Reliability prediction analysis and reliability improvement of the satellite power control unit

Costly and critical equipment that can operate in harsh space condition such as radiation and temperature should be designed with high reliability and redundancy because of the maintenance difficulties. To prevent satellite loss due to equipment failure, the design should be developed by considering the results of qualitative and quantitative reliability analysis performed at design phases. In this way, the risks that will adversely effect the product life are eliminated or minimized and the reliability targets determined at the conceptual design phases are ensured. In this thesis, the current reliability approaches in space systems were researched and new reliability improvement solutions were presented. The space-qualified Power Control Unit (PCU), one of the most vital units in a satellite, was developed with reliability analysis parallel to the design by means of PTC-WQS software. In addition to numerical reliability improvements, qualitative requirements such as prevention of single point failure and failure propagation were also considered. The details of the three proposed reliability improvement methods have been demonstrated in practice on the circuit board based on the DC/DC buck converter topology and GKB design with internal redundancy is presented meeting rigorous reliability targets. First of all, part stress analysis, which controls the electrical and thermal stress levels on the electronic components, was performed to extend component life, increase reliability and enhance the end-of-life performance of equipment. Secondly, to provide a cost-effective solution, the structure of redundancy was determined by drawing reliability block diagrams and the design was ensured to meet the quantitative reliability target thanks to Reliability Prediction Analysis. Finally, Failure Mode, Effects and Criticality Analysis (FMECA) was carried out parallel with the PCU design process to eliminate or minimize potential failures and their effects, increase the rate of failure detection, and prevent single point failure and failure propagation. To achieve this, a new hybrid FMECA approach which is free from subjectivity and has functional, hardware and interface features with adjustable resolution, has been proposed and demonstrated in practice.