Article Preview
Top1. Introduction
A systematic classification of generic methods for reducing technical risk is crucial to safe operation, engineering designs and software. However, this very important topic has not been covered with sufficient depth in the reliability and risk literature. For many decades, the focus of reliability research has been primarily on reliability prediction rather than reliability improvement and risk reduction. Acquiring the relevant knowledge and data related to the failure mechanisms and quantifying all types of uncertainty, necessary for a correct prediction of the time to failure is a formidable task. However, this task does not need to be addressed if the focus is placed on the reliability improvement and risk reduction instead of reliability prediction. The reliability improvement and risk reduction methods do not normally rely on reliability data or on a detailed knowledge of physical mechanisms underlying possible failure modes. They derive their strength from generic laws, invariants and patterns associated with increased reliability and reduced risk. As a result, these methods are particularly useful in developing new designs, with no failure history and with insufficiently researched failure mechanisms. Recent work on formulating generic principles and methods for improving reliability and reducing technical risk has been done in (Todinov, 2015). The present paper contributes to the exiting work an important reliability improvement and risk reduction method referred to as ‘the method of separation.’
Harmful interaction of factors critical to reliability and risk is a major source of failures. Separation is the act of disuniting risk-critical factors. Separating risk-critical factors to reduce the harmful interaction is therefore a major avenue for improving reliability and reducing risk. Surprisingly, the method of separation has not yet been discussed as a risk-reduction tool. Despite that some techniques used in engineering are clearly instances of separation, they have never been recognised as such and have never been linked with this method.
Thus, deliberate weak links and stress limiters have already been used for preventing the stresses from reaching dangerous levels. The deliberate weak links are consciously designed weak points that are easily replaced (Eder, 2008) and usually protect expensive devices. As it will be demonstrated later, this technique is essentially an instance of separation on a parameter but it has never been n linked with the method of separation.
Another example can be given with the concept ‘barrier’ (Svenson, 1991; Leveson, 2011; Hollangel, 2016). Barriers have also been used as accident prevention tools and protection measure mitigating the consequences from an accident. A classification of barriers has been proposed in (Eder and Hosnedl, 2008). Despite that barriers are also instances of separation, no link has ever been made with the method of separation. Barriers distancing triggers from hazards reduce the likelihood of an accident while barriers distancing hazards from targets reduce the consequences given that accident has occurred.
It is only recently that the method of separation has been suggested in (Todinov, 2015) as a potential risk reduction tool but the discussion of the separation method was very limited and did not cover mechanisms through which separation achieves reliability improvement and risk reduction. Furthermore, no classification of separation techniques has been proposed.
One of the ways to make engineering systems more efficient is to increase the temperature, pressure and speed of the operating fluids and reduce the cross sections of components to reduce weight. However, the increased temperature, pressure and speed of the fluids accelerate the degradation of the components while reducing the cross sections leads to increased stresses and stress amplitudes which increase the risk of failure due to fast fracture and fatigue. Similarly, increasing the efficiency of manufacturing requires increasing the speed of operations which leads to low precision and unreliability.