The impact of electricity price forecast accuracy on the optimality of storage revenue

Grid connected electrical energy storage could enable large numbers of intermittent renewable generators to be deployed in the UK. Many studies investigate the revenue which could be achieved through arbitrage assuming perfect foresight of electricity prices. In practice, storage operators will not have perfect foresight and will have to devise operational strategies using price forecasts. This paper investigates the impact of forecast accuracy on the optimality of storage revenue. The optimal revenue available is determined using linear programming and historic electricity prices. The results are compared to those found using dynamic programming and electricity price forecasts with increasing percentage error. A small scale lithium ion battery and a large pumped hydro energy storage (PHES) device are compared. The results show that revenue reduces at an increasing rate with increasing forecast error. The PHES device is more sensitive to forecast accuracy than the lithium ion battery. For both technologies, with a maximum error of 30%, 80% of the optimal revenue can be achieved. With increased capacity and significantly increased power rating, the lithium ion battery becomes more sensitive to price forecast accuracy.