Solar energy, as an emerging renewable clean energy, has been rapidly growing for 15 years all over the world and is expected to grow 15% annually until 2020. In 2015, at least 40 GW of Photovoltaic (PV) systems were installed, achieving 178GW current solar power installation worldwide. In the next five years, 540 GW cumulative capacities are expected to be installed worldwide and US contributed 6.5 GW PV installations in 2015. US electricity demand is expected to be dominated by solar power by 2050 or even earlier. The widespread deployment of PV will not only contribute to a reduction in greenhouse gas emission, but can also mitigate the worldwide fossil fuel depletion.
As the number of PV systems increases, the mass of PV waste will increase as well, adding a new source to the existing waste stream. The amount of End-of-Life (EoL) PV will approach 13.4 million ton worldwide, including approximately 5.5 million ton located in the US by 2025. PV contains high value, toxic, and energy-intensive materials. In addition, the market price of some materials utilized in the thin-film and crystalline PV technologies has drastically increased in the recent years.
There is a strong need of coordinating the information to optimize the reverse logistics planning in a photovoltaic (PV) recycling network in the U.S. Two major tasks are included: 1) locating PV Recycling Centers (PVRC); 2) allocating Transportation Companies (TC) shipping PV installation sites (PVIS) to PVRC. One contribution of this dissertation is to decide the optimal number, as well as the location of PVRC by minimizing the overall cost. Another contribution is to determine the optimal distribution scheme to minimize the transportation cost among TC, PVIS, and PVRC.
In order to accomplish the two tasks, a mathematical modeling framework was developed to facilitate PV recycling in an economically and environmentally feasible manner. The framework included two mathematical models: 1) Multi-Facility Optimization Model; 2) Optimal Distribution Model. The multi-facility optimization model included the transportation module, the economic module, and the environmental module. The model identifies the geographical location of the prospective PVRCs by minimizing the total costs in different scenarios. While in the Optimal Distribution Model, a static and a dynamic optimization algorithm was applied for conducting the optimal solution accurately and efficiently.
To show the efficacy of the proposed framework, case studies for recycling EoL PV in California were performed. Historical PV installation data in the region was utilized to gather information about the amount of the prospective end-of-life (EoL) PV waste generation in CA. In order to integrate the temporal and the spatial dispersion of PVISs in CA, a three-phase recycling plan was proposed. For well displaying the geographical results, Geographic Information System (GIS) was utilized to visualize the installation data, optimized location of the PVRCs, and the optimal distribution scheme. The proposed generic framework provided a great insight for decision makers about the trade-offs among various scenarios by considering cost, environmental impact, and investment risk on PV recycling planning.