Your failure data already knows why things break. Weibull is how you read it.
Two days on the method, not on a product. You leave able to tell whether an asset is failing early, failing randomly, or wearing out, what each one changes about the maintenance plan you run today, and how to do the analysis in whatever you already have on your desk.
Four things you can do on the Monday
This is a working course. Every method is applied to failure data on the day, in the software you will use afterwards.
Read a β and act on it
Fit a distribution to a real failure set and say which failure pattern you are in. Then defend the maintenance decision that follows from it.
Handle the data you actually have
Suspensions, censoring and incomplete records are the normal case in a CMMS. Learn the methods that survive messy history instead of assuming clean data.
Set a replacement interval that pays
Work out where a time-based interval earns its cost and where it destroys remaining life. Stop replacing components that were not going to fail.
Know when the fit is wrong
Goodness of fit and model selection, so you can tell a defensible result from a curve that happens to pass through the points.
Two days, ten sessions
Day one builds the method. Day two puts it under load with censored data, growth testing and repairable systems.
- Introduction to Weibull analysis. Where it fits in a reliability program and what it will and will not tell you.
- Probability and statistics concepts. The working minimum, no more.
- The Weibull distribution and its properties. Shape, scale, and what β means on a real asset.
- Estimating the parameters. Graphical and numerical methods, and when they disagree.
- Goodness of fit and model selection. Proving the fit, and the limits of the test.
- Censoring and incomplete data. Maximum likelihood and regression on records with gaps.
- Reliability growth and accelerated life testing. Estimating parameters when you cannot wait for failures.
- Repairable systems. Why a repaired asset is not a new one, and what that breaks.
- Your data, your questions. Bring a failure set. Work it with the trainer.
- Recap and close. Where this feeds next, and what to read.
Bring your own failure history
The strongest session on this course is the one where you fit a distribution to your own equipment. Export a failure set from your CMMS before you come. If you cannot get data out, tell us at booking and we will supply a set from equipment close to yours.
Who this is for
Reliability and maintenance engineers
Anyone who sets or defends a replacement interval, a PM frequency, or a spares holding.
Asset and maintenance managers
If you approve maintenance strategy, this is the analysis your engineers should be showing you.
Anyone heading into RBD or plan optimisation
Weibull is the entry point. The failure models you build here are the inputs to everything downstream.
What you need before you start
A working understanding of basic probability and statistics. No prior Weibull experience is assumed. If you have used Weibull++ before and want the advanced material, say so at booking and the day two content is weighted accordingly.
Weibull is the first step in the analysis chain
The failure models you build on this course are the required input to the two courses that follow. Most teams take the three together.
Weibull Analysis
2 days. Build defensible failure models from real history.
RBD Modelling
2 days. Apply those failure models to a whole system and find what actually limits it.
Maintenance Plan Optimisation
3 days. Simulate, find the bad actors, and rebuild the plan around them.
Learn the analysis. Then run it in whatever you already have.
Most Weibull courses are product training with the method bolted on. When the licence lapses or the one person who had it moves on, the capability goes with it. This course teaches the method first, then shows you two ways to run it. You pick the one that fits your site.
Excel
Median rank regression, plotting and interval setting in a spreadsheet you already own and already have approval to use. Enough to start Monday, on the assets you were arguing about last week.
Your existing suite
Already running a reliability platform? The method transfers directly. We cover where the commercial tools earn their cost, and where they do not.
Learn the method. Then put AI to work on it.
Understand the analysis first, leave with the templates, then use AI to do in hours what used to take days. The course closes with a working session on exactly that, inside your site’s IT and confidentiality rules.
What AI accelerates
Pulling failure events out of messy CMMS exports and free-text work orders, flagging likely suspensions, and drafting the data preparation that eats most of an analysis day.
What stays with the engineer
β is a judgment about physical failure behaviour. AI drafts and sorts. The engineer decides, and the decision stays defensible in front of the budget holder.
What you leave with
A prompting guide written for reliability work, a prompt library for life data analysis, and a setup checklist for using AI within site IT and data rules.
We run our own operation this way. The AI workflows taught here are the ones we use to deliver client work, from OEM manual extraction to failure data preparation.
Three ways to take it. One seat or the whole crew.
Every course runs in all three formats. On site is priced as an engagement rather than per head, so crew size is your call. Public and online cohorts confirm once a small minimum books, and every booking is protected: if a date does not confirm, you choose the next cohort, the online session, or a full refund.
Public classroom
Runs in our training room in Maroochydore on the Sunshine Coast. Book one seat or several and work alongside engineers from other operations. Dates confirm once the cohort reaches its minimum.
Live online
Same instructor, same exercises, same data, delivered live rather than recorded. Confirms on a smaller cohort than the classroom, and it is the route in from anywhere outside the Sunshine Coast.
On site at your operation
We come to you and the whole course runs on your equipment and your failure history. Priced as an engagement, not per head.
Taught by the engineers who do this work on sites
Not career trainers. The person at the front of the room fits distributions to real failure data for operating mines and plants, and brings those data sets into the course.
Delivered by the engineering bench, not a training department
Courses are taught by HolisticAM reliability engineers who deliver this same work for clients between courses. Whoever stands at the front ran this analysis on an operating site recently, so the examples are current and the answers are practical.
Proven in delivery
Training is a delivery arm of the consultancy. The most recent full program trained 26 engineers and maintainers at a major Pacific gold operation, delivered live online, with certificates of completion issued.
It gets used on the Monday
Within days of that delivery, attendees were sending their own analyses of live plant equipment back to the trainer for review. That is the outcome every course here is built for.
The session was informative, well-structured and incredibly helpful. The knowledge gained will certainly enhance my RCA facilitation skills at work.
This course was terrific and beneficial. It opened my mind to expand and think differently. I enjoyed every moment.
Across 27 external participant surveys from our Apollo RCA courses, self-rated knowledge moved from 3 to 8 out of 10.
Before you book
Do I need to know statistics?
A working understanding of basic probability and statistics. If you can read a histogram and you know what a mean and a standard deviation are, you will keep up. Day one rebuilds the working minimum before it goes near a distribution fit.
Do we need to buy software?
No. The course teaches the method, then shows you two ways to run it: in Excel, or in a commercial suite if your site already has one. Nothing here requires a licence, and nothing you learn stops working when a licence lapses.
Can AI just do the analysis for us?
It can fit a curve today, and the result will look plausible. Plausible is the danger: without the method you cannot tell a sound fit from a fluent one, and you cannot defend either to the person signing the maintenance budget. The course teaches the method precisely so your team can put AI to work on the slow parts and audit everything it produces. That is why the AI session closes the course instead of replacing it.
Can you run this at our site?
Yes, and on site is the strongest version of this course because the team works its own failure data and the discussion stays on your equipment. On site is priced as an engagement rather than per head, so a team of three is as welcome as a team of thirty.
What if our CMMS data is a mess?
That is the normal case and day two is built for it. Censoring, suspensions and incomplete records are covered as the standard condition rather than an edge case. Bring the data as it is.
Can I do it online?
Yes. Every course runs live online as well as in the classroom and on site. Online is delivered live with an instructor, not recorded, and you work the same exercises on the same data. If you are one engineer rather than a team, this is usually the fastest way in.
Where do public courses run?
In our training room in Maroochydore on the Sunshine Coast. If travel does not make sense from where you are, the live online cohort is the same course with the same instructor, and on site delivery brings it to your operation.
Should we do RBD Modelling as well?
If the goal is a rebuilt maintenance plan rather than better analysis on its own, yes. Weibull builds the failure models, RBD applies them to the system, and Maintenance Plan Optimisation turns that into the plan. Ask about running the three as a block.
Stop replacing components that were not going to fail
Two days, your own failure data, and a method you can defend to the person who signs off the maintenance budget.