What assumptions sit behind the temperature forecast?
Review recent measurements and a future cooling-command scenario. See where input quality, timing or an unknown command prevents a numerical illustration.
The bundled model was trained on a synthetic reactor. Its outputs are conditional illustrations and remain unqualified for a field process. Agreement between predictors cannot establish sensor calibration or permission to act.
Review a scenario
Pasted inputs are processed in this browser. The authored app has no capture-upload endpoint. Use the offline kit for original files or sensitive records. Examples are synthetic; no vendor export or operating-site use has been validated.
Preparing local review…
Input-record review
| Sample | Record review | Reasons |
|---|
Conditional outputs and unknown outcomes
Times are measured from the last source sample. The first output estimates the next source tick. A scenario command affects the following output. Predictor difference is descriptive; it is not a confidence interval or error bound.
| Time | Neural illustration | Linear illustration | Input basis | Actual outcome |
|---|
Hashes, declarations and report assumptions
Inspect or edit the capture and profile
Map five completed source pairs: temperature in K and recorded cooling command as a fraction from 0 to 1. Keep source and receipt timestamps distinct. A future planned or assumed command is not delivery/readback evidence. Edits invalidate the displayed worksheet.
Pasted editor text is reviewed with LF line endings.
Keep the evidence separate
1 / Usable records
Quality, unit, source, clock and age checks use your entered mapping. Passing them does not authenticate the source or calibrate the sensor.
2 / Conditional illustration
Known scenario inputs allow mathematical propagation. An unknown command stops subsequent outputs. The future process outcome stays unknown.
3 / Qualified model
The bundled synthetic model has no verified field-process qualification. A different process needs its own data, identification and engineering acceptance.
Why the biased example matters
The retained study includes a temperature bias that gives both predictors plausible inputs. The worksheet cannot recover true temperature from their agreement. Keep original exports and independent calibration evidence with the engineering handover.
The comparison also exposed unstable online adaptation. This worksheet uses fixed weights and a linear reference. It has no adaptation, command selection or plant write path.
Inspect the synthetic window mapping ↗