Industrial Use Case 6

UC6: Olfactory Sensing Arrays

From tailored sensing materials to integrated 64×128 digital olfactory hardware with cross-reactive chemical fingerprint readout.

Use case 6

UC6

Deploy materials & generate chemical intelligence

64×128 digital olfactory array

Use Case 6 advances tailored sensing materials from formulation to fully integrated hardware systems.

Led by Oliveris Tech Incubator Ltd (Ireland), UC6 establishes a high-density 64×128 device-level digital olfactory array capable of capturing multi-dimensional, cross-reactive chemical fingerprints rather than isolated molecular signals.

To train and validate the system, UC6 introduces a Closed-Loop Training Platform (CLTP)—a controlled environment that generates rich multimodal process-fingerprint datasets under simulated operating disturbances. This closed loop connects AI material screening directly with hardware readout integration and experimental validation.

Ultimately, UC6 provides real-time decision support for industrial operators, significantly improving process availability, heat exchanger efficiency, and resource recovery in harsh operating conditions.

Lead partner
Oliveris Tech Incubator
Consortium partners
Oliveris Tech Incubator, University of Sheffield, TU Eindhoven, FLOWPHYS

Paired use case

UC5: Analyte-sensitive coatings for digital olfaction

Discover & optimise sensing materials

64×1288192 pixelsENIG-on-FR4CLTPUQmulti-domainOLV

Target KPIs

  • 64×128 IDC assembled (8192 pixels); 8×8/16×16 coupon sub-arrays for ramp-up
  • ≥30% increase in response-manifold volume vs random coating placement
  • Macro-F1 ≥0.85 with calibrated UQ (ECE ≤0.05)
  • Array-level LoD ≤1–5 ppm on ≥2 VOC classes; t90 ≤5 s median
  • Drift ≤2%/24 h with RH/T compensation
  • ≥90% coating yield; <5% pixel-to-pixel variance on ≥8×8 coupons
  • EIS/Nyquist feature error ≤10% vs surrogate on ≥3 analyte conditions
  • CLTP fingerprint datasets — carbon capture ↔ geothermal transferability

Validation: WP13 (definition) → WP14 (in-silico screening) → WP15 (experimental validation). Reports published as D14.x and D15.x deliverables become public.

UC5 + UC6 integrated system

Chemical state monitoring through AI-discovered analyte-sensitive materials. The array is the validation vehicle — commercial value is replacing laboratory analysis, manual sampling, and delayed process diagnostics.

UC5 discovers materialsUC6 deploys arraysSciFM learns chemistryWorks across domains
UC5: Functional Coatings — formulation track →

Learning hierarchy

Level 1

Material response

Functional coatings interact with the chemical environment and generate electrical responses.

Level 2

Sensor fingerprint

Multiple cross-reactive sensors produce a unique multidimensional response pattern.

Level 3

Chemical state recognition

AI identifies the underlying chemical and process state from the fingerprint.

Level 4

Industrial intelligence

Chemical-state information becomes actionable process knowledge.

Level 5

Decision support

Operators act on AI-generated intelligence to optimise processes.

Industrial targets

Carbon capture solvents

Amine plants (MEA, DEA, MDEA, Piperazine) — degradation produces NH₃, aldehydes, ketones, organic acids, sulphur compounds.

Today: expensive, laboratory-based, periodic sampling.

  • Solvent health & remaining solvent life
  • Corrosion risk
  • Capture efficiency
  • Continuous degradation chemistry fingerprinting

Geothermal brines

Dissolved silica, chlorides, carbonates, sulphates, trace metals — chemistry drives scaling, corrosion, and mineral precipitation.

Today: periodic sampling with limited real-time process visibility.

  • Scaling propensity & brine stability
  • Corrosion risk
  • Mineral extraction performance
  • Digital-twin process intelligence

What the AI learns

Carbon capture

  • Solvent loading state (lean, partially loaded, rich)
  • Degradation pathways (oxidation, thermal, heat-stable salts)
  • Contamination signatures (SOx, NOx, particulates, metals)
  • Foaming & operational instability precursors
  • Corrosion-related chemical environments
  • Capture efficiency & regeneration effectiveness

Geothermal brines

  • Brine chemistry state (low / moderate / high salinity)
  • Scaling pathways (silica, carbonate, sulphate deposition)
  • Contamination signatures (chemicals, drilling residues, trace metals)
  • Mineral precipitation precursors
  • Corrosion environments (low pH, chloride-rich, sulphide species)
  • Heat exchanger efficiency & plant availability

Why digital olfaction

AspectConventionalSimuLingua
Detection modelOne sensor → one analyteCross-reactive arrays
CalibrationCalibration-basedFingerprint-based
TransferabilityLimitedMulti-domain
TimingLaboratory-dependentReal-time
InterpretationStatic thresholdsAI learning
DataSingle variableMultimodal

Project updates subscription

Get milestones, publications, events, consortium news and other SimuLingua project information — straight to your inbox.

Project coordination

Get in touch with the SimuLingua project.

FLOWPHYS AS coordinates the Horizon Europe action. Reach out for scientific, consortium or press enquiries across our nine partners.

Project enquiries

HORIZON-RIA · GenAI4EU

1 June 2026May 2030

HORIZON-CL4-INDUSTRY-2025-01-DIGITAL-61

Follow the project on our social networks