Information processing device and program

The patent segments the complex CFD calculation into multiple simplified calculation stages. First, a rough air environment prediction is performed using simplified models, then progressively refined through intermediate calculations, and finally adjusted with detailed parameters. This multi-stage segmentation approach maintains prediction accuracy while significantly reducing total calculation time compared to performing a complete CFD analysis from scratch.

The patent performs preliminary air environment predictions using simplified models before conducting detailed CFD calculations. By pre-calculating baseline air flow patterns, temperature distributions, and humidity fields using faster computational methods, the system prepares initial conditions that accelerate subsequent detailed simulations, thereby reducing the overall time required for accurate prediction.

The patent employs pre-built standardized model templates that represent common air conditioning systems, building layouts, and environmental conditions. Users can quickly create customized models by copying these templates and modifying parameters such as device positions, room dimensions, and operational settings, rather than constructing models from scratch. This significantly improves ease of model creation while maintaining calculation speed.

The patent develops universal model components that can be applied across multiple different scenarios. Standardized elements such as generic air conditioner models, heater models, window models, and wall models are designed to work in various building types and configurations. These multi-functional components reduce the effort required to create new models, as users can reuse and adapt existing universal elements rather than developing custom models for each situation.

The patent applies local quality by incorporating detailed element placement information selectively in regions where it most impacts air environment prediction accuracy. For example, precise positioning data is prioritized for air conditioning devices, heat sources, and critical measurement zones, while less critical areas use simplified representations. This selective approach improves overall estimation accuracy without proportionally increasing model complexity across the entire system.

The patent implements dynamic model complexity adjustment based on prediction requirements. The system can adaptively simplify or detail model elements depending on the specific prediction task, available computational resources, and required accuracy level. This allows the model to maintain high accuracy when needed while reducing complexity for routine predictions, balancing both requirements effectively.