Anooshirvan Mostmand
Integrating BIM, BEM, and Deep Learning for Enhanced Energy Forecasting in Industrial Facilities: A Multi-Model Approach for Predictive Analytics and Real-Time Visualization.
Rel. Matteo Del Giudice, Enrico Macii, Anna Osello, Edoardo Patti. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Energetica E Nucleare, 2024
Abstract: |
Effective energy management in industrial facilities is crucial for achieving sustainability and operational efficiency. This thesis presents an advanced framework consisting of two complementary components: a predictive energy modeling system and the development of a real-time visualization application for facility management. The research is focused on enhancing energy forecasting capabilities and providing practical visualization tools to improve decision-making processes. The first component of the study involves the development of machine learning models to forecast energy consumption in the PCMA San Benigno facility located in Turin Italy, which is analyzed as a single energy-consuming entity. The models developed include a Fully Connected Neural Network (FCNN), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Bidirectional LSTM (Bi-LSTM). These models predict energy consumption based on simulation in Building Energy Modeling (BEM) software, capturing interactions between HVAC systems, environmental factors, and operational conditions. The integration of Building Information Modeling (BIM) and BEM is crucial in representing the dynamic interactions between internal building systems and external environmental conditions. This integration contributes to improving the accuracy of energy forecasts, supporting informed strategic decisions regarding facility operations. The second component focuses on the development of a real-time visualization application for multi-zone energy monitoring and decision support. This interactive tool enables facility managers to visualize energy consumption and other critical metrics across different operational zones of the facility. Unlike the forecasting models, which treat the facility as a holistic unit, the application provides detailed zone-specific insights, dividing the building into multiple zones to reflect individual energy performance metrics. Additionally, a Convolutional Neural Network (CNN) is integrated to forecast BEM features, offering real-time insights into zone-specific energy dynamics. This capability supports targeted optimization of HVAC operations and fosters efficient energy management throughout the facility. Together, these two components form a comprehensive energy management framework. By combining advanced predictive modeling with interactive visualization, the approach facilitates both strategic planning and operational decision-making. The findings highlight the significant potential of integrating data-driven forecasting with real-time visualization to achieve proactive and intelligent energy management in complex industrial environments. |
---|---|
Relatori: | Matteo Del Giudice, Enrico Macii, Anna Osello, Edoardo Patti |
Anno accademico: | 2024/25 |
Tipo di pubblicazione: | Elettronica |
Numero di pagine: | 105 |
Informazioni aggiuntive: | Tesi secretata. Fulltext non presente |
Soggetti: | |
Corso di laurea: | Corso di laurea magistrale in Ingegneria Energetica E Nucleare |
Classe di laurea: | Nuovo ordinamento > Laurea magistrale > LM-30 - INGEGNERIA ENERGETICA E NUCLEARE |
Aziende collaboratrici: | NON SPECIFICATO |
URI: | http://webthesis.biblio.polito.it/id/eprint/34373 |
Modifica (riservato agli operatori) |