Neural network dynamic differential control for long-term price guidance mechanism of flexible energy service providers

Abstract

Price guidance in the flexible energy electricity market is not mature. The imbalance between supply and demand in current electricity markets is mainly caused by uncontrolled flexible energy generation; the electricity supply with load changes in the traditional electricity market is no longer applicable to the current flexible energy electricity market. The work proposes neural network dynamic differential control as the long-term price guidance mechanism for flexible energy service providers. This mechanism alleviates the imbalance between supply and demand in the future electricity market. The neural network dynamic differential control method combines the long-short term memory network and the stochastic dynamic differential method. The proposed method integrates the effects of environmental temperature, holidays, and government regulation on demand. This mechanism enables the demand of flexible energy customers to equal the generation capacity of flexible energy, customers to minimize costs in a long-term electricity market and flexible energy to be completely absorbed. This mechanism directly controls the electrical energy demand and indirectly controls the indoor temperature of the building cooling model through price guidance signals. The experimental results show that the building cooling model can save 4.84% of the cost and 65.03% of the potential cost savings.