Your workbench for engineering optimization

One intelligent workspace to connect your engineering simulation, run optimizations, and turn results into decisions.

Define

From simulation to optimization setup

Describe your engineering problem and Polaris translates it into the workflow required to get started.

Plan

Build a strategy around your goals.

Discuss what you want to achieve and Polaris recommends an approach adapted to the problem.

Run

Track your progress

Track live progress while Polaris monitors the run, surfaces important developments.

Decide

Turn results into decisions.

Identify promising candidates and decide whether to refine the search further.

Airfoil Design

Clear chat

Your local runner is connected. One evaluator registered:

Airfoilairfoil

Let's use it for a new study

Problem update

Connected to airfoil evaluator

+ variable thickness (continuous [0.06, 0.18])

+ variable camber (continuous [0, 0.05])

+ variable le_radius (continuous [0.006, 0.02])

+ variable te_wedge (continuous [4, 14])

+ variable alpha_cruise (continuous [0, 8])

+ objective cd_cruise (minimize)

+ objective mass (minimize)

+ constraint mass (lte 98)

~ evaluation: manual_script → external_command

~ name: "New Project" → "Airfoil Design"

✓ Applied automaticallyRevert

Five variables, two objectives, one constraint. Review the bounds on the right — I'll work inside whatever you set.

Noted — camber capped at 0.04.

What would you like to do with your simulation?

Understand the design space — see how different variables and outputs behave.

Find better designs — search for designs that improve specific objectives.

Find better designs

Then sample the space broadly first and refine on whatever looks promising — 200 evaluations at about 15 s each. I've set that up on the right; change either part.

Sampling is done and the search has started. Two evaluations have errored, both at thickness < 0.085 — thinner than your solver will handle. Leave it with me; I'll flag anything worth interrupting for.

50 min later
Run 1 finished — 200 evaluations, 12 errored. Improvement stopped after about 140, so I've opened the results.

mark the designs that are on the front

Visualization change

Added a Pareto plot with the front marked.

+ plot Pareto Plot with Front (pareto_front)

✓ Applied automaticallyRevert

Three non-dominated feasible designs — marked on the plot and in the candidates table. Below 104 counts of drag every kilo you save costs a lot of drag — open any design to see its pressure field and solver log.

+ UploadAsk the AI Engineer…|

Airfoil Design

3be3ad00-55f2-44c6-9297-14adcbbbc27a

Export

Define

Experiment

Evidence

Memory

fin_count	wall_t	pressure_drop	wall_temp_max	status
0.176	0.006	82.0	85.1	feasible	1 img 1 log
0.171	0.020	92.0	83.4	feasible	1 img 1 log
0.093	0.019	104.0	81.7	feasible	1 img 1 log
0.131	0.008	123.1	81.5	feasible	1 img 1 log
0.152	0.030	82.8	86.7	feasible	1 img 1 log
0.126	0.031	82.1	85.7	feasible	1 img 1 log

Designd042a4✕

OverviewImages (1)Logs (1)

cp_field

Written by your evaluator · 412 KB

cd_cruise

82.0

mass

85.1

thickness

0.176

camber

0.006

Polaris

From simulation to decision, in one intelligent workspace.

01Define

02Plan

03Run

04Decide

Simplicity

Engineering is complex. Using the tools to solve it shouldn't be.

Control

You set the direction. Polaris supports the process without taking decisions out of your hands.

Transparency

Every recommendation, action, and change remains visible, explainable, and auditable.

Your workbench for engineering optimization

One intelligent workspace to connect your engineering simulation, run optimizations, and turn results into decisions.

Define

Plan

Run

Decide

Simplicity

Control

Transparency

Run your next optimizationwith Polaris.

Run your next optimization
with Polaris.