A comparative study of two novel reliability/availability calculation methods applied to a hydroelectric power plant

In this work we compare two novel methods for monitoring reliability/availability of hydroelectric power system. Both approaches are based on the construction of a dependency matrix that represents the functional relationship between systems, subsystems and equipment. In the first (qualitative) approach, given a state vector indicating the reliability/availability of each component classified as available or unavailable, a scanning algorithm determines whether the system is available, degraded or unavailable. The second (quantitative) approach propagates the initial availability/reliability state of the components, given in a scale ranging from 0 to 1, using a Markovian iterative method that converges to the state of the system which is also given in the same availability/reliability scale. Both approaches were tested with a case study consisting of a simplified hydroelectric power plant. The results shows good agreement of both approaches in common situations and suggests using the quantitative method whenever reliable time dependent data are available.