Using measured daily meteorological parameters to predict daily solar radiation
Introduction
Renewable energy sources such as solar have emerged as effective alternatives to fossil fuels. Optimal design of solar systems requires exact prediction of the solar energy [1], [2], [3], [4]. To this aim, several empirical methods have been developed to avoid performing costly in-situ solar radiation measurements [5], [6]. Some of the well-known methods in this area are auto-regression, Markov chain, or robust optimization techniques [7], [8], [9]. Among the empirical methods, machine learning has been widely used to solve real world problems [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28]. Artificial neural networks (ANNs) are well-known machine learning systems that have been utilized to predict the solar radiation [2], [3], [4], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40]. As typical examples in this context, Rahimikhoob [41] successfully applied ANN for predicting the global solar radiation. Alam et al. [42] employed ANN to predict beam solar radiation. Mellit et al. [43] conducted a comprehensive study on application multi-layer perceptron of ANN for estimating total solar radiation data. However, a limitation of ANNs is that they cannot always be converted into explicit forms to provide the details of the prediction process [44].
Genetic programming (GP) [45] is a new approach with notable simulation capabilities. It is a division of genetic algorithms (GA) that generates computer programs rather [44]. For the last decade, GP has been used to formulate complicated engineering problems [44], [46], [47], [48], [49], [50], [51], [52]. Gene expression programming (GEP) [53], [54] is a recent branch of GP evolving programs of various sizes and shapes. Compared to other soft computing methods, it has been barely applied to energy related problems [2], [3], [55], [56], [57]. Although ANNs are used to predict the “daily” solar radiation [41], [58], [59], [60], [61], [62], none of the existing studies have focused on GP-based analysis of daily solar radiation. Therefore, this study presents the GEP technique for predicting the daily solar radiation in Iran. A regression analysis was later preformed to benchmark the proposed model.
Section snippets
Methodology
GP is an optimization technique to create computer programs. The process is inspired by the biological evolution of living organisms [45]. Generally, in GP, the main goal is to find a program that connects inputs to outputs (see Fig. 1) [55]. The solutions derived by conventional GP are shown as tree structures [63]. However, there are other branches of GP that produce models in different shapes (see Fig. 2) [44], [55]. Among different variants, linear variants have a higher speed permitting
The methodology for predicting daily solar radiation
The soft computing tools commonly follow similar steps to develop a prediction model for the daily solar radiation [3], [55], [67]. A similar approach was also considered herein.
Performance verification
It is known that |R| > 0.8 is a good indicator of accuracy of a regression model [69]. Moreover, the values of RMSE and MAE should be minimum. It can be observed from Fig. 5 that the GEP model has acceptable performance both for the training (RTraining = 0.866, RMSETraining = 1.032 kW h/m2/d, MAETraining = 0.817 kW h/m2/d) and testing (Rtesting = 0.881, RMSEtesting = 1.202 kW h/m2/d, MAEtesting = 0.940 kW h/m2/d) data. Besides, new measures were checked for more verification of the models on the testing [70], [71].
Sensitivity analysis
As discussed before, the effect of all of the considered parameters (i.e., Tave, Tmin, Tmax, H, P, W, and E) on DS is well understood. Ignoring any of these three parameters for the model development resulted in a model with poor performance. Herein, a sensitivity analysis was conducted to provide a more in depth understanding of the contribution of these important parameters to the prediction of DS. A common approach for the sensitivity analysis in the GP-based modeling is to obtain the input
Conclusion
In this study, a new GEP approach is presented for the prediction of the daily solar radiation. A comprehensive database containing data collected for about 20 years was used to develop the GEP model. The model predicts the daily solar radiation with an acceptable accuracy and outperforms the developed regression-based model. GEP does not need a predefined function for the modeling of the solar radiation. As expected, the solar radiation is more affected by average of air temperature, wind speed
References (71)
Prediction of direct and global solar irradiance using broadband models: validation of REST model
Renewable Energy
(2006)- et al.
An integrated artificial neural networks approach for predicting global radiation
Energy Convers. Manage.
(2009) - et al.
A hybrid computational approach to estimate solar global radiation: an empirical evidence from Iran
Energy
(2013) Modeling global solar radiation using Particle Swarm Optimization (PSO)
Sol. Energy
(2012)- et al.
New sunshine-based models for predicting global solar radiation using PSO (particle swarm optimization) technique
Energy
(2011) - et al.
Determination of the length of hydraulic jumps using artificial neural networks
Adv. Eng. Soft.
(2012) - et al.
Sensitivity analysis of the artificial neural network outputs in simulation of the evaporation process at different climatologic regimes
Adv. Eng. Soft.
(2012) - et al.
Artificial neural network predictors for mechanical properties of cold rolling products
Adv. Eng. Soft.
(2012) Predicting of torsional strength of RC beams by using different artificial neural network algorithms and building codes
Adv. Eng. Soft.
(2010)Artificial neural network based modeling of performance characteristics of deep well pumps with splitter blade
Energy Convers. Manage.
(2006)
An architectural framework for the construction of hybrid intelligent forecasting systems: application for electricity demand prediction
Eng. Appl. Artif. Intell.
Evolutionary computation in power systems
Int. J. Power Energy Syst.
Hydroelectric power plant management relying on neural networks and expert system integration
Eng. Appl. Artif. Intel.
A fuzzy-based approach for strategic choices in electric energy supply The case of a Swiss power provider on the eve of electricity market opening
Eng. Appl. Artif. Intel.
Forecasting next-day price of electricity in the Spanish energy market using artificial neural networks
Eng. Appl. Artif. Intel.
Hybrid System for fouling control in biomass boilers
Eng. Appl. Artif. Intel.
Energy demand prediction using GMDH networks
Neurocomputing
Annual electricity consumption forecasting by neural network in high energy consuming industrial sectors
Energy Convers. Manage.
Prediction of principal ground-motion parameters using a hybrid method coupling artificial neural networks and simulated annealing
Comput. Struct.
A simplified model for generating sequences of global solar radiation data for isolated sites: using artificial neural network and a library of Markov transition matrices approach
Sol. Energy
Univariate and multivariate forecasting of hourly solar radiation with artificial intelligence techniques
Sol. Energy
Forecasting based on neural network approach of solar potential in Turkey
Renewable Energy
Use of artificial neural networks for mapping of solar potential in Turkey
Appl. Energy
Time series modeling and large scale global solar radiation forecasting from geostationary satellites data
Sol. Energy
Estimation of solar radiation over Turkey using artificial neural network and satellite data
Appl. Energy
Forecasting of preprocessed daily solar radiation time series using neural networks
Sol. Energy
The potential of different artificial neural network (ANN) techniques in daily global solar radiation modeling based on meteorological data
Sol. Energy
Estimating global solar radiation using artificial neural network and air temperature data in a semi-arid environment
Renewable Energy
Computation of beam solar radiation at normal incidence using artificial neural network
Renewable Energy
Methodology for predicting sequences of mean monthly clearness index and daily solar radiation data in remote areas: application for sizing a stand-alone PV system
Renewable Energy
Stepwise approach for the evolution of generalized genetic programming model in prediction of surface finish of the turning process
Adv. Eng. Soft.
A machine code-based genetic programming for suspended sediment concentration estimation
Adv. Eng. Soft.
Target detection in SAR imagery by genetic programming
Adv. Eng. Soft.
Prediction of shear strength of FRP-reinforced concrete beams without stirrups based on genetic programming
Adv. Eng. Soft.
Modeling of transfer length of prestressing strands using genetic programming and neuro-fuzzy
Adv. Eng. Soft.
Cited by (33)
STO-DAMV: Sequential topology optimization and dynamical accelerated mean value for reliability-based topology optimization of continuous structures
2023, Computer Methods in Applied Mechanics and EngineeringShort term multi-steps wind speed forecasting for carbon neutral microgrid by decomposition based hybrid model
2023, Energy for Sustainable DevelopmentA review on global solar radiation prediction with machine learning models in a comprehensive perspective
2021, Energy Conversion and ManagementA comprehensive review of hybrid models for solar radiation forecasting
2020, Journal of Cleaner ProductionCitation Excerpt :The authors also justify the choice of CS algorithm by performing a comparison with other optimization algorithm DE and ACO and showed that CS provides an accurate performance compared to the DE and ACO. Mousavi et al. (2015) proposed a new artificial approach for daily global solar radiation prediction on horizontal surface in Iran. The proposed hybridization method integrates the ANN model with Simulated annulling (SA) to form ANN-SA ensemble prediction model.
New daily global solar irradiation estimation model based on automatic selection of input parameters using evolutionary artificial neural networks
2019, Journal of Cleaner ProductionCitation Excerpt :On the other hand, recent studies made an effort to use parameters that are routinely recorded at a large number of meteorological stations such as E. F. Alsina et al. in (Alsina et al., 2016) where the used parameters were: top of atmosphere irradiation, day length, number of rainy days, average rainfall, latitude and altitude. In another important work (Mousavi et al., 2015), the authors proposed a prediction model using daily air temperature, relative humidity, atmospheric pressure, wind speed and earth temperature providing good accuracy in the prediction of solar irradiation. Furthermore, we note that all works used one or two hidden layers, which is sufficient for modeling and estimating solar irradiation (Al-Alawi and Al-Hinai, 1998; Benghanem et al., 2009; Premalatha and Valan Arasu, 2016).