Hazmat Segregation Compliance Check

Hazmat segregation compliance checking is a quality gate on the load plan. Before a trailer or unit load device is sealed, a classification model compares commodity placement against ADR and IATA segregation tables. It flags incompatible pairs, cites the UN numbers involved, and names the segregation rule ID that was breached. If placement is unknown, the check returns empty rather than inventing a pass. A supervisor still seals the load.

The job is narrow. It does not replace full dangerous-goods classification, SDS review, or carrier acceptance. It answers one question: given the planned positions of known hazmat commodities, do those positions violate published segregation rules for the mode of transport?

Load planners optimize for cube, weight, and stop sequence. Segregation sits in a different rulebook. ADR for road and IATA Dangerous Goods Regulations for air define which hazard classes and UN numbers may not share a vehicle, hold, or defined separation distance. A plan that meets axle limits and door order can still place oxidizers next to flammables, or stack incompatible packages in the same bay.

Manual review is slow and uneven. Planners scan labels, remember recent training, and hope the booking system already blocked bad combinations. Booking systems often stop at commodity acceptance, not at 3D placement. Two acceptable shipments can become non-compliant once they share the same trailer. Late swaps, cross-dock reloads, and multi-stop mixes make that gap worse.

The cost of a miss is operational and regulatory. Carriers can refuse the load, inspectors can detain the unit, and the shipper inherits incident risk if incompatible goods travel together. Catching segregation violations at seal time is late for replanning, but earlier than rolling the doors.

Inputs are structured. The model needs each hazmat line item with UN number, proper shipping name or class/division where required, packing group when relevant, quantity or package count, and a placement coordinate or bay/slot ID from the load plan. Mode context matters: ADR road rules and IATA air rules are not interchangeable, and the check must use the table that matches the move.

The classification step maps each commodity into the segregation taxonomy used by the chosen rule set. Compatibility is then evaluated pairwise (and, where rules require it, against group or zone constraints) using the planned positions. Distance and barrier rules are applied when the regulation distinguishes "away from," "separated from," or "segregated from" rather than a simple co-load ban.

Outputs are either empty or a list of violations. Empty means placement is unknown or no rule breach was found given the data provided. Unknown placement must not be treated as compliant. A violation record should include both UN numbers (or UN and incompatible class), the segregation rule ID from ADR or IATA, a short plain-language reason, and the load-plan locations involved so a planner can move packages without re-reading the full table.

Supervisor seal remains mandatory. The model is a quality control signal on the plan, not an electronic signature or carrier approval. If the check is red, the seal should wait until the plan is corrected or an authorized exception is documented outside this use case.

Related building blocks often sit upstream or beside this check. Commodity identity and class confirmation belong with hazmat compliance classification. Feasible repositioning after a flag is a job for a 3D load plan optimizer. Confirming that what was planned is what was loaded is closer to CV load completeness verification. Weight and axle constraints after a reshuffle can be rechecked with an axle weight distribution estimator.

Labelmaster is widely used for dangerous-goods labeling, documentation, and regulatory reference content. Teams often treat its materials and tools as the human-facing source of truth for how a commodity should be marked and what rules apply. A segregation compliance check does not replace that stack. It consumes UN and class data that labeling and booking workflows should already have validated, then tests placement. If Labelmaster-driven classification is wrong upstream, segregation output will be wrong for a different reason.

Descartes provides transportation and global logistics software that frequently includes dangerous-goods and compliance-oriented capabilities in broader TMS and shipping workflows. In many networks, Descartes-side booking or DG modules gate whether a commodity can ship at all. Segregation checking starts after acceptance: the commodity is allowed on the network, but the specific trailer or ULD layout may still be illegal. Integrating the check at load-plan finalize, not only at order create, closes that gap.

SAP GTS (Global Trade Services) is used for trade compliance, restricted-party screening, and related regulatory controls in SAP landscapes. Hazmat segregation is not the same as export control, but enterprise shippers often keep product and compliance attributes in the same master-data world GTS touches. The practical pattern is to pull authoritative UN and hazard attributes from the ERP/compliance master into the load plan, then run segregation as a yard or warehouse quality step before seal. Treating GTS as a placement engine is a category error; treating it as a trusted attribute source is usually correct.

Inspektlabs focuses on vehicle and asset inspection using computer vision. That is complementary, not redundant. Segregation compliance on the plan is a rules check on coordinates and UN numbers. Vision inspection can support physical condition, damage, or completeness cues at the yard. Pairing plan-level segregation with visual load verification reduces two different failure modes: illegal adjacency on paper, and "planned but not loaded as planned" on the dock.

Vendor boundaries matter for liability. Labeling vendors and trade-compliance suites do not automatically certify a sealed trailer. Vision vendors do not interpret ADR tables. The use case owns the segregation decision object: violations with UN numbers and rule IDs, or empty when placement cannot be assessed.

Data quality dominates accuracy. Missing UN numbers, free-text hazard descriptions, or bay IDs that do not map to a geometry model force an empty result. Prefer empty over a green seal recommendation. Version the rule pack explicitly (ADR edition, IATA DGR edition) so auditors can see which table produced a rule ID.

Scope the mode. A road trailer check must not silently apply air segregation distances, and air ULD checks must not assume ADR vehicle language. Multimodal moves need a clear primary rule set per leg, or separate checks per leg.

Keep humans in the seal path. Present violations in planner language with locations and rule IDs, not only model scores. Log who acknowledged a fail, what plan revision followed, and whether the recheck cleared. Do not auto-seal on model silence when placement inputs were incomplete.

Measure quality outcomes that match the use case: segregation violations caught before seal, false alarms investigated against the cited rule ID, and incidents or carrier refusals tied to incompatible co-loads after deployment. Those metrics stay in the quality outcome lane. Throughput gains from fewer refused loads are secondary effects, not the definition of success.

Success is a repeatable pre-seal quality control: every hazmat load plan with known placement either clears with no segregation violations under the stated ADR or IATA rule pack, or returns actionable violations naming UN numbers and segregation rule IDs. Unknown placement yields empty, and the supervisor still seals only after process rules say the gap is closed. The check sits beside labeling, trade compliance, TMS acceptance, and dock verification. It does not absorb them.