Solar Tracking System by Utilized Optimized Algorithm Based Deep Learning

Al-Muswe, Aya Fadel Abed; Al-Amri, Izz Kadhum Abboud; Al-Turfi, Mohammed Nasser; Alkhayyat, Ahmed

doi:10.1109/iiceta54559.2022.9888545

Cited by 3 publications

(6 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors in [102,106,116,117,121,125,126,129,131,132] used the Min-Max scaling approach to normalize their data. On the other hand, the authors in [111,133] used the standard scaler approach to normalize the data to have a zero mean. The various works reviewed also stated that data normalization helped improve the performance of the proposed learning algorithms.…”

Section: Which Data Preprocessing Methods Are Applied To Solar Tracke...mentioning

confidence: 99%

“…Adam optimizer was used in various works, [111,129,132,133]. The authors in these studies used the Adam optimizer to maintain an elementary moving average of the gradient parameters.…”

Section: Review Of Optimization Methodsmentioning

confidence: 99%

“…Further, the authors in [105,107,110] used LDR sensor readings and solar power data as inputs to train the proposed DL model. Additionally, the authors in [98,109,129,133] used astronomical data to train their models to predict the sun's path. Arif et al [123] used astronomical data, energy performance data and LDR readings to train a neuro-fuzzy-based controller for a PV solar panel tracking system.…”

Section: What Is the Nature Of Datasets Used In Solar Tracker-based D...mentioning

confidence: 99%

“…Setiawan et al [98] used 𝑅𝑀𝑆𝐸 to evaluate the forecasting performance of an ANFIS-based model. Similarly, the researchers in [100,133], used 𝑅𝑀𝑆𝐸 as a measure of the predictive performance of an MLP-based model. Hijawi & Arafeh [105], used 𝑀𝐴𝑃𝐸, 𝑀𝑃𝐸, and 𝑅 to measure the error size of the proposed neuro-fuzzy-based controller for a dual-axis solar tracker.…”

Section: Which Performance Metrics Are Used To Evaluate Solar Trackin...mentioning

confidence: 99%

“…The WEC was preferred due to its ability to reduce time-series noise and perform feature extraction. Similarly, Al-Muswe et al[133] used a DHL model by combining a Deep-RNN and LSTM to predict the sun's position.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Deep Learning Techniques for Photovoltaic Solar Tracking Systems: A Systematic Literature Review

Phiri

Mulenga

Zimba

et al. 2023

Preprint

View full text Add to dashboard Cite

Although photovoltaic (PV) panels are extensively used to convert solar energy into electric energy, the continuous change in the sun's angle with reference to the earth's surface limits their capacity to collect sufficient energy. To improve efficiency, solar trackers are used to constantly adjust the PV panels towards the sun to maximize energy capture. There has been an increase in the use of deep learning (DL) in solar tracking systems as it has proven to be one of the most efficient techniques owing to its success in related fields. Although there are several studies on different designs of solar tracking systems, a synthesis of the state-of-art knowledge is lacking in the literature. Therefore, this study carried out a review of the DL methods used in solar tracking systems. Specifically, dataset usage, preprocessing methods, feature engineering methods, DL algorithms and the performance metrics used in the identified studies. The review considered the studies published from 2012 to 2022. In the initial search, 5,724 articles were selected from 7 digital libraries. Only 37 academic papers were included in the review based on the inclusion criteria. The results revealed that deep hybrid learning models were the most popular among researchers. Further, the study identified research challenges and future directions relating to the availability of DL-based PV solar tracking with respect to datasets, image data, data normalization, data decomposition, and feature engineering methods. This work will be resourceful to current and future researchers in addressing the trends and challenges related to the application of DL in PV solar trackers.

show abstract

Section: Which Data Preprocessing Methods Are Applied To Solar Tracke...mentioning

confidence: 99%

“…Adam optimizer was used in various works, [111,129,132,133]. The authors in these studies used the Adam optimizer to maintain an elementary moving average of the gradient parameters.…”

Section: Review Of Optimization Methodsmentioning

confidence: 99%

Section: What Is the Nature Of Datasets Used In Solar Tracker-based D...mentioning

confidence: 99%

Section: Which Performance Metrics Are Used To Evaluate Solar Trackin...mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Deep Learning Techniques for Photovoltaic Solar Tracking Systems: A Systematic Literature Review

Phiri

Mulenga

Zimba

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

A Multistage Hybrid Deep Learning Model for Enhanced Solar Tracking

Mulenga,

Phiri,

Simukonda

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Solar tracking helps maximize the efficiency of solar applications, such as photovoltaic (PV) solar panels. In the recent past, machine learning (ML) techniques have been extensively used to implement automatic solar tracking. However, applying predictive models in solar trackers is a non-trivial task due to the randomness and non-linearity of meteorological data, limiting their ability to clearly represent the underlying data patterns. Most existing predictive models take a monolithic approach to addressing limitations related to meteorological data, thereby limiting their performance. Therefore, this paper proposes a deep hybrid learning (DHL) model to enhance solar tracking performance. Furthermore, the proposed model improves feature representation in the data by using combined normalization methods and conversion of numerical data to images. In a nutshell, the model integrates sine and cosine transformations (SCT) to reveal cyclical patterns in the data, sigmoid and minimum-maximum data transformations to scale the data to a Gaussian distribution, and Gramian Angular Fields (GAF) to convert tabular data into 2D image representation to take advantage of the feature extraction capability of a convolutional neural network (CNN). The model also utilizes longterm short-term memory (LSTM) and gated recurrent units (GRU) for both spatial and temporal feature extraction. The results show that the aggregation of the above-mentioned methods significantly enhances solar tracking. The proposed hybrid model outperforms existing methods on a publicly available dataset, achieving outstanding performance with MAE, MAPE, and RMSE scores of 0.0073, 1.4635, and 0.0097, respectively.

show abstract