Apparatus for generating temperature prediction model and method for providing simulation environment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing temperature prediction models face challenges in accurately predicting indoor temperatures due to the complexity of variables involved, such as building materials, window number, wall thickness, season, date, and time, leading to deteriorated prediction accuracy and the difficulty in deriving optimal equations.
Innovation Solution
An apparatus that generates a temperature prediction model by setting and optimizing hyperparameters through a process of training and updating based on the difference between predicted and actual temperatures, allowing for the reflection of various variables and improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If equations are updated by expressing relationships between variables and temperature, then the model can reflect various building variables, but the complexity increases and prediction accuracy deteriorates when the number of variables increases
Solution Approach 1:
The patent replaces the traditional mechanical/mathematical equation-based temperature prediction system with an artificial neural network system. Instead of using complex analytical equations that become intractable with multiple variables, the invention uses a neural network that can automatically learn and represent complex non-linear relationships between building variables and temperature without requiring explicit mathematical formulation. This substitution resolves the contradiction by maintaining adaptability to multiple variables while avoiding the complexity and inaccuracy of traditional equation-based approaches.
2Manufacturing precision
If humans check the difference between actual and simulated temperature data to update equations, then the model can be adjusted, but it is difficult to derive the optimal equation due to reliance on human intuition
Solution Approach 1:
The patent implements self-service by enabling the neural network model to automatically update its own parameters through backpropagation and gradient descent algorithms. Instead of relying on human experts to manually adjust equations based on intuition, the system autonomously learns from the difference between actual and simulated temperature data, iteratively optimizing its internal parameters to minimize prediction errors. This resolves the contradiction by achieving high prediction accuracy through automated optimization rather than manual adjustment.
Solution Approach 2:
The patent incorporates feedback mechanisms where the difference between actual temperature measurements and simulated temperature predictions is continuously fed back to the neural network. This feedback drives the automatic adjustment of model parameters through gradient-based optimization algorithms, enabling the system to learn from errors and progressively improve prediction accuracy. The feedback loop replaces human intuition-based adjustments with systematic, data-driven optimization.
3Device complexity
If traditional equation-based methods are used to predict temperature, then the model structure is simple, but the prediction accuracy deteriorates due to the complexity of multiple variables
Solution Approach 1:
The patent fundamentally changes the parameters and structure of the prediction model by transitioning from traditional physics-based equations to a data-driven neural network architecture. This parameter change enables the model to capture complex non-linear relationships between multiple building variables and temperature that cannot be adequately represented by simple analytical equations. The neural network's ability to learn complex patterns from data resolves the contradiction by prioritizing prediction accuracy over structural simplicity.
Data Source
AI summary
An apparatus for generating a temperature prediction model is disclosed. The apparatus for generating a temperature prediction model includes the temperature prediction model configured to provide a simulation environment, and a processor configured to set a hyperparameter of the temperature prediction model, train the temperature prediction model, in which the hyperparameter is set, so that the temperature prediction model, in which the hyperparameter is set, outputs a predicted temperature, update the hyperparameter on the basis of a difference between the predicted temperature, which is outputted from the trained temperature prediction model, and an actual temperature, and repeat the setting of the hyperparameter, the training of the temperature prediction model, and the updating of the hyperparameter on the basis of the difference between the predicted temperature and the actual temperature by a predetermined number of times or more to set a final hyperparameter of the temperature prediction model.