Markovian-based stochastic operation optimization of multiple distributed energy systems with renewables in a local energy community

摘要

Local energy communities (LECs), as locally and collectively organized mull-energy systems, are expected to play an important role in energy transition, since they enable deployment of sustainable energy technologies and consumer engagement, bringing various benefits to the community users and contributing to the overall energy and climate objectives. An LEC may consist of multiple distributed energy systems (DESs), which, interconnected through local grid and heating network, can share power and thermal energy with no costs for community's users. This paper focuses on stochastic daily operation optimization of multiple DESs with renewables in an LEC. The problem is to find the optimized operation strategies of energy devices in each DES, and decide the amount of electrical and thermal energy to be shared among DESs with the objective to minimize the total expected net energy and CO2 emission cost of the LEC, while meeting given day-ahead demand of community's users. The problem is challenging because of the intermittent and uncertain nature of renewable generation and the coupling of energy devices and energy processes intra and inter DESs. To address these issues, a stochastic mixedinteger linear programming model is established with uncertain renewable generation modeled by a Markovian process to avoid the difficulties and drawbacks associated with scenario-based methods. The problem is solved by using branch-and-cut. Numerical testing results show that the total expected cost of the LEC is reduced by the integrated management of the DESs as compared to the costs attained under other operation modes where there are no interconnections among DESs, demonstrating the potential benefits that can be achieved with LECs through the optimized management of local energy resources aiming to foster efficient use of the available energy. Results also highlight the benefits of the stochastic approach as compared with the deterministic one.