Development of a sensor for corona current measurement under high-voltage direct-current transmission lines

Abstract

A sensor-based distributed measurement system is designed for collecting and monitoring the corona current under the extremely high-voltage direct-current transmission lines. It is highly significant for researching the corona characteristics of extremely high-voltage direct-current transmission lines to reduce the corona effect, operating loss, and environmental impact. A new shared memory strategy has been studied based on field-programmable gate array and ARM. Based on field-programmable gate array technology, the fiber-optic transmission scheme for wide-frequency corona current high-speed signal has been set up to ensure the safety and the reliability of corona current measurement system. The proposed system has been used in China’s state grid high-voltage direct-current test base and the in situ power transmission projects. Based on the experimental results, the proposed measurement system demonstrates that it can adapt to the complex electromagnetic environment under the transmission lines and can accomplish the accurate, flexible, and stable demands of the electric-field measurement.

Keywords

Sensor corona current high voltage direct current data acquisition

Introduction

The electromagnetic environment (such as audible noise, radio interference, and total electric field) caused by corona discharge not only restricts the conductor selection and structural design of transmission line parameters but also often brings environmental issues. With the improvement of the energy saving and environmental protection requirements of the transmission line, only relying on measurement and analysis of corona effect, such as audible noise, radio interference, and corona loss, has not met the engineering requirements to study the electromagnetic environment of transmission lines. We need to continue in-depth study of the problem of corona discharge, such as the relationship between corona current and corona effect aspects.

The basic mechanism of the corona discharge can be obtained by measuring the corona current. However, because of the weak signal and wide bandwidth of the corona current, the value of corona current is difficult to be measured accurately. The traditional corona current measurement system bandwidth is low (usually 2 MHz or less), and measuring performance cannot meet the voltage level and gradually increased research needs. Only a few papers have reported work on the measurement of the corona current by far. In Greason et al.¹ and Kucerovsky et al.,² the authors used an optically decoupled data acquisition module to measure the corona current generated in a pin-plane gap, but its maximum sampling rate is 16 kHz and can only cover a bandwidth of 8 kHz. In Liu et al.,³ a corona loss monitoring system is proposed. It used an optical power supply electronic current transformer to acquire the corona current signals and transmitted the signals by optical fibers. This method can measure current in ultra-high-voltage (UHV) environment safely. But, the measuring bandwidth of the system is still limited to 7.5 kHz. In Wang and Zhang,⁴ a commercial low-inductance resistance is used to measure the corona current. After improving the sampling rate of the data acquisition system, the signal acquisition width can reach up to 4.5 MHz. However, besides the sampling rate, the distributed parameters of the resistance sensor also affect the integrity of the corona current to a great extent, which is less well considered in the present methods. Moreover, relevant studies on the corona current measurement described in Mihailovic et al.,⁵ Bull et al.,⁶ Zhang et al.,⁷ and Kucerovsky et al.⁸ are mostly carried out under a test voltage that is less than 500 kV. The corona current data available mostly have been measured in the experimental environment, and the data are generally not verified in high-voltage direct-current (HVDC) transmission lines. With the development of the UHV transmission technology, the anti-breakdown requirement of the measurement system becomes much more rigorous under the UHV environment, and it is necessary to find a new method to measure the UHV corona current.

The object of this article is to describe the development of a new sampling resistance sensor for corona current measurement in a wide band under UHV environment which the impedance could keep invariable when the frequencies vary between the ranges of 0 and 10 MHz. To start with, a new structure, which is specially designed according to the UHV dielectric breakdown and wide-band measuring requirements of the sensor, is introduced. Then, the synchronous dynamic random-access memory (SDRAM) controller and the related first input first output (FIFO) settings accordingly have been introduced during the development of the wide-frequency corona current measurement system, and the large-capacity cache problems involved in the wide-frequency corona current data acquisition system have effectively been solved based on Field-Programmable Gate Array (FPGA) and advanced RISC machines (ARM) technologies. Finally, the high-voltage discharge and frequency characteristic tests prove the UHV dielectric breakdown performance and the wide-band measuring characteristics of the designed sensor, respectively. The test results in an actual ultra-high-voltage direct-current (UHVDC) transmission line demonstrate that the sensor can accurately measure the corona current under the UHV environment.

Theoretical overview

Principle of sensor measurement

If there is corona phenomenon generating in transmission line, it will have corona current in it. Corona current will cause power loss. In general, corona current is small and unit length line corona loss is also less severe, but the transmission line is so long that corona loss cannot be neglected. Corona-caused power loss can be calculated by the corona current, and corona-equivalent circuit is essentially equivalent to a nonlinear resistor–capacitor (RC) circuit as shown in Figure 1.

Figure 1.

The equivalent circuit model of corona discharge.

In Figure 1, V_on is the corona onset voltage, C₁ is the geometric capacitance of the conductor configuration, C₂ is the additional nonlinear capacitance due to corona, C₃ is the capacitance representing the charge loss to the air, and G is the nonlinear conductance representing corona loss. The measuring equipment is placed between the power generation and the transmission conductor. When the voltage of high-voltage DC generator is below the corona onset voltage V_on, C₂ = C₃ = G = 0. Only capacitance charging current with the wire-to-ground and a small amount of leakage current exist. When the voltage of high-voltage DC generator is above the corona onset voltage V_on, the nonlinear conductance G will increase rapidly, while C₂ and C₃ increase very slowly, and subsequently, large corona current appears with increasing voltage. The corona current on the conductor is equal to the current through the resistor minus the leakage current of insulators. The quantity of leakage current is negligible compared with the corona current. Therefore, only corona current is measured through the high-frequency sensor and its data are recorded.

Dielectric breakdown performance

The surface electric-field intensity of the sensor working under the UHV environment should be lower than the dielectric breakdown intensity of air. According to the Peek equation,^9,10 the breakdown intensity of the air around the conductor can be given by

E_{o n} = 30 m δ (1 + \frac{K}{\sqrt{r δ}})

(1)

where E_on and K are the empirical constants depending on the nature of the applied voltage; r is the radius of the conductor (cm); m is the roughness factor of the conductor, which is usually between 0.75 and 0.85; and δ is the atmospheric correction coefficient given by

δ = \frac{0.386 P}{273 + t}

(2)

where t is the atmosphere temperature (K) and P is the atmospheric pressure (mmHg).

When t is 298 K, the breakdown intensity E_on is 30 kV/cm. So, the sensor designed surface electric-field intensity should be lower than 30 kV/cm when the voltage reaches up to 800 kV on the transmission lines. Meanwhile, other environmental factors, such as humidity and temperature, also affect the breakdown intensity to some extent.^11–15 However, comparing with the air density, the effects of these factors are not significant.

Corona discharge pulse current features

To capture the corona current characteristics of the transient signal correctly and avoid the frequency aliasing, the overall scheme of the broadband domain corona current measurement system was formulated, combined with the time-frequency characteristic of the corona current pulse signal and system requirements. Corona discharge process can produce high-frequency pulse current, which contains many higher harmonic that can cause interference to radio communications and other signal.

Corona discharge has distinction for different electrode polarity. In the negative corona, after the air ionization, electron is excluded from electrodes, and the positive ions will be gathered on electrode surface. When the electric field continues to strengthen, positive ions will be sucked into the electrode to form a pulse current and flow to both sides of the wire. With negative charge continues to spread into the surrounding space, electrode surface electric-field intensity increases, and the next corona ionization process will happen and next pulse current. A series of pulse group will appear for the circulation. When normal polarity corona occurs, negative ion or electron is sucked into the electrode, and the same repetitive pulsed corona current will be formed.

With corona discharge,^16–18 a series of impulse current will inject transmission lines. Pulse current is double-exponential distribution with time.^18–21 It can be represented by the following formula

i (t) = K I_{p} (e^{- α t} - e^{- β t}), t \geq 0

(3)

where I_p is the corona current amplitude (mA); K, α, and β are the constant, corona current pulse rise time, and fall time depends on α and β, respectively. The positive and negative corona current time-domain expressions are

i (t) = 2.335 I_{p} (e^{- 0.01 t} - e^{- 0.0345 t})

(4)

i (t) = 1.33 I_{p} (e^{- 0.019 t} - e^{- 0.285 t})

(5)

The frequency domain expression of pulse current is

| I (j ω) | = \frac{k I_{p} (β - α)}{\sqrt{(α^{2} + ω^{2}) (β^{2} + ω^{2})}}

(6)

The time-domain and frequency-domain characteristics of positive and negative corona pulse current are shown in Figure 2(a) and (b), respectively.

Figure 2.

Corona pulse current features: (a) time-domain characteristics and (b) frequency characteristics.

Through the corona current, time–frequency characteristic can be known that positive corona pulse current has larger amplitude than the negative and has a longer duration. Frequency domain feature reflects that pulse current has a wide frequency range. It also illustrates the necessity to research corona effect from the perspective of corona current.

Design of the corona current sensor

The data acquisition module of UHV broadband domain corona current measurement system is different from the general data acquisition system. The common data acquisition system only emphasizes the validity and accuracy of the signal collection results, and do not require high requirements for the complexity of the environment and the instantaneity of the acquisition signal. Due to the wide domain, corona current data acquisition system is mainly used for corona current pulse signal acquisition under UHV environment, and it has very high requirements for the data sampling rate and data sampling accuracy. So, the higher performance requirements are put forward. High-speed and high-precision data acquisition is the key technology for corona current data acquisition system. The corona current pulse signal has the characteristics such as short rise time, high pulse amplitude, and rich frequency composition. Therefore, higher and stricter requirements have been proposed for data storage of the corona current pulse signal. In addition, the flexible high-capacity storage and the cache settings is an important consideration.

Corona current measurement system framework

In this article, digital broadband domain corona current measuring system is introduced for data acquisition, storage, and transmission. The corona current distribution measuring system structure is shown in Figure 3.

Figure 3.

Scheme of HVDC corona current distribution measurement system.

The measurement system is mainly composed of high-voltage (HV) local area and safe location measurement area. The HV local area, which contains different space position distribution of corona current sensor, is used to complete data collection and storage. In the HV local area, the high precise no-inductance resistor is applied as the current sensor.^22–25 The resistor is connected in series with the bundle conductors in the test lines. The optically powered electronic current transformer is used to convert and transmit the current signal. In addition, the signal is acquired by the data acquisition card. Independent power supply is used to provide energy for electronic components within the system. The acquired data are sent by the photoelectric conversion unit and transmitted to safe location measurement area through outdoor optical fiber. The safe location measurement area is used to complete the data reception work and implement communication with upper computer. The collected corona current data can be processed and saved in computer terminal.

Design of the block diagram of the data acquisition system

As the corona current data acquisition system is mainly used for corona current pulse signal acquisition under extremely high-voltage direct-current (EHVDC) environment, the accuracy of the data sampling is very high, so the higher performance requirements are required. High-speed and high-precision data acquisition for corona current signal acquisition is the key technology of data acquisition system. So, data acquisition system becomes the core of the corona current measurement system. It includes data acquisition module, large-capacity data storage module, and data transmission module. Figure 4 is the block diagram of the data acquisition system. Data acquisition module realizes the corona current signal conditioning, trigger and filtering, and to achieve parameter acquisition process control and the upper machine operation; data storage module uses to realize high-speed data caching and large-capacity data storage; data transmission module is mainly to realize digital transmission of corona current signal by optical fiber.

Figure 4.

Block diagram of the data acquisition system.

Parallel alternating sampling structure is used for broadband domain corona current measurement system. The input analog signal is sent to multiple channels at the same time after signal conditioning and then multiple analog-to-digital converter (ADC) chips are used for parallel sampling. ADC converter works according to their sampling clock, and the sampling rate of the whole system is the sum of the ADC sampling rate, so as to realize the function of high-speed data acquisition.

In the broadband domain corona current measurement system, FPGA is the main controller. The ADC transmits data collected to the FPGA internal cache. FPGA can set ADC data conversion module and data output sequential circuits, so as to realize the control of the ADC and data acquisition.

In the process of high-speed data acquisition system storage, both storage capacity and storage rate are two important aspects that should be considered. The application of large-capacity memory SDRAM can greatly improve the data storage capacity. High-speed data caching module consists of asynchronous FIFO. High-speed sampling data from the ADC are cached through high-speed data caching module. It can realize the striding clock domain data processing. The use of FIFO between high-speed ADC and SDRAM can improve the seamless connection between the ADC and mass storage. The collected data are stored in the SDRAM memory through the data storage module and are transmitted through the USB interface in data transmission module. In order to solve the problem of inconsistency between memory chips’ reading/writing speed and transmission interface module clock frequency in the process of data transmission, read data cache also uses the asynchronous FIFO module. The proposed circuit of data acquisition system is shown in Figure 5.

Figure 5.

PCB of the data acquisition system.

Software design of the measurement system

Software’s design needs to consider the user’s demands. The software is mainly for the operator and data analyst. They have different demands for the software. The operators are required for setting of parameters and real-time data collection, as well as real-time surveillance for the equipment running conditions. For data analysts, relevant analyses and processing will be done to the data stored in the database after finishing the measurement. Therefore, software structure is built based on the demand analysis of users, as shown in Figure 6.

Figure 6.

Structure chart of the measurement system software.

Measuring software consists of control layer and data layer. Control layer, locating at the upper layer, which is composites of interaction surface of human and machine, is the core layer of the two, supplying operations to the users so as to realize the real-time control to the corona current. Data layer, locating in the lower layer, which is oriented for data analysis personnel, realizes the real-time storage of data.

Based on the proposed structure, the measurement software used in this system is developed as shown in Figure 7. Main interface of software consisted of seven parts: (1) power control area, used to control on and off of the equipment; (2) acquisition control area, controlling the data acquisition work way; (3) parameter setting area, setting up the data acquisition parameters; (4) curve display, showing real-time distribution of electric field under transmission lines; (5) data storage box, showing the data storage paths; (6) system status display, monitoring the battery status; and (7) menu and tool bars, containing all operations for the users.

Figure 7.

Software interface of the measurement system.

Design of the overall structure

High-voltage prevention design is required to protect the high-frequency resistance sensors. If the resistance sensor breaks down, collecting device will be directly accessed into the test line to be damaged. Therefore, four 400 Ω non-inductive resistors are used in the front of the equipment with parallel manner for the high-frequency signal acquisition. When anyone of them breaks down, there will be other parallel resistances in the circuit to ensure the safety of the collecting device.

The corona current sensor is working in a complex electromagnetic environment. The performance of the sensor will be affected by corona discharge. Therefore, electromagnetic shielding must be designed to restrain the interference of electromagnetic field. When circuit is wrapped with metal shell which separated it from electric field, there is no electric field in the shell no matter whether the shell is grounding. The material of the shielding box for the measurement system is aluminum with 10 mm thickness. The interference will be reduced and the reliability will be improved for corona current measurement equipment.

The prototype of the measuring device is shown in Figure 8. Design and application of grading rings are used to avoid corona discharge at both ends of the sensor. The determination of grading ring parameters is given in reference. It can make the sensor surface electric-field intensity below 20 kV/cm under 800 kV extremely high-voltage (EHV) environment and meet the needs of practical engineering.

Figure 8.

Measurement equipment: (a) prototype equipment and (b) general structure of the equipment.

Sensor performance evaluation

Impedance characteristics test

Corona current measuring equipment collects corona current signal on the conductor through sampling resistance sensor. The choice of sampling resistance depends on the change range of the conductor current in the test process. In order to ensure accurate measurement of current signal and retain a certain measurement margin, a large number of experiments were tested. In the end, four 400 Ω high-power precision non-inductive resistances were chosen and collected the corona current signal in the parallel form.

It is evident from Figure 9 that the impedance of the sensor almost keeps constant when the frequency is less than 30 MHz. So, the sensor is able to measure the corona current whose bandwidth is within 30 MHz. The sensor impedance changes rapidly when the test frequency is above 30 MHz as a result of lumped capacitance and inductance. Fortunately, the corona current signals whose frequencies are over 30 MHz occupy only a small portion, which can be filtered beforehand in the signal processing stage.

Figure 9.

The sensor impedance test result.

High-voltage discharge test

The corona current measurement system has been used in the China’s Electric Power Research Institute EHVDC test base. Corona current measurement equipments’ installation and distribution are shown in Figure 10. In the corona discharge test, 800 kV voltage was applied to high-voltage generator; corona discharge around the resistance sensor was observed using Ofil’s SuperB corona camera. Figure 11 shows the result of corona discharge of the resistance sensor under 800 kV EHVDC environments. It shows that the photon count is 0. Therefore, the sensor surface voltage gradient is not more than E_on, and the sensor meets the engineering design requirements.

Figure 10.

Measuring equipment distribution: (a) corona cage and (b) testing tower.

Figure 11.

Sensor high-voltage test results: (a) positive pole and (b) negative pole.

Experimental results

Test platform

To validate and test the proposed corona current measurement system, the same-tower double-circuit HVDC experimental lines were erected by the China Electric Power Research Institute at Beijing. The transmission line, which is shown in Figure 12, is 100 m long, and the model of the transmission line is LGJ-95/20 with four bundles of 40 cm spacing. The minimum distance from the ground to the conductor is 7 m and the separation distance between the positive pole and the negative pole is 6 m. The test meteorological parameters are shown in Table 1.

Figure 12.

Actual test lines.

Table 1.

Meteorological parameters.

Locations	Altitude (m)	Air pressure (kPa)	Temperature (°C)	Relative humidity (%)
Beijing	27	101.3	21.3	26

Corona current measurement at test HVDC transmission line

The test voltage applied to the positive pole was 0, 100, 125, 150, 175, 200, 225, 250, 275, and 300 kV. Correspondingly, the voltage applied to the negative pole was 0, −100, −125, −150, −175, −200, −225, −250, −275, and −300 kV. After using bipolar voltage mode, the change of V-I characteristics of corona current is shown in Figure 13. Corona current increases gradually and exponentially with the rise of the line voltage grade; the positive corona current strength is slightly higher than the negative under the same voltage grade.

Figure 13.

The change of V-I characteristics of corona current.

The complexity and diversity of noise in corona current measurements of HVDC transmission lines leads to challenges in estimating corona performance. Using an offline noise suppression method for HVDC corona current measurements is that given that the corona current and background noise processes coexist within the same frequency band, we develop a novel multifaceted filtering approach for HVDC corona currents.

Figures 14 and 15 show background signal under 0 kV and corona current signal under positive 300 kV on HVDC test lines in Beijing, respectively. By comparison of Figures 14 and 15, it can be found that relatively high narrowband signals of 3.325, 7.538, 7.987, 9.005, 9.337, 9.675, 10.351, 11.253, and 12.837 MHz exist both under 0 and +300 kV. After long-term monitoring, these signals which have been confirmed are radio signals and broadcast signals from space; so, they can be removed using filter. However, collected corona current signals in addition to containing the narrowband interference signals still contain a lot of interference signals such as pulse interference and random noise; so, further data processing is needed to extract the effective corona current signals from the background noise. The corona current signal after data processing is shown in Figure 16. A lot of noise and narrowband signal in original signal have been effectively filtered, and noise reduction effect is obvious. It can promote further research for HVDC corona characteristics.

Figure 14.

Background signal: (a) time domain and (b) frequency domain.

Figure 15.

Corona current signal under positive 300 kV: (a) time domain and (b) frequency domain.

Figure 16.

Corona current signal under positive 300 kV after data processing: (a) time domain and (b) frequency domain.

Conclusion

In this article, a broadband corona current measurement system to measure the distributed corona current under EHVDC transmission lines for fulfilling the demands of extensibility, flexibility, and stability has been presented. The proposed system comprises HV local area and safe location measurement area. Data acquisition system in HV local area is the core of the corona current measurement system based on FPGA and ARM. Based on FPGA technology, the fiber-optic transmission scheme for broadband corona current signal has been set up to ensure the safety and the reliability of corona current measurement system. Sensor’s performance evaluation shows the sensor has flexible and robust performance in complex environments under EHVDC transmission lines. The experimental results under in situ HVDC transmission lines demonstrate that the proposed measurement system can adapt to the complex electromagnetic environment and accurately collect the corona current data.

Footnotes

Academic Editor: Yonghe Liu

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by National Natural Science Foundation of China (61273165).

References

Greason

Kucerovsky

Bulach

. Non-invasive optoelectronic system for measurement of electrostatic discharge (ESD) induced phenomena. In: Thirtieth industry applications conference, Orlando, FL, 8–12 October 1995, pp.1264–1270. New York: IEEE.

Kucerovsky

Greason

Doyle

. Optical decoupling in measurements of corona currents. In: IAS annual meeting, 1999. Thirty-fourth industry applications conference, Phoenix, AZ, 3–7 October 1999, pp.756–763. New York: IEEE.

Liu

You

Wan

. UHV AC corona loss measurement and analysis under rain. In: Proceedings of the 9th international conference on the properties and applications of dielectric materials, Harbin, China, 19–23 July 2009, pp.130–133. New York: IEEE.

Wang

Zhang

The measurement method for corona discharge current under high-voltage environment. IEEE T Instrum Meas 2008; 57(8): 1786–1790.

Mihailovic

Petricevic

Stojkovic

. Development of a portable fiber optic current sensor for power systems monitoring. IEEE T Instrum Meas 2004; 53(1): 24–30.

Bull

Jaeger

NAF

Rahmatian

A new hybrid current sensor for high-voltage applications. IEEE T Power Deliver 2005; 10(1): 32–38.

Zhang

Qin

. A new electro-optic hybrid current-sensing scheme for current measurement at high voltage. Rev Sci Instrum 1999; 70: 3755–3758.

Kucerovsky

Greason

Doyle

. Data acquisition system for the measurements of corona currents. In: Thirty-third industry applications conference, St. Louis, MO, 12–15 October 1998, pp.2014–2018. New York: IEEE.

Peek

FW.

Dielectric phenomena in high voltage engineering. New York: Wiley, 1929.

10.

Tan

Modified peek formula for calculating positive DC corona inception electric field under variable humidity. IEEE T Dielect El In 2012; 19(4): 1377–1382.

11.

Cui

Wang

Yuan

. Relative localization in wireless sensor networks for measurement of electric fields under HVDC transmission lines. Sensors 2015; 15(2): 3540–3564.

12.

Maruvada

PS.

Influence of ambient electric field on the corona performance of HVdc transmission lines. IEEE T Power Deliver 2014; 29(2): 691–698.

13.

Modeling and control of extended multiterminal high voltage direct current systems with three-wire bipole structure. Int T Electr Energy Syst 2015; 25(10): 2036–2057.

14.

Wang

Kundur

Yuan

. Noise suppression of corona current measurement from HVdc transmission lines. IEEE T Instrum Meas 2015; 65(2): 264–275.

15.

Fang

Cui

Zhou

. Impact factors in measurements of ion-current density produced by high-voltage DC wire’s corona. IEEE T Power Deliver 2013; 28(3): 1414–1422.

16.

Merbahi

Yousfi

Gardou

JP.

Electric and spectroscopic analysis of surface corona discharges in ambient air and comparison with volume corona discharges. IEEE T Plasma Sci 2012; 40(4): 1167–1176.

17.

Liu

Effects of wind condition on hydrophobicity behavior of silicone rubber in corona discharge environment. IEEE T Dielect El In 2016; 23(1): 385–393.

18.

Bian

Chen

. Influence of aged conductor surface conditions on AC corona discharge with a corona cage. IEEE T Dielect El In 2011; 18(3): 809–818.

19.

Cui

. Comparison between the audible noises generated from single corona source under DC and AC corona discharge. CSEE J Power Energ Syst 2015; 1(3): 23–30.

20.

Lü

You

Liu

. AC conductors’ corona-loss calculation and analysis in corona cage. IEEE T Power Deliver 2012; 27(2): 877–885.

21.

Asano

Yatsuzuka

Yamaki

DC corona discharge of a metal filament particle within parallel-plate electrodes. IEEE T Ind Appl 2000; 36: 87–92.

22.

Depari

Flammini

Sisinni

Measurement of resistance and capacitance of MOX sensors with high sampling rate. IEEE T Instrum Meas 2012; 61(9): 2483–2491.

23.

Yuan

Qinghua

Liu

. Development and application of high-frequency sensor for corona current measurement under ultra high-voltage direct-current environment. IEEE T Instrum Meas 2012; 61(4): 1064–1071.

24.

H-f

Qian

Dong

. Development of multi-band ultra-high-frequency sensor for partial discharge monitoring based on the meandering technique. IET Sci Meas Technol 2014; 8(5): 327–335.

25.

Zheng

. A tripole HVDC system based on modular multilevel converters. IEEE T Power Deliver 2014; 29(4): 1683–1691.