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 popularity of microgrids is increasing considerably because of their environmental and technical advantages. However, the major challenge in microgrid integration is its financial feasibility due to high capital costs. To address this obstacle, renewable energy incentive programs, which are the motivation of this study, have been proposed in many countries. This paper provides a comprehensive evaluation of the technical and financial feasibility of a campus microgrid based on a techno-economic analysis using the Microgrid Decision Support Tool, which was implemented to support decision-making in the context of microgrid project investment. A method for microgrid design aiming to maximize system profitability is presented. The optimal microgrid configuration is selected depending on financial indices of the project, which directly address the returns on an investment. Most importantly, this analysis captures all the benefits of financial incentives for microgrid projects in California, U.S., which presents a key difference between the California market and other markets. The impact of incentives and uncertain financial parameters on the project investment is verified by sensitivity analysis. The outcomes show that the optimal configuration generates significant electricity savings, and the incentives strongly determine the financial feasibility and the optimal design of a microgrid.

Stand-alone microgrids integrating renewable energy sources have emerged as an efficient energy solution for electrifying isolated sites, such as islands and remote areas. The design of a microgrid involves various influential factors, including technological development, economic feasibility, and environmental impacts, based on the conditions and regulations of a particular site. This paper proposes a comprehensive microgrid design framework based on power system analysis and techno-economic analysis. The obtained optimal microgrid configuration satisfies both the design objective and power system performance regulations. The proposed design approach focuses on using practical data and can adapt to any microgrid design problems based on the local characteristics of a specific site. The practicality and effectiveness of the design framework are validated by applying it to the design of a stand-alone microgrid for Deokjeok Island in South Korea. The case study results justify the importance of considering site-specific characteristics and the impacts of power system conditions on the optimal microgrid design.

Herein, we propose a novel method for identifying power congestion and renewable energy source (RES) power curtailment in power grids by using deep neural network (DNN)-based optimal power flow (OPF) analysis. Synthetic data for load demand and RES power generation are used to obtain the OPF solutions by using an OPF solver. RES locations are selected based on an analysis of the congestion cases, and the DNN is trained by using the load demand and RES power as inputs and the OPF solution as the output. Thereafter, post-processing is performed on the DNN-OPF output by using network information. The final results show the accuracies of identified values on RES curtailment, line loading, generation dispatch schedule, and total generation operating cost. The IEEE 39-bus system was adopted as a case study to validate the proposed model. The results demonstrate that the proposed scheme is much faster and more suitable for transmission systems than conventional OPF methods. Furthermore, the normalized root-mean-squared error was less than 1%, and the computational time was more than 30 times faster than conventional OPF analysis.