# Electrically de-icing cable-stayed bridges

Discussion in 'Architecture & Engineering' started by Peter Dow, Feb 17, 2020.

1. ### Peter DowRegistered Senior Member

Messages:
215
What started off as my politically-motivated blog post last week, I've since been elaborating on, in an engineering design style, so I thought I should get some professional help.

Please Register or Log in to view the hidden image!

Please Register or Log in to view the hidden image!

BBC: “Falling ice causes first Queensferry Crossing closure”

Keeping such bridges open even in icing conditions is really not rocket science. What, to me anyway, is the obvious solution – to pass an electrical heating current through the bridge’s support cables – doesn’t seem to be “obvious” to other research scientists and engineers whose “Thermal Systems” for melting the ice are reviewed here.

I suggested this simple solution, outlined the calculations required and warned of some dangers in an email to the Queensferry Crossing bridge authorities and contractors in March 2019, but as usual, the authorities ignore solutions until there is a political price to be paid for continuing to ignore solutions in a pig-headed, in-denial kind of way that politicians like to get away with, if they possibly can.

There follows a link to a PDF of the email I sent the bridge authorities last year – hopefully you can click the link and open and / or download the PDF so you can read it.

Queensferry falling ice hazard solution – electrically-heated cable stays

Deicing power for 70km of cables
@ 100W/m = 7MW = household electricity within a 3 mile radius of the bridge.
@ 250W/m = 17.5MW = household electricity within a 5 mile radius of the bridge.

Cable strands

Please Register or Log in to view the hidden image!

Some strands in the cable are better situated for heating the cable than other strands, depending on their position in the cable as I have labelled them alphabetically, beginning with the label “A” for the centre strand (which is the worst strand for heating the outside of the cable, where the ice would be) and labelling the outer strands last in alphabetical order, which are best for heating the outside of the cable.

The cable strands are by convention named here using the format – “(Number of strands in the cable)-(Letter)”. Thus the centre strand in the 55-strand cable is named as “55-A”, the 6 strands immediately surrounding the sole 55-A are all named of type “55-B”.

For each strand in the cable we can assign a factor of heating capacity.

Please Register or Log in to view the hidden image!

For the 55-strand cable, total heating capacity factor assigned is 48.

For the 55-strand cable, there are a total of 24 strands which have utility for heating the cable – 6 of the 55-F type name strands, 6 x 55-Gs and 12 x 55-Hs. The 31 other strands (the 55-A to 55-Es) are not needed for heating per se, though could carry electrical currents whether by design or otherwise.

We can tabulate for each strand label, the heating power fraction and percentage, according to each strand’s heating capacity factor as a fraction of the cable’s total heating capacity factor.

Please Register or Log in to view the hidden image!

Please Register or Log in to view the hidden image!

For the 61-strand cable, the total heating capacity factor assigned is 54.

For the 61-strand cable, there are a total of 24 strands which have utility for heating the cable – 6 x 61-Gs, 12 x 61-Hs and 6 x 61-Is. There are 37 other strands – the 61-A to 61-Fs.

Please Register or Log in to view the hidden image!

Please Register or Log in to view the hidden image!

For the 73-strand cable, the total heating capacity factor assigned is 54.

For the 73-strand cable, there are a total of 30 strands which have utility for heating the cable – 12 x 73-Hs, 6 x 73-Is and 12 x 73-Js. There are 43 other strands – the 73-A to 73-Gs.

Please Register or Log in to view the hidden image!

See my blog post for details for 85-, 91- and 109- strand cables.

Last edited: Feb 17, 2020

to hide all adverts.
3. ### Peter DowRegistered Senior Member

Messages:
215
VSL SSI 2000 Stay Cable System
There are a number of options available in the VSL SSI 2000 Stay Cable System so these figures cannot be confirmed without sight of the Queensferry Crossing engineering design specifications (or by actually measuring the cables, which I am unable to do!).

Please Register or Log in to view the hidden image!

For now, I am assuming for simplicity that the required maximum heating power in watts/metre is the same as the stay pipe diameter in mm. This is not far off the maximum heat radiation from the sun on such a stay pipe, square on to the sun, at midday, midsummer, on a cloudless day – or more than enough heat to melt any ice in short order!

Please Register or Log in to view the hidden image!

It is now possible to tabulate for each cable-label strand, the maximum heating power per metre and assuming a strand resistance of 0.001137 ohms per metre, what the maximum strand current would be.

Please Register or Log in to view the hidden image!

Cable voltages and power
To calculate the cable voltages and power and to calculate the total maximum power to heat all the cables of the Queensferry Crossing accurately, I will need to know how many of each size of cable and their lengths.

Direct Current Heating
Those theoretical differences between strand situations only matter for direct current heating if it is possible electrically to isolate strands from each other. The strands are attached via steel wedges to a steel anchor head, which, for now, effectively connects all the strands together electrically.

Please Register or Log in to view the hidden image!

Cable anchorages

Please Register or Log in to view the hidden image!

Teflon/PTFE-coated glass fibre fabric sheaths to electrically isolate the strands from the anchor head. The outer strands are for heating. The inner strands are for signals.

It should be possible to insert Teflon/PTFE-coated glass fibre fabric sheaths between the wedges which grip the strands we wish to insulate and to isolate from the anchor head and from each other, unless and until they are connected to electrical heating or signal circuits.

The signal circuits could be used to report to the power supply control electronics at one end of the cable, the output of heating current sensors at the other end of the cable, to help to detect current leakage faults in the cable strands’ insulation, to implement a residual current device, to trigger safety power-cut-outs or circuit-breakers, most notably.

Teflon is a good insulator and is used for thread seal tape illustrating the properties of lubrication of the wedge to its housing cone required. The glass fibre fabric should provide strength under compression and a superior dimensional stability versus creep under load that a pure Teflon sheath may suffer from.

Clearly the sheath would have to remain thick enough to insulate against the highest voltage difference which might appear between the heating strands and the anchor head.

Such sheaths would likely not be available as an off-the-shelf product in the required dimensions (though general purpose PTFE-coated fibre glass cloth is commonly available) and would likely require to be custom manufactured, tested and proved in the laboratory.

So isolating the strands for DC heating purposes presents technical challenges. It would be very convenient if the outer strands could be preferentially used for heating purposes without having to isolate the strands electrically etc. but to achieve that we must consider using not direct current but alternating current instead.

Alternating Current Heating
The skin effect observed with alternating current changes matters in that with increasing frequency the heating current will tend to distribute towards strands nearer the surface of a cable. However if too great a frequency is used then the skin effect will increase the resistance of even the most superficial strands so much that inappropriately high and difficult to insulate against voltages would be required to obtain the required heating power.

Assuming that the appropriate AC frequency can be determined for preferentially heating the superficial strands of the Queensferry Crossing stay cables, although there would be no need to isolate the strands from the anchor head, there then presents the challenge of isolating the anchor heads and anchorages so that the current is not dissipated through the bridge instead of heating the cables as required.

Having isolated the cables for heating purposes, one may then wish later to reconnect the cables electrically to the rest of the bridge and disconnect the heating power supplies for lightning protection purposes. Certainly, one would not wish to encourage a lightning strike to find its way to ground via the bridge’s cable deicing power supplies!

Tower ice
To prevent the bridge piers or towers (with non-conducting concrete surfaces) from icing up, they could be surface fitted with new electrical heating trace cables which are then appropriately electrically-powered for deicing when necessary.

Ideally, such additional heating elements would have been embedded into the surface of the piers at construction time. Too late for that now.

Another option to consider is heating the hollow piers from within. However, considering the considerable mass and thickness of the piers their surfaces would have to be kept above freezing temperature all winter long. Heating the piers from within, there simply wouldn’t be time to allow the piers to get freezing cold because there was no icing then suddenly heat them from the inside to deice a sudden incidence of icing.

So heating from within bridge piers would use more electricity, though the cost shouldn’t be prohibitive – surplus grid electricity is a common occurrence at times of high wind power generation, so the electricity grid managers should offer a very low price for such electricity (just the grid connection charge) – plus it should be a lot safer upgrade from the point of view of bridge users – far less chance of things falling onto the road during the fitting of the piers’ internal heating elements.

Last edited: Feb 17, 2020

to hide all adverts.
5. ### (Q)Encephaloid MartiniValued Senior Member

Messages:
20,855
How well do the ice collars and ice sweepers work? My understanding is that what was used on other bridges with the same problems.

to hide all adverts.
7. ### Peter DowRegistered Senior Member

Messages:
215

The technicians have to be dangled on a cable in icy cold weather to deploy the chains and they have to retrieve them for re-use.

What if a technician falls to his death or drops a chain onto someone below, killing them? It's not a fool-proof method.

Also, the Queensferry Crossing stay cables are protected by a stay pipe made of HDPE, which could take some scratching by metal collars / chains scraping the ice off.

I prefer electrical heating which can be fully automated.

(Q) likes this.
8. ### Peter DowRegistered Senior Member

Messages:
215
This expandable E-glass sleeving, expands from a relaxed internal bore of 15mm to a maximum bore of 38mm and insulates to 500V when not expanded, which is a useful size while relaxed to accommodate the strand and while expanded to accommodate the wedges.

The insulation should cope with the highest DC voltage of about 100 Volts, used to power the longest and highest heating capacity factor strands, albeit that this sleeving is inappropriately resin-coated and would therefore likely require to be custom adapted, the resin cleaned off and re-coated with PTFE, tested and proved in the laboratory. Perhaps wrapping the wedges in PTFE thread seal tape is all that is required to supplement the product as supplied for satisfactory performance? A promising avenue for research.

Please Register or Log in to view the hidden image!

DC Power Supplies
Not forgetting DC power supplies and I have noticed a comprehensive range of 3kW to 10kW DC power supplies here that I think will do nicely, an average of about a dozen power supplies per cable (more for the longer cables, fewer for the shorter cables), about 3500 power supplies required to de-ice all 288 cables.

Please Register or Log in to view the hidden image!

Where to store the cable power supplies?
Let’s examine the option of storing the cable heating power supplies in the towers, racked next to the anchorages of the cables which they will be heating. There might just be enough room to squeeze in another half a tonne of power supplies for the 4 cables per floor (assuming their racks are securely attached to the tower walls), 12 tonnes worth of power supplies for all 24 floors per tower, for all 3 towers!

Even at 94% efficiency for switch mode power supplies, each tower’s cable power supplies could be generating at most about 0.4 MW of waste heat energy. A new massive extractor fan fitted into the roofs of the towers would be required to cool the inside of the towers while the DC power supplies are heating the cables.

Considering how cramped the insides of the towers are already, the daunting cooling problem, not to mention the risk of a tower fire destroying all of a tower’s power supplies at one time, it looks to be much the better option to install the cable power supplies on the deck, next to the deck anchorages to allow them to be supplied with power.

Please Register or Log in to view the hidden image!

Heating the towers may be as simple as a big electric heater on the ground floor, the warm air rising up the insides of towers in between the open stairways and scaffolding.

Last edited: Feb 22, 2020
9. ### Peter DowRegistered Senior Member

Messages:
215
The stay cables penetrate the surface of the deck, as can be clearly seen in this next photograph, taken during construction.

Please Register or Log in to view the hidden image!

Therefore best access to the anchor heads, to attach the cable heating power supplies, may be from inside the deck, where the power supplies themselves should be stored too.

Last edited: Feb 23, 2020
10. ### Peter DowRegistered Senior Member

Messages:
215
DC Circuit Diagrams
Locating all the electrics at the deck anchorages, while leaving the strands earthed at the tower anchorages, offers advantages for design, development, installation, commissioning and servicing.

Please Register or Log in to view the hidden image!

Circuit Diagram – 2 heating strands, 1 power supply

Please Register or Log in to view the hidden image!

Heating strands pair current balance detector

The window detector circuit compares the isolated power supply’s potential with respect to earth to detect the expected balance of current and voltage in the heating strands pair. If an imbalance fault develops then the safety switch is used to cut the power.

Last edited: Feb 25, 2020
11. ### Peter DowRegistered Senior Member

Messages:
215
"Effectively"? Perhaps I should have used the word "incidentally".

Please note, however, that when introducing a design requirement to conduct large electrical currents between strand pairs at the tower anchor heads (see DC Circuit Diagrams) the incidental electrical connection at the wedges may be of insufficiently or unreliably low resistance and should be supplemented with an ultra-low resistance connector between the strand ends, to avoid faults developing from excessive resistance heating at the wedges.

12. ### Peter DowRegistered Senior Member

Messages:
215

Please Register or Log in to view the hidden image!