Why it matters
Without AI, solvent choices risk irreversible damage, hazardous exposure, and missed loan deadlines as solvent systems drift or misalign with materials, undermining reversibility and insurance claims.
From design to safety
Design
AI-guided solvent selection
The AI reads solubility parameter data, material class, and gel behavior to propose compatible solvent systems, gels, and wipes for proposed treatments based on historical outcomes; it also specifies parameter windows for acetone, ethyl acetate, and cyclohexane, plus acceptable ratios and temperature limits.
- Teas chart
- Solvent gels
- Pemulen
Build
System integration and validation
Lab tests verify parameter mappings, gel viscosity, and wipe compatibility, recording outcomes in a material safety and treatment notebook for audit and traceability; protocol sheets include solvent purity, batch IDs, and chain-of-custody notes.
- Teas chart
- Pemulen
- Evolon
- Cleaning tests protocol binder
Operation
Live cleaning decisions
Real-time AI checks parameter windows, triggers cleaning tests, and updates treatment proposals and condition reports for client approval and archival note in the vault; it logs solvent totals, gel loads, and wipe counts.
- Solvent gels
- Evolon
- Free solvents
Maintenance
Calibration and stock control
Regular calibration of parameter models, stock checks for solvents, gels, and wipes, and updates to the TEAS chart database with version control and audit trails; spare parts and calibration certificates are linked to each project file.
- Teas chart
- Solvent gels
- Free solvents
Safety
Safety and compliance monitoring
The system flags hazardous solvent levels, COSHH compliance, and reversibility checks, with sign-off by the accredited conservator for loan compliance and public exhibition safety; it maintains PPE logs and fume-hood performance checks.
- Cleaning tests protocol binder
- Pemulen
- Evolon
Part by part
- 1
Teas chart
- Watch for
- drift in solubility parameters, misreadings, incomplete parameter coverage across material types during pilot phase
- AI action
- readings compared to material profiles; alert if out of tolerance; update proposed solvent set
- 2
Solvent gels
- Watch for
- gel viscosity drift, phase separation, incompatibility with fabric or fiber; trend shift during long sessions
- AI action
- monitor viscosity data and gel flow; trigger alternative gel formulation if out of range
- 3
Pemulen
- Watch for
- emulsion stability loss, thickening inconsistencies in shear tests and storage
- AI action
- assess polymer dispersion; recommend alternative thickener or dilution when viscosity deviates
- 4
Evolon
- Watch for
- fiber damage risk from solvent exposure, compression marks on delicate textiles during wipe
- AI action
- predict wipe-surface interaction; adjust wipe density and solvent load in real time
- 5
Free solvents
- Watch for
- inventory drift, temperature excursion, improper labeling, cross-contamination risk
- AI action
- track stock conditions; issue re-stocking and handling reminders to onsite teams
- 6
Cleaning tests protocol binder
- Watch for
- missing test records, incomplete approvals, outdated methods, unaudited protocols
- AI action
- flag missing signatures; enforce current protocol version and link to treatment proposals
Worked example
Catching a solvent mismatch before a treatment bounce
Illustrative scenario with sample data.
A private studio evaluates a fragile varnish layer on an oil painting and tests a solvent–gel pair guided by solubility parameters to avoid swelling.
Signal in
material_id: varnish_油, parameter_set: A, gel: gel_01, solvent: acetone, temperature: 22C, humidity: 45%, test_id: T-2026-07, max_temp: 25C, min_humidity: 40%, batch_id: B-0426, purity: 99.8%, storage: sealed cabinet
System out
alert: parameter mismatch detected; recommended alternative gel and solvent; update treatment proposal with new test plan; log in condition report and PACR records; archive old protocol version
Human sign-off: Head conservator approves new protocol and updates the condition report and treatment proposal; signs off on reversibility plan and archive entry.
How to roll it out
Phase 1
Phase 1
Install TEAS chart database, calibrate gel viscosity sensors, and train staff on protocol and sign-off procedures.
Phase 2
Phase 2
Pilot in two small projects, log outcomes, and refine AI thresholds and test forms.
Phase 3
Phase 3
Full rollout across projects, ongoing QA, and integration with condition reporting and accreditation records.
Where people stay in charge
- Only a qualified conservator or restoration scientist may approve solvent protocol changes or sign off on treatment reports.
- The AI may not override reversibility requirements or PACR standards without human validation.
- All cost-sensitive decisions and safety flags must be reviewed by the head of conservation and documented in the treatment proposal.
Questions owners and crews ask
What data does the AI use to design solvent cleaning?
The AI uses solubility parameter data, material class, gel behavior, and past treatment outcomes to propose compatible solvents, gels, and wipes, then records decisions in the treatment proposal and condition reports.
How is reversibility ensured in this system?
Reversibility is checked through testing on reference samples, documenting outcomes, and requiring human sign-off before any irreversible step is adopted.
Who signs off on the final cleaning protocol?
The accredited conservator or head of conservation signs off on the final cleaning protocol and updates to the condition report.
What happens if a solvent or gel fails a test?
The AI triggers an alternate formulation, documents the rationale, and routes the updated treatment proposal for human approval.