Automatic leak detection system in water pipes network

• Grant number: 17/00798-3
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Duration: March 01, 2019 - August 31, 2020
• Field of knowledge: Engineering - Sanitary Engineering - Basic Sanitation
• Principal Investigator: Antonio Carlos Oliveira Júnior
• Grantee: Antonio Carlos Oliveira Júnior
• Host Company: Stattus4 Cidades Inteligentes e Sustentabilidade Ltda
• CNAE: Fabricação de aparelhos e equipamentos de medida, teste e controle ; Desenvolvimento e licenciamento de programas de computador não-customizáveis ; Tratamento de dados, provedores de serviços de aplicação e serviços de hospedagem na internet
• City: Sorocaba
• Pesquisadores principais: Anderson Fraiha Machado
• Associated researchers: Ana Maria Frattini Fileti ; Thiago Ragozo Contim
• Associated research grant: 15/01100-4 - Water leak detection system using machine learning, AP.PIPE
• Associated scholarship(s): 19/07594-0 - Automatic leak detection in water pipes network, BP.TT; 19/08046-6 - Automatic leak detection in water pipes network, BP.TT

Abstract

The present research has as objective the development of a prototype of a platform of detection of water leakage, which will be composed by: sensors distributed in the field; a software of artificial intelligence; and the web/mobile for the management of the sensors and data visualization. Brazil is the country that has the largest reserve of fresh water in the world, but about 20% of the cities are in emergency situation due to lack of water. This fact makes concessionaires worry about distribution efficiency as one of their main operational objectives. Traditional methods of leak detection under the soil depend on: the rods equipped with a pickup device; the acoustic; and the geofones; performing the inspection hearing on the lines of the pipe. In this case, the effectiveness of the detection depends intrinsically on the experience of the professional. The project methodology is structured in the experimental research, whose results will be the prototype pilot - Platform auto-detection of water leakage in the distribution network, composed of sensors distributed in the field, interconnected by a Mesh network, sending data to a remote server responsible of the characterisation of the samples as leaking or not leaking. Such classification is done by an Artificial Intelligence software configured in an architecture that allows interaction and continuous learning. These are the challenges presented, and that they design an operational improvement significant, in view of the difficulty of detecting leaks in the field, added to the scarcity of skilled labor, and territorial coverage. The operational costs for the execution of the tasks of leak detection are costly for the concessionaire, which in turn are not exercised on a large scale, which generates a waste of resources. (AU)