← All work

A live window into a robotic manufacturing process that used to be a black box.

A research system connecting IoT enabled wire arc additive manufacturing to a SQL database, so a build can be queried and watched while it is still running.

Runs on IoT sensors SQL Information model Real time visualisation Adaptive workflows
Research WAAM Process Intelligence
Status Research Data path Sensor to SQL Latency Real time Post hoc reconstruction None

The challenge

Wire arc additive manufacturing produces a continuous stream of sensor data while it builds. Almost none of it was structured, and none of it was queryable while the build was in progress.

Researchers reconstructed what happened afterwards from logs. Anything that went wrong mid build was discovered far too late to change the outcome.

01 Sensor output had to land in a structured store, not a file to be parsed later.
02 The process had to be observable during the build, not after it finished.
03 The system had to keep working as the process itself was changed.

The system

01

Model

An information model maps what the machine reports onto a schema a researcher can query without knowing the hardware.

02

Stream

Sensor output is written directly to a SQL database as the build runs, rather than batched into files afterwards.

03

Observe

A command centre visualises, sorts, and filters the live data, so anomalies are visible while the part is still on the machine.

04

Adapt

Read and write workflows adjust as process parameters change, so the instrumentation does not have to be rebuilt each time.

How it works under the hood

Structuring the data at capture time is what makes everything downstream possible. Once sensor output is relational, querying a build becomes an ordinary database question rather than a parsing exercise, and the visualisation layer is a thin client over it.

Observable while it runs.

The difference between a log and an instrument is timing. A log tells you what happened; an instrument lets you act before it finishes happening.

What this means for your business

Machines that already talk

Most industrial equipment emits far more data than anyone stores. Capturing it in a queryable form is usually the whole project.

Query instead of parse

Once the data is relational, questions that took a custom script become a query anybody on the team can write.

Instrument the process

Real time visibility turns a post mortem culture into an intervention one, which is where the savings actually sit.

Instrument what you already run.

We build and run these systems ourselves, so you are hiring people who operate the thing rather than specify it. Tell us what your equipment reports and what you cannot see today.

Book a call

Thirty minutes on what runs manually today. We follow up by email and send a proposal with scope and a fixed price before any work starts.

Or hire us through Upwork ↗