CPU Temperature Prediction Using Vector Distance Controller Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing temperature management systems using linear prediction models face challenges in accurately replicating the dynamic characteristics of CPU temperature across varying operating conditions, leading to degraded control performance and overshoot when the actual conditions deviate from the defined model.

Innovation Solution

The system divides the operating area into multiple ranges and assigns multiple controllers with different prediction models, selecting the controller with the closest matching conditions to the current operating conditions using a weighted vector distance calculation to predict future CPU temperatures and determine the manipulated variable for the cooling device.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a single linear prediction model is used for temperature management, then the device complexity is low, but the control precision degrades when operating conditions deviate from the model's defined conditions

Engineering Contradiction:

Improveprediction model structureVSAvoidtemperature control precision

Core Design Contradiction:

Device complexityVSManufacturing precision

Solution Approach 1:

The operating area is divided into multiple operating ranges based on power consumption and intake air temperature. Multiple linear prediction models are established, each corresponding to a specific operating range. The controller selects the appropriate model based on current operating conditions, thereby maintaining high temperature control precision across varying conditions without requiring a single complex nonlinear model.

Inventive Principle:

Principle #1Segmentation

2Manufacturing precision

If multiple prediction models are established for different operating ranges, then the temperature control precision is improved, but the device complexity increases

Engineering Contradiction:

Improvetemperature control precisionVSAvoidprediction model structure

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The system dynamically selects the appropriate prediction model based on current operating conditions (power consumption and intake air temperature). The controller calculates the vector distance between current conditions and model conditions to determine the most suitable model. This dynamic adaptation allows the system to maintain high precision without permanently increasing structural complexity.

Inventive Principle:

Principle #15Dynamics

3Reliability

If the prediction model conditions are set to match current operating conditions, then the control performance is optimized, but the adaptability to varying operating conditions deteriorates

Engineering Contradiction:

Improvecontrol performanceVSAvoidadaptability to operating conditions

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

Multiple prediction models are established, each with specific operating conditions (power consumption and intake air temperature). The controller can universally apply any of these models depending on the current operating range, making the system adaptable to various conditions while maintaining optimized control performance for each specific condition through the selected model.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10061365B2Temperature prediction system using prediction models, temperature prediction method using prediction models, and non-transitory computer readable recording medium having therein program for temperature prediction using prediction models

Publication Date: 2018.08.28 FUJITSU LTD

AI summary

A temperature management system includes a detection unit for detecting a temperature of a heat generator, power consumption, and an intake air temperature of the electronics device; a control unit for controlling a manipulated variable to be given to the cooling device so that the temperature of the heat generator becomes close to a target value, wherein the control unit includes controllers assigned respectively to operating ranges of the electronics device and each controller includes a prediction model for predicting a future temperature of the heat generator under conditions set for the corresponding operating range, and a vector distance between a first vector for a current state of the electronic device and a second vector for conditions included in the prediction model is calculated to select the controller that corresponds to the prediction model of the shortest vector distance.