Peak Shaving

The primary financial driver of peak shaving is the reduction of demand charges—fees levied by utilities based on the highest average power draw (in kW) over a 15-30 minute interval during a billing cycle. For a fleet operator, a single spike in simultaneous charging can set this charge for the entire month.

• Mechanism: The Battery-Aware Scheduler caps the aggregate fleet charging load below a configurable kW threshold.
• Impact: A 500 kW peak reduction can save $5,000–$15,000 monthly in commercial tariffs.
• Trade-off: Requires over-provisioning of charging time to spread loads, which may reduce asset utilization if not optimized.

Peak shaving algorithms align charging events with off-peak tariff windows when electricity prices are lowest. This is distinct from simple scheduling; the system dynamically forecasts task demand and state of charge to ensure operational readiness is maintained while shifting non-urgent charging to midnight or early morning slots.

• Dynamic Pricing: Integrates real-time utility API feeds to adjust to price spikes.
• Constraint: Must respect Minimum Charge Thresholds to avoid stranding agents.
• Example: A warehouse fleet shifts 80% of its energy consumption from a 2:00 PM peak ($0.18/kWh) to a 2:00 AM trough ($0.06/kWh).

Beyond cost, peak shaving prevents local transformer overload and voltage sag in facilities with limited electrical infrastructure. When a fleet of fast-charging AMRs connects simultaneously, the instantaneous inrush current can trip breakers or exceed the facility's sanctioned load limit.

• Active Load Management: The orchestrator staggers charge initiation by milliseconds to avoid inrush overlap.
• Infrastructure Deferral: Effective peak shaving can postpone costly electrical panel upgrades (often $100k+).
• Relationship: Works in tandem with Load Shifting to create a flat, predictable facility load profile.

Modern peak shaving relies on predictive load forecasting rather than simple static caps. The system uses historical data and operational schedules to predict when a natural peak is likely to occur and pre-emptively discharges or idles non-critical loads.

• Baseline Model: A rolling average of facility load establishes the threshold that triggers shaving.
• Machine Learning: Regression models predict peak probability based on shift changes and order volume.
• Pre-cooling Analogy: Similar to thermal storage, the system "pre-charges" agents aggressively just before a predicted peak window closes.

Aggressive peak shaving often forces higher C-Rates during narrow off-peak windows, which accelerates Battery Degradation. A sophisticated peak shaving engine co-optimizes for both tariff avoidance and battery longevity.

• Cost Function: The optimizer weighs the $/kWh savings against the $/cycle degradation cost.
• Thermal Constraints: Shaving logic respects Battery Thermal Models to prevent derating during fast off-peak charging.
• Outcome: Avoids the trap of saving $500 on electricity while causing $1,000 in premature battery wear.

In advanced deployments, peak shaving extends to Vehicle-to-Grid (V2G) discharge, where idle fleet vehicles export power back to the facility during critical peak intervals. This transforms the fleet from a passive load into an active distributed energy resource.

• Bi-directional Chargers: Required hardware to invert DC battery power to AC grid power.
• State of Energy (SoE) Logic: Discharge is halted at a strict Energy Buffer to preserve mission capability.
• Revenue Stream: Participation in utility demand response programs generates direct payments for exported kW.

A complementary energy management strategy that moves fleet charging operations from periods of high electricity cost or demand to periods of lower cost or higher renewable energy availability. While peak shaving focuses on reducing the maximum power draw at any single moment, load shifting is concerned with when energy is consumed. Together, they form the foundation of demand-side management. A fleet might shift charging to overnight hours (load shifting) and then stagger individual charge sessions within that window to avoid a new peak (peak shaving).

An optimization routine that determines when, where, and for how long each agent in a fleet should charge to meet operational demands while respecting constraints. Key inputs include:

A mathematical component within a scheduling optimizer that assigns a cost value to energy consumption. For peak shaving, this function heavily penalizes consumption during high-tariff periods and when aggregate fleet demand approaches a facility's demand charge threshold. The function incorporates time-of-use electricity rates, demand charges measured in $/kW, and battery degradation costs to produce a total cost signal that the optimizer minimizes.

A defined period of time, often dictated by energy tariffs or facility schedules, during which an agent is permitted or scheduled to connect to a charging station. Peak shaving algorithms use charging windows as soft or hard constraints, dynamically adjusting which agents charge within a given window to keep total instantaneous power draw below the target peak. Narrow windows increase scheduling complexity but offer greater cost control.

An optimization engine that finds feasible schedules and routes for a fleet by treating battery capacity, charge rates, and station availability as hard constraints in a mathematical programming model. For peak shaving, the solver adds a maximum aggregate power constraint across all active chargers. Common approaches include mixed-integer linear programming (MILP) and constraint programming (CP), which can guarantee that no schedule violates the facility's peak power limit.

The algorithmic process of prioritizing and sequencing multiple agents waiting for access to a limited number of charging stations. During peak shaving events, the queue manager may defer low-priority agents or reduce charge rates across active sessions to stay under the power threshold. Policies include: