In short. Automatic dispatch runs only as many sets as the load needs, while keeping enough reserve to survive the largest credible event. Thresholds, delays, priority rotation and failure replacement must be tuned together, or the plant will either hunt on and off or run short of margin.
Dispatch rules
| Rule | Purpose |
|---|---|
| Start threshold | Start another set when load exceeds a percentage of running capacity for a set time |
| Stop threshold | Stop a set when remaining sets can carry load with margin for a set time |
| Hysteresis | Gap between start and stop thresholds to avoid cycling |
| Priority and rotation | Balance run hours and keep the best sets ready |
| Reserve rule | Maintain capacity to lose the largest running set |
| Failure replacement | Start the next available set if one trips |
Sequence of a start
Demand detected → start command → crank → run → voltage and frequency acceptable → sync check → breaker close → load ramp → sharing active. Each step has a timeout. If a set fails to start, the next set is called.
Sequence of a stop
Stop condition → unload ramp → breaker open → cool-down → stop → return to standby. Do not stop sets in a way that leaves the plant below reserve.
Timing issues
- A start delay that is too long may cause overload before help arrives.
- Too short causes starts on brief peaks.
- Large steps, such as motor starts, are not met by dispatch alone; they need enough running capacity or sequencing. See large-motor starting.
Engine considerations
Running engines at very low load for long periods causes wet stacking and fouling. Dispatch should avoid prolonged light loading, and plan periodic higher-load running. See load-bank testing.
Illustrative example
A plant with three 1,000 kVA sets runs one set to 80 percent load, then starts a second after a 10 second delay. When load falls below 40 percent of two-set capacity for 5 minutes, the second set unloads and stops. If the running set trips, the next set starts, and the PMS sheds low-priority loads until it takes over. See PMS architecture.
The sequence on a mains failure (illustrative timings)
| Step | Event | Typical time |
|---|---|---|
| 1 | Mains failure detected and confirmed | 1 to 3 s |
| 2 | All standby sets receive start command | Immediate |
| 3 | First set at rated voltage and frequency closes to dead bus | 8 to 15 s |
| 4 | Other sets synchronise and close | 5 to 10 s each |
| 5 | Load added in priority steps | Seconds between steps |
| 6 | Surplus sets stop if demand is low | After minimum run time |
Dispatch rules
- Order of start by priority and run hours.
- Start a further set when loading passes the upper threshold for a delay time.
- Stop a set only if the others can carry the load with margin.
- Keep the N+1 standby available.
Fail-to-start handling
If a set fails to start in the allowed time, the next set is called without waiting for operator action, and an alarm records the fault. If capacity is insufficient, automatic load shedding limits the load to what is available.
Dead-bus closing safety
Only one breaker may close to a dead bus. Hardware interlock plus logic ensures this. See dead-bus and live-bus logic.
See N+1 design.
The master engineering question
At each load level, how much margin does the running fleet have against the worst credible event, and how long does the next set need to arrive?
What to send us
Send load profile and set data. Send your application details.
Questions engineers ask
Why do sets keep starting and stopping?
Thresholds or delays are too close. Increase hysteresis and delay.
Should I rotate priority?
Yes, usually, to balance wear. Keep sets in good health ready.
How much reserve?
Defined by the redundancy level and worst credible event.
Can dispatch use load forecasts?
Yes, via an EMS, within PMS safety limits.
Values are illustrative. Tune on site.