Digital Engineering · Star Industry

One model of the vehicle, useful at every phase.

An integrated architecture trade-study engine for launch vehicles and satellites. Every variant is a single file that each discipline model reads and writes — so one description of the vehicle stays consistent from the first trade study to the production baseline. Trade studies run in minutes, not weeks.

Request a Demo → See Worked Example

One file per architecture

Coupled convergence loop

Ranked KPIs across variants

SysON round-trip

CONVERj — architecture dashboard

One Model · Every Phase

Continuous integration for hardware.

The same architecture file underpins the whole programme — models, dashboard and the SysON bridge all read and write it. No re-modelling between phases, no drift between tools; each model can be swapped for a higher-fidelity version as the design matures.

01 · Trade Studies

Explore the space

Rank dozens of architectures on payload, cost and T/W — decide with evidence, not spreadsheets.

02 · Design

Size the vehicle

Coupled propulsion, tank, mass and cost models converge automatically for each variant; the optimiser closes the Δv budget.

03 · Qualification

Prove compliance

Every requirement checked against the live model, with the SysON product tree kept authoritative both ways.

04 · Production

Hold the baseline

Mass and cost resolved down to the part, and tracked as the design matures toward build.

How It Works

One file. A model chain. A clean SysON seam.

The architecture file is the single source of truth during computation — each model reads its inputs by name and writes only its own section, so the numbers can never fall out of step.

Source of truth

One file per architecture, with clear section ownership — meta, mission, propulsion, tank, mass, performance, cost, KPIs. Each model reads by dotted path and writes only what it owns.

File

The conductor

Runs the models in dependency order from the N2 interface diagram, iterating coupled loops — e.g. tank ⇆ COPV — until residuals converge below tolerance, automatically, for every variant.

Convergence

SysON bridge

A config-driven bridge seeds the architecture from the SysON product tree and writes converged outputs back — without ever owning the tree. The round-trip is only needed at the start and the end.

SysON

Case Study · The Trade-Study Problem

Replace the workbook stack.

Architecture decisions hinge on tightly coupled models. Running them by hand is slow, lossy and opaque — and the design is never done, it's re-baselined. Every iteration shifts the numbers you were just tracking, and everything downstream moves with them. CONVERj replaces the workbook stack with one architecture file and one convergence loop — so trade studies run in minutes, not weeks, and stay current as the design moves.

Trade-study task

Excel workbook stack

CONVERj

Source of truth

Numbers copied by hand between spreadsheets. A single stale cell invalidates the study.

Single architecture JSON. Models read inputs via dotted paths, write outputs to owned sections.

Convergence loops

Tank sizing and COPV placement iterate against each other. No one has patience for it by hand.

Conductor runs the full chain to convergence. The optimiser names the violated constraint, not "failed".

Comparing variants

Ten architectures means ten workbooks. KPIs live in different tabs, different formats.

Cross-variant ranking by any metric, filter, or custom expression. Live in the compare view.

SysON alignment

The authoritative product tree and the workbook diverge within days of the last review.

Round-trip with SysON: import requirements, validate per loop, push the updated product tree back.

Worked Example · ARCAS

Sub-orbital sounding rocket. Converged in 3 minutes.

A 1959 USAF / NASA double-base solid sounding rocket. A heritage benchmark with documented apogee, mass and trajectory data — the right vehicle to back-check a Phase A trade-study toolchain. Architecture imported from SysON → full model chain → trajectory matched within 0.2 % of heritage data.

64 km

Target apogee

4.5 kg

Payload

85°

Launch elevation

Wallops

Launch site

As Imported · 238.5 km

First run overshoots.

+174.5 km vs target. The configuration was sized for a different reference trajectory — CONVERj names the violated constraint, not just "failed". The optimiser converges from there.

ARCAS first-run trajectory in CONVERj — apogee 238.5 km, overshooting the 64 km target — // First run — 238.5 km apogee, target 64 km

Optimised · 64.1 km ✓

Bisection finds the answer.

+0.1 km vs target. Bisection on a propellant scaling factor — full chain re-runs each iteration. Trajectory matched heritage data within 0.2 %. Inert mass scales sub-linearly (k^0·85 default) — a one-line schema knob.

ARCAS optimised trajectory in CONVERj — apogee 64.1 km, matching the 64 km target — // Optimised — 64.1 km apogee, 0.2 % vs heritage

Scope

Orbital, suborbital, satellite — one tool.

The architecture's class drives which models run and how the app renders it. A per-architecture list selects the disciplines that matter for each vehicle.

Launch · Orbital

1–4 stages to LEO / SSO

Liquid or solid propulsion, sized by the full model chain and closed by a 3DOF performance model that flies the vehicle to orbit.

Launch · Suborbital

Sounding rockets & hoppers

Solid-motor or liquid-feed stacks flown by the trajectory engine to apogee. Heritage-validated against the ARCAS sounding rocket.

Satellites

Microsat → largesat

The launch-vehicle chain swaps for a bus-sizing pipeline: power, comms link budget and microgravity disturbance models.

The App · Six Working Views

The same machinery, driven from a browser.

Load a project, edit the architecture, run the models, watch the KPIs update.

Mission

Requirements in, Δv out

Define the mission and its requirements; each is mapped to a KPI and auto-verified against the live model.

Architecture

Configuration & KPIs

Live vehicle diagram redrawn from the schema, editable choices, KPI cards with pass / fail dots.

Optimise

Size to the mission

Size the vehicle to the mission automatically — the optimiser closes the Δv budget.

Product Tree

Resolved to the part

Declared vs computed mass and cost for every part, with divergences beyond tolerance flagged.

Trajectory

3DOF ascent to orbit

Ground track, staging events, max-Q and burn times fall straight out of the flown trajectory.

Compare

Rank every variant

Cross-architecture ranking by any KPI, filter or custom expression — live.

Compare & Trade-Off

Rank every variant. Live.

Run many architectures through the same convergence loop, then weight the KPIs that matter to your mission — mass, Δv, T/W, apogee, cost, microgravity, anything you choose. The ranking updates as you slide each weight, so the solution space is something you steer, not something you receive.

Cross-variant ranking by any metric, filter, or custom expression.
Live weighting sliders — explore the solution space against any priority mix without re-running.
Score breakdowns showing each KPI's contribution to the final ranking.
Add a new KPI by editing one config file; every architecture picks it up on the next run.

CONVERj trade-off view — multiple architectures ranked with weighted KPI sliders — // CONVERj — trade-off view

KPIs & Extensibility

Consistent KPIs. New models by config.

Every architecture reports the same KPI set — max payload, total Δv, wet mass, T/W at liftoff, total cost, cost per kg, payload fraction and convergence status — so any two variants are directly comparable.

Add by config, not by code. A new model drops into a model folder with its read/write contract named — the conductor auto-detects it. A new KPI is one entry in the mapping file, and every architecture picks it up on the next run. One config file, zero code changes.

See It Live

Try CONVERj for yourself.

Two ways in, both on request: a hands-on demo instance with a limited feature set you can explore at your own pace, or a live walkthrough of the full tool with our team.

Hands-on

Limited-feature demo

Request access to a live demo instance and explore the workflow, dashboard and compare view at your own pace.

Request demo access →

Guided

Full walkthrough

Book a live walkthrough of the full tool with our team — run against a mission close to yours.

Book a walkthrough →

Engagement Model

Three phases. Bespoke scope.

Every engagement is scoped and priced individually — tell us your mission and we shape the work around your programme. A typical trade study runs in three phases; the durations below are indicative and flex to match your cadence.

Phase 1 · Scope & Seed

2 weeks

Ingest your mission requirement.

We ingest your mission requirement, stand up the systems-engineering baseline, and seed a candidate architecture. Outputs: scoping memo, initial requirements set, baseline architecture.

Phase 2 · Trade Study

4 – 8 weeks

Run every variant of interest.

We run the model chain across every variant of interest — propellant, staging, engine count. You receive ranked KPIs, convergence reports, and the live CONVERj dashboard.

Phase 3 · Handover