Partager

Techniques d'intelligence artificielle pour des systèmes CVC économes en énergie (en anglais)

Les avantages de l’intelligence artificielle (IA) pour les systèmes de chauffage, de ventilation et de conditionnement d'air (CVC) ont été l’un des sujets brûlants de la Conférence internationale sur les technologies émergentes pour les systèmes CVCR durables et intelligents, coparrainée par l’IIF et qui s'est tenue en Inde, à Kolkata, les 27 et 28 juillet 2018.

The IIR cosponsored a conference in Kolkata (India) on “emerging technologies for sustainable and intelligent HVACR systems”.It was held on July 27-28, 2018. Issues related to energy, new refrigerants and smart buildings in India more particularly were discussed.


The benefits of Artificial Intelligence (AI) for Heating, Ventilation, and Air-Conditioning (HVAC) systems was one of the hot topics of the conference.


Artificial Intelligence (Artificial Neural Networks, Machine Learning Algorithms, Genetic Algorithms, Fuzzy Systems, etc.) has proved to be quite useful in areas such as robotics, pattern recognition, forecasting, medicine, power systems, manufacturing, optimization and signal processing.

  • According to N. Dutta and T. Das1, significant advances have been made in the past decades on the application of AI techniques for HVAC systems design, control, management, optimization, and fault detection and diagnosis.

    During the conference, they presented a multi-layered Artificial Neural Network (ANN) model to estimate the heating and cooling loads of buildings for efficient HVAC system design. Errors reported in the model are well within the acceptable limits.

  • According to P. Adhikary et al2, current building climate control systems often rely on pre-determined maximum occupancy numbers, coupled with temperature sensor data to regulate HVAC. However, rooms and zones in a building are not always fully occupied. Real-time occupancy information can potentially be used to reduce energy consumption. They propose an ANN-based occupancy detection solution to address the need for real-time in-building occupancy information. The proposed solution can track real-time location of tagged occupants. Based on the findings, they present operable strategies for optimizing HVAC-related building energy consumption by using occupancy information.

The authors consider that ANN models may be used as an alternative method in engineering analysis and predictions. ANNs mimic somewhat the learning process of a human brain. They operate as a "black box'' model, requiring no detailed information about the system. Instead, they learn the relationship between the input parameters and the controlled and uncontrolled variables by studying previously recorded data, like the way a nonlinear regression might perform.

The presented ANN modeling can predict the HVAC occupancy load of a building with acceptable accuracy. The authors stress that it is certainly more economical to be able to investigate the behavior of energy consuming systems without having to construct and experiment on several systems or use expensive models. The application of ANNs has shown that it is possible to model such systems with a minimum amount of input data, thereby providing the designer of such systems with the flexibility to test several systems quickly. The significant reduction in estimation times is the major benefit of the present method.


*Papers from this conference are available in Fridoc database : à compléter une fois la fiche-mère en ligne


1 DUTTA N., DAST., Artificial Intelligence techniques for energy-efficient H.V.A.C. system design. Available in Fridoc.

2 ADHIKARY P., BANDYOPADHYAY, MAZUMDAR A., ANN-based occupancy detection solution for energy efficient HVAC control: a case study. Available in Fridoc.