*5.1. A Case Study for Di*ff*erent Operating Modes*

The automatic control was initially adjusted to reproduce the results obtained from simulations with measured data. Inputs with the scrap baskets as well as maximum and minimum mass flows were set to match those documented for an industrial 140 t DC furnace. After adjusting the automatic control to match the simulation on the basis of measured data for the operation chart, a scenario was tested where the oxygen used for the furnace operation was taken from a different source with an oxygen content of 40% instead of the 99% used for the initial case, with the remaining fraction consisting of nitrogen in both cases. The mass flows for oxygen lancing, burners, and post-combustion were adjusted to have the same mass flow of pure oxygen. Due to the decreased oxygen content, this led to an increased total mass flow and more nitrogen being injected together with the oxygen. The operation chart was automatically adjusted to reach the same tapping temperature and maximum carbon content, resulting in an increased energy consumption (both electrical and chemical) and an increased tap-to-tap time. The following cases were studied:


The oxygen shown in the following discussion is the actual mass of pure oxygen; therefore, between Case 1 and Case 2 the total oxygen consumption increased by 6.5%, whereas, due to the increased nitrogen fraction, the nitrogen carried into the furnace with the injected oxygen for Case 2 was 64 times that of Case 1. Figure 6 shows the measured electrical power (real) during the heat compared to the automatic control for Case 1 and Case 2. The time was normalized using tap-to-tap time of the measured heat, the power was normalized using the maximum measured value.

**Figure 6.** Electric power comparison.

Although the measured value fluctuated, the automatic control gave constant values. This had little impact on the overall consumption if the mean measured power was selected for automatic control. As can be seen from the reduction in power to zero, for the measured heat, the second basket was charged at about 0.23 process time, whereas automatic control charging occurred at roughly 0.3 for Case 1 and slightly later for Case 2. The time when the second basket was charged depended on the initial density of the scrap as well as other parameters, which could vary between baskets, and the progression of the process was not reproduced exactly here. Therefore, the reduction of power and mass flows associated with the charging of the second basket occurred slightly later in the simulations when compared to the measured operation chart. At about 0.65 and 0.7, the power was reduced for the flat bath phase for Case 1 and Case 2, respectively.

Figure 7 shows the consumption of natural gas denoted by CH4 and injected coal denoted by C for the measured heat and Cases 1 and 2. Mass flows and time were normalized the same way as in Figure 6. For the first basket, the burners both in Case 1 and Case 2 showed almost identical behavior as was measured. Due to the delayed charging of the second basket, burner operation was delayed as well for the second basket, with a larger delay for Case 2 as meltdown of the initial basket was slower when compared to Case 1. Carbon lancing started slightly later than for the measured operation chart and the mass flow was reduced at around 80% of the process time.

**Figure 7.** Natural gas and injected coal comparison.

Figure 8 shows the normalized mass flows of post-combustion and lanced oxygen. The oxygen mass flow for natural gas combustion is not shown as its profile was similar to that of the natural gas shown in Figure 7. The post-combustion oxygen followed a profile comparable to that of the burners and showed good agreement with the measured values. With the automatic control, oxygen lancing was increased for a short period before the second basket was charged. At roughly 60% process time, the mass flow was increased to its maximum value and reduced again at around 80% for the flat bath phase until tapping. The measured value showed a smoother progression during meltdown and fluctuated more during the flat bath phase; however, the general profile and the total oxygen consumption could be reproduced during automatic control.

**Figure 8.** Oxygen comparison.

Figure 9 shows the progression of the simulated melt temperature for the measured operation chart and Cases 1 and 2. The delayed charging of the first basket was visible between 0.2 and 0.4 process time, as the temperature drop associated with the charging of cold scrap occurred later under automatic control. Although the temperature profile of Case 1 closely followed that of the measured case and reached tapping temperature almost simultaneously, the increased energy demand for Case 2 was visible in the lower temperature during the process and the delayed achievement of the desired tapping temperature. Overall, the automatic control implemented was able to reproduce the progression of the heat compared to the measured operation chart and indicate what impact a different oxygen source would have under otherwise similar conditions, showing the increased tap-to-tap time and consumption of electrical and chemical energy when the same operating strategies were applied in both cases.

**Figure 9.** Melt temperature comparison.

Table 2 shows the resulting consumption of electric power, oxygen, natural gas, and coal, as well as the extracted off-gas and tap-to-tap time for the two cases relative to the values measured for the complete heat. Again, the increases in electrical and chemical energy used for Case 2 were visible.


**Table 2.** Calculated performance indicators for the study cases.
