Every O&M quotation lists cleaning, string measurement and a thermal survey. Almost none of them mention pigeons. What actually gets damaged under a tilted array, how mesh and cable entries should be done, and why this work sits in nobody's scope.
Every rooftop maintenance quotation lists module cleaning, string measurement and an annual thermal survey. Almost none of them mention pigeons. Two years after commissioning, the pigeons are usually why the array under-performs and why the roof drain is blocked, and nobody has a line item to point at.
A tilted array on a flat roof creates a sheltered, dry, warm cavity that is close to ideal for pigeons and mynas. What follows is predictable. Nesting material packs in under the modules and blocks the air flow the modules rely on for cooling, so cell temperature rises and output falls across a whole area of the roof. Droppings on the glass are opaque and acidic, and because they are localised they create hot spots on the cells they cover rather than a general loss. Feathers and nesting material wash into the roof drains until the roof ponds during a monsoon downpour. The acid attacks anodising on frames and the galvanising on the structure. And there is a genuine occupational health issue, because dried droppings and nest dust carry respiratory pathogens and your technicians are the people who have to work in it.
Rats chew DC cable insulation, and this is the most under-appreciated fire risk on a rooftop plant. The damage happens where the cable is easiest for the animal to reach and hardest for a person to see, in the tray under the array or at the entry into the plant room. The result is an earth fault if you are lucky and a parallel arc if you are not, and a DC arc does not go out by itself. Rodents also nest inside inverter and combiner enclosures because they are warm, and one rat across two busbars ends the day.
Wind-blown soil collects along the low edge of a tilted array on an RCC roof, and within a couple of monsoons there is enough of it for grass and small plants to establish. Roots into a waterproofing membrane are an expensive repair. Moss and algae do the same thing more slowly on a shaded, damp roof. Wasps build in combiner boxes and inverter enclosures, which turns routine inspection into something nobody volunteers for. On ground-mounted plants the vegetation problem is larger and it brings snakes with it.
Because it is not electrical work. Bird proofing, drain clearing and vegetation control fall in the gap between the solar contractor's scope and the factory's own housekeeping contract, and anything in that gap does not get done. The remedy is to write these tasks into the AMC explicitly with a stated frequency, or to write them into the factory maintenance schedule with a named owner. Either is fine. Assuming the other party has it covered is the failure mode.
A workable schedule for a Bangladeshi factory roof: clear the roof drains before the monsoon and again mid-season, inspect under the array for nesting at the same two visits, walk the cable trays for chew damage quarterly, and pull vegetation while it is still seedlings rather than after the roots are into the membrane.