Abstract
The reliability question of the components, specifically the process control valves, became an important point to be investigated in nuclear power plants and others areas such as oil refinery or offshore oil platform, considering security and extension life of the plant. Therefore, the necessity of improvements in monitoring and diagnosis methods started to be the extreme relevance in the maintenance predictive field, as mentioned by Guimarães Carneiro (2007), establishing as main goal the reliability and readiness of the components systems. The development of nonintrusive monitoring and diagnostics method makes it possible to identify malfunctions in plant components during normal plant operation. The purpose is to present an analysis and diagnostic of a water flow control valve that charges a steam generator, simulating part of the secondary circuit of a power water reactor. This facility is part of the laboratory of tests of propulsion equipment of the Brazilian Navy, in São Paulo – Brazil (2007). The methodology of this project is based on the graphic analysis of two parameters; the valve actuator air pressure and the valve plug displacement. The data come from intelligent positioner; it is a kit of Delta V™ Automation System (2010). The analysis is implemented on detection of anomalies approach using Expert Systems by Fuzzy Logic technique based on rules and knowledge base. A distinct approach as mentioned by Choudhury ( 2005), Desborough (2001), Gerry (2001) or Camila Pinto (2010). Once the baseline measurements of control valves are taken, it is possible to detect fault symptoms, leaking, attrition, clamping, damage, etc. The system of monitoring and diagnostic was designed at the MATLABTM platform “FUZZY LOGIC TOOLBOX”.
