Dynamic load management is what allows a site to install more charging bays than its supply could serve simultaneously at full rate. It works because vehicles rarely all draw maximum current at the same moment.
The engineering interest is entirely in the minutes when they do, and in what happens when the system coordinating them stops working.
What it measures
Dynamic management needs to know the total site load, including everything that is not charging: lifts, chillers, kitchens, lighting. That comes from a meter at the supply point rather than from the chargers themselves.
The available charging budget is the supply capacity minus that measured load, minus a safety margin. Without the measurement, a system can only divide a fixed allowance, which is useful but not dynamic.
How the limit reaches the vehicle
The controller allocates a share to each active charger, and each charger communicates its limit to the vehicle by adjusting the control pilot duty cycle. A compliant vehicle reduces its draw accordingly.
That makes the limit an offer rather than an enforcement. Vehicles vary in how quickly they respond, and a few respond unexpectedly, which is why allocation should retain headroom rather than assign to the last ampere.
The fallback decides whether it is safe
If the coordinating system becomes unreachable, chargers must fall back to a limit that is safe with every bay active. Reverting to full rating on a supply-constrained site will trip the protection.
That fallback value is a design decision, not a default. Set it too low and a network outage becomes a customer complaint; too high and it becomes an electrical event.
Phase awareness recovers capacity
Single-phase chargers on a three-phase supply load one phase each. Allocation that ignores which charger sits on which phase can hit a per-phase limit while other phases sit idle.
Phase-aware allocation is more complex and recovers capacity that a simple total-current budget leaves unused, which on a constrained site is the difference between several additional bays and none.
Central or distributed
Allocation can be computed centrally, in the cloud or in a site controller, or negotiated between chargers locally. Central is simpler and depends on the link; distributed continues working when the link is down and is harder to reason about.
Which is appropriate depends on how tolerable a connectivity loss is at that site, which is a commercial question as much as a technical one.
Prioritisation is a business rule
Once capacity is scarce, someone charges slower. Whether that is the visitor rather than the tenant, or the long-dwell vehicle rather than the one leaving soon, is a decision the operator should make and the system should implement.
Systems that impose their own fairness model rather than expressing yours tend to be overridden, which defeats the purpose.