DC String Cable Installation: Burial, Conduit or Drainage Pipe

In a utility-scale solar park, the DC string cables are buried once, usually when the EPC budget is tightest, and then have to work for the whole life of the plant. The three common methods, laying cables directly on sand bedding, pulling them through closed conduit, or running them in perforated drainage pipe, differ far more in how they fail than in what they cost to build. My position is that a well-designed perforated drainage pipe is the most reliable DC string cable installation method over the plant’s life, because it is the only one of the three that deals with both rodents and standing water. The cheaper options each leave one of those doors open, and the bill arrives in operation.

Why the DC string cable installation method matters more than its price

The Solar Bankability project ranked technical risks in operating PV plants by their combined repair and lost-production cost. In the cabling category, improper installation came out on top, with a cost impact of the same order as the leading PV module risks ^[1]. Its failure catalog reads like a list of trench problems. Among installation defects it lists bending radius too small, cable conduit not closed and DC underground cables not protected against mechanical damage. Among operational failures it lists cables damaged by rodents, which the report says happens mostly underground where cables are installed without conduits ^[1].

The reason a single cable defect is expensive lies in how inverters protect themselves. SMA’s transformerless inverters, as one example, measure the insulation resistance of the connected PV strings before connecting to the grid and refuse to connect if it is below the threshold ^[2]. SMA notes that the fault can be weather-dependent, appearing with rain or dew where cable insulation or plug connections are defective, which makes it hard to locate ^[2]. A ground fault that develops during feed-in is worse: the fault current from the grid side can exceed 1 kA and damage the power modules, and SMA states that such damage is not covered by the inverter warranty ^[2]. The current array design standard, IEC 62548-1:2023, revised its safety clause to cover array insulation monitoring and earth fault detection ^[3]. The fault will be detected, and on a transformerless inverter detection usually means the inverter stays offline.

The economics are therefore asymmetric. Until the fault is found, a damaged string can cost the output of every string on that inverter, plus the time to trace a defect that may only appear when the ground is wet, plus excavation if the cable is directly buried. Having worked across development, EPC management and asset management, I treat the cable installation method as a lifetime-cost decision, not a line item to value-engineer at tender.

Option 1: Direct burial on sand bedding

Direct burial is the lowest-capex method: open trench, a layer of sand, cables, more sand, warning tape and backfill. It has genuine merits. No pulling force is applied to the cable, and in free-draining ground any water that reaches it drains away, so the cable sees the intermittent wetting its standard was written for. Its weaknesses are mechanical and biological.

Rodents. Sand bedding protects insulation from sharp stones; it does nothing against teeth. This is the failure mode the Solar Bankability catalog ties specifically to cables buried without conduit ^[1]. In European farmland the pressure can be high. The common vole moves from grassland into agricultural land when grassland carrying capacity is exceeded, and during multiannual outbreaks its populations can exceed 1,000 individuals per hectare ^[4]. Solar parks built on former arable land or grassland sit in exactly this habitat. Rodent-repellent sheath compounds are available ^[5], but a repellent is a deterrent, not a barrier.

Damage before energization. Rodent and mechanical damage does not wait for the plant to age. Cables lying in an open trench or in fresh, loose backfill are exposed from the day they are laid, and a gnawed or crushed section can already be present at commissioning. In my view this is why direct burial can cost yield from the first year of operation, not only late in life.

Approvals. Direct burial is not part of the EN 50618 type approval. Prysmian, for example, declares its TECSUN (PV) H1Z2Z2-K suitable for direct burial on the basis of internal testing and field experience, states that this test is not part of the EN 50618 type approval, and requires burial guidelines to be followed ^[6]. If direct burial is chosen, the cable supplier’s written approval for that method, with its conditions, belongs in the project file.

Repairs. Every fault on a directly buried cable means locating it, excavating, and making a joint that becomes a new weak point in the circuit.

Option 2: Pulling through closed conduit

Closed conduit, typically corrugated HDPE duct, solves the rodent and mechanical problems as long as the ends are sealed. It introduces two new ones: pulling tension and trapped water. Conduit selection decides how serious the second one becomes.

Pulling tension and bending radius

PV string cables are thin, flexible single cores, and their allowed tensile loads are small. Prysmian’s TECSUN (PV) data sheet lists a maximum tensile load of 50 N/mm² during installation and 15 N/mm² in operation, which for a 6 mm² cable is 300 N and 90 N ^[6]. Three hundred newtons is roughly the weight of a 30 kg mass.

The pulling force rises along the duct and sharply in bends. In a bend, friction is proportional to the force already in the cable, so the harder the pull, the harder the cable is pressed against the inner wall, and the force grows exponentially, the capstan effect ^[7]. For a single 90° bend the force multiplier is e^(μπ/2): about 1.3 at a friction coefficient of 0.16 and about 2.2 at 0.5, the range of friction values used in that paper’s examples ^[7]. Two or three bends between the trench and the mounting structure compound this quickly, and a hand pull gives the crew no indication of when the limit has been passed.

Over-tensioned, abraded or tightly bent cables do not always fail at once. A stressed or nicked insulation can pass a dry commissioning test and surface later as intermittent insulation alarms in rain or dew, the pattern SMA describes for defective cable insulation ^[2]. My position, based on that mechanism, is that tension damage typically becomes visible after the first year of operation rather than at handover.

Water with no way out

A closed duct is not a watertight vessel. Water gets in through couplings, through open ends at the risers, and by condensation, and in a basic sealed HDPE duct laid without fall it has nowhere to go. Wiring rules address this directly. BS 7671, which is based on the IEC 60364 series, requires in regulation 522.3.2 that where water may collect or condensation may form in a wiring system, provision is made for its escape, and NICEIC guidance recommends drainage points and ventilation in conduit systems that are not designed to be gas-tight ^[8].

The nuclear industry reached the same conclusion: U.S. NRC aging guidance for inaccessible medium-voltage cables notes that cables in conduits can be exposed to condensation and wetting, and calls for periodic inspection for water in manholes and conduits, with draining as needed ^[9]. Sandia’s review of submerged cable research identifies moisture as a significant aging stressor and describes water treeing, a form of insulation damage driven by water and electrical stress that is believed to be permanent even after the cable dries out ^[10]. That evidence comes from medium-voltage AC cable and should not be transferred one to one to 1.5 kV DC, but the direction of the risk is the same: insulation that sits in water for years carries an aging stressor that intermittently wetted insulation does not.

What AD7 and AD8 actually mean for PV cables

The presence-of-water classes come from IEC 60364-5-51. AD7, immersion, covers the possibility of intermittent partial or total covering by water. AD8, submersion, covers equipment that is permanently and totally covered ^[11].

EN 50618 and IEC 62930 ^[12], the two PV cable standards, define PV cables for an AD7 condition ^[13]. Their water-related requirement is a long-term DC insulation test of 10 days in 85 °C water at 1.8 kV DC ^{[6, 13]}. Neither standard qualifies a cable for permanent submersion. H1Z2Z2-K variants marked AD8 are on the market, but Prysmian, which sells one, describes the usual AD8 marking as a manufacturer self-declaration without a reference standard, based on a test designed for AC cables rated up to 450/750 V, a situation “far removed from the reality of photovoltaic installations” ^[5]. The test usually borrowed comes from EN 50525-2-21, Annexes D and E: up to 100 days in 50 °C water under AC voltage ^[13]. Some manufacturers add proprietary DC water tests; Prysmian’s runs at 1,800 V DC in 70 °C water for more than 12,000 hours ^[5]. Cable engineers have also reported that water in the field, combined with the move to 1,500 V systems, is exposing weaknesses in some PV system components ^[13].

The accurate version of the point is this. AD8 is a recognized class in the IEC 60364 wiring rules, but the European and international PV cable standards do not qualify cables for it. A standard H1Z2Z2-K sitting permanently in a flooded duct is operating outside its qualified condition, and an AD8 marking is only as good as the test behind it. Where water can stand around the cable, ask the supplier for the AD8 test method, voltage type, temperature and duration.

Option 3: Perforated drainage pipe

A perforated, or slotted, corrugated pipe is the approach I recommend for most sites, for reasons that follow from the two failure drivers above. The pipe wall is a physical barrier against rodents and digging, as a closed conduit is. The perforations let water that enters the pipe drain into the bedding, so the cable experiences intermittent wetting, the AD7 condition its standard covers, instead of standing water. The initial capex is higher than direct burial, but in my view that difference is small next to a single inverter-level fault hunt over the plant’s life.

The honest limits are those of any drain, and they decide whether the solution works:

• A drain collects water as well as releasing it. Subsurface drains are designed to take in groundwater along their length and carry it to an outlet ^[14]. In clay, or where groundwater rises above the pipe, a drainage pipe without fall and outlet fills like a closed duct. It needs a continuous fall to a discharge point above the groundwater level, or it should not be used.
• Fine soil enters through the perforations. FAO drainage guidance notes that water drag carries soil particles into the pipe through the perforations, which can never be fully prevented, only slowed or stopped with an envelope around the pipe ^[14]. In silty or fine sandy soils, specify a geotextile-wrapped pipe or a free-draining surround.
• Roots can penetrate drainpipes ^[14]. Keep runs away from hedges and tree lines.
• Open ends defeat the barrier. Slots do not stop a rodent that walks in through an open pipe end. Cap or mesh the ends at risers.
• Pulling still applies. If cables are pulled through the pipe, every tension rule from the conduit option applies. Generous pipe diameter, short runs, few bends, a lubricant compatible with the cable sheath and measured pulling force keep it under control.

Decision table: direct burial vs closed conduit vs drainage pipe

Installation QA checklist

Design and procurement

• Choose the method per area from a soil and groundwater survey, not one method for the whole site.
• For drainage pipe, show fall, outlet points and outlet levels on the drawings, with outlets above the expected groundwater level. Specify an envelope for fine soils.
• Obtain the cable supplier’s written approval for the installation method, its maximum tensile load and minimum bending radius and, where AD8 is claimed, the test standard, AC or DC, voltage, temperature and duration ^{[5, 6]}.
• Size the pipe so that cables pull without jamming, and limit the number of bends per run.

Installation

• Keep the trench bottom free of stones, with sand bedding below and above the cables or pipe.
• Measure pulling force with a dynamometer or a tension-limited winch on representative runs and keep it below the manufacturer’s installation limit. Use proper pulling grips.
• Protect cables from sharp edges at pipe entries and respect the minimum bending radius at risers.
• Cap or seal duct and pipe ends against rodents and debris on the day of installation.
• Keep the time cables spend in open trenches short and backfill promptly.

Testing and handover

• Test insulation resistance on every string before connection. The IEC 62446-1 commissioning regime includes a PV array insulation resistance test and, as an additional test, a wet insulation resistance test ^[15].
• Record the values per string as a baseline for later comparison.
• Use the wet insulation test on sections where damage is suspected.

Operation

• Track inverter insulation events and compare them with rainfall; a weather-linked pattern points to a cable or connector defect ^[2].
• Inspect drainage outlets after heavy rain and keep them clear.
• Repeat insulation measurements on suspect strings before the defects liability period ends.

Conclusion

The cheapest trench is cheap only on the day it is backfilled. Direct burial leaves cables open to rodents and later digging. A closed duct protects the cable but adds pulling risk and can hold the cable in water its standard never qualified it for. A perforated drainage pipe, designed as a drain with fall, outlets, envelope and sealed ends, addresses both failure drivers. It costs more at construction, and it is the option I recommend when a plant is judged on its whole operating life rather than on its tender price.