Desulphurization Plant Monitoring and Fault Detection Using Principal Component Analysis

Abstract

Reliability, safety and efficiency of power plants become increasingly important due to the demand for cost-efficient energy production and tightening environmental regulations. Equipment malfunctions and faults are typical in industry but may lead to reduced production, shutdown of the plant or fatalities at the worst. Certain types of equipment faults induce exceptional behaviour that can be detected on the monitored variables and diagnosed before the severely damaging effects have occurred. Early intervention is often more cost-effective than allowing the equipment to fail. In this study, principal component analysis with different monitoring indices is used to monitor a desulphurization plant removing the sulphur dioxide from the flue gas of a CHP plant. Contributions of variables to the monitoring indices are checked during a special event. The approach is tested during normal process operation and during a period with a malfunctioning pump. The results show that the approach has potential for the early detection of an incipient fault.