Photovoltaic (PV) Panel Waste Volumes-4

Both scenarios mentioned in our previous article assumed a 30-year average panel lifespan and 99.99% loss probability after 40 years. In environmental impact analysis of a 30-year panel lifespan (for example, in life cycle assessments), these evaluations can be made more clearly. In the models developed, it is assumed that PV panels are removed for renewal and modernization by 40 years at the latest. Therefore, it can be assumed that the durability of PV panels is compatible with the average experience of building and construction products such as facade elements or roof tiles. These are also traditionally assumed to have a lifespan of 30-40 years.

Model Scenario In the early loss scenario, the following damage assumptions can be made based on literature analysis and expert opinion:

• It is assumed that 0.5% of PV panels (in terms of installed PV capacity in MW) reach the end of their useful life due to damage during transportation and installation stages, • Due to poor installation, 0.5% of PV panels will become waste within two years, • 2% will become waste after ten years, • 4% will be discarded after 15 years due to technical failures. The early loss scenario includes panel replacement failures such as broken glass, broken cells or ribbons, and cracked back sheets with insulation defects. However, only panels with serious functional or safety defects requiring complete replacement are included, while other defects such as those reducing power output or causing panel color fading are disregarded. In the early loss scenario, the shape factor was calculated through regression analysis between data points in the literature and was also evaluated as early failures.

The resulting alpha shape factor for the early loss scenario is 2.4928, which is lower than the literature values presented. This is because if a panel lasts longer, it contains early defects that lead to higher losses in the first 30 years and lower losses in later life. For each scenario (normal loss and early loss), the failure probability value (alpha) is multiplied by the weight of panels installed in a particular year according to the Weibull function.

Since a larger alpha value is used in the regular loss scenario, the curve rises smoothly and intersects with the early loss scenario curve at the 30-year nominal lifespan point. In accordance with the Weibull function and due to assigned different alpha parameters, regular loss and early loss scenarios have opposite effects after 30 years. Therefore, the regular loss scenario shows a higher loss probability after 30 years (see Figure 1). This study is the first to measure PV panel waste on a global scale and across different PV technologies. This means that the scenarios described here must be accepted as order of magnitude and directional rather than high accuracy or precision due to simple assumptions and lack of statistical data. Moreover, they encourage the need for further assessment. This box provides a brief overview of three major areas of uncertainty that could affect the study's findings and conclusions. Uncertainty regarding cumulative installed PV capacity through 2050 is an input factor for the model and is therefore not further considered here. First and foremost, existing data on PV panel failure modes and mechanisms represents only a small fraction of the number of panels installed worldwide. This means that some of the basic assumptions carry certain uncertainties and will need to be corrected as more data becomes available. The rapid development of PV materials and designs adds another level of complexity and uncertainty to the projections. Moreover, failure does not mean that a panel will enter the waste stream in a particular failure year. This is because some failures may not be detected immediately or may be tolerated for years.

For example, if a PV panel still produces some output, even if lower than when first commissioned, replacement may not be economically justified. Therefore, existing data on the various determinants of the end of life of a PV panel are typically linked to system elements that are non-technical and very difficult to predict.

The final major uncertainty relates to the basic assumptions used to model the probability of PV panel losses versus the life cycle of panels using the Weibull function. The literature for calculating Weibull shape factors for the regular loss and early loss scenarios of this study can be reviewed. It is assumed that early losses in the early loss scenario remain constant in the future. In other words, no learning aimed at reducing early losses is taken into account. The model also excludes PV systems that are repowered.

In summary, this study develops two scenarios;

It is intended to account for the above uncertainties regarding regular loss and early loss mechanisms and the estimated timing of panel failures. To better predict future potential PV panel waste streams, national and regional decisions on PV waste stream regulation should include a monitoring and reporting system. This provides advanced statistical data to strengthen waste stream estimates and provide a consistent framework for policy regulations. The modeling above shows PV panel waste projections by country through 2050. The following section summarizes the model's findings.

PV Panel Waste Projections

Today the total annual e-waste in the world is 41.8 million tons. By comparison, according to the modeled early loss scenario, cumulative PV panel waste will not exceed 250,000 tons by the end of 2016. This amount represents only 0.6% of today's total e-waste, but the global amount of waste panels from PV will increase significantly in the coming years. In the regular loss scenario, PV panel waste is calculated as 43,500 tons as of end of 2016, with an increase expected to reach 1.7 million tons in 2030. By 2050 there could be an even more striking increase to approximately 60 million tons.

The early loss scenario projection estimates that the total PV waste stream will be much higher, with 250,000 tons alone by the end of 2016. This projection will rise to 8 million tons in 2030 and a total of 78 million tons in 2050.

This is because the early loss scenario assumes a higher percentage of early PV panel failures than the regular loss scenario. Based on the best information currently available, this report suggests that actual future PV panel waste volumes will likely fall somewhere between the regular loss and early loss values. Annual PV panel waste through 2050 is modeled in Figure 3, showing the development of PV panel end-of-life and new PV panel installations as a ratio of the two projections. This ratio, for example, starts at less than 5% at the end of 2020 (meaning in the early loss scenario, 220,000 tons of annual waste compared to 5 million tons in new installations). However, over time it increases to 4-14% in 2030 and could reach 80-89% in 2050. At this point, 5.5-6 million tons of PV panel waste (depending on scenario) is projected compared to 7 million tons in new PV panel installations.   A characteristic of the Weibull curve shape factors for the two modeled scenarios is that both scenarios' estimated waste intersect. The scenario predicting more waste panels in a given year changes later. The intersection is projected to occur in 2046. This modeling feature can be seen in Figure 8, which shows the PV panel waste volume for the early loss scenario result in 2050, comprising more than 80% of the volume of new installations. The comparable figure for the regular loss scenario exceeds 88% in the same year.     Cemil Koyunoğlu Yalova University Faculty of Engineering Department of Energy Systems Engineering