Kurniawan, Amar (2025) Perancangan Sistem Monitoring Efisiensi dan Keandalan Secara Real-Time Pada Portable Solar Power Plant Berbasis Decision Tree Regression. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5009211040-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5009211040-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Transisi energi global menuju sumber daya terbarukan menuntut sistem pembangkit listrik yang efisien, andal, dan adaptif terhadap kondisi lingkungan. Penelitian ini berfokus pada perancangan sistem monitoring efisiensi dan keandalan secara real-time pada portable Solar power plant berbasis algoritma decision tree regression untuk mengoptimalkan performansi sistem konversi energi surya. Sistem ini dikembangkan dengan integrasi internet of things (IoT) menggunakan mikrokontroler ESP32, sensor PZEM-017, dan piranti ukur temperatur serta irradiansi untuk mendapatkan parameter input dan output secara berkelanjutan dengan solar power plant didesain portable bisa dipindahkan saat tidak digunakan. Model machine learning digunakan untuk mengevaluasi pengaruh variasi state of charge (SOC) baterai terhadap efisiensi sistem serta nilai keandalan berdasarkan parameter statistic seperti mean squared error, Root mean squared error, dan koefisien determinasi. Penelitian ini juga mencakup kalibrasi sensor berbasis metode karakteristik statik dan dinamika untuk menjamin akurasi dan presisi data. Hasil dari model prediktif menunjukkan korelasi yang signifikan antara variabel lingkungan dan performa keluaran dari model decision tree yaitu, Sebesar R² = 0.7101, mse= 0.8 dalam memprediksi arus panel, sedangkan dalam prediksi arus baterai didapatkan nilai performansi R² =0.55 mse=3.26 dengan variasi nilai SOC tidak berpengaruh secara signifikan terhadap performansi, dan didapatkan nilai keandalan terhadap parameter output nya diatas 80% sedangkan untuk subsistem panel nilai keandalan tertinggi diperoleh pada parameter arus panel yaitu 81.52%, Untuk hasil efisiensi diperoleh hasil rentang efisiensi panel surya adalah 0.35-92.77% dan efisiensi SCC adalah 0.74% - 99.36%.
=======================================================================================================================================
The global energy transition towards renewable resources requires power generation systems that are efficient, reliable, and adaptable to environmental conditions. This research focuses on the design of a real-time efficiency and reliability monitoring system for portable solar power plants based on decision tree regression algorithms to optimize the performance of solar energy conversion systems. This system was developed with the integration of the Internet of Things (IoT) using an ESP32 microcontroller, PZEM-017 sensor, and temperature and irradiance measurement devices to continuously obtain input and output parameters with a portable solar power plant that can be moved when not in use. A machine learning model was used to evaluate the influence of battery state of charge (SOC) variations on system efficiency and reliability based on statistical parameters such as mean squared error, root mean squared error, and coefficient of determination. This study also includes sensor calibration based on static and dynamic characteristic methods to ensure data accuracy and precision. The results of the predictive model show a significant correlation between environmental variables and the output performance of the decision tree model, with R² = 0.7101 and MSE = 0.8 in predicting panel current, while in predicting battery current, the performance values obtained were R² = 0.55 and MSE = 3.26, with variations in SOC values not significantly affecting performance. and the reliability value for the output parameters is above 80%. For the panel subsystem, the highest reliability value is obtained for the panel current parameter at 81.52%. For efficiency results, the range of solar panel efficiency is 0.35–92.77%, and SCC efficiency is 0.74%–99.36%.
Actions (login required)
 |
View Item |