A Further Study of the Performance of a Glass-Melting Tank Furnace Fired by Coke Oven Gas
JSGT 1946 V30 T013-T030
This paper is divided into four sections. Section 1 deals with questions raised in an earlier paper concerning the methods of making allowance for furnace temperature changes when fitting regressions to the observations. It is concluded (a) that the amounts of heat associated with the temperature changes which normally occur are so small that they can be ignored in a full analysis, thereby reducing arithmetical labour, (b) that using the data associated with the constant KTin the Furnace Performance Formula is not appropriate. Section 2. Further data are presented for the furnace discussed in the earlier paper. The utility of plotting Q, the Thermal Performance Index, is emphasised, and is shown to direct attention at the earliest moment to changes in performance which require a technical explanation. It is shown that the more formal analysis of the data leads to similar conclusions, and in particular that the lower values of Q obtained during the earlier weeks of the melting campaign now discussed, relative to the corresponding values of the previous campaign, are associated with a lower value of the quantity of gas required per day to maintain the new furnace at a temperature of 1400° with zero glass load. Section 3. Attention is drawn to the usefulness of the coefficient b of the formal analysis as an estimate of the amount of fuel needed to produce a ton of glass in the furnace concerned. If adjustment of fuel supply be possible, then this constant gives guidance as to the magnitude of the fuel change which should be made to accord with inevitable changes in the demand for glass. Section 4. A detailed analysis of data obtained daily for a period of six weeks shows that there are marked differences in the amount of glass melted which are not accompanied by corresponding changes in the fuel consumed. This suggests a possibility of further economies in fuel.
B.P. Dudding
