Real Data PoC — Steel Energy Decision Layer

From industrial energy data to governed decisions.

This Proof of Concept processes real industrial steel energy data through the virtauto Decision Layer and generates traceable ALLOW, HOLD and BLOCK outputs.

DATASET UCI Machine Learning Repository — Steel Industry Energy Consumption · DAEWOO Steel Co. Ltd, Gwangyang, South Korea · 35,040 data points · CC BY 4.0

Load Data

Analyze

Decision Layer

Business Case

INSTRUCTION: Download the CSV file from the UCI Repository: steel+industry+energy+consumption.zip → unzip → upload Steel_industry_data.csv here.

⚡

Upload Steel_industry_data.csv

Drag & drop or select file
35,040 rows · 9 features · approx. 2.6 MB

Loading dataset...

// Overview

Data points analyzed

—

Loading period...

Avg. consumption / interval

—

kWh per 15-minute interval

Peak consumption

—

Maximum interval value

Anomalies detected

—

Consumption > 2σ above mean

—% of windows

Idle periods est.

—

Consumption < 20% of peak

Total CO₂ equivalent

—

tCO₂ in measurement period

// virtauto Decision Layer

ALLOW windows

—

Admissible optimization windows

HOLD windows

—

Insufficient or non-admissible context

BLOCK windows

—

Maximum load or hard constraint

Estimated savings

—

Based on ALLOW windows only

// Sample Decision Trace

Load dataset to generate a real decision trace.

// Consumption Profile

// Energy consumption over time sample

// Load type distribution

// Avg. consumption by weekday

// Power factor distribution

// Weekday vs. weekend consumption

// Anomaly detection — Top 10

#	Timestamp	Consumption	Z-Score	Power Factor	Load Type	Severity

// Advisory Interpretation

VIRTAUTO ADVISORY

deterministic local summary · no backend required

Load the dataset first, then use one of the preset questions. This PoC uses local deterministic analysis for explainable advisory output.

// Savings potential calculator

Adjust parameters

Installed capacity 500 kW

Annual operating hours 6,000 h

Energy price 12 ct/kWh

Optimization potential 9%

Number of lines / sites 4 lines

Projected savings potential

Annual consumption —

Annual energy cost —

Optimization potential —

Annual savings —

CO₂ reduction est. —

virtauto ROI share est. —

Total annual savings —