The oxidation state or partial oxygen pressure (pO2) of the glass melt influences many glass melt and glass product properties such as fining and foaming behavior, radiant heat transfer, forming characteristics via (a color-dependent) cooling rate, and the glass color of the final product. For these reasons, an on-line system has been developed to measure the partial oxygen pressure of the glass melt, based on an electrochemical cell, using stabilized zirconia as solid electrolyte. The system consists of a disposable sensor of which the tip (containing the electrochemical cell and a thermocouple) is dipped in the glass melt A water-cooled lance is used to protect the connector and wiring from the hot furnace atmosphere.
In an extensive industrial test program, the sensor has been evaluated in the glass melt in the feeder section and in the glass melt underneath the batch blanket in the batch charging area of an industrial furnace producing green container glass.
In the feeder section the sensor lifetime is approximately 2 weeks at temperatures between 1100 and 1200°C. A good correlation was found between the measured cell EMF values (related to pO2) and the analyzed Fe2+/Fetot ratio in the ready product The furnace operator used the EMF signal for batch recipe adaptations and considered the continuous availability of the EMF value as an indicator for the glass melt oxidation state a major improvement compared to the analyzed Fe2/Fetot value, which was available only once a day. The batch sensor signal corresponds well to the feeder sensor signal, but had a lead time of about 9 h, meaning that the measured redox state of the met in this location has predictive value for the product color.
Model calculations show that a simulated 10% redox offset in the batch recipe can be reduced to a 5.6% redox offset of the final product using the feeder sensor and to only a 1.8% redox offset of the final product using the batch sensor. An optimal redox correction system was assumed for the calculation (e.g., the possibility of direct addition of reducing or oxidizing agents to the batch being charged).
The batch sensor has a high potential for control purposes because of its location early in the melting process. However, the much more severe circumstances (relatively high melt temperatures of 1350-1450°C and a more reactive fresh melt) make a reliable measurement of more wan 3 days difficult as a result of a limited sensor lifetime. Future research activities will therefore be focused on the development of a robust redox sensor and integration of its signal in a software control model, indicating the necessary redox corrective measures in the batch recipe.
|Name||Ceramic Engineering and Science Proceedings|