Understanding the impact of failures on modern astronomical instruments

The size and complexity of ground-based astronomy instrumentation is continuously increasing in order to make most efficient use of the existing and new facilities. The logic of this trend is clear, since the high cost of building and maintaining telescopes leads to high demand for their use, maximising the science output is essential. Instrument multiplex and operational modes are therefore ever-increasing. In highly complex instruments, it can be difficult to interpret the meaning of the term “failure” since the loss of individual channels or modes is not the same as a total loss of science capability. In this paper we explore the relationship between the instrument reliability and the delivered science. We also discuss the need to incorporate a good understanding of the reliability in the instrument design. We consider the difference between the inherent availability of the instrument, the operational availability and the scientific output. At the UKATC we have developed a process and set of tools to analyse an instrument design to determine both the inherent and operational availability using a combination of Failure Mode, Effect and Criticality Analysis (FMECA), Reliability Diagrams, and Fault Tree Analysis. These can then be considered in terms of the instrument productivity to determine strategies for redundancy, maintenance and repair. We also review the state of reliability analysis within the ground-based community and compare it with the requirements for space astronomical instrumentation. The methodology and tools developed are intended to be compliant with the requirements for space products.