| The widely application of tungsten and the development of economic causes the increasingdemands of tungsten year by year, which led to over-exploitation of our national exploredtungsten mine, the resource situation is not optimistic. How to use raw tungsten mine efficientlyand recycling second tungsten resources became an important issue in tungsten smelting industry.Off-line detection and manual operation are use for traditional detection and control of the ionexchange process, it’s a long time, and easily missed, labor-intensive method. In order to improvethe efficiency of ion exchange, reduce labor intensity and production costs, researching a set ofmethods of process parameters online detection and automatic control flow switching becomesnecessary. In this paper, soft measurement techniques and predictive control method was used forachieving automatic detection and control of the tungsten ion exchange process, which was basedon the ion-exchange theory and technology as well as referenced to the results of previous studies.The efficiency of tungsten ion-exchange was improved, and accurate switching betweenprocesses was realized. Several aspects of research and discussion in this paper as follows:1) Ion exchange process was divided into four distinct phases for analyzing the microscopicprocess of ion exchange. Fick law was used for describing the dynamic process of ion exchangeprocess that at various stages, and the ion-exchange process model of column process wasestablished with the boundary conditions of ion exchange thermal equilibrium principle. Differentparameters were set to test the process model and the predicted results of the model andexperimental data were contrasted, according to the error, to correct the parameters of the liquidfilm thickness in the model which was difficult to measure.2) Explored the on-line detection technique of process parameters on the basis of thecharacteristics of the tungsten ion exchange process. According to the physical and chemicalproperties of sodium tungstate solution, the soft measurement solutions with the temperature-conductivity parameters for the auxiliary variables was adopted, and focused on building andsolving the soft measurement model. Contrasted the different organizational forms in softmeasurement model of neural network, and analyzed the advantages and disadvantages of variousforms, the final network form was used with component concentration, temperature for the inputand conductivity for the output. Using the form of neural network inverse mapping method withmulti-constraints to solve the parameters of soft measuring which is to be measured, to improvethe hit rate and prediction accuracy of the soft sensor model.3) Analyzing the requirements of automated process control based on the manual operationprocesses of tungsten ion exchange process. Building the control system based on the PC/104embedded board and some function modules. Based on the specific hardware systems, to build and configure the appropriate embedded software platform at different levels.4) According to the ion exchange process control features, explore the predictive controlmethod in the ion exchange process control applications. Proposed predictive control scheme andthe single input variable predictive control algorithm preliminary simulation. |
