Permanent Star connection

You can only do it if the motor is wound for it. If it is, you must consult with the motor manufacturer for the connections. There are some industry standards but they may or may not be followed. So nothing that anyone tells you will be universally true.

Now as to feasibility. When you connect a motor that was designed for Delta in a Star pattern, you are reducing the effective winding voltage to 57% of normal (the square root of three). This reduces the torque to the square of that amount, so you end up with 33% torque. The frequency remains the same and and therefore the motor's mechanical power is reduced to 33% of normal as well. So the 40% figure was incorrect.

An AC motor only uses the energy required to move the load, plus whatever inefficiencies are in the motor itself, as defined by losses. In a modern motor, losses have already been reduced a lot more than most people realize. Most energy efficient designs now run at 95% efficiency or better. So your total losses will only be 5% f the absorbed power to start with. Of that 5%, about 25% are iron losses, i.e. magnetic permeability resistance. These losses are related to the applied voltage so when you reduce the voltage, you reduce these losses. So running at 57% voltage will theoretically reduces the iron losses by a similar percent (although some are fixed as well). So that means of the 5% total losses, no more than 43% of 25% will be reduced. So factor that out: .05 x .25 = .0125, x .43 = .005375 or 0.54% of the absorbed will be saved by running the motor in Star. Keep in mind that the entire scheme only works if the motor is already only 33% loaded. So if you have a 10kW motor, and you are only needing 3.3kW, you can connect it in Star and save .0178kW or 18W of energy. If for any reason your load changes and you need more than 33% power, you will not have it and the motor will stall.

Of course when the motor isn't running you are saving 100% energy!

In terms of application, this is normally applied to constant load applications, for example, a fan, where the load is low because of limited air needs. Any job with sudden impact loads above the normal would make motor stalling render it unsuitable.

This can be quite successful if load conditions are within motor limitatioons and effects quite a large drop in current.

The energy aspect is correctly described by j Raef (GA) but you may also consider two other important differences:

1 Since the motor is now 'undervolted' compared to its nameplate relationship, the slip curve is substantially flattened, this means that the motor, for a similar load, will run slower. This, if acceptable, will reduce energy demand.

2. This flettened slip curve leaves the motor susceptible to any extra load causing an increase in slip but, if 'overloaded' will still draw less than nameplate current but now be operating inefficiently and in danger of burnout. So, careful motor protection may be needed if reconnected to the 'new' condition.

3. Reduced current and VA in 'normal' state gives a supply pf advantage which could also be useful.

Many observers see the energy saving caused by point 1 don't recognise the change in speed, attributing the reduction to motor losses instead, hence the hype around voltage optimisers etc in motor loads. If your supply is higher than your motor rating, the resultant 'overfluxing' may mean there are greater gains in STAR connection from a motor loss viewpoint but this is marginal, at best, in my opinion.