COST-BASED ENERGY MANAGEMENT FOR COOPERATIVE MICROGRIDS

Abstract

As microgrid (MG) systems are becoming more prevalent, the future smart distribution systems may be envisioned to comprise of multiple interconnected MGs that can run as autonomous entities. Under such scenarios, energy can be wheeled among MGs in order to alleviate their deficits locally, using the surplus energy of others for mutual benefits. The obvious advantage emanates from exploiting the spatiotemporal and technological diversities in the multiple MGs. However, new interfaces between MGs are necessary and an intelligent energy management (EM) is inevitable for determining optimal energy exchanges between MGs and that with the grid. The presence of intermittent energy sources, consumer demands and time-varying electricity prices, further accentuates the need for optimal bidding strategies and intelligent EM in such an interconnected MG-based system. In this thesis, several new strategies for the coordinated EM of multiple cooperative MGs and the distribution network are developed. Increasing injection of renewable energy (RE) sources might lead to periods of time in which there is a combination of a very high injection of RE and low demand for electricity. As a consequence, prices become depressed to the point that they do not cover the variable operating costs of the generating units. To deal with the risk (i.e. low profits or losses) that stem from the uncertainty of RE sources, a risk-control scheme is developed that can control the trade-off between maximizing the optimal expected profits of MGs and its variability. MGs are modelled as autonomous entities that schedule their local generation and storage units, and gain benefits through energy trading. Within this framework, two criteria are proposed for modelling the interactions between MGs – (i) power loss reduction criteria and (ii) prioritizing criteria. The subsequent power imbalances, which could not be compensated by local power exchanges, are being traded by the central controller in the market. For this purpose, a detailed model is developed for cooperative MGs’ power scheduling and bidding problem in competitive electricity markets. The framework thereby, allows the MGs to share their resources and collectively interact with the grid as one entity in both day-ahead and realtime markets. The proposed framework, thereby, leverages the diversities across different MGs to compensate for the intermittent attribute of RE sources and enables the operators A Abstract ii to achieve desired trade-offs for different conflicting objectives, e.g., maximizing the expected profits of MGs while minimizing their bid deviations, RE and load curtailments. Thereafter, in order to take into account, the heterogeneous nature of the MGs, a new multi-period optimal dispatch scheme based on game theory, is developed for a distribution system with clustered MGs. The developed methodology starts with a selected number of switches, which generate various topologies. First of all, load flows are performed to ascertain the feasibility of these topologies and then for each feasible topology, the proposed stochastic game-theoretic framework is applied. Results corroborate the effectiveness of the proposed approach in improving the payoffs of the MGs, while ensuring that all network constraints are within their specified limits. Finally, this thesis considers the problem of the unavailability of the distributions of the RE power productions. To enable the MG operation optimization design, the MG aggregator can fit the forecasting error data into appropriate distributions based on the historical data and forecast data over a sufficiently long period. However, the complexity increases when the distributions of the RE productions are unavailable. To deal with this, a worst-case transaction mechanism is proposed that only requires the information on the deterministic bounds on the total RE harvested at the MGs, across the scheduling horizon. Specifically, this chapter exploits the concept of using local MG generation capacities to support other MGs, under the worst-case scenarios of intermittent RE generation. The developed methodologies have been applied to three test systems – (i) three MG system, (ii) five MG system and (iii) modified IEEE 33-bus distribution system with five MGs. The efficacy of the stochastic game theory framework is also tested on the IEEE 33- bus system and the PG&E 69-bus system partitioned into four and six MGs, respectively. Results corroborate the effectiveness of the proposed cooperative approaches in yielding potential environmental and economic benefits to both consumers and operators, compared to the case where each MG operate individually only in coordination with the grid. The developed approaches further ensure the autonomy of each MG by utilizing only the power exchange requests as the optimal control signals between MGs