Construction sequencing automation

Installation sequencing is the part of planning that decides what can physically be built before something else. It is normally done by an experienced engineer reading drawings. This page sets out how that judgement can be encoded and applied consistently across an entire plant model.

Short answer

Construction sequencing can be automated when the logic is derived from engineering data rather than generated from text. Every item is assigned to a physical process area, equipment is sequenced per unit through erection and grouting, and piping, electrical and instrumentation work in an area is gated on the equipment in that area being ready. Durations, crew sizes and resources come from project engineering and staffing data, and the result is written as a dependency-correct network that imports into Primavera P6.

Last updated 20 August 2026 / Sensali, Perth, Western Australia

What installation sequencing actually decides

A construction sequence answers three questions for every item on a plant: when can work start, what has to be finished first, and how many people are needed. On an oil & gas or LNG facility that has to hold across thousands of items spanning structural steel, piping, electrical and instrumentation.

The rule set

  • Equipment is sequenced per physical unit, through erection and then grouting.
  • Piping, electrical and instrumentation are sequenced per process area rather than per item in isolation.
  • Area gating: bulk disciplines in an area cannot begin until every piece of equipment in that area is ready.
  • Relationships are finish to start with lag where the methodology requires it, so the network behaves correctly when Primavera P6 schedules it.
  • Durations and crew sizes are derived from project engineering and staffing data at a 1:1 basis rather than estimated.

Because the rules are explicit, they are applied identically across the whole model. The same model always produces the same sequence, which is not true of two planners working the same scope by hand.

Why process areas come first

Sequencing logic is meaningless until items are grouped into the physical systems they belong to, such as inlet receival, acid gas removal or mercury removal. Sensali derives that grouping from model containment and connectivity data combined with the project engineering documentation, measured at 98.8% accuracy on the models processed so far. Area assignment is what makes gating possible.

Why prompt based tools produce unusable sequences

CriterionPrompt based systemsEngineering driven sequencing
InputText promptNavisworks model and project engineering data
Model understandingNoneGeometry, hierarchy, containment, connectivity
DependenciesInferred and unverifiableDerived from process areas and construction methodology
SequencingGeneric templateApplied per unit and per process area
RepeatabilityVaries between runsDeterministic rules across the whole model
OutputText or spreadsheetPrimavera P6 XER, ready to import

A language model asked to write a construction sequence has no access to the geometry, no knowledge of which vessel sits inside which area, and no way to prove that a stated dependency is real. It produces text that reads like a schedule. Encoding the methodology instead means every relationship can be traced back to a physical fact about the plant.

Handling model revisions

A revised model normally means the schedule is reworked by hand. When the sequence is generated from the model, a revision is reprocessed through the same pipeline and the differences fall out of the new activity network instead of being hunted down manually.

Where this fits

Sequencing is stage four of the Sensali pipeline. The stages before it are described on Navisworks to Primavera P6, and the stage after it on Primavera P6 XER export.

Run this on one of your completed projects

We are working with a small number of EPC and project controls teams to validate Sensali. NDA available before any model or project data is shared.

Book a call →