Reliability Engineering

Component Reliability vs System Reliability

Reliability Engineering splits naturally into two areas viz component reliability and system reliability. For our purposes, components may be defined as items which are not repairable, which are replaced rather than repaired. Systems on the other hand are defined as assemblies of several components.

Component Reliability

Analysing failures in time and predicting failures is a matter of applying statistical techniques. In this regard, Weibull Analysis is one of the most useful. Often taught, seldom used, the technique is used by progressive companies to refine product designs and, where failures may be economically unavoidable, to develop optimal maintenance regimes.

It is often said that companies do not have the data to perform Weibull Analysis or to predict failures. This is often not true. There is usually a considerable amount of statistical information in the heads and notebooks of persons in various positions in the company. While this data may not be perfect, it is often very useful and usable. It must always be remembered that the cost of perfect information is infinite. If one is to wait for perfect data, one will never do any analysis. The other point to remember is that analysis begets data, not the other way around. This is one of the counter-intuitive facts of statistical analysis. When one proceeds with an analysis, and the persons contributing begin to see the value of what one is doing, this creates the motivation to glean more data from the system and to preserve it for future use. Data collection systems often fail, because the data is not used and the collectors see no point and lose heart. (Or a new software system is installed and the data from the old system is not transferred – but that is another story……)

Sparse or inadequate data can be supplemented by data synthesis. This is a technique for collecting the qualitative and subjective experience of a group of persons and optimally quantifying it. This is sometimes essential to fill in gaps in an analysis, for example a risk analysis to predict the probabilities and consequences of operator error.

System Reliability and Availability

System Reliability and Availability may be predicted (within limits) by constructing models of the system. The model will consist of components linked together in a reliability sense. By applying various mathematical techniques, including simulation, a prediction of the resulting reliability and availability of the system may be obtained. Modifications may then be introduced, such as adding redundancy, to improve the reliability and or the availability of the system. This technique is particularly valuable in refining the design of heavy mechanical systems, such as mineral processing plants. Provided the limitations in the accuracy of the technique is understood, valuable insights into plant performance can be gleaned in the design stage.

Reliability Engineering as the Foundation for Modern Maintenance Practice

Modern Maintenance Engineering, has as its foundation, the principles of Reliability Engineering. This is because unless one understands the principles of plant and component degradation which lead to the need for maintenance, one cannot maintain optimally. And Reliability Engineering provides this understanding.