Reality vs. Expectations for ESP Reliability Analysis

Abstract

If you're reading this, then it's likely that your first reaction to the paper's title was something along the lines of ‘oh no, not another paper on reliability’. Believe me, I totally understand: ESP reliability – something that is fundamentally important to both ESP suppliers and end users – often gets treated in technical papers with unnecessarily complex and arcane mathematics. Fundamentally, however, it can be and should be kept reasonable so that everyone can discuss it using the same concepts and terms. This paper, therefore, will not try to create complicated new metrics. It will use existing runlife analysis techniques to analyze Perenco's entire ESP history and to highlight how different measures can be used in different ways. At each step, the paper will compare the results of the various measures to expectations in order to gain some real evidence-based insight into ESP reliability, which is often surprising. Next, the paper will create logical conclusions based on this evidence that potentially apply to general ESP reliability behavior. And finally, based on those conclusions, the paper will address the implications on how to manage ESP's. Before getting started, it's important to emphasize a couple points about reliability analysis. The first point is that there is no single evaluation that will tell you everything you want to know about reliability. There are many different analyses possible because there are many different questions to address on reliability, such as: ‘which ESP's are at risk?’ ‘are we improving?’, ‘did this change make a difference?’, or ‘how many ESP failures should I plan for?’ The second point is that, because of this inherent diversity in analysis, and despite my feeling that the subject can be kept at a level most people can understand, there will always be a certain level of complexity to statistical evaluation. It takes time to explain and understand, and this can be difficult to condense into a 5-minute ‘elementary school level’ presentation, as is often requested.