Peak Shaving

The core mechanism of peak shaving involves time-shifting energy consumption from high-demand intervals to low-demand periods. This is achieved by charging Battery Energy Storage Systems (BESS) during off-peak hours when electricity prices are low and discharging them during peak windows. Unlike load shedding, which simply drops demand, peak shaving maintains operational continuity by relying on stored energy. The strategy directly targets the duck curve phenomenon, where net load drops midday due to solar generation and ramps steeply in the evening.

For commercial and industrial (C&I) consumers, a significant portion of the electricity bill consists of demand charges—fees based on the highest 15-minute average power draw during a billing cycle. Peak shaving algorithms monitor real-time load and dispatch stored energy the instant consumption approaches a predefined demand threshold, capping the maximum kilowatt draw from the grid. This can reduce demand charges by 30-70% without altering operational behavior.

Modern peak shaving relies heavily on Lithium-ion BESS due to their fast response times (sub-100ms) and high energy density. The Battery Management System (BMS) continuously monitors State of Charge (SoC) and State of Health (SoH) to ensure the system can meet the predicted peak load. Advanced systems use Model Predictive Control (MPC) to optimize discharge schedules based on load forecasts, ensuring sufficient reserve capacity is maintained for unexpected spikes.

In microgrids and off-grid sites, peak shaving coordinates with diesel or natural gas generators. Rather than sizing a generator for the absolute maximum load, the system uses a smaller generator running at its optimal efficiency point while the battery handles transient peaks. This prevents wet stacking in diesel generators—a condition caused by prolonged low-load operation—and significantly reduces fuel consumption and maintenance costs.

Unmanaged simultaneous charging of electric vehicle fleets creates massive, short-duration power spikes that can overload distribution transformers. Peak shaving algorithms dynamically adjust EVSE charging rates using Open Charge Point Protocol (OCPP) commands. By staggering charge sessions and temporarily reducing amperage during the facility's non-EV peak loads, the system prevents transformer overload and avoids costly infrastructure upgrades.

Beyond behind-the-meter savings, aggregated peak shaving resources can participate in wholesale markets. A Virtual Power Plant (VPP) orchestrates distributed batteries to reduce load simultaneously, providing capacity services to grid operators. This transforms a cost-saving measure into a revenue stream. The dispatch must be highly reliable, often requiring sub-second response to automated generation control (AGC) signals to qualify for frequency regulation markets.

A financial optimization strategy directly enabled by peak shaving. Commercial and industrial electricity tariffs often include demand charges based on the highest 15-minute average power draw during a billing cycle. A single spike can represent 30-70% of a monthly bill. Demand charge management uses battery energy storage or load curtailment to cap this peak, delivering immediate cost reduction without necessarily reducing total energy consumption.

The primary hardware asset for modern peak shaving. A BESS charges during off-peak, low-cost periods and discharges during peak demand windows. Key performance parameters include:

• C-Rate: Determines how quickly the battery can discharge relative to its capacity
• Round-trip efficiency: Typically 85-95% for lithium-ion systems
• Depth of Discharge (DoD): Dictates usable capacity without accelerating degradation Utility-scale BESS installations are increasingly co-located with EV fleet depots to manage coincident charging peaks.

An advanced control algorithm that optimizes peak shaving decisions in real time. Unlike simple rule-based controllers, MPC solves a finite-horizon optimization problem at each time step, incorporating:

• Forecasted load profiles
• Time-of-use electricity prices
• Battery state of charge constraints
• Degradation costs This predictive capability allows the system to pre-charge storage before anticipated peaks, rather than reacting after the spike has already incurred demand charges.

A real-time power allocation algorithm that prevents local circuit overloads during peak shaving events. When multiple EV chargers share a constrained electrical panel, dynamic load balancing distributes available capacity based on:

• Individual vehicle State of Charge (SoC)
• Driver departure time priorities
• Transformer thermal limits This ensures peak shaving at the building level doesn't inadvertently trip branch circuit breakers or exceed cable ampacity ratings.