Chance-constrained optimal dispatch of integrated energy systems based on data-driven sparse polynomial chaos expansion

In integrated energy systems, stochastic variations of different energy types of loads and the increasing penetration of renewable energy generation have resulted in considerable uncertainties. These uncertainties pose significant challenges to the economics and safe operation of integrated energy systems. Conventional deterministic methods of optimal dispatch overlook the effects of uncertainties, while stochastic optimization, though accounting for uncertainties, often yields conservative solutions that may adversely affect the economic operations. Chance-constrained optimization can effectively deal with uncertainties and provide more flexibility in balancing operational risks and benefits by expanding the feasible region. Hence, this paper proposes a chance-constrained optimal dispatch method for integrated energy systems and employs data-driven sparse polynomial chaos expansion method to enhance solving efficiency. First, the proposed chance-constrained optimization aims to ensure the optimal economic operation with an affordable security confidence level that balances safety and economics compared with deterministic and stochastic optimizations. The introduced data-driven sparse polynomial chaos expansion method enables the fast computing of the output response using only historical data, i.e., without knowledge pertaining to the distribution functions. Moreover, an improved iterative verification structure is proposed, which further improves the convergence speed and accuracy. Finally, the advantages and feasibility of the proposed method are verified using a test case and compared with those of deterministic and stochastic optimization. Results show that the proposed method successfully reduces the operational cost and controls violation probability.