In short. Load management is the set of rules and controls that decide which loads a generator carries, in what order they connect after a transfer, which are shed when demand exceeds capacity, and how they are restored. A well-designed scheme lets a correctly sized generator carry a site whose simultaneous peak or motor starting would otherwise overload it, and protects a paralleling plant when a set is lost. It is usually cheaper than the next size of generator.
The problem load management solves
Generator capacity is bought for the steady load, with a margin. Three things routinely exceed it: simultaneous motor starting after a transfer, when every contactor that dropped out re-closes at once; the coincident peak, when chillers, compressors and the process line are all at full demand in the same afternoon; and, on a multi-set plant, the loss of one set. Buying a generator large enough for all three means running it lightly loaded for most of its life, which wastes capital and fuel and causes wet stacking. Load management lets the plant be sized for what matters and handles the exceptions by control.
Step one: classify the loads
Every scheme begins with a load list in which each feeder is assigned a priority. Four levels cover almost every facility we see in Telangana and Andhra Pradesh.
| Priority | Description | Typical examples | Treatment on generator |
|---|---|---|---|
| 1. Safety and mission-critical | Loss creates a safety hazard or an unrecoverable loss | Fire pumps, emergency lighting, operating theatres, UPS input, control systems | Connected first, never shed |
| 2. Operationally essential | Loss stops production or service | Process line, cold rooms, servers, water supply, main lifts | Connected second, shed only as a last step |
| 3. Deferrable | Can wait minutes to hours without harm | Comfort air conditioning, battery charging, secondary pumps, workshop power | Connected last, shed first |
| 4. Non-essential during outage | Not needed while on generator | Decorative lighting, canteen equipment, non-urgent machinery | Not connected on generator supply |
The classification should be agreed with the people who operate each area and signed off, because it is a decision about what the business is willing to lose for a few hours. Building the list with kW, starting kVA and power factor for each item is described under load inventory.
Step two: sequence the connection
After a transfer to generator, loads are connected in priority order with a delay between steps so that voltage and frequency recover before the next step. For most sets between 100 and 1,000 kVA, 5 to 15 seconds between steps is sufficient; large motors or transformer energisation may need longer, and step sizes should be checked against the set’s load acceptance, which on a CPCB IV+ set with aftertreatment may be more limited than on an older engine. See transient response. Sequencing is implemented with timers on contactor coils or feeder breakers, or in a PLC or master controller that closes feeders in order on a “generator on load” signal. Soft starters or drives on individual large motors may remove the need to sequence at all.
Step three: shed when capacity is exceeded
Shedding is the automatic disconnection of lower-priority loads when the generator cannot carry the demand. The triggers and the response must be decided and set.
- Overload by kW. The controller sheds the lowest priority block when output exceeds a threshold, typically 90 to 100 percent of rating for a set time, and the next block if it is still high.
- Underfrequency. A fast, large overload pulls engine speed down before a kW relay can respond. A trigger at around 48 to 49 Hz with a short delay sheds immediately and is the primary protection against a stall.
- Loss of a set in a paralleling plant. When a generator breaker opens unexpectedly, the controller sheds to a pre-calculated table rather than waiting to measure the overload.
- Low reserve. A plant controller may refuse to connect deferrable load when spare capacity is below what the largest expected step needs.
Shedding blocks should be sized so that one or two steps resolve the condition. The relay engineering is under load shedding.
Step four: restore in a controlled sequence
Restoration is the step most often left to chance, and it is where the second trip happens. When capacity returns, loads are reconnected in priority order with the same inter-step delays as the initial sequence, and each restoration is followed by a hold of a minute or more. A scheme that sheds at 100 percent and restores at 95 percent will cycle; restore only when load is well below the shed threshold with the shed block’s expected demand added, for example below 70 percent.
A worked example
Worked example (illustrative): a food processing plant has a 500 kVA, 400 kW generator. Priority 1 load is 60 kW, priority 2 is 220 kW including a 45 kW refrigeration compressor, priority 3 is 110 kW of air conditioning and workshop supply, and priority 4 is 40 kW. The total of 430 kW exceeds the set, but priorities 1 to 3 total 390 kW, acceptable for a few hours if the set is in good condition at site ambient. The sequence connects priority 1 at transfer, priority 2 after 10 seconds, the compressor on its own step 15 seconds later because its starting kVA is the largest single transient, and priority 3 after a further 15 seconds. Shedding drops priority 3 at 100 percent kW for 10 seconds or at 48.5 Hz instantly, and drops the compressor only if that has not cleared the overload. Restoration of priority 3 waits until load has been below 280 kW for two minutes. The plant runs on a 500 kVA set instead of the 625 kVA set it would otherwise have bought.
Load management in multi-generator plants
On a paralleling plant the same four steps apply, with one addition: the number of sets online changes the capacity, so the scheme must know which sets are available. The master controller holds a table of what each combination of sets can carry, sheds to it immediately when a set trips, and uses it to decide whether a deferrable block can connect now or must wait for another set. See generator paralleling and master controls.
UPS within the scheme
A UPS bridges the gap while the generator starts, but its rectifier can draw a large step when it begins recharging batteries after the transfer. Most industrial UPS allow recharge current to be limited or delayed on generator supply, and that setting belongs in the plan, as does a check of the UPS input tolerance for generator frequency and voltage variation. See UPS and non-linear loads.
Common mistakes
- Allowing all loads to reconnect at once after transfer; the combined starting current trips the generator on underfrequency and the site is dark again.
- Shedding only on kW overload with a long delay; a large step stalls the engine before the relay times out.
- Setting the restore threshold just below the shed threshold; the scheme cycles the air conditioning on and off every minute.
What to send us
Send the load list with kW and starting method for each major feeder, the generator nameplate, a photo or drawing of the switchboard showing which loads are on which feeders, a record of any trips during outages, and a short note of which areas the business can do without for a few hours. If a UPS is installed, send its rating and whether it has a generator mode. We will produce a priority table, a connection sequence and shedding and restoration settings. Use the application engineering enquiry.
Questions engineers ask
What is the difference between load management and load shedding?
Load shedding is one function: disconnecting selected loads when demand exceeds capacity. Load management is the whole scheme, which also includes classifying loads, sequencing their connection after a transfer, deciding when to restore shed loads, and in multi-set plants coordinating with how many generators are running.
Can load management let me buy a smaller generator?
Often, yes. If the overload cases are simultaneous starting or a short coincident peak, sequencing and shedding handle them and the set can be sized for the sustained essential load. It cannot help if the essential load itself exceeds the set; the classification exercise tells you which case you are in.
Should the UPS be shed to protect the generator?
No. The UPS feeds the loads with the least tolerance for interruption and belongs in the highest priority tier. What should be managed is its battery recharge current, which can be limited or delayed when the UPS detects generator supply, so that it does not add a large step at the worst moment.
How fast must load shedding act?
For a loss-of-set event or a large sudden overload, within a few hundred milliseconds, which is why underfrequency and breaker-status triggers are used rather than kW measurement alone. For a slow overload a delay of 5 to 10 seconds is appropriate so that transient peaks do not cause unnecessary shedding.
Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.