The intelligent forecasting of the performances in PV/T collectors based on soft computing method

https://doi.org/10.1016/j.rser.2016.11.225Get rights and content

Abstract

Solar energy has been widely used in various aspects as the greatest promising and pollution free energy comparing with other available resources in nature. Photovoltaic-thermal (PV/T) is the most generative technology, which has been invented to utilize electrical energy and heat from the solar system. The article presents a novelty of using Extreme Learning Machine (ELM) into the air type PV/T technology. For this purposes, two air type PV/T designs were fabricated and practiced for a cooling fin design in the collector and finally, collected the experimental data, which was adapted to estimate electrical and thermal efficiency for the PV/T system. Then, the results of ELM prediction model were compared with Genetic Programming (GP) and Artificial Neural Networks (ANNs) models. The experimental result was accommodated to improving the predictive accuracy of the ELM approach in comparison. Further, outcome results indicate that developed ELM models can be used satisfactorily to formulate the predictive algorithm for PV/T performances. The ELM algorithm made a good generalization, which can learn very faster comparing with other conventional popular learning algorithms. The results revealed that the improved ELM model is a well fitted tool to predict the thermal and electrical efficiency with higher accuracy.

Introduction

Solar is the boundless source of energy in terms of abundance and environmental aspect [1]. The various solar technologies have been developed with new innovative and diverse features since last decades ago [1,2]. Solar thermal and photovoltaic system are the two important application area of solar energy [3]. Among the solar thermal technologies, photovoltaic-thermal (PV/T) is most famous as a co-generation process, which can generate thermal and electrical energy concurrently. An energy distribution system for PV/T model is illustrated in Fig. 1.

Solar energy has been received by two different ways including thermal and photovoltaic system [4], shown in Fig. 1. A typical commercial photovoltaic cell can convert 4–17% of incoming irradiation into useful electrical energy, whereas more than 50% can be utilized to generate thermal energy from the useful collector area [5]. For this purpose, several solar thermal collectors with the affluent design concept have been employed in the recent years [6], [7], [8]. PV/T system is upgrading day by day to purvey the thermal energy as the thermal efficiency (TE) in collector depends on the collector types and photovoltaic internal-external factors [9]. Some reviews summarize a recent advancement and application of flat plate PV/T that have been practiced on different configurations with several improvement methods [4], [5], [10], [11], [12], [13]. These reviews reveal the progression of collector design in recent years and the future required modification.

At the early stage of PV/T air heating development, professor Böer has built a ‘Solar One’ house in 1973/1974 at the University of Delaware, USA [14]. After him, in the early 1980s a pioneer research group of Hendrie [15], was carried out the details mathematical derivation and a logistic idea of PV/T air heating. An unglazed PV/T collector was initiated at the University of Patras, Greece [16], which was basically a liquid and air type heat extraction process. Thereafter, different parameters were analyzed widely, which was closely related to the enhancement of a collector performance. A prototype PV/T air heating system was built at Politecnico di Milano, Italy [17], and carried out to explore the air flow, collector tilt, and air gap. The application of heat augmented material has been improved remarkably due to their higher thermal performance. A study of compound parabolic concentrated (CPC) and conventional PV/T systems were followed by Elsafi et al. [18]. Therefore, the CPC and fins have been designed and fabricated in a prototype double pass PV/T system [19]. The experiment was extended with a details analysis for single and double fan operating system. A suspended thin flat metallic sheet (TFMS) and fin system for the collector were individually explored by Tonui et al. [20]. Most studies were related to the optimizing the collector variables parameters for improving the PV/T performance. Such as, a PV/T air collector models with various collector area's and different length to width ratios were widely described [21], [22]. A hybrid PV/T collector with a dual channel feature for the various working fluid was investigated [23]. Four air types hybrid PV/T models have been tested to analysis the TE and EE by Amori et al. [24].

The water type PV/T collector is the most popular system which has been widely studied [6], while air type collector has more advantages and simpler operating system. The objective of PV/T air collector is attaining the maximum heat gain and spontaneous reduction of PV cell surface temperature by controlled air flowing. The air type collector has been investigated to improve the PV module's EE by moving air [4]. In the application, the collected heat can be utilized widely in domestic and pre-water heating, air heating, low-temperature heating [25]. Pre-water heating is growing very fast and employed extensively in domestic, commercial buildings in most countries of the world [26]. These studies explore the number of methods to improve the PV/T performance. The method consists of several mechanical and thermodynamics systems, for an example, galvanized steel was used as a heat absorber material in the numerical modeling in the thermal collector [27]. In addition, different technique and designs have been proposed in the area of air type PV/T collector. Integration of insert devices in the collector have been widely discussed to improve the PV/T performance [20], [25], [27], [28], [29]. A heat sink (fins) made by aluminum (Al), brass, nickel and a copper material was installed to study the fin materials and shapes [30]. Therefore, the several dynamic models and CFD (Computational Fluid-Dynamics) analysis have been developed for the flat-plate Solar Air Collector (SAC) [31], [32], [33].

Various PV/T collector models have been proposed for its remarkable benefits within the last few years [2]. A significant research has been conducted to analysis PV/T market development, especially for PV and concentrating solar power plant [34]. Soft computing-based prediction techniques could be a suitable replacement to avoid the expensive test of PV/T. The prediction model is efficient to predict the outputs against some input variables for which they are not trained on. Application of artificial intelligence procedure in the context of heat transfer analysis has been gradually increasing, such as an Artificial Neural Network (ANN) method which gives a satisfactory prediction to calculate the thermal performance of SAC [35], [36]. In comparison with ANN method, other conventional methods have some limitation in terms of accuracy, especially for non-linear data [37]. A sequence of prediction model has been modified by Varol et al. [29], and employed three different soft computing tactics, including ANN, Adaptive-Net-work-Based Fuzzy Inference System (ANFIS) and Support Vector Machine (SVM) to carry out to analyze the SAC performance. In a case study which conducted by Esen et al. [38], with the aim of space heating, the SVM method was applied to Ground Couple Heat Pump (GCHP) as an intelligent approach. Another research work was intended to adopt the finding of modeling's system efficiency by means of Least-Square Support Vector Machine (LS-SVM) [39], Wavelet Neural Network (WNN) and ANN [40]. Thermal energy distribution model of water type PV/T was developed and ANN has been used to calculate Optimum Power Operating Point (OPOP) [41].

Nowadays, application of new computational approach for determining the optimal values, functions and solving the real problems are getting abundant attention by researchers in various engineering disciplines the different scientific areas. Among these techniques, Neural Network (NN) has been recently introduced as a most popular computational approach in a different engineering field. When classical parametric methods are incapable to effectively dealing with the complex non-linear problems, the ANN provides acceptable results. Regarding ANN training, Support Vector Machine (SVM), Back Propagation (BP), Hidden Markov Model (HMM) are the most popular algorithms. A popular algorithm called Extreme Learning Machine (ELM) was introduced for Single Layer Feed-Forward Neural Network (SLFN) by Huang et al. [42]. In addition, ELM requires less training time than NN due to the quick learning process and robust performance [43]. Consequently, a number of investigations have been performed effectively of ELM applications to solve the difficulties in various scientific fields [44], [45], [46], [47], [48], [49]. Mostly, the ELM is a powerful algorithm comparing with conventional algorithms like BP for faster learning response. Moreover, ELM attempts to get the little training error and norm of the weights.

The current article shows a strong motivation to introduce estimation model for electrical and thermal efficiency in the PV/T system by using a soft computing approach, ELM. Heat transfer to the different layers of two proposed PV/T models was practiced experimentally and introduced ELM as an efficient prediction algorithm to get the thermal and electrical performance with more accuracy. Latter, ELM results have been also compared with the GP and ANNs results by data training to get possible output from the passive design system.

Section snippets

Design and material

To improve the thermal performances in the thermal collector, several geometrical design concepts were found from previous literature study. Two models, named as model A and model B were investigated to analysis results. Both models have the similar parameters with PV module, air flow channel, thin film metallic sheet (TFMS), and thermal insulator except the presence of fin with a different dimension in model B. Longitudinal fins employed in model B was mounted between PV panel rear surface and

Results and discussion

Thermal heat gain is directly linked to the air flow rate in the PV/T system, with other condition such as solar radiation, surrounding condition and the rate of heat loss. The maximum heat gain was achieved for the highest operating air flow in the channel in case of the model B, also ensured the higher EE and TE. The flowing air can mitigate the temperature of the PV module, which affects negatively on module efficiency. The TE shown in Fig. 10, Fig. 11, is the sum of collector TE and

Conclusion

The present paper reports a prediction model of EE and overall TE based upon the experimental test data from two PV/T models. The first section of article describes the design and method for two PV/T models with a different configuration. Then, the experimental procedure was induced to measure the temperatures, solar radiations, mass flow rate. A methodical pathway of ELM predictive model was investigated for determining the performance of PV/T designs. ELM model predicted well the PV/T’s

Acknowledgement

The authors would like to acknowledge the Ministry of Higher Education of Malaysia and The University of Malaya, Kuala Lumpur, Malaysia for the financial support under SATU joint reseaech scheme: RU018J-2016 and postgraduate research grant: PG 239-2014B.

References (83)

  • S.R. Reddy et al.

    A review of PV–T systems: Thermal management and efficiency with single phase cooling

    International Journal of Heat and Mass Transfer

    (2015)
  • K.W. Böer et al.

    Solar conversion under consideration of energy and entropy

    Solar Energy

    (2003)
  • A.M. Elsafi et al.

    Comparative study of double-pass flat and compound parabolic concentrated photovoltaic–thermal systems with and without fins

    Energy Conversion and Management

    (2015)
  • J.K. Tonui et al.

    Improved PV/T solar collectors with heat extraction by forced or natural air circulation

    Renewable Energy

    (2007)
  • M. Farshchimonfared et al.

    Channel depth, air mass flow rate and air distribution duct diameter optimization of photovoltaic thermal (PV/T) air collectors linked to residential buildings

    Renewable Energy

    (2015)
  • M. Farshchimonfared et al.

    Full optimisation and sensitivity analysis of a photovoltaic–thermal (PV/T) air system linked to a typical residential building

    Solar Energy

    (2016)
  • D. Su

    Dynamic performance analysis of photovoltaic–thermal solar collector with dual channels for different fluids

    Energy Conversion and Management

    (2016)
  • K.E. Amori et al.

    Field study of various air based photovoltaic/thermal hybrid solar collectors

    Renewable Energy

    (2014)
  • R. Liang

    Performance evaluation of new type hybrid photovoltaic/thermal solar collector by experimental study

    Applied Thermal Engineering

    (2015)
  • K. Touafek et al.

    Design and modeling of a photovoltaic thermal collector for domestic air heating and electricity production

    Energy and Buildings

    (2013)
  • Y. Tripanagnostopoulos

    Aspects and improvements of hybrid photovoltaic/thermal solar energy systems

    Solar Energy

    (2007)
  • Y. Varol

    Forecasting of thermal energy storage performance of phase change material in a solar collector using soft computing techniques

    Expert Systems with Applications

    (2010)
  • A.M. Elsafi et al.

    Comparative study of double-pass flat and compound parabolic concentrated photovoltaic–thermal systems with and without fins

    Energy Conversion and Management

    (2015)
  • L.A. Tagliafico et al.

    Dynamic thermal models and CFD analysis for flat-plate thermal solar collectors–A review

    Energy Reviews

    (2014)
  • A. Khelifa

    Modeling and detailed study of hybrid photovoltaic thermal (PV/T) solar collector

    Solar Energy

    (2016)
  • K. Kramer et al.

    The interaction of standards and innovation: Hybrid photovoltaic–thermal collectors

    Solar Energy

    (2013)
  • M. Caner et al.

    Investigation on thermal performance calculation of two type solar air collectors using artificial neural network

    Expert Systems with Applications

    (2011)
  • A. Sözen et al.

    Determination of efficiency of flat-plate solar collectors using neural network approach

    Expert Systems with Applications

    (2008)
  • C.-S. Ong et al.

    Building credit scoring models using genetic programming

    Expert Systems with Applications

    (2005)
  • H. Esen

    Modeling a ground-coupled heat pump system by a support vector machine

    Renewable Energy

    (2008)
  • H. Esen

    Modelling of a new solar air heater through least-squares support vector machines

    Expert Systems with Applications

    (2009)
  • H. Esen

    Artificial neural network and wavelet neural network approaches for modelling of a solar air heater

    Expert systems with applications

    (2009)
  • M.B. Ammar et al.

    Artificial neural network based control for PV/T panel to track optimum thermal and electrical power

    Energy Conversion and Management

    (2013)
  • Q. Yu

    Bankruptcy prediction using extreme learning machineextreme learning machine and financial expertise

    Neurocomputing

    (2014)
  • X. Wang et al.

    Online sequential extreme learning machine with kernels for nonstationary time series prediction

    Neurocomputing

    (2014)
  • L. Ghouti et al.

    Mobility prediction in mobile ad hoc networks using extreme learning machines

    Procedia Computer Science

    (2013)
  • D.D. Wang et al.

    Fast prediction of protein–protein interaction sites based on extreme learning machines

    Neurocomputing

    (2014)
  • R. Nian

    Extreme learning machine towards dynamic model hypothesis in fish ethology research

    Neurocomputing

    (2014)
  • P.K. Wong

    Modeling and optimization of biodiesel engine performance using kernel-based extreme learning machine and cuckoo search

    Renewable Energy

    (2015)
  • M. Farshchimonfared et al.

    Channel depth, air mass flow rate and air distribution duct diameter optimization of photovoltaic thermal (PV/T) air collectors linked to residential buildings

    Renewable Energy

    (2015)
  • J.K. Tonui et al.

    Air-cooled PV/T solar collectors with low cost performance improvements

    Solar Energy

    (2007)
  • Cited by (33)

    • An extensive review on the latest developments of using oscillating heat pipe on cooling of photovoltaic thermal system

      2022, Thermal Science and Engineering Progress
      Citation Excerpt :

      Consequently, to control the PV panel temperature within the manufacturer’s operating range, it is essential to remove the generated heat from the photovoltaic panels by adapting cooling PV approaches. Cooling methods are applied to boost the efficiency of PV and PV/T panels, as displayed in Fig. 1 [11–28]. This study presents a detailed review regarding the research outcomes of the cooling approach using oscillating heat pipes (OHPs) or pulsating heat pipes (PHPs).

    • A parametric assessing and intelligent forecasting of the energy and exergy performances of a dish concentrating photovoltaic/thermal collector considering six different nanofluids and applying two meticulous soft computing paradigms

      2022, Renewable Energy
      Citation Excerpt :

      The soft computing prediction methods, which predict the output performance of a system based on different input parameters, are promising techniques to avoid the expensive experimental tests of the system ([26,27]). Mojumder et al. [28] applied three soft computing techniques (Extreme Learning Machine (ELM), Genetic Programming (GP), and Artificial Neural Networks (ANNs)) to present a model for a typical PVT performance based on the experimental data. Their results demonstrated that the ELM model is faster and more accurate than the two other models for predicting the electrical and thermal efficiencies of the PVT. Ammar et al. [29] applied the ANN model to develop the thermal energy distribution of a PVT and predict the optimum power production point.

    View all citing articles on Scopus
    View full text