How is automation applied to refinery operations?

At-a-Glance: Refinery automation integrates DCS/PLC/SIS, analyzers, APC/MPC, and real-time optimization to maximize throughput, cut energy and giveaway, and protect assets. The core is constraint-based control that keeps units on-spec and on-limit with high uptime and low emissions.

I. Objective Definition and Key KPIs

• I.1 Objective
  • 1.1 Operate each unit at its active constraints with stability and safety, converting planning targets into stable, on-spec production at minimum OPEX and emissions.
  • 1.2 Close the loop between planning (LP), real-time optimization (RTO), advanced process control (APC/MPC), and basic regulatory controls (PID) to drive sustained benefit.
• I.2 KPIs (typical targets are estimated)
  • 2.1 Throughput: barrels per calendar day (bpcd) by unit; refinery utilization > 92–96%.
  • 2.2 On-stream factor/uptime: > 98% for critical units; SIS demand rate minimized.
  • 2.3 Energy: Specific Energy Consumption (SEC) 2.5–4.5 GJ/ton crude; furnace efficiency > 88–92%; steam balance losses < 2%.
  • 2.4 Quality: Product giveaway < 0.10–0.25 RON (gasoline), sulfur giveaway < 5–10 ppm, ASTM D86 cutpoint error < 2–4 °C.
  • 2.5 Emissions: CO2 intensity < 25–35 kg CO2e/bbl; flaring < 0.05–0.15% of fuel gas; stack O2 1.5–3.0% with CO < 50–150 ppm.
  • 2.6 Control performance: % loops in auto > 85–90%; MPC service factor > 85%; analyzer uptime > 95%; IAE trending down week-over-week.
  • 2.7 Reliability: Compressor surge margin > 10–15%; valve stiction index < 2%; instrument bad-actor count trending down.
  • 2.8 Economics: APC benefit $/day vs. baseline; RTO profit delta > 0.5–1.5 $/bbl; hydrogen network cost $/MSCF minimized.

II. Critical Parameters and Target Ranges

II.A Key Control Formulas Used in Refinery Automation

• 2.A.1 PID law (velocity form):

  \( u(t) = u_{0} + K_c \left[ e(t) + \frac{1}{T_i}\int_0^t e(\tau)\,d\tau + T_d \frac{de(t)}{dt} \right] \)

• 2.A.2 Integral of Absolute Error (IAE) to monitor loop performance:

  \( \mathrm{IAE} = \int_{t_0}^{t_1} |e(t)|\,dt \)

• 2.A.3 Valve sizing relation (liquid, non-choked):

  \( Q = C_v \sqrt{\dfrac{\Delta P}{G_f}} \)

• 2.A.4 Energy balance for heaters/exchangers:

  \( Q = \dot{m}\, C_p\, \Delta T \)

• 2.A.5 Distillation cutpoint inference (simplified):

  \( T_{\text{cut}} = a_0 + \sum_i a_i\, T_i + \sum_j b_j\, \Delta P_j \)

• 2.A.6 MPC objective (quadratic program, simplified):

  Minimize \( J = \sum (y - y^{\ast})^\top Q (y - y^{\ast}) + \Delta u^\top R \Delta u \) subject to \( y = G u \), and constraints \( y_{\min} \le y \le y_{\max} \), \( u_{\min} \le u \le u_{\max} \)

• 2.A.7 Compressor surge margin:

  \( \mathrm{SM} = \dfrac{\dot{m}_{\text{oper}} - \dot{m}_{\text{surge}}}{\dot{m}_{\text{oper}}} \times 100\% \)

III. Step-by-Step Procedure / Workflow / Checklist

III.1 Architecture and Strategy

• 1.1 Define automation layers and roles:
  • 1.1.1 Field layer: smart instruments, control valves (positioners with diagnostics), drives (VSD), analyzers.
  • 1.1.2 Control layer: PLC for package equipment; DCS for process PIDs, sequences, interlocks; SIS for safety.
  • 1.1.3 Optimization layer: APC/MPC per unit; RTO for site-wide economics; scheduling interface; historian/MES.
  • 1.1.4 Visibility layer: KPIs dashboards, control performance monitoring (CPM), alarm management, digital shift log.
• 1.2 Establish control philosophy documents:
  • 1.2.1 Control narratives, cause-and-effect, shutdown keys, IPLs, and constraint definitions per unit.
  • 1.2.2 Alarm philosophy: priority, shelving rules, rationalization, alarm rates (< 6 alarms/10 min steady-state).
• 1.3 Cybersecurity zoning and conduits with least privilege; patch and backup regime aligned to SIS/DCS policies.

III.2 Instrumentation and I/O Rationalization

• 2.1 Critical measurement selection:
  • 2.1.1 Redundant transmitters for high-SIL services; add valve position feedback and travel sensors on critical valves.
  • 2.1.2 Online analyzers for sulfur, RON/RVP, H2 purity, and tail gas; soft sensors for properties with slow analyzers.
• 2.2 Signal engineering:
  • 2.2.1 Sampling rates: fast loops 100–500 ms (PLC/turbomachinery), general PIDs 1–3 s, slow inferentials 10–60 s.
  • 2.2.2 Filtering: use first-order filters where process noise dominates; avoid over-filtering near constraints.
• 2.3 Valve hardware quality: high-resolution positioners, low deadband; specify equal-percentage trims for temperature/flow control.

III.3 Regulatory Control and Loop Tuning

• 3.1 Map P&IDs to control loops; implement cascade, feedforward, and ratio controls for disturbance rejection.
• 3.2 Tune PIDs with step tests or relay auto-tuning; use IMC or lambda tuning targeting closed-loop time constant 3–5× process deadtime.
• 3.3 Implement constraint overrides: high selectors for pressure/temperature limits; bumpless transfer logic.
• 3.4 Commission sequences: startup/shutdown, heater light-off, coker drum switch-over, compressor start permissives.

III.4 Advanced Process Control (APC/MPC)

• 4.1 Identify controlled variables (CVs), manipulated variables (MVs), and constraints; include economic weights from planning.
• 4.2 Collect plant tests covering safe MV moves; build, validate, and cross-validate models; include disturbance variables (DVs).
• 4.3 Commission MPC with move suppression; enforce hard limits; schedule setpoints to keep units on active constraints.
• 4.4 Add inferentials: cutpoint, sulfur, RON, H2 purity; train with lab/analyzer data; implement bias tracking with data quality checks.

III.5 Real-Time Optimization (RTO) and Utilities

• 5.1 Construct site-wide LP/NLP with validated yields, energy prices, product values, hydrogen and steam networks.
• 5.2 Reconcile data (gross error detection) hourly; run RTO every 15–60 minutes; pass optimal targets to MPC/DCS.
• 5.3 Implement dedicated optimizers for furnaces (O2/CO), steam headers (letdown minimization), and hydrogen (compressor load sharing, purge optimization).

III.6 Asset Health and Alarm Management

• 6.1 Deploy CPM: loop oscillation index, valve stiction index, % time in auto, IAE trend, MV travel/range metrics.
• 6.2 Analyzer management: calibration schedules, validation against lab, outlier rejection, uptime > 95% with MTTR tracking.
• 6.3 Bad-actor program: monthly top-20 list for loops/instruments; root cause and corrective actions tied to CMMS.

III.7 Training, MOC, Sustainment

• 7.1 High-fidelity operator training simulator for startups/upsets; update annually with plant test data.
• 7.2 Management of Change: narratives, alarm rationalization, SIS proof tests, APC model refresh cadence (quarterly/after revamps).
• 7.3 KPI governance: weekly APC/RTO performance review; monthly energy and emissions review; quarterly benefit audit.

IV. Risk & Mitigation (HSE, Reliability, Redundancy)

• IV.1 HSE/SIS
  • 1.1 SIS integrity: proof test intervals; bypass management; independent sensor/logic/actuator paths; IPL verification.
  • 1.2 Heater safety: flame failure detection, LEL/O2 analyzers voting, permissives and purge verification.
• IV.2 Cybersecurity
  • 2.1 Network segmentation; application whitelisting; backups tested; incident response drills.
  • 2.2 Strict data diode/DMZ for business-to-control data; RTO/MES integration hardened.
• IV.3 Measurement/Analyzer Risks
  • 3.1 Analyzer drift/fouling: auto validation, redundancy, fallback to inferentials with bias limits.
  • 3.2 Bad data propagation: gross error detection, data quality flags, MPC soft-sensor weighting.
• IV.4 Mechanical/Control Risks
  • 4.1 Valve stiction/air failure: position feedback alarms; maintenance triggers on travel index/deadband.
  • 4.2 Compressor surge: high-speed antisurge PLC; surge detection using fast pressure ratio and flow; trip logic tested.
  • 4.3 Oscillation/cycling: anti-windup, output rate limits, decoupling, proper filter time constants.
• IV.5 Human Factors
  • 5.1 HMI design: high-performance graphics, situational awareness KPIs, alarm flood suppression, clear MPC states.
  • 5.2 Procedures: clear rules for manual override, MPC on/off criteria, and post-upset recovery sequences.

V. Optimization Levers

• V.1 Constraint-Based Operation
  • 1.1 Keep active constraints engaged: furnace duty, column ?P, regenerator temp/O2, compressor limits.
  • 1.2 Use move suppression and target back-offs dynamically based on variability and risk posture.
• V.2 Energy and Emissions
  • 2.1 O2 trim and CO control with bias-to-safety logic; sootblowing optimization by ?P/stack O2/CO trends.
  • 2.2 Heat integration: MPC manipulates pumparounds/reflux to maximize preheat train recovery subject to fouling limits.
  • 2.3 Steam network optimization: minimize letdowns, prioritize backpressure turbogenerators.
• V.3 Quality Giveaway Reduction
  • 3.1 Blend APC with in-line analyzers and inferentials; adaptive biasing with lab grabs; minimize RON/sulfur giveaway.
  • 3.2 Distillation inferentials: multi-point temperature/pressure analytics to lock cutpoints tightly.
• V.4 Hydrogen and Offgas Networks
  • 4.1 Optimize compressor loading, purifier cuts, make-up rates; purge minimization with H2 purity inferentials.
  • 4.2 Route offgas to most economic sink (fuel gas, recovery) within flare minimization constraints.
• V.5 Reliability-Driven Control
  • 5.1 Anti-fouling strategies: monitor ?P/UA trends; MPC moderates severity to extend run length; schedule cleanings.
  • 5.2 Valve/loop health integrated to maintenance: trigger work orders when stiction index or travel exceeds thresholds.
• V.6 Data Analytics and Digital Twins
  • 6.1 Hybrid soft sensors for sulfur/RON/H2 to cover analyzer downtime.
  • 6.2 Scenario testing on dynamic simulators to validate new control strategies before field deployment.

VI. Verification & Monitoring Plan

• VI.1 Daily
  • 1.1 Control room huddle: MPC status (service factor, active constraints, CV tracking), % loops in auto, top alarms.
  • 1.2 Energy/emissions: furnace O2/CO dashboards, flare rate, steam header imbalances.
  • 1.3 Quality: real-time analyzer vs. lab bias; product giveaway report.
• VI.2 Weekly
  • 2.1 APC benefit tracking vs. baseline; RTO convergence rate; hydrogen/steam optimizer savings.
  • 2.2 CPM: IAE changes, oscillation index, valve travel/stiction, bad-actor list and corrective actions.
  • 2.3 Analyzer uptime/MTBF/MTTR review; calibration and validation plan adherence.
• VI.3 Monthly/Quarterly
  • 3.1 Model maintenance: re-identify MPC models if MAPE > target (e.g., > 5–10%); refresh inferentials.
  • 3.2 Alarm KPIs: standing alarms, flood events, average alarm rate; rationalize and adjust limits/hysteresis.
  • 3.3 Economic audit: $/bbl benefit, SEC trend, CO2 intensity, flare intensity; reconcile with planning.
• VI.4 Tests and Proofs
  • 4.1 SIS proof testing per SIL; compressor antisurge functional tests; heater permissives.
  • 4.2 Disaster recovery: backup restore tests; cybersecurity tabletop exercise.
• VI.5 Documentation & Governance
  • 5.1 Control narratives/as-builts current; MOC records for all changes; training logs maintained.
  • 5.2 KPI dashboards archived; variance analysis and actions tracked to closure.

Summary Implementation Checklist

• • Control philosophy, alarm philosophy, cybersecurity zones approved.
• • Instruments/analyzers specified, installed, validated; valves with diagnostics.
• • PIDs tuned; sequences commissioned; SIS proof-tested.
• • APC/MPC built, tested, commissioned; inferentials validated.
• • RTO online with reconciliation; targets flowing to APC.
• • CPM/analyzer management active; KPI governance cadence in place.
• • Training simulator deployed; MOC and sustainment plan active.