Font Size: a A A

Method Research For System Identification And Control Of Hydraulic Cylinder Piston Based On Neural Network

Posted on:2007-09-25Degree:MasterType:Thesis
Country:ChinaCandidate:Z WangFull Text:PDF
GTID:2120360182484014Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The sea is one of the most important sources where we can get the resource, but for it is very difficult to exploit, so there are only a few ways such as building sea platform, among these ways, the platform is the one of the most important ways right now.The chief request is to be safe and stable .in order to build a safe platform, it must be tested completely from every aspect, so we produce a similar platform which has the same characteristic as the real platform but smaller than it, and then we use a piston to simulate the random ice curves to push the platform, the feedback of the platform will help us to find the defects of the platform's design and correct or optimize them.In the whole experiment system, the pushing process is the core, because only the simulation meets a high precision, the results is reliable to use. The control mode of the original piston control system is classical PID control system, which include a control box and control computer .Because its precision and stability can not be satisfying, hence the whole control system must be changed and updated.Based on the servo piston system of Changchun experimentation machine academe ,the paper narrates the necessity and the process of system identification .New system environment is dSpace which produced by the dSpace company in German .This system environment is based on MATLAB/Simulink ,there is no gap when you want to transfer between the MATLAB and dSpace .The transfer from VB to dSpace relies on system identification ,for the original system is unknown to us ,if we want to use the piston system correctly and efficiently ,we must identify the piston so that we can know the piston's specialty .The paper narrates the NN and system identification theory and the realization process of system identification in detail, and the results show that, the new system improves much more than the original system and the basic requirements are fulfilled.
Keywords/Search Tags:System Identification, Neural Network (NN), DSpace, MATLAB/Simulink
PDF Full Text Request
Related items