Client
Hindalco Alumina Division
Industry
Metals & Mining
Implementation
14 Weeks
Algorithmic Scheduling Unlocks 22% Higher Kiln Utilization
Inquizity implemented a hybrid AI-optimization system to automate Hindalco’s multi-grade kiln scheduling, overcoming sequence dependencies and energy inefficiencies that manual Excel planning couldn’t resolve.
Project Specifics
-
Grade Complexity
-
Energy Optimization
-
Changeover Intelligence
-
Operator Adoption
Scope
-
4 rotary kilns
-
120+ specialty alumina grades
-
9-month historical data
Technical Stack
-
MILP core with heuristic adapters
-
SAP PP-PI integration
-
Digital twin for thermal modeling
Timeline
- Design: 3 weeks
- Pilot: 6 weeks
- Scale: 5 weeks
The Challenge
Hindalco’s specialty alumina production faced critical bottlenecks:
- Sequence Chaos: 120+ product grades with complex changeover rules (e.g., 8hr refractory cooling between certain batches)
- Excel Gridlock: Monthly scheduling took 3 weeks across 4 planners, with 14+ spreadsheet versions
- Hidden Costs: Suboptimal kiln sequencing led to $1.8M/year in preventable energy waste"*
Our Approach
We co-designed a solution that respects both chemistry and operations
01
Digital Twin
Mirroring kiln thermodynamics and material properties
02
Hybrid Scheduler
MILP for optimal sequences + heuristics for operator preferences
03
Override Dashboard
Let planners adjust algorithms with tracked rationale
Project Specifics
-
Grade Complexity
-
Energy Optimization
-
Changeover Intelligence
-
Operator Adoption
Technical Deep Dive
Core Algorithms
Automatic mapping of your unique constraints – from chemical batch dependencies to union break schedules – into optimization parameters.
Integration
SAP PP-PI for order data
OSIsoft PI for real-time kiln telemetry
Change Management
12 onsite training sessions
Custom alert thresholds for metallurgists
Results
Metric |
Before |
After |
|---|---|---|
|
Kiln Utilization |
68% |
83% |
|
Scheduling Time |
21 Days |
4 Days |
|
Energy Cost/Ton |
$18.70 |
$15.90 |
