithm

•Pareto Optimal solutions

•Pipe cleaning as an alternative to adding new pipes

•Including pumping cost into the optimization

•Including tank cost into the optimization

•Importance of redundancy in the system