Pré-despacho de carga em usinas termoelétricas considerando a gestão da manutenção via lógica fuzzy

Abstract

This thesis presents a new proposal for load pre-dispatch considering the technical conditions of the engines of thermoelectric power plants by combining several maintenance and diagnostic techniques and using computational intelligence based on fuzzy logic. Diagnosis of the technical conditions of the engines is done using lubricant analysis, vibration analysis, and thermography. With this data and with the statistical analysis it is possible to predict when an engine can fail and consider this in the load pre-dispatch. To increase engine reliability and power supply, a Maintenance Management Program (MMP) is developed using management tools, applying only 4 TPM (Total Predictive Maintenance) pillars and combined them with predictive maintenance and diagnostics, thus allowing to reduce failures of plant equipment. Some results achieved after the implementation are: reduction of the annual cost of maintenance, reduction of corrective maintenance, increase of the MTBF (Mean Time Between Failures) and decrease of MTTR (Mean Time To Repair) in all areas. In addition, the proposed pre-dispatch scheme ensures the demanded power with a high degree of reliability and quality.