Process plants drift. Oil fields produce more water each year. Cooling loops accumulate biology. Feedstock specifications get relaxed when a cheaper supplier appears. Debottlenecking projects push throughput 15% beyond the original design.
Filtration is specified once, at the beginning, against the cleanest and most optimistic version of the process that will ever exist.
What actually changes
In our experience of returning to installations after five to ten years, four changes recur:
- Flow rate has increased, typically 10 to 25% above original design, through debottlenecking
- Solids loading has increased, often severalfold, through feedstock change or equipment wear upstream
- Fluid chemistry has shifted — higher chlorides, different pH, new additives that affect material compatibility
- Maintenance resource has decreased, making systems that needed weekly attention progressively unviable
Designing with that in mind
Size the vessel envelope generously even when the current duty does not need it. A housing with spare element positions costs a fraction more today and absorbs a throughput increase without a capital project. The same logic applies to nozzle sizes, which are almost impossible to change later.
Choose metallurgy against where the chemistry is heading, not where it is. The incremental cost of duplex over 316L at fabrication is far smaller than the cost of replacing a corroded vessel in year seven, plus the production it takes with it.
Instrument for trending from day one. A differential pressure transmitter feeding the DCS costs very little and, five years on, is the only record that explains why the system behaves differently than it did at commissioning.
And design for the maintenance team that will exist, not the one that exists now. If a system needs weekly manual intervention to work, it will eventually not work.