Control method, computer-readable recording medium storing control program, and air conditioning control device
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing air conditioning control systems face challenges in maintaining accurate temperature uniformity across different areas of a room due to changes in room layout, seasonal variations, and time of day, leading to user discomfort and high calculation costs for regular relearning of prediction models.
Innovation Solution
A control method that involves obtaining trained models for each area, detecting radiation temperatures, blowing air to predicted unevenness areas, and re-training models based on subsequent temperature readings to improve prediction accuracy and user comfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If regular relearning of prediction models is performed to maintain accurate temperature uniformity, then temperature control precision is improved, but calculation costs and processing time increase
Solution Approach 1:
The prediction model performs self-evaluation by calculating a confidence level that indicates its own reliability. When the confidence level falls below a threshold, the system triggers relearning automatically, eliminating the need for external monitoring and manual intervention. This self-service mechanism reduces processing overhead while maintaining temperature control precision.
Solution Approach 2:
The system implements a feedback loop where the confidence level output from the prediction model feeds back into the decision-making process for relearning. This feedback mechanism allows the system to dynamically adjust when relearning should occur based on actual model performance, optimizing the balance between temperature control accuracy and computational resource usage.
2Adaptability or versatility
If prediction models are frequently updated to adapt to room layout changes and seasonal variations, then adaptability is improved, but calculation costs increase
Solution Approach 1:
The prediction model autonomously evaluates its own confidence level to determine when it has become unreliable due to environmental changes. This self-service approach allows the system to adapt to room layout changes and seasonal variations only when necessary, rather than through frequent scheduled updates, thereby reducing calculation costs while maintaining adaptability.
Solution Approach 2:
Instead of continuous or frequent model updates, the system employs periodic relearning triggered by confidence level thresholds. This periodic action based on actual model degradation rather than fixed time intervals optimizes the balance between adaptability to environmental changes and computational energy consumption.
3Ease of manufacture
If a simple control system is used to reduce device complexity, then ease of manufacture is improved, but temperature uniformity control precision deteriorates
Solution Approach 1:
The system changes the parameter of model confidence evaluation from an implicit complex process to an explicit calculated value. By introducing a confidence level parameter that quantifies model reliability, the system maintains simple device architecture while achieving precise temperature uniformity control through intelligent parameter-based decision making.
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
This approach effectively suppresses temperature unevenness by proactive air conditioning control and reduces the need for continuous relearning, enhancing user comfort while minimizing calculation costs.
Implementation Method 1
a radiation sensor 80 that detects radiation temperature in each area
Implementation Method 2
blowing air to the area where temperature unevenness is predicted to occur
Data Source
AI summary
A control method for causing a computer to perform a process includes: obtaining a trained model for each area of a plurality of areas; detecting a first radiation temperature for each of the areas; blowing blowout air to an area where temperature unevenness is predicted among the plurality of areas on a basis of the detected first radiation temperature for each of the areas and the trained model; detecting a second radiation temperature for the area after the blowout air is blown; and executing re-training of the trained model on a basis of a label related to temperature unevenness created on a basis of the second radiation temperature and the second radiation temperature.