Project Cassandra
Public capability explorer

From evidence to a reviewed decision

Welcome to the Thermal Cortex.

Here is the step-by-step workflow of ARG-BLC1, designed specifically for industrial chillers.

Each stage builds on the previous one. Four seconds per step; the completed architecture stays in view.

ARG-BLC1 architecture1 / 6

Use the layer controls to explore each stage.

A public component view of ARG-BLC1. Forecasting branches from checked data; diagnosis and prediction meet at reviewed evaluation. Visual nodes are illustrative.
01of 06

Workflow step

Data collection

Collect available temperature, pressure, flow and power readings on one shared timeline.

Explore layer by layer

Use arrow keys to move between stages.

Read all six stage explanations

01

Data collection

Bring the evidence together.

An investigation begins with available temperature, flow, electrical and operating records arranged around a shared timeline.

  • Identify what each measurement represents.
  • Keep units, timestamps and equipment boundaries visible.
  • Distinguish measured values from calculated quantities.

Do the available records describe the same equipment and period?

02

Measurement quality

Check what the evidence can support.

Missing, repeated, stale or conflicting readings can change an interpretation. The quality of the evidence sets the limits of the analysis.

  • Check continuity and agreement between records.
  • Make gaps and uncertain measurements explicit.
  • Resolve questionable readings before drawing a conclusion.

Which observations are reliable, and which still need verification?

03

Thermodynamic context

Connect readings to physical behaviour.

Heat transfer and electrical demand become more useful when considered with operating state, load and the conditions around the equipment.

  • Use a consistent heat and power boundary.
  • Read temperatures and flow together.
  • Compare observations under relevant operating conditions.

Does the physical context explain the observed behaviour?

04

Anomaly investigation

Turn a deviation into a question.

An unusual pattern is a starting point for investigation. A symptom alone does not establish a component fault or its cause.

  • Compare a deviation with its operating context.
  • Keep observations separate from possible explanations.
  • Identify the evidence that could confirm or reject a hypothesis.

What additional evidence would distinguish the possible explanations?

05

Forecasting

Explore a possible future with context.

A forecast asks how behaviour may develop under stated conditions. Its assumptions and uncertainty belong alongside the expected trend.

  • State the conditions behind the scenario.
  • Keep the prediction horizon and uncertainty visible.
  • Revisit the interpretation when conditions change.

Under which conditions is this outlook useful?

06

Reviewed improvement

Keep the decision with the responsible person.

A proposed improvement becomes a decision through engineering review. Evidence, practical constraints and the judgement of the responsible person belong together.

  • Review the proposal against the supporting evidence.
  • Consider operating constraints and expected trade-offs.
  • Let the responsible person decide whether and how to proceed.

Is the evidence sufficient for a person to make an informed decision?