Real-time Chemical Process Data Compression Algorithm And Its Application Based On The Wavelet Transformation

The developments in the computer and the sensor technologies have explosively increased the quantity of process data measured in the chemical engineering production process. Abundant data need to be stored because they bring a lot of useful information that will benefit many sorts of the running and controlling tasks. However, these data inevitably carry much redundant and irrelevant information resulted from the measuring and the computing errors and even out of some human factors. Moreover, with the development of the network technique, the remote and network-based monitoring and controlling strategies become popular in the industry production process. And they both require large amount of the process data transmitted in real time via the network in order to share the data resources. Thus it is necessary to develop the real-time process data compression technique to reserve the major characteristic information of the process, to save the data storing space, as well as to reduce the data stream and to improve the speed and the efficiency of the data transmission. Accordingly, it is necessary to develop the data compression technique which would serve the mill-wide automation system and improve the performance concerning precision and real time in automatic technology.There are three main methods for process data compression up to now, i.e. piecewise linear compression, vector quantizer compression and transform compression. Among them, wavelet transformation is regarded as the most promising process data compression method for its many good characters, especially those in the data processing area.This thesis applies wavelet transformation to the real-time data compression in the chemical engineering, especially to the two different classes of production process -the continuous process and the batch process. The main contributions include the following aspects:1) The purpose, the meaning, the characteristics and the state of the art of the data compression are summarized first. Typical data compression methods are introduced with the advantage and disadvantage analysis.2) The characteristics of the continuous process are analysed. A kind of real-time data compression algorithm based on the harr wavelet is proposed and applied to the TE model - the typical simulating model of the continuous process. The result shows the efficiency and the advantages of the algorithm. And the denoising ability of the wavelet compression algorithm for the compressed data is discussed.3) The characteristics of the batch process are analysed. A kind of real-time data compression algorithm based on the db2 wavelet is proposed. The experiments in the rubber banburying process prove the feasibility and the practicability of the algorithm.4) The above two proposed algorithms are compared via the simulations of the same chemical engineering data compression. The influence to the quality of data compression caused by the wavelet with different filter length is studied. The simulation and the theoretical analysis show that the longer the filter is, the better fidelity the reconstructed signal gets, but the greater the boundary effect becomes.Finally, the work of the thesis and the further research directions are summarized.