Daily operation of multi-energy systems based on stochastic optimization considering prediction of renewable energy generation

Nowadays, the energy crisis is one of the most critical challenges facing countries. To deal with this crisis, instead of independently optimizing each of the energy carriers (electricity, gas, heat etc.), all carriers are considered simultaneously by a concept called multi-energy system, which not only leads to economic benefits but also has environmental benefits. Here, a mixed-integer linear programming model is proposed to minimize the total daily cost of a local multi-energy system including the cost of energy exchange with the main grid, the cost of natural gas, and carbon emission costs. A polynomial neural network model is used to forecast the hourly wind speed and radiation of the next day. Also, a probabilistic scenario-generation and scenario reduction method is utilized to generate the possible scenarios from the probability density function. The simulation results show that the proposed model using neural network prediction and stochastic optimization increases the total cost of the multi-energy system by 12% in the worst-case. Sensitivity analysis of loads, electricity prices, and gas prices have been used to investigate the behaviour of variables on energy hub operating costs.

Automated summary

A mixed-integer linear programming model is proposed for minimizing the total daily cost of a local multi-energy system, using neural network prediction and stochastic optimization. The model reduces the total cost of the multi-energy system by 12% in the worst-case scenario. Sensitivity analysis of loads, electricity prices, and gas prices are conducted to investigate the variables' impact on energy hub operating costs.