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.

This article addresses a new aspect of the problem of sensor network design, namely the concept of flexibility of sensor network. Algorithms based on graph theoretical concepts and MINLP methods are developed for analyzing the flexibility of a given sensor network, designing a flexible sensor network, and upgrading a sensor network to improve its flexibility. Several examples are reported to illustrate the presented algorithms. Using the proposed approach, one can obtain a flexible sensor network, which is able to ensure the observability of all the key variables even under some cases in which the original flowsheet changes.

Because industrial process is often under changes, it is necessary to reconcile the measurement of the dynamic process. There are many problems in the present dynamic data reconciliation methods, such as low calculation efficiency, difficulty of treatment for the reconciliation of input variables. Based on the analysis of the characteristic of dynamic data reconciliation, an integral approach is presented which integrates finite element collocation method, filtering technique and robust method. The finite element collocation method can reduce the amount of discrete model constrained equations to decrease the problem complexity without any loss of measurement information. The filtering technique can eliminate random errors in input variables effectively and no lag or signal distortion will be introduced. Monte Carlo method is used to test the efficiency of the robust method for gross error detection. The calculation results show that the integral approach can improve the calculation efficiency greatly- and can deal with the gross error of abnormal type properly.

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.