ML Fan Speed Control for Quiet Computing Cooling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Computing devices face a challenge in balancing fan speed to effectively cool components while minimizing noise, which can distract users and interfere with audio outputs, especially when fan speed rapidly increases or decreases.

Innovation Solution

A computing device equipped with processing devices, temperature sensors, and a fan tachometer uses a machine learning model to generate a fan control signal based on performance data, including temperature and fan speed data, to dynamically control the fan speed, employing a thermal control PID algorithm and a fan control PID algorithm to maintain optimal cooling while reducing noise.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Temperature

If fan speed is increased to cool components effectively, then cooling performance is improved, but noise increases and causes user distraction

Engineering Contradiction:
Improvecomponent temperatureVSAvoidfan noise
Core Design Contradiction:
TemperatureVSObject-generated harmful factors

Solution Approach 1:

The patent implements dynamic fan speed control by transitioning from static to dynamic operation. The system continuously monitors temperature and adjusts fan speed in real-time, allowing the fan to operate at optimal speeds for current thermal conditions rather than running at constant high speed, thereby reducing noise while maintaining effective cooling when needed

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent employs feedback control mechanisms where temperature sensors continuously monitor component temperatures and feed this information back to the control system. The system also monitors fan noise levels and uses this feedback to adjust fan speed, creating a closed-loop control system that balances cooling performance with noise reduction based on actual operating conditions

Inventive Principle:
Principle #23Feedback

2Speed

If fan speed changes rapidly to respond to temperature changes, then cooling responsiveness is improved, but noise distraction increases due to sudden volume changes

Engineering Contradiction:
Improvecooling responsivenessVSAvoidsudden noise changes
Core Design Contradiction:
SpeedVSObject-generated harmful factors

Solution Approach 1:

The patent applies cushioning principles by implementing gradual fan speed transitions instead of abrupt changes. The system anticipates the need for speed changes and implements them progressively over time, cushioning the transition to prevent sudden noise spikes that would distract users while still responding effectively to temperature changes

Inventive Principle:
Principle #11Beforehand cushioning (Prior cushioning)

Solution Approach 2:

The patent implements periodic monitoring and adjustment of fan speed based on temperature thresholds and time-based smoothing. Rather than responding instantaneously to every temperature fluctuation, the system uses periodic sampling and time-averaged responses to smooth out rapid changes, maintaining cooling effectiveness while reducing abrupt noise variations

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS11907032B2Generating fan control signal for computing device
Publication Date: 2024.02.20 MICROSOFT TECHNOLOGY LICENSING LLC
  • US11907032B2 patent drawing
  • US11907032B2 patent drawing
  • US11907032B2 patent drawing

AI summary

A computing device is provided, including one or more processing devices, one or more temperature sensors, a fan, and a fan tachometer. The one or more processing devices may be configured to execute an application program. While executing the application program, the one or more processing devices may be further configured to collect performance data including temperature data received from the one or more temperature sensors and fan speed data received from the fan tachometer. The one or more processing devices may be further configured to generate a fan control signal at least in part by applying a machine learning model to the performance data. The one or more processing devices may be further configured to control the fan according to the fan control signal.