It’s been a dismal year for our PV array; after a best-ever January, production has fallen off and stayed down, 20% lower than expected.
At first I thought the low generation was due to dirty panels, so I stepped up my cleaning routine. Then I thought it was due to shading, so we had some more branches trimmed. Neither of these remedies addressed the production shortfall, so finally I resolved to have our installer investigate. I began gathering numbers on our monthly production.
But this takes time, a commodity as rare around here as free electricity. I finally gathered the necessary documentation and submitted it for review in July.
In various dark periods in my past, I have done some customer support work and some system testing work, and I know that the vast majority of technical problems reported by users are their own damn fault. Here is a classic example, from my semester on the graveyard shift in the Mac lab in college about 100 years ago: a thick-necked football player who probably went on to make more money in his first two years out of school than I’ve made ever waved me over to complain that his mouse wasn’t working — when he pushed it forward, the cursor on the screen moved down, and when he pulled it back, the cursor moved up. He’d spent enough time pecking his way through the semester’s various essay assignments to have acquired some mastery of the pointing device, so its apparent misbehavior was discomfiting. Every mouse movement broke his concentration, as he had to consciously override well-established muscle memory to put the cursor where he wanted it. This required so much mental effort that it never occurred to him to pick up the mouse and rotate it 180°. (The cord is supposed to come out of the top, you know.)
So I fully expected to be put off by our installer with inquiries about cleaning, weather, and tree cover, and I was working up solid answers for each in hopes of minimizing the delay of any sane CS agent’s justifiable evasion techniques.
Then I realized that the best way to identify a generation shortfall is not by measuring multiple months’ worth of production data for comparison to previous years or to other systems in the neighborhood… the best way to identify a generation shortfall is to take a spot reading when the system should be maxed out.
My PV array was designed to produce maximum power during the peak hours of the peak season: noon-6pm, May through October. If at 2pm on a bright July afternoon the system wasn’t cranking out 95+% of its rated maximum of 2500 watts, then something must be wrong.
Giving the system every possible advantage, I washed the panels first, then waited for them to dry. The inverter showed a measly 1800 watts. We were missing 28% of our generation capacity.
I sent a brief email to the installer. Within a couple hours I got a reply: someone would be out to inspect the system the next day.
(Tune in next week for Part II)