If flow and return water is so contaminated that it is akin to thick mud then, yes there could be a problem but I would think that such flow and return water as you envisage might be pretty much insulated providing the the breaks you mention are a suitable length

confirmation of electrical separation between the second (new) supply and the building's existing electrical supply.

My first question is what do the DNO mean - separation of the live conductors (Ls&N) or separation of everything, including PE? The former is usual in many situations (e.g. terrace of houses sharing metallic water/gas supply pipework). The latter will be much more difficult to achieve, in many situations (e.g. steel framed buildings) impossible. Trying to have two separate equipotential zones in the same single occupation building might be 'interesting'.

PME often confuses matters - with some DNOs occasionally only providing an earth terminal for the first supply and insisting subsequent ones were TT'd - but there are alternatives - e.g. having one common earthing point outside the building and separate N & PE conductors from there.

   - Andy.

The water is unlikely to be live unless something has gone badly wrong, it is however likely to be at or near earth potential - and it is as above the separation of the two earthing systems that is tricky, but even before you start  the disparate systems will be connected by adventitious earthing to the order of tens to hundreds of ohms anyway via their substation electrodes and anything metal that sits on the ground.

Pure water is actually a good insulator (many megohms per cubic cm), and where you have to that can be maintained with propylene glycol antifreeze and ion exchange resins, but the addition of typically required  corrosion inhibitors and brass pipe fittings tends to spoil that ideal somewhat. Even so,  if a few feet of 22mm PEX pipe can be tolerated, the water column resistance of tap water mixed to the recommended dose with some Fernox F1 or similar can be kept to tens of K ohms or higher, so negligible coupling.

(The corrosion inhibitor is an alkali so makes the water more conductive than straight out of the taps, to the tune of TDS of 300-800 ppm equivalent or perhaps a few hundred to a k or two ohms per cm cube .)

Mike.

Is the water within plastic pipework in public water supply network an electrical hazard?

No, as by the time it is impure enough to be a significant conductor it is too contaminated to be safe to drink, but wastewater maybe, depending what is in it. The salt water from water softeners being cycled and swimming pool wastes are two examples that need a bit more thought. Even so by the time the liquid column is longer than a few hundred times its diameter, it is a pretty good resistance.

Incidentally we use water with controlled doses of washing soda in plastic tubes as fail to safe high voltage bleed resistors in the pulse power lab at work . Very cheap, easy to maintain and a visible indicator of faults.

Mike,

Whilst "saving" £50k sounds like a good idea, it may not be in the longer term. How much is your significantly large heat pump and installation costing? Is it likely that you may want more power in the future, for car charging for example? Have you discussed the tariff for the power fully, large users can get better prices? How much extra will the second supply cost for installation compared to the one you already have. Have you considered the fact that your own local substation is likely to give you better supply security, and that if you need a generator makes connection much simpler? In the bigger picture £50k is not a lot of money when you potentially use 500 kVA, with the way prices are going it could be only about 100 hours usage!