1

2006 IEEE PES Transmission and Distribution Conference and Exposition Latin America, Venezuela

A New Measurement Method of the Dynamic Contact Resistance of HV Circuit Breakers

M. Landry, A. Mercier, G. Ouellet, C. Rajotte, J. Caron, M. Roy, and Fouad Brikci

Index Terms-- Circuit breaker testing, Circuit breakers, Contact mechanical factors, Contact resistance, Contacts, Timing, Time measurement, Time domain analysis, Time domain measurements, Test equipment.

developed over 10 years ago to assess the condition of the arcing contacts without dismantling the breaker. This method is no longer widely used since the interpretation of the resistance curve remains ambiguous. Previously published test results usually depicted several spikes [1-3] in the resistance curve which could be the result of a partial contact part during the contact movement. The following paper presents a new dynamic-contactresistance measurement method that has been validated by field tests which were performed on air-blast and SF6 gas circuit breakers. The new method is based on the breaker contact resistance measurement during an opening operation at low speed. After reviewing the characteristics of the dynamic resistance curve and the measuring system and parameters, the paper deals with relevant values that can be extracted from the resistance curve for detecting contact anomalies wear and/or misalignment. Finally, case studies are presented and test results are discussed. II. MEASURING SYSTEM AND SENSORS

HE design of modern high-voltage puffer-type SF6 gas circuit breakers is based on the switching of two parallel contact sets. First, the low-resistance silver-plated contacts or the main contacts are specifically designed to carry the load current without any excessive temperature rise. Second, following the main contact part, the tungsten-copper arcing contacts are finally opened, thus initiating arc quenching and current interruption. To assess the condition of the breaker contacts, the main contact resistance measurement is usually performed. However, the static resistance measured when the breaker remains in a closed position does not give any indication of the condition of the arcing contacts. To evaluate the latters condition, an internal inspection can be done, but timeconsuming and costly maintenance procedures must be followed in order to securely handle the SF6 gas and arc byproducts. It should be remembered that excessive arcingcontact wear and/or misalignment may result in a decrease of the circuit breakers breaking capacity. The dynamic contact resistance measurement (DRM) was
M. Landry, A. Mercier, G. Ouellet, C. Rajotte, J. Caron, M. Roy HydroQubec F. Brikci, Ph.D. Zensol Automation Inc.

I. INTRODUCTION

For dynamic contact resistance measurements (DRMs), three signals must be recorded: - the injected current (IDC) of at least 100 A in order to minimize relative noise level; - the voltage drop (VD) across the breaker contacts; - the breaker contact travel curve. Since the new DRM method presented in this paper will be performed during an opening operation at low contact speed when the breaker is off-line, some commercial acquisition units with the following features may be used: - 3 analog inputs with at least 12-bit resolution and appropriate range of voltage inputs; - a sampling frequency of 10 kHz; - a total acquisition time of 30-100 s; - connection to a portable computer for calculation of the instantaneous contact resistance (VD/IDC), data analysis and interpretation using dedicated software. Finally, the following sensors are required: - Hall-effect current sensor allowing accurate measurement of both the current amplitude and the abrupt current variation at the arcing contact part that corresponds to the complete breaker contact opening; - linear or rotary contact travel sensor depending upon the breaker technology.

1-4244-0288-3/06/$20.00 2006 IEEE

III. MEASURING PARAMETERS A. Closing operations DRMs during closing operations are not generally useful since the measurement must be performed during a transient state, i.e. from open to closed contacts. There are two main reasons why the measurement in this condition is impractical: - the abrupt resistance variation from infinity (open contacts) to the arcing contact resistance is difficult to measure, making the resistance level of the arcing contact difficult to detect; - the transient DC current at the moment of arcing contact touch generates undesired noise level and therefore jeopardizes the measurement. B. Opening operations at low contact speed DRM should be rather performed during opening operations at low contact speed ( 0.002-0.2 m/s). Fig. 1a shows superimposed typical resistance curves of two consecutive measurements at rated speed on break A (Table I). The two traces have been synchronized by superimposing instants of the main contact part which is identified as tm in the Fig. 1a graph. Note that no filtering has been applied.

contact resistance may be plotted as a function of the contact travel (section C).

b)
100

50 ms

Arcing contact part

R (micro-ohms)

At low speed Contact speed=0.2 m/s Main contact part tm t 0.15 m/s

Fig. 1b. Comparison of the dynamic contact resistance curves according to the conventional (at rated speed) and the new (at low speed) methods b) At low speed on break A

a)
500

0.5 ms

Arcing contact part

Resistance (micro-ohms)

At rated speed Main contact part tm No reproducible measurements

For break B (Table I), Fig. 1c depicts another DRM curve that was recorded at the rated contact speed. Several spikes can be observed. Moreover, it is absolutely impossible to identify the main contact part. The presumed main contact part is indicated based on other measurements at low contact speed. As for break A, it is anticipated that this phenomenon is caused by partial contact part due to high contact speed and acceleration. At low contact speed, the DRM curve is far smoother and the main contact part can be easily identified (Fig. 1d).

c) R (micro-ohms)

Arcing contact part At rated speed 500 2 ms Presumed main contact part

time

Fig. 1a. Comparison of the dynamic contact resistance curves according to the conventional (at rated speed) and the new (at low speed) methods a) At rated speed on break A

At the rated speed, it can be observed that the resistance curves are not reproducible from one test to another. Moreover, this phenomenon is more marked in the vicinity of the arcing contact part. During the validation test program, it was observed that this behaviour is completely random. On the contrary, for the same breaker A, Fig. 1b shows two dynamic contact resistance curves obtained at low contact speeds of 0.2 and 0.15 m/s. The two traces have also been synchronized by superimposing instants of the main contact part. Except for the fact that the curves exhibit different instants of the arcing contact part which are due to measurements at different contact speeds, the two resistance curves appear to be almost identical. To eliminate these time deviations resulting from the contact speed, the dynamic

t
Fig. 1c. Comparison of the dynamic contact resistance curves according to the conventional (at rated speed) and the new (at low speed) methods c) At rated speed on break B

d) R (micro-ohms)
500

5s At low speed

Arcing contact part

Contact speed= 0.002 m/s

Main contact part t

Fig. 1d. Comparison of the dynamic contact resistance curves according to the conventional (at rated speed) and the new (at low speed) methods d) At low speed on break B Relatively new fixed contact (F1)

It must be pointed out that partial contact part does not occur when a high current is interrupted since electromagnetic forces are exerted on the contacts, maintaining them together until final contact separation. Therefore, it is assumed that the low-speed DRM more adequately simulates the actual operating conditions of an in-service HV circuit breaker.

IV. PARAMETERS TO BE EXTRACTED FROM THE DYNAMIC RESISTANCE CURVE A. Contact wear algorithm A contact wear algorithm was developed for the new DRM method. Fig. 2b depicts the contact resistance curve for different contact sets of an HV air-blast circuit breaker: one relatively new fixed contact (F1) and four moving contacts in different stages of wear (Fig. 2a): a new contact (M1), a slightly worn contact (M2), a worn contact (M3) and a seriously damaged contact (M4), thus forming 4 complete contact sets (F1-M1, F1-M2, F1-M3 and F1-M4). These contact sets were mounted in a laboratory test set-up comprising a vertical computer-numerical-control milling machine, thus allowing the contacts to be closed and opened at a relatively constant and low contact speed. For each contact set, Fig. 2c shows the curves of the cumulative area beneath the dynamic contact resistance curves of Fig. 2b. The area value (Ar) just before the beginning of the vertical slope corresponds to the maximum value reached just before the arcing contact part.
New moving contact (M1) Slightly worn moving contact (M2)

Worn moving contact (M3)

Seriously worn moving contact (M4)

Fig. 2a. Wear contact analysis by evaluating the area beneath the dynamic contact resistance curve for different contact sets a) View of the fixed and moving contacts

b) R (m
0.5 s 1 m 2 4 3 1 Contact set F1-M3 Contact set F1-M4 t

a)
R and Contact travel

Arcing contact part D p

t0 t Contact travel curve Dynamic resistance curve

P a D a

1: 2:

Contact set F1-M1 Contact set F1-M2

3: 4:

R p

Main contact part Time

Fig. 2b. Wear contact analysis by evaluating the area beneath the dynamic contact resistance curve for different contact sets b) Graph of the dynamic contact resistance curves

c) Ar (m.s)
0.5 s 1 ms 5.4 m.s 3.9 m.s 2.8 m.s 4 3 1: 2: Contact set F1-M1 Contact set F1-M2 3: 4:
t

b)
2 1 2.7 m.s t

Arcing contact part

R (milliohms)

R a * D a Main contact part R p Contact travel (mm) D p D a P a

Contact set F1-M3 Contact set F1-M4

R a

Fig. 2c. Wear contact analysis by evaluating the area beneath the dynamic contact resistance curve for different contact sets c) Graph of the cumulative area beneath the dynamic contact resistance curves

For the different contact sets, the Ar value is: - 2.7 m.s for the new contact set F1-M1; - 2.8 m.s for the slightly worn contact set F1-M2; - 3.9 m.s for the worn contact set F1-M3; - 5.4 m.s for the seriously damaged contact set F1-M4. These Ar values provide an excellent assessment of the actual condition of the contact sets. In fact, the Ar value increases based on contact wear. The seriously damaged contact is clearly identified since the Ar value (i.e. 5.4 m.s) is twice that for the new contact set (2.7 m.s). B. Graph of the contact travel curve and resistance curve Fig. 3a depicts a typical dynamic resistance curve during an opening operation at low speed where t0 corresponds to the beginning of the breaker contact motion. In most breaker operating manuals, the procedure for performing such a low speed opening is given. It is always relevant to superimpose the travel curve of the breaker contact in order to extract diagnostic parameters related to the position of both the main contacts and the arcing contacts. These parameters are: - Rp (): Average main contact resistance - Dp (mm): Main contact wipe - Da (mm): Arcing contact wipe - Pa (mm): Position of the breaker contacts at the arcing contact part

Fig. 3a. & Fig. 3b. Parameters to be extracted from the dynamic contact resistance curve a) Contact resistance and contact motion as a function of time b) Contact resistance as a function of contact travel

C. Graph of the resistance curve as a function of the contact travel To compensate for the fact that the dynamic resistance curve is measured at a low contact speed that is not necessarily constant for the two test series (Fig. 1a), the contact resistance graph must be plotted as a function of the contact travel (Fig. 3b) in order to evaluate two additional parameters for diagnosing the arcing contact conditions: - Ra (): Average arcing contact resistance = ( Ri=1,N) / N (Fig. 3b), N= Number of samples in the interval Da - Ra*Da (m.mm): Area beneath the resistance curve as a function of the contact travel (Fig. 3b) The latter parameter provides a criterion for evaluating the global breaker contact wear and/or contact alignment status. Once the graph is plotted, all diagnostic parameters can be deduced, including those in section B. Since this graph can be considered as complete for diagnosing the breaker contact condition, it will be given for each case study presented in the following section.

V. CASE STUDIES The new DRM method was validated in the field on SF6 gas circuit breakers. Three case studies are presented in the following section. Table I summarizes the measurement results for which abnormal values are highlighted. A. Case study No. 1: One break of a 315-kV capacitor-bank SF6 gas circuit breaker Fig. 4 presents the DRM results on a break (Break A, Table I) of a 315-kV capacitor-bank SF6 gas circuit breaker which has performed 2492 operations. Based on this graph and the results listed in Table I, it can be deduced that the arcing contacts are in excellent condition. In fact, the Ra value of 185 is almost constant throughout the contact motion. The global criteria Ra*Da is also relatively low, i.e. 3.6 m .mm. In addition, the main contact part can be easily detected.

equipment. An investigation of the breaker failure revealed that an arcing contact tip appeared to have broken off during an opening operation and thus impaired the subsequent closing operation. In the fall of 2002, the DRM was performed. Based on the Fig. 5a graph, the parameters defined in section IV were extracted and listed in the case study No. 2 row in Table I. The instantaneous arcing contact resistance reaches an abnormal peak of 1 m while the average value (Ra) of 420 could be interpreted as normal. The most relevant factor is the product Ra*Da that reaches 10.3 m.mm, thus suggesting a contact anomaly. As mentioned in section IV, this factor represents the cumulative area beneath the resistance curve, thus summing the resistance variations or the contact wear during arcing contact opening.

a) R (milliohms)

Break B

3 Break A

R (milliohms)

1 Main contact part 0 5 15 25 35 Contact travel (mm)

0 0

10

20

30 50 40 Contact travel (mm)

Fig. 4. DRMs on break A TABLE I SUMMARY OF DRM RESULTS

b)

Break C

Break A: Break of a 315-kV capacitor-bank SF6 gas circuit-breaker Break B: Break of a 120-kV capacitor-bank SF6 gas circuit-breaker Break C: Same as break B, except that arcing contacts were overhauled Break D: Break (with internal restrike) of a 230-kV SF6 gas reactor circuitbreaker Break E: Same as break D, but without internal restrike

15

25

35 55 45 Contact travel (mm)

Fig. 5a. & Fig. 5b. Dynamic contact resistance measurements on one break of a 120-kV capacitor-bank SF6 gas circuit-breaker a) Dynamic resistance curve before contact dismantling b) Dynamic resistance curve after contact overhaul

B. Case study No. 2: One break of a 120-kV capacitor-bank SF6 gas circuit breaker Case study No. 2 (Fig. 5) presents the DRM results on a break (Break B, Table 1) of a 120-kV capacitor-bank SF6 gas circuit breaker which has performed 687 operations. In February 2000, a major failure occurred on this circuit breaker which caused important damage to the surrounding

Photos of the moving and fixed arcing contacts of the tested break are shown in Fig. 5c.

Ra*Da of 60 m.mm is the highest value that was ever obtained during the validation test program. The break was dismantled and arcing traces on both the moving and fixed arcing contacts as well as on the supporting tube of the main contacts were observed.

a) 3
Break D

R (milliohms)

20

40 60 Contact travel (mm)

b) 3
Break E 2

R (milliohms)

1 Main contact part 0 0 20 60 40 Contact travel (mm)

Fig. 5c. Dynamic contact resistance measurements on one break of a 120-kV capacitor-bank SF6 gas circuit-breaker c) View of the damaged moving and fixed arcing contacts

On the moving arcing contact, it can be observed that one arcing contact tip is off center. This abnormality caused damage to the fixed arcing contact (see right-hand side photo). It is believed that this condition occurred due to a misalignment of the arcing contacts at the break assembly. After an arcing contact overhaul and careful contact alignment, the DRM was performed one more time. Fig. 5b presents the measurement results that showed that the arcing contact condition was definitely restored. In fact, the Ra value of 173 is low. Furthermore, the low Ra*Da value of 3.4 m.mm indicates that the arcing contact is in excellent condition. C. Case study No. 3: One break of a 230-kV reactor SF6 gas circuit breaker Fig. 6a presents the DRM results for break D (Table I) for which an internal breakdown occurred without a major failure. In this case, the Ra value is about 2 mwhich indicates very severe damage to the arcing contacts. The global value

Fig. 6. DRM results on breaks of a 230-kV reactor SF6 gas circuit-breaker a) Resistance curve following an internal restrike of the break D b) Normal break E

For comparison purposes, Fig. 6b gives the DRM results for a normal break (Break E, Table I) of the same circuit breaker. Based on the curves and the extracted value in Table I, the arcing contacts of this break are clearly in excellent condition. In fact, the Ra value of around 100 is almost constant from the main contact part up to the arcing contact part. VI. CONCLUSION This paper presents a new dynamic contact resistance measurement method performed during opening operations at low contact speed aimed at evaluating the breaker condition without dismantling it. Compared to the DRM curves at the rated contact speed, the new method allows reproducible curves to be obtained which are easy to analyze and interpret. Three signals must be measured: the injected DC current that must be produced by a stable source, the voltage drop across the breaker contacts and the contact travel.

To extract the diagnostic parameters, a dedicated software program was developed in order to plot the dynamic resistance curve as a function of the contact travel, i.e. m versus mm. Six vital diagnostic parameters values are therefore determined: - average main contact resistance; - average arcing contact resistance; - main contact wipe; - arcing contact wipe; - position of the breaker contact at the arcing contact part; - and the cumulative area beneath the resistance curve. The last parameter is the most relevant one since it allows the overall contact wear and/or contact alignment status to be assessed. Moreover, values obtained from different breaker technologies can be compared. For example, values of about 3 m.mm indicate healthy breaker contacts while values of about 10 m.mm indicate faulty contacts. The three case studies presented in this paper prove that the new DRM method provides vital information about the breaker contact condition. Without dismantling the breaker, the maintenance crew can thus plan maintenance work for specific breakers for which the DRMs reveal contact anomalies. VII. REFERENCES Papers from Conference Proceedings (Published):
[1] [2] [3] Salamanca F., Borras F., Eggert H., Steingrber W., Preventive Diagnosis on High-Voltage Circuit Breakers, Paper No. 120-02, 1993 CIGRE Symposium, Berlin. Kumar Tyagi R., Singh Sodha N., Condition-Based Maintenance Techniques for EHV-Class Circuit Breakers, 2001 Doble Client Conference. Ohlen M., Dueck B, Wernli H., Dynamic Resistance Measurements A Tool for Circuit Breaker Diagnostics, 1995 Stockholm Power Tech International Symposium on Electric Power Engineering, Vol. 6, p. 108113, Sweden.

from the IEEE Power Engineering Society in 1986. He is a senior member of the IEEE Power Engineering Society, member of the Canadian IEC Technical Committee 17 on switchgear, and a registered Professional Engineer in the province of Qubec. Andr Mercier received his B.Sc. in Electrical Engineering from Laval University (Qubec City) in 1977 and, in 1979, completed the credits for a M. Sc. degree in microcomputers. For 1 years, he worked with GENIFAB Inc. in hardware and software design and production of control and data acquisition systems for various industries. In 1980, he joined the development and research centre of Qubec Iron Titanium (QIT, 500-MW smelter and mill). For ten years, he designed and built many real-time process control and data acquisition systems (both software and hardware) using microprocessors, PLC, PC, VAX, etc. He was also involved in contracts concerning data acquisition for external companies. In 1990, he joined IREQ as a research scientist working on projects related to circuit breakers, focusing on the monitoring and controlled switching aspects. He is involved in some IEC and CIGRE working groups, and recently in the CIGRE A3.07 working group where he was a very active member. He designed and installed the first controlled system to energize unloaded high-power transformer taking into account the residual flux. He teached the design of industrial electronic controls using digital and microprocessor circuits at cole Polytechnique de Montral. He is a member of the Ordre des Ingnieurs du Qubec. Gilles Ouellet was born in Qubec, Canada, on February 20, 1953 and received his diploma in physics laboratory techniques from CEGEP La Pocatire, Qubec in 1974. Specialized Training: Course on high-voltage, heavy-current measuring techniques given by Dr. Ryzsard Malewski of IREQ specializing in data measurement and acquisition. Expertise in circuit simulation using EMTP, FLUX2D, PHI3D, etc. and various electric and magnetic field calculation tools. Technical experience: Responsible for the selection and testing of various sensors for instrumenting a GEC-Alsthom PK-PKV type circuit breaker. These sensors are now being used to monitor two breakers installed in IREQ's high-power laboratory in 1993, at some Hydro-Quebec substations and at some New-York Power Authority substations. Involved in the development of a new method for measuring the contacts resistance of breakers. Involved in the development of a capacitor-divider system used at Hydro-Qubec's Rivire Sainte-Anne substation and now in Mexico, in a joint project with BG CHECO. Designed several capacitor banks with strong di/dt (200 kA in 1.8 ms) used to test the oxy-metallic blocks of the new sacrificial surge arrester for the James Bay system. Cooperation requested by manufacturers such as McGraw-Edison and Raychem during the development of the metal-oxide surge arrester. Involved also with CIGRE Working Group 3.4.11 on Surge Arresters in the testing of metal-oxide blocks. Co-author of a new welding-arc control method based on acoustic-wave modulation, now marketed by a high-tech firm in Qubec. Participated in simulation studies of short-circuit-current interruption in the absence of current zero, which led to selection of the circuit breaker to be used for Hydro-Qubec's series compensation system. Author of fifty reports and contributor to five IEEE publications. Jacques Caron received his B.Sc A. in Electrical Engineering from Sherbrooke University in 1975. In 1979, he obtained his M.Sc in Energy from INRSnergie, University of Qubec. Since 1979, he has been working in the Maintenance department of Transnergie, the Transmission Division of HydroQubec. He is currently involved in the technical support of circuit-breakers maintenance activities. These breakers are from various vintages and technologies, like vacuum, oil, airblast and SF6, most of them rated for operation down to -50C, with voltages ranging from 12kV up to 765kV. Michel Roy received the BSEE degree from Laval University in 1972. From 1972 to 1990, he worked at IREQ as a test engineer. He was involved on the

VIII. BIOGRAPHIES
Michel Landry (S'75, M'77, SM'90) was born in Qubec, Canada, on August 23, 1952. He received his B.Sc.A. in Electrical Engineering from Sherbrooke University in 1975. In 1977, he obtained his M.Sc. in Energy from INRS-Energie, University of Qubec in Montral. From 1977 to 1979, he was responsible for testing electrical power apparatus for the high-power laboratory at Hydro-Qubec's Research Institute (IREQ). In 1979, he joined the Electrical Equipment department at IREQ and, until 1983, participated in the development of mathematical methods and computer programs for the selection of electrical components of IREQ's synthetic platform designed to perform synthetic breaking tests according to IEC and ANSI standards. From 1983 to 1985, he was involved in a research and test program to qualify the required electrical performance of 2-MV air-blast circuit-breakers for Hydro-Qubec's James Bay power system. He also contributed to the development of a new post-arc technology applied to power circuit breakers. From 1985 to 1990, he was engaged in research related to the design of a new SF6 puffer breaker for low temperature application (-50oC), which earned the 1991 Mritas prize in Engineering awarded by the Ordre des Ingnieurs du Qubec. He is now involved in many research projects related to breaker interrupting performance and in-service condition monitoring for which MONITEQ, an on-line monitoring system for HV circuit breakers, earned an R&D 100 given by the prestigious R&D Magazine of Chicago. He has authored or co-authored more than 35 international publications, one of which earned a prize paper award

8 construction of synthetic test station and the test techniques for HV and EHV circuit breakers. From 1990 to 1994, he was the chief engineer for the development of the Hydro-Qubec standards on HV and EHV circuit breakers and their certification. In 1995, he joined the Maintenance departement in Hydro-Qubec. He is mainly responsible for SF6 circuit breakers and GIS substations. He is registered Professionnal Engineer in the province of Qubec. Dr. Fouad Brikci, Ph.D., is the president of Zensol Automation Inc., one of the leading manufacturers of circuit breaker analyzers in the world. Dr. Brikci was the first to introduce on the market the concept of TRUE computerized test equipment in the field of circuit breaker analyzers. As a former university teacher and CNRS researcher in France, Dr. Brikci has developped experience in the fields of Electronics, Automation and Computer science. Most activities were focused on the industrial application of computers. Among his achievements are the development of a major automated system made for paint manufacturers, development of fully computerized measuring systems for quality control manufactured by circuit breaker manufacturers, laboratories and maintenance services of electric utilities. Dr. Brikci holds a Ph.D. degree in Electronics and a Master in Sciences in EEA (Electronics, Electrotechnics and Automation) from the University of Bordeaux, France.