Your Reliability Engineering Professional Development Site
One of the first things to do when faced with a set of numbers is to plot them. A histogram is often the first choice, maybe a dot plot. Up your data plotting skills and let your data provide a bit more information by using a box plot. [Read more…]
A common practice I’ve seen in organizations is to deal with field failures when they occur. This may occur when the mistaken assumption that no failure will occur due to ‘such an excellent design.”
Ben Franklin may not have been thinking about future product failures, yet his quote:
By failing to prepare you are preparing to fail.
implies we need to prepare ourselves and our organization to deal with field failures. Having clear processes to deal with field failures is a best practice. [Read more…]
As reliability professionals, we are in the business of estimating or forecasting the reliability performance of our product, equipment, or system. While we use a range of tools to analytically make these estimates, sometimes we do not have sufficient data or information.
One method is to ask another person that has knowledge of the particular technology, use conditions, or whatever is hampering our work. If you ask two people you most like will get two different answers. If you ask 10, 10 different answers.
One way to work with a group of subject matter experts is to conduct a structured communication technique called the Delphi Method. [Read more…]
At an early concept meeting discussing the technical strategy for the new product, the engineering teams were at an impasse. The decision matrix balanced out with three distinct options. Product reliability differed slightly with each option yet presented risks just as the considerations of cost, complexity, feature set, and time to market.
The project manager, the leader of the development program, asked a few questions, asked for input from the director of engineering, and selected a path forward.
The team accepted the decision. The project went well. Yet, I’ve often wondered how did she know which option to select. I also learned to trust her judgment on difficult decisions. [Read more…]
What if you knew all the possible outcomes for your product’s reliability performance due to component variations, for example? What if you knew the future with enough certainty to make a difference?
Building on brainstorming, what-if analysis involved using models or prototypes that allow you to change something and see how it alters the output or performance. What if we change this support bracket from iron to aluminum? What if we swap out this 100 ohm resistor for a 200 ohm one?
As a curious engineer you could spend many, many hours conducting what-if based experiments, so there is a bit more to this idea then just a random walk of changes. [Read more…]
Over the past week, I received a couple of interesting questions. One concerned assuming a Weibull beta value for an accelerated life test plan. The second involved assuming expected life models for elements within a reliability block diagram.
In both cases, we faced incomplete data and uncertainties, yet felt the need to assume some values in order for the math to work out. We do make assumptions in order to solve problems. We also can make mistakes that lead to unwanted consequences. [Read more…]
One of the enjoyable parts of reliability engineering work is the consistent need to learn. We learn how new materials, designs, applications, and systems work, and fail. Sometimes we learn through proactive characterization studies, sometimes via unwanted field failures.
Failures will occur, it is what we learn from them that matters. The ability to gather and remember the lessons learned is a common and ongoing need for every organization. We are not very good at it, in general. [Read more…]
Graphs contain information and often tell a story. Our interpretation of the graphic can be aided or hindered by the design or style of the plot. Cleveland and McGill (1984) studied graphical perception and found the use of dot plots to aid viewers to understand the data’s message clearly.
The nature of a dot plot is like a bar chart, yet without the bars. Less ink, just a dot to indicate count or position along an axis permits conveying information simply. Due to its simplicity, it also permits adding additional information useful for comparisons or spotting trends, and more. [Read more…]
Uncertainty is another word for risk. Reliability uncertainty or risk is neither good nor bad, it just a bit unknown. Until we know the outcome, the eventual reliability performance, we will not know the impact.
So, how do we deal with reliability uncertainty? Will our product or system work as expected over time, or will it fail? Let’s examine a few of the common approaches in use and when and why the approach is effective. [Read more…]
You may have heard of the 80/20 rule. The idea is that 80% of the wealth is held by 20% of the population. As an Italian economist, Vilfredo Pareto made this observation that became generalized as the
Pareto Principle: 80% of outcomes are due to 20% of causes
For field returns, for example, we may surmise that 80% of the failures are due to 20% of the components, for example. This principle helps us to focus our work to reduce field failures by address the vital few causes that lead to the most, or most expensive, failures. [Read more…]
How is it that some people continue to get better at managing meetings, designing complex test plans, making presentations, or solving problems? How in general do people improve their performance over time at something? [Read more…]
Carefully considering the acceleration factor (AF) is essential when conducting an accelerated life test. Like warp drives shortening the distance, accelerated life tests (ALT) attempt to shorten time. Think of the warp factor and acceleration factor as being similar, well, sort of. [Read more…]
There is a type of error when conducting statistical testing that is to work very hard to correctly answer the wrong question. This error occurs during the formation of the experiment.
Despite creating a perfect null and alternative hypothesis, sometimes we are investigating the wrong question. [Read more…]
Some time ago when talking with someone I just met, the conversation turned to what we did for a living. I mentioned being a reliability engineer, and his response: “Oh, yes, we do reliability”. Curious, as I’m not sure that I ‘do reliability’, we then talked about what he meant.
The conversation revealed that they had a list of tasks that they accomplished for each product under development. They did tests and reviews of the results. A lot of testing. They did FMEA and HALT. He believed the engineers did derating or stress/strength calculation. He didn’t know about process stability with vendors or internal manufacturing lines.
They did stuff, which meant they did reliability.
A common question when setting up a hypothesis test is concerning sample size. An example, might be: How many samples do we need to measure to determine the new process is better than the old one on average?
While this seems like a simple question, we need a bit of information before we can do the calculations. I’ve also found that the initial calculation is nearly always initiated a conversation concerning the balance of sample risks, the ability to detect a change of a certain size and the constraints concerning the number of samples. [Read more…]
