Study on Pantograph–Rigid Catenary Separation Through Simulation Experiments and the Dynamic Characteristics of DC Arcs

Abstract

The pantograph–catenary system is a critical component of the traction power supply network. Due to hard points on the overhead contact line and vibrations of the pantograph, pantograph–catenary separation may occur, leading to offline DC arc events. To investigate the characteristics of DC arcs generated during pantograph–catenary separation in metro systems, this study constructs a laboratory platform that simulates the offline process and analyzes the electrical characteristics, optical intensity, and arc-burn duration under different electrode separation conditions. First, a DC pantograph–catenary offline arc simulation platform is developed using a contact wire, a carbon-strip pantograph slider, and a linear motor, enabling slider movement in both horizontal and vertical directions. Second, offline discharge experiments are conducted to compare the discharge process and electrical arc characteristics with and without horizontal slider motion. Finally, arc luminosity and burn duration are measured under various electrode separation configurations, and the influence of voltage level, current level, and electrode material is examined. Experimental results reveal a significant polarity effect, where the arc burn duration is notably longer when the contact wire serves as the cathode than when the carbon slider serves as the cathode. At the instant of separation, the high electric field intensity within the micro-gap triggers pronounced “peak phenomena” in both arc resistance and power, accompanied by abrupt voltage surges and transient current dips. Furthermore, the introduction of horizontal motion modulates the arcing process, causing the stable arcing voltage to follow a distinctive trend of a slow increase followed by a gradual decrease, which differs from static separation characteristics. Finally, this study demonstrates that voltage levels exert a more dominant influence on arc luminosity and duration than current levels, while the maintenance voltage of the arc channel remains significantly lower than the air breakdown voltage.

1. Introduction

As of August 2025, a total of 54 cities in China have put urban rail transit systems into operation, comprising 331 lines with a cumulative operational mileage of 11,210.7 km. In metro systems, rigid overhead conductor rails are widely adopted due to their advantages in spatial efficiency, convenient maintenance, and improved safety performance. Compared with flexible catenary systems, rigid suspension offers lower installation height requirements, simpler structural design, reduced construction cost, and easier routine maintenance. However, as train speeds increase and severe wheel–rail excitations occur, the stable operation of pantographs is significantly affected, leading to more frequent pantograph–catenary separation events.

Extensive research on arcing phenomena in rigid overhead conductor rails has been conducted by scholars both domestically and internationally. Wang [1], through the synchronous acquisition of arc electrical data and high-speed imagery combined with image processing and three-color pyrometry, concluded that under negative polarity and high-current conditions, arc energy is higher; current is the dominant factor influencing arc behavior; and an external magnetic field can effectively shorten arc duration. Song [2] developed an improved magnetohydrodynamic (MHD) arc model and validated it through COMSOL6.2 simulations and spectroscopic experiments, showing that the arc column exhibits the highest temperature at its center, with temperature increasing with current and decreasing with gap length. Integrating COMSOL modeling with experimental verification, Ma [3] demonstrated that contact condition and current are key factors influencing pantograph–catenary arcing and further established a predictive model capable of identifying arcing or weak current-collection points with high accuracy. Tang [4] constructed a distributed-parameter train-network model, improved the Habedank arc model, and combined ATP-EMTP simulations with field testing, concluding that pantograph–catenary arcs produce strong electromagnetic radiation in specific frequency bands that interfere with onboard equipment. Tang also proposed effective mitigation measures, including recommended speed limits and the installation of ferrite rings. Hao and Gao [5], based on secondary development of Fluent using magnetohydrodynamic theory and dynamic mesh technology validated through experiments, found that during pantograph lowering, oppositely rotating vortices form around the arc, giving rise to thermal contraction effects, and faster lowering speeds result in a greater rate of change in arc voltage with increasing separation distance. Qiao [6] enhanced the Mayr arc model using energy balance and the transverse arc-blowing theory and validated it through Simulink simulations and laboratory experiments. Their results indicate that increasing train speed and separation gap strengthen airflow cooling, which increases arc dissipation power and significantly raises both ignition and extinction voltage peaks. Using a rotating-wheel test rig with series inductance, Gabriella Crotti et al. [7,8] generated controllable DC pantograph–catenary arcs and recorded their stochastic characteristics and volt-ampere properties at an equivalent speed of 139 km/h. Surajit Midya et al. [9,10,11], utilizing a specialized wheel-based apparatus capable of simulating varying gaps and multidimensional pantograph motion, showed that parameters such as traction current, speed, and power factor substantially influence the physical properties and DC components of AC and DC arcs. Wang, Wei et al. [12,13] constructed an experimental platform capable of simulating multidimensional pantograph motion, and through the combined use of high-speed imaging and infrared thermography, identified the morphological evolution and electrical characteristics of pantograph–catenary arcs, providing essential experimental data for simulation research. Based on arc-blowing theory, Lu, Chen et al. [14,15] proposed a modified Habedank model and concluded that higher train speeds increase arc dissipation power, while larger separation gaps or arc lengths lead to increased arc voltage. Scholars including Kaisheng Peng, Zhigang Liu, Fei Lin, Zhou Hongyi, Huang Xiang, and Chen Yu subsequently applied the Habedank arc model to electrical sectioning, phase separation, and train-network scenarios of traction power supply systems, investigating the variations in arc energy, voltage, current, and arc length under different conditions, as well as their impacts on traction voltage, current, and harmonic behavior [16,17,18,19,20,21].

The generation of a pantograph–catenary arc is essentially a transient process in which the inter-electrode contact state evolves from a mechanical connection to a plasma channel. From a theoretical perspective, when the pantograph strip and the contact wire separate due to fluctuations in current collection quality, the contact pressure drops sharply, leading to a surge in contact resistance. This creates a localized high-temperature thermal effect at the contact point, triggering thermionic and field emissions from the electrode surface, which subsequently break down the gap medium to form an arc. Currently, the Cassie and Mayr models, along with their modified versions, are widely adopted in academia to describe this energy balance process; specifically, the persistence of the arc depends on the dynamic equilibrium between the input power and the dissipation power (including heat conduction, convection, and radiation). Given that metro DC power supply systems are characterized by high current and low voltage, their arc evolution differs from classic micro-current discharges. This process often bypasses the distinct glow-discharge stage and transitions directly into a stable arc discharge phase. Therefore, an in-depth investigation into the influence of horizontal running speed and electrical parameters on arc energy (light intensity) and arcing duration under dynamic separation conditions can provide a crucial theoretical basis for revealing the extinction mechanism of high-current pantograph–catenary arcs and optimizing current collection quality in metro systems.

To examine the dynamic behavior of separation arcs within rigid catenary systems, an experimental framework was established, integrating a pantograph, contact-wire electrodes, and a linear motor. This study prioritizes the evolutionary properties of the arc, specifically evaluating fluctuations in voltage and current, alongside variations in luminosity and discharge duration.

2. Experimental Setup

An experimental platform was constructed using a contact wire, a carbon strip, and a linear motor. The contact wire used in the experiments is a CTSM150 copper–tin alloy conductor, while the carbon strip is a metallized carbon strip identical to that used on CRH380A EMUs. The contact wire is shown in Figure 1c. For laboratory convenience, a 15 cm section of the carbon strip was cut and perforated for installation, serving as the simulated pantograph in the experimental platform. The carbon strip is shown in Figure 1d. The equivalent circuit diagram and the experimental apparatus are shown in Figure 2. The various components of the experimental apparatus are shown in Figure 1.

To simulate pantograph–catenary separation during train operation, a horizontal–vertical motion mechanism was adopted, allowing the pantograph to move horizontally relative to the contact wire while also achieving vertical separation. This configuration more closely reflects the actual operating conditions of railway trains and enhances the stability of the test carriage at different motion speeds [22]. In this study, vertical separation of the carbon strip alone is referred to as static separation, whereas simultaneous vertical separation and horizontal motion are defined as dynamic separation. A schematic diagram of the experimental apparatus is shown in Figure 3.

Since the experimental voltage is relatively low, the voltage signal was measured using a low-voltage probe to capture the pantograph–contact wire voltage transmitted to the ground. The current waveform was obtained by measuring the voltage across a shunt resistor and calculating the corresponding current. The optical detection module employed a Hamamatsu high-speed photodiode detector (model S5971) (Hamamatsu, Beijing, China) coupled with an optical fiber. The optical probe was positioned near the arc region to collect light signals, which were transmitted through the fiber to the photosensitive surface of the photodiode. The detector converted the received light signal into an electrical signal, which was recorded by an oscilloscope, thereby transforming the variations in arc luminosity into changes in voltage. The experimental detection apparatus is shown in Figure 4.

3. Arc Characteristics of Pantograph–Catenary Separation Discharge

Under experimental conditions with a DC supply voltage of 24 V, a current of 60 A, and a resistive load of 0.4 Ω, the discharge phenomenon and corresponding waveforms during pantograph–catenary separation were recorded. The vertical separation speed was set to 5 mm/s.

3.1. Voltage and Current Variation During Pantograph–Catenary Separation Arcing

3.1.1. Voltage and Current Variation During Static Pantograph–Catenary Separation Arcing

The discharge phenomenon and the corresponding voltage and current waveforms are shown in Figure 5.

At the moment of pantograph–catenary separation, an arc is generated between the two electrodes. During arc ignition, the voltage exhibits a steep rise, reaching 16 V (rising from 0 V to 16 V instantaneously). Following this initial surge, the voltage recedes to approximately 10 V, subsequently undergoing a steady rise as the increasing gap during pantograph separation sustains the arc. Upon the moment of arc extinction, a final abrupt transition occurs, with the voltage escalating sharply until it reaches a supply level of 24 V. The arc current follows an opposite trend to the arc voltage. At ignition, the current first drops sharply and then rebounds. As the arc continues to burn, the current gradually decreases, and at the moment of arc extinction, it abruptly falls to 0 A.

3.1.2. Voltage and Current Variation During Dynamic Pantograph–Catenary Separation Arcing

With a vertical separation speed of 5 mm/s for the carbon strip and a horizontal running speed of 200 mm/s along the contact-wire direction, the voltage and current waveforms of the separation arc were obtained, as shown in Figure 6.

The overall dynamic separation discharge process is similar to that of static separation. The main difference lies in the behavior during the stable discharge phase: in dynamic separation, the arc voltage initially increases slowly and then gradually decreases, while the arc current first decreases slowly and then rises to a steady level. At the beginning of arc stretching, repeated discharges occur due to unstable contact conditions between the pantograph and the contact wire.

3.2. Electrical Characteristics of Pantograph–Catenary Separation Arcs

The evolution of an arc is a complex physical process that is often modeled as a nonlinear resistance. The static arc resistance and arc power, obtained from the measured arc voltage and current data, are shown in Figure 7 and Figure 8.

Prior to the detachment of the pantograph–catenary interface, the narrowing contact area between the contact wire and carbon strip triggers an escalation in contact resistance. This leads to an elevated current density at the interface, which generates significant thermal energy. At the precise moment of separation, sharp peaks are observed in both arc resistance and power. Within a minimal timeframe, thermionic and strong-field emissions cause a rapid surge in plasma density between the electrodes, initiating the arc. During this initial phase, the narrow gap and intense electric field result in vigorous arc combustion. As the distance increases toward the point of extinction, the electric field strength diminishes, causing the arc resistance to approach infinity while the arc power undergoes a continuous decline.

4. Experimental Study on Arc Duration and Arc Luminosity During Pantograph–Catenary Separation

4.1. Light Intensity Characteristics of Discharge During Pantograph–Catenary Separation

4.1.1. Arc Light Intensity Characteristics Under Static Pantograph–Catenary Separation

The measured arc luminosity and the corresponding arc voltage and current waveforms are shown in Figure 9.

By comparing the arc current waveform on the left side of Figure 7, it can be seen that the acquisition timing of the arc luminosity corresponds to the measured arc current waveform. At the moment of arc initiation, the luminosity exhibits a sharp peak. As the pantograph and contact wire gradually separate, the arc luminosity fluctuates significantly during the stable discharge stage and finally drops abruptly when the arc extinguishes.

4.1.2. Arc Light Intensity Characteristics During Dynamic Pantograph–Catenary Separation

The dynamic separation arc light-intensity test was carried out with a pantograph–catenary vertical separation speed of 5 mm/s and a carbon strip horizontal moving speed of 200 mm/s. The voltage, current and light-intensity waveforms of the dynamic decontact arc are shown in Figure 10.

The overall waveform of the arc light intensity during dynamic pantograph–catenary separation is similar to that of the static separation arc. However, the light intensity of the dynamic separation arc is more stable, increases slowly as the gap widens, and is much lower than that of the static burning arc.

4.2. Influence of Voltage and Current Levels on the Burning Duration and Light Intensity of Decontact Arcs

4.2.1. Influence of Current Level on the Burning Duration and Light Intensity of Decontact Arcs

The supply voltage was maintained at DC 24 V, and the separation speed was maintained at 5 mm/s. By inserting resistors of different values, tests on separation-arc light intensity and burning duration were carried out with loop currents of 20 A, 40 A, and 60 A. The phenomena observed at the same instant for different groups are shown in Figure 11.

During pantograph–catenary separation, intense arc burning occurs, and the severity of the burning phenomenon tends to increase with the current level. When the current rises to 60 A, pronounced metal spattering caused by the high temperature is observed. The corresponding data waveforms are shown in Figure 12.

As the current level increases, the arc burning duration gradually becomes longer, erosion becomes more severe, and the light intensity recorded by the sensor increases. Compared with lower current levels, at 60 A, the arc light intensity exhibits a sharp peak at the initial stage of arc generation, which coincides in time with the occurrence of metal spattering. The comparison of arc burning time and light intensity is given in

Table 1. .Statistical data on arc burning time and arc light intensity during current variation.

	60 A	40 A	20 A
Arc burning time (Mean ± SD)	175 ± 11.8 ms	125 ± 12 ms	55 ± 11 ms
Arc optical intensity (Mean ± SD)	13 ± 1.89 V	7.5 ± 1.59 V	1.5 ± 0.71 V

.

When the current decreases, both the arc duration and the arc light intensity decrease. The reason for this is that as the current flowing through the electrodes drops, the associated Joule heating is reduced, the thermionic emission mechanism is weakened, and consequently, the severity of arc burning diminishes.

4.2.2. Influence of Voltage Level on the Burning Duration and Light Intensity of Decontact Arcs

The current was kept at 40 A, and the static separation speed was maintained at 5 mm/s. Separation-arc light intensity and burning duration tests were carried out at voltage levels of 24 V, 20 V, and 16 V. The phenomena observed at the same instant for different groups are shown in Figure 13.

During pantograph–catenary separation, intense arc burning occurs. As the voltage level increases, the severity of the burning phenomenon tends to increase, and the luminous area of the arc expands significantly. The corresponding data waveforms are shown in Figure 14.

The duration of the arc is profoundly influenced by fluctuations in the pantograph–catenary voltage; for instance, a reduction in voltage from 24 V to 16 V results in a sevenfold decrease in arc persistence. Furthermore, experimental data indicates that higher levels of voltage lead to a conspicuous surge in recorded light-intensity values. A detailed comparative analysis of these burn durations and optical intensities is presented in

Table 2. Statistical data on arc burning time and arc light intensity during voltage variation.

	24 V	20 V	16 V
Arc burning time (Mean ± SD)	125 ± 12 ms	70 ± 16 ms	22 ± 8.79 ms
Arc optical intensity (Mean ± SD)	7.5 ± 1.59 V	1.2 ± 0.6 V	0.3 ± 0.71 V

.

When the voltage decreases, both arc duration and arc light intensity decrease. The reason for this is that as the voltage applied across the electrodes drops and the electric field strength in the same gap is reduced, thereby weakening field emission and reducing the severity of arc burning. Comparing Table 2 with Table 1, for the same proportional increase or decrease in voltage and current, the influence of voltage on arc burning time and light intensity is much greater than that of current.

4.3. Influence of Train Operating Speed on Arc Duration

At a supply voltage of DC 24 V, a loop current of 60 A, and a pantograph–catenary vertical separation speed of 5 mm/s, the horizontal running speed of the carbon strip was varied to obtain the relationship between the arc burning time during pantograph–catenary separation and the horizontal running speed of the pantograph. The results are shown in Figure 15.

The arc burning time during dynamic pantograph–catenary decontact is shorter than that during static decontact; the burning time tends to decrease as the horizontal running speed increases. The reason for this is that in static decontact, the arc root position remains essentially unchanged, whereas during dynamic separation, the arc root moves with the horizontal displacement of the carbon strip. The horizontal motion of the strip enhances air convection, and thus, increases the heat dissipation power: the higher the horizontal running speed, the greater the heat dissipation power and the shorter the arc burning time.

5. Conclusions

In this paper, an experimental platform for pantograph–catenary decontact arcs was built in which the carbon strip has two degrees of freedom, both horizontal and vertical. The DC pantograph–catenary arc electrical characteristics, arc light-intensity characteristics, and burning duration were investigated. The experiments show the following results:

(1)

The arc burning time is longer when the contact wire acts as the cathode than when the carbon strip acts as the cathode.

(2)

In a DC pantograph–catenary system, when the voltage and current levels are increased or decreased by the same factor, the influence of the voltage level on the arc light intensity and burning duration during pantograph–catenary separation is greater than that of the current level.

(3)

At the instant of pantograph–catenary separation, the voltage exhibits a significant abrupt rise before receding to a stable arcing voltage. Conversely, the current shows a trend of a sharp drop followed by recovery at the moment of arc ignition, eventually decreasing to 0 A upon arc extinction.

(4)

At the moment of separation, due to the extremely small gap distance and high electric field intensity, both arc resistance and arc power exhibit pronounced “peak phenomena.”

(5)

As the offline gap increases, the electric field intensity decreases, causing the arc resistance to increase gradually towards infinity. Meanwhile, the arc power continues to decline amidst fluctuations until the arc is extinguished.

(6)

Compared to static separation, the voltage during the stable arcing phase of dynamic separation (with horizontal running speed) displays a trend of a slow increase followed by a gradual decrease, while the current slowly decreases before rising steadily.

(7)

The voltage required to maintain the arc channel is much lower than the air breakdown voltage, which explains why the arcing phenomenon is more persistent than simple air dielectric breakdown.