2000

Malm, Timo; Kivipuro, Maarit

The aim of the safety validation process is to prove that the product meets the safety requirements. Safety validation of complex programmable systems has become an increasingly common procedure since programmable systems have turned out to be useful also in safety related systems. However, a new kind of thinking related to the whole life cycle of the programmable product is needed and new validation methods (analysis and testing) to support the old methods are inevitable. This means that methods such as failure mode and effect analysis (FMEA) are still applicable, but they are not sufficient. Methods are needed also to guarantee the quality of the hardware and software. The main validation methods are analysis and tests, and usually both are needed to complete the validation process. Analysis is very effective tool to validate simple systems thoroughly, but a complete analysis can be ineffective against failures of modern programmable electronics. Large programmable systems can be so complicated that a certain strategy in the validation process is necessary to keep the resources required reasonable. A good strategy is to start as early as possible and at the top level (system level). It is then possible to determine the safety critical parts by considering the safety requirements, categories (according to EN 954), safety integrity levels (according to IEC 61508), and the structure of the system. The critical parts are typically parts that the system rely on and which have some properties which cannot be seen clearly at the top level. A newly arising problem is that large programmable systems are becoming difficult to realise and the analysis is often difficult to understand. Figures can often illustrate the results of the analysis better than huge tables. However, there is no all-purpose excellent illustrating method, but the analyser needs to draw figures so that the main subject is well brought out.