Influence of mechanical vibration on typical oil–paper insulation partial discharge defects in power transformers

1 Introduction

Power transformers are critical components in electrical transmission and distribution systems, and their reliable operation is essential for the stability and safety of power grids. Statistical investigations have shown that insulation-related failures account for a significant proportion of transformer faults (Fasani, 2025; Jia et al., 2022; Zhou et al., 2023a), among which oil–paper insulation degradation plays a dominant role. Partial discharge (PD) is a typical early manifestation of insulation defects and is widely recognized as both an indicator of insulation deterioration and a mechanism that accelerates insulation aging. Therefore, PD detection and analysis have become indispensable tools for transformer condition assessment and fault diagnosis.

Extensive studies have been carried out to investigate PD characteristics in oil–paper insulation under electrical and thermal stresses. Various PD defect models, such as needle–plate, surface discharge, and floating electrode defects, have been widely employed to reveal discharge mechanisms and characteristic parameters, including partial discharge inception voltage (PDIV), discharge magnitude, repetition rate, and phase-resolved patterns (Jianxun et al., 2021; Gao et al., 2019; Li et al., 2022; Fahmi et al., 2021; Cui et al., 2017). These studies have laid a solid foundation for understanding PD behavior and for developing diagnostic methods based on electrical measurements. However, most existing investigations are conducted under static mechanical conditions, implicitly assuming that the insulation system remains mechanically stationary during operation (Pattanadech and Muhr, 2016; Zhou et al., 2023b; Zhang et al., 2022; Yang et al., 2024).

In practice, mechanical vibration is an inherent operating condition of power transformers (Chen et al., 2022; Wu et al., 2023; Lin et al., 2025; Shao et al., 2011). During normal operation, continuous vibration is generated mainly by core magnetostriction and electromagnetic forces acting on windings. The vibration intensity may further increase under load variations, DC bias, or mechanical looseness. Long-term vibration is known to cause mechanical fatigue, structural loosening, and insulation damage. Meanwhile, vibration may also interact directly with the oil–paper insulation system, introducing additional physical processes that are not present under static conditions.

Recent studies have suggested that mechanical vibration can significantly influence the behavior of insulating oil and internal defects. Mechanical excitation may alter the motion of gas bubbles (Yan, 2021; Sun et al., 2024), conductive particles (Shi, 2025), or weakly constrained components in the oil gap. It may also induce periodic variations in electrode spacing, leading to fluctuations in the local electric field (Wang et al., 2019; Hao et al., 2025). These effects indicate that PD behavior under vibration cannot be fully explained by electrical stress alone (Wu, 2006). Building on these findings, several recent investigations have specifically examined the effects of mechanical vibration on oil-paper insulation degradation. For instance (Guo et al., 2023), reported the gas production characteristics of needle-plate defects under typical vibration conditions (Guo et al., 2024), while (Guo et al., 2023) further analyzed the suspension discharge behavior in similar environments (Guo et al., 2024). Although these works have advanced the understanding of macroscopic chemical responses and isolated defect discharge pheno mena, a notable gap remains in systematic experimental studies that simultaneously address the electrical PD evolution characteristics and compare the vibration sensitivity of multiple typical defect types.

Moreover, the conventional phase-resolved partial discharge (PRPD) analysis framework is fundamentally built upon the assumption that PD is governed exclusively by the power-frequency electric field (Saleh and Algwari, 2023; Liu et al., 2020; Wang et al., 2019). This assumption inherently overlooks the potential modulation effects that the mechanical vibration phase may exert on discharge pulses, and there is currently a lack of established methodologies to capture such coupled dynamics.

To address these gaps, this paper presents an experimental study on the influence of mechanical vibration on typical oil–paper insulation PD defects in power transformers, with a focus on electrical parameter evolution, comparative defect sensitivity, and underlying micro-physical mechanisms. An electro-mechanical coupled experimental platform was established, enabling the simultaneous application of controlled mechanical vibration and high-voltage electrical stress. Two representative PD defect models, namely, needle–plate and floating electrode defects, were investigated under various vibration frequencies and acceleration levels. Key PD parameters, including PDIV, discharge repetition rate, discharge magnitude, and phase-resolved characteristics, were quantitatively analyzed. Notably, a Vibration Phase-Resolved Partial Discharge (V-PRPD) method is introduced to innovatively map the PD pulses to both the electrical and mechanical vibration phases, providing a novel analytical dimension for identifying vibration-modulated PD behaviors.

The results demonstrate that mechanical vibration affects PD behavior in a defect-dependent manner. Floating electrode defects exhibit high sensitivity to vibration, characterized by a significant reduction in PDIV and discharge magnitude and a pronounced increase in discharge repetition rate. In contrast, needle–plate defects show relatively lower sensitivity, with vibration exerting both suppressive and promotive influences depending on vibration intensity. Further analysis indicates that these differences arise from distinct dominant micro-physical mechanisms: vibration-induced electric field distortion plays a major role in floating defects, whereas needle–plate defects are jointly influenced by electric field variation and bubble dynamics disturbance in the oil gap. Unlike previous studies that primarily correlated vibration with macroscopic dissolved gas analysis, this study elucidates the micro-mechanistic electrical behaviors governing PDIV and discharge magnitude variations.

The findings of this study provide experimental evidence for the interaction between mechanical vibration and partial discharge in oil–paper insulation and highlight the importance of considering mechanical operating conditions in PD-based transformer diagnostics.

2 Experimental setup

To investigate the PD characteristics of oil–paper insulation under vibration, a Mechanical-electrical test platform was established. The schematic diagram of the experimental setup is shown in Figure 1. The platform consists of three main subsystems: a high-voltage (HV) generation system, a mechanical vibration system, and a PD detection system.

The mechanical vibration system consists of a power amplifier (Model: NF BP4610) and a vibration exciter (Model: VT-150). By regulating the output of the power amplifier, the amplitude and frequency of the vibration generated by the exciter can be precisely controlled. The exciter is rigidly connected to the bottom of the defect model. The high-voltage electrode is supported by an insulating bracket to isolate it from the model casing; this configuration is designed to simulate the relative vibration between transformer windings and electrodes. The AC voltage was generated by a power frequency test transformer (100 kV/50 kVA) and applied to the model through the protection resistor. The data acquisition system primarily comprises a digital oscilloscope and a PD detector. A measuring impedance was used in conjunction with the PD detector (Model: DX9121b) to acquire Phase-Resolved Partial Discharge (PRPD) patterns. Simultaneously, the oscilloscope was employed to monitor both the PD pulse signals and the mechanical vibration waveforms.

To represent typical partial discharge defects in oil–paper insulation of power transformers, two representative defect models were employed in this study: a needle–plate defect and a floating electrode defect.

Needle–plate discharge and floating electrode discharge are two common forms of partial discharge defects encountered during transformer operation. In this study, corresponding needle–plate and floating electrode models were constructed to represent typical discharge scenarios in oil–paper insulation. The needle–plate model simulates tip-dominated discharges caused by strong electric field concentration, which commonly occur when sharp protrusions, metallic burrs, or foreign conductive particles exist in the insulation structure. In contrast, the floating electrode model represents operating conditions such as suspended sharp conductors or poorly fixed leads, whose discharge behavior is generally more sensitive to mechanical vibration.

To realistically simulate typical partial discharge defects in transformers, both defect models were implemented in an oil-filled tank, as shown in Figure 2. In the floating electrode model, a high-voltage needle electrode is arranged opposite to a floating needle electrode. Both needle electrodes have a curvature radius of 0.2 mm, and the gap distance between the two needle tips is fixed at 1 mm. The floating needle electrode is electrically isolated from ground, forming a floating potential under the applied electric field, while the high-voltage needle electrode is connected to the power-frequency high-voltage source.

In the needle–plate discharge model, a circular insulating pressboard with a diameter of 25 mm and a thickness of 1 mm is mounted on a copper plate electrode, which is reliably grounded. A high-voltage needle electrode with a tip curvature radius of 0.2 mm is positioned above the pressboard, forming a 1 mm oil gap between the needle tip and the pressboard surface. This configuration generates a highly non-uniform electric field near the needle tip, promoting tip-dominated partial discharge activity.

Both electrode configurations are installed inside a cylindrical oil-filled tank with a height of 70 mm and a base diameter of 104 mm. The tank is filled with transformer insulating oil to ensure consistent dielectric conditions during the experiments. By maintaining identical gap distances and comparable electrode tip geometries, the influence of defect type and mechanical vibration on partial discharge characteristics can be systematically investigated.

During the experiments, the high-voltage electrode was mechanically isolated from the oil tank by an insulating support and introduced into the discharge model from the top. The grounded electrode was integrated with the base of the discharge model and rigidly connected to the vibration table through bolts, ensuring reliable mechanical coupling. When sinusoidal vibration excitation was applied by the vibration exciter, only the grounded electrode moved periodically together with the base, while the high-voltage electrode remained stationary. As a result, a controlled relative displacement between the two electrodes was generated, which effectively simulates the vibration-induced relative motion of insulation defects under actual transformer operating conditions.

At the initial stage, the applied voltage was gradually increased until the discharge model reached the partial discharge inception voltage (PDIV). At this voltage level, the detectable ultrasonic signals were sparse and insufficient for quantitative analysis of discharge characteristics. Therefore, the applied voltage was further increased to exceed the PDIV, so that stable partial discharge activity could be sustained, allowing sufficient ultrasonic signals to be acquired for subsequent analysis.

In the electro-mechanical coupled experiments designed in this study, mechanical vibration was treated as the primary independent variable. A constant power-frequency AC voltage was applied by the high-voltage test system, while the vibration parameters were adjusted using a power amplifier. The selected vibration frequencies and acceleration amplitudes were determined based on the mechanical vibration characteristics of power transformers under practical operating conditions, with reference to existing literature and engineering experience.

Regarding vibration frequency selection, the mechanical vibration of oil-immersed transformers during normal operation is mainly induced by periodic magnetostriction of the core and electromagnetic forces acting on the windings. Consequently, the vibration components are concentrated around 100 Hz and its harmonic frequencies. Previous simulation and experimental studies have shown that transformer vibration energy is mainly distributed near 100 Hz and its low-order harmonics, particularly 200 Hz and 300 Hz (Cui et al., 2022; Xie, 2023). Therefore, vibration frequencies of 100 Hz, 200 Hz, and 300 Hz were selected in this study to represent the dominant vibration characteristics of operating transformers.

For vibration amplitude selection, both the vibration characteristics encountered in practical transformers and the controllable operating range of the experimental platform were considered. Since transformer vibration amplitudes vary depending on transformer rating, structural design, installation conditions, and operating conditions, it is difficult to define a universal vibration amplitude applicable to all transformers. Therefore, acceleration amplitudes ranging from 0.25 g to 1 g were selected to cover representative operating conditions from relatively weak vibration to comparatively strong mechanical excitation. Four acceleration levels (0.25 g, 0.5 g, 0.75 g, and 1 g) were adopted with an interval of 0.25 g to systematically investigate the influence of vibration intensity on partial discharge behavior.

To further analyze partial discharge (PD) characteristics under electro-mechanical coupling, a vibration phase–resolved partial discharge (V-PRPD) pattern is introduced as an analytical representation. Conventional phase-resolved partial discharge (PRPD) patterns are widely used to characterize the phase distribution and magnitude of discharge pulses within a power-frequency voltage cycle and are effective for discharge type identification and discharge intensity evaluation. However, this method implicitly assumes that PD behavior is predominantly driven by the electric field and does not incorporate mechanical excitation into the analysis framework, which limits its ability to characterize the relationship between discharge behavior and mechanical stress.

To address this limitation, the V-PRPD pattern incorporates the mechanical vibration signal as a reference phase, enabling the correlation between discharge pulse magnitude and vibration phase within a unified coordinate system. The construction procedure is as follows. First, the synchronously acquired signals are filtered and preprocessed. Then, phase calculation is performed on the vibration signal to obtain the vibration phase corresponding to each discharge pulse. Finally, the discharge pulse magnitudes are mapped according to their associated vibration phases, forming a V-PRPD pattern with vibration phase as the horizontal axis and discharge magnitude as the vertical axis.

Using the V-PRPD pattern, the distribution characteristics of partial discharge activity over the vibration cycle can be intuitively analyzed. Differences in discharge pulse magnitude and discharge occurrence density at different vibration phases can be compared, thereby revealing the influence of mechanical vibration on PD behavior. This approach provides an effective analytical tool for investigating partial discharge mechanisms under electro-mechanical coupling conditions.

3 Partial discharge characteristics of floating defects under mechanical vibration

3.1 Influence of mechanical vibration on the discharge inception characteristics of floating defects

In the floating electrode configuration, when one needle electrode is in a floating state, the electric field distribution within the insulation gap becomes inherently unstable. Compared with the needle–plate electrode, the floating electrode not only exhibits more severe electric field distortion, but is also prone to slight positional displacement under external mechanical disturbances. As a result, its partial discharge (PD) characteristics demonstrate a higher sensitivity to mechanical vibration.

For floating electrode defects, the application of mechanical vibration has a pronounced influence on the discharge inception characteristics, leading to a significant reduction in the partial discharge inception voltage (PDIV). As shown in Figure 3, when a vibration frequency of 300 Hz with an acceleration amplitude of 0.25 g is applied, the PDIV decreases by up to 73.8%, representing the maximum observed reduction under the tested conditions.

To further investigate the correlation between mechanical vibration and floating electrode discharges, a vibration excitation of 100 Hz and 1 g was applied to the floating defect model. Partial discharge ultrasonic signals and vibration signals were synchronously acquired. Based on the correspondence between ultrasonic pulse amplitudes and vibration phases, multiple datasets were collected and used to construct the vibration-phase-resolved partial discharge (V-PRPD) patterns of the floating defect. The resulting V-PRPD pattern is shown in Figure 4.

The V-PRPD results indicate that the ultrasonic pulse amplitudes of floating defects tend to concentrate around the vibration waveform peaks under 100 Hz excitation. Specifically, a higher density of ultrasonic PD pulses is observed near the vibration phase of 90°, whereas fewer pulses appear around 270°. When the vibration phase approaches 90°, the relative distance between the two needle tips reaches its minimum, resulting in the most severe electric field fluctuation and facilitating PD occurrence. In contrast, at a vibration phase near 270°, the needle tips are farthest apart, leading to a weaker electric field distortion and a lower probability of discharge initiation.

3.2 Influence of mechanical vibration on the discharge development characteristics of floating defects

The partial discharge behavior of floating electrode defects exhibits higher sensitivity to mechanical vibration. Therefore, in the investigation of the influence of vibration on the discharge development characteristics of floating defects, different vibration conditions were applied. The corresponding partial discharge characteristic parameters of floating defects under various vibration states are shown in Figure 5.

The application of mechanical vibration has a pronounced influence on the discharge repetition rate of floating electrode defects. As shown in the figure, a clear increasing trend in the repetition rate is observed under 100 Hz vibration. Under 200 Hz and 300 Hz vibration conditions, the repetition rate exhibits a rapid increase with increasing vibration amplitude, followed by a tendency to saturate.

For the 100 Hz vibration, the continuous increase in the repetition rate can be attributed to the fact that mechanical vibration causes the electric field strength between the electrodes to exceed the PD inception threshold. As the vibration amplitude increases, the duration during which the electric field remains above the threshold becomes longer, resulting in more frequent discharge occurrences. In contrast, under 200 Hz and 300 Hz vibration, the saturation effect of vibration amplitude on the repetition rate arises because higher-frequency vibration prolongs the duration when the inter-electrode distance reaches its minimum, while the vibration-induced relative displacement remains relatively small. Consequently, further increases in vibration amplitude no longer lead to a significant enhancement in the discharge repetition rate.

Mechanical vibration also leads to a significant reduction in both the average and maximum discharge magnitudes. Under the same acceleration amplitude, increasing the vibration frequency does not result in a pronounced additional decrease in discharge magnitude. However, when the acceleration amplitudes are 0.5 g, 0.75 g, and 1 g, the reduction in discharge magnitude is substantially greater than that observed at 0.25 g. Among all tested conditions, the vibration at 300 Hz and 1 g has the most significant impact on the average discharge magnitude, which is reduced by 88.9% compared with the case without vibration.

With the applied voltage kept constant, the PRPD patterns of floating electrode discharges obtained under different mechanical vibration conditions are shown in Figure 6. The PRPD patterns of floating electrode PD are significantly influenced by mechanical vibration. Without vibration, PD events occur less frequently and are mainly concentrated around the peaks and troughs of the AC voltage waveform.

The application of mechanical vibration promotes more frequent discharge activity, and the PRPD patterns exhibit a sheet-like distribution under vibration excitation. Meanwhile, mechanical vibration leads to a reduction in the average discharge magnitude. Specifically, vibrations at 100 Hz and 200 Hz reduce the average discharge magnitude by 73.5% and 73.6%, respectively. These results indicate a strong correlation between floating electrode PD behavior and mechanical vibration.

3.3 Analysis of the influence of mechanical vibration on floating electrode partial discharge

The floating defect model consists of a high-voltage needle electrode facing an ungrounded floating needle electrode. Owing to the absence of rigid mechanical constraint, the floating electrode possesses a relatively high degree of freedom in its displacement. As a result, the local electric field formed between the two needle tips becomes more intense and highly sensitive to mechanical vibration.

In the absence of mechanical vibration, the distance between the two needle tips can be regarded as being determined by a static gap d0 together with microscale perturbations Δh within the dielectric medium. Under this condition, the local electric field can be approximately expressed as Equation 1:where β_f is the electric field distortion coefficient between the floating electrodes, which is related to the geometric structure of the two needle electrodes.

E f t = β f U d 0 + Δ h (1)

The influence of mechanical vibration on floating partial discharge essentially arises from the vibration-induced displacement of the high-voltage electrode, which leads to a change in the relative position between the high-voltage needle and the floating needle. Consequently, the electric field in the insulation gap is jointly affected by electrical stress and mechanical vibration. Under electro-mechanical coupling conditions, the electric field strength between the floating electrodes can be expressed as:where κ_f denotes the conversion efficiency of vibration acceleration into equivalent displacement at the floating electrodes. Owing to the structural characteristics of the floating electrode, both coefficients are larger than those of the needle–plate electrode configuration.

E f v t = β f U d 0 + κ f a 2 π f v 2 sin 2 π f v t (2)

In addition, the electric field distortion factor β_f in Equation 2 is theoretically a function of the instantaneous distance d between the two needle electrodes. According to the approximation theory for highly non-uniform electric fields, βf can be expressed as Equation 3:where d ₀ denotes the electrode gap under static conditions, p is an exponent related to the curvature radius of the needle tip, typically ranging from 1 to 2, and d represents the instantaneous electrode spacing.

β f = d 0 d p (3)

Under the maximum vibration condition considered in this study, the displacement amplitude of the electrode is approximately 4 μm. Compared with the static gap of d0 = 1 mm, the relative variation in electrode spacing Δd/d0 is only about 0.4%. Accordingly, the relative fluctuation of the electric field distortion factor Δβf/βf is estimated to fall within the range of 0.4%–0.8%. This level of variation is much smaller than the electric field distortion directly induced by mechanical vibration. Therefore, in the present model, βf is reasonably treated as a constant.

The significant reduction in PDIV of floating electrode discharges under mechanical vibration is mainly attributed to the large electric field distortion factor and the periodic relative displacement of the floating electrode induced by vibration acceleration. The dynamic compression of the oil gap reduces the effective breakdown distance, resulting in enhanced local electric field distortion and consequently lowering the external voltage required to trigger partial discharge. Moreover, due to the structural characteristics of the floating electrode, the inhibitory effect of mechanical vibration on bubble formation is relatively weak, and thus vibration does not impose an obvious suppressive influence on discharge initiation.

The discharge repetition rate of floating electrode PD under vibration exhibits a clear dependence on vibration frequency. At 100 Hz, mechanical vibration periodically modulates the electrode spacing, causing fluctuations in the electric field strength. As the duration during which the field exceeds the discharge threshold increases with vibration amplitude, the discharge triggering probability continuously rises, leading to a monotonic increase in repetition rate with acceleration amplitude. In contrast, at higher vibration frequencies of 200 Hz and 300 Hz, the effective duration of enhanced electric field fluctuations becomes longer. High-frequency discharge activity is already triggered at a relatively low acceleration amplitude of 0.25 g. With further increases in vibration intensity, the repetition rate no longer rises significantly due to energy release through discharge channels and saturation effects, resulting in a characteristic “rapid increase followed by stabilization” behavior. In addition, high-frequency vibration imposes stronger disturbance on discharge channels, which enhances discharge sensitivity but weakens discharge sustainability, causing the overall repetition rate to be slightly lower than that under low-frequency vibration.

The reduction in discharge magnitude is primarily caused by the periodic fluctuations in electrode spacing and local electric field induced by mechanical vibration, which allow floating discharges to be triggered prematurely before the electric field reaches a high level, thereby producing lower-energy discharge pulses. Meanwhile, vibration-induced disturbances render the discharge channel unstable, hindering the development of high-energy discharges and further reducing discharge magnitude. A comparison across different vibration frequencies shows that the decrease in discharge magnitude becomes slightly more pronounced as vibration frequency increases. This is because higher vibration frequencies correspond to shorter vibration periods; although the displacement amplitude is smaller, the electric field fluctuates more frequently, leading to repeated discharge triggering at lower instantaneous field strengths. As a result, the energy of individual discharges is further reduced, and the overall discharge magnitude gradually decreases with increasing vibration frequency.

4 Partial discharge characteristics of needle–plate defects under mechanical vibration

4.1 Influence of mechanical vibration on the discharge inception characteristics of needle–plate defects

Under an alternating electric field, the needle–plate electrode configuration exhibits a highly non-uniform electric field distribution, with pronounced field concentration at the needle tip. As a result, partial discharge is readily initiated in the vicinity of the needle electrode. Mechanical vibration induces periodic variations in the electrode spacing, leading to fluctuations in the local electric field distribution. When the instantaneous electric field strength exceeds the discharge inception threshold, partial discharge occurs. Figure 7 illustrates the influence of mechanical vibration on the par-tial discharge inception voltage (PDIV).

Experimental results indicate that mechanical vibration exerts a dual influence on partial discharge behavior in the needle–plate electrode configuration. When a relatively low vibration level is applied, the PDIV increases. The maximum increase of 5.6% is observed at a vibration frequency of 300 Hz and an acceleration amplitude of 0.25 g. As the vibration amplitude further increases, the PDIV begins to decrease. When the acceleration amplitude exceeds 0.75 g, the PDIV becomes lower than that under the non-vibration condition. The most pronounced reduction, reaching 11.1%, occurs at a vibration frequency of 100 Hz and an acceleration amplitude of 1 g.

At the discharge inception stage, low-level vibration disturbs the formation and aggregation of gas bubbles in the oil gap, hindering the development of bubble bridges between electrodes and thereby suppressing discharge initiation. This effect leads to an increase in PDIV at an acceleration amplitude of 0.25 g. With further increases in vibration intensity, the vibration-induced disturbance of the electric field becomes dominant. The resulting enhancement of electric field distortion increases the probability that the local field strength exceeds the discharge inception threshold, facilitating discharge initiation and causing a reduction in PDIV.

Figure 8 presents the V-PRPD pattern and the discharge phase distribution of the needle–plate defect under a vibration condition of 100 Hz and 1 g. The ultrasonic PD pulse amplitudes of the needle–plate defect are relatively uniformly distributed over the sinusoidal vibration cycle at 100 Hz, forming a band-like scatter pattern. However, a noticeable peak in discharge occurrence is observed within the vibration phase range of 180°–240°.

His characteristic indicates that, for the needle–plate defect, the discharge behavior remains predominantly governed by the electric field distribution. Meanwhile, mechanical vibration influences the discharge probability by periodically modulating the electrode gap distance, thereby introducing secondary effects on PD occurrence.

These results indicate that the discharge behavior of the needle–plate defect remains predominantly governed by the electric field, with mechanical vibration exerting a limited influence on the discharge triggering instant. Owing to the structural characteristics of the needle–plate configuration, the electric field distortion induced by vibration-related variations in electrode spacing is insufficient to dominate the discharge process. Consequently, PD initiation does not exhibit a strong dependence on the vibration phase, and discharge events are randomly distributed throughout the vibration cycle without a clear preference for any specific phase interval.

Unlike the floating defect discussed previously, the needle–plate defect exhibits both suppressive and promotive effects under mechanical vibration. This dual behavior primarily originates from the structural differences between the needle–plate and floating electrode configurations.

Under low-amplitude vibration conditions, the relatively large effective surface area of the plate electrode introduces stronger disturbances to bubble motion in the insulating oil. Such disturbances hinder the stable migration and accumulation of bubbles along the electric field lines, preventing the formation of bubble aggregation regions. As a result, an increase in PDIV is observed, a phenomenon that does not occur in the floating defect model. As the vibration amplitude increases, the influence of electrode relative displacement on the geometric electric field distribution becomes increasingly significant. Electric field distortion and the enhancement of instantaneous field peaks gradually become the dominant factors. Consequently, the discharge-suppressing mechanism associated with bubble disturbance becomes secondary, making partial discharge more likely to occur and leading to a decrease in PDIV.

4.2 Influence of mechanical vibration on the discharge development characteristics of needle–plate defects

Mechanical vibration acts as a periodic mechanical disturbance to tip discharge, influencing the electric field distribution at the oil–paper insulation tip as well as the relative distance within the discharge model. Meanwhile, vibration also disturbs the behavior of gas bubbles in the insulating oil, thereby affecting the partial discharge process.

By analyzing the measured PD signals, characteristic parameters of partial discharge under different vibration conditions can be obtained. The discharge repetition rate, average discharge magnitude, and maximum discharge magnitude are all important indicators for characterizing PD behavior. The PD characteristic parameters under different vibration conditions are presented in Figure 9.

With the application of mechanical vibration, the discharge repetition rate of the needle–plate defect exhibits an overall increasing trend. However, when the vibration acceleration increases from 0.25 g to 0.5 g and from 0.75 g to 1 g, a decrease in the repetition rate is observed. This indicates that the influence of vibration on discharge activity is not strictly monotonic with vibration intensity.

As the vibration level increases, the maximum discharge magnitude shows an increasing tendency. The average discharge magnitude exhibits a more pronounced increasing trend under low-frequency vibration at 100 Hz, whereas a decreasing tendency is observed under vibration frequencies of 200 Hz and 300 Hz.

The PRPD patterns of the needle–plate defect under different mechanical vibration conditions are shown in Figure 10. It can be seen that mechanical vibration has a relatively limited influence on the PRPD patterns of the needle–plate defect. The discharge pulses remain mainly concentrated around the voltage peaks, forming a triangular distribution. After vibration is applied, some low-amplitude discharges appear during the negative half-cycle of the applied voltage. Under 100 Hz and 200 Hz vibration, the average discharge magnitude increases by 34.3% and 33.0%, respectively. Overall, the PD behavior of the needle–plate defect exhibits a weak correlation with mechanical vibration.

4.3 Analysis of the influence of mechanical vibration on needle–plate partial discharge

The needle–plate electrode configuration represents a typical highly non-uniform electric field discharge model, in which strong electric field concentration occurs in the needle tip region and serves as the dominant trigger for partial discharge. Under conditions without mechanical vibration, the geometric structure between the needle and plate electrodes remains stable, and the inter-electrode gap can be regarded as a constant value d0. In this case, the electric field distribution is mainly determined by the applied power-frequency voltage. Owing to the sharp geometry of the needle electrode, significant local electric field distortion occurs near the needle tip, and partial dis-charge is therefore preferentially initiated in this region. The local electric field strength near the needle tip can be approximately expressed as Equation 4:where β _n is the electric field distortion factor between the needle and plate electrodes, which is related to the electrode geometry and the inter-electrode gap distance.

E n t = β n U d 0 (4)

From the perspective of electric field variation, the influence of mechanical vibration on partial discharge in the needle–plate configuration essentially arises from the periodic change in the discharge gap caused by sinusoidal vibration. This periodic variation in the electrode spacing leads to a corresponding periodic fluctuation in the electric field strength within the oil gap. Under mechanical vibration, the instantaneous electric field strength between the needle and plate electrodes can be expressed as Equation 5:where κ _n denotes the efficiency coefficient for converting vibration acceleration into equivalent displacement in the needle–plate electrode system, f _v is the vibration frequency, and a is the vibration acceleration.

E n v t = β n U d 0 + κ n a 2 π f v 2 sin 2 π f v t (5)

Similar to the analysis conducted for the floating electrode defect, the electric field distortion factor βn of the needle–plate configuration also follows the variation law associated with the instantaneous discharge gap. Under micrometer-scale relative electrode displacement, the fluctuation of βn is less than 0.8%. Therefore, in this section, βn is likewise approximated as a constant.

The introduction of mechanical vibration leads to periodic and slight variations in the needle–plate electrode spacing, which in turn causes periodic distortion of the local electric field distribution. As the vibration acceleration increases, the instantaneous electrode displacement becomes more pronounced, resulting in stronger compression or stretching of the oil gap and causing the local electric field strength to temporarily exceed that under static conditions. When the acceleration amplitude remains constant, the displacement amplitude of the electrode is inversely proportional to the square of the vibration frequency. Consequently, with increasing vibration frequency, the relative electrode displacement decreases rapidly, and the degree of vibration-induced local electric field distortion gradually weakens.

In addition to changes in the electric field distribution, the generation, migration, and aggregation of gas bubbles in the oil gap also play an important role in the partial discharge behavior of the needle–plate defect. When the electrodes remain stationary, the electric field is highly concentrated at the needle tip, causing dissolved gas in the insulating oil to precipitate and form microbubbles. Driven by the electric field force, these bubbles migrate along the electric field lines. Over time, multiple bubbles gradually aggregate under the action of the electric field, forming localized regions with reduced dielectric strength within the electrode gap. When the applied voltage reaches the corresponding threshold, partial discharge is initiated preferentially within the bubble aggregation region. At this stage, no complete dielectric breakdown of the insulation structure occurs; instead, non-bridging partial discharge takes place within the localized weak insulation region formed by the bubble aggregation.

As illustrated in Figure 11, when sinusoidal vibration excitation is applied to the electrodes, the oil gap between the electrodes is subjected to periodic mechanical disturbance. This mechanical disturbance significantly affects the motion behavior of gas bubbles in the oil, interfering with their directional migration along the electric field lines and hindering the formation of a stable bubble aggregation region. Instead, part of the bubbles are dispersed toward the edges of the insulating paperboard, suppressing the development of bubble clusters within the electrode gap. As a result, the formation of localized weak insulation regions is inhibited, leading to an increase in the threshold of the partial discharge inception voltage (PDIV) and exhibiting an overall suppressive effect on partial discharge activity.

The application of mechanical vibration induces relative motion between the needle and plate electrodes, resulting in periodic distortion of the electric field strength within the discharge gap. This electric field fluctuation facilitates the initiation of partial discharges and generally leads to an increase in discharge repetition rate. Meanwhile, mechanical vibration also disturbs the originally stable discharge channel and interferes with the formation and aggregation of gas bubbles in the oil gap, exerting a certain suppressive effect on discharge development. When the vibration acceleration amplitude or frequency varies, the influence of vibration on bubble generation and aggregation may become more pronounced, which can locally reduce the discharge repetition rate. Nevertheless, since vibration-induced electric field distortion remains the dominant factor affecting discharge initiation, the overall discharge repetition rate tends to increase after vibration is applied, although local decreases may occur under specific vibration conditions.

After mechanical vibration is applied to the electrode model, the relative position between the needle tip and the oil–paper interface undergoes dynamic variation. Owing to the highly non-uniform electric field near the needle tip, vibration promotes further local field concentration, enhancing the electron ionization process and leading to a transient increase in discharge energy. Consequently, the maximum discharge magnitude exhibits an increasing trend. At lower vibration frequencies, the discharge channel remains relatively stable, and the disturbance induced by vibration is limited. Under such conditions, vibration mainly promotes partial discharge activity, resulting in more regular discharge behavior and the occurrence of stable medium-to-high magnitude discharges, thereby increasing the average discharge magnitude. When the vibration frequency increases to 200 Hz and 300 Hz, the relative electrode motion becomes more rapid, and discharge events occur more frequently. The accumulated energy is released through multiple low-energy discharges, leading to a reduction in the average discharge magnitude and an overall decrease in discharge energy.

5 Conclusion

This paper systematically investigated the influence of mechanical vibration on the electrical partial discharge (PD) characteristics of typical oil–paper insulation defects (needle-plate and floating electrode) in power transformers. Distinct from previous studies that primarily focused on macroscopic gas production characteristics under vibration, this work delves into the electrical parameter evolution and the underlying micro-physical mechanisms governing defect-specific PD behaviors. The main conclusions are as follows:

• Mechanical vibration affects PD inception and development in a highly defect-dependent manner. Floating electrode defects exhibit high sensitivity to vibration, showing a continuous reduction in PDIV and discharge magnitude due to severe vibration-induced geometric electric field distortion. In contrast, needle–plate defects show a unique “suppression-then-promotion” trend: low-amplitude vibration suppresses discharge by disturbing bubble dynamics in the oil gap (increasing PDIV), while high-amplitude vibration promotes discharge via electrode displacement and subsequent electric field intensification.
• The proposed Vibration Phase-Resolved Partial Discharge (V-PRPD) methodology provides a novel analytical dimension for vibration-modulated PD behaviors. The V-PRPD patterns reveal that floating electrode discharges exhibit evident phase clustering with respect to the mechanical vibration cycle, reflecting their high susceptibility to vibration-induced perturbations. Conversely, the discharge distribution of needle–plate defects remains relatively uniform across the vibration phase, further confirming their lower mechanical sensitivity.
• The comparative analysis highlights that the electrical PD evolution under electro-mechanical coupling cannot be generalized across different defect structures. The distinct micro-mechanisms—bubble dynamics disturbance for needle-plate defects versus geometric electric field distortion for floating defects—provide a more comprehensive theoretical basis for multi-defect diagnosis. These findings underscore the necessity of incorporating mechanical operating conditions and advanced V-PRPD features into PD-based condition assessment for power transformers.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Yiqi Liu, South China University of Technology, China

Reviewed by: Tao Zhao, North China Electric Power University, China

Peter Mwinisin, University of Bologna, Italy