← Career
PROJECT 05 / INDUSTRIAL IoT

Container Dryer Control

Turned intuition-driven pepper drying into a repeatable, data-based recipe

ROLE
Solo across the entire dev stack
PERIOD
2023.09 — 2024.08
STATUS
a dozen units · in stable operation
Architecture
  1. 01 건조기 펌웨어
  2. TCP 02 수집 · 파싱
  3. 03 레시피 제어
  4. 04 태블릿 제어 · 관제

The expert finds the recipe; the system reproduces it

What temperature-humidity curve makes good peppers is food science, not software. A food-science research director owned the recipe; the system collected and visualized the data that let him find it, then reproduced the confirmed recipe identically every run.

Early fault detection from vibration and temperature

A failure mid-drying ruins the whole batch. Watching temperature anomalies and vibration together surfaced a weakening airflow pattern on one side — the machine was stopped, a faulty circulation fan found and replaced before any batch was lost. Predictive maintenance confirmed in action, not in slogan.

Migrating the tablet app from React Native to Flutter

The tablet app began in React Native; performance, library support, and update upkeep on field tablets prompted a re-evaluation, and it moved to Flutter. The judgment carried forward — the crane safety app was later built Flutter-first.

Local control always wins

When the on-site tablet and remote control could conflict, local control was made unconditionally dominant — remote never overrides the person standing at the machine.

Status

Deployed at a pepper-drying site and in stable operation since. Talks to the firmware over raw TCP, parsing binary packets directly.