A Methodology for the Efficient Application of Controlled Switching to Current Interruption Cases in High-Voltage Networks

Transcription

1 A Methodology for the Efficient Application of Controlled Switching to Current Interruption Cases in High-Voltage Networks C. D. TSIREKIS Hellenic Transmission System Operator Kastoros 72, Piraeus GREECE F. D. KANELLOS Hellenic Transmission System Operator Kastoros 72, Piraeus GREECE G. J. TSEKOURAS Dept. of Electrical & Computer Science Hellenic Naval Academy Terma Hatzikyriakou, Piraeus GREECE Abstract: - Transients produced upon the opening operation of a circuit-breaker may be harmful for the network elements and the switching device. A modern countermeasure for the reduction of such transients is controlled switching. Fundamental requirement for all controlled switching applications is the precise definition of the desired switching times. In this paper a new methodology is proposed for the calculation of the optimum switching instants, taking into account circuit-breaker s statistical scatters in the contact operation time and the slope of the contact gap voltage withstand characteristic. Key-Words: - Controlled Switching, Switching Transients, ATP-EMTP Simulation, Circuit-Breaker 1 Introduction Controlled switching is a technique that automatically adjusts the circuit-breaker mechanism in such a way that switching operation takes place at a point-on-wave which minimizes switching transients, such as the phase-to-earth overvoltage, the inrush current and the transient recovery voltage (TRV) across the breaker poles [1, 2, 3]. One of the most significant requirements for proper controlled switching performance is to reduce the statistical variations of contact operation times, since they may affect the success of this method [1, 2, 3, 4]. Circuit-breaker technology has improved these statistical scatters, allowing thus utilities and manufacturers to achieve contact operation times quite close to the preferable ones. This means that a precise definition of the desired switching times is required, taking into account the effects of parameter changes (such as the trapped charge in a capacitor bank or the impedance of a load) and the circuit-breaker characteristics (such as the statistical variations of the contacts operating times and the contacts gap characteristic of dielectric strength) on the optimum switching instant. In this paper a methodology for the calculation of the optimum circuit-breaker opening instants is proposed. Specific restrictions, such as contacts gap voltage withstand characteristic and variations in contacts operating times are considered. With the aid of numerical techniques, the optimum switching instants can be easily calculated. 2 Operation Principles 2.1 Controlled Switching Arrangement In general, a typical controlled switching arrangement consists of the following main parts: The circuit-breaker, with an independent pole operation capability. Devices for the measurement of the instant values of the circuit-breaker currents. A controller, which after receiving the switching command, processes the signals from the measuring devices, determines the suitable reference phase angles and sends the breaking commands to each breaker pole, so that the breaking occurs at the optimum instant. According to this, a typical controlled switching layout is shown in the following figure: SOURCE SIDE INTERFACE CONTROLLER LOAD SIDE Fig.1: Main parts of a typical controlled switching arrangement. ISBN:

2 2.2 Operation in an Ideal Circuit-Breaker For the analysis of each part of the operation sequence during the controlled opening, the definition of the ideal circuit-breaker is introduced. Its main characteristics are the following: No arc is developed during the opening operation. Therefore, currening occurs simultaneously to the mechanical separation of the contacts. The probability of reignition is zero. The sequence of each partial operation during controlled opening of an ideal circuit-breaker is illustrated in the next figure. In this figure and all the next ones, t and T represent time instants and time intervals, respectively: T w + Τdelay t comm and t start Toperating Current across contacts Breaking command (User) Operating command to poles Breaking Fig.2: Operation sequence during controlled opening of an ideal circuit-breaker T operating corresponds to the time interval between the sending of the opening command from the controller (t start ) and the starting of the contacts mechanical separation. Since T operating is very short, it can be neglected and considered that current breaking occurs instantly and simultaneously to the command sending from the controller and the contacts mechanical separation (t start = ). Therefore, the corresponding time delay T delay for each phase is given by the following equation: T delay = T w (1) 2.3 Application to a Real Circuit-Breaker In practical applications, contact gap has a finite dielectric strength. This affects the controlled opening operation. In a real HV circuit-breaker, currening (instant in Fig.2) does not occur simultaneously to the instant of the contacts mechanical separation (instant t separate ). The current flow after the instant of t separate is kept through the electric arc developed between the opening contacts and is broken at one of the next physical current zeros, since at those instants the energy absorbed by the arc, maintaining it, is zero. However, as current approaches the physical zero points, the arc comes to an instable mode. As a result, current becomes zero slightly before a physical zero instant (current chopping). Therefore, opening switching transients are independent of the phase angle corresponding to the instant of contacts mechanical separation and controlled opening cannot reduce this kind of transients directly. Immediately after the arc extinguishing, a transient recovery voltage (TRV) is installed across the contacts gap. The initial rate of rise of recovery voltage (RRRV) may be quite high, especially in inductive currents interruption [5]. As a result, it is strongly probable for TRV to exceed once or multiple times gap dielectric strength, with the subsequent occurrence of a number of reignitions. Each reignition, is equivalent to a temporary reclosing of the circuit, which generally occurs under more adverse conditions than the closing with zero initial conditions, due to the accumulated energy. In any case, reignitions must be avoided, so that more dangerous transients, such as voltage escalation, are prevented. The key for the prevention of restrikes is the relation between the magnitude of TRV and the gap dielectric strength. The slope of the latter (Rate of Rise of Dielectric Strength - RRDS), which depends on the contacts separation velocity and the gap dielectric behavior, may be quite smaller than RRRV. However, it appears earlier (at the instant of contacts mechanical separation), while TRV is established at the instant of electric breaking, as it is illustrated in the next figure: Breaking current Current Chopping t separate Grid phase voltage Characteristic of gap dielectric strength T arcing Current's "physical zero" Transient recovery voltage (TRV) Fig.3: Successive inductive current interruption. No reignition due to a long arcing duration (T arcing ) As mentioned previously, electric breaking occurs slightly before one of the current zeros after contacts mechanical separation. Therefore, a maximum arcing duration (T arcing ) is needed for the prevention of a reignition. This means that contacts mechanical separation should occur immediately after a current zero, so that even when electric breaking occurs near the first current zero, arcing ISBN:

3 duration T arcing is quite long (just shorter than a halfcycle) and the gap dielectric strength is high enough to avoid reignition. All the above are illustrated in Fig. 3 and 4, for inductive current interruption without and with reignition, respectively: Breaking current Current Chopping Grid phase voltage Characteristic of gap dielectric strength t T arcing separate Current's "physical zero" Reignition Transient recovery voltage (TRV) Fig.4: The same inductive current interruption, but with reignition, due to the short arcing duration (T arcing ) The general result is that the aim of the controlled switching application in currening cases is the avoidance of reignitions. The sequence of the various operations is illustrated in Fig.5: t command T w + Τdelay Tarcing t separate Current across contacts Breaking command from User Operating command to breaker poles & starting of contacts separation Currening Fig.5: Operation sequence during controlled opening of a real circuit-breaker The velocity of contacts separation The RRDS due to the gap s stochastic nature. The third and the fourth time intervals are quite short, so that their statistical deviations can be neglected. Deviations of Τ w and Τ delay contribute to a total deviation to the instant of starting of contacts movement. Finally, the last two deviations contribute to the slope of dielectric strength characteristic. Therefore, the basic statistical deviations which should be always considered, despite of their source, can be condensed to the following two: Deviation of the starting instant of contacts movement ( Τ), depending only on the performance of the controller. Deviation of RRDS, depending only on the performance of the circuit-breaker. The maximum deviations can be estimated with a great probability (up to 99.99%) by the manufacturers. In particular, a controller with a maximum Τ of ±2 ms is easily constructed. The same is valid for a maximum Τ of ±1 ms, but with higher cost, while there is no reference for a reduction of Τ below ±0.7 ms. On the other side, a deviation of RRDS due to contacts velocity does not exceed ±5%, while the deviation due to the gap s stochastic nature is higher (up to ±20%). The existence of a deviation Τ causes a bilateral parallel shift to the gap dielectric strength characteristic, creating an area of possible instants of contacts mechanical separation as wide as 2 Τ, as shown in Fig.6. This results in the reduction of the arcing duration (T arcing ). The deviation of RRDS contributes to a further increase of the probability of a reignition, as shown in the same figure. Breaking current 2.4 Statistical Deviations A variety of factors, like environment conditions (ambient temperature, humidity κλπ.), the utilization frequency of the circuit-breaker and the whole device, the age of the arrangement and the corresponding equipment strain and wear, contribute to the appearing of statistical deviations in the operational characteristics of the controlled switching arrangement, such as the following ones: The waiting duration Τ w The time delay Τ delay The time between the sending of the command from the controller to the circuit-breaker and the starting of contacts movement The time between the starting of contacts movement and their mechanical separation T T Transient Recovery Voltage (TRV) Fig.6: Effect of Τ and RRDS deviation to the instant of contacts mechanical separation 3 Calculation Methods From the preceded analysis, it became obvious that calculation of the optimum switching instants in ISBN:

4 real networks is quite complicated. In practically all cases, the currening instant (opening instant) does not coincide with the instant of mechanical separation of the circuit-breaker contacts (target instant). The presence of statistical deviations leads to the claim that instead of a simple optimum opening instant and the respective target instant, it is more realistic to talk about an optimum window of target instants. Furthermore, three-phase operation of the network leads to the final conclusion that we should talk about optimum combination of target instants windows. From all the above mentioned, it is obvious that a systematic calculation of the optimum target instants windows combinations, taking into account all network parameters, circuit-breaker electrical characteristics and various statistical deviations. It is also necessary to assess the controlled switching performance for the prevention of reignitions. Both of these aims are met successfully through the application of the algorithms described in the next paragraphs. 3.1 Basic Steps of the Algorithm The first step of the algorithm run is the precise calculation of the waveforms of breaking currents and TRVs, as they are derived after successful (without reignition) breaking of the currents in the preceded phases. It should be noted, that except the waveform of the breaking current of the first phase, the waveforms of the rest phases depend in general on the instants of electric breaking in the preceded phases. However, if the peak value of the breaking current is well higher than the chopping level, the electric breaking in each phase occurs exactly at the chopping instant and consequently, it is independent of the instant of contacts mechanical separation. Therefore, for each combination of values of chopping level and the electric parameters of the grid, the waveforms of the current and TRV in each phase are precisely defined and independent of the instant of contacts mechanical separation in this case. This is not valid if chopping level is at least of the same order of magnitude with the peak value of the breaking current, since electric breaking may occur even simultaneously to the instant of contacts mechanical separation. As a consequence, in this case the waveforms of the breaking current of the two lasing phases, as well as of all TRV, are functions of the instants of contacts mechanical separation in the preceded phases. In the next stage, the user defines the variation step of the possible instants of contacts mechanical separation (t step ), which is set equal for the three phases for minimization of the number of parameters, as well as the total number of steps. In addition, the user defines the vectors of the possible values of RRDS, of the deviation of the velocity of contacts separation ( υ) and of the deviation of the slope of gap dielectric strength characteristic due to its stochastic nature ( s). Finally, the user defines the shape of the gap dielectric strength characteristic through appropriate polynomial factors. Considering that the left end of the optimum instants window of contacts mechanical separation is the first discrete instant after current zero (equal to the sum of current zero instant and t step ), the calculation of the optimum instants window is based on the right limit of the gap dielectric strength characteristic (see Fig. 6). Therefore, for the slope of this characteristic the deviation given are those which provided for the right limit of the instants window of the contacts mechanical separation, since this is the most adverse one. For instance, if the dielectric strength characteristic is straight, its formula is: Vright(t,tseparate) = RRDS (1 υ s) (t tseparate) (2) In the above formula, t separate is the step-varied instant of contacts mechanical separation. Starting from current zero instant in the first phase (t o ), the instant of contacts mechanical separation increase gradually by t step, until gap dielectric strength characteristic intersects the rectified TRV waveform. Assuming that this takes place for an instant of contacts mechanical separation t separate = t last. The optimum instants window for contacts mechanical separation for the first phase is then [t o, t last - t step ] and the mean value of this window combined with maximum T is used for its designation. Considering that deviation Τ is bilateral, the absolute value of the maximum Τ, so that the reignition is avoided, is derived from the following equation: t last t step t o T = (3) 2 In case that chopping level is of the same order of magnitude with the breaking current peak value, the process in this stage is modified, so that the left end of the optimum instants window of contacts mechanical separation does not necessarily coincide to current zero. The optimum instants window starts from the first instant of contacts mechanical separation which does not lead to reignition. The same process is repeated for the rest phases. Finally, the minimum value among the maximum allowed Τ derived for the three phases is chosen as ISBN:

5 the maximum allowed Τ for the whole system. The combination of the optimum instants windows of contacts mechanical separation, as well as the maximum allowed DT, depend on the combination of the values of the grid parameters, the chopping level (only when it is well lower than current peak value), the circuit-breaker dielectric characteristics and the corresponding deviations. Any common programming package (Matlab, Mathcad, C, Fortran etc.) can be used for the application of this algorithm. breaking current. 3. Calculation of voltages and currents in the equivalent network after the instant. 4. Results of step 3 are superimposed to those of step 1, for the total expressions of the voltages and currents after the breaking of the first phase. 5. The process is repeated for the breaking of the rest phases, using the voltages and currents after step 4 after the breaking of the preceded phase instead of the respective steady-state quantities. 3.2 Calculation Methods As mentioned in the previous sections, for the application of the algorithm of the optimum instants of contacts mechanical separation, the calculation of the breaking current in each phase is required. This calculation can be performed via either analytical or numerical methods. Analytical methods comprise the solution of systems of differential equations describing the transient behaviour of the network. The Current Injection Method is such a method, which is proposed for the application of the algorithm. On the other side, the most conventional method for the numerical solution of the network during transient conditions is the use of programs like ATP/EMTP [6, 7]. The application of both methods is described in the next paragraphs Current Injection Method The widely known Current Injection Method is used for the network solution in currening cases, assuming a linear network [4, 8]. According to this technique, the calculation of the transients is based on the fact that the current elimination at the instant of its breaking ( ) is equivalent to the injection of the same magnitude and opposite polarity to the breaking current, as shown in Fig.7. Therefore, the transient voltages and currents in all network places are derived from the superimposition of their instant values which they would have without the breaking and the respective values obtained after the aforementioned current injection at the instant. The application of the method can be summarized in the following steps: 1. Calculation of the steady-state voltages and currents in various network places before the breaking of the first phase. 2. Development of an equivalent network after the replacement of voltage and current sources with open- and short-circuits, respectively. The breaker pole to open is replaced with a current source, connected at the instant, with the same magnitude and opposite polarity to the Fig.7: Current injection method Breaking current Injected current Superimposition of the two currents Result (actual current) The advantage of Current Injection Method against the direct network solution is that there is no need for the (often hard) calculation of the initial conditions, since their effect is taken into account through the superimposition. However, its results are valid only for networks consisting exclusively of linear elements. Another disadvantage is that the calculation of the transients are usually quite complicated, due to a plenty of reasons, like the unsymmetrical three-phase networks in the intermediate steps (differening instant in each phase), the inductive and/or capacitive coupling between the three phases, as well as the existence of elements with distributed parameters or complicated models (lines, cables, transformers). For these reasons, the application of the Current Injection Method is just used for the comprehensive approach of the relative switching phenomena and the assessment of the performance of controlled switching application to a specific currening case. In any case, the use of numerical methods is necessary for a more precise assessment Use of ATP/EMTP Program For the numerical solution of the network in transient conditions, the widely known ATP/ΕΜΤΡ program is used. The model of Statistical Switch included in this program is the most suitable circuitbreaker model for controlled switching simulations. ISBN:

6 The use of this model causes automatically the repeated run of the simulation for different combinations of instants of contacts mechanical separation in each phase, following a uniform probability distribution with user-defined mean time and standard deviation (see Fig.8), since the interest is focused on the most adverse transients and not on the probability of their occurrence. Fig.8: Probability distribution of opening time T f(t): Density function, T : Mean time, σ: Standard deviation For instance, in the usual case of the desired instants window of contacts mechanical separation corresponds to a 50 Hz half-cycle (10ms), it is derived that σ = ms. According to the general principle followed by ATP/ΕΜΤΡ, electric breaking is achieved at the first instant after contacts mechanical separation at which current criterion is met. As current criterion is defined the condition, at which the absolute instant value of breaking current is lower than a user-defined current margin, according to the following figure. It is obvious, that the above current margin represents the chopping level. i SWITCH σ Τ 3 σ f(t) Τ + 3 σ Current forced to zero in next step Current margin t Current margin Switch opens Current Fig.9: Illustration of application of current criterion in an EMTP switch during current breaking In the majority of the currening cases, a number of automatic re-runs between and is enough for the precise assessment of the optimum instants windows of contacts mechanical separation. The use of any similar numerical software, such as ATP/EMTP, aims to the precise evaluation of the performance of controlled switching application, but it does not ensure the thorough investigation of the Τ problem. For this reason, it would be better that such software is used in a second stage, after the application of an analytical method, such as Current Injection Method to a simplified network model, which is more suitable for the identification of the possible issues. 4 Conclusions In this paper, a new methodology has been proposed, for the calculation of the optimum switching instants for current interruption cases. The methodology is based on the Current Injection Method that eliminates the need for exhaustive simulations for an initial assessment of the performance of a possible controlled switching application. Circuit-breaker s characteristics, like contact operation time scatter and deviation of the slope of the contact gap voltage withstand characteristic are taken into account in this method. References: [1] CIGRE WG13.07, Controlled Switching of HVAC Circuit-Breakers - Planning, Specification and Testing of Controlled Switching Systems, Electra No 197, pp , August [2] CIGRE WG13.07, Controlled Switching of HVAC Circuit-Breakers - Guide for Application Lines, Reactors, Capacitors, Transformers, 1 st Part: Electra No 183, April nd Part: Electra No 185, August [3] CIGRE Task Force , Controlled Switching: A State-of-the-Art Survey, 1 st Part: Electra No 163, December nd Part: Electra No 164, February [4] C.D. Tsirekis, N.D. Hatziargyriou, B.C. Papadias, Controlled Switching Based on the Injection Method, International Conference on Power Systems Transients (IPST 97), Vol. II, pp , Rio de Janeiro, Brazil, June [5] CIGRE WG13.02, Interruption of Small Inductive Currents - Chapter 3, Part A, Electra No 75, March [6] Leuven EMTP Center, ATP Rule Book, June 1993, Leuven (Belgium). [7] Bonneville Power Administration, EMTP Theory Book, Oregon (Portland), [8] W.M.C. Van Den Heuvel, B.C. Papadias, Interaction Between Phases in Three-Phase Reactor Switching, 1 st Part: Electra No 91, pp , Dec nd Part: Electra No 112, pp , May ISBN: