What kind of fluctuations are present at the line voltages (3p 4wire) when the load is welding machine... What kind of fluctauations/disturbances in input voltages can be present at the worst case when the load is welding machine (I donot know about the power rating/specs about welding machines)........lets say if harmonics are present what kind of harmonics/if the input voltage wave gets disturbed?....what is the worst possible care should be taken when neutral is connected in these kind of loads?....

As usual, the questions on this site are way too general to get any kind of accurate answers. If I make some broad assumptions, I would guess that the worst case welding topology would simply have a diode rectified input to an internal capacitor. With this type of welder it could create a significant distortion power factor on the line depending on the manufacture's choice of capacitance value versus weld power. A value in the range of .6pf and a current crest factor above 3 might be conceivable. At this point it would depend on the source impedance to determine the voltage fluctuation. A zero source impedance would create no voltage variation. You could run simulations to get an idea of the different amounts of harmonics by varying the source impedance, input capacitance and output load (which could be modeled as resistive).
I do not know of any 3P welding power sources that utilize the neutral conductor. All that I am aware of are designed to work off of a Delta distribution and for the most part draw balanced current from the primary.

"Sag " and "flicker "are very common in case of welding.

I once was sent to Bangkok to fault find a telephone exchange that had blown up. Literally every piece of equipment in the exchange had failed, including the PC's and the airconditioning.
The root cause was a welding company that had, illegally, attached its feeders to the external pole mounted 3-phase supply transformer.
Our equipment had been tested to the ANSI C62.41 recommendation with the knob turned up to "11", ie the source voltage was the absolute max the Kryten pulse generator could produce, about 8kV, and it passed. There was also an input overvoltage protection mechanism for mains OVP. Lord only knows what was the actual disturbance at the input to the exchange.

Carole, did they connect up to the telephone CO's transformer or on the distribution side of their transformer (i.e. did they add their one transformer ?)