How to Reduce Energy Consumption of HVAC Systems Using AI-based Control Systems

Introduction

• Presenter: Conserve It and hosted by Backs Asia
• Speaker: Chirayu Shah, VP of Operations at Conserve It
• Topic: Using smart machine learning AI-based chiller plant control and optimization systems to reduce energy consumption in HVAC systems
• Audience Interaction: Audience poll regarding kilowatt autonomous efficiency for plant rooms

Chirayu Shah's Background

• Experience: 19 years in the automation and controls industry
• Expertise: Technical product development, business development in smart buildings, IoT solutions, building automation, and energy efficiency

Key Points

Background and Challenges

• Buildings account for almost a quarter of global carbon emissions
• The UN Paris Agreement targets to reduce carbon emissions
• HVAC systems consume almost half of a building's total energy
• Common challenges include inefficient HVAC operation, system downtimes, lack of automation, and introduction of renewable energy sources

Solution Approach Using AI

• Utilizes a centralized chiller plant and AI-based optimization system (Plant Pro)
• AI is used to generate real-time control algorithms and predictive modeling
• Machine learning techniques help in creating highly accurate predictive models

Plant Pro System

• Components: Hardware and software for chiller plant optimization
• Capabilities: Integrates with BMS, performs continuous commissioning, generates automated alerts, and uses AI for control algorithm generation
• Security: Operates within the customer's secure network without sending data off-site

AI and Machine Learning Techniques

• Techniques: Multivariate polynomial regression, large-scale non-linear programming solvers, and smart sequencing
• Functions: Predictive control, smart sequencing (optimum staging, sequencing, and load balancing), and smart flow optimization
• Example: Real-time adjustments to achieve optimal efficiency

Case Study in Southeast Asia

• Setup: 3 water-cooled chillers, 4 primary and condenser pumps, unique condenser side system with radiators and VSD fans
• Results: 18.8% energy savings, reduced manpower hours leading to significant cost savings
• Feedback: Improved efficiency without compromising on performance or comfort

Q&A Session Highlights

• Integration with BMS: Simple Ethernet or serial connection, no need to replace existing systems
• Baseline Derivation: Uses independent energy meter data, IPMVP methodology
• Challenges with OEMs: None, due to open technology and proven interoperability
• Prediction in Control: Predictive models help in making well-informed decisions to ensure optimal system performance
• Customer Concerns: System respects operational constraints while optimizing performance

Conclusion

• Encouragement to explore and interact with Conserve It and other exhibitors at Backs Asia
• Assurance of follow-up for unanswered questions
• Emphasis on the importance of AI and machine learning in improving HVAC system efficiency