Five systems, five identifiers, one machine. The maintenance system knows it by an equipment number and the financial system knows it by an asset tag. The manufacturer knows it by serial number, the condition monitoring platform by an instrument tag, and the regulator by a unit and station designation.
All five refer to the same compressor. Nothing in the estate says so.
What is asset configuration drift?
Asset configuration drift is the accumulated divergence between what an asset was specified to be and what it actually is after years in service. It is normal, it is not a failure, and it is why two units commissioned on the same day need different work five years later.
What causes an asset to diverge from its specification?
Five things, all of them ordinary. A field change is applied. A part is superseded and the replacement is not quite the same, which nobody flags because the work order closed cleanly. A repair is done correctly but differently from the manual. Work is deferred with an approval and then deferred again, and an exception granted for good operational reasons is never revisited.
None of that is bad practice. It is what operating equipment looks like. The problem is that what is written down does not keep up, so work gets planned against a specification that stopped being accurate years ago.
Why are fleet averages a poor guide to any single unit?
Because the averages describe a population that no longer exists.
Two identical trucks leave the depot on the same day and diverge immediately, because wear is not a function of distance alone. It is a function of the loads carried, the terrain covered, and the duty cycle demanded. A bearing that fails early on one unit and late on another usually has its explanation sitting in the trip and cargo history, and that explanation only exists if the trips and the cargo sit in the same structure as the parts.
The same holds for a compressor station, a machining centre, or an airframe. Same specification at commissioning, different history since, different work required now.
Why does one asset appear five times across your systems?
Because each system was built to answer its own questions and assigned its own identifier to do it. None of them was designed to agree with the others, and reconciling them by hand is a project rather than a task.
The consequence is that the most basic question about an asset is expensive to answer. What is the actual configuration of this unit today, when did it last run, and what condition is it in? What work is outstanding, is the unit compliant, and can anyone show why? A technician, an engineer, and an accountant will each answer differently, and each will be partly right.
What does resolution produce on the first pass?
More than most organizations expect, and usually before any prediction is made.
Connecting the maintenance, financial, engineering, and condition sources for a defined set of assets, then resolving identity across them, produces a consistent and scored inventory. Duplicate assets surface, along with assets still carried on the books that were disposed of years ago and assets in service that were never recorded at all. So do units whose configuration on paper does not match the parts actually fitted to them. All of it in the same pass.
Scoring then makes the inventory honest. Applied per attribute rather than per asset, it can state that a given compressor is well understood mechanically and thinly evidenced for compliance. That is a specific piece of work someone can own. Scored as one number for the whole asset, it would have been either condemned or flattered, and the useful detail would have been lost either way.
Poor data quality costs organizations an average of $12.9M annually according to Gartner. In an asset-intensive operation, a meaningful share of that sits in the gap between the asset on paper and the asset in the field.
One twin per asset, attribute by attribute
The PolyPhaze white paper Best decision. Best action. covers the twelve agents, the five phases, and how condition becomes a prediction that can be checked. Download the full asset twin ebook for the detail behind it.
Frequently asked questions
What is as-maintained configuration?
As-maintained configuration is the actual current build of an asset, including every field change, part substitution, repair, and granted exception applied since commissioning. It differs from the as-designed specification, and it is the record maintenance and compliance work should be planned against.
Why does the same asset have different IDs in different systems?
Each system assigns identifiers for its own purpose: an equipment number in the CMMS, an asset tag in the financial system, a serial number from the manufacturer, an instrument tag in condition monitoring, and a regulatory designation. Nothing in the estate states that they describe one machine until entity resolution establishes it.
Are fleet averages useful for maintenance planning?
They are useful for budgeting and capacity planning and poor for scheduling work on an individual unit. Two assets with the same specification diverge based on load, terrain, and duty cycle, so the average describes a population rather than any machine in it.