Integration of Gaussian Processes and Particle Swarm Optimization for Very-Short Term Wind Speed Forecasting in Smart Power

Integration of Gaussian Processes and Particle Swarm Optimization for Very-Short Term Wind Speed Forecasting in Smart Power

Miltiadis Alamaniotis (Purdue University, West Lafayette, USA) and Georgios Karagiannis (University of Durham, Durham, UK)
DOI: 10.4018/IJMSTR.2017070101


This article describes how the integration of renewable energy in the power grid is a critical issue in order to realize a smart grid infrastructure. To that end, intelligent methods that monitor and currently predict the values of critical variables of renewable energy are essential. With respect to wind power, such variable is the wind speed given that it is of great interest to efficient schedule operation of a wind farm. In this article, a new methodology for predicting wind speed is presented for very short-term prediction horizons. The methodology integrates multiple Gaussian process regressors (GPR) via the adoption of an optimization problem whose solution is given by the particle swarm optimization algorithm. The optimized framework is utilized for the average hourly wind speed prediction for a prediction horizon of six hours ahead. Results demonstrate the ability of the methodology in accurately forecasting the wind speed. Furthermore, obtained forecasts are compared with those taken from single Gaussian process regressors as well from the integration of the same multiple GPR using a genetic algorithm.
Article Preview

1. Introduction

Integration of renewable energy in the power grid is one of the cornerstones in building the smart power system of the future (Farhangi, 2010). Renewable energy is not only a sustainable source of energy, but most importantly, it may contribute in greener and less polluted cities of the future (Brenna et al., 2012). Therefore, utilization of renewable energy has profound benefits that may not go overlooked. With respect to energy sources, solar and wind are the most prominent and promising energy sources (Dincer, 2000).

Wind power is produced by the operation of wind mills. The driving force behind the wind power production is the wind intensity as it is expressed in terms of speed. In order to fully exploit the wind speed and produce adequate amount of power, the wind mills are grouped together in an area of close geographic vicinity. The group of wind mills as a whole consists of a “wind farm” that may be seen as the equivalent of a power plant, which uses conventional fuel (Papathanassiou and Boulaxis, 2006).

As opposed to conventional power plants the wind farm does not constantly produce the same amount of power. The reason behind that is the nature of the driving force; wind speed is a stochastic variable and cannot be controlled by human means (Aksoy et al., 2005). As a result, scheduling wind power production is a very challenging task and difficult to fully exploit. For instance, during consumption peak hours, when there is a great need for excess power, wind farms might not produce any power because of the lack of wind. In contrast, wind power may be available during times in which the load demand is very low, e.g., after midnight. In addition, the lack of efficient solution for large scale electricity power, results in wasting the generated from wind power.

Smart power systems come to fill the gap in efficient utilization of wind power. They are the result of the integration of power systems with information technologies (Alamaniotis et al., 2010). The overall idea is that use of information in power systems may compensate for the lack of physical storage (Alamaniotis and Tsoukalas, 2013). One of the crucial tools in implementing smart power systems is anticipation (Alamaniotis and Agarwal, 2014; Tsoukalas and Gao, 2008). Anticipation promotes planning and subsequent scheduling of production and consumption activities; in other words, it allows the intelligent management of the power system.

With respect to wind power production, anticipation may be adopted for wind speed forecasting. Speed forecasting allows wind farm operators to schedule the operation of the wind mills and estimate the amount of produced energy at specific time of the day. In addition, it assists 1) the system operator to schedule the operation of the plant units, and 2) the market operator to determine the cost of power ($/Kwh). Overall, wind speed forecasting is a great tool for the efficient and economically operation of power system (Wang et al., 2004).

In this paper, a new methodology for wind power forecasting is being presented. The methodology aims in predicting the wind speed in very short-term prediction horizon. It should be noted that there is a high variety of methods that exist in the literature that use tools from artificial intelligence and statistics (Cadenas and River, 2010; Du et al., 2008; Li and Shi, 2010;Lei at al., 2014; Soman et al., 2010). However, most of those methods deal with the problem of short term forecasting, e.g., a day ahead forecasting, while they require a huge amount of data. Our methodology, aspires in solving the problem of predicting the wind speed for a very short ahead of time horizon, while using a minimal amount of previous recorded data. Furthermore, it aims in capturing the dynamics of the wind in very short term ahead of time, and subsequently predicting any abrupt changes in wind speed.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 7: 2 Issues (2019): Forthcoming, Available for Pre-Order
Volume 6: 2 Issues (2018): Forthcoming, Available for Pre-Order
Volume 5: 4 Issues (2017)
Volume 4: 4 Issues (2016)
Volume 3: 4 Issues (2015)
Volume 2: 4 Issues (2014)
Volume 1: 4 Issues (2013)
View Complete Journal Contents Listing