WO2008031364A1

WO2008031364A1 - Method for sampling a current or voltage profile and for forming sampling values, in particular for use in protection and control devices for energy transmission systems

Info

Publication number: WO2008031364A1
Application number: PCT/DE2006/001414
Authority: WO
Inventors: Andreas Jurisch; Torsten Kerger
Original assignee: Siemens Aktiengesellschaft
Priority date: 2006-08-07
Filing date: 2006-08-07
Publication date: 2008-03-20
Also published as: CN101506668A; DE112006003960A5; CN101506668B

Abstract

The invention relates to a method for sampling a current or voltage profile (U(t)) and for forming sampling values (U'(n)) having a temporal relationship with a predefined system clock (G). According to the invention, an auxiliary clock (H) which runs freely in relation to the system clock is used to form auxiliary sampling values (U(n)), the time offset of the sampling time of each auxiliary sampling value in each case in relation to the system clock is determined, time offset values (Δt(n)) specific to auxiliary sampling values are determined, and the auxiliary sampling values (U(n)) are re-sampled using the time offset values (Δt(n)) specific to auxiliary sampling values, with the sampling values (U'(n)), which have a temporal relationship with the system clock, being formed in the re-sampling.

Description

description

Method for sampling a current or voltage profile and for forming samples, in particular for use in protective or conducting devices for energy transmission systems

The invention relates to a method, in particular for use in protective or guiding devices for energy transmission systems, with the features according to the preamble of claim 1.

^■ Such a method is known from German laid-open specification 199 33 684. In this method, a high voltage or high current signal is sampled using an internal clock which is synchronized with the globally available GPS (Global Positioning System) signal. The GPS signal thus provides a system clock to which the samples refer. The samples are subsequently evaluated for protection purposes, in particular for differential protection purposes.

The invention has for its object to provide a method with which can be formed at low equipment or hardware costs samples that have low jitter and are very low noise.

This object is achieved on the basis of a method of the type specified according to the invention by the characterizing features of claim 1. Advantageous embodiments of the method according to the invention are specified in subclaims.

According to the invention, it is provided that auxiliary samples are formed with an auxiliary clock signal relative to the system clock, the time offset of the sampling instant of each auxiliary sample is determined relative to the system clock, and auxiliary sample individual time offset values are determined, and the auxiliary samples resampling using the sub-sample individual time offset values at which the samples related to the system clock are formed.

A significant advantage of the method according to the invention is the fact that overall less hardware is required for generating the samples than in previous sampling systems, namely because the used for sampling auxiliary clock runs freely and thus on a phase tracking, for example on the basis of a PLL (Phase Locked Loop) circuit, can be completely dispensed with. The scanning thus takes place with a freewheeling clock and the resampling, for example, "only" by calculation using the determined time offset values.

Another significant advantage of the method according to the invention is that the samples are very low in jitter and noise. It has been found by the inventor that a sampling clock derived from a given system clock and phase-coupled with it may have a relatively large jitter and lead to a relatively large noise of the sampled values. At this point, the invention starts by providing that such a phase coupling or phase tracking is dispensed with; instead sampling is performed independently of the system clock and a high-quality, especially low-jitter, auxiliary clock free-running over the system clock is used for an upstream auxiliary scan. The AbtastZeitpunkte the auxiliary samples are thus compared to the system clock usually arbitrarily offset in time or phase-shifted; in order to still produce a ^'temporal reference to the system clock, the Hilfsabtastwerte are then subjected to re-sampling or a resampling, and using the hilfsabtastwertindividuellen time offset values, which have been previously recorded at the Hilfsabtasten for each auxiliary sample. As a result, samples are formed with the correct time base, but are significantly less jitter and less noise, because the system clock is used only in the context of downstream, for example, only numerically carried out, "resampling" (resampling), but not for the primary "physical see" sampling.

Preferably, the system clock used is the GPS signal, the Galileo signal, a real-time SNTP signal, a system clock of an SDH network or a system clock signal derived with one or more of these signals.

As an auxiliary clock signal, the output signal of a local free-running oscillator is preferably used. The auxiliary clock signal can be generated for example with a Sammelschienenschutz- central unit and fed with this in connected to scanning devices. The scanning devices may for example be integrated in transducers which are connected to one or more energy transmission lines.

Alternatively, the auxiliary clock signal can be derived from the signals of a real-time transmission network, in particular a real-time ETHERNET network.

The time offset values are formed, for example, with a counter which counts the clock pulses of a predetermined count clock and which is reset by pulses of the system clock and whose count is read out as a time offset value when an auxiliary sample is taken with the auxiliary clock. The count is illustratively a measure of the amount of time that has elapsed between the last pulse of the system clock and the acquisition of the respective auxiliary sample.

The samples formed by the method described are further processed, for example, into current or voltage vectors, preferably when the sampled values increase Protective purposes in protection or control devices are further processed.

The invention also relates to an arrangement for sampling a current or voltage profile and for forming samples having a temporal reference to a predetermined system clock.

With regard to such an arrangement, the object of the invention is to achieve that the samples have a low jitter and are as low-noise as possible.

According to the invention, this object is achieved by an arrangement with an auxiliary scanning device and a resampling device: the auxiliary scanning device is supplied with the system clock and with an auxiliary clock free-running relative to the system clock and forms auxiliary samples and auxiliary sample time-offset values with the auxiliary clock; which indicate the time offset to the system clock for each sampling time. Subsequent to the sub-sampler is the resampler, which resamples the sub-samples of the sub-strobe using the sub-samples of the sub-samples at which the samples relating to the system clock are formed.

With regard to the advantages of the arrangement according to the invention, reference is made to the above statements in connection with the method according to the invention.

Preferably, the auxiliary sampler is included in a current or voltage converter when forming samples relating to a power transmission system. The resampling device can for example form part of a field device, in particular a protective device.

The auxiliary scanning device and the resampling device are preferably connected to each other via a data transfer bus or a data transmission network.

If a busbar protection central unit is present in the energy transmission system, it is considered advantageous if the busbar protection central unit forms the auxiliary clock and transmits this to the auxiliary scanning device.

With regard to a pointer measurement, it is considered advantageous if a resampler is connected to the resampler, which forms current or voltage vectors with the samples.

The invention will be explained in more detail with reference to exemplary embodiments; thereby show by way of example

FIG. 1 shows a first exemplary embodiment of an arrangement for generating sampled values, by means of which the method according to the invention is also explained by way of example,

FIG. 2 shows by way of example a sampled voltage signal,

FIG. 3 schematically shows the reconstruction of a voltage signal,

FIG. 4 shows a resampling process with an interpolation filter in a schematic representation,

6 shows the amplitude response associated with an impulse response of an interpolation filter with a Blackman window, FIG. 7 shows the amplitude response of the interpolation filter according to FIG. 6 in the passband, FIG. 8 shows exemplary sampling times, FIG. 9 shows a table with relevant mathematical properties of the interpolation filter according to FIG. 6, FIG.

FIG. 10 shows a flowchart which summarizes the method for forming samples once more by way of example, and

Figure 11 shows another embodiment of an inventive arrangement.

FIG. 1 shows an arrangement 10 for scanning a current or voltage profile and for forming a corresponding current or voltage vector. The arrangement is explained by way of example for the case in which a voltage signal U (t) is sampled and a voltage vector U is to be formed.

The arrangement 10 has an auxiliary scanning device 20, to the input E20a of which the voltage signal U (t) to be sampled is fed. A further input E20b of the auxiliary sampling device 20 is connected to an auxiliary clock generator 30 in conjunction, which on the output side generates a freewheeling auxiliary clock or auxiliary clock signal H and feeds it into the auxiliary scanning device 20.

The auxiliary scanning device 20 has on the input side an A / D converter 40 whose analog input is connected to the input E20a of the auxiliary device 20 and thus supplied with the voltage signal U (t). Downstream of the A / D converter 40 is a measured value generator 50 which transmits samples U (n) formed by the A / D converter 40 to a downstream measured value feed 60.

The measured value generator 50 is also connected to a memory or latch module 70 in connection, on the input side, a counter 80 is connected upstream. The counter 80 is connected with its reset input R80 to a GPS receiver 90, which generates a GPS signal as a system clock or system clock signal G and feeds it into the counter 80. The system For example, clock G generates a short pulse per second.

A clock input T80 of the counter 80 is connected to a signal generator 100, which feeds a clock signal or counting clock T into the clock input T80 of the counter 80.

An output A20 of the auxiliary scanning device 20 is formed by an output A60 of the measured value memory 60; this is connected to a resampling device (or resampling device) 200, which is followed by a pointer-forming device 210.

The resampler 200 resamples the auxiliary samples U (n) of the A / D converter 40 and produces, on the output side, resampled voltage values U '(n) which are subsequently converted by the pointer generator 210 to the voltage vector U; the voltage vector U thus refers to the system clock G of the GPS receiver 90 and forms a so-called synchro-pointer. By the term synchro pointer is meant that the pointer U is related to the system clock.

The arrangement according to FIG. 1 is operated, for example, as follows:

With the GPS receiver 90, the system clock G is generated on the output side, which is fed to the counter 80. At the same time the counting clock T is fed to the signal generator 100 in the counter 80, which then begins to count. Whenever the system clock G generates a pulse, the counter 80 is reset.

The count Z of the counter 80 is then stored in the memory module 70 when it is triggered by the free-running auxiliary clock H. As soon as the auxiliary clock H generates its n-th clock pulse, both the A / D converter 40 is thereby triggered and a sample value U (n) is generated as At the same time, the respective counter reading Z (n) is recorded in the memory module 70 and transmitted to the measured value generator 50 as an auxiliary-sample-individual time offset value .DELTA.t (n). If the counting clock has a clock rate fT = 1 GHz, the time offset value Δt (n) can be determined relative to the pulses of the system clock G with a time accuracy of 1 ns.

The counter 80 and the memory module 70 thus together form a counter which forms the time offset values Δt (n) by counting the clock pulses of the count clock T, resetting each time a pulse of the system clock G is present, and outputting its count Z as the time offset value as soon as the auxiliary clock H an auxiliary sample U (n) is recorded.

The measured value generator 50 transmits the respective auxiliary sampling value U (n) of the A / D converter 40 together with the assigned auxiliary sample time offset value Δt (n) to the measured value memory 60, which stores the auxiliary sampling value U (n) with the associated auxiliary sample individual time offset value Δt (n) as measured value pair (U (n); Δt (n)) stores.

Although the auxiliary sample U (n) has not been sampled at the system clock G, it nevertheless has a known phase relationship with the system clock G, namely, for each auxiliary sample U (n) the respective time reference to the system clock G or to the each preceding pulse of the system clock G is defined by the auxiliary sample individual time offset value Δt (n). In addition to the counter value Z, Δt (n) or with the

Time offset value .DELTA.t (n) also an absolute time TAS are stored, indicating the time or the time of the last pulse of the system clock absolute; this makes it possible to absolutely determine the sampling timing TAH for each auxiliary sample according to: TAH = TAS + Z * 1 / fT The resampler 200 reads out the auxiliary samples U (n) along with the associated sub-sample individual time offset values Δt (n) from the sampler 60 and performs "computational" resampling using the system clock G. As part of this resampling process, sampled samples U '(n) which relate to the system clock G, for which a resampling rate which corresponds to an integer multiple of the system clock G is preferably used, and the samples U' (n) can then be used in the pointer-forming device 210

Form voltage vector U, which forms a so-called synchro-pointer. The term "synchro-pointer" is - as already explained - in this context, a pointer to understand that is synchronized with the predetermined system clock G.

FIG. 2 shows an example of the voltage signal U (t) sampled by the freewheeling auxiliary clock H at a fixed sampling rate.

In order to be able to calculate a pointer U as accurate as possible for characterizing the fundamental wave of the voltage signal U (t), the sampling frequency of the auxiliary clock H is preferably set such that it is an integer multiple of the frequency of the fundamental wave of the voltage signal U (t). In order to meet this requirement, for example, a frequency measurement of the fundamental wave of the voltage signal U (t). A multiple of the fundamental frequency determined in this way is subsequently used as auxiliary clock H.

As already mentioned, a resampling method is used to match the samples U (n) to the system clock G. The resampling method is based on the theory of Shannon's sampling theorem. This theorem states that any ideal band-limited signal x can be interpolated using the si function:

" ⁼ - ^∞ equation 1 τ ._ ,. sin (π 'f _A ' t) with hJt) = - * - ^ - '-

FIG. 3 illustrates this state of affairs on the basis of a detail of a sine function with the indicated si-interpolation functions, which enable a reconstruction of sampled band-limited signals. So you can calculate an ideal band-bound function given in points 3, 4, 5, ..., etc. at any other place by solving Shannon's equation for those points.

There are two ways to interpret this function for a resampling application:

The first possibility is to lay the interpolation function - as shown in FIG. 3 - with its extreme value through the given sampled values. However, this variant is not very practical for a resampling application, since here an infinite sum over all si functions would have to be calculated for each sample.

According to a second variant, the fact that the si function assumes the value 0 at the location of all other samples, except for the sample to be interpolated, is taken advantage of. That is, if the signal having the si function is interpolated at the locations of the new samples, then only one si function needs to be considered.

If the si function is now positioned at the location of the new sample with the period of the new sample interval and calculated at the locations of the given samples, then the new sample can be determined. FIG. 4 illustrates this situation on the basis of a resampling by means of an interpolation filter. Of course, this interpolation should of course not be performed directly with the si function, since this function has a non-zero value not only in a finite time interval. However, simply cutting off the function at the fifth zero crossing on both sides of the extreme value provides a frequency response which deviates significantly from the ideal lowpass in the frequency domain. Here it is better to limit the si function with a window function to a finite range. In the present example, the Blackman window function is used for this. Here you reach one

Locking damping ratio of min. 6OdB. Better values could be achieved with the well-known Kaiser function; however, this is at the expense of the window size. Since 60 dB can be considered sufficient, preferably no Kaiser window should be used.

Figure 5 shows the interpolation function used in more detail. The curve denoted by "xi" represents the exemplarily selected Blackman window function. The curve denoted by "si ^w " shows the non-time limited si function, and the curve labeled "hsi" is the time-limited interpolation function by means of the window function.

FIG. 6 shows the amplitude response associated with the impulse response shown using the example of a sampling rate reduction from 10 kHz to 5 kHz. As can be seen in FIG. 6, a band limitation of the interpolated signals takes place to the new half sampling frequency. The waviness in the Durσhlassbereich is a maximum of 0.2 10 ^"4 .

FIG. 7 shows the characteristic in the passband. It can be seen in FIG. 7 that on average the ripple in the entire passband is significantly smaller than 10.sup.- ⁴ , which means that all accuracy requirements can be met in the passband as well Calculate the function value for each time point t ₀ using the following equation:

,,, / \ sin (π - f _A - (L - n - T,))

and w (t) = 0.42 - 0.5 • cos (2π (t ₀ -n ■ T _Λ )) + 0.08 • cos {4π (t ₀ ^{-n ■} T _Λ ))

This equation is the solution of a least squares estimator that fits the interpolation function at the desired location t ₀ into the received sample stream. The function w (t) is the Blackman window function.

FIG. 8 shows the realization of the resampling for the selected resampling method: At the points in time marked with a cross, a sample is present in the original sample stream on the basis of the auxiliary clock H. At the points marked with a circle should be rescanned. For resampling, use is made of a sampling rate which is synchronized with the system clock G and preferably corresponds to a multiple, for example an integer multiple, of the pulse rate of the system clock G. By the term "GPS sync pulse" is exemplified one of the pulses of the GPS signal, of which usually occurs one per second.

The ratio of the new to the old sampling frequency is 3/5 in this example. If the sample is to be calculated at time t = 0, then an interpolation filter whose extreme value is exactly at the original sample value 0 is used. The data window in this case extends exactly w / 2 samples into the past and w / 2 samples into the future. For the calculation of the new sample at t = 3 an interpolation filter must be used, the extreme value has been postponed by exactly 3.5 T _Ä the original Abtastin- tervalls T _A to the right. In sum, therefore, five different filter coefficient sets are needed to implement the resampling in this example. The formulas in the table according to FIG. 9 describe the proposed algorithm in detail.

FIG. 10 once again shows the exemplary process flow as a flowchart.

Decisive for the accuracy of the determined synchro-pointer U is the exact determination of the fundamental signal frequency f of the input signal U (t). For this purpose, preferably the frequency f is determined from pointers of several consecutive data windows according to the following formula:

From the development of the phase angle φ (t), the

Derive the validity of the frequency measurement f. The determined frequency measurement value f is valid when the phase angle develops almost linearly. If sudden changes in the course of the measured phase angle are detected, the data window for the frequency measurement must be repositioned. With a now valid frequency measurement value f, the resampling is performed a second time in a second step. The pointer thus obtained can then be used as a synchro pointer. In the recirculation steps described, the method is continued further and further.

The concrete technical implementation of the resampling method in the resampling device 200 can be carried out, for example, in a manner as described in "A flexible sampling rate conversion method" (Smith, J. O. and

Gösset, P .; 1984; Proceedings of the International Conference on Acoustic, Speech, and Signal Processing, Sand Diego, Volume 2, pages 19.4.1.-19.4.2; New York. IEEE Press).

FIG. 11 shows a further exemplary embodiment of an arrangement according to the invention with which the method according to the invention can also be carried out. FIG. 11 shows a power transmission line 500 to which two transducer units 510, 520 are connected. The two measuring transducer units 510 and 520 are each equipped with an auxiliary scanning device 20, as has already been explained in connection with FIG.

The two transducer units 510 and 520 communicate with a downstream protection device 530; the connection between the protection device 530 and the two transducer units 510 and 520 is ensured by a data bus or a data transmission network (eg, Ethernet network) 540, which may be equipped, for example, with a switch 550.

As can also be seen in FIG. 11, the two measuring transducer units 510 and 520 as well as the protective device 530 are subjected to a system clock G, which is predetermined, for example, by the GPS signal.

The protection device 530 is equipped with a resampler 200 and a pointer generator 210; the two components may correspond to the resampler 200 and the pointer generator 210 of FIG.

The arrangement according to FIG. 11 can be operated as follows:

The 510 and 520 transducers are used for power and power

Voltage measurements Ul (t) and Il (t) and U2 (t) and 12 (t) of the power transmission line 500 detected and analog / digital converted. In this case, according to an auxiliary clock H, auxiliary sampled values Il (n), 12 (n), Ul (n) and U2 (n) are formed which, together with the respectively assigned auxiliary sampled value individual time offset values Δt (n), are sent via network 540 to Protective device 530 are transmitted. The auxiliary clock H can be individually generated in each of the transformer units 510 and 520. or are specified externally for both transducer units 510 and 520; for example, the auxiliary clock H is formed by a busbar protection central unit 600 and fed into the transducer units 510 and 520. Alternatively, the auxiliary clock H can be derived from the data signals of the network 540, which is preferably a real-time transmission network, in particular a real-time ETHERNET network.

The protection device 530 subjects the auxiliary samples Ul (n),

U2 (n), Il (n) and 12 (n) resampling and generates on the output side samples Ul '(n), U2' (n), II '(n) and 12' (n), with the system clock G are synchronized. With these samples Ul '(n), U2' (n), II '(n) and 12' (n), the pointer-forming device 210 of the protection device 530 forms sync pointers IJL, 12, Ul and U2_.

The synchro-pointers IJL, V2_, UJL_ and U2 can be evaluated within the protection device 530 or outside the protection device 530, for example in a downstream control system or the like, as to whether an error, in particular a short circuit, has occurred on the energy transmission line 500 , With the synchro-pointers IJL, JC2_, UJL and U2 error signals can be generated.

Claims

claims

A method for sampling a current or voltage profile (U (t)) and for forming samples (U '(n)) which have a temporal reference to a predetermined system clock (G), characterized in that - with a relative to System clock freewheeling auxiliary clock

(H) auxiliary samples (U (n)) are formed, the skew of the sampling time of each

Auxiliary sample is respectively determined relative to the system clock and auxiliary sample individual time offset values (Δt (n)) are determined and

-substituting the auxiliary samples (U (n)) using the auxiliary sample individual time offset values (Δt (n)) at which the samples (U '(n)) relating to the system clock with respect to time are formed.

2. Method according to claim 1, wherein the GPS signal, the Galileo signal, a real-time SNTP signal, a system clock of an SDH network or a system clock signal derived with one or more of these signals is used as the system clock.

3. Method according to one of the preceding claims, characterized in that a clock signal of a local free-running oscillator (30) is used as an auxiliary clock.

4. The method according to any one of the preceding claims 1 - 2, characterized in that the auxiliary clock is obtained from a clock signal of a real-time transmission network, in particular a real-time ETHERNET network.

5. Method according to one of the preceding claims, characterized in that the time offset values (Δt (n)) are formed with a counter (70, 80) which counts the clock pulses of a predetermined counting clock (T) and which are each represented by pulses of the

System clock (G) is reset and their count (Z) is read out as a time offset value (.DELTA.t (n)) when the auxiliary clock (H) an auxiliary sample (U (n)) is recorded.

6. Method according to one of the preceding claims, characterized in that current or voltage vectors (U) are formed with the sampled values.

7. An arrangement for sampling a current or voltage curve (U (t)) and for forming samples (U '(n)), which have a temporal reference to a given system clock (G), d a d u c h e c e n c i n e t e that

- one with the system clock (G) and with a free-running relative to the system clock auxiliary clock (H) applied Hilfsabtasteinrichtung (20) for forming Hilfsabtastwer- th (U (n)) and for forming hilfsabtastwertindividuellen ^'time offset values (At), each of the Specify the time offset of the sampling instant of the associated auxiliary sample relative to the system clock, and

- resampling means (200) arranged downstream of the auxiliary scanning means (20) and resampling the auxiliary samples (U (n)) of the auxiliary scanning means (20) using the auxiliary sample time offset values (Δt) at which the samples (U '(n )), which relate in time to the system clock (G).

8. Arrangement according to claim 7, characterized in that the auxiliary scanning device forms part of a current or voltage converter (510, 520) of an energy transmission system.

9. Arrangement according to one of the preceding claims 7-8, in that a re-sampling device (200) forms part of a field device (530), in particular a protective device, of an energy transmission system.

10. An arrangement according to any one of the preceding claims 7-9, wherein auxiliary scanning means and resampling means are interconnected via a data transfer bus or a data transmission network (540).

Arrangement according to one of the preceding claims 7-10, characterized in that the auxiliary sampling means comprises counting means (70, 80) constituting the skew values (Δt (n)) by counting the clock pulses of a predetermined counting clock (T) Counter reading (Z) in each case resets in the presence of a pulse of the system clock (G) and outputs as a time offset value their respective count (Z) when the auxiliary clock (H) an auxiliary sample (U (n)) is recorded.

12. Arrangement according to one of the preceding claims 7-11, in that a pointer-forming device (210) is connected to the resampling device (200), which is connected to the

Samples (UMn)) Current or voltage vector (U) forms.