Estudo da conveniência da identificação do fator de atrito e da rugosidade em redes de distribuição de água através do método transiente inverso com algoritmo genético

Abstract

Water is an essential natural resource for all life forms and their distribution should be carried out with quality and without waste. One of the methods of optimization of water distribution systems is based on the simulation (modeling) of real hydraulic networks using computational models that can predict their behavior in various situations the logo of its useful life. The main steps of this method are calibration; operation and control; design and optimization; and route networks. The idea of ​​this work is to improve one such technique, the calibration parameters of the pipes identification process (friction factor, surface roughness, diameter and the like) existing in networks in which they are considered unknown. The methodology adopted is the Transient Inverse Method (MTI) with optimization of the solution selection technique by Genetic Algorithm (GA). The main objective is to calibrate the main parameters (friction and roughness factor) to analyze the convenience of two techniques: calculation of the friction factor by the method by MTI-AG and from the roughness (using the formula Swamee, 1993 ). The study is carried out for two taken synthetic networks of literature, but represent real systems. The conditions for the analysis are: two maneuvers valves (slow and abrupt) responsible for the transient event, the monitoring of transient loads in only 20% of the nodes of the networks and the use of two types of solution selection of genetic algorithm (with elitism and without elitism). The experiments are carried out through three programs: the first calculates the permanent conditions, the second provides transient loads through the method of characteristics (MOC) in the search for solutions of the equations of motion in the transient flow and the third, who works so combined with others, selects the best results through iterations performed according to the genetic algorithm, a technique inspired by the mechanisms of evolution of living beings. The results indicate that regardless of the conditions imposed is initially more efficiently identify the friction factor from the absolute roughness than calibrate this factor by MTI-AG due to the great variability of friction factors for the occurrence of the transient event.