Systems engineering · 2024–2025
A bioartificial liver, made testable before it was built.
A five-person senior-design team turned a clinically ambitious device into a six-module safety pipeline and a computational twin they could simulate, challenge, and explain.
The question
Biology, fluid mechanics, controls, and clinical safety all had to agree before the design could be credible.
The outcome
The team produced a four-cartridge architecture and a digital twin that exposed flow, clearance, and safety assumptions at one-minute resolution.
Why it matters
AI compressed the build, but framing made the tools useful. Once the team could name the modules, inputs, transitions, and failure conditions, model-assisted coding became an engineering loop instead of prompt-driven guesswork.
- Role
- Industry mentor; system framing and technical guidance
- Collaborators
- Five UGA biomedical engineering students; faculty advisor Dr. James Kastner
- Duration
- Two semesters
- Period
- 2024–2025
The decisive artifact was the decomposition: six bounded modules, explicit transitions, and a model that made every assumption inspectable.
From plasma separation through the final safety return gate.
Across plasma and hepatocyte compartments.
Two parallel cartridges across two serial stages.
System map
Six gates between blood and patient.
Open each module to inspect its role. The sequence is fail-closed: the final return path stays unavailable until the monitored conditions clear their thresholds.
01Plasma separator+
Splits plasma from the cellular stream so treatment can happen without exposing cells to the reactor.
02Pump regulation+
Maintains flow and pressure inside explicit safety limits before material enters the cartridges.
03Cell metabolism+
Models diffusion and hepatocyte conversion across plasma and cell compartments with ten coupled equations.
04Sampler+
Creates an observable checkpoint for the modeled outlet conditions.
05Mixer+
Recombines treated plasma with the bypassed cellular stream.
06Return monitor+
Keeps the return path closed until the measured conditions satisfy the safety gate.
Consequential decisions
Where the work changed direction.
These choices shaped the architecture, the evidence, or the way the system could be used.
Simulate before touching the bench
The team used simulation to narrow the design space before committing scarce lab time. Literature became model inputs: transport constants, compartment volumes, reaction rates, and clinical thresholds.
Split for safety, stack for clearance
Two parallel cartridges reduced flow through each unit and protected the hepatocytes from shear. A second serial stage compounded clearance when one pass was not enough.
Make the return gate fail closed
The final module checks toxin levels and operating conditions before treated blood returns to the patient. Safety logic became part of the architecture, not a note added at the end.