1. Introduction As operating time increases, the insulation of capacitors gradually deteriorates, eventually ending its lifespan due to insulation breakdown. Changes in capacitor insulation characteristics can be assessed by measuring its electrical parameters. The conditions for partial discharge depend on the electric field distribution within the insulation device and the electrophysical properties of the insulating material. Therefore, measuring the characteristics of partial discharge can evaluate the insulation quality of pulse capacitors with different structural designs and manufacturing processes, reveal their defects, and even determine the operating field strength of the insulating dielectric and predict the capacitor's lifespan based on the results of partial discharge characteristic tests. The operating field strength of the dielectric in a pulse capacitor is higher than that of the same capacitor under AC operating conditions. Because the impregnating agent withstands a lower field strength than the insulating film, partial discharge may occur between layers, at points of local electric field concentration, and in areas with residual air bubbles. Radiation, heating, and chemical reactions at the partial discharge point gradually damage the insulation, leading to slow insulation deterioration until breakdown. For a long time, most electrical equipment has operated under power frequency AC voltage, and partial discharge research has also focused on AC voltage operating conditions, with very little research conducted on partial discharge under DC conditions. Because the operating conditions of pulse capacitors vary depending on their application, there is no unified testing standard. The authors conducted extensive partial discharge experiments on all-film pulse capacitors, attempting to determine the relationship between partial discharge parameters and capacitor characteristics such as lifespan and breakdown voltage. 2. PD Detection Test Method Partial discharge in power frequency high-voltage equipment often occurs within a certain phase range of AC voltage. Pulse capacitors operate in a charge-discharge manner and do not have a specific phase corresponding to AC voltage; instead, considering the characteristics of the continuous charging phase, applying DC high voltage to the capacitor and then measuring its partial discharge is closer to reality. Therefore, for the measurement of partial discharge of pulse capacitors, the authors chose to conduct the measurement under DC voltage using an electrical detection method. Figure 1 shows the main circuit wiring and the principle block diagram of the detection system. The low-voltage side voltage is supplied through an isolation transformer, filter, and regulated power supply. The current sensor CT is connected to the grounding side of the test sample CX. The method for determining the calibration factor of the detection system still adopts the method specified in the AC voltage test standard (i.e., injecting charge between the terminals of the test sample). The interference level measured by the entire device within the test voltage range is 8–10 pC. [align=center] T1-Voltage Regulator T2-Isolation Transformer T3-High Voltage Test Transformer R-Current Limiting Resistor D-Silicon Stack CT-Current Sensor CK-Coupled Capacitor CX-Test Sample[/align] The structure of the test sample capacitor is as follows: three layers of 15μm double-sided roughened polypropylene film, aluminum foil as electrodes, and S oil as impregnating agent. The element compression coefficients are 0.78, 0.81, 0.84, and 0.86 respectively, and the capacitance of a single test sample element is 0.63μF. The following PD parameters are selected and recorded in the test: (1) The applied DC voltage value U1 when the first PD amount Q0 is greater than the specified amount. Due to the interference level limitation of the test device, Q0 = 10pC is taken in the test. (2) The number of discharges Nt within the specified time (the unit of t is min). It refers to the number of times the discharge amount is greater than Q0 within the specified time t under a certain DC voltage. The amount of Nt depends on the PD development process and electric field recovery process in the medium. This parameter is closely related to the insulation performance, and the smaller the value, the better. (3) Number of PD discharges (N0) within a specified time (1 min) during self-discharge. This refers to the number of discharges greater than Q0 within a specified time when the charging power supply is disconnected after the element is charged to the specified value and the capacitor self-discharges. If the leakage resistance of the capacitor element is small, the self-discharge speed will be faster, and the voltage drop across the capacitor will be faster, which will lead to a decrease in N0. (4) Maximum apparent discharge quantity Qm. This refers to the maximum apparent discharge quantity of the test sample PD within a specified time. The test method is as follows: the voltage is increased to the specified value at a rate of 500V/s, stabilized for a period of time (0-10 min), and then the capacitor self-discharges (1-2 min), and finally the entire charge is released through a large resistor. 3 Test Results 3.1 Repeatability Test of PD Parameters The repeatability of the measured parameters is detected by repeated tests on the same test sample. In each test, after the capacitor element is charged to the specified voltage value, the capacitor is allowed to self-discharge for 2 min, and then grounded through a large resistor. U1, N0 and Qm are recorded in the test. The test results show that the first obvious partial discharge of the element almost always occurs during the voltage increase process. The reason for this may be that under varying voltage, the voltage across the PP film, S oil, or void impurities in the capacitor element's dielectric is distributed according to the dielectric constant. The oil and air gaps experience higher electric field strength, while the breakdown field strength of air and oil is lower than that of the PP film. The experiment yielded the following results: When the stable voltage U ≤ 11kV, the measured PD parameters U1, N0, and Qm of the tested element are all within a certain range, with relatively small dispersion. Therefore, these parameters can be used to reflect the PD characteristics of the capacitor element. However, when U > 11kV, due to potential damage to the insulation performance, the PD characteristics deteriorate, and the dispersion of PD parameters becomes larger. 3.2 Comparative Experiment of PD Testing for Capacitors with Different Lifespans Two groups of capacitor samples were selected for PD testing: one group consisted of elements that had undergone 1000 repeated charge-discharge cycles at a certain voltage (charging voltage 15kV), and the other group consisted of elements that had not undergone lifespan testing. The aim was to identify the difference in PD parameters between the two groups through the PD test. The comparative experiment method was: after voltage boosting, the capacitor self-discharged for 1 minute, during which PD parameters were measured. The experimental results are shown in Table 1. In the table, N1 represents the number of times the PD charge is greater than Q[sub]0[/sub] in the first minute under stable voltage. As can be seen from Table 1, the PD detection of the test specimens with different degrees of aging after life test and the test specimens without life test has the following differences: (1) The U1 (average 4kV) of the aged test specimens is much smaller than the U1 (average 7kV) of the untested components. The first PD always occurs during the voltage rise process. After repeated charge and discharge tests of the aged test specimens, the insulation material of the test specimens is degraded under the action of strong electric field and large current. The carbonized particles or bubbles in the film and oil increase. During the voltage rise process, PD is easily generated in the bubbles and impurities, causing U1 to drop. It can be seen that the size of U1 can reflect the quality of insulation with the same structure. (2) The N1 of the aged test specimens is usually greater than N0, and N0 is close to 0; while the N1 and N0 of the untested test specimens are not significantly different. 3.3 Comparison of PD of specimens with different compression coefficients Table 2 shows the statistical results of PD parameters of multiple specimens measured under the condition of applying a stable voltage for 4 minutes. Voltage U1 is taken in 3 ranges. N4 in the table represents the number of times the PD is greater than Q[sub]0[/sub] within 4 minutes after stable voltage application. From Table 2, it can be seen that: (1) Correspondence between compression coefficient K and U1 value. Generally speaking, the U1 value of capacitor elements with smaller K is slightly higher. For example, in the specimens with K = 0.78 and 0.81, U1 > 10kV accounts for 60% of the total number of tests, while in the specimens with K = 0.84 and 0.86, it accounts for only 18%. The reason for this may be that in specimens with larger K values, the three layers of film in the dielectric and the film and aluminum foil are more tightly bonded; during vacuum drying and impregnation, the penetration of S oil in elements with larger K values ​​is more difficult, and bubbles are difficult to remove; during the voltage increase process, partial discharge is more likely to occur in the air gap, so the U[sub]1[/sub] value is smaller. (2) Within a similar range of applied stable voltage (10.5-12kV), the number of discharges is also related to the value of K. For example, if a stable voltage of 12kV is applied for 4 minutes, the N4 of the test samples with K = 0.78 and 0.81 is much greater than that of the test samples with K = 0.84 and 0.86. The N4 of the former is greater than N0, indicating that the insulation performance of the test sample with a smaller K value is inferior to that of the test sample with a larger K value under stable withstand voltage conditions. The life test results confirm this, and the charge-discharge life test results with a charging voltage of 15kV are shown in Figure 2. (3) Determination of the critical voltage when a large number of PDs appear in the element. The test results show that as the applied DC voltage increases, the number of PDs and the amount of discharge in the test sample will increase. The results of repeated charge-discharge tests confirm that when the capacitor element has more than 30 PDs of more than 10pC per minute during the stable DC voltage stage, the pulse capacitor will accelerate insulation degradation and greatly reduce its life when operating in this state. Let the minimum applied stable DC voltage at which a large number of partial discharges (PDs) occur be the critical voltage UPD. Determining the UPD value for the partial discharge test of the pulse capacitor is of great reference value for the design of capacitor structural parameters. Figure 3 shows the relationship curve between UPD and t. As can be seen from the figure, UPD decreases rapidly as t increases. When the applied voltage is below 9kV, even if t>10min, the test sample will not show a large number of PDs. Since UPD is closely related to the voltage application time t, it is necessary to select a UPD for a stable voltage application time. Figure 4 shows the probability of a large number of PDs occurring within the fourth minute of 20 test samples under stable voltage application conditions. When the applied voltage reaches 12kV, the probability of a large number of PDs occurring is 47%. At this time, the voltage is less than the capacitor breakdown voltage and will not cause damage to the capacitor insulation. If the applied voltage corresponding to this time is set as the critical voltage, the average UPD of this batch of test samples can be selected in the range of 10-12kV, and the voltage application time during the test can be selected to be about 4min. 4 Conclusions (1) The PD characteristics of a pulse capacitor element can be characterized by parameters such as the applied DC voltage value U1 when the first PD value exceeds the specified value, the number of discharges Nt within a specified time under the specified DC voltage, the N0 within the self-discharge time, and the maximum apparent discharge Q0. (2) Capacitors with good insulation performance have higher U1 values ​​and smaller Nt values, and exhibit the phenomenon of Nt≈N0, and also have a longer lifespan. This can be used to judge the insulation performance of the pulse capacitor element. (3) Based on the life test results and PD test results, a suitable test voltage UPD is selected. Under the action of this voltage, a limited number of PDs occur. Measuring the amount and number of partial discharges within a fixed time under this voltage helps to determine the quality of insulation.