Published: June 2026 Technical Level: Advanced Category: Power Systems Design
The interaction between NEC Article 625 load management provisions and the practical engineering of large-scale EV charging installations determines whether a major EVSE deployment can be accommodated within existing electrical infrastructure or requires a utility service upgrade. NEC 625.42 permits the feeder and service load calculation for EVSE to be based on the managed load limit established by an automatic load management system, rather than the sum of EVSE nameplate ratings — a provision that can reduce the calculated service load by 60 to 90 percent for large installations where vehicles typically remain connected for many hours and require only a fraction of their maximum charging rate to meet their departure energy requirement. This paper develops the NEC 625.42 load management methodology in detail, the technical requirements for an automatic load management system that qualifies for the calculated-load credit, the smart charging control architectures that satisfy these requirements, and the engineering analysis that determines the optimal managed load limit for a given installation — balancing the infrastructure cost savings from a lower managed load against the probability that the system will fail to meet some drivers' charging requirements during high-demand periods.
The fundamental tension in commercial EV charging design is between the nameplate capacity required to serve all vehicles at their maximum charging rate simultaneously and the actual average power demand that results when vehicles are managed to charge at rates that meet their departure time and energy requirements without exceeding the available electrical capacity. A parking garage with 200 Level 2 EVSE ports at 7.2 kW each has a nameplate connected load of 1,440 kW — a demand that would require a large utility service upgrade and several years of lead time in most jurisdictions. But if the typical vehicle is parked for 8 hours and requires 30 kWh of charging energy (a typical workday session), the average power required per vehicle is only 30/8 = 3.75 kW — half the maximum Level 2 rate. If 80 percent of the 200 vehicles are parked simultaneously at any time, the average aggregate power demand is 160 × 3.75 = 600 kW — 42 percent of nameplate.
Under NEC 625.42(B), the feeder and service may be sized for 600 kW rather than 1,440 kW, provided an automatic load management system limits the aggregate demand to 600 kW. The 840 kW difference determines whether the project requires a utility service upgrade: if the existing service can accommodate 600 kW of additional EV demand, no upgrade is needed; if it cannot accommodate 1,440 kW, the project would otherwise be infeasible.
NEC 625.42(B) establishes four requirements for an automatic load management system that qualifies for the calculated-load credit. First, the system must be listed for the purpose — a requirement that currently encompasses OCPP-compliant EV energy management systems listed to UL 3141 or equivalent. Second, the system must automatically reduce the charging rate of connected vehicles when the aggregate demand approaches the managed limit, without requiring manual intervention by a building operator. Third, the system must be set to a maximum aggregate load that does not exceed the serving electrical infrastructure's rated capacity. Fourth, the system must be tamper-resistant — the managed load limit must not be adjustable by the EVSE users through the EVSE interface.
The listing requirement is the most significant practical constraint: not all smart charging software platforms have achieved the listed status required for the 625.42(B) credit, and the engineer must verify the listing before specifying the platform as the basis for the reduced service sizing. Major platforms with current UL 3141 listing include ChargePoint CPM (Cloud Platform Manager), EVCS SmartCTRL, and several utility-managed EVSE programs. The listing status should be verified against the current UL certified products list at the time of design, as listings may change.
The optimal managed load limit is determined by the probability that the system will be unable to meet all drivers' charging requirements — a "missed session" where a vehicle departs with less than the requested energy — at a given aggregate demand limit. This probability is a function of the distribution of session arrival times, parking durations, and energy requirements across the vehicle population.
For a workplace installation where vehicles arrive between 7 AM and 9 AM, depart between 4 PM and 6 PM, and require an average of 25 kWh with a standard deviation of 10 kWh, the managed load that produces a missed session rate below 5 percent can be estimated from queuing theory or from Monte Carlo simulation of the session demand. The analysis typically shows that the optimal managed load is 25 to 35 percent of nameplate capacity for well-managed workplace charging installations, and 40 to 60 percent for retail or public charging where parking durations are shorter and more variable.
Open Charge Point Protocol (OCPP) version 2.0.1 is the dominant communication standard for managed EVSE in the North American market, supported by all major EVSE manufacturers and most energy management system platforms. OCPP defines the message exchange between the charge management system (CMS) server and individual EVSE for session initiation and termination, charging rate adjustment (the SetChargingProfile message), energy metering, fault reporting, and firmware update management.
The smart charging model in OCPP 2.0.1 uses charging profiles to communicate the charge management system's dispatch commands to each EVSE. A charging profile specifies the maximum charging rate (in amps or watts) as a function of time, for a specified session or for the EVSE's default behavior. The CMS updates each EVSE's charging profile in real time as the aggregate demand control algorithm adjusts individual rates to maintain the managed aggregate below the limit. The EVSE executes the profile by modulating its output current via the pilot signal (for AC Level 2 EVSE) or its power electronics control (for DC fast chargers), maintaining the requested rate within the accuracy specified by the charging profile (typically ±10 percent).
The aggregate demand control algorithm runs at the CMS server level, executing on a control cycle of 5 to 30 seconds (shorter cycles provide faster response to demand spikes; longer cycles reduce communication load on the network). At each cycle, the algorithm: (1) collects the current charging rate of each connected vehicle from the EVSE energy meter; (2) computes the aggregate demand; (3) if the aggregate exceeds the limit, identifies vehicles in the lowest priority tier and reduces their charging rate by a specified increment; (4) if the aggregate is below the limit and headroom is available, identifies vehicles in the highest priority tier that have been rate-limited and increases their rate toward the EVSE maximum.
The priority assignment — which vehicles are rate-limited first — is the engineering decision with the most direct impact on user satisfaction. A departure-time-based priority algorithm assigns priority proportional to the ratio of remaining required energy to remaining available charging time: a vehicle with 20 kWh remaining and 2 hours to departure requires 10 kW average and is assigned higher priority than a vehicle with 20 kWh remaining and 6 hours to departure requiring only 3.3 kW average. This algorithm maximizes the number of sessions completed successfully under the managed demand constraint.
The 2026 NEC (NFPA 70, 2026 edition) includes revisions to Article 625 that address bidirectional EV charging (V2G and V2H) for the first time with specific requirements for the interconnection of vehicle-to-grid capable EVSE with the premises wiring. Article 705 is referenced for the interactive system interconnection requirements, and new language in 625.54 clarifies the anti-islanding requirements for V2G capable EVSE to ensure that the vehicle battery cannot unintentionally energize the premises wiring during a utility outage. Engineers designing V2G installations in jurisdictions that have adopted the 2026 NEC must verify that their EVSE equipment, load management system, and service entrance configuration meet the new requirements.
The analysis in this paper connects to several companion studies in this library. Readers concerned with the upstream and downstream engineering will find Electric Vehicle Charging Infrastructure develops a closely related aspect of the same problem, while Electric Vehicle Charging Infrastructure extends the treatment into an adjacent domain. For the broader methodological context, Electric Vehicle Charging Infrastructure Deployment provides complementary depth.
NEC 625.42 permits the feeder and service load calculation for EVSE to be based on the managed load limit established by a qualifying automatic load management system rather than on the sum of charger nameplate ratings, and the central conclusion of this paper is that this single provision is what determines whether a large EV deployment can be accommodated within existing infrastructure, because it can reduce the calculated service load by 60 to 90 percent for installations where vehicles dwell for hours and draw only a fraction of their maximum charging time. The smart-charging architecture developed here, built on the OCPP communication framework and a demand-control algorithm, is the engineering means by which the managed-load limit is enforced reliably enough to satisfy the code. For the practicing engineer, the operative takeaway is that the load-management system must meet the qualifying-system requirements of NEC 625.42 to be creditable in the load calculation, and that the managed-load limit then becomes the governing design parameter, so the demand-control algorithm and its communication infrastructure must be specified with the same rigor as the conductors and overcurrent devices they allow to be downsized.
[1] NFPA 70, National Electrical Code, Article 625, 2023 edition, NFPA, 2023.
[2] Open Charge Point Protocol (OCPP) 2.0.1, Open Charge Alliance, 2022.
[3] UL 3141, Standard for Energy Management Equipment, UL, 2022.
[4] SAE International, SAE J1772, SAE, 2017.
[5] SAE International, SAE J3068: Electric Vehicle Power Transfer System using a Three-Phase Capable Coupler, SAE, 2018.
[6] IEEE Standard 2030.5-2018, Smart Energy Profile Application Protocol Standard, IEEE, 2018.
[7] EPRI, Electric Vehicle Smart Charging: Technical and Policy Considerations, EPRI 3002019900, 2022.
[8] NFPA 70, National Electrical Code, Articles 705, 2023 edition, NFPA, 2023.