Do you mean 440 phase to phase on 3 phase, or L-N (as if any one phase)? The former is 10% up on a nominal 230/400 and  I would expect anything well designed to ride that out, medical or not, while  the latter would suggest some serious fault,
PV will reverse the direction of the voltage drop in the wiring to it when generating, so long thin feeder runs are unwise, as they drive up the voltage at the end where the PV connects and that can be the limit on how much you can push in.. (*)
Mike

edited for typos and clarity.

Edit
* this is also another  reason it is recommended that PV circuits have their own submain if not back to the origin then at  least back to a point where the supply impedance is low enough and/or  there are enough loads pulling down so that the voltage rise under generation conditions does not become  an issue.

PV inverters are meant to disconnect if the grid voltage goes beyond normal limits - so in theory shouldn't be the cause of voltages exceeding 253V L-N (230V+10%) or 440V L-L. The rules are getting a bit more involved these days (so things are more likely 'ride through' some slightly dodgy events rather than everything suddenly shutting off together which tends to cause even more serious problems for the grid)  which might mean there are a few occasions when normal limits are apparently exceeded but generally things should remain within bounds. Similarly most inverters are adjustable and it might not be beyond the realms of possibility that someone's upped a limit (perhaps to allow for voltage drop on the connection circuit, with the intention that things still shouldn't go above 253V at the point of connection) but again if it's been done properly it shouldn't be too much of an issue.

  - Andy.

Werl... It is possible, especially with a campus like a hospital that loads between the embedded generator (of any type) would mask a rising voltage at the point of supply. Such as if the transformer was tapped at a low ratio historically, and the local HV is now running higher than nominal due to reduced load.

Or if there's a "voltage optimiser" in series and someone forgot to adjust the G59/G99 interface settings on the downstream SSEGs (which would be a breach of the connection agreement).

I think that if there were significant numbers of scanners keeling over due to excess voltage it would at least make it to the local newspapers.

Wouldn't a hospital which is big enough to have an MRI scanner have its own transformer?

Such as if the transformer was tapped at a low ratio historically, and the local HV is now running higher than nominal due to reduced load.

That's a fair point. Historically it's no unknown for transformers supplying installations with a hefty "base load" and long lengths of LV distribution cables to be tapped to something higher than normal - to counteract the "inevitable" voltage drop in the distribution system. If however local PV (or other on-site generation) effectively reduces demand from the grid, the voltae drop along the distribution system would also drop - leading to higher than intended voltages at appliances. Not really the fault of the PV though - the same might come about just by conventional energy efficiency measures.

    - Andy.

I have just had a CT scan at Musgrove Park hospital, Taunton and enquired of the technician as to the electrical arrangements. Was told that the machine has its own supply from the nearby substation and the voltage fluctuations are seldom a problem.

It is designed to work from 340 volts (about the lower limit of a European nominal 380 volt supply) up to 440 volts (about the upper limit of a UK nominal 415 volt supply)

We fitted one in a chest cancer depatment, it had been spec'd to have a huge supply cable, 4 core 240mm iirc, yet the power requirements of the unit were quite small. Anyway, in the large control / input panel, there was a transformer that weighed around 100kg. I had nothing to do with the internal wiring, but presumed it was acting like an isolating transformer, to give separation from the mains supply.

Some designs can be very thirsty when actually scanning - consider page 7 of this manual from Siemens showing 120kW per source - but of course most of the time it is not illuminating the patient at all, during set up and positioning, and if even it is, not always at absolute full power - so other data suggests that only about 20kW of cooling is needed. The risk of a long thin supply is if supply droops too far during those very occasional calls for full steam ahead....
( I am aware of versions that have been repackaged for field hospitals, where reliable power supply is more of an issue, and these need very careful management to avoid electrical problems..)

Mike

Yes, a CT scanner was deployed to Camp Bastion.

That second link mentions average power of 20 kW, but averaged over what? The max consumption is 125/140 kVA.

It's a wonder that the hospital lights don't dim when they are scanning!

Averaged over the scan - which is a series of 'slices' where the source at is rotated about the patient and data acquired, and then the patient is moved in or out of the scanner fractionally and another slice of the scan is taken, In the meantime some very clever software is used to stretch and scale the 'images' to form the view that is presented to the radiographer. between doing this, the source is turned right down or off. It is possible in some designs to have the patient moving steadily and a helical scan, then the load is more constant, but the image stitching problem is harder.

It would require very large capacitors to store enough energy as DC to store enough energy for a whole scan and smooth this out, so it s easier to put a bursty load across the mains,  and to live with the need for a dedicated sub-main that looks like it is being under-run,  on average.

Compared to those hospitals that have a cyclotron in the basement for the preparation of nuclear medicines, it is still not a huge load -those can be in the low hundreds of kW, but tend to be run at full power for signficant  periods.

Mike,