**7. Outlook**

The necessity to reduce the global greenhouse gas emissions leads to an intensified research and usage of alternative insulating gases with lower greenhouse potential [18,41]. One main challenge of the application of SF6 alternatives will remain: the safe and reliable operation of gas-insulated systems requires a decent knowledge of the partial discharge behavior in order to achieve a reliable risk analysis out of PD measurements. Therefore, the partial discharge physics, such as pulse amplitude

and pulse distance, was compared between SF6 and one possible alternative for medium and high voltage equipment: synthetic air under higher gas pressures. The presented investigations were all carried out at a gas pressure of *p* = 0.5 MPa. The aim of this section is to underline the applicability of the presented experimental procedure to other insulating gases and to give an outlook for future research. In contrast to the presented results of the impulse currents, these investigations will also include the pulseless currents.

#### *7.1. Inception Voltage (Voltage Rising Test)*

The inception voltages of current impulses determined with a VRT as described in Section 5.1 showed the differences expected from the comparison of the effective ionization coefficients between the two insulating gases (Figure 15). Due to the lower dielectric strength of synthetic air, its inception voltage is lower than in SF6. The differences in between both polarities of the protrusion are more prominent in SF6. This might be an effect of the differences in the electron affinity and the ionization energies. Since the starting electrons at a positive protrusion must be generated by detachment processes, a lower number of free electrons was present in the vicinity of the protrusion in SF6 compared to synthetic air. This could be justified comparing the ionization energies of both gases [26] and the low lifetime of free electrons in SF6 [21]. Hence, the voltage could increase further, and the differences in the discharge inception voltage at a positive protrusion were higher comparing both gases.

**Figure 15.** Inception of partial discharge impulse currents depending on the polarity of the protrusion in synthetic air and SF6, *p* = 0.5 MPa.

#### *7.2. Amplitude of Partial Discharge Impulses*

Comparing a single partial discharge impulse at inception voltage, it was evident that the PD behavior was significantly different (Figure 16a). Besides the higher PD amplitude at inception voltage, the time constants for the rise and decay of the impulses were different due to the significantly different effective ionization coefficients (Figure 16b).

This resulted in a higher converted charge for gases with a lower ionization coefficient. The comparison of Figure 16a,b may lead to the assumption that the rise times were dependent on the voltage polarity, especially for synthetic air as the insulating medium. This has to be evaluated in more detail during further investigations.

**Figure 16.** Comparison of single PD impulses in SF6 and synthetic air *p* = 0.5 MPa.

Besides the evaluation of a single PD impulse, the comparison of the peak values is of interest, because they can be an indicator for the detectability of PD with state-of-the-art PD measurement techniques [9]. PD impulses at a negative protrusion in synthetic air were approximately one order of magnitude higher than the impulses in SF6, despite the applied voltage (Figure 17). Whereas the peak current amplitude for a negative protrusion in SF6 was roughly constant, it increased in synthetic air with increasing electric field strength. At a positive protrusion, the peak currents increased significantly in SF6 with increasing voltage, as already described in Section 5.2.1. At twice the inception voltage, there was no significant difference between the impulse amplitude in SF6 and synthetic air. In contrast to SF6, the PD impulses at a positive protrusion in synthetic air were in the same order of magnitude as the ones at a negative protrusion. An increasing amplitude with increasing electric stress as in SF6 could not be observed. The described small subsequent current impulses in SF6 (Figure 11b) could not be observed in synthetic air. Hence, the partial discharge physics seemed to be different, and no change of the type of PD could be observed with varying electric stress.

**Figure 17.** Mean value of partial discharge peak current and minimum/maximum in dependence of the applied voltage and polarity at a gas pressure *p* = 0.5 MPa.

#### *7.3. Time Difference between Subsequent Sequences*

Besides the investigated variations in the impulse current amplitudes, the time differences between subsequent sequences were analyzed. At inception voltage, the time differences in synthetic air were lower than the ones observed for SF6, independent of the voltage polarity (Figure 18).

**Figure 18.** Comparison of time differences between subsequent sequences in synthetic air and SF6 depending on the applied voltage and the polarity of the protrusion, *p* = 0.5 MPa.

Increasing the voltage led to lower time differences between subsequent impulses; the differences between the two investigated gases were less pronounced and in the range of several microseconds. Considering several PD impulses within one recorded sequence (Δ*t* <1 μs), a significant amount could only be observed in SF6 for positive polarity (Figure 11). This behavior was only weakly developed in synthetic air and could only be observed for a few sequences at a negative protrusion at three times the inception voltage *U*i.

The comparison of SF6 with synthetic air at 0.5 MPa showed crucial differences in the discharge behavior. The different effective ionization coefficients, resulting from the different abilities of the gases to act as an electron scavenger, led to different PD current amplitudes, time differences between subsequent impulses, and time constants of the current impulses. This was explicitly shown by the comparison of single PD current impulse amplitudes (Figure 16).

It was confirmed that this type of investigation was not only applicable for SF6-insulated systems, but also for alternative insulating gases.
