A strategy of developing on-line optimization intelligent systems based on combining flo wsheeting simulation and optimization package with artificial neural networks (ANN) is presented in this paper. A number of optimization cases for a certain chemical plant are obtained off-line by using PROCESS-Ⅱ or other flowsheeting programming with optimization. Then, taking these cases as training examples, we establish a neural network systems which can be used on-line as an optimizer to obtain setpoints from input data sampled from distributed control system through gross errordetecton and datareconciliation procedures./Such an on-line optimizer possesses two advantages over nonlinear programming package: first of all, there is no convergence problem for the trained ANN to be used online; secondly, the frequency for setpoints updating is not limited because only algebraic calculation rather than optimization is required to be carried out on-line. Here two key problems of implementing ANN approaches to the on-line optimization are discussed: how to improve the prediction accuracy of ANNs models for meeting the optimization requirements. Results from an actual fractionation unit of a FCC plant in a refinery showed a 0.5%-1.0% increase in the total recovery of light oil products. Details of the strategy used are described.

In two related papers we define the concept of structural flexlbility and apply it specifically to a class of heat integrated distillation processes. A process ts structurally feasible if the type of equipment it contains and how it is currently configured allow it to meet specified design requirements. It is structurally infeasible if it does not. Size of equipment is not at issue. The first paper develops a simple rnethod to permit an estimate to be made of a structural flexibility index for a heat integrated distillation process with a fixed configuration. Multieffect columns are permitted. This work shows, that processes that differ only in their heat integration can have very different indices for structural flexibility. The second paper presents a synthesis metfiodotogy to discover the heat integrated arrangements which maximize the structural flexibility index for the process, Only the simpler arrangements of such columns are considered where they are placed one above the other and without multieffecting. The papers present algorithms by example. Insights are stressed.

In petrochemical processes, it is necessary to ensure the reliability and validity of process data for admimstrative management and many technical applications. These include process control and online srmulation and optimization. Steady-state onlme data reconciliation and new methods for data classifying, steady-state identification and gross error handling were adopted to develop an online data reconciliation system (DRS). The DRS has bee successfully implemented with the online optimization system in the crude oil distillation unit at Lanzhou petroleum refinery in China Using reliable and valid data achieved through online data reconciliation, the online optimization results showed a profit increase of more than 4 million RMB.

A wavelet-based method is proposed for steady-state detection in continuous processes. In this method, process trends are extracted from the measured raw data via wavelet-based multi-scale processing. The process status is then measured using an index with value ranging from 0 to 1 according to the wavelet transform modulus of the extracted process signal. Finally, a steady state is identified if the computed index is small (close to zero). The determination of a characteristic scale for performing steady-state detection was also studied. Compared with the existing approaches for steady-state detection, this method has better precision for detecting changes in process due to the good localization property of wavelet transform, and is more suitable for on-line applications. In this paper, the method is described in detail, and has then been applied to the crude oil unit of a refinery, and to the recausticizing plant of a chemical pulp mill.

Dynamic data reconciliation can supply more accurate data for dynamic optimization, dynamic fault diagnosis, and control by means of incorporating process information in some mathematical model. It will be an ill-posed inverse problem if the sensitive input variables are unmeasured; here, the sensitive input variable is defined as the variable that, if it is unmeasured, can only be estimated through the differentiation of other measured variables. In such a case, existing methods cannot obtain correct and usable data effectively. To address the problem, based on the principle of regularization, the wavelets are adopted to construct regular operators. And, a new approach is proposed to determine the optimal scale level corresponding to the optimal approximate operator in which the prior statistical information of the signal is utilized. The algorithm can deal with the estimation of unknown sensitive input variable effectively. The results show that more accurate estimation of the sensitive input variable can be obtained by using the proposed method as compared with the one obtained by using existing collocation methods based on polynomials.