An Optimization Coal Procurement and Inventory Model for Power Generation Enterprises Based on Data-driven Chance Constraints

Optimization of coal procurement and inventory for power generation enterprises are of great significance for guaranteeing power supply and ensuring generation income. The requirements for safe coal inventory level have been clearly put forward by the energy administrative authority of our country. However, no existing research has ever focused on the probabilistic model and corresponding optimization strategy for the violation risk of inventory caused by the uncertainties of power generation and transportation capacity. Aiming at this problem, this paper presents an optimization coal procurement and inventory model for power generation enterprises based on data-driven chance constraints and proposes a corresponding solution method. Firstly, with consideration of the uncertainty of power generation and transportation capacity, the data-driven chance constraints for inventory are established and converted to soluble constraints of conditional value at risk (CVaR). Furthermore, based on the convexity of CVaR to decision variables, a piecewise linear approximation method for CVaR constraints is proposed. A power generation enterprise which owns 10 coal power plants is selected for case study. The optimization results show that with consideration of the chance constraints, the violation risk of power coal inventory is restricted within the allowable range; the proposed piecewise linear approximation method for CVaR constraints can make the model scalable and reduce the model’s scale with a high accuracy.