Energy as a hidden variable
For secondary steel producers, energy is the largest variable cost — often exceeding 30% of total production cost. Yet most plants track energy consumption at the monthly utility bill level. By the time the bill arrives, the decisions that drove consumption are weeks old. There is no feedback loop, no ability to connect a specific heat or rolling pass to its energy cost.
This plant had smart meters installed on all major equipment — electric arc furnaces, rolling mills, compressors, and lighting circuits — but the data sat in a standalone SCADA dashboard that nobody in production or finance ever looked at. The ERP and the energy data lived in separate worlds.
Bridging SCADA and ERP
The integration was simpler than expected. The SCADA system exposed a Modbus TCP interface. An IoT edge agent polled meter readings every five minutes and wrote them as structured records into the ERP's asset module. Each meter reading is linked to the active production order running on that equipment at that timestamp.
- Per-heat energy cost — every electric arc furnace heat tagged with its kWh and equivalent cost
- Shift comparison dashboard — shift supervisors see energy intensity vs. production output in real time
- Peak demand alerts — automatic notification when combined demand approaches the contracted peak limit
- Idle consumption tracking — energy drawn during non-production periods flagged separately
- Monthly variance report — planned vs. actual energy cost by cost centre, down to individual equipment
Where the savings came from
The first month of data revealed three patterns nobody had seen before. First, the night shift consistently ran 18% higher energy intensity per tonne than the day shift — traced to a practice of running furnaces at higher power settings to compensate for cooler ambient temperatures. Second, three compressors were running continuously during lunch breaks when production was stopped. Third, peak demand was being triggered by a scheduling pattern that could be shifted by forty minutes without affecting output.
We did not need more energy. We needed to see where we were wasting it. The data was always there. We just never connected it to anything useful.
Native IoT, no third-party middleware
The platform's native IoT connector handled the Modbus integration without any additional software purchase. The edge agent runs on a ₹4,000 industrial Raspberry Pi mounted in the switch room. All analytics run inside the ERP's existing reporting layer. The plant did not buy a separate energy management system — they used the 100+ modules already in the platform they were paying for.