A regime-dependent artificial neural network technique for short-range solar irradiance forecasting

Solar power can provide substantial power supply to the grid; however, it is also a highly variable energy source due to changes in weather conditions, i.e. clouds, that can cause rapid changes in solar power output. Independent systems operators (ISOs) and regional transmission organizations (RTOs) monitor the demand load and direct power generation from utilities, define operating limits and create contingency plans to balance the load with the available power generation resources. ISOs, RTOs, and utilities will require solar irradiance forecasts to effectively and efficiently balance the energy grid as the penetration of solar power increases. This study presents a cloud regime-dependent short-range solar irradiance forecasting system to provide 15-min average clearness index forecasts for 15-min, 60-min, 120-min and 180-min lead-times. A k-means algorithm identifies the cloud regime based on surface weather observations and irradiance observations. Then, Artificial Neural Networks (ANNs) are trained to predict the clearness index. This regime-dependent system makes a more accurate deterministic forecast than a global ANN or clearness index persistence and produces more accurate predictions of expected irradiance variability than assuming climatological average variability.     

Short term solar irradiance forecasting using a mixed wavelet neural network

In modern smart grids and deregulated electricity markets, accurate forecasting of solar irradiance is critical for determining the total energy generated by PV systems. We propose a mixed wavelet neural network (WNN) in this paper for short-term solar irradiance forecasting, with initial application in tropical Singapore. The key advantage of using wavelet transform (WT) based methods is the high signal compression ability of wavelets, making them suitable for modeling of nonstationary environmental parameters with high information content, such as short timescale solar irradiance. In this WNN, a combination of the commonly known Morlet and Mexican hat wavelets is used as the activation function for hidden-layer neurons of a feed forward artificial neural network (ANN). To demonstrate the effectiveness of the proposed approach, hourly predictions of solar irradiance, which is an aggregate sum of irradiance value observed using 25 sensors across Singapore, are considered. The forecasted results show that WNN delivers better prediction skill when compared with other forecasting techniques.     

Use of neural networks for the creation of hourly global and diffuse solar irradiance data at representative locations in Greece

In this work, a new approach is tested by applying neural networks treatment to meteorological time-series data sets, recorded during 1991–2000 at certain Greek locations, in order to create fully appropriate solar data information. Neural networks, in this case, are used for creating missing mean, maximum and minimum global and diffuse solar irradiance hourly data, when educated with other known meteorological time-series hourly values. For this purpose, hourly data of air temperature, relative humidity, sunshine duration, clouds’ octals, as well as local latitude are used with regard to these sites. Neural networks’ education process outputs are checked against known hourly values of solar irradiance, based upon the mentioned meteorological hourly raw data necessary for this action recorded at the National Observatory of Athens, the actinometric station at the Technological Education Institute (TEI) of Piraeus, and six other locations. Selection of these sites is representative of the climatic conditions in Greece, from north to south and east to west. Following the same scheme, the produced hourly global and diffuse mean hourly solar irradiance values are in a very good agreement (p<0.01) with actual measurements.     

Forecast of solar irradiance using recurrent neural networks combined with wavelet analysis

In this paper, artificial neural network is combined with wavelet analysis for the forecast of solar irradiance. This method is characteristic of the preprocessing of sample data using wavelet transformation for the forecast, i.e., the data sequence of solar irradiance as the sample is first mapped into several time-frequency domains, and then a recurrent BP network is established for each domain. The forecasted solar irradiance is exactly the algebraic sum of all the forecasted components obtained by the respective networks, which correspond respectively the time-frequency domains. Discount coefficients are applied to take account of different effect of different time-step on the accuracy of the ultimate forecast when updating the weights and biases of the networks in network training. On the basis of combination of recurrent BP networks and wavelet analysis, a model is developed for more accurate forecasts of solar irradiance. An example of the forecast of day-by-day solar irradiance is presented in the paper, the historical day-by-day records of solar irradiance in Shanghai constituting the data sample. The results of the example show that the accuracy of the method is more satisfactory than that of the methods reported before.     

Modeling solar still production using local weather data and artificial neural networks

A study has been performed to predict solar still distillate production from single examples of two different commercial solar stills that were operated for a year and a half. The purpose of this study was to determine the effectiveness of modeling solar still distillate production using artificial neural networks (ANNs) and local weather data. The study used the principal weather variables affecting solar still performance, which are the daily total insolation, daily average wind velocity, daily average cloud cover, daily average wind direction and daily average ambient temperature. The objectives of the study were to assess the sensitivity of the ANN predictions to different combinations of input parameters as well as to determine the minimum amount of inputs necessary to accurately model solar still performance. It was found that 31-78% of ANN model predictions were within 10% of the actual yield depending on the input variables that were selected. By using the coefficient of determination, it was found that 93-97% of the variance was accounted for by the ANN model. About one half to two thirds of the available long term input data were needed to have at least 60% of the model predictions fall within 10% of the actual yield. Satisfactory results for two different solar stills suggest that, with sufficient input data, the ANN method could be extended to predict the performance of other solar still designs in different climate regimes.     

Wavelet transform and neural networks for short-term electrical load forecasting

Demand forecasting is key to the efficient management of electrical energy systems. A novel approach is proposed in this paper for short term electrical load forecasting by combining the wavelet transform and neural networks. The electrical load at any particular time is usually assumed to be a linear combination of different components. From the signal analysis point of view, load can also be considered as a linear combination of different frequencies. Every component of load can be represented by one or several frequencies. The process of the proposed approach first decomposes the historical load into an approximate part associated with low frequencies and several detail parts associated with high frequencies through the wavelet transform. Then, a radial basis function neural network, trained by low frequencies and the corresponding temperature records is used to predict the approximate part of the future load. Finally, the short term load is forecasted by summing the predicted approximate part and the weighted detail parts. The approach has been tested by the 1997 data of a practical system. The results show the application of the wavelet transform in short term load forecasting is encouraging.     

Predicting total solar irradiation values using artificial neural networks

This study explores the possibility of developing an artificial neural networks model that could be used to predict monthly average daily total solar irradiation on a horizontal surface for locations in Uganda based on geographical and meteorological data: latitude, longitude, altitude, sunshine duration, relative humidity and maximum temperature. Results have shown good agreement between the predicted and measured values of total solar irradiation. A correlation coefficient of 0.997 was obtained with mean bias error of 0.018 MJ/m² and root mean square error of 0.131 MJ/m². Overall, the artificial neural networks model predicted with an accuracy of 0.1% of the mean absolute percentage error.     

Combining solar irradiance measurements,satellite-derived data and a numerical weather prediction model to improve intra-day solar forecasting

Isolated power systems need to generate all the electricity demand with their own renewable resources. Among the latter, solar energy may account for a large share. However, solar energy is a fluctuating source and the island power grid could present an unstable behavior with a high solar penetration. Global Horizontal Solar Irradiance (GHI) forecasting is an important issue to increase solar energy production into electric power system. This study is focused in hourly GHI forecasting from 1 to 6 h ahead. Several statistical models have been successfully tested in GHI forecasting, such us autoregressive (AR), autoregressive moving average (ARMA) and Artificial Neural Networks (ANN). In this paper, ANN models are designed to produce intra-day solar forecasts using ground and exogenous data. Ground data were obtained from two measurement stations in Gran Canaria Island. In order to improve the results obtained with ground data, satellite GHI data (from Helioclim-3) as well as solar radiation and Total Cloud Cover forecasts provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) are used as additional inputs of the ANN model. It is shown that combining exogenous data (satellite and ECMWF forecasts) with ground data further improves the accuracy of the intra-day forecasts.     

Generation of synthetic daily global solar radiation data based on ERA-Interim reanalysis and artificial neural networks

Four variables (total cloud cover, skin temperature, total column water vapour and total column ozone) from meteorological reanalysis were used to generate synthetic daily global solar radiation via artificial neural network (ANN) techniques. The goal of our study was to predict solar radiation values in locations without ground measurements, by using the reanalysis data as an alternative to the use of satellite imagery. The model was validated in Andalusia (Spain), using measured data for nine years from 83 ground stations spread over the region. The geographical location (latitude, longitude), the day of the year, the daily clear sky global radiation, and the four meteorological variables were used as input data, while the daily global solar radiation was the only output of the ANN. Sixty five ground stations were used as training dataset and eighteen stations as independent dataset. The optimum network architecture yielded a root mean square error of 16.4% and a correlation coefficient of 94% for the testing stations. Furthermore, we have successfully tested the forecasting capability of the model with measured radiation values at a later time. These results demonstrate the generalization capability of this approach over unseen data and its ability to produce accurate estimates and forecasts.     

Daily solar irradiation estimation over a mountainous area using artificial neural networks

In order to design both active and passive solar energy systems, radiation data are needed for the studied location. The implementation of such renewable energy systems is especially important in places like natural parks, where acoustic and fossil fuel derived contamination has to be completely avoided. Measure of solar radiation is usually accomplished by means of radiometric station nets with a low spatial resolution. To estimate the radiation in sites located away from the stations, different interpolation/extrapolation techniques may be used. These methods are valid on places where the spatial variability of radiation is not significant, but becomes less accurate if complex terrain areas are present in between the radiometric stations. As an alternative, artificial intelligence techniques have been used in this work, along with a 20 m resolution digital model of terrain. The inputs to the network have been selected using the automatic relevance determination methodology. The data set contains 3 years’ data of daily global radiation measured at 12 different stations located in the north face of the Sierra Nevada National Park in the surroundings of Huéneja (Granada), a town located in the South East of Spain. The stations altitude varies from 1000 to 1700 m. The goal of this work has been to estimate daily global irradiation on stations located in a complex terrain, and the values estimated by the neural network model have been compared with the measured ones leading to a root mean square error (RMSE) of 6.0% and a mean bias error (MBE) of 0.2%, both expressed as a percentage of the mean value. Performance achieved individually for each of the stations lies in the range [5.0–7.5]% for the RMSE and [−1.2 to +2.1]% for the MBE. Results point out artificial neural networks as an efficient and easy methodology for calculating solar radiation levels over complex mountain terrains from only one radiometric station data. In addition, this methodology can be applied to other areas with a complex topography.     

Study of hourly and daily solar irradiation forecast using diagonal recurrent wavelet neural networks

An accurate forecast of solar irradiation is required for various solar energy applications and environmental impact analyses in recent years. Comparatively, various irradiation forecast models based on artificial neural networks (ANN) perform much better in accuracy than many conventional prediction models. However, the forecast precision of most existing ANN based forecast models has not been satisfactory to researchers and engineers so far, and the generalization capability of these networks needs further improving. Combining the prominent dynamic properties of a recurrent neural network (RNN) with the enhanced ability of a wavelet neural network (WNN) in mapping nonlinear functions, a diagonal recurrent wavelet neural network (DRWNN) is newly established in this paper to perform fine forecasting of hourly and daily global solar irradiance. Some additional steps, e.g. applying historical information of cloud cover to sample data sets and the cloud cover from the weather forecast to network input, are adopted to help enhance the forecast precision. Besides, a specially scheduled two phase training algorithm is adopted. As examples, both hourly and daily irradiance forecasts are completed using sample data sets in Shanghai and Macau, and comparisons between irradiation models show that the DRWNN models are definitely more accurate.     

Modeling of solar radiation using remote sensing and artificial neural network in Turkey

Artificial neural networks (ANNs) were used to estimate solar radiation in Turkey (26–45°E, 36–42°N) using geographical and satellite-estimated data. In order to train the Generalized regression neural network (GRNN) geographical and satellite-estimated data for the period from January 2002 to December 2002 from 19 stations spread over Turkey were used in training (ten stations) and testing (nine stations) data. Latitude, longitude, altitude, surface emissivity for ?_4, surface emissivity for ?_5, and land surface temperature are used in the input layer of the network. Solar radiation is the output. Root Mean Square Error (RMSE) and correlation coefficient (R²) between the estimated and measured values for monthly mean daily sum with ANN values have been found as 0.1630 MJ/m² and 95.34% (training stations), 0.3200 MJ/m² and 93.41% (testing stations), respectively. Since these results are good enough it was concluded that the developed GRNN tool can be used to predict the solar radiation in Turkey.     

Modelling of solar energy potential in Nigeria using an artificial neural network model

In this study, an artificial neural network (ANN) based model for prediction of solar energy potential in Nigeria (lat. 4–14°N, log. 2–15°E) was developed. Standard multilayered, feed-forward, back-propagation neural networks with different architecture were designed using neural toolbox for MATLAB. Geographical and meteorological data of 195 cities in Nigeria for period of 10 years (1983–1993) from the NASA geo-satellite database were used for the training and testing the network. Meteorological and geographical data (latitude, longitude, altitude, month, mean sunshine duration, mean temperature, and relative humidity) were used as inputs to the network, while the solar radiation intensity was used as the output of the network. The results show that the correlation coefficients between the ANN predictions and actual mean monthly global solar radiation intensities for training and testing datasets were higher than 90%, thus suggesting a high reliability of the model for evaluation of solar radiation in locations where solar radiation data are not available. The predicted solar radiation values from the model were given in form of monthly maps. The monthly mean solar radiation potential in northern and southern regions ranged from 7.01–5.62 to 5.43–3.54 kW h/m² day, respectively. A graphical user interface (GUI) was developed for the application of the model. The model can be used easily for estimation of solar radiation for preliminary design of solar applications.     

Clear-sky solar luminous efficacy determination using artificial neural networks

Under cloudless conditions, the effect of atmospheric variables, such as turbidity or water vapour, on luminous efficacy is an important source of variability, often limiting the use of simple empirical models to those sites where they were developed. Due to the complex functional relationship between these atmospheric variables and the luminous efficacy components, deriving a non-local model considering all these physical processes is nearly impossible if standard statistical techniques are employed. To avoid this drawback, the use of a new methodology based on artificial neural networks (ANN) is investigated here to determine the luminous efficacy of direct, diffuse and global solar radiation under cloudless conditions. In this purpose, a detailed spectral radiation model (SMARTS) is utilized to generate both illuminance and solar radiation values covering a large range of atmospheric conditions. Different input configurations using combinations of atmospheric variables and radiometric quantities are analyzed. Results show that an ANN model using direct and diffuse solar irradiance along with precipitable water is able to accurately reproduce the variations of the three components of luminous efficacy caused by solar zenith angle and the various atmospheric absorption and scattering processes. This proposed model is considerably simpler than the SMARTS radiation model it is derived from, but still can retain most of its predicting power and versatility. The proposed ANN model can thus be used worldwide, avoiding the need of using detailed atmospheric information or empirical models of the literature if radiometric measurements and precipitable water data (or temperature and relative humidity data) are available.     

Estimation of global solar radiation using artificial neural networks

This paper introduces a neural network technique for the estimation of global solar radiation. There are 41 radiation data collection stations spread all over the kingdom of Saudi Arabia where the radiation data and sunshine duration information are being collected since 1971. The available data from 31 locations is used for training the neural networks and the data from the other 10 locations is used for testing. The testing data was not used in the modeling to give an indication of the performance of the system in unknown locations. Results indicate the viability of this approach for spatial modeling of solar radiation.     

A dynamic compact thermal model for data center analysis and control using the zonal method and artificial neural networks

Full-scale data center thermal modeling and optimization using computational fluid dynamics (CFD) is generally an extremely time-consuming process. This paper presents the development of a velocity propagation method (VPM) based dynamic compact zonal model to efficiently describe the airflow and temperature patterns in a data center with a contained cold aisle. Results from the zonal model are compared to those from full CFD simulations of the same configuration. A primary objective of developing the compact model is real-time predictive capability for control and optimization of operating conditions for energy utilization. A scheme is proposed that integrates zonal model results for temperature and air flow rates with a proportional–integral–derivative (PID) controller to predict and control rack inlet temperature more precisely. The approach also uses an Artificial Neural Network (ANN) in combination with a Genetic Algorithm (GA) optimization procedure. The results show that the combined approach, built on the VPM based zonal model, can yield an effective real-time design and control tool for energy efficient thermal management in data centers.     

A 24-h forecast of solar irradiance using artificial neural network: Application for performance prediction of a grid-connected PV plant at Trieste, Italy

Forecasting of solar irradiance is in general significant for planning the operations of power plants which convert renewable energies into electricity. In particular, the possibility to predict the solar irradiance (up to 24 h or even more) can became - with reference to the Grid Connected Photovoltaic Plants (GCPV) - fundamental in making power dispatching plans and - with reference to stand alone and hybrid systems - also a useful reference for improving the control algorithms of charge controllers. In this paper, a practical method for solar irradiance forecast using artificial neural network (ANN) is presented. The proposed Multilayer Perceptron MLP-model makes it possible to forecast the solar irradiance on a base of 24 h using the present values of the mean daily solar irradiance and air temperature. An experimental database of solar irradiance and air temperature data (from July 1st 2008 to May 23rd 2009 and from November 23rd 2009 to January 24th 2010) has been used. The database has been collected in Trieste (latitude 45°40′N, longitude 13°46′E), Italy. In order to check the generalization capability of the MLP-forecaster, a K-fold cross-validation was carried out. The results indicate that the proposed model performs well, while the correlation coefficient is in the range 98-99% for sunny days and 94-96% for cloudy days. As an application, the comparison between the forecasted one and the energy produced by the GCPV plant installed on the rooftop of the municipality of Trieste shows the goodness of the proposed model.     

Application of artificial neural networks for the wind speed prediction of target station using reference stations data

In this study, artificial neural networks (ANNs) were applied to predict the mean monthly wind speed of any target station using the mean monthly wind speeds of neighboring stations which are indicated as reference stations. Hourly wind speed data, collected by the Turkish State Meteorological Service (TSMS) at 8 measuring stations located in the eastern Mediterranean region of Turkey were used. The long-term wind data, containing hourly wind speeds, directions and related information, cover the period between 1992 and 2001. These data were divided into two sections. According to the correlation coefficients, reference and target stations were defined. The mean monthly wind speeds of reference stations were used and also corresponding months were specified in the input layer of the network. On the other hand, the mean monthly wind speed of the target station was utilized in the output layer of the network. Resilient propagation (RP) learning algorithm was applied in the present simulation. The hidden layers and output layer of the network consist of logistic sigmoid transfer function (logsig) and linear transfer function (purelin) as an activation function. Finally, the values determined by ANN model were compared with the actual data. The maximum mean absolute percentage error was found to be 14.13% for Antakya meteorological station and the best result was found to be 4.49% for Mersin meteorological station.     

Estimation of solar radiation components incident on Helwan site using neural networks

Field testing carried out for solar energy applications is costly, time consuming and depends heavily on prevailing weather conditions. Adequate security and weather protection must be provided at the test site. Measurements may also suffer from delays that can be caused by system failures and bad weather. To overcome these problems the need for accurate model becomes evermore important. To achieve such prediction task, an artificial neural network, ANN, model is regarded as a cost-effective technique superior to traditional statistical methods. In this paper, Levenberg optimization function is adopted to predict insolation data in different spectral bands for Helwan (Egypt) monitoring station. The predicted values were then compared with the actual values and presented in terms of usual statistics. The results hint that, the ANN model predicted infrared, ultraviolet, and global insolation with a good accuracy of approximately 95%, 93% and 96%, respectively. In addition, ANN model was tested to predict the same components for Aswan over an 11 month period. The predicted values of the ANN model compared to the actual values for Aswan produced an accuracy of 95%, 91% and 92%, respectively. Data for Aswan were not included as a part of ANN training set. Hence, these results demonstrate the generalization capability of this approach over unseen data and its ability to produce accurate estimates.     

Experimental analysis of dynamic characteristics on the PEM fuel cell stack by using Taguchi approach with neural networks

This study determines the optimum operating parameters for a proton exchange membrane fuel cell (PEMFC) stack to obtain small variation and maximum electric power output using a robust parameter design (RPD). The operating parameters examined experimentally are operating temperatures, operating pressures, anode/cathode humidification temperatures, and reactant flow rates. First, the dynamic Taguchi method is used to obtain the maximum and stable power density against the different current densities, which are regarded as the systemic inputs considered a signal factor. The relationship between control factors and responses in the PEMFC stack is determined using a neural network. The discrete parameter levels in the dynamic Taguchi method can be divided into desired levels to acquire real optimum operating parameters. Based on these investigations, the PEMFC stack is operated at the current densities of 0.4–0.8 A/cm². Since the voltage shift is quite small (roughly 0.73–0.83 V for each single cell), the efficiency would be higher. In the range of operation, the operating pressure, the cathode humidification temperature and the interactions between operating temperature and operating pressure significantly impact PEMFC stack performance. As the operating pressure increasing, the increments of the electric power decrease, and power stability is enhanced because the variation in responses is reduced.