Reliability Comparison of Uninterruptible Power Supply (UPS) System Configurations

The paper presents the reliability study of Uninterruptible Power Supply (UPS) system configurations. The five main UPS system design configurations namely Capacity, Isolated Redundant, Parallel Redundant, Distributed Redundant, and System plus System Redundant configurations were considered, and comparisons on the resultant system's reliability parameters were discussed in detail. The results obtained from this study can play an essential role in the selection, maintenance and application of the UPS systems. UPS reliability parameters such as Failure Rates, Mean Time Between Failures (MTBF), and Availability can be estimated from the proposed system's reliability estimation methods. The suggested reliability estimation methods in this paper are the Reliability Block Diagram (RBD) and the Monte-Carlo Simulation methods. Reliability Block Diagram (RBD) approach is a reliability estimation method where the system's functional block diagram is constructed to show how the system is partitioned and described how it is expected to operate. In this method, the components in the system is represented by a block that are considered to exist in one of the two possible states; operational or failed. The component's reliability parameters (i.e. failure and repair rates) are assigned to every block and the system's reliability estimation was conducted according to the resultant block diagram and the operating conditions of the system. The Monte'Carlo simulation method is a technique in which the failure and repair rates of the component in the system are simulated by the use of random number generators. The system's reliability and availability calculation are applied to each sample to produce the deterministic reliability parameters that try to mimic the result that would be obtained from field trials. The reliability model of the system has to be constructed first and the component's failure and repair rates have to be determined. The component's reliability and availability parameters can then be calculated and simulated using the random number generator, stating its limits and the number of simulations. In both methods, reliability levels of each configuration were estimated using the reliability parameters of the components that made up the system. Several reliability parameters that are useful in estimating the reliability of UPS systems were determined and discussed.