In short. The energy management system (EMS) is the control layer that decides which source supplies the load, when the generator starts and stops, when the battery charges or discharges, and how much reserve is held. Its objectives conflict, so they must be ranked: equipment safety first, then system stability, then resilience, then economics. A good EMS also defines what happens when a meter, network link or controller fails. Poor logic can waste the value of expensive hardware.
What the EMS sees
A hybrid EMS collects live data from every source and the site load: generator status, kW, kVAr and fuel level; battery state of charge, temperature, available charge and discharge power; PV production; grid voltage, frequency and import; and the load itself. It also uses slower information such as the tariff schedule, the time of day, forecast solar output and the reserve policy. From this it issues setpoints to the generator controllers, the battery PCS and the PV inverters, and commands to breakers and load-shedding relays. The quality of its decisions is limited by the quality of this data, so metering accuracy and communication reliability are engineering requirements, not afterthoughts.
Ranking the objectives
The objectives an owner asks for usually include resilience, fuel saving, maximum solar use, lower demand charges and long battery life. They pull against each other: maximum solar use and fuel saving want the battery cycled hard, battery life wants it cycled gently, and resilience wants it kept full. The EMS needs an explicit hierarchy.
| Priority | Layer | Examples of rules |
|---|---|---|
| 1 | Safety and equipment limits | Never exceed PCS current, cell temperature or generator rating; never discharge below minimum state of charge |
| 2 | System stability | Maintain voltage and frequency; keep enough online reserve for the largest step |
| 3 | Resilience | Hold the outage and black-start reserve; keep the generator ready |
| 4 | Economic optimisation | Shift solar, shave peaks, minimise generator hours |
| 5 | Asset life | Limit cycling depth and high state-of-charge dwell time |
A rule in a lower layer can never override one above it. Cost saving should never push a generator below its minimum load or spend the reserve.
Typical control functions
Generator dispatch
Load-dependent start and stop, with thresholds, delays and minimum run and off times so engines are not short-cycled. On multi-set sites this includes choosing which set runs to balance hours, as in multi-DG sequencing.
Battery charge and discharge
The state-of-charge window, reserve band, charge source priority and charge rate, described in BESS charge and discharge control.
PV curtailment
Limiting PV output so the generator stays above minimum load and the battery is not overcharged.
Load management
Shedding non-essential loads when sources are short, and restoring them in stages.
A dispatch rule written properly
Vague rules cause most commissioning problems. “Use the battery at night” is not a rule. A usable rule reads like this (illustrative): when the generator is running, site load is below 35 percent of generator rating for ten minutes, and battery state of charge is above 60 percent, transfer load to the battery and stop the generator after a five-minute cool-down; restart the generator when state of charge falls to 30 percent or load exceeds 80 percent of PCS rating for more than thirty seconds. The numbers come from the site design; the point is that every condition, threshold and delay is stated and testable.
Dispatchable energy (kWh) = (SoC − SoCreserve) × Nameplate capacity (kWh)
SoC is the present state of charge and SoCreserve the protected reserve level, both as fractions. With a 1,000 kWh battery at 70 percent and a 30 percent reserve, 400 kWh is available to economic dispatch (illustrative); the EMS must never plan beyond it.
Failure modes and fallback
Every EMS must have a defined answer to these questions: what happens if the EMS itself fails, if the network between EMS and controllers fails, if a meter reading freezes, if the battery trips offline, if PV disappears, or if a generator controller loses communication. The safe answer is usually that each device falls back to a self-contained mode: generators to their own AMF and load-sharing logic, the battery to a standby or grid-forming hold mode, and PV to a fixed low limit or off. These fallback modes must be tested at commissioning by actually pulling cables and stopping the EMS. See degraded control modes.
Cybersecurity and change control
An EMS connected to a remote monitoring platform or the internet is a route into the site’s power system. Access should be authenticated and logged, the control network segregated from office IT, and remote write access limited to named people. Setting changes should be recorded and approved, because a single change to a reserve threshold can undo the site’s resilience design without anyone noticing.
Common mistakes
- Setting economic objectives without a written priority order. The EMS spends the reserve to save a few litres of diesel.
- Leaving fallback behaviour to manufacturer defaults. A communication fault stops the generator or trips the battery at the wrong moment.
- Testing each device separately but never the whole sequence. The plant fails its first real outage.
- Allowing remote setting changes without logging. Thresholds drift and nobody can explain later why the system behaved as it did.
What to send us
Send the single-line diagram, the make and model of every generator controller, battery PCS and PV inverter, and any existing control philosophy or EMS documentation. Tell us what the owner wants the system to achieve and in which order, and describe any incident where the controls behaved unexpectedly. Through the application engineering enquiry we can review the logic or write a control philosophy.
Questions engineers ask
Is the EMS the same as the generator’s power management system?
No. A power management system coordinates generators: synchronising, load sharing, sequencing and shedding. The EMS sits above it and decides the overall energy strategy across generators, battery, solar and grid, issuing setpoints the PMS and other controllers carry out.
What communication protocol should the EMS use?
Modbus TCP is the most common for generator controllers, inverters and meters, with IEC 61850 on larger switchgear. Critical trips and interlocks should remain hardwired so that a network failure cannot disable protection.
Can the EMS forecast solar and load?
Many can, using weather data and historical profiles to decide in advance how much to charge or when to start a generator. Forecasting improves economics, but the reserve rules must hold even when the forecast is wrong.
How do I know the EMS is working well?
Track generator hours, average generator load, battery cycles, curtailed PV and reserve violations each month against the design intent. A dashboard nobody reviews is not a control system.
Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.