Four exact formats
B1 QTB, B1 Hot, B2 QTB, and B2 Hot sheets with decimal and unit validation.
Production calculator preview
Auditable equation effect
ΔYP and ΔTS are shown together as exact table values. Chemistry uses the corresponding hot-rolling equation, while cooler and speed effects use the selected QTB equation. The local app separately tests these relations against close actual production cases.
Exact equation values
Editable engineering summary
Equation effects, future local learning, and your manual engineering adjustment remain separate. In this safe public preview, manual values are stored only in this browser; production learning remains in the local app.
Hot-rolling equation source
| Element | Increment | Equation ΔYP | Equation ΔTS | Learned ΔYP | Learned ΔTS | Manual YP | Manual TS | Final ΔYP | Final ΔTS | Why |
|---|
Size-wise QTB effect
| Size | Effect | Basis | Equation ΔYP | Equation ΔTS | Learned ΔYP | Learned ΔTS | Manual YP | Manual TS | Final ΔYP | Final ΔTS | Why |
|---|
Same cooler count · size-wise
| Size | Reference speed | Decrease | Equation ΔYP | Equation ΔTS | Learned ΔYP | Learned ΔTS | Manual YP | Manual TS | Final ΔYP | Final ΔTS | Why |
|---|
Auditable baseline
The temporary calculator uses these unchanged baseline equations. In the local app, a new calculator workbook can be uploaded and is activated only after its YP/TS formulas reproduce Excel's saved results and pass range checks.
Protected local workflow
This prevents a public link from exposing production spreadsheets, measured heats, or write access to your model.
B1 QTB, B1 Hot, B2 QTB, and B2 Hot sheets with decimal and unit validation.
Suspicious results are held for review instead of silently influencing the model.
Accepted actual results make small corrections and batch candidates must pass repeated validation—never a sudden jump.
Element and process relations use close comparison cases and report when evidence is insufficient.
All future equation, regression, chart, and format changes will be made in the local project. A new preview can be mounted only when you request it.
Deviation transparency preview
The local app creates this explanation for each saved actual result. This public page shows the layout only; it contains no production record and saves nothing.
Illustrative feedback case
The point is checked for units, physical limits, and residual outliers. If accepted, one result can nudge the reference by no more than 1.5 MPa. A wider change waits for 20 accepted points, repeated held-out validation, and a 10% blend.
Common calibration shift · medium evidence. YP and TS moved together, so a general testing offset or an unrecorded production change is more plausible than immediately changing one chemistry coefficient.
Edge-of-data check · low evidence. Inputs near the observed limits are flagged as extrapolation risks, not claimed as causes.