International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
Predicting Solar Energy Consumption in a Commercial building using Machine Learning Divya Rani1, Harsh Kumar2, Kunal Singh3 , Ajay Solanki4 1, 2 , 3 Undergraduate students , 4 Assistant Professor
Department of Mechanical Engineering , Delhi Technological University , New Delhi , India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This research was aimed at investigating
weather data, forecast weather data, and building-specific datasets .
different modeling approaches to predict solar power consumption in the building industry in India. LSTM neural networks, and linear regression , two machine learning models were used for evaluating efficiency. Namely, to create models, the analysis of historical records, weather forecasts, and building-specific datasets was performed. The demographic of the region, location, calendar factors, and historical consumption trends were extracted to feature consumption. LSTM networks can well catch complex temporality patterns while having low interpretability, while the linear regression model can assume causation but have reduced accuracy. As shown, the LSTM model was more successful over fitting data with 97.12% accuracy. However, the study shows that both approaches are possible and provide valuable insights into India urban expansion and sustainable energy solutions. The research brings to light the importance of using advanced predictive modeling solutions to address the energy problem and a greener environment in India.
Thus, this study aims to explain the applicability and productivity of LSTM neural networks and linear regression for prescriptive energy following India's urban development. Research presents considerable conclusions for usage of energy forecasting methods creatively tailored for the building industry in India, enabling policymakers and stakeholders to address energy-related concerns and support ecological sustainability..
2. LITERATURE REVIEW "Predicting Energy Consumption in Residential Buildings Using Advanced Machine Learning Algorithms" by Dinmohammadi, F., Han, Y., & Shafiee (2023)[1] The article provides a review of predicting building energy consumption with a focus on machine learning methods. Several machine learning algorithms that are currently widely used in predicting building energy consumption and performance, such as artificial neural networks, support vector machines and Gaussian-based regressions, and clustering, are introduced in this paper. The review underscores the need for accurate energy prediction for efficient decision making on energy consumption and intelligent renovation of buildings. The study recommends a framework to select the best machine learning method for various cases and building usage. Finally, the conclusion of the review argues the current hurdles associated with machine learning on energy prediction and their limitations and possible research avenues for improvement using machine learning in predicting and benchmarking energy.
Key Words: Solar Energy , Energy Consumption , Commercial building , Machine learning , Predictive Modeling
1.INTRODUCTION India’s building industry faces major challenges in regulating energy use and reducing environmental impacts due to rapid urbanization and population growth. The sector is characterized by a high carbon footprint and requires immediate measures to reduce these negative impacts by increasing energy efficiency and use of renewable energy sources. One of the possible ways to tackle these problems is accurate estimation of solar power usage using predictive modeling approaches. This study examines linear regression and long-short-term memory neural network effectives in predictive modeling approaches of solar power consumption in the Indian building sector. It is necessary to examine the extracted key features, such as the demography and geography, and calendar, and the history of the building’s energy consumption, which helps to understand the time and environmental patterns of energy use. To do this, the following is used: primary measurement data, historical
© 2024, IRJET
|
Impact Factor value: 8.226
"Energy consumption prediction by using machine learning for smart building: Case study in Malaysia" by Mel Keytingan , M. Shapi et al. (2021)[2] This study explored the use of machine learning for the development of a predictive model for energy consumption in smart buildings. The model training, data analysis, and model performance evaluation all took place in the Microsoft Azure Machine Learning Studio. The paper compared models based on three predictive techniques: k-nearest neighbor, support vector machine,
|
ISO 9001:2008 Certified Journal
|
Page 2713