Dynamic process models
Represent mass and energy balances, equipment dynamics, operating envelopes and disturbances in documented simulation environments.
Independent research · Process Control Engineering · Kenya
Engineering better control for dynamic systems.
NURU is an independent process-control engineering research lab startup. We investigate how first-principles models, advanced control, reinforcement learning and digital-twin methods can improve the way complex processes are understood and controlled.
We distinguish research plans from verified findings. Experimental methods are evaluated in simulation before any claim is made about safety or live-plant performance.
Process-control research loopSimulation first
NURU
CONTROL LAB
CONTROL LAB
01 / MODELDynamics + constraints
02 / CONTROLPID · MPC · RL
03 / EVALUATERobustness + safety
04 / REPORTReproducible evidence
FoundationProcess modelsApproachControlled testsStandardEvidence first
Research focus
Process control is the core. AI is a tool, not the claim.
The lab focuses on control problems with explicit dynamics, operating constraints and measurable objectives. We compare learning-based ideas against conventional engineering baselines.
Advanced and learning-based control
Study PID and model-predictive control alongside reinforcement-learning policies under the same constraints and evaluation protocol.
Simulation-to-operation methods
Investigate how engineering specifications and sensor data can be used to build, validate and monitor process digital twins.
Initial programme
Build a benchmark that engineering teams can inspect.
The initial technical focus is a constrained simulated process-control problem, with reproducibility and clear comparison against baseline controllers as first priorities.
WORKSTREAM 001 / EARLY-STAGE RESEARCH
Learning-based control of a simulated chemical process
Develop a dynamic process model, establish conventional control baselines and evaluate a learned policy on reference tracking, disturbance rejection, constraint violations and operating cost.
Research question
Can a learning-based controller improve multi-objective process performance without sacrificing constraint compliance or robustness?
The answer must come from controlled experiments—not an unsupported performance claim.
Current public status: research programme and implementation work are being developed. No peer-reviewed result or validated live-plant deployment is claimed on this website.
Research workflow
From physical understanding to measured performance.
A process-control study should be auditable from assumptions through results.
Model
Define process dynamics, parameters, constraints and disturbance scenarios.
Baseline
Establish and document conventional control strategies for fair comparison.
Evaluate
Measure tracking, robustness, constraint violations and resource use.
Release
Publish methods, code, settings and limitations when results are ready.
Independent lab. Transparent status.
NURU is presented as an independent research-lab startup. This website does not claim that it is a registered nonprofit, accredited institute, or a university laboratory. Research leadership at NURU does not by itself establish eligibility for any external programme.