Font Size: a A A

Research And Development On The Quality Management System Of The Rolled Steel Section MES

Posted on:2014-01-04Degree:MasterType:Thesis
Country:ChinaCandidate:Y LuFull Text:PDF
GTID:2251330422466034Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Due to the characteristic of multistage mixing production with the characteristic ofcoexistence of disperse and continuation from the Iron and steel industry, together with thecomplicated production processes and the diversified information the productive processincludes the physical changes and chemical changes. We need not only to consider about thelimit from the process and the resource in the production process but also to consider about thetemperature, energy consumption, size precision, composition deviation, the quality of theproduct and so on. This paper raised the management system which faced to the workshops inthe factory---Manufacturing Execution System (is short for MES).This paper designed and realized a MES quality control system which was to controlthe quality of steel products. This system which combined the RBF neural network expert systemand the quality trace part is called hot rolling product quality test system. This paper improvedthe RBF neural network on its kernel function and the set of the hidden nodes. And it makes theimproved RBF neural network to combine with the chemical properties test, the physicalproperties test, dimensional accuracy test and the surface of the products test. It used RBF neuralnetwork expert system to accomplish the test of the four items to verify whether the product waseligible. At the same time, it assumed a forecast which could get the physical properties resultfrom the calculation of chemical properties. It realized the assumption by RBF prediction neuralnetwork. This paper combined the prediction-test neural network and made it the kernel part ofthe quality test expert system.At the same time, due to the―black box‖problem from the explanation facility in theconventional neural network expert system and the need from the work of steel industry qualitytest this system made quality trace to the test result after the statistic analysis of the quality data.According to the characteristic of the production processes I designed the tree-form‘back ruleswhich is on basis of coding. It could trace all the information during the life cycle of the product.The trace was divided into lengthways tracing and lateral tracing. The aim of lengthways tracingis to lock the problematical link and to trace the responsible process. The aim of lateral tracing isto obtain the procedure parameter from the life cycle of the product.
Keywords/Search Tags:manufacturing execution system, radial basis function neural network, multiquadratic, quality management, quality trace, expert system
PDF Full Text Request
Related items