Refinery Operator Combines Active AESA and Passive RF for Layered Defense

Context and Challenge

A large refinery operation running continuous, high-throughput processing faced an increasingly complex airspace risk profile. The site’s footprint included densely packed process units, flare stacks, tank farms, and perimeter access roads—an environment where a small unmanned aircraft could pose disproportionate consequences. Beyond safety concerns, the facility also managed operational continuity requirements that allowed limited downtime for upgrades or testing.

Several factors made aerial detection particularly difficult:

• High electromagnetic activity from industrial equipment, radios, and nearby infrastructure raised the baseline noise floor.
• Cluttered radar geometry around tall structures created multipath reflections and shadowed zones.
• Variable drone profiles (from small consumer craft to higher-end platforms with custom payloads) meant no single sensor modality was reliable across all scenarios.
• On-site constraints limited the number of elevated mounting locations and made wide-area coverage challenging without disrupting operations.

Prior to improvements, the security team relied heavily on visual reporting and fixed cameras near critical assets. That approach delivered intermittent coverage and minimal early warning—especially at night, in haze, or when drones approached low and fast. The refinery needed a detection capability that could provide early, consistent awareness while minimizing false alarms that would distract operators and degrade trust in the system.

Approach and Solution

The strategy centered on layered sensing, combining active AESA radar for wide-area detection with passive RF monitoring for emitter-based identification and intent cues. The intent was not redundancy for its own sake, but complementary coverage: when one mode was limited by the environment or the drone’s behavior, the other could still contribute actionable signal.

Designing for “Layered, Not Duplicated” Coverage

A practical design principle guided the architecture: each sensor type should answer a different part of the detection problem.

• Active AESA radar was tasked with:

  • Detecting and tracking objects regardless of whether they were transmitting RF
  • Maintaining continuous surveillance across approach corridors and open areas
  • Providing range, bearing, and track continuity for response planning

• Passive RF monitoring was tasked with:

  • Detecting control links, telemetry, and downlinks where present
  • Helping classify signals into likely drone activity versus background RF
  • Offering early cues about direction of arrival and potential operator proximity when emissions were detectable

The refinery’s goal was to turn drone detection into an operational workflow: not simply “alarm on drone,” but assess, verify, and respond with confidence.

Step 1: Site Survey and Threat Modeling

A multi-disciplinary team from security, operations, and safety conducted a survey to identify:

• Most likely approach paths (along roads, from adjacent open land, and from over-water vectors where applicable)
• Critical zones requiring earliest warning (control rooms, tank farms, flare areas, and utility corridors)
• Known RF hotspots that could complicate passive sensing
• Radar line-of-sight constraints created by tall equipment and terrain features

The assessment prioritized coverage continuity and alarm quality over maximum range on paper. In practice, consistent detection inside a refinery’s cluttered environment often benefits more from smart placement than from raw sensor performance.

Step 2: AESA Radar Placement for Persistent Tracks

Active AESA radar units were positioned to create overlapping coverage. Overlap served two purposes:

1. Reducing shadowed areas behind structures
2. Improving track stability when reflections and clutter fluctuated with weather and process conditions

Configuration focused on tuning detection thresholds to the environment. The team emphasized:

• Clutter suppression calibrated to the refinery’s geometry
• Track persistence settings to avoid “track popping” during brief occlusions
• Operational profiles for different conditions (day/night, heavy rain, high-activity maintenance periods)

This helped shift radar alerts from frequent, low-confidence contacts to fewer, higher-confidence tracks—critical for keeping response teams focused.

Step 3: Passive RF Monitoring for Classification and Corroboration

Passive RF sensors were deployed with attention to:

• Antenna placement away from the loudest on-site emitters where feasible
• Coverage of likely operator areas outside the fence line
• Filtering baselines to account for known legitimate emissions (facility radios, industrial telemetry, nearby commercial traffic)

The passive layer was not treated as a guaranteed detector; rather, it was framed as a context engine. When RF emissions were present, they could corroborate radar tracks, help prioritize alerts, and support faster verification.

Step 4: Fusing Alerts Into a Single Workflow

A key improvement was integrating both detection modes into a unified operating picture, so the security team could quickly answer:

• Is there a track (active detection)?
• Is there a correlated RF signature (passive corroboration)?
• Does the behavior match a drone profile (speed, altitude changes, loitering patterns)?
• What is the confidence level, and what response step is appropriate?

The fusion logic emphasized correlation, not duplication. For example:

• A radar track with no RF was treated as potentially autonomous flight, a non-emitting craft, or an object requiring visual verification.
• RF detection with no radar track was treated as either distant activity outside radar coverage, non-line-of-sight propagation effects, or benign interference—prompting targeted checks rather than immediate escalation.
• A radar track with matching RF directionality was prioritized as high-confidence and routed into a faster response path.

Step 5: Operational Playbooks and Training

Technology alone was not considered sufficient. The refinery implemented playbooks aligned to alert confidence:

• Monitor: low-confidence alerts, background checks, camera cueing
• Verify: dispatch visual confirmation, thermal camera handoff, perimeter observation
• Escalate: high-confidence correlated detections near critical zones, with defined safety and communications steps

Training focused on preventing overreaction while ensuring speed when it mattered. Operators were coached to interpret sensor data as decision support, not as a standalone verdict.

Results

The layered approach produced practical improvements in detection quality and response coordination. While exact performance depends on local conditions, the refinery observed several outcomes that were consistently repeatable:

• Earlier awareness of aerial approaches, especially along predictable corridors where radar tracking was most stable
• Improved alert confidence when passive RF corroborated radar tracks, reducing time spent debating whether an alert was real
• Faster verification through automated camera cueing and clearer prioritization of which contacts mattered most
• Reduced operator fatigue by shifting from frequent ambiguous alerts to fewer, better-contextualized events
• Better after-action clarity, with recorded tracks and RF detections supporting review, tuning, and procedural refinement

Importantly, the refinery also gained resilience against common failure modes:

• When drones emitted no detectable RF, active radar still provided detection and tracking.
• When radar performance was temporarily degraded by clutter, weather, or geometry, passive RF sometimes still provided early indication of activity.
• When both were available, correlation offered a high-confidence pathway for rapid decision-making.

Key Takeaways

• Layered sensing is about complementarity. Active AESA radar and passive RF do different jobs; pairing them reduces blind spots and improves decision quality.
• Placement and tuning matter as much as sensor selection. Industrial clutter, reflections, and RF noise can defeat a “spec sheet” design without careful site-specific configuration.
• Correlation reduces false alarms and speeds response. The operational value comes from fusing tracks and emissions into a single, prioritized workflow.
• Workflows are part of the system. Clear playbooks, training, and verification steps prevent both complacency and overreaction.
• Expect variability and plan for it. Drones may be non-emitting, automated, or flown in challenging geometries; a layered design provides resilience across scenarios.

By combining active AESA detection with passive RF context, the refinery moved from reactive spotting to structured, early-warning surveillance—strengthening protection around high-value assets without overwhelming operators or disrupting continuous operations.