Coal Plant Retirement and Repurposing Optimization Model : Technical Report (including the GAMS Source Code)

Overview & Synopsis

Retirement and repurposing of coal-fired power plants are a critical part of the energy transition. This is the core part of programs like Accelerating Coal Transition (ACT) that is being implemented in several countries around the world. While there are established planning models that comprehensively capture both capacity entry and exit decisions at the system level (i.e., across all generation types to meet systemwide demand), there are nuances when it comes to coal retirement. Sites of incumbent coal plants can be repurposed to switch to a different fuel (e.g., biomass), or to install renewable and energy storage facilities, depending on the available land and system needs. There are social dimensions including keeping part of the workforce in such repurposed plants. There are other issues around stranded asset value of coal plants that still have productive life left, availability of financing and allocation of concessional finance for retirement and repurposing activities. There is currently no model/tool that can capture these issues to make balanced decisions on retirement and repurposing of coal generating units. The retirement and repurposing decisions are typically being made based on technical criteria such as age and efficiency at a project level. The proposed modeling framework is developed to address coal (or any fossil fuel) fleetwide retirement and repurposing decisions cast as a multi-year mixed integer linear programming model taking into consideration technical, economic and policy considerations. It makes retirement and repurposing decisions at a generating unit level to maximize the net revenue for the entire fleet given a target generation level and/or CO2 emissions cap. The latter target(s) could be derived from an upstream systemwide planning model making it an important linkage with the plan, or could be a policy/program target as is often the case for CO2 targets under ACT. Decision variables in the model include retirement decisions, repurposing options that maximize value of the site, employment, finance structure and dispatch of coal units. The model formulation including the source code (written in General Algebraic Modeling System) is presented in this paper. The working of a prototype model has been illustrated using a set of simple examples and also a case study for the ESKOM (South Africa) coal fleet.