Praficy Consultant

We are seeking a skilled Data Analytics & Machine Learning Engineer to lead the development of a data-driven early warning and analytics system for detecting and mitigating bridging of scrap metal during the melting process in induction furnaces.

The solution will leverage GE Proficy tools (Operations Hub, Historian, and CSense) to ingest, visualize, and analyze process and batch data. The goal is to reduce downtime, improve energy efficiency, and scale the solution across multiple furnaces within the organization.

1. Pilot Implementation

Initiate proof-of-concept on Induction Furnace #1 (Johnstown plant), known for high bridging frequency.

2. Data Acquisition & Integration

Configure GE Proficy Historian to collect time-series data (voltage, current, power, frequency, ground faults).

Integrate batch metadata from M3, DASH, or similar systems (Batch ID, timestamps, chemistry, operator details, logs).

3. Data Processing & Alignment

Align batch records with furnace data.

Apply sequencing logic using refractory liner heat counts for correct chronological ordering.

4. Data Collection Strategy

Collect datasets from:

30 regular heats and 30 sintering heats from Furnace #3.

Known bridged heats for pattern recognition.

Differentiate sintering heats due to unique thermal characteristics.

5. Heat Segmentation

Segment melting cycles into phases (Charging, Melting, Superheating, Tapping).

Use electrical signal patterns and timing markers.

6. Data Cleaning & Preprocessing (GE CSense)

Perform noise filtering, outlier handling, missing data interpolation.

Apply preprocessing specific to heat types (regular, sintering, bridged).

7. Machine Learning Development

Build predictive ML models in GE CSense:

Model A 30 regular heats

Model B 30 sintering heats

Identify early warning signs of bridging via time-series behavior and chemistry patterns.

8. Furnace Electrical Behavior Analysis

Study electrical parameter trends across furnace types and chemistries.

Calibrate models for generalization across plants.

9. Early Warning & Real-Time Alerting

Define model-driven or threshold-based logic for real-time bridging risk alerts.

Tailor alerts by heat type for operator actionability.

10. Root Cause Analysis

Correlate bridging incidents with electrical anomalies, chemistry, timing irregularities, and operator behavior.

11. Actionable Recommendations

Provide real-time and data-backed recommendations for operational improvements.

Include chemistry adjustments and sintering heat management.

12. Visualization & Reporting

Develop dashboards in Operations Hub or Power BI for:

Real-time metrics

Historical trends

Alerts

Batch chemistry & process phases

13. Scalability & Documentation

Document all data flows, model logic, assumptions, and procedures.

Build framework to roll out solution across furnaces with local customization.

14. Energy Consumption Analysis

Track & analyze energy usage by phase and batch type.

Identify inefficiencies to support cost savings and sustainability initiatives.

Proven experience with GE Proficy Historian, Operations Hub, and CSense.

Strong background in machine learning, data preprocessing, and signal analysis.

Understanding of induction furnace operations or similar industrial processes.

Expertise in data integration, visualization tools (Power BI preferred), and root cause analysis.

Familiarity with batch processing, manufacturing data systems (M3, DASH), and time-series data.