Solar Panel Lifetime Detection using Deep Learning Network Based on Temperature and Humidity Sensors

Main Article Content

Yinghua Zheng
Yuanyuan Zheng

Abstract

Solar photovoltaic (PV) performance is reduced due to the increase in panel temperature. Solar PV panels must be kept at an ideal temperature to work at their peak and have the longest useful life possible. The current generation of temperature sensors has a slow response time, poor resolution, and poor accuracy. More precision, a larger dynamic range, and very high sample rates are all provided by fiber-optic sensors. This study develops a revolutionary deep learning-based method for predicting the lifespan of solar panels in order to maintain efficiency and maximise their usefulness. At first, the temperature was and humidity (TH) data are collected from the solar photovoltaic panel using the fibre-optic sensor and Sensirion SHT15 sensor. The primary device of the solar PV panel lifetime detection system is Raspberry Pi which is used to store the data collected by different sensors. These data are transferred to cloud server using Raspberry Pi. Based on the gathered data the deep learning-based Bi-LSTM network is used to detect the panel lifetime using the threshold value. Furthermore, the GSM module will notify consumers if any alterations are made to the solar panel. The efficacy of the proposed model was assessed utilising the precise criteria, such as sensitivity, accuracy, and specificity. The proposed method’s accuracy of 95.2% is higher than that of conventional DL networks. By using certain measures like specificity, sensitivity, and accuracy, the suggested Bi-LSTM improves on classic CNN and LSTM by 1.78% and expands the accuracy rate range by 4.20%. The proposed method’s accuracy of 95.2% is higher than that of conventional DL networks. Compared to conventional CNN and LSTM, the suggested Bi-LSTM improves overall overall trend by 4.20% and 1.78%. respectively.

Article Details

Section
Special Issue - High-performance Computing Algorithms for Material Sciences