If you point a .45 calibre revolver with only one bullet at your head and pull the trigger, the probability is that you will survive.
Seriously, there is leakage current to ground from both insulation resistance and from the capacitive effect of the cnductors, insulation and sheath. You may have a ground detection system and if it is permanently connected rather than push to test, it will provide a path to ground.
The point is that even an ungrounded system is actually a high impedance grounded system. Whatever current is flowing in this high impedance distributed connection will mostly flow in your body when you touch a hot bus. If this current is much more than 10 or 15 ma. your next of kin will be notified.
Then there are always the devices that have developed quite a bit of leakage current to ground. And there is your playful friend who pushes the ground detector button while you're touching the bus bar. There is also the posibility of a non lethal shock causing yuor hand to involuntarily bridge the buss bars and light up a phase to phase arc.
As in, "The shock was survivable, he died from the burns."
Don't play Russian Roulette with a semi automatic pistol. Don't touch the bus bars.
yours

What voltage are we talking about?

The OP said 450.
That's usually enough for condolances.
respectfully

waross is absolutely correct, there is no such thing as an ungrounded system, only very high impedance grounded systems.  If the only grounding impedance is the capacitive coupling between the conductors and ground, a ground fault might only draw a few 10s or 100s of mA, but through your body that would be more than enough to seriously ruin your day.

http://www.allaboutcircuits.com/vol_1/chpt_3/3.html

Sorry, I missed the reference to 450.  

I agree with other responses - the ungrounded system will still be capacitively coupled to ground, and you will likely get shocked.

I investigated an accident a couple of years ago when an electrician became a ground path for an "ungrounded" 4160 V system.  Left a couple of very small burns on each hand - and a dead body.  Never work alone.

Thanks for the responses. I would, in reality, never touch a live wire but I used that reference to find out how a ungrounded system does its job.
I was thinking that it would be a better system of power transfer for residential use. I guess it would be very expensive to have a return back to the source. And not worth it if it does not provide a high level of safety.

Ungrounded systems used to be popular in industrial settings, not for safety reasons, but for service continuity. A single ground fault did not shut down the system.  

But there are a lot of other problems with ungrounded systems that caused them to lose favor.

dpc (Or any body else)
If you could mention a few of those problems I would appeciate it.
Or, a link with the pros and cons.
Other then that the extra wire, it seems to me to be an ideal system. (I am not an expert with power distribution)

I have seen a lot of shrimp and lobster boats that typically had two 60 KW sets, or a 20 KW and a 60 Kw. Only one set would be used at a time. The voltages would be about 127-220 or 120-208.
If the voltmeter is connected line to neutral the engineer will set the voltage to 120 volts. If the volt meter is connected line to line the engineer will try for 220 or 230 volts.
All systems were grounded.

Ungrounded systems:
There is less chance of starting a fire with a well-maintained ungrounded system, but if a ground develops and is not repaired, you have lost that advantage.
There is a school of thought that fears through the hull currents and accelerated corrosion with a grounded system.
Not so. A grounded system should not have or cause currents in the hull or machinery. Furthermore, to cause corrosion the current must leave the hull or enter the hull. A self-contained electrical system on a vessel should have only one connection to ground. Under normal circumstances this will not cause any current to flow in the hull or in the machinery. In the event of a ground-fault, current will flow only long enough to trip the breaker. Usually a small part of a second.

Shore power:
On shore problems.
It is possible but not likely for a bad shore power connection to cause corrosive currents. There are conditions of unbalance that can develop on a shore power system that can theoretically cause the neutral voltage to be different than the ground potential with resulting hull damaging currents. This is not common, except, if a shore facility has a system problem, then all vessels with grounded systems would be at risk.
On ship problems.
If a vessel connects to a three phase shore power source and loads one or two phases heavily, or if for any reason there is a heavy neutral current this can cause a voltage rise on the neutral. As a result, a significant part of the neutral current may flow through the ground connection and through the hull. A longer shore power cable and more impedance in the cable will cause a greater percentage of the neutral current to flow in the ground path (through the hull)

For these reasons, I will not say that ungrounded systems on vessels are wrong. Personally I would favor using isolating transformers if I anticipated problems with shore power. Another option is a grounding switch that is closed at sea but opened when on shore power. If another person favors an ungrounded system or a grounding switch to avoid possible shore power problems I consider it a difference of opinion, but not something that is wrong.

Currents causing corrosion:
I am familiar with a repair facility used for shrimp and lobster boats. It was built by driving sheet piling and then dredging out one side and filling the other side. The boats tie along side. There is a lot of welding going on. There will sometimes be three boats stem to stern. Sometimes there will be one or two more tied on the outside. For years the standard practice was to leave the welding machines on the shore and ground them to the sheet piling. One welding cable would be taken onboard from each machine and the hull and the ocean would form the return path to the welding machine. There was never any noticeable corrosion on any vessel from this practice.
After 8 or 10 years the sheet piling was damaged in a storm. It was found to be so badly corroded that much of it had to be replaced. I don't know if the sheet piling had zincs. I don't know if the corrosion was due to the welding or an absence of zincs, probably both.
The standing orders now (which are sometimes followed) are to connect the ground to the vessel rather than to the sheet piling.
I am still seeing new and strange electrical equipment and connections on vessels.
The last surprise was a 4 Kw diesel gen-set with a sea water cooled alternator.
I hope this helps your understanding of some of the reasons for grounding and not grounding vessel electrical systems frankiee. I have tried to keep it as non technical as possible and to avoid any terms or jargon that may not be familiar to you.
respectfully

The IEEE Green Book states there are two principal advantages of ungrounded systems:
First is operational: The first ground fault on a system causes only a small ground current to flow, so the system may be operated with a ground fault present.
Second is economics: No expenditures are required for grounding equipment or grounded system conductors.

Numerous advantages are attributed to grounded systems, including greater safety, freedom from excessive system overvoltages that can occur on ungrounded systems during arcing, resonant or near-resonant ground faults, and easier detection and location of groundfaults when they do occur.

I am not familiar with grounding issues on vessels, but I assume the main concern would be to keep the power on as long as possible, especially on sea. (Keep boat running with maybe a ground fault present)

Failure seldom stops us, it is the fear for failure that stops us - Jack Lemmon

Make the best use of Eng-Tips.com
Read the Site Policies at FAQ731-376

The problems with ungrounded systems are pretty well documented but just some highlights:

The general problem is the risk of overvoltage.  When one phase goes to ground, the system remains in service but voltage on the unground phases goes to 480V to ground instead of a nominal 277V.  

Also, due to the capacitive nature of the grounding, an arcing fault can result in a "repetitive restrike" that keeps adding voltage charge, resulting in voltages many times above the nominal 480V.  This results in a lot of insulation failures.  

In practice, many facilities just ignore the fault detection when one phase goes to ground.  When another phase goes to ground they then have a nasty phase-to-phase fault that will probably cause more damage than a ground fault on a well protective solidly-grounded system.

If starting from scratch in an industrial facility, I would go with a high-resistance grounded 480V system. In a commercial facility, I'd go with solidly-grounded.

Hello RalphChristie and
I believe that the main reason that ungrounded systems are chosen for vessels is unique to vessel applications.
In a normal on shore distribution system the ground path from a service to the supply transformer is in parallel with the neutral. This ground system includes the grounding system at the service, the grounding at the distribution transformer and often the multiple protective earthing on a common neutral.
As low as the impedance of the ground connection is, the dynamic impedance of a properly installed neutral is typically low enough that virtually all the neutral current flows in the neutral and very little of the neutral current flows in the ground system.
Not so in marine installations. The contact of a hull with the ocean is an extremely low impedance connection.
Consider a shore power installation. There may be a feeder running the length of the anchorage. Taps will be taken from this feeder to service various vessels.
The port authority may exersize little or no control over the load balance of vessels connected to shore power.
For example consider the shore power system is 120/208 volt 3 phase and that the vessel on the end of the line uses a single phase system. This vessel has a lot of single phase loading. The resulting neutral current is approximately equal to the line current. As a result, there is a voltage drop in the long supply circuit not only in the line conductors but also in the neutral. If this vessel has a grounded system much of the current will flow through the hull rather than through the neutral conductor. If the ship-board system is ungrounded there will be no hull currents.
Now consider the adjacent vessel. This vessel may have a perfectly balanced electrical system with zero neutral current, however if it is connected to the common feeder the neutral potential will be elevated from ground because of the neutral voltage drop in the feeder and if the vessel system is grounded then hull currents will flow. This will happen as long as the neutral connection to shore power is intact even if the vessel is not taking any power from the shore power system.
Hence, ungrounded systems, isolating transformers and ground connection switches.
I agree with most of the previous posts regarding normal on-shore installations, but I understand that possible hull corrosion is the main concern, and the main reason for using ungrounded systems at sea.

Quote:

Second is economics: No expenditures are required for grounding equipment or grounded system conductors.

How old is the green book and how long since it has been updated to recognise NEC requirements for equipment grounding?
respectfully

A followup on vessel corrosion. I spoke today with an owner and past manager of the repair facility that I mentioned.
Vessel corrosion is mitigated by sacrificial Zinc pads attached to the hull.
As I mentioned, the vessels in this area typically use grounded internal electrical systems. I was told that if a vessel is on shore power for more than about a week, the zincs are watched for accelerated corrosion. In a couple of weeks on shore power the loss of zinc may be noticable.
In regards to the sheet piling wharf; It is protected by 36 x 24 Lb. sacrificial zinc pads. The zinc pads are replaced every six months because of the accelerated loss of zinc caused by welding currents. The policy of using a ground cable to the vessel rather than to the sheet piling was difficult to enforce and the practice has reverted to grounding the welding machines to the sheet piling and semi-annual replacement of the zincs. That's over 1700 lbs a year of new zinc pads.
respectfully.

Frankiee,

I've seen the insulation resistance from phase to earth go as low as 5kohm and have spoken to people that have seen it lower still on marine networks.  You're a braver man than I if you are intending to touch a phase conductor on a IT system.

Check out these docs....
http://www.mikeholt.com/files/PDF/ect172.pdf

http://www.schneider-electric.com/cahier_technique/en/pdf/ect178.pdf

The entire Chinese medium voltage distribution network is an ungrounded 10kv system, 50hz. They make it a practice that once a single phase to ground fault is detected at the substation, they have 2 hours to go out and patrol the lines to locate it and remove it.

Shanghai is now using more and more underground distribution system.  There has been a review if they should change to a grounded system like most others in the world.  I heard one complaint about the ungrounded system is that with more and more cable system, they are saying that the cable charging current is getting to be difficult to manage.  I could not readily see the connection.  Can you guys tell me why it is so ?

Having the 2 hour rule also means that they must find a way to quickly locate the single phase to ground fault location.  There has been all kinds of crazy schemes claiming that they can use a unique measurement technicque to locate the fault.  The latest claims to involve the 'wavelet' technology.  Do you guys know about this and if it works, how does it work ?

Overhead distribution systems are probably more forgiving of the overvoltage issues that trouble ungrounded systems.  With increased undergrounding, I predict a lot of outages due to cable insulation failures, arc-overs, unexplained "mystery faults", etc.  

Time-domain reflectometers can be used to search for faults on cables, but probably not while it is energized.