In microgrids, forecasting solar power output is crucial for optimizing operation and reducing the impact of uncertainty. To forecast solar power output, it is essential to forecast solar irradiance, which typically requires historical solar irradiance data. These data are often unavailable for residential and commercial microgrids that incorporate solar photovoltaic. In this study, we propose an hourly day-ahead solar irradiance forecasting model that does not depend on the historical solar irradiance data; it uses only widely available weather data, namely, dry-bulb temperature, dew-point temperature, and relative humidity. The model was developed using a deep, long short-term memory recurrent neural network (LSTM-RNN). We compare this approach with a feedforward neural network (FFNN), which is a method with a proven record of accomplishment in solar irradiance forecasting. To provide a comprehensive evaluation of this approach, we performed six experiments using measurement data from weather stations in Germany, U.S.A, Switzerland, and South Korea, which all have distinct climate types. Experiment results show that the proposed approach is more accurate than FFNN, and achieves the accuracy of up to 60.31 W/m² in terms of root-mean-square error (RMSE). Moreover, compared with the persistence model, the proposed model achieves average forecast skill of 50.90% and up to 68.89% in some datasets. In addition, to demonstrate the effect of using a particular forecasting model on the microgrid operation optimization, we simulate a one-year operation of a commercial building microgrid. Results show that the proposed approach is more accurate, and leads to a 2% rise in annual energy savings compared with FFNN.

This paper presents a method to seek the PI controller parameters of a PMSG wind turbine to improve control performance. Since operating conditions vary with the wind speed, therefore the PI controller parameters should be determined as a function of the wind speed. Small-signal modeling of a PMSG WT is implemented to analyze the stability under various operating conditions and with eigenvalues obtained from the small-signal model of the PMSG WT, which are coordinated by adjusting the PI controller parameters. The parameters to be tuned are chosen by investigating participation factors of state variables, which simplifies the problem by reducing the number of parameters to be tuned. The process of adjusting these PI controller parameters is carried out using particle swarm optimization (PSO). To characterize the improvements in the control method due to the PSO method of tuning the PI controller parameters, the PMSG WT is modeled using the MATLAB/SimPowerSystems libraries with the obtained PI controller parameters.

Microgrids can be considered as controllable units from the utility point of view because the entities of microgrids such as distributed energy resources and controllable loads can effectively control the amount of power consumption or generation. Therefore, microgrids can make various contracts with utility companies such as demand response program or ancillary services. Another advantage of microgrids is to integrate renewable energy resources to low-voltage distribution networks. Battery energy storage systems (BESSs) can effectively compensate the intermittent output of renewable energy resources. This paper presents intelligent control schemes for BESSs and autonomous energy management schemes of microgrids based on the concept of multi-agent systems. The proposed control scheme consists of two layers of decision-making procedures. In the bottom layer, intelligent agents decide the optimal operation strategies of individual microgrid entities such as BESSs, backup generators and loads. In the upper layer, the central microgrid coordinator (MGCC) coordinates multiple agents so that the overall microgrid can match the load reduction requested by the grid operator. The proposed control scheme is applied to Korea Power Exchange’s Intelligent Demand Response Program.

At the present time, power-system planning and management is facing the major challenge of integrating renewable energy resources (RESs) due to their intermittent nature. To address this problem, a highly accurate renewable energy generation forecasting system is needed for day-ahead power generation scheduling. Day-ahead solar irradiance (SI) forecasting has various applications for system operators and market agents such as unit commitment, reserve management, and biding in the day-ahead market. To this end, a hybrid recurrent neural network is presented herein that uses the long short-term memory recurrent neural network (LSTM-RNN) approach to forecast day-ahead SI. In this approach, k-means clustering is first used to classify each day as either sunny or cloudy. Then, LSTM-RNN is used to learn the uncertainty and variability for each type of cluster separately to predict the SI with better accuracy. The exogenous features such as the dry-bulb temperature, dew point temperature, and relative humidity are used to train the models. Results show that the proposed hybrid model has performed better than a feed-forward neural network (FFNN), a support vector machine (SVM), a conventional LSTM-RNN, and a persistence model.

The microgrid and demand response (DR) are important technologies for future power grids. Among the variety of microgrid operations, the multi-agent system (MAS) has attracted considerable attention. In a microgrid with MAS, the agents installed on the microgrid components operate optimally by communicating with each other. This paper proposes an operation algorithm for the individual agents of a test microgrid that consists of a battery energy storage system (BESS) and an intelligent load. A microgrid central controller to manage the microgrid can exchange information with each agent. The BESS agent performs scheduling for maximum benefit in response to the electricity price and BESS state of charge (SOC) through a fuzzy system. The intelligent load agent assumes that the industrial load performs scheduling for maximum benefit by calculating the hourly production cost. The agent operation algorithm includes a scheduling algorithm using day-ahead pricing in the DR program and a real-time operation algorithm for emergency situations using emergency demand response (EDR). The proposed algorithm and operation strategy were implemented both by a hardware-in-the-loop simulation test using OPAL-RT and an actual hardware test by connecting a new distribution simulator.