Predicción de fallas mecánicas en motores de combustión interna hyundai- h21/32 en la empresa termoeléctrica Celec Ep Central Sacha.

Abstract

The project aims to predict potential failures early in reciprocating motors used for power generation and thus prevent severe and costly damage. It was carried out for the Electric Corporation of Ecuador (CELEC) in its thermoelectric plant located in La Joya de los Sachas. The plant has 12 Hyundai 9H21/32 power units that burn fuel oil and diesel. The units have corrective and preventive maintenance for hours of operation, but there is no adequate predictive maintenance supervision that reduces component wear, improves efficiency and therefore the economic aspect. It was decided to use vibration and ultrasound analysis due to their efficiency and little intervention during the operation of the machines. The information on the operation and adequate operating ranges was collected, choosing a sample of 4 MDUs and using a portable signal detection equipment called Windrock that has a digital screen, pressure, ultrasound and vibration sensors. The signals obtained were compared with a database made based on other projects and between past measurements. Each signal was separated into events that present similarities and differences in their early or late action, detection of short or large impacts, lack or intermittence of signals. In this way, reports of operating failures are obtained before and after each preventive maintenance and check the efficiency of the predictions. A practical manual for the use of the Windrock equipment, a comparative database applicable in software and explanatory videos were made