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.

A knowledge-based system for aiding in the construction of practical heat exchanger networks called SPHEN is presented in this article. The created network configurations feature low utility cost, low capital investment, rather stabie heat recovery during operation and rational distribution of pressure drops among the network branches. SPHEN consists of four component parts: network construction, simulator, flowsbeet diagram generation and user's feedback subsystem. So SPHEN is characterized by its hybridity, reflected in knowledge representation and abstraction, solving methods and implementation languages. The hybrid system is implemented through disk files and data files connecting different parts into one working space. An example of redesigning an energy recovery network of a large-scale crude oil unit is illustrated and the results obtained are successful.

Dealing with multidimensional problems has been the "bottle-neck" for implementing wavenets to process systems engineering. To tackle this problem, a novel multidimensional wavelet (MW) is presented with its rigorously proven approximation theorems. Taking the new wavelet function as the activation function in its hidden units, a new type of wavenet called multidimensional non-orthogonal non-product wavelet-sigmoid basis function neural network (WSBFN) model is proposed for dynamic fault diagnosis. Based on the heuristic learning rules presented by authors, a new set of heuristic learning rules is presented for determining the topology of WSBFNs. The application of the proposed WSBFN is illustrated in detail with a dynamic hydrocracking process.

In this paper, Wavelet Transform is applied to improve the on-line prediction of crude oil distillation (COD) qualities. First, the wavelets-based multi-scale analysis method was applied to treat the process data effectively. According to their unique characteristics with multi-scale Wavelet Transform, different variations are detected and modified, such as steps, peaks, noises, abnormal sudden changes, and so on. Using the process data with noises discarded and abnormal sudden changes treated effectively, the COD quality prediction can be sure of safety and high accuracy. Next, for some COD units on-line quality analyzers are available, and a new correcting strategy is taken in this paper to give effective update. The update is accomplished based on process data trends extraction. At last, the data processing and the on-line update strategy are introduced to a prediction system of COD product qualities to improve its prediction accuracy. The improved prediction system has been applied in a real plant, and the results are satisfactory.

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.