Refining PSO Applied to Electric Energy Cost Reduction in Water Pumping - page 05

Next, the comparative graphs of the pressure evolution in the reference nodes are displayed.


Figure 6: Pressure variation at node 15: comparison of techniques


Figure 7: Pressure variation at node 20: comparison of techniques


Figure 8: Pressure variation at node 25: comparison of techniques


Figure 9: Pressure variation at node 30: comparison of techniques



By analysing the network from Carrijo (2004) using the PSO-R, the performance of the optimizer was obtained in terms of the final operation cost compared with that of the genetic algorithm. In economic terms, when compared with the fixed rotation scenario, the PSO-R results in a reduction of 63.5% in operational cost.

The PSO-R performed 888000 evaluations of the objective function (216 particles with 3000 iterations in the binary phase, and 120 particles and 2000 iterations in the continuous phase) for the presented network, where the stop criterion for both phases was the absence of improvement of the objective function for at least 100 consecutive iterations.

From an operational point of view, the method proved to be effective in determining the operational rules for the assessed network, which resulted in high energy savings by reducing the operational cost of the pumping systems. Furthermore, the optimized routine resulted in better use of the variable level reservoir, whose emptying and filling movement assured a better water quality.

Moreover, the optimized routine exhibited less variable behaviour for the evaluated pressures at the relevant nodes of the studied system. The less abrupt pressure variation softens the stresses on the pipes, which subsequently increases the useful life of the pipes due to less fatigue processes.

By assessing the optimization method in a continuous phase, the results indicated that the lower limit for the specific rotations requires further evaluation because in several cases, the frequent use of a specific low rotation may lead to an undesirable phenomenon, such as cavitation, excessive vibration, or even overheating the equipment.

Finally, the use of computational tools to operate delivery systems can be extremely useful when used with clear knowledge of hydraulic issues and computational methods involved in the problem. The indiscriminate use of computational models could create serious issues for system operations, and therefore, there is a need to provide tools to the operators that can unite the operation practice with the ease of the computational tools.