Securing a Solar Farm or Battery Site: Cable Theft, Remote Sites and No Mains Power
A solar farm is an unusual security problem, because it combines the worst features of two very different sites. It has the asset profile of an industrial installation: high-value copper, inverters, transformers, switchgear. And it has the operating profile of a remote field: nobody there, no mains supply where you need it, patchy signal, and a perimeter measured in hundreds of metres across open ground. The standard industrial answer, a fenced compound with mains cameras and a monitored alarm, assumes a compact site with power and a comms room. Most renewable sites in the UK have neither. Last updated: November 2026. What gets stolen, and when Cable, overwhelmingly. DC string cable and AC runs between inverters and the substation. It is copper, it is accessible, and on a site under construction or mid-repower it is often lying in runs rather than buried. Theft of cable is also the most damaging per pound stolen, because the replacement cost is dwarfed by the lost generation while the site is down. Inverters and optimisers, particularly on smaller sites and rooftop arrays. Batteries, on BESS sites, though these are heavy enough that it tends to be organised rather than opportunistic. Copper earthing, which is quick, quiet, and leaves a site unsafe as well as poorer. Tools and plant during construction, which is a different problem with the same answer as any construction site. Timing is predictable. The highest-risk windows are during construction, during maintenance or repowering when cable is exposed, and on long weekends and holidays when nobody is scheduled on site for several days. Why the fence is not the answer Most sites have a deer fence or a stock fence. It keeps out livestock and nothing else. Upgrading to a security fence around forty acres is an enormous spend for a modest gain, because anyone willing to cut a fence will cut a better fence. The leverage is not the perimeter. It is detection, identification and response, in that order, and all three are harder on a remote site than the brochure suggests. The power problem, honestly Here is the irony that catches people out: a solar farm cannot usually power its own cameras. The array feeds an inverter and goes to the grid. There is no low-voltage supply sitting at the gate, and taking an auxiliary feed means an electrician, a DNO conversation, and a permanent install that nobody wants to do for a camera. On a site under construction, the array is not generating at all. Which means cameras on these sites are nearly always independently solar powered, with their own panel and battery, exactly like a remote farm camera. And they have to survive the same British winter, which is the thing that actually determines whether a camera is still working in February. The sizing arithmetic is in our solar CCTV through a UK winter guide, and it applies here unchanged. The connectivity problem Most sites have 4G of some sort. The question is whether it is good enough at the point where the camera needs to be, which is frequently the far corner of the site rather than the gate. Three approaches, in order of how often they are the right answer: 4G per camera. Simple, independent, and a SIM per camera. Fine for two or three cameras. Expensive and fiddly at ten. One connection, local links back to it. A single 4G or 5G router at the best signal point, with cameras linking to it over long-range wireless. On a site measured in hundreds of metres across open ground with no obstructions, this is where WiFi HaLow is genuinely strong: sub-GHz, long range, low power, and no SIM per device. Our guide to choosing between 4G, LoRaWAN and HaLow sets out the trade-offs. Fixed line. Occasionally available at the substation. Rarely where the cameras are. What to actually cover Not the perimeter. Four things: The access point. There is almost always exactly one vehicle entrance, because that is how the site was built. Everything that leaves heavy goes out through it. A camera with plate capture here is the highest-value device on the site by a wide margin, and our ANPR guide covers getting the angle right. The inverter stations and substation compound. The concentrated value. Cable runs during works only. Temporary cameras during the window where cable is exposed, removed afterwards. This is where most of the loss happens and it is a defined period. The approach track. Not the public road, which raises the questions covered in our signage and footpaths guide, but the private track in. A vehicle on that track at 2am has no business there, and that is a much cleaner detection rule than motion in a field. Detection rules that work on an empty site An unmanned site has one enormous advantage over a farm: anybody there is an exception. On a working farm, motion at the gate is the postman, the vet, the milk tanker and a dog. On a solar farm out of hours, motion on the access track is either maintenance, which is scheduled, or a problem. That makes alerting genuinely practical rather than something you mute after a fortnight. Set it up as: person or vehicle detection, on the access track and compounds, outside scheduled working hours, going to somebody who can act. AI classification matters here because sheep are frequently grazed under the arrays, and a motion alert per sheep is useless. And then somebody has to respond This is the part that gets skipped, and it is the part that decides whether any of it works. An alert at 02:40 to a phone in an office that opens at nine is not security. Before fitting anything, settle: who gets the alert, what do they do, and who goes? For most sites the honest answer is a keyholding or response contract, or an arrangement with the landowner, who is frequently a farmer living two fields away and is by far the fastest person to the gate. A camera system with no response plan still has value, because recorded evidence recovers equipment and supports insurance. But it is a recovery tool, not a prevention one, and it should be sold…