Study Of Intelligent Grading System For Grain Purchasing Based On Fuzzy Neural Network

Grain is the substance basic for man to live,It play a key role in countryeconomy and other industries. Now, When grain departments purchase the grain,there are differences between the quality examinations for kinds of grain. But theyall include moisture content and unit weight. That is, Moisture content and unitweight are the fixed items of quality examination. Therefore, this paper will studythe Intelligent Grading System for Grain from these two sides.This system unites on-line moisture content and unit weight examination,which is convenient, has a reasonable configuration and has changed the traditionalgrading measures through manual examining and computation. The system cansatisfy the need of purchase, make sure a highly exact precision. Therefore, thestudy has good economic and social profits and has extending and applying worth.Main research work is as follows:1. According to the request of grain grading by moisture content and unit weight, the paper has designed the Intelligent Grading System for Grain. Which includes feeding equipment of electromagnetic libration, weight sensor, digital temperature, capacitive moisture sensor, approaching switch and signal exchanged equipment so on. The hardware electrocircuit is designed by the method of functional modules, which mainly includes importing and adjusting module, RC/F and V/F exchanging module, CPU, key-press display, serialcommunication and electrical source module. As to every functional module, we have done the particular designing analysis;2. The paper has improved the structure of moisture sensor and designed a forked and single piece plane capacitance sensor. Moreover, it also analyses the reasonability and feasibility;3. Basing on the simulative instruments software Labview, the paper exploits the work software of the system. Which includes parameter design, data collection, data processing, data management and grading so on. This software has a friendly user interface, can display the test process on real-time and modify the controlling parameter on-line. Therefore, it can predigest the test process and data processing;4. On the base of analyzing the questions that lie in the present grain grading and contrasting kinds of data processing methods, the paper applies fuzzy theory and neural network and builds a simulation model. This model takes the outputs of the fuzzy system as the imports of the BP network. In the paper, first we use 100 training samples to training the model, and then test it by 32 testing samples. Through the test it proves that the modeling is impersonal, accurate and credible;5. After finishing the software and hardware design of the system and the debugging of the control electrocircuit, we have applied it in practice. Through the practical application, we further prove the precision of the system. The result tells us that the has powerful abilities of non-linear identifiability, knowledge representation and fault-tolerance. Therefore, it makes the system has enough precision and reliability to realize the testing of moisture content and unit weight and intelligent grading. The paper suggests and realizes the conception of applying the FNN in the...