ADAPT

Peak Shaving and Demand Response

Peak shaving optimization reduces electricity costs by up to 30% through optimal battery storage scheduling that minimizes demand charges and shifts consumption from peak to off-peak periods.

Understanding the Problem

Peak shaving reduces peak demand charges (often 30-70% of commercial electricity bills) by optimally scheduling battery storage and load shifting. The problem formulates as MILP optimization determining when to charge/discharge batteries, shift flexible loads, and curtail demand while respecting battery capacity, power limits, and operational constraints. Classical MILP solvers (Gurobi, CPLEX) provide exact solutions for typical building-scale problems. Quantum computing exploration remains early with 5-10 year horizon due to current hardware limitations, but problem structure (mixed-integer quadratic) positions it for future quantum advantage as technology matures.

THE CHALLENGE

What Makes it Hard

Optimizing battery energy storage system (BESS) charge/discharge schedules and controllable load management to minimize peak demand charges and total electricity costs while maintaining operational requirements and battery health.

WHO FACES IT

Facility managers at commercial and industrial sites seeking 20-30% electricity cost reductionEnergy managers optimizing battery storage systems to maximize demand charge savings and ROIUtility demand-side management programs coordinating customer load shifting for grid stabilitySustainability officers using peak shaving to enable renewable integration and reduce carbon footprintBuilding operators in high-demand-charge markets (manufacturing, data centers, cold storage)

Multi-period optimization balancing immediate cost savings against battery degradation and future operational needs

Real-time operational constraints including battery state-of-charge limits, power capacity, and minimum run times

Uncertainty in demand forecasting, renewable generation, and electricity pricing requiring robust or stochastic optimization

BUSINESS IMPACT

Peak shaving delivers 20-30% electricity cost reduction with 3-5 year payback, targeting demand charges that represent 30-70% of commercial bills.

Demand charges can account for 30-70% of commercial electricity bills.

Demand Charges

30-70%

Of C&I Bills^[1]

Demand charges can account for 30-70% of commercial and industrial electricity bills.

Cost Reduction

20-30%

Energy Savings^[2]

Businesses implementing peak shaving with battery storage achieve 20-30% energy cost reductions.

ROI Timeline

3-5yr

Payback Period^[3]

Typical C&I battery storage systems pay for themselves in 3-5 years through reduced demand charges.

How We Solve It

Classical optimization formulates peak shaving as MILP: continuous variables for battery charge/discharge power, binary variables for operational decisions, constraints on battery capacity/power limits, and objective minimizing demand charges plus energy costs. MILP solvers like Gurobi and CPLEX solve typical building-scale problems (15-minute intervals, 24-96 hour horizons) in seconds. Stochastic programming or robust optimization handles demand uncertainty. Model Predictive Control (MPC) enables real-time receding horizon optimization.

Heterogeneous
Hybrid Compute

What We Bring

MILP optimization for battery scheduling with demand charge minimization using Gurobi, CPLEX, or open-source solvers

Stochastic programming incorporating demand forecast uncertainty and renewable generation variability

Model Predictive Control (MPC) for real-time receding horizon optimization adapting to actual conditions

Multi-objective optimization balancing cost reduction, battery health preservation, and operational constraints

FUTURE POSSIBILITIES

The
Quantum Horizon

Quantum computing shows theoretical promise for large-scale peak shaving with complex constraints but remains 5-10 years from practical advantage. Current classical MILP solvers handle production-scale problems efficiently, limiting near-term quantum value proposition.

Exploratory Work

Near-term (2025-2027) quantum advantage unlikely for peak shaving due to efficient classical solvers and limited problem scale (hundreds to thousands of variables vs. millions needed for quantum advantage). Medium-term (2027-2030) possible advantage for very large multi-building or microgrid optimization coordinating thousands of batteries and flexible loads. Long-term (2030+) fault-tolerant quantum computers may enable real-time stochastic optimization at utility scale with massive uncertainty scenarios. Current quantum hardware (5,000 qubits) insufficient for practical peak shaving optimization. Organizations should focus on proven classical MILP approaches (Gurobi, CPLEX) for immediate ROI, monitor quantum developments for future exploration, and participate in academic research if strategically aligned. Quantum realism score: 2/5 (5-10 year horizon, limited near-term practical advantage).