RS monogramRussell Schmidt
Lightbox image, just a zoomed in version of the last picture. Hit Escape to exit and return to the last page.

Power Management Algorithms for EV Charging Sites Part 1

Versions will be posted on Halloween 2025 Substack and Medium

Power Management Algorithms for EV Charging Sites: A Technical Deep Dive

When deploying EV charging infrastructure, one of the most critical—yet often overlooked—decisions is choosing the right power allocation algorithm. With electrical service upgrades expensive and time-consuming, most charging sites operate under power limits that require intelligent, automated distribution across multiple charging sessions.

The Power Constraint Problem

Consider a common scenario: a location has 100kW electrical service supporting six 30kW DC fast chargers. When all six ports are occupied, naive power distribution would attempt to deliver 180kW — nearly double the available capacity. The charging management system must decide how to allocate the available 100kW across active sessions.

The algorithm choice significantly impacts user experience, energy throughput, and operational efficiency. Let's examine the main approaches.

1. Balanced (Equal) Charging

Algorithm: Divide available power equally among all active charging sessions.

Implementation:

Available Power ÷ Number of Active Sessions = Power per Session
100kW ÷ 6 sessions = 16kW per session

Advantages:

  • Fairness: Every user receives identical treatment
  • Simplicity: Easiest to implement and explain; sessions start when vehicles plug in
  • Predictability: Users can estimate charging time upon arrival
  • Optimal for long dwell times: Works well when vehicles have hours to charge

Disadvantages:

  • Inefficient at low power limits: Higher percentage of capacity lost to transactional friction (resistance, heat, onboard components) when power is divided among many sessions
  • Suboptimal charging curves: EVs charge most efficiently at higher power levels early in the session
  • Underutilized capacity: If some vehicles can't accept the allocated power (due to battery limitations), that power goes unused

Best Deployment Scenarios:

  • Workplace charging where fairness is paramount and dwell times are long (6+ hours)
  • Retail locations with similar dwell times and relatively high power limits
  • Sites with homogeneous vehicle types fleet charging, especially during peak times to manage energy costs
  • High-capacity sites where individual allocations remain above 10-15kW

2. First In, First Out (FIFO)

Algorithm: Allocate maximum available power to sessions in order of arrival.

Implementation:

Session 1: 30kW (or vehicle maximum)
Session 2: 30kW (or remaining capacity)
Session 3: 30kW (or remaining capacity)
Sessions 4-6: 0kW (queued)

Advantages:

  • Optimal charging efficiency for vehicles receiving power
  • Clear queue management with predictable wait times
  • Maximizes early session charging speeds when batteries accept highest power

Disadvantages:

  • Communication critical: Risk of users experiencing "plug in and nothing happens" without clear queue status, leading to repeated plug cycling attempts
  • Potential for gaming early arrival for preferential treatment; stopping other sessions to advance in the queue
  • Inefficient utilization depending on implementation details, you can see unused capacity as in the example above (10kW )

Best Deployment Scenarios:

  • Highway corridor charging where speed is critical
  • Fleet depots with scheduled charging windows
  • High-utilization sites where queuing is acceptable

3. Last In, First Out (LIFO)

Algorithm: Prioritize most recently arrived sessions for power allocation.

Implementation:

Newest session receives maximum power
Older sessions receive reduced or suspended power

Advantages:

  • Optimal for long-distance travel: Newer arrivals likely have lowest state of charge and highest energy needs
  • Immediate service for new arrivals
  • Optimizes for charging curve More energy allocated to vehicles with presumably lower state of charge

Disadvantages:

  • Extreme unfairness to early arrivals in mixed-use scenarios
  • Communication critical: Same risk as FIFO—users may experience "plug in and nothing happens" without proper status indication
  • Potential for indefinite delays for early sessions
  • Encourages poor user behavior (repeated reconnection attempts)

Best Deployment Scenarios:

  • Fleets with large variances between routes optimized for charging curve efficiency around staggered arrivals
  • Possible for emergency prioritization scenarios Gets vehicles out of low state of charge faster

4. Ranked Order (Priority-Based)

Algorithm: Assign power based on predetermined user or session priorities.

Priority Factors:

  • User subscription level (premium vs. standard)
  • Vehicle type (fleet vs. personal)
  • Historical usage patterns
  • Payment method or membership status
  • 3rd party software dynamically managing charging

Implementation:

Priority 1 users: Maximum available power
Priority 2 users: Remaining power allocation
Priority 3 users: Minimal or no power

Advantages:

  • Revenue optimization through tiered service levels
  • VIP customer satisfaction
  • Fleet software integration: Enables third-party fleet management systems to control charging order and priorities
  • Operational control for site owners

Disadvantages:

  • Difficulty implementing with L2s most Level 2 chargers cannot identify vehicles; requires vehicle telematics
  • Potential management burden managing tiers, levels
  • Additional point of failure if reliant on 3rd party software to manage charging order
  • Risk of creating "second-class" user experience

Best Deployment Scenarios:

  • CPOs with membership tiers capture some value from premium users
  • Mixed public-private networks favor company vehicles over visitors
  • Fleet operations with telematics integration enabling automated priority management
  • Hotel or resort charging with guest prioritization
  • Corporate campuses with employee vs. visitor priority

5. Departure Time Optimization

Algorithm: Allocate power to minimize total site dwell time or meet user-specified departure deadlines.

Implementation:

User inputs desired departure time and target charge level
Algorithm calculates minimum power needed to meet targets
Allocates power to most time-constrained sessions first

Advantages:

  • User-centric approach with explicit time management
  • Optimal site throughput when combined with accurate departure predictions
  • Reduces overstaying by completing charges near departure time
  • Utility integration: Ideal for demand response programs and time-of-use optimization when departure times are flexible
  • Perfect for long dwell times: Works best when vehicles have extended parking periods

Disadvantages:

  • Difficulty implementing with L2s most Level 2 chargers cannot identify vehicles; requires vehicle telematics
  • Requires user input accuracy (often unreliable)
  • Complex optimization calculations
  • Vulnerable to gaming (false urgent departure times)

Best Deployment Scenarios:

  • Fleet charging with overnight dwell times and preconditioning enable overnight charging and then stopping and restarting sessions ahead of departure to maximize range
  • Workplace charging with known work schedules and utility partnerships; optimimze charging around shift change and time of use
  • Utility demand response programs where charging can be shifted to off-peak hours or discharging scheduled during peak demand

6. Dynamic Load Balancing

Algorithm: Continuously adjust power allocation based on real-time vehicle acceptance rates and charging curves.

Implementation:

Monitor actual power consumption vs. allocated power
Redistribute unused power to sessions that can accept it
Adjust allocation as vehicles move through charging curve phases

Advantages:

  • Maximum utilization of available power capacity
  • Adapts to vehicle charging characteristics
  • Improves overall site efficiency

Disadvantages:

  • Not as useful for L2s lack of dynamism in L2 charging curves
  • Complex implementation requiring real-time monitoring
  • Frequent power fluctuations may confuse users
  • Requires sophisticated communication with vehicles

Best Deployment Scenarios:

  • High-utilization DC charging sites where efficiency is critical
  • Mixed-vehicle environments with varying charging capabilities
  • Sites with expensive demand charges requiring precise power management
  • Sites with microgrids where a controller is managing a site's power

7. Predictive Algorithms

Algorithm: Use machine learning to predict session duration and optimize power allocation accordingly.

Implementation:

Analyze historical patterns: vehicle type, arrival time, user behavior
Predict session duration and power requirements
Allocate power to maximize site throughput or revenue

Advantages:

  • Anticipatory optimization based on predicted behavior
  • Continuously improving through machine learning
  • Can optimize for multiple objectives (throughput, revenue, satisfaction)
  • Works best with stable patterns: Ideal when operations don't expect radical surprises or usage pattern changes

Disadvantages:

  • Requires significant historical data
  • Complex implementation and maintenance
  • Predictions may be inaccurate for unusual sessions or during pattern disruptions

Best Deployment Scenarios:

  • Mature sites with extensive usage data
  • Network operations with cross-site learning
  • High-value locations justifying algorithmic complexity

8. Highest SOC First

Algorithm: Prioritize vehicles with the highest current state of charge for power allocation.

Implementation:

Query vehicle SOC levels via communication protocols
Allocate maximum power to vehicles with highest charge levels
Lower SOC vehicles receive reduced or no power

Advantages:

  • Fleet readiness optimization: Ensures some vehicles reach 100% charge quickly for immediate deployment
  • Emergency preparedness: Maintains a subset of fully-charged vehicles for critical missions
  • Faster session completion: High SOC vehicles finish charging sooner, freeing up capacity

Disadvantages:

  • The "oligarchy algorithm": Gives power to those who already have the most, creating inequity
  • Poor utilization of charging curves: Misses the high-efficiency early charging phase for low SOC vehicles
  • Potential for stranded low SOC vehicles if power remains limited
  • Difficulty implementing with L2s most Level 2 chargers cannot identify SOC; requires vehicle telematics

Best Deployment Scenarios:

  • Emergency services requiring rapid fleet readiness
  • Islanded grid scenarios with limited energy budgets requiring minimum fleet availability
  • Critical infrastructure where some vehicles must maintain high readiness levels
  • Military or first responder fleets with emergency deployment requirements

9. Lowest SOC First

Algorithm: Prioritize vehicles with the lowest current state of charge for power allocation.

Implementation:

Query vehicle SOC levels via communication protocols
Allocate maximum power to vehicles with lowest charge levels
Higher SOC vehicles receive reduced or no power

Advantages:

  • The "socialist algorithm": Allocates energy to vehicles most in need
  • Optimal charging efficiency: Takes advantage of high power acceptance at low SOC levels
  • Maximizes daily energy throughput: Vehicles with highest intraday usage get priority
  • Prevents stranded vehicles: Ensures no vehicle is left with insufficient charge

Disadvantages:

  • Delayed completion for vehicles approaching full charge
  • Potential inefficiency if low SOC vehicles have limited power acceptance
  • Complex SOC monitoring requirements across the fleet
  • Difficulty implementing with L2s most Level 2 chargers cannot identify SOC; requires vehicle telematics

Best Deployment Scenarios:

  • Fleet operations with varying daily usage patterns
  • Shared mobility services where vehicle availability is critical
  • Delivery, rideshare, autonomous fleets with high daily energy consumption and ability to take advantage of dwell times for cleaning
  • Transit systems requiring equitable charge distribution across vehicles

Hybrid Approaches

Most real-world deployments combine multiple strategies:

Time-Based Hybrid:

  • Balanced charging at low kW during peak hours for cost optimization
  • FIFO allocation during off-peak for fastest charging

Tiered Service:

  • Premium users get priority allocation
  • Standard users receive balanced sharing

Adaptive Algorithms:

  • Start with balanced allocation
  • Switch to FIFO when site utilization exceeds threshold

Selection Criteria

Choose algorithms based on:

Site Characteristics:

  • Usage patterns (peak vs. distributed)
  • User types (fleet vs. public)
  • Electrical constraints
  • Physical layout

Business Objectives:

  • Revenue optimization
  • User satisfaction
  • Network throughput
  • Operational simplicity

Technical Capabilities:

  • Communication infrastructure
  • Real-time monitoring capabilities
  • Integration with payment systems
  • Regulatory compliance requirements

Implementation Considerations

Communication Requirements:

  • Vehicle-to-infrastructure for real-time power adjustment
  • User interface for departure time input
  • Backend systems for priority management

Fallback Strategies:

  • Default to balanced allocation if primary algorithm fails or communications disrupted
  • Manual override capabilities for operators
  • Emergency prioritization for critical users

Performance Monitoring:

  • Track algorithm effectiveness through key metrics
  • User satisfaction surveys
  • Energy throughput analysis
  • Revenue per charging session

Conclusion

There's no universal "best" charging algorithm—the optimal choice depends on specific site characteristics, user needs, and business objectives. The choice between "oligarchy" (Highest SOC First) and "socialist" (Lowest SOC First) algorithms particularly highlights how technical decisions reflect operational philosophies and fleet management strategies.

Successful deployments often start with simpler approaches (balanced or FIFO) and evolve toward more sophisticated algorithms as usage patterns become clear and technical capabilities mature. SOC-based algorithms require vehicle communication capabilities but offer powerful fleet optimization when that infrastructure is available.

The key is matching algorithm complexity to operational requirements while maintaining user experience standards. As EV adoption accelerates and charging sites become more constrained, intelligent power management will increasingly differentiate successful charging networks from those struggling with user satisfaction and utilization efficiency.

Consider starting with balanced allocation for predictability, but build infrastructure to support more sophisticated algorithms as your network grows and usage patterns emerge. The charging algorithm may seem like a technical detail, but it's actually a critical product decision that shapes every user interaction with your infrastructure.