Refinery Utility Air System Integration

REFINERY UTILITY AIR SYSTEM INTEGRATION

Service Air, Instrument Air, Hazardous Process Interfaces, Continuity and Plant-Wide Utility Segregation

Worked figures are illustrative engineering examples unless identified as measured project data.

1. Engineering Answer — Treat Refinery Air as Multiple Utilities, Not One Header

A refinery compressed-air system should not be treated as a single plant-air network serving every user at one pressure and one quality level. Service air, instrument air, maintenance air, workshop users, analyzer or packaged-equipment utilities and temporary shutdown demand can have different consequences when pressure falls or contamination reaches the point of use. The integration task is therefore to preserve clear utility boundaries while allowing practical source sharing where it is engineered, monitored and recoverable.

The most important refinery design principle is functional segregation. Instrument air supports control valves, actuators and other control functions whose loss can affect unit stability or safe shutdown. Service air supports general-purpose users such as tools, cleaning and maintenance. These two demand classes should not be allowed to compete without deliberate pressure hierarchy, priority logic and isolation. A large discretionary service-air event must not quietly consume the pressure margin required by control users.

A refinery system also crosses multiple process units, pipe racks, utility corridors, workshops and remote packaged systems. Long-distance distribution, isolation valves, treatment equipment, local regulators and temporary shutdown connections all influence delivered pressure. Capacity at the compressor house therefore has to be reconciled with what reaches the user during actual simultaneous operation, not only with compressor nameplate flow.

2. Scope and Engineering Boundaries

Its job is to make the compressed-air interfaces explicit and verifiable.

3. Required Asset — Refinery Utility Segregation Requirements

It establishes what each user class receives, how it is isolated, what pressure/quality evidence is required and what happens when system capacity becomes constrained.

4. Build the Refinery Demand Model by Unit and Operating State

Refinery air demand changes with operating state. Normal steady production, unit start-up, catalyst handling, maintenance, regeneration activity, flare or utility upsets, turnaround work and restart can all change the air balance. A credible system model therefore combines user inventory with operating scenarios rather than assuming a single fixed diversity factor.

5. Demand Assessment — Connected Load Is Not Coincident Demand

Each user should be described by a documented flow basis, required pressure, duty cycle, operating state, criticality and whether the flow is continuous, cyclic or event-driven. Summing every connected device at full consumption usually overstates normal demand, while ignoring shutdown and restart scenarios can understate critical peaks.

A practical engineering model can express scenario demand as:

Qscenario = Σ (Qi × Ni × Fi × Si) + Qleak + Qtemporary

where Qi is flow per user, Ni is quantity, Fi is duty/diversity factor, Si is a scenario factor, Qleak is the measured or justified leakage allowance and Qtemporary is any planned maintenance or temporary demand. The factors should come from operating knowledge, equipment data and measurement rather than generic percentages.

6. Service Air vs Instrument Air — Pressure Hierarchy

The refinery should maintain enough pressure margin that lower-priority service demand cannot force instrument-air users below their verified requirement. This can be achieved through separate headers, pressure-control valves, priority valves, staged source control, or combinations of these approaches. The correct arrangement depends on the existing plant architecture and operating philosophy.

A useful principle is that higher upstream pressure is not a substitute for segregation. If the network repeatedly requires a large compressor-room pressure increase to keep remote instrument users stable during service-air peaks, the system likely has a distribution, restriction, control or demand-management problem that should be resolved at its source.

7. Pressure-Budget Engineering

The compressor discharge setpoint should be the result of a documented pressure budget, not a habit. Start with the lowest acceptable pressure at the critical user and add only the verified losses between that point and the compressor header.

If the plant is operating materially above this calculated requirement, investigate the reason before accepting the higher pressure as permanent. Artificial demand, leaks and energy use often increase as system pressure rises.

8. Refinery Distribution Architecture

Large refineries often require a combination of main utility headers, unit sub-headers, local rings and protected branches. The architecture should support isolation without creating long single-path restrictions. Cross-connections should be documented because a normally closed valve that is opened during an upset can change flow direction, treatment boundaries and pressure behavior.

9. Air Treatment and Quality Zoning

Refinery compressed-air quality should be defined by the most demanding protected user on each branch, not by a single plant-wide assumption. Instrument air generally requires stronger moisture and contamination control than ordinary service air. Local specialist packages can require additional treatment or tighter conditions. The integration plan should therefore identify treatment boundaries, sampling points and cross-connection risks.

10. Hazardous Process-Area Interfaces

Compressed-air integration in refinery process areas must respect the project’s area classification, ignition-control requirements, equipment approvals, hot-work controls and operating procedures.

The utility engineer should document where compressors may be located, how intake air is protected from process releases, where exhaust or heat can be rejected, where condensate is routed, and which temporary connection methods are permitted. These are interface decisions, not generic compressor rules.

11. Compressor-House and Environmental Interfaces

12. Redundancy, Continuity and Recovery

Refinery instrument air should be treated as a continuity utility. The required redundancy philosophy depends on process criticality, compressor train arrangement, treatment redundancy, header topology and the time available before critical users become unstable. Redundancy is not achieved merely by installing an extra compressor if a single dryer, filter bank, power supply or isolation valve can still interrupt the whole system.

13. Storage — Use It for Defined Transients, Not Permanent Deficits

Receiver capacity can support short transients or provide a controlled response window, but it cannot compensate indefinitely for insufficient compressor capacity or undersized distribution. Storage calculations should use absolute pressure and a defined allowable pressure decay between initial and minimum usable pressure.

For a simplified planning check, usable free-air inventory is proportional to vessel volume multiplied by the absolute pressure difference divided by the chosen reference pressure. The final design should use the project’s thermodynamic basis, actual temperature assumptions and vessel requirements.

14. Turnaround and Temporary-Air Integration

Turnarounds can create some of the largest temporary service-air demands in a refinery. Portable tools, cleaning, testing and contractor equipment may be added faster than the permanent system was designed to accommodate. Temporary demand should therefore have a separate intake process: expected flow, pressure, duration, location, hose/manifold size, quality requirement, connection point and responsibility.

15. Monitoring Architecture

Trend data should be reviewed by operating state. A system can look healthy on an average trend while repeatedly dipping below the required pressure during short events.

16. Failure Modes and Troubleshooting

17. Worked Refinery Integration Scenario

Consider a refinery where two process units share an instrument-air header and the central utility system also feeds service-air users. Normal instrument-air demand is 4.8 m³/min, normal service-air demand is 5.5 m³/min and miscellaneous measured leakage is 1.2 m³/min. A turnaround adds a temporary maintenance demand estimated at 5.0 m³/min during peak tool use.

If the temporary demand is simply added to the existing service-air header, the total operating requirement can rise from 11.5 m³/min to approximately 16.5 m³/min before contingency margin. If the permanent plant can deliver only 14.5 m³/min at the required header pressure, the system does not have sufficient sustained capacity for that turnaround state. Raising the pressure is not a valid cure.

A better response is to segregate the turnaround demand using a dedicated temporary compressor or temporary trunk sized for the maintenance load, leaving the permanent plant to protect instrument and routine service demand. The utility map is then updated with the approved temporary source, isolation points, hose/manifold capacity, pressure verification points and removal date.

Decision lesson — refinery compressed-air integration is a priority and boundary-management problem as much as a compressor-capacity problem.

18. Energy and Capacity Optimisation

Eliminate inappropriate service-air blow-off before adding compressor capacity.

Repair leakage by zone so losses can be assigned to the responsible unit or header.

Reduce system pressure only after confirming the minimum pressure at critical instrument-air users.

Keep treatment filters and dryers within acceptable pressure-drop ranges.

Sequence compressors so base-load machines operate efficiently and trim capacity follows measured demand.

Use temporary sources for short campaigns rather than permanently oversizing the base plant where justified.

Investigate persistent high header pressure as an integration or distribution problem, not simply an operating preference.

19. Commissioning and Acceptance

20. Change Control — What Must Trigger Reassessment