One of the surprising results of the precipitous decline of photovoltaic panel pricing (which has lowered the cost of energy in sunny places), is that the tempered glass used to prevent the paper-thin PV silicon wafers from being broken by hailstones, or rocks thrown by rowdy teenagers, is now nearly as expensive as the PV cells themselves.
Greg Sittler tells me that one solution to this is to protect the PV panels with chicken wire, which absorbs some 10% of the insolation, suspended some distance above them. Chicken wire is tossed-rock-proof and hailstone-proof, and typically withstands a few decades of weathering. Galvanized 20-gauge chicken wire is about US$2.90/kg (from wireclothman), which is about US$1.30 per square meter for the “fine” 1-inch (25-mm) mesh size, or about US$3 per square meter from Ace Hardware. This seems like a very good solution to me. (You might need two staggered layers in order to reliably stop small rocks.)
Chicken wire is made of galvanized mild steel, which mostly deforms plastically upon impact, thus being broken by a sufficient number of repeated impacts. However, its hexagonal structure is better suited to absorbing impacts than a square mesh, because it’s inherently stretchy, though not quite so stretchy as a knit-fabric structure would be; a knit-wire structure would be able to absorb much more impact energy for a given amount of wire by virtue of experiencing more macroscopic deformation, just as a hexagonal mesh can absorb more than a square mesh. Alternative materials that occur to me include music wire, nylon, PTFE, UHMWPE, rubber, polycarbonate, glass fiber, basalt fiber, polyimide, polyamide-imide, glass-fiber-reinforced polyamide-imide, and polyurethane.
The ideal material for this purpose would combine a high mechanical energy capacity like rubber (2–9 kJ/ℓ), nylon (0.3–2 kJ/ℓ), or ASTM A228 music wire (11–14 kJ/ℓ); excellent ultraviolet resistance like steels, PTFE, glass fiber, and basalt fiber; and low price, like nylon. Many of the polymers listed above can be heavily filled and mixed with UV-absorbing and free-radical-scavenging components in order to improve their UV resistance, but generally not to the decades of endurance in thin fibers needed for maintenance-free PV operation.
An alternative, then, suggested by Greg, is to combine virtues by using fairly stiff but UV-proof fibers like glass fiber, mild steel, or PTFE to form “trampoline panels” that are suspended around the edges by springs made from a material with a higher mechanical energy capacity. These springs might be coils, knit fabric, cantilevered leaf springs, zigzag fibers, foam blocks, or in some other shape, but at any rate they can be protected from the sun, so they can be made of cheap materials like nylon, polyurethane, or perhaps mild steel.
Both the springs and panels might need to resist creep, which might require using a high-melting material rather than an organic polymer, though teflon and polyimide might be good enough. Also, though, it might be the case that creep is not a real concern here, because the normal relaxed loading scenario is a tiny fraction of what the protective layer must be designed to tolerate during an impact. So, for example, according to Du Pont’s Teflon PTFE Properties Handbook, at room temperature teflon creeps by 100% in a few hours under a 10-MPa load (close to its ultimate strength), but takes hundreds of hours to creep by 1% under a 3.5-MPa load.
It might be helpful to use multiple different fiber sizes, like the ripstop nylon used in parachutes. A small rock of 10 mm diameter might typically weigh 3 g and be hurled at some 15 m/s, thus carrying some 300 mJ. Stopping it within 100 mm thus requires at least 3 N of force along its direction of motion. Suppose it strikes a single strand of the net, which deforms to catch it; then these 3 N might be 15 N in each direction along the strand, so the wire must withstand some 15 N.
The fiber diameter needed to resist this rock depends on the material chosen (for the trampoline panels, if those are used). Teflon’s ultimate strength is about 10 MPa; that of rubber, about 16 MPa, depending on fillers; that of polyimide (Kapton), about 200 MPa; that of mild A36 steel, about 500 MPa, though its yield stress is lower, around 200; that of nylon, about 900 MPa; that of music wire, gel-spun UHMWPE, or E-glass fiber, about 3 GPa; that of S-glass or basalt fiber, about 5 GPa.
So, depending on the fiber chosen, you might need a fiber of 1.4 mm (of teflon, suggesting that teflon may be too weak for this) or of 60 μm (of basalt fiber) to stop the small rock.
But consider a larger rock, 100 mm in diameter, weighing 3 kg, thus carrying some 300 J of energy. Stopping it in the same 100 mm requires 3 kN of force, or perhaps 15 kN if it is being stopped by a single strand, requiring a 44-mm-diameter teflon bar or a 2-mm-diameter basalt-fiber rope. If the strands are 10 mm apart then perhaps we can ensure that it strikes at least 27 of them (9 in each of three basket-weave directions), so perhaps the load is only some 600 N each, requiring teflon fibers of “only” 8 mm diameter, converting the rock shield into a very effective and expensive PTFE sunshade, or 400-μm basalt rope.
So, the ripstop approach would be to make, say, 90% of the fibers thin, weak, and cheap, to catch the small rocks, and the other 10% stronger, either by making them thicker or using a stronger material.
The stronger “ripstop” or “reinforcement” cables can be thick enough to carry a thick UV-protection layer, made, for example, of carbon-black-filled teflon. Even UV-protection fillers in a polymer might slow the degradation of such a thick cable to a tolerable degree during the panels’ design lifetime.
For example, you could use 0.2-mm (AWG 32, much thinner than usual 20-gauge 0.8-mm chicken wire) galvanized mild steel wire spaced 10 mm apart, then 3-mm (AWG 8) music wire every 100 mm, which you would also have to galvanize. If the weave goes in three directions, this works out to 300 m of thin wire (30 g) and 30 m of the thick wire (2 kg) per square meter.
Unfortunately at onlinemetals.com, 0.125-inch music wire costs US$15.79 per pound, or US$34.80/kg, 11 times the price (by weight) of chicken wire; so 2 kg/m² is US$70 per square meter; while PV modules, including the glass, currently only cost about US$40 per square meter, so this is still too expensive. (MSC offers a similar price; it’s not just onlinemetals.)
Given that, though, I’m pretty sure it’s possible to relax the problem to get the cost down to a reasonable level.
Probably what I need to quickly vet materials for such uses is a cost per newton meter: the cost for a meter of cable thick enough to resist a load of one newton. Music wire at US$15.79/pound times 7.9 g/cc divided by 3 GPa is about 90 microdollars per newton meter, while mild steel at US$2.90/kg times 7.9 g/cc divided by 500 MPa is about 46 microdollars per newton meter. (Maybe the bulk metal is cheaper than chicken wire, though.) Nylon rope costs US$81 for 250 feet of ⅜" 3050-pound-test braided rope, or in SI units, 76 m of 10-mm 1380-kg-test braided rope; that’s about 80 microdollars per newton meter, but not UV-resistant.
Amazon suggests “Campbell galvanized steel wire rope, 7×19 strand core, 3/16" bare OD, 250' length, 840 lbs breaking strength” for US$61.92 (“+$209.82 shipping & import fees” to Argentina), which is 4.8 mm bare OD, 76 m, 3.7 kN, and is claimed to weigh 16.5 pounds. From the weight presumably the cross section is only about 12 mm², so as a solid round steel rod its diameter would be 4.0 mm, the rest I suppose being air between strands. The strength would thus work out to 300 MPa, so either it’s a bit underrated or the spool weighs a lot; one buyer claims that the spools he’s tested broke at over 1000 lbs (4.4 kN), which works out to 360 MPa. Taking them at their word, though, this works out to 220 microdollars per newton meter, not including shipping; considerably pricier than the other alternatives.