Integrated solar PV and piezoelectric based torched fly ash tiles for smart building applications with machine learning forecasting

Abstract

Solarphotovoltaic (PV) power generation projects are rapidly increasing since the necessity for electricity has increased in developing countries like India. Still, the complexity of technical requirements narrows the financial margins of the construction industry. Limited researchers have investigated PV manufacturing, planning and techniques for electricity generation. Additionally a new method of generating power via the use of a human population has been developed. A piezoelectric transducer is used to detect the vibration. When a footstep applies pressure to the piezoelectric transducer, the pressure or force is converted to electrical energy. This experimental research investigates the photovoltaic sensors and piezoelectric transducers incorporated floor tiles in a residential building for power generation. The performance of the sensors were analysed experimentally and analytically by using machine learning models such as ARIMA time series model, Support Vector Machine (SVM) and k-nearest neighbour (KNN). The prediction models were validated through assessment matrix such as Root Mean Square Error (RMSE). From the experiments, it was observed that voltage is high during the day when a load is applied, and a low amount is acquired at night when no load is applied. And from the analytical study, the ARIMA time series model has the highest prediction accuracy. Error analysis was done to demonstrate the efficiency of the proposed models. The accuracy of the developed models can be further increased, which is subject to future research.