01First focus
Mechanical
structures · mechanisms · thermal · fluids
- loads and stiffness
- fatigue
- heat paths
- flow
- kinematics
Hardest problem
Tolerance stacks and fatigue life, where small errors compound over thousands of cycles.
Engineering
Intent → engineered physical system.
A user should eventually be able to start with “I want to build this.” Terranoux is being built to carry that intent through the engineering work required to make it real.
Building towardThis page describes the system Terranoux is building, not what it does today.
From one sentence
What Terranoux should help determine.
- 01
Understand
- What the object or system needs to do
- Constraints
- Engineering requirements
- 02
Design
- Possible architectures
- Component choices
- Materials
- 03
Model and test
- CAD and geometry
- Simulations
- Failure modes
- Experiments
- 04
Build
- Prototypes
- Manufacturing processes
- Bill of materials
- Suppliers
- Estimated cost
- 05
Close the loop
- Validation
- Iteration
The loop
Eight stages. Every one can send you back.
- 01
Understand
Turn intent into requirements and constraints.
What must the thing do, in what environment, for how long, at what cost? Vague intent becomes measurable requirements, and the requirements that conflict are surfaced early.
- requirements
- environment
- cost targets
- standards
- output
- Requirements with acceptance criteria
- sends it back
- A requirement turns out to conflict with another.
- 02
Design
Generate and compare engineering approaches.
Several architectures, not one. Each is compared against the requirements on performance, risk, cost and how hard it will be to make.
- architectures
- trade-offs
- components
- risk
- output
- Ranked candidate designs
- sends it back
- No candidate design satisfies the constraints.
- 03
Model
Represent geometry, materials, components and system behaviour.
The chosen design becomes geometry, material assignments, component selections and behavioural models, all linked back to the requirements they serve.
- geometry
- materials
- tolerances
- interfaces
- output
- Engineering model
- sends it back
- The model is missing a behaviour that matters.
- 04
Simulate
Test ideas computationally before spending physical resources.
Structural, thermal, fluid, electrical and dynamic analysis where credible models exist. Failures found here are cheap.
- physics
- failure modes
- margins
- sensitivity
- output
- Predicted behaviour with known limits
- sends it back
- Simulation shows insufficient margin.
- 05
Experiment
Run targeted physical experiments where simulation is insufficient.
When the model is uncertain, a small, bounded experiment answers the specific question. Not a full prototype, just the measurement that matters.
- uncertainty
- test design
- instruments
- evidence
- output
- Measurements that update the model
- sends it back
- A measurement disagrees with the model.
- 06
Prototype
Produce testable physical iterations.
A physical build made with the process closest to production that the stage allows, so what is learned carries forward.
- fabrication
- assembly
- lead time
- cost
- output
- Physical iteration
- sends it back
- The prototype cannot be built as designed.
- 07
Validate
Measure whether the system actually meets its requirements.
Every requirement from the first stage is tested against the physical result. Passing simulation is not passing validation.
- test plans
- acceptance
- margins
- reliability
- output
- Evidence against each requirement
- sends it back
- A requirement is not met in testing.
- 08
Manufacture
Translate the validated design into a repeatable production process.
Processes, tooling, tolerances, bill of materials, suppliers and quality control, so the thousandth unit matches the one that was validated.
- processes
- BOM
- suppliers
- quality
- output
- Production-ready design package
- sends it back
- The process cannot hold a tolerance at cost.
Reasoning across
Eleven things that have to agree.
Engineering decisions are rarely about one domain. The value is in holding all of them at once.
- Requirements
- What must be true, and how it will be checked.
- Geometry
- Shape, fit, clearances and interfaces.
- Materials
- Properties as processed, not as catalogued.
- Components
- Off-the-shelf parts, ratings and availability.
- Physics
- Structural, thermal, fluid, electrical, dynamic.
- Tolerances
- What variation the design can absorb.
- Cost
- Per unit, per volume, per change.
- Manufacturability
- What a real process can make repeatably.
- Suppliers
- Who can make or provide it, how fast.
- Testing
- Which measurement settles which question.
- Iteration
- What changes when evidence disagrees.
Not just generation
Engineering requires reality.
Generating a plausible design is the easy part. A physical system has to survive all of this, and the only way to know is to check: simulation, tools, experiments, measurements and iteration.
Physics
loads, dynamics, energy
Tolerances
stack-ups, fits, clearances
Material behaviour
yield, creep, fatigue
Heat
dissipation, expansion, derating
Force
impact, vibration, contact
Wear
friction, abrasion, life
Manufacturing
process limits, access, cost
Cost
materials, labour, volume
Testing
the result that settles it
Example workflow · illustrative
The inspection vehicle, start to finish.
One hypothetical project through all eight stages, including the part where testing proves the first design wrong.
Intent
Illustrative example
Build a lightweight autonomous inspection vehicle.
“A portable inspection platform for rough terrain. One person should be able to carry it.”
One sentence of intent. Everything that follows has to trace back to it.
No such vehicle has been built. Values show the kind of reasoning involved.
One connected representation
Change one thing. See what else moves.
Requirements, design decisions, models, simulations, experiments, components, cost and manufacturing are linked. That is the difference between a reasoning system and a folder of files.
Change one constraint
Pick a change. A connected model can trace what it affects. A pile of separate files cannot.
Illustrative. Uses the example inspection vehicle from the homepage.
Where Terranoux can go
Disciplines.
Directions the engineering model is being designed to cover. None is offered as a supported capability today.
02Direction
Electrical
electronics · power · controls · embedded systems
- power budgets
- component selection
- thermal derating
- control loops
- firmware
Hardest problem
Keeping electrical, thermal and mechanical models consistent as a board changes.
03Direction
Materials
selection · properties · fabrication · testing
- property data
- processing effects
- compatibility
- cost and availability
Hardest problem
Published properties rarely match the part that comes out of a specific process.
04Direction
Manufacturing
processes · tooling · tolerances · assembly
- process capability
- design for manufacture
- assembly sequence
- inspection
Hardest problem
Knowing what a specific machine, shop or supplier can actually hold, not what a catalogue says.
05Direction
Robotics
mechanisms · sensing · control · physical interaction
- actuation
- sensing
- dynamics
- contact
- safety
Hardest problem
Contact and friction, where simulation and reality diverge fastest.
A design is only an idea until reality agrees.