For four years now, a group of dedicated University of Toronto engineers, programmers, mathematicians and statisticians have been hard at work on your behalf. Their efforts to bring a major advance in decision-making to maintenance professionals is now a reality.

EXAKT: The CBM Optimizer, an uncommonly fine software package for optimizing your condition-based maintenance decisions is now available to engineers who need to renew, repair or replace equipment — and who need to be sure they’re doing it at the lowest possible cost.

But before revealing all the exciting details about EXAKT,.let’s take a moment to compare good maintenance decisions with not-so-good maintenance decisions. —

Good Maintenance Decisions.
A good maintenance decision is when we replace equipment not too soon, and not too late — and at the same time, make sure our long term costs are the lowest possible. To do this, it’s necessary to trade off the cost of too many preventive maintenance replacements against the cost of equipment failing in service.

Sometimes the Age-Based technique is replaced by a Condition-Based system. For example, oil analysis. When the analysis shows certain levels of wear-metals in the equipment, then it’s a signal to take action. Other condition data may result from tracking changes in...

• vibration
• temperature
• fuel consumption
• pressure drops
• codified visual assessments
• current-draw...

...to name a few.

If you’re using one or more Age-Based or Condition-Based approaches you’re certainly on the right track.

Except for a big "however".

The problem is... how can we deal with and utilize this wealth of information? It is, indeed, a "wealth" of information but how can you harness it to extract the maximum use from it? It is, after all, information that if properly handled can lead to replacement decisions that would minimize one’s long-term costs.

That question brings us to...

Better Maintenance Decisions.
Clearly then, the answer is to find some way to organize and analyze data-streams from both the age-based and the condition-based systems — a joint analysis of the inputs would certainly yield a more realistic and accurate lifetime profile

That software of course, is EXAKT.

What It Takes
To Use
EXAKT

One or two things are required before EXAKT can be put to work in your organization. One of them is data...and the other is a degree of effort to install and learn the routines associated with using a sophisticated software package like this one.

DATA
Clearly, to use EXAKT, you’ve got to have data. To unleash the full power of EXAKT on a given type of equipment, you need to have records on the failures and preventive replacements on these units over time — lifetime data — and for this same unit, its history in terms of condition monitoring data.

An easy way to see where you stand in terms of data is to use the flow chart that follows. It’s a good guide to the answer.

MODEL-BUILDING
EXAKT requires that a mathematical model be built for each type of equipment you wish to analyze, that is, for every group of equipment units that have similar characteristics — make & model, size, power, and so on.

You need not start from scratch. EXAKT will build the model for you, with your help, and using your data.

One way to go about this task is to assign the model-building exercise to someone in-house. That person would be comfortable with statistics — an engineer, a statistician, or an O.R. professional. The in-house model-builder would be trained by a recommended consulting organization to handle model-building in your organization. He or she will gain a complete understanding of both the conceptual material as well as the steps required for constructing the models.

Model-building happens only once, at the beginning of your use of the software, plus occasional alterations when and if conditions change.

The rest of the time new condition-monitoring results are simply fed into the program — (the process can be automated) — and replacement decisions are produced for each unit. The decisions are calculated according to a strategy that makes sure you don’t have too many replacements, or too few...and that results in the "lowest-cost" solution.

Another model-building approach is to have us do it for you. Once we build your models, the chances are that in-house personnel can take it from there, for modifications and up-dates.

THE EXAKT DEMO
The enclosed demo version of the software is an integral part of this information package. We urge you to install it and go through the routine that accompanies it.

The demo exercise will give you an excellent chance to assess the product and to learn what’s involved with it. You might also come up with some questions you’d like to pass along to us.

Here’s What
EXAKT
Will Bring To Your Decision-Making

EXAKT brings rewards far greater than the effort you will have put into learning it or setting it up. The program’s output is, in a word, prodigious.

To appreciate everything EXAKT can do, you’ll have to go through the demo program, but here are just a few of the many features of this valuable software package.

OPTIMAL STRATEGY
For the everyday maintenance practitioner who just wants to get a simple answer to a simple question, i.e., "Should we keep on running, or should we replace right now?", he or she will turn to the EXAKT Optimal Strategy chart —

This chart is the "nub" of EXAKT’s output. It enables the engineer to immediately find the answer — to replace immediately, to wait for the next inspection, or to monitor closely because of incipient failure — on a clear unambiguous graph. EXAKT assumes that regular inspections are being carried out, so the position of the dot on the chart will provide one of three possible actions related to the inspection interval policy.

One of them — Replace Before Next Inspection — is a warning signal which says although you’re "getting close" to a possible failure event — the risk of failure combined with the benefits of running a while longer means that you may be able to delay replacement. It’s an invaluable help in planning because it allows you to forecast the maintenance workload and the down-time eventualities in the near-term.

The optimal strategy chart. A disarmingly simple diagram that, in fact, delivers the most profound and extraordinarily useful output of the entire EXAKT software package!

SENSITIVITY
Another graph, Sensitivity of Optimal Policy reveals, for your particular operation, just how important accurate cost data is to the ultimate replace-or-don’t-replace decisions.

This chart allows you to get comfortable with two things...one is how the cost-data precision affects the hazard level (the level at which a replacement is recommended), and the other is whether you should plan less, or more, preventive replacements.

In the chart we’ve illustrated the cost ratios of failure replacements to preventive replacements run from near-zero to a high of 16. The curve shows that at a low cost ratio, say in the 2-to-4 range, where the costs of failure replacement are higher — but not that much higher than a preventive replacement, one would plan less preventive replacements. Where the cost ratio is high, one would plan to replace preventively more often.

Sounds useful? It surely is!

PROBABILITY
Still another fine feature of EXAKT is a table — its long name is The Markov Chain Model Transition Probability Matrix.

This table shows the probabilities of going from one state (for example...light contamination, medium contamination, heavy contamination, very heavy contamination) to another, between inspections. In the words of one of our principal researchers... } The table provides a formal quantitative estimate of the probability that the equipment will be found in a particular state at the next inspection, given its state today~ .

Referring to the table, we can see that the co-variate under analysis is iron, and that four levels of readings have been established. (That establishment would have occurred in the model-building phase we covered earlier).

The four levels indicate the severity of the condition — such as light contamination to very heavy contamination If the second level of severity where readings are 11.5333 to 23.0667 indicate, say, a "medium" state, the table would tell you that there is a 76% (0.756024) probability of the measurement being in the same range at the next inspection, a 24.6% of it having worsened to the next higher level of severity (23.667-to-34.6), and a 100% chance of it having reached the worst range (above 34.6) before the next inspection.

COVARIATES
The underlying mathematics of EXAKT includes a Proportional Hazards Model .

Although space limits the possibility of our fully explaining the approach here, the model is in fact the backbone of the program in that it deals with the co-variates being monitored. Co-variates associated with wear-metals, for example will include measurements of iron, chrome, copper and so on, in the equipment’s oil.

Examination of the co-variates’ values can be an important part of the output that reliability engineers will want to see. Suffice it to say that EXAKT provides a wealth of data in tabular and graphical form to allow just about any imaginable measure. Parameter-estimates, Standard Errors, Wald (a goodness-of-fit test), p-values, 95% confidence intervals, expected values — they’re all there!

AGE-BASED WEIBULL
ANALYSIS

One of the main components of EXAKT is the Weibull module.

If you don’t have condition data, this module can be effectively used to deliver vital decision-making results on its own.

This display shows several things. The chart reveals a medium-gray (colored green in the real product) that illustrates the costs that are associated with preventive replacements. The "dot" pin-points the recommended, and lowest-cost, point for replacement...every 6297 hours. The darkest section (red) shows the not-preventively-replaced (i.e., failure-replaced) components’ contribution to the total cost of replacement.

The table indicates that if you abide by the recommended replacement time, about 70% of your replacement costs will be preventive-replacement related, and it shows that 92% of your replacements will be done on a preventive basis, and 8% on a run-to-failure basis.

There’s even more to this chart, but these few words answer the question — how useful can this analysis be?...and the answer is VERY!.

EXAKT
Technical Specifications.

Connecting with your data base or CMMS — Computerized Maintenance Management System should be "no problem". Your data base program almost certainly has an ODBC driver, and that’s the key to compatibility with EXAKT.

EXAKT also has its own database — of the "Access" type — which is used for temporary storage of in-process calculations.

EXAKT runs in Windows ‘95 or NT, and requires a minimum of 16mb of RAM. It will occupy about 15mb of hard-disk space.

Is EXAKT For You?
Here’s a Quick Way To Find Out

Mathematically
Inclined?

The PHM (Proportional Hazards Model) used in the program is the Weibull failure model with time-dependant covariates to incorporate the condition information. Its hazard function h(t) has the general form of this formula —

A Markov chain model is used to describe the behaviour of the condition over time. This, together with the PHM, forms a complete statistical model which is used together with cost information to calculate the corresponding optimum component replacement strategy. From this, replacement decisions are obtained for individual components that depend on their ages and the most current information.