Automatic EV load management: a five-box map and commissioning tests

An automatic EV load management system is not simply a charger setting. It watches one or more electrical conditions, makes a permitted decision, and sends an instruction to chargers.

It leaves someone responsible when the result is not what the site expected. A property can have capable equipment and still have an unclear system if the team cannot say what is watched, what is allowed to change, or who can intervene.

This guide covers how to plan one and how to test it. It names the code section, the listing standards and the protocol that matter, and it gives a five-box map and test cases. It does not choose a vendor or set electrical limits. Those decisions belong to a qualified designer, a licensed electrician, the authority having jurisdiction and the utility.

What to take away

Define the boundary the system will manage

Begin with a plain sentence: "This system observes ___ and may adjust ___ within ___." The blanks force useful decisions. The observed point might be a charger group, a panel, or a feeder, and the record should name which one.

The adjustable item might be a session limit, an enabled state, or a start time. The third blank names the limit the logic must respect, such as a reviewed site cap in amperes.

Do not substitute a marketing label for this sentence. "Dynamic," "smart," and "managed" can describe many different functions. A useful project record identifies the control boundary, the equipment at that boundary, the setting owner, and the evidence used to confirm the current arrangement.

The code basis: NEC 625.42

Under the 2023 National Electrical Code, section 625.42 allows the load of EV supply equipment to be limited by an energy management system instead of the sum of the charger ratings. The energy management system must meet the requirements of Article 750. The limit it enforces on the service or feeder then becomes part of the reviewed design.

The 2023 text also lets the energy management system sit inside one piece of equipment or inside a listed system of several pieces. Check the wording in the NFPA code book, and check the edition your jurisdiction has adopted, because an earlier edition may read differently.

Illustrative example: a feeder rated for 100 A with a measured building load of 70 A leaves 30 A of headroom. The controller caps the chargers' combined draw at 30 A. When measured load falls, it raises the cap only after the load has stayed below a set level for a set time. For example, that time might be 5 minutes. The 5-minute value is an illustration, not a code or vendor setting.

Know the control approaches

Most apartment systems use one of three approaches, and the boundary sentence should say which:

Check listings, the utility and the permit

Three outside checks come before procurement.

Draw the signal-to-action map

The clearest map has five boxes and arrows between them:

Add a name to every arrow. A reading may arrive over a local connection, a networked gateway, or a metering platform. An instruction may change a setpoint, pause a session, or place equipment in a defined state. The plan should state what happens when a signal arrives late, is missing, or conflicts with another condition.

This map is not a wiring diagram. It is an operating explanation that the property manager, installer, and support provider can all use during a service call.

Set an operating objective that can be checked

"Avoid problems" is too vague to test. A usable objective says what the team maintains and which tradeoffs are allowed, such as keeping operation within a reviewed site limit or following an approved schedule.

A site may state that sessions are managed within a documented group limit. It should not claim that every vehicle will receive a certain amount of energy by a certain time unless the system and the support process can deliver that.

Separate automatic action from site approval

Automation can carry out a configured rule. It does not settle whether that rule suits the electrical installation, the utility arrangement, the code requirement, or the resident agreement.

For every automatic action, list the source of authority: a reviewed design document, an approved operating policy, a vendor configuration record, or another controlled reference. If the team cannot name the source, treat the action as an open question.

Choose a safe fallback before an outage

Every automated arrangement needs a defined response for communication loss, sensor failure, provider outage, or uncertain data. The approved design sets the real values. For example, a fallback might hold each charger at a fixed low current, such as 16 A, or stop new sessions until the link returns. Those values are illustrations.

Write the decision path in advance. Name who may declare a temporary mode and how that decision is recorded. The site should also know what evidence ends the temporary mode: restored communication, a completed test, a provider confirmation, or review by the responsible professional.

Commission the behavior, not just the hardware

Commissioning should show that the system behaves as described. A test plan includes normal operation, a managed condition, a communication-loss case, a recovery step, and evidence the team can find later.

Each case needs a pass mark written before the test:

Record what was tested, which assets took part, the settings in place, the expected behavior, the observed behavior, and any limits on the test. A screenshot alone is rarely enough because it may omit the time, the configuration, or the asset identity.

After commissioning, repeat a focused test when a change could affect behavior, such as a change to a control group, a software update, a meter or gateway swap, or new equipment.

Make change control routine

The configuration at opening is a starting point. A workable change process records the request, the owner, the reason, the affected assets, the source of approval, the implementation time, the test, and the result. For consequential changes, add a rollback plan.

Treat access changes with the same care as a visible settings change. The outgoing administrator should not be the only person able to find a configuration or its support history.

Assign operating roles and build a handover packet

Keep the duties separate even when one company holds several roles. A property manager may own resident communication and records. A service provider may handle diagnostics. A licensed electrician evaluates any work that changes the installation. A software administrator controls account access.

A new manager or provider should be able to understand the system from the handover packet alone. It should contain the boundary description and signal-to-action map, asset identifiers, the latest approved settings reference, named contacts, commissioning evidence, open issues, and a dated change and incident record.

Common questions

Is automatic load management the same as a charger scheduling feature?

Not necessarily. Scheduling may be one input or action in a broader system. A load-management plan should identify the monitored condition, the decision logic, the controlled asset, and the fallback state.

Can a property copy another site's settings?

No. A setting depends on the boundary, equipment, approvals, design assumptions, and operating rules at that site. Use another site only as a prompt for questions.

What should be reviewed after a controller or network change?

Review the signal path, affected assets, access roles, fallback behavior, test evidence, support contacts, and the resident-facing message. Involve the responsible technical party when the change affects the installation or the approved design.