Simulation and reasoning: intelligent building thermal problem detection

This paper discusses a model for integrating a rigorous thermal simulation with computational reasoning. This model is used to build an intelligent computer-aided system that assists designers throughout the design process. The model uses rigorous hierarchical thermal simulation modules linked to several databases. In addition, Artificial Intelligence Techniques are used to build a multilevel reasoníng structure for both the initial and the final design. Neural networks are utilized for the initial design stage problem detection where problems are incomplete. The neural networks are integrated with a multi-knowledge reasoning structure that utilizes the blackboard framework. This structure is used for problem detection in the intermediate to final design stages and for advice throughout the design process.