Canopy Platform
Configuring and solving optimal control (trajectory optimisation) problems using collocation
We built the state-of-the-art, so you don’t have to
At Canopy Simulations we observed that team after team will re-invent the same laptime simulation technology, with varying degrees of success. We realised that we could end the wasteful “not invented here” mentality, and get vehicle dynamicists back to focusing on what they do best; finding lap time.
We combined our decades of expertise to create the most advanced car model, simulations and data analysis tools yet seen in Formula One or any other motorsport division. Then we unleashed their true potential using the power of cloud computing.
The Canopy Platform has been built around these tools, enabling anyone to leapfrog the competition to have best-in-class simulation capability at a fraction of the cost of developing in-house.
Our Simulations
Dynamic lap / Dynamic multi-lap
The holy grail of lap simulation. A fully dynamic simulation with no driver model, achieving the maximum performance of the car, including all of its dynamics, without any assumptions about how it is driven.
Dynamic Lap solves every state of the car model at every point on the track simultaneously to give a dynamically consistent state trajectory which traverses the lap in minimum time subject to any number of algebraic or integral constraints on its behaviour.
With support for multiple laps, powertrain thermal and tyre thermal, Dynamic Multi-Lap allow you to optimise entire stints.
Secondary lap sensitivities
The computation of local sensitivities around the lap help us to evaluate complex engineering trade-offs. Additional vector outputs are computed for lift and drag coefficients, ride heights, mass, power and grip.
When we’re deciding whether to fit a component or make a setup change, we typically gain in some places, but lose in others. Secondary Lap Sensitivities allow you to identify where these gains and losses originate.
Channel inference
Working backwards from telemetry data, this simulation uses the vehicle model to estimate the value for channels which were not measured directly. It uses the quasi-static equilibrium of the car and an inverse tyre model to make these deductions.
Channel Inference doesn’t have to be run on track data; any lap simulation output can be staged as telemetry. In this way it is possible to look at the effect of small changes in car setup on balance.
For example if two inference sims are run with different rear anti-roll bar rising rates on the same telemetry we could examine delta rLatBalF in the results to see how this might affect car balance at particular points around the circuit.
While this is a simplification, because we wouldn’t achieve exactly the same vCar-gLat-gLong around the circuit after the setup change, it can be easier to interpret than a lap simulation, because the lap driven doesn’t change.
Quasi-static lap
Despite being the traditional workhouse of simulation in Formula 1, Canopy have pushed forward the state of the art of simulation to such an extent that this is largely obsolete. This simulation finds the quasi-static equilibrium of the car at every point around the track which is consistent with the speed and acceleration of the points surrounding it.
Taking approximately 1min to simulate an average lap, this is one of the fastest (not to mention most high-fidelity) quasi-static lap simulations in the world. However, the majority of our customers prefer to wait another couple of minutes for the vastly improved realism obtained from Dynamic Lap.
Generate racing line
Given a car and a track definition containing only the x-y coordinates of the track edges, Generate Racing Line will find the racing line which enables the given car to lap the track in the shortest time.
Straight Sim
Straight Sim is a straight-line simulation which finds the static equilibrium state (with no acceleration) of the car at a series of speeds from 0 to 100m/s.
Tyre thermal dynamic lap
After correlating to track or rig data using Thermal Replay, the thermal dynamics of the tyre ribs, through thickness elements, wheels, and brakes can be computed.
Combined with the effect of temperature on grip and cornering stiffness, Dynamic Lap can find the optimal driving style to get the most out of the tyres. This allows us to make informed decisions regarding the thermal effect of setup changes.
Combined with strategy weights this tool can also be used to find tyre saving opportunities around the lap while maintaining a competitive advantage in places which are critical for overtaking.
This is a computationally demanding exercise with 250,000 unknowns, but thanks to the powerful cloud computing resources available, up to 200 laps can be computed in parallel in around 25 minutes.
This tyre thermal model can be run in real time in the driver-in-loop simulator so drivers experience the feedback effect of their driving on tyre temperature.
Thermal replay
In order to speed up correlation to your measured powertrain or tyre temperatures this tool enables you to input telemetry data, run it through the thermal model, and examine your fit to track data.
In conjunction with explorations you can sweep through thousands of combinations of parameters in minutes to help you quickly dial in your correlation.
If you’re interested in powertrain temperatures and battery derating, then you can run these parameters through Dynamic Multi-Lap to perform a full stint energy and thermal optimisation. Alternatively if you are concerned about tyre temperatures, then this puts you in a good place to run Tyre Thermal Dynamic Lap.
Limit Sim
Limit Sim is the cornering version of Straight Sim. User can customise the gLat-gLong angle for which quasi-static equilibrium is achieved. The simulation will then return the maximum cornering performance for the customised condition (e.g. late entry, mid corner, early exit…) through the speed range.
Virtual 4-post RIG
he car is run over a defined road profile by feeding the road heights into the four corners (wheels) of the car in continuous time.
The results show the dynamic response of the car to the given road profile, enabling analysis of contact patch load variation and exploration of how changes in damping and compliance will affect ride characteristics.
Ride Sim
Ride Sim is our simulation for assessing car performance over track bumps. Ride Sim will firstly compute the optimal racing line for a given car, before driving the car around this racing line over a fully featured track bump profile. Users can then assess the ride performance of the car at the limits of performance, taking into account the inertial loads, the impact of aerodynamics on ride and all other effects on the vertical dynamics we would want to include.