Model predictive control
On a crude unit, one controller coordinates furnace outlet temperature, side-draw cuts, and overflash against tower pressure limits.
Industries / Refining
Refinery advanced process control (APC) coordinates crude, vacuum, coker, reformer, FCC, and hydrotreating units so each runs closer to its limits at stable quality. Atlas Prediction Control restores and sustains APC performance inside the control systems you already run, from the first assessment to ongoing monitoring.
Request a Plant Performance Scan2 to 4 hours, remote, no cost.
Process Environments
Refinery Applications
Refinery units have interacting variables, changing feed conditions, and operating constraints. APC coordinates these variables to sustain performance.
| Unit | Interacting variables and limits | APC objective |
|---|---|---|
| Crude distillation | Furnace outlet temperature, side-draw cuts, overflash, tower pressure, crude switches | Maximize throughput and cut quality through feed changes |
| Vacuum distillation | Furnace temperature, vacuum level, wash oil rate, product draws | Maximize gas oil recovery within heater and tower limits |
| Delayed coker | Furnace coil temperature, drum cycle timing, fractionator draws, compressor limits | Push throughput and yield against heater and compressor constraints |
| Catalytic reformer | Reactor inlet temperatures, hydrogen to hydrocarbon ratio, octane | Hold octane at target with minimum giveaway, using quality estimation |
| FCC | Reactor temperature, regenerator limits, wet gas compressor, main fractionator | Maximize conversion at the active constraint |
| Hydrotreater and hydrocracker | Reactor temperatures, hydrogen supply, product sulfur | Hold sulfur spec while minimizing hydrogen and fuel use |
Operational Indicators
These operating conditions can indicate an opportunity to improve control performance.
Engineering Capabilities
Engineering support covers the control models, underlying loops, operating constraints, and ongoing monitoring needed to sustain APC performance.
On a crude unit, one controller coordinates furnace outlet temperature, side-draw cuts, and overflash against tower pressure limits.
A soft sensor infers reformate octane or product sulfur between lab samples, so control acts on a reading available every minute.
Feedforward from crude properties and decoupling on side-draw loops steady the base layer during a crude change.
Calculated heater duty, tower flooding margin, or compressor surge margin give the controller a limit it can honor.
CLPM ranks loops by time in manual and valve health, and Atlas Advisor™, where deployed, ties APC and historian signals to root cause.
Remote Resident™ rechecks models after catalyst changes, turnarounds, and crude slate shifts.
Working With Atlas
Assessment starts with the units where constraints bind most often and where operators hold the largest margins.
A Plant Performance Scan (2 to 4 hours, remote, no cost) scores the unit. A Plant Audit (2 to 4 days onsite) adds a financial estimate. You provide engineer time and recent operating data. Atlas provides the score, findings, and loss estimate.
Atlas designs and runs step tests in live operation, then builds or recalibrates the models. You provide operations access and test windows. Atlas provides test design, analysis, and models.
Atlas commissions the controller or soft sensor with your operators and validates it in live operation. You provide operator and engineering time and DCS integration support. Atlas provides tuning, validation, and handover.
Control Loop Performance Monitoring (CLPM) tracks loop and controller health against live data. A 90-day pilot on an agreed set of loops carries no fee.
Remote Resident™ keeps models, constraints, and loops aligned as conditions change.
Technical Questions
Start With an Assessment
Request a Plant Performance Scan. You receive a score out of 100, key findings, and an indicative annual loss figure.
Request a Plant Performance Scan2 to 4 hours, remote, no cost.