Seeking AI-Powered Digital Solutions for Air Conditioning Energy Savings

Air conditioning (AC) is a major energy consumer in both residential and commercial buildings, accounting for nearly 10% of global electricity use and significantly contributing to peak power demand in many regions. With rising concerns over climate change, carbon emissions, and escalating energy costs, governments and industries are actively seeking advanced solutions to enhance energy efficiency without compromising indoor comfort.

Despite advancements in inverter technology and refrigerant efficiency, many AC systems still rely on traditional Proportional-Integral-Derivative (PID) control logic. The PID adjusts compressor speed, expansion valve opening, and fan operation solely based on temperature differences, often causing overcorrection and fluctuations between overcooling and overheating.

Some of key challenges to be addressed include:

• Static Operating Parameters: Lack of coordination between room conditions, user preferences, and external climate factors.
• Inefficient Heat Load Management: Lack of predictive forecasting and real-time responsiveness heat load variations.
• Over-Reliance on PID Control: Adjustments based solely on instantaneous temperature differences rather than learning-based approaches.
• Limited Cloud-Terminal Coordination: Lack of seamless integration between unit-level operations and centralized cloud analytics.

Haier is seeking an AI-powered digital solution that dynamically adjusts AC operating parameters based on real-time heat load and user comfort requirements through intelligent cloud-terminal coordination.

The proposed solution should integrate predictive, AI-powered control and cloud-terminal coordination to reduce power consumption while maintaining thermal comfort, targeting a comprehensive energy-saving effect of over 20% for both cooling and heating. Key requirements include:

• Data-Driven Intelligent Control:
  • Move beyond traditional PID control by integrating predictive, AI-driven algorithms that dynamically adjust parameters.
  • Incorporate real-time heat load estimation, user comfort and behavior learning, and external climate conditions for optimal adjustments.
• Cloud-Terminal Collaboration:
  • Adaptive control by coordinating cloud-based intelligence with real-time terminal-side optimization.
  • Support both cloud-connected operation and offline control.
• Hardware & Integration:
  • Incorporate a miniaturized terminal-side module that can be seamlessly integrated into AC units.
  • Equipped with an AI chip (with an MCU operating frequency preferably below 300MHz and hardware acceleration computing power under 0.5 TOPS)
  • Ensure an effective service life of at least 10 years.

For further details, please refer to the attached "Proposal Template".

• Haier is open to receive proposals from SMEs, start-ups, and academic institutions.
• Candidates with practical experience in implementing reinforcement learning algorithms are prioritized. The algorithm should demonstrate environmental adaptability, with capabilities for self-iteration and self-learning.
• Proposals from higher education institutions (IHLs) should demonstrate a Technology Readiness Level (TRL) of 7 or above.
• Solution should be open for global market adoption.
• With an estimated market potential of 5 million units per year, successful proposals could lead to discussions on scaling up manufacturing depending on collaborator technology status.

The product or solution development is expected to take between 1 to 3 years, depending on the TRL, application suitability, and collaboration viability.

A three-phase approach will be adopted for this challenge:

• Phase 1: Proof of Concept (POC) development
• Phase 2: Demonstration of working prototype
• Phase 3: Full functional product development

Note: A separate business collaboration contract may be established between Haier and the selected solution provider, contingent upon Haier HQ’s evaluation of the proposal and its determination that the solution is suitable for the next phase of development. The contract will subsequently proceed according to its own timeline.

Resources to be Provided:

• In-Kind Support: Access to relevant data, testing and pilot sites, existing product components, and mentorship from Haier product engineers.
• Monetary: Funding to support co-development or trial costs may be provided, subject to Haier HQ’s evaluation of the proposal and determination of its feasibility for the next phase of development. Upon meeting these criteria, a separate project contract will be signed with Haier HQ during the product development phase separately.