Wind speed is one of the primary renewable sources for clean power. However, it is intermittent, presents nonlinear patterns, and has nonstationary behavior. Thus, the development of accurate approaches for its forecasting is a challenge in wind power generation engineering. Hybrid systems that combine linear statistical and Artificial Intelligence (AI) forecasters have been highlighted in the literature due to their accuracy. Those systems aim to overcome the limitations of the single linear and AI models. In the literature about wind speed, these hybrid systems combine linear and nonlinear forecasts using a simple sum. However, the most suitable function for combining linear and nonlinear forecasts is unknown and the linear relationship assumption can degenerate or underestimate the performance of the whole system. Thus, properly combining the forecasts of linear and nonlinear models is an open question and its determination is a challenge. This paper proposes a hybrid system for monthly wind speed forecasting that uses a nonlinear combination of the linear and nonlinear models. A data-driven intelligent model is used to search for the most suitable combination, aiming to maximize the performance of the system. An evaluation has been carried out using the monthly data from three wind speed stations in northeast Brazil, evaluated with two traditional metrics. The assessment is performed for two scenarios: with and without exogenous variables. The results show that the proposed hybrid system attains an accuracy superior to other hybrid systems and single linear and AI models.

No abstract

Solar irradiance forecasting has been an essential topic in renewable energy generation. Forecasting is an important task because it can improve the planning and operation of photovoltaic systems, resulting in economic advantages. Traditionally, single models are employed in this task. However, issues regarding the selection of an inappropriate model, misspecification, or the presence of random fluctuations in the solar irradiance series can result in this approach underperforming. This paper proposes a heterogeneous ensemble dynamic selection model, named HetDS, to forecast solar irradiance. For each unseen test pattern, HetDS chooses the most suitable forecasting model based on a pool of seven well-known literature methods: ARIMA, support vector regression (SVR), multilayer perceptron neural network (MLP), extreme learning machine (ELM), deep belief network (DBN), random forest (RF), and gradient boosting (GB). The experimental evaluation was performed with four data sets of hourly solar irradiance measurements in Brazil. The proposed model attained an overall accuracy that is superior to the single models in terms of five well-known error metrics.