Part 1 of this article, which appeared in the September's 2017 edition of Electrical Tester online and is available by clicking here, looked at the basics of time and travel measurements for circuit breakers. This second and final part presents a practical case study, and also discusses how to proceed when the circuit breaker manufacturer provides little or no information to aid time and travel measurement.
Robert Foster - Application engineer
Case study: Siemens SPS2-38-40-2 Circuit Breaker
To investigate how transducer placement affects travel measurements, several transducers, rotary and linear, were attached to a Siemens SPS2-38- 40-2 SF6 dead tank circuit breaker fitted with an FA2.20 mechanism. This breaker is rated at 38 kV and is capable of interrupting 40 kA. It has one break per phase and is gang operated (See Figures 1 and 2).
Fig.1: Siemens SPS2-38-40-2 nameplate
Fig. 2: Siemens SPS2-38-40-2 circuit breaker with FA2.20 mechanism
Siemens recommends measuring motion with a linear transducer attached to an actuating arm on the mechanism (See Figure 3).
Fig.3: Transducer connection Linear A
The company states that 80 mm of travel at the mechanism is equal to 120 mm of travel at the contacts – that is, a multiplication factor of 1.5 must be used to determine the true contact motion. It is also stated that for the close operation, the speed calculation points of contact touch and 10 ms before should be used, and for the open operation, contact separation and 10 ms afterwards. In addition to the standard transducer connection, four more transducer connections were made: one linear transducer was connected to the end linkage arm on the third phase and three rotary transducers were attached to the rotating splines that drive the interrupter (See Figure 4). The same speed calculation points were used for all connections.
Fig. 4: Transducer connections, from left to right Linear Linkage, Rotary C, Rotary B, and Rotary A
To compare the different transducer measurements, at first no conversion factors were used on the two linear transducer connections and a factor of 1º equals 1 mm was used for the rotary transducer. Figure 5 shows the motion traces from the three different connections for a close operation; all three are scaled at 10 mm per division.
Trace A on the graph in red is the linear transducer connected directly to the mechanism as recommended by Siemens; this will be referred to as Linear A. Trace B in black is the rotary transducer and is connected to the rotating spline on B Phase; this will be referred to as Rotary B. Trace C in blue is the linear transducer connected to the end of the interconnect linkage near the rotating spline on C phase and will be referred to as Linear C.
For all graphs, the bottom of the curve is the fully open position and the top of the curve is the fully closed position. The timing for each phase is also shown – a thin line is open and a thick line is closed. The timing results show that all three phases are relatively in sync with 0.3 ms difference between the slowest and the fastest, and a close time of about 48 ms. As expected, the travel measurements vary widely due to the different connection points. The stroke for Linear A is 78.9 mm, Rotary B is 59.0 mm (59°) and Linear C is 106.5 mm. From these stroke measurements, the travel dependent parameters such as velocity, overtravel, penetration, rebound etc., will also vary by transducer placement.
Fig. 5: Close operation with no conversion, Linear A in red (center trace), Rotary B in black, and Linear C in blue (top trace)
An examination of the open operation (see Figure 6) shows similar variance in stroke values for the different connections.
Fig. 6: Open operation with no conversion, Linear A in red (center trace), Rotary B in black, and Linear C in blue (top trace)
As mentioned earlier, the circuit breaker manual states that 80 mm of motion at the mechanism is equivalent to 120 mm of contact movement, therefore with a transducer stroke of 78.9 mm the contact stroke is 118.35 mm. Since the other transducers were measuring motion on the same operation, the ratios for the other linear transducer and the rotary transducer can be calculated. Since 59.0° = 118.35 mm for Rotary B and 106.5 mm = 118.35 mm for Linear C, the conversion factors are 2.003 mm/° and 1.1099 mm/mm respectively. With this information to hand, the circuit breaker was tested again and the appropriate conversion factor applied to each transducer. The results are shown in Figures 7, 8, 9 and 10.
Fig. 7: Close operation with conversion factors applied, LinearA in red, Rotary B in black, and Linear C in blue
Fig. 8: Close operation with conversion factors applied, bottom to top: Linear A in red, Rotary B in black, and Linear C in blue
Fig. 9: : Open operation with conversion factors applied, Linear A in red, Rotary B in black, and Linear C in blue
Fig. 10: Open with conversion factors applied, bottom to top: Linear C in red, Rotary B in black, and Linear C in blue
These graphs show that even with three different transducer attachment points and two different types of transducer, very similar results are obtained as long as the correct conversion factor is applied. The maximum difference between stroke values is 0.2 mm on the close operation and 0.5 mm on opening. Penetration, overtravel, and rebound are also very close for the three different measurements.
An interesting observation is that the travel trace of Linear A, the linear transducer connected directly to the mechanism, shows oscillation throughout the entire movement, and Linear C, the linear transducer connected to the end of the linkage, shows slight oscillatory movement at the beginning and end of travel. Most likely the flex in the travel rod and the connection of the rod to the transducer accounts some of this movement. Additionally, since Linear A is connected directly to the mechanism, the vibrations of the mechanism affect the motion throughout the entire travel. Since the close operation requires more energy – closing the breaker and charging the opening spring – this effect is more apparent on the close operation.
Another interesting observation is that the velocity is different for each connection. This is in part due to the calculation points being based on contact touch and separation, so the variance in the timing will affect where on the curve the velocity is calculated.
On Linear A the vibrations may also affect the velocity calculations. If one of the speed calculation points is on the crest of an oscillation and the other calculation point is in the trough of an oscillation, the speed value obtained can be very different from the value obtained when the points fall on a neutral part of the oscillation. This effect was confirmed by examining several consecutive operations and observing that the speed values obtained from Linear A varied by 0.16 m/s or 3%, whereas the values from the other two connections varied by a magnitude less. Play in the linkages also affects velocity calculations.
A final thing to consider is that a linear conversion factor was assumed – that is, a conversion constant was used. Comparing Rotary B to Linear C, they align better at the beginning and at the end of travel. In the middle of the movement they diverge slightly and, since this is the portion of the curve where velocity is calculated, it follows that the speeds will diverge slightly as well. If the geometries were analysed and a conversion table was built for both connections, they would be likely to overlap throughout most of the travel and the velocities would align more closely.
Fig. 11: Close operation with conversion factors applied, bottom to top: Rotary A in red, Rotary B in black, and Rotary C in blue
Fig. 12: Close operation with conversion factors applied Rotary A in red, Rotary B in black, and Rotary C in blue
The motion traces of the three different rotary transducers can be examined to see how the same connection can be placed at different distances from the mechanism – that is, at different points along the interconnecting linkages – and yield similar results. Figures 11 and 12 show the results from a close operation. Trace A on the graph in red is the rotary transducer connected directly to the rotating spline that drives the interrupter in A phase; this will be referred to as A phase. Similarly, trace B in black is the rotary transducer connected to Phase B and will be referred to as B phase. Lastly, trace C in blue is the rotary transducer connected to Phase C and will be referred to as C phase.
Once again all three traces are very similar with a variance in stroke of only 1.2 mm between the shortest and the longest phase. Note that A phase begins to move approximately 0.5 ms before phases B and C, which can be expected since it is the closest connection to the mechanism. A and C phases produce very smooth traces throughout the motion but B phase has some oscillations in the first 20 ms of travel. These oscillations are most likely due to a mechanical delay. B phase is pushed by the linkage from A phase and then it has to push the linkage to phase C. Any mechanical play in the connections between B and the other two phases will result in small perturbations.
The velocities of A and C are fairly close but B phase is 0.2 m/s slower. This is probably caused by two factors: firstly, the oscillations in the trace can cause different speed points to be taken as already mentioned; secondly, the timing of the three phases is slightly different and the speed calculation point is based on contact touch.
Careful observation of the speed calculation points reveals that they do not line up in time.
Observing the open operation in Figure 13 and Figure 14 shows even more consistency between the different phases.
Fig. 13: Open operation with conversion factors applied, bottom to top: Rotary A in red, Rotary B in black, and Rotary C in blue
Fig. 14: Open operation with conversion factors applied, Rotary A in red, Rotary B in black, and Rotary C in blue
All three traces practically lie on top of each other with no deviation until the contacts reach the closed position. The stroke of the different phases is closer with only 0.5 mm difference between the shortest and the longest phase.
Once again the velocities of the three phases are different but observing the contact times and speed calculation points shows that the velocity is calculated at slightly different points on each curve thus changing the values slightly. If the speed calculation points are changed to reference below closed and a differential, then B and C travel at the same velocity while A phase travels slightly slower as it has to push the other two phases.
What to do if little or no information is available from the manufacturer
Occasionally the manufacturer may not provide appropriate information for travel measurements and it is then left to the technician to decide what type of transducer to use, where to connect it, what conversion factor/table to use (if any) and which speed calculation points to use for determining the velocity of the contacts.
Careful consideration should be taken before attaching a transducer and once a method is determined, the same attachment and measurement parameters should be used in the future to permit trending of the results. Although the travel recordings obtained without access to manufacturer’s information will provide valuable data and can be used for future reference, the values obtained may not necessarily be comparable to the factory test reports or parameter limits.
Once again, in order to avoid damage to the transducer and its accessories, no part of the transducer, mounting bracket, or travel rod (if used), must be in the direct path of any moving parts of the circuit breaker.
The first thing to look for in deciding where to connect the transducer is whether direct connection to the contacts or contact actuating arm is possible. If it is, a linear transducer can be connected and the correct stroke, velocities and other parameters will be measured without the need of a conversion table. If direct connection to the contacts is not possible, which is often the case, then a location that is very close to the contacts, with the minimum number of linkages between the connection point and the contacts, should be selected.
Either a linear or rotary transducer can be used. IEC 62271-100 states that the mechanical characteristics can be recorded with a travel transducer at “convenient locations on the drive to the contact system where there is a direct connection, and a representative image of the contact stroke can be achieved”.
Connecting directly to the mechanism can cause unwanted vibrations and influence the results, so this should be avoided if possible. If an indirect connection is used then there are two options: create a conversion table/factor, or measure the absolute value of the motion, in either length or angle, and trend the results with the transducer connected in the same spot during future testing.
If a conversion factor or table is to be used, the connection points and linkages can be examined and measured to develop a trigonometric function that relates the transducer movement to the contact motion. The function can also be determined from the mechanical drawings of the circuit breaker.
If the stroke of the contacts is known, another, less accurate, method of creating a conversion factor is to assume a linear relationship between the connection point and the contacts. The known stroke of the contacts can be divided by the measured stroke of the transducer to create the conversion factor. This value can then be used to measure the travel characteristics for initial fingerprint measurements and for future testing.
It should be noted that if the relationship between the connection point and the contacts is nonlinear, other stroke-dependent parameters such as velocity, overtravel, rebound and so on, may not be correct. Additionally, if the initial measurement is made when there are issues with the circuit breaker – that is, when the stroke is not correct – the subsequent measurements will also be incorrect. If there are other circuit breakers of the same type available, it is beneficial to compare measurements to verify the correction factor.
If no speed calculation points are provided by the manufacture then it is recommended to use contact touch and 10 ms before for the closing operation, and contact separation and 10 ms after for the opening operation. This will ensure that the velocity is measured in the critical arcing zone of the interrupter. Once again, it is worth stressing that once the method of transducer connection, conversion factor, and speed calculation points have been decided upon, they should be employed throughout the life of the circuit breaker to allow trending of results.
Circuit breakers are key elements in electrical transmission and distribution networks all over the world. IEEE C37.09 states that “travel-time curves shall be obtained for all outdoor circuit breakers with an interrupting time of three cycles or less.” To verify that the circuit breaker will operate effectively when called upon to protect assets in the network, time and travel analysis must be performed.
When determining what type of transducer should be used, where it should be connected, what conversion factor should be applied and what speed calculation points should be used, the first step is to consult the manufacturer’s manual. If it contains no directions or the directions are unclear, the next step is to contact the manufacturer.
If this option is not available either, the technician performing the testing must decide how to proceed. If possible, a direct connection between the transducer and the contacts should be adopted, but if this is impractical, a connection point that is near the contacts with a minimum amount of linkage, which can accurately represent the travel of the contacts, should be used. If the geometry of the circuit breaker is known, a conversion factor or table can be created to accurately calculate the stroke and parameters that are dependent on the stroke.
Even if the original measurement points used by the manufacturer are not known, valuable data can still be obtained with a transducer as long as it is placed in a sensible location. In fact, even if motion is measured at different points on the circuit breaker, as long as the correct conversion factor is applied, the results will be very similar.
If an accurate conversion factor or table cannot be created, the absolute value of the transducer stroke and its parameters can be measured during commissioning or when the circuit breaker is known to be in good condition. The values obtained can then be trended over time to track any changes in the movement or operation of the circuit breaker.
Finally, once one type of connection and conversion factor is chosen, all future measurements should be made using the same setup so that the results can be correctly trended.