Advanced Control of Active Distribution Networks Integrating Dispersed Energy Storage Systems

Due to the increased penetration of Distributed Generations (DGs) in distribution networks, the system control and operation may become quite different from the case of traditional network. Most DGs can only provide intermittent power to the Active Distribution Networks (ADNs) due to the intermittent nature of the resources. Moreover, ADN utilities usually do not own DGs, and have difficulty in controlling directly DGs output powers. The main problem related to the considerable connection of DGs is usually associated to the node voltage quality and line congestion mitigation. Within the above context, the motivating factors for this thesis are supported by the issues related to optimal operation and control of ADNs integrating stochastic and non-stochastic DGs. One of the most promising near-term solution is offered by using distributed Energy Storage Systems (ESSs) which can perform their full role to guarantee a more flexible network. Indeed, the availability of ESSs allows, in principle, to: (i) actively control the power flows into the grid, (ii) indirectly control the voltage profiles along the network feeders and (iii) locally balance the hour/daily and weekly load variations. In this thesis, ESSs are assumed to be the only controllable devices in ADNs. As a result, DGs can be indirectly controlled by means of ESSs. First, this manuscript presents control-oriented model for ESSs. In this respect, the accurate estimation of ESS behavior is utmost important. A generic charge representative model for any ESSs is proposed. Moreover, an improvement of the most common electric equivalent circuit models for the two selected ESSs with different characteristics (namely supercapacitors and batteries) is provided for the development of specific control schemes. They are based on the modeling of redistribution of charges that characterizes the dynamic behaviors of the two devices during long time charging/discharging and relaxation phases. Second, this manuscript presents advanced control/scheduling algorithm for ADNs. The operation and control of ADNs can be achieved either centrally or in a decentralized way. The amount of information to be centrally treated would considerably grow due to the number of generation equipment’s inserted into the grid and the stochastic operation nature of some of them. This consideration introduces the idea that some ADN operation problems, such as voltage control or line congestion mitigation, can be solved in a distributed manner which would help to relieve the information processing burden and to enhance the system security while preventing unwanted event from propagating through the grid. Therefore, the decentralized schemes are considered subdividing the network into quasi-autonomous areas. To this end, given a set of ESSs optimally located in a balanced and radial ADN, this thesis proposes a network partitioning strategy for the optimal voltage control of ADNs. Thus, the network is decomposed into several areas; each under the control of one ESS which has maximum influence on its corresponding area. Based on this clustering, decentralized scheduling strategies and real-time decentralized control algorithms for the clustered ADNs are proposed. The proposed zonal control capability focuses on voltage control and line congestion management. In both proposed decentralized scheduling and real-time control algorithms the communication among different areas is defined using the concept of Multi-Agent Systems.