Structure and Module Design of Solder Joint Quality System Based on Neural Network A-890

Structure and Module Design of Solder Joint Quality System Based on Neural Network
The structure of the SMT solder joint quality analysis and evaluation expert system based on neural network is composed of the actual needs to solve the problem of SMT solder joint quality analysis and evaluation. The structure of the traditional expert system is expanded, and the learning mechanism of samples is introduced as the knowledge acquisition module of the system. . The system mainly includes interpretation interface module, neural network learning module, knowledge base module, database and its management module, inference engine module, control interface, etc.
(1) Interpretation interface module: Interpretation interface is a user interface, which is an interface for question input, result output and man-machine dialogue.
(2) Neural network learning module: Through the learning of samples, the required connection weight distribution is obtained to complete the acquisition of knowledge. The BP algorithm and other methods can be used to realize the transformation of implicit domain knowledge into the knowledge base.
(3) Inference engine module: Inference engine proposes a method of using knowledge to solve problems. Parallel reasoning can be realized through methods such as forward numerical calculation of neural network.
(4) Knowledge base module: The knowledge in the knowledge base is obtained by the learning system, which is the basis for the inference engine to complete inference and problem solving. The knowledge base can be continuously updated, which is represented by learning new samples based on the original knowledge, and obtaining new network parameter distributions that reflect more knowledge and experience.
(5) Database and its management module: On the one hand, the database is used for users to query related data, and on the other hand, it provides learning samples of neural network. The knowledge base of the expert system is updated by extracting learning samples from the database. In addition, the system should also provide an interface for reading library files, that is, a data reading module, to implement reading operations on database files. The database management module has complete database operation functions and can be used as an independent module.
(6) Other modules: The system also includes other sub-modules that are not shown, such as the standardized processing module for sample data.

Leave a Reply

Your email address will not be published. Required fields are marked *