A grid-tied plant shuts down when the grid drops, on purpose, and no installer can switch that off. What anti-islanding does, why generators complicate it, and what a hybrid inverter and a battery actually change.
No. A grid-tied plant, which is what nearly every net-metered rooftop in Bangladesh is, stops the moment the grid drops. The modules are still in full sun and the inverters are in perfect health. They disconnect deliberately, and there is no setting an installer can change to prevent it.
This surprises people who have just spent money on a solar plant and then sat through an outage watching it do nothing. It is worth understanding why, because the reason also tells you what to do about it.
If an inverter kept exporting into a feeder the utility had switched off, it would energise a line that a lineman has isolated, tested and believes to be dead. That is the hazard the industry calls islanding, and every grid-tie inverter certified to IEC 62116 or IEEE 1547 is required to detect the loss of the grid and disconnect within a fraction of a second.
The same requirement explains a second thing owners notice. When the power comes back, the plant does not resume instantly. The inverter watches the returning supply for a set observation period, commonly a few minutes, and only reconnects if voltage and frequency have stayed inside limits throughout. On a day of repeated load-shedding those reconnection delays add up, and the plant's daily generation figure will look worse than the weather alone would explain.
Anti-islanding behaviour and the reconnection parameters are part of what the distribution utility approves when it grants the interconnection. They are set at commissioning, recorded, and they are not the contractor's to adjust afterwards.
Most Bangladeshi factories switch to a diesel or gas generator when the grid goes. This is where an ordinary grid-tied plant becomes an active problem rather than merely an idle one.
The inverter cannot tell a generator from the grid. It sees a supply within its voltage and frequency window, waits out its reconnection timer and starts producing. Solar output then displaces load from the generator, and if generation exceeds what the site is drawing, power flows backwards into the set. Generators are not designed to absorb it. The controller trips on reverse power if it is fitted with the protection, and the engine overspeeds if it is not.
There are three accepted ways to handle it, and every site with standby generation needs one of them shown on its single line diagram.
A hybrid system does form an island, safely, by first opening a contactor that separates the site from the utility. Once that separation is physically made and proven, the inverter can energise a designated backup board from the battery and the array without any possibility of back-feeding the network. That is the whole difference: a grid-tied plant cannot prove separation, and a hybrid one can.
Four things are worth knowing before assuming this solves everything.
Backing up an entire factory is rarely the right answer, and the cost of trying makes the whole project look bad. Decide instead what genuinely cannot stop: site and emergency lighting, security and access control, the server room and network, effluent treatment blowers where the biology depends on them, a cold room, the fire pump controller, and any process that spoils a batch if it halts mid-cycle.
Put that list on a dedicated backup board and size the hybrid inverter and battery around the board. Everything else stays on the grid-tied plant, which continues to do the job it is good at, which is producing energy cheaply when the grid is up.
One warning on cost. Separating those circuits onto a backup board means rewiring part of the distribution system, and on an existing factory that electrical work is frequently a larger line in the quotation than the battery itself. Get it surveyed and priced before deciding, because it changes which option is actually the cheapest.