KC Engineering / Capabilities
Three capabilities, one team.
Aerospace testing, machine learning and data, and controls. Most projects combine them, and senior engineers lead each one.
Plan, run and report a test campaign.
We handle the whole campaign: test plan, model design, instrumentation, tunnel runs, data reduction and the final report. One team is accountable from start to finish.
We have run tests in subsonic, transonic, supersonic and water tunnels, including setups where a computer feeds simulated flight loads to a model held in the tunnel.
A / Plan
Test plan
Design of experiments and a test matrix that fits the budget.
B / Build
Model and instruments
Model design, sensors and data acquisition layout.
C / Run
Tunnel execution
Force and moment, pressure, PIV, PSP, DIC and schlieren measurements.
D / Report
Data and reporting
Reduced data, uncertainty estimates and a written report.
Related papers
- 2022Experiments on Vortical Interactions on Generic Multi-Swept Wing Configurations
- 2023Nonlinear, Rate-Dependent Evaluation of Flight Damping Derivatives of the Orion Crew Module
- 2016Cyber-physical flexible wing for aeroelastic investigations of stall and classical flutter
- 2017Developing a reduced order model from structural kinematic measurements of a flexible finite span wing in stall flutter
Turn sensor data into predictions.
We build neural networks and reduced-order models from measured data, and put them into the instruments and control loops our clients already use. The result is software that runs on the target hardware.
Examples: a wave-elevation predictor for a wave-energy client, surrogate models for unsteady aerodynamic data, and low-order models of fluid flow for control design.
A / Predict
Surrogate models
Feedforward, LSTM and GRU models that stand in for slow simulations or measurements.
B / Reduce
Reduced-order models
POD, DMD and sparse identification of the dominant flow behavior.
C / Tools
Engineering software
PyTorch and Streamlit tools delivered to client engineers.
D / Deploy
Real-time inference
Models exported to LabVIEW real-time and FPGA targets.
Related papers
- 2026Wave Estimation Tool v1.1 - LSTM + Space-Time POD
- 2017Developing a reduced order model from structural kinematic measurements of a flexible finite span wing in stall flutter
- 2024Leading edge vortex dynamics on finite aspect ratio swept wings exhibiting large amplitude oscillations
- 2008Low-dimensional modelling of a transient cylinder wake using double proper orthogonal decomposition
Design controllers and test them on hardware.
We design controllers from measured data and physical models, then test them on real-time hardware. Methods include adaptive, model-reference, reduced-order and hybrid physics-plus-ML control.
Applications include flow control, wing flutter suppression, vortex-shedding regulation and wave-energy converters.
A / Identify
System identification
Models of a system built from test data.
B / Reduce
Model-based control
Controllers designed from low-order models.
C / Combine
Physics plus ML
Physical models corrected with machine-learned terms.
D / Deploy
Real-time hardware
LabVIEW real-time, FPGA and embedded targets, with hardware-in-the-loop tests.
Related papers
- 2014Closed-Loop Flow Control of a Forebody at a High Incidence Angle
- 2018Feedback Flow Control: A Heuristic Approach
- 2017Robust LQR control for stall flutter suppression: A polytopic approach
- 2022Efficiency analysis of the cycloidal wave energy convertor under real-time dynamic control using a 3D radiation model
Technical disciplines
What the work draws on.
Eight disciplines behind the three capabilities.
Illustrative visualization · generic planform · move pointer to orbit
- 001 Testing
Experimental aerodynamics
Subsonic, supersonic and water-tunnel campaigns: PIV, PSP, DIC, force and moment, pressure, schlieren.
PIVPSPDICSchlieren
- 002 Testing
Computational fluid dynamics
RANS, hybrid RANS/LES and DES, from CAD and meshing through to unsteady results, on DoD HPCMP systems.
RANS / DES / LESDoD HPCMPMeshing
- 003 ML
Machine learning
PyTorch models for engineering prediction, with Streamlit tools and export to real-time targets.
PyTorchLSTM / GRUStreamlit
- 004 MLControls
Reduced-order modeling
POD, DMD, SINDy and Koopman methods that turn complex physics into small models for control.
POD / DMDSINDyKoopman
- 005 TestingControls
Fluid-structure interaction
Dynamic stall, leading-edge vortices, gust response and flutter on flexible structures.
AeroelasticityGust responseDynamic stall
- 006 Controls
Closed-loop control
System identification, state estimation and feedback design on real-time hardware, including PID, MRAC and adaptive control above 1 kHz.
System IDMRACLabVIEW RT / FPGA
- 007 Controls
Active flow control
Synthetic jets, fluidic oscillators, plasma actuators and blowing to manage separation, lift, drag and shocks.
Synthetic jetsOscillatorsPlasma
- 008 TestingMLControls
Technical advisory
Independent review for program offices, R&D leaders and acquisition teams.
Program reviewRFP supportR&D strategy
Ways to work with us
How projects are structured.
Three common formats.
-
A - Study
Scoped technical study
A fixed-scope investigation: trade study, feasibility analysis, test design or machine-learning proof of concept.
Typically 6 to 16 weeks
-
B - Program
Cooperative R&D
Multi-year research as prime or subcontractor: test campaigns, hardware and controller deployment, with a full record of the work.
Typically 12 to 48 months
-
C - Advisory
Advisory retainer
Ongoing independent review and expert access for program offices and R&D leaders.
Monthly retainer
Have a hard technical problem?
Describe it and we will tell you whether and how we can help.