State-of-Charge Scheduling Under Time-of-Use and Ancillary Service Constraints

🎯 Learning Objectives

✓ Calculate optimal SoC trajectories using linear programming under TOU pricing and regulation reserve requirements
✓ Design a 15-minute resolution SoC schedule for a 2 MW / 4 MWh BESS co-aggregated with 3 MW solar PV under CAISO market rules
✓ Analyze trade-offs between revenue from energy arbitrage versus ancillary service penalties due to SoC deviation
✓ Explain how battery degradation models constrain feasible SoC excursions in multi-day scheduling horizons
✓ Apply IEEE 1547-2018 and FERC Order No. 2222 compliance checks to a proposed SoC schedule

📖 Why This Matters

In mining operations, remote sites increasingly deploy solar + battery microgrids to replace diesel generators — but without smart SoC scheduling, batteries waste opportunity: charging during expensive peak hours or missing high-value ancillary service payments. This lesson bridges blasting engineers’ understanding of *system reliability* and *timing precision* (like delay sequencing in blast design) to the equally critical timing discipline required in BESS scheduling — where milliseconds matter for frequency response and hours matter for cost-optimal charging.

📘 Core Principles

SoC scheduling rests on three interdependent layers: (1) Physical layer — battery dynamics governed by power balance (dSoC/dt = -P_batt/(E_max·η)), capacity fade, and thermal limits; (2) Market layer — TOU rate structures (e.g., CAISO’s 5-tier ‘Super Off-Peak’ to ‘Peak’ periods) and ancillary service performance metrics (e.g., NERC BAL-003-1 accuracy thresholds); (3) Aggregation layer — how DERMS platforms translate aggregated BESS+PV signals into ISO-compliant resource adequacy commitments. Crucially, unlike static blast design, SoC scheduling is *receding-horizon*: decisions today affect flexibility tomorrow — analogous to burden-spacing interactions across successive blast rounds.

📐 Optimal SoC Update Equation

The discrete-time SoC evolution model forms the backbone of all scheduling formulations. It links power dispatch to state evolution while embedding efficiency and capacity constraints.

💡 Worked Example

Problem: A 2.5 MW / 5 MWh lithium-iron-phosphate battery starts at 60% SoC at 6:00 AM. It discharges at 2.0 MW for 90 minutes under 92% round-trip efficiency. What is its SoC at 7:30 AM?

1. Step 1: Convert time to hours → 90 min = 1.5 h

2. Step 2: Compute energy discharged → E_out = P × t = 2.0 MW × 1.5 h = 3.0 MWh

3. Step 3: Account for efficiency loss → Effective energy depletion = E_out / η = 3.0 / 0.92 ≈ 3.26 MWh

4. Step 4: Initial stored energy = 0.60 × 5 MWh = 3.0 MWh

5. Step 5: Final stored energy = 3.0 − 3.26 = −0.26 MWh → violates physical limit → SoC clamped at 0%

Answer: The SoC reaches 0% at ~7:23 AM; full depletion occurs before 7:30 AM, confirming the need for constraint-aware scheduling.

🏗️ Real-World Application

At Newmont’s Boddington Gold Mine (Western Australia), a 12 MW / 24 MWh BESS co-aggregated with 22 MW solar uses SoC scheduling to meet both site load and WA’s Wholesale Electricity Market (WEM) contingency FCAS requirements. The scheduler prioritizes SoC maintenance between 25–75% during daylight to preserve headroom for fast up-regulation bids, while shifting 85% of charging to off-peak (11 PM–6 AM) — reducing grid import costs by AUD $1.2M/year and earning AUD $480k/year in FCAS payments. Violations of SoC deadband (±2% from target) trigger automatic rebalancing via curtailed solar export — mirroring how misaligned delay timings in multi-row blasts cause premature confinement loss.

✏️ Student Exercise

Given: A 1 MW / 2 MWh BESS (η = 94%, SoC_min = 10%, SoC_max = 90%) co-located with 1.5 MW solar (forecast: 0.8 MW avg 9–11 AM; 0 MW 1–3 PM). TOU rates: $28/MWh (off-peak), $112/MWh (peak), $62/MWh (shoulder). Ancillary service contract requires ≥80% SoC at 3 PM to bid into 4 PM regulation market. Task: Construct a 4-hour SoC schedule (9 AM–1 PM, 15-min intervals) maximizing net value. Assume solar curtailment incurs $0 penalty; grid import/export has no congestion cost.