Industrial Process Data Correction Method

Data rectification is a key technology of process industry integrated automation, utilizing the redundancy of measured data and mechanism of the production process; it filters the gross errors of measured data by statistical methods. To improve the quality of the measurements and estimate unmeasured variables, the measured data is adjusted systematically, and the potential uncertainty of the model is amended, according to the optimization theory. For different models, the corresponding data rectification algorithms are also different. So, it is urgent to study efficient data rectification algorithms for different process models to meet the needs of industrial applications. In this thesis, data rectification algorithms of linear dynamic and nonlinear dynamic process model were researched, the main works are listed as follows:(1) To solve the problem of large-scale computation caused by system dimension. Large-scale system decomposition coordination algorithm was introduced to reduce the system's dimension and variables classify. And the simulation of dynamic systems shows that this approach is effective.(2) Standard deviation or covariance matrix is supposed to be known in traditional algorithms of nonlinear dynamic data reconciliation, which is out of accord with practical application. An improved algorithm was proposed, standard deviation was estimated and updated though moving horizon window approach.(3) Step input signal can't be recognized efficiently in traditional algorithms of gross error detection; it results in miscarriage of justice and lag around chopped spot. An improved algorithm based on 3σapproach and step judgments logic was proposed. The simulation shows that the algorithm is efficient.