Pipe QC Traceability: Lab and Production Line in One System
A web system for a ductile iron pipe plant that traces every pipe: from the chemical analysis of the metal in the lab, through the production stages and mechanical tests, to the sticker and the test certificate.
PythonFlaskSQLAlchemyAI
Context
In a pipe plant, accepting or rejecting a batch depends on numbers from the lab and numbers from the production line. When those numbers live in scattered paper and Excel files, the question "where did this pipe come from and why was it accepted?" becomes a manual search.
Scope I disclose publicly
- A Flask application on SQLAlchemy, with login and per-role permissions.
- Chemical analysis per ladle and furnace, with an accept or re-test decision.
- Pipe registration, production stages, mechanical tests, production orders and the warehouse.
- QR stickers for pipes, test certificates, scheduled reports and KPI alerts.
- Non-conformance tracking and defect reasons.
- An AI assistant and a daily summary, with an Arabic and English interface.
Decisions and trade-offs
- The pipe number is the key: every stage and every test is linked to it, so tracing is a query, not a search.
- Schema changes through Flask-Migrate, instead of a hand-kept list of ALTER TABLE statements that could only ever add columns.
- An audit log of changes, because a quality decision has to stay reviewable.
- The trade-off: server-rendered pages instead of an SPA, which is simpler to run inside a plant network.
The repository is public: the code, the models and the routes can be read directly. The screenshot uses demo data.
Check the public sourceDisclosure: A system delivered to a plant. I do not name the client and publish no real production data.