End-of-life PV module volumes: when the feedstock actually arrives

A module installed in 2008 reaches its design end of life around 2033. A module installed in 2015 arrives around 2040. This means the feedstock available to a recycling plant today reflects installations from the late 1990s and 2000s, plus everything that failed early — and in most markets that early-failure stream is a substantial share of what is actually available.

The practical consequence is that national installation statistics describe the feedstock of the 2030s and 2040s, not what will turn up at your gate next year. For a plant being planned now the useful inputs are different: the size and vintage of the installed base in your catchment area, the rate at which modules are being replaced early or damaged, and current practice for dealing with them.

Four sources produce module waste well before the design life expires: physical damage from hail, storm and thermal events; manufacturing defects and warranty replacements; repowering projects that replace working modules with higher-output ones to make better use of existing land and grid connections; and handling damage during transport and installation.

Repowering is the one that most often surprises planners. A site being repowered produces its entire module stock in a few months, at one address, with clean logistics — which is a very different feedstock profile from a steady trickle of domestic replacements. If your business plan assumes steady volume but your first contract is a repowering project, the line has to suit a campaign, not a continuous process.

Published global projections vary widely because they depend on assumed module lifetimes, on installation scenarios and on how much of the early-loss stream is counted as waste. Modelling by IRENA and the IEA photovoltaic power systems programme, published in 2016, put cumulative end-of-life PV panel volume at roughly 1.7 to 8 million tonnes by 2030 and up to about 78 million tonnes by 2050, under its scenarios.

Those figures are useful for showing that the material is coming. They are not useful for sizing a plant. For that, the inputs are local: the installed base within economic transport distance, its vintage, the replacement and damage rate you can observe, and the permitted capacity you can actually obtain.

1. 01

   Establish the local installed base

   Capacity and vintage within the distance you can economically collect from. This, not the global figure, is the upper bound on your feedstock.

2. 02

   Identify the campaign sources

   Are there repowering, decommissioning or replacement programmes within reach? A campaign changes the required through-put for a short period.

3. 03

   Decide whether to design for peak or for average

   Designing for the peak wastes capital in the quiet months. Designing for the average and outsourcing the peaks requires a partner arrangement.

4. 04

   Make the line scalable

   A plant that can add a station or a shift without rebuilding is worth more than one specified to a single number. Ask how the line grows.

5. 05

   Check the permitted volume

   In many markets the permit, not the equipment, is the binding constraint. Confirm the through-put your permit allows before sizing a machine to exceed it.

The end-of-life volume modelling referred to above was published by the IEA Photovoltaic Power Systems Programme in its work on end-of-life management of photovoltaic panels, alongside IRENA. Later task reports update the assumptions and the ranges move accordingly, which is why the useful move is to read the current report rather than to rely on a number quoted in a summary.