Predictive Fan Control for Data Center Powertrain Cooling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current thermal control systems for powertrain components in data centers fail to anticipate or predict temperature swings due to high-frequency power cycling, leading to excessive heat generation and reduced component life.
Innovation Solution
Implementing a thermal management system that uses artificial intelligence and machine learning models to predict power load fluctuations, allowing for dynamic adjustment of cooling subsystems such as fan speed and liquid circulation rates to mitigate temperature swings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If cooling measures are activated by preset thresholds, then the cooling system can be controlled based on temperature levels, but the system cannot anticipate or predict immediate temperature increases during high-power cycling periods, resulting in large temperature swings
Solution Approach 1:
The AI/ML model performs preliminary action by predicting future temperature changes based on power load patterns before the actual temperature swings occur. The system analyzes historical power consumption data and load characteristics to anticipate thermal events, allowing the cooling system to be proactively adjusted rather than reactively responding to threshold breaches.
Solution Approach 2:
The system implements feedback by continuously monitoring power load data and using AI/ML models to predict temperature trends. The predicted temperature information feeds back into the cooling control decision-making process, enabling dynamic adjustment of cooling measures based on anticipated thermal conditions rather than static threshold-based control.
2Temperature
If fans or cooling measures are increased in speed during high-power periods, then excessive heat can be mitigated, but the cooling system cannot immediately decrease when power drops, causing excessive cooling and energy waste during low-activity periods
Solution Approach 1:
The AI/ML model predicts upcoming low-power periods by analyzing power load patterns and trends. By anticipating the timing and duration of power reductions, the system can proactively reduce cooling measures before the power drop occurs, preventing excessive cooling and associated energy waste during transition periods.
Solution Approach 2:
The cooling system transitions from static threshold-based control to dynamic predictive control. The cooling measures are continuously adjusted based on real-time power load data and AI-predicted temperature trends, allowing the system to adapt cooling intensity to match actual thermal demands during both high-power and low-power periods.
3Device complexity
If preset temperature thresholds are used to control cooling, then the control logic is simple, but the system reacts too slowly to high-frequency power cycling, causing large temperature variations that affect component life
Solution Approach 1:
The system replaces simple mechanical threshold-based control logic with an AI/ML-based predictive control system. Instead of using fixed temperature thresholds that require complex hysteresis and deadband logic to prevent oscillation, the AI model directly predicts temperature outcomes and optimizes cooling control decisions, simplifying the overall control architecture while improving performance.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Reduces temperature fluctuations and extends the life of powertrain components by anticipating power usage changes, thereby improving system reliability.
Implementation Method 1
Fans, or other cooling measures of a cooling sub-system, such as liquid cooling sub-systems are then turned on or increased in speed to mitigate the excessive heat
Data Source
Figure 1A
Figure 1A
Figure 1B
AI summary
A power component includes at least one device processor configured to: obtain power output data, wherein the power output data comprises: a power characteristic associated with a high frequency component of a load; and a power characteristic associated with a low frequency component of the load. A power component may obtain a trained thermal management artificial intelligence (AI) and/or machine learning (ML) model. A power component may be based at least on the power output data and the trained thermal management AI and/or ML model, infer a cooling sub-system setting. A power component may set the cooling sub-system in accordance with the cooling sub-system setting.