André-Michel Ferrari — Accendo Reliability New Contributor

Definition of Reliability

The concept of Reliability is often misused, misunderstood, and misinterpreted. Reliability in its academic root, is defined as the probability that a system will perform its intended function in a specified mission time and within specific process conditions. So, it is in essence a probability.

The time variable is crucial in calculating the reliability of a system. Reliability is the Probability of Success. And 1 minus the Probability of Failure.

$$ \displaystyle \large R\left(t\right) = 1 – F\left(t\right) \:\:\:\: (where \: F = Failure \: Probability) $$

Introduction to Reliability or Life Analysis

The basis of the model is a statistical distribution. An example of the process is illustrated in Diagram 1 where a number of “n” centrifugal pumps are run on a test bench until they fail. Each time interval to failure is recorded and once all the failure records are collected, a normalized frequency graph can be constructed. This frequency graph helps define the statistical distribution that best represents the life cycle of a typical component in the population. Using specific mathematical transformations applied to the distribution, the probability of failure after a defined mission time can be derived. And much more information as we will discover in this article.

Introduction

Reliability Engineering analytics come with graphs and visual representations providing the analyst with various types of key information. Some graphs are more popular than others. This article highlights the 4 main graphs used in Reliability Engineering Analysis.

They are Reliability, Probability of Failure, Probability Density and Failure Rate. Less common types are not covered here. Such as Contour Plots. The basis of the graphs in this article is the Weibull Distribution. In other words, all the graphs derive from this very popular distribution. The same concepts applies to other distributions.

Introduction

Sometimes, non reliability practitioners assume that a Reliability Analysis is just about “tossing” a distribution at a set of data. It is a little bit more complex than this. There needs to be some sort of engineering validation on the final outcome. The question to be asked is: “does this life model used make engineering sense?” This article highlights the importance of choosing the correct statistical distribution. The alternative leads to flawed decision making and can be detrimental especially when dealing with critical equipment or million-dollar decisions. A worked example is also provided.

Introduction

This article is a highly simplified overview of Equipment Maintenance Strategies. Its objective is to categorize the types of maintenance approaches for the lay person. Sometimes, explaining the basics using simple terminologies, initiates dialogues with others that are outside of our field of expertise. And as such improve collaboration between teams. How about making this accountant who somewhat controls the maintenance budget, understand what maintenance is really about? On the other hand, that same accountant can educate us on their work practices.

Crow-AMSAA Analysis Overview

Crow-AMSAA Plots have a variety of names, such as Reliability Growth Plots or Duane Plots. The term “Crow” comes from Dr Larry H. Crow, who enhanced James T. Duane’s pioneering launch of this methodology, which was developed in the early 1960s (1). Crow successfully applied the method in the US Army Materials System Analysis Activity (AMSAA). The technique has blossomed into large amounts of new applications in industry such as but not limited to:

• Making reliability more visible and manageable hence contributing to Reliability Improvement Programs,

• Monitoring Design Optimization and Quality Performance,

• Catastrophic or adverse event occurrence predictions and trends which in essence are the objective of this article.

In practical terms, The Crow-AMSAA technique involves plotting, most commonly, cumulative failures vs cumulative time on a log-log scale resulting in straight line plots (2). The line slope value (or Beta value) indicates improving, deteriorating, or constant failure occurrences. Due to the straight-line nature of the plots, future failure forecasts can be estimated. In plain words, based on the current trend, when is the next failure expected to occur? This method handles mixed failure modes, so it is, therefore, suitable for the complex nature of the generating units. [Read more…]

I have often heard employees complain they have no data to initiate a proper Reliability Improvement Program. This is not always true. And to no fault of theirs. They just don’t know how to use what they already have in terms of records. If you are running an operation, you should at least have production records – i.e. how much you are producing on a daily basis. If you don’t have this, then maybe you should not be in business at all. This article looks at ways to initiate a Reliability Program using the Barringer Process Reliability (BPR) methodology. The greatest advantage of this methodology is that it only requires productions records as an input.  That is how many units of production the plant produces on a daily basis.  For example, the barrels of crude oil processed per day in a refinery. Or the hectoliters of beer brewed in a brewery daily.

Failure Patterns according to Moubray

In his book Reliability Centered Maintenance¹, John Moubray highlights 6 patterns of failure. However, one needs to be careful about how those patterns are interpreted and used. Or misused. These 6 failure patterns are as follows:

• A: Bathtub Pattern
• B: Age Related or Wear Out Pattern
• C: Fatigue Pattern
• D: Initial Break-in Pattern
• E: Random Pattern
• F: Infant Mortality Pattern

Definition of a PF Curve

A PF curve is a graphical tool used in the field of maintenance and reliability. It essentially illustrates a component’s health degradation over its lifetime. As well as a visual guide on when to conduct appropriate action to minimizes operational risks related to unplanned failures. It is essentially a planning tool.

Concurrent or simultaneous failures can happen with redundant or spared systems. This means that both spared equipment can fail at the same time leaving the operator with no production output. For example, we have two alternating pumps operating in a parallel configuration. Each one acts as a spare and at any one time can take over if the other one fails. This article is based on a question I was asked during a recent industry presentation. I thought the example was interesting and informative enough to share with the Maintenance and Reliability community.

Assets or components in an operation can be dependent on each other. The network of equipment contributing to the operation output is complex and intricate. A RAM model helps account for the complexities including dependencies.

The fundamental purpose of Reliability, Availability, and Maintainability (RAM) modeling is quantifying system performance, typically in a future interval of time. A system is a collection of items whose coordinated operation leads to the output, generally a production value. The collection of items includes subsystems, components, software, human operations, etc. For example, an automobile is a system. Its sub-components being the drivetrain, engine, gearbox, etc. In RAM models, it is crucial to account for relationships or dependencies between items. This helps determine the final output of the system. In various industries, RAM models have proven to be effective as cost avoidance or decision-making tools, as well as their ability to confirm or counter stated assumptions by internal stakeholders.