In short. A generator is a source; the power system decides what happens next. Continuity depends on a chain of mains sensing, AMF logic, engine start, generator validation, transfer or paralleling, switchgear, protection, distribution and the load, and every interface must behave correctly. Integrated power design starts from the operating sequence the site needs when any source changes state, then selects the ATS, panels, breakers and controls that deliver that sequence and proves it by testing.
Why a good generator can still fail the plant
We regularly attend sites where the generator itself is healthy and the outage still happened. The set started but the AMF panel never sent the transfer command because mains sensing was wired to the wrong side of a fuse. The set took load but tripped when every motor restarted at once. Two sets were installed for redundancy but shared one control supply that failed. In each case the engine and alternator were fine; the system around them was not designed or tested as a whole.
Integrated power engineering treats the generator as one element in a chain and designs the chain.
The elements and what each one does
| Element | Function | Typical design question |
|---|---|---|
| Mains sensing and AMF logic | Detect abnormal mains and run the standby sequence | Which conditions count as failure, and with what delays? |
| Automatic transfer switch | Move load between mains and generator | Open, closed or delayed transition, and what rating and utilisation category? |
| Paralleling and synchronising | Run several generators on a common bus | How many sets, what redundancy, which priority? |
| Load sharing | Divide kW and kVAr between paralleled sets | Isochronous or droop, and how is reactive load shared? |
| Load management | Admit, shed and restore loads | What starts first, and what is dropped if a set trips? |
| Switchgear | Switch and protect power flow | Fault level, breaker ratings, interlocks |
| Transformers | Link voltage levels | Inrush on generator supply, vector group, neutral |
| Protection | Isolate faults selectively | Does the nearest breaker clear the fault with generator fault current? |
| Earthing and neutral | Fault return path and safety | Where is the neutral earthed with each source? |
| Monitoring | Show source, breaker and load state | What alarms reach whom, and over which link? |
Each element has its own guide: automatic transfer switches, AMF panels, generator paralleling, protection coordination and the earthing and neutral architecture guide.
Begin with the sequence, not the ATS rating
The usual starting question is what rating of ATS is needed. The better starting question is what must happen, step by step, when a source changes state. For a single standby set with an AMF panel, the normal sequence is mains monitoring, detection of failure after a confirmation delay, engine start, generator voltage and frequency validation, transfer, load supply, mains return detection, a stability delay, retransfer and a cool-down run before stopping. Each step has a timing, a condition and a fault response.
t_total = t_detect + t_start + t_stabilise + t_transfer
Here t_detect is the mains failure confirmation delay, t_start is the cranking and run-up time to rated speed, t_stabilise is the time for voltage and frequency to settle within limits, and t_transfer is the switching time of the ATS or breaker. As an example (illustrative), a confirmation delay of 3 seconds, a start of 5 seconds, stabilisation of 2 seconds and a transfer of 0.5 seconds gives a supply gap of about 10.5 seconds. Loads that cannot tolerate that gap need a UPS; loads that can tolerate it do not. The real values come from the controller settings and the manufacturer’s start performance.
Design for the abnormal states
A serious design defines a response for each credible condition, not only for normal operation.
- Mains fails and the generator fails to start: alarm, retry limit, and whether a second set is called.
- Generator starts but voltage or frequency is out of limits: transfer must be blocked.
- The ATS fails to transfer: alarm and manual procedure.
- One set trips in a parallel plant: shed load to what the remaining sets can carry.
- Mains returns unstable or with wrong phase sequence: retransfer must wait.
- A breaker is racked out for maintenance: interlocks must still prevent paralleling of unsynchronised sources.
- A communication link fails: the plant must fall back to a safe local mode.
Single set, multiple sets and grid paralleling
Single standby set
One set with an AMF panel and a changeover or ATS covers most commercial buildings and small factories in Hyderabad. The design work is in correct sizing for motor starting and in a transfer arrangement that is interlocked and tested.
Multiple paralleled sets
Several sets on a common bus give redundancy, allow capacity to follow load and keep engines efficiently loaded. Before a breaker closes, voltage, frequency, phase sequence and phase angle must match, and afterwards the sets must share kW and kVAr in proportion to rating. Synchronising panels and load sharing cover this.
Paralleling with the grid
Closed-transition transfer, peak lopping and export all require operation in parallel with the DISCOM supply. These need DISCOM permission, suitable protection such as reverse power and loss-of-mains relays, and careful study, so they are designed case by case.
Proving the system works
An integrated system is accepted on a test of the whole chain. The commissioning test should open the mains supply, let the system detect, start, transfer and carry real or load-bank load, then restore mains and observe retransfer and cool-down, while recording times, voltage and frequency. Failure injection, such as a simulated failed start or a tripped set in a paralleled plant, shows whether the abnormal state responses work. Routine starts without transfer prove only that the engine runs.
Related photographs and videos
Common mistakes
- Selecting the ATS by current rating alone without considering its transition type and fault withstand, which can leave a switch that welds or opens under fault current.
- Designing only the normal sequence, so the system has no defined response when a set fails to start or trips under load.
- Letting all loads restart at once on transfer, which causes a voltage and frequency dip that trips the generator.
- Ignoring neutral earthing when a second source is added, which creates circulating currents or defeats earth fault protection.
- Accepting a weekly no-load start as a test, which never exercises transfer or load acceptance.
What to send us
Send the single-line diagram, the load list with motors and their starting methods, the existing generator and panel nameplates, and a description of what must happen when mains fails and returns. Tell us whether you need redundancy, paralleling or grid paralleling, and share photographs of the main LT panel and existing changeover. Send an application engineering enquiry and we will review the sequence and architecture. The system design guide sets out the full method.
Questions engineers ask
What is the difference between an AMF panel and an ATS?
The AMF panel contains the logic that detects mains failure and controls the generator start, stop and transfer sequence. The ATS is the switching device that physically moves the load between sources. Many panels combine both, but they are separate functions with separate failure modes.
When should we parallel two generators instead of buying one large set?
Paralleling makes sense when you need redundancy, when load varies widely so one large set would often run lightly loaded, or when future growth is expected. It adds synchronising panels and controls, so for a small steady load a single correctly sized set is usually simpler.
How long does a standby generator take to supply the load?
Typically around ten to fifteen seconds from mains failure, made up of the confirmation delay, engine start, stabilisation and transfer. The exact figure depends on the controller settings and engine, and it should be measured at commissioning.
Can our generator run in parallel with the electricity board supply?
Technically yes, with synchronising equipment and suitable protection, but it requires DISCOM approval and a protection study. For most standby duties an open transition transfer that never parallels with mains is simpler and is what we recommend unless there is a clear reason otherwise.
How should an integrated power system be tested?
By a full mains failure simulation under load that runs the complete sequence, including retransfer, with times and readings recorded, plus tests of the main failure responses. This should be done at commissioning and repeated periodically.
Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.