Study On Intelligent Design Of Automotive Radiator Based On Hybrid Case-based Reasoning

As an important part of the cooling system, automotive radiator has a great influence on the engine’s power, stability, durability and economy. When the radiator is designed, we should ensure that the engine works in the optimum temperature. And prevent from appearing too cold or overheating. The radiator design is more compact, more effective and lighter. This article is to study on the more effective design of radiator. Using CBR and improved Genetic Algorithms in radiator design, it can manage the knowledge and experience of radiator design more effectively; can use the enterprise production ability of existing equipments more fully; and can make the development cost much lower and the development cycle much shorter. In brief, using CBR and improved Genetic Algorithms in radiator design has very importantly theoretical and practical engineering application value for radiator design.By analyzing the theory of radiator design and the improved Genetic Algorithms, we use jf factor as the objective function to measure the stand or fall of radiator. In Visual Studio2008, we develop a platform of automotive radiator optimized design. We use jf factor as the objective function to measure the stand or fall of radiator. The structure of radiator could be optimized by the improved Genetic Algorithms. And optimization results were entered into the database in order to be inquired conveniently by designer, and displayed or printed out by crystal.Case-based Reasoning is a kind of intelligent design methods which solves the current design problem based on past examples or experience. CBR includes case-representation, case-retrieval, case-adaptation and case-maintenance four parts. It has a lot of advantages such as simple knowledge-acquisition, convenient system-maintenance, reasonable solving-logic, high solving-efficiency, and self-learning ability, etc. In CBR technology, case retrieval is the most key technology, and it also is the core to reflect the intellectualization of design. Case retrieval is a case search process whose purpose is to match case from case database, and find out the same or similar cases of solving problem. This paper uses the fuzzy membership function, nearest neighbor (K-NN) method and the concept hierarchy tree, and determines a hybrid intelligent design retrieval technology which mixed fuzzy theory, nearest neighbor method and concept hierarchy tree. Combining CBR based on hybrid retrieval technology with the radiator design knowledge. Based on Visual Studio2008and SQL Server2005, the intelligent design system of automotive radiator is built up. This system makes the radiator parameters as its case databases, and weights feature weight by hybrid intelligent design retrieval technology. It can quickly retrieve the satisfaction of similar instances by Hybrid Reasoning matching.