JPS6346648B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a frequency detection method, and more particularly to a frequency detection method for an alternating current input amount in a power system.

Until now, the zero point approximation method has been known as a method for detecting the frequency of alternating current in power systems, but the detection accuracy is not sufficient when detecting the frequency using this method, so it is necessary to improve the detection accuracy. has the disadvantage that the sampling frequency must be set high.

FIG. 1 shows the relationship between the AC input waveform and the sampling values when frequency is detected by the zero point approximation method. As shown in the figure, the AC input waveform is sampled at fixed time intervals ΔT, and T1 to T34 are sampling points, and _v1 to T34 are sampling points.
v ₃₄ indicates the sampling value. In this zero point approximation method, the zero point is found approximately. If the product of two consecutive sampling values is negative, it is determined that a zero point exists between the sampling points corresponding to those sampling values, and the zero point is determined by linear approximation from these sampling values. Therefore, the sampling values v ₁ , v ₂ , v ₅ ,
Each set of v ₆ , v ₉ , v ₁₀ ... is used for zero point approximation.

Here, the times Δt ₁ and Δt ₂ from the zero point determined by the linear approximation to the sampling points T1 and T2 can be obtained by simple proportional calculation as follows.

That is, Δt ₁ = |v ₁ |・ΔT/(|v ₁ |+v ₂ ) ...(1) Δt ₂ =v ₂・ΔT/(|v ₁ |+v ₂ ) ...(2). It is clear that Δt ₅ , Δt ₆ , Δt ₉ , and Δt ₁₀ can be similarly determined as follows.

Δt ₅ = v ₅ · ΔT / (v ₅ + | v ₆ |) ...(3) Δt ₆ = | v ₆ | · ΔT / (v ₅ + | v ₆ |) ... (4) Δt ₉ = | v ₉ ｜・ΔT/(｜v ₉ ｜+v ₁₀ ) ……(5) Δt ₁₀ =v ₁₀・ΔT/(｜v ₉ ｜+v ₁₀ ) ……(6) Therefore, the frequency is detected once per cycle It is clear that the period can be expressed as Δt ₂ +Δt ₉ +7·ΔT. The term 7·ΔT is the time between sampling points T2 and T9. Therefore, the detection frequency can be obtained from the following equation (7).

= 1/(Δt ₂ + Δt ₉ + 7・ΔT) ...(7) Also, if the frequency is detected once every half cycle, the half cycle is expressed as Δt ₂ + Δt ₅ + 3・ΔT, so the detection frequency is can be obtained from the following equation (8).

=1/{2(Δt ₂ +Δt ₅ +3・ΔT)} ...(8) By finding one cycle or half cycle in this way, it is possible to detect the frequency every cycle or every half cycle. However, it is clear that the frequency detection accuracy depends on Δt ₂ , Δt ₅ , Δt ₉ , etc. In other words, in order to improve the calculation accuracy of Δt ₂ , Δt ₅ , Δt ₉ , etc. and increase the frequency detection accuracy, the frequency of the sampling frequency must be increased. Furthermore, in the case of the above method, since the detection speed and the detection accuracy are in a contradictory relationship, the conventional method has the disadvantage that it is not possible to detect the frequency at high speed and with good accuracy.

An object of the present invention is to provide a frequency detection method that can detect frequencies with high detection speed and high accuracy.

For this purpose, in the present invention, each time a zero point is detected, the time from that zero point to the zero point before an arbitrary multiple of cycles in the past is determined, and the frequency is detected from this time.

The present invention will be explained below with reference to FIGS. 1 to 3.

First, before going into a specific explanation of the present invention, the principle of the present invention will be explained. Δt ₁ and Δt ₂ obtained in the above equations (1) and (2) are obtained by dividing ΔT proportionally by the absolute values of v ₁ and v ₂ , and the sum of Δt ₁ and Δt ₂ is equal to ΔT. This relationship also holds true in equations (5) and (6), and the sum of Δt ₉ and Δt ₁₀ found here is equal to ΔT. However, when calculating the sum of Δt ₂ and Δt ₉ in equation (7), it does not become ΔT. By the way, if the sampling frequency is set to 2n times the frequency of the alternating current (n is 4 in the example in Figure 1), then if the frequency of the alternating current is constant at the rated value, the sampling phase will periodically change to the same phase. If you calculate the sum of Δt ₂ and Δt ₉ , this should be ΔT, but in reality there are minute phase differences, conversion errors during analog to digital conversion, and linear approximation errors. By |(Δt ₂ +
An error of Δt ₉ )−ΔT| occurs. At this time, the error ratio α to ΔT is as follows.

α = | (Δt ₂ + Δt ₉ ) − ΔT | / ΔT ...(9) The error α in equation (9) is the term Δt ₂ + Δt ₉ in equation (7),
This is included in the term Δt ₂ + Δt ₅ in equation (8),
Assuming that this error is constant for these two equations, and finding the frequency calculation accuracy for each equation, we get:
The denominator of equation (7) (Δt ₂ +Δt ₉ +7ΔT) is approximately twice the value of the denominator of equation (8) (Δt ₂ +Δt ₅ +3ΔT). Since this value includes the error α, it can be said that frequency detection according to equation (7) with a large denominator value is more accurate than that according to equation (8). If this logic is extended, the detection accuracy will be improved as the interval required for detection becomes longer. However, it is clear that as the detection accuracy improves, the detection speed decreases inversely. The only way to maintain detection accuracy above a certain level and to perform frequency detection every half cycle is to perform detection as in the present invention.

Now, the present invention will be specifically explained with reference to FIG.
For example, if the length of the section required for detection is 2 cycles, the data required for frequency detection immediately after sampling point T18 is Δt ₂ and Δt ₁₇ , as well as
This is the time between sampling points T2 and T17. Therefore, the frequency in this case is obtained from the following equation (10).

= 1/1/2 (Δt ₂ + Δt ₁₇ + 15ΔT) ...(10) Similarly, the data required for frequency detection after half a cycle is Δt ₆ , Δt ₂₁ , and the time between sampling points T6 and T21. Therefore, the frequency detected immediately after the sampling point T22 is =1/1/2(Δt ₆ +Δt ₂₁ +15ΔT) (11). Furthermore, sampling points T26, T30, T
The frequencies detected immediately after 34 are each expressed by the formula
It is obtained as (12), (13), and (14).

=1/1/2 (Δt ₁₀ +Δt ₂₅ +15ΔT) ...(12) =1/1/2(Δt ₁₄ +Δt ₂₉ +15ΔT) ...(13) =1/1/2(Δt ₁₈ +Δt ₃₃ +15ΔT) ... ...(14) That is, the frequency is detected every half cycle of the minimum interval, but since the detection interval is lengthened to 2 cycles, it is possible to perform high-speed detection and high-accuracy detection. .

If the general formula is extracted from formulas (10) to (14), it can be expressed as follows.

= 1/1/2 (Δt′+Δt″+15ΔT) ……(15) This equation (15) is for the case where the detection interval is 2 cycles, but it can be similarly applied in the case of 3/2 cycles as follows. can be expressed as

=1/2/3(Δt′+Δt″+11ΔT) (16) Furthermore, the general formula in the case where the detection interval is set to N times the half cycle as a reference is as follows.

=1/N/2(Δt'+Δt''+KΔT) =N/2(Δt'+Δt''+KΔT)...(17) In this case, the coefficient K of ΔT can be expressed as a function of N.

K=3+4(N-1)...(18) Therefore, the frequency can be found generally from equations (17) and (18), but equation (18) is determined by the sampling frequency. . The above equation (18) is for the case where the frequency of the alternating current amount is 50 Hz and the sampling frequency is 400 Hz. If the AC amount is 50 Hz and the sampling frequency is 300 Hz or 600 Hz, K = 2 + 3 (N-1), K = 5 + 6 (N-1), and the same applies for other sampling frequencies. K can be found as a function of N.

Although the present invention is as described above, one possible application of the present invention is, for example, a frequency relay.
FIG. 2 shows an outline of the flow in that case.

According to this, step 1 is whether or not there is a setting change request.
If a setting change request is made, the setting values of N and frequency are set in step 2, and then the process returns to step 1. If it is determined in step 1 that there is no request for setting change, then in step 3 the sampling value v _i obtained at the sampling point is fetched and stored. Next, in step 4, the currently captured sampling value v _i and the sampling value v _i-1 one sample before are integrated, and whether the integration result is positive or negative is determined. This integration and the determination of the polarity of the integration result are, of course, performed to detect the presence of a zero point and to approximate the zero point to a straight line. If the integration result is determined to be positive in step 4, the counters K ₁ to K _o are incremented by one step in step 5 to count the number of ΔT, and then in step 12 the subsequently obtained sampling values are loaded. They will be waiting for you as long as possible.

Also, if the integration result is determined to be negative in step 4, the sampling values around the zero point are
In step 6 _, Δt _{1 ,}
After Δt ₂ has been determined, Δt ₁ and Δt ₂ are stored in step 7. The frequency is determined in step 8 from this Δt ₁ and Δt ₂ (expressed as Δt2 _o ) immediately after the zero point N times half cycles ago. After this, in step 9, the counter _K1 is cleared, and the remaining counters _K2 to _K0 are incremented by one step, and then in step 10, it is determined whether the frequency obtained in step 8 is greater than or equal to the set value. That's why. If it is greater than the set value, an output to that effect is issued in step 11. In this case, the frequency obtained in step 8 is used for display and recording, as well as for detecting variations from the set value, and the variation detection output is used as a control output for various controls externally. be.

FIG. 3 shows the general configuration of the frequency relay. As shown in the figure, the AC input is sampled at a predetermined period by a sample and hold circuit 15 via an analog filter 14 for aliasing errors, and then converted into a digital value by an A/D converter 16. This digitized sampling value is taken into an arithmetic processing unit 18 such as a computer via an input device 17, and is subjected to the processing as described above. The detected frequency values and judgment results obtained through this process are output to the output device 19.
It is designed to be taken out to the outside via a holder and used for display and recording, as well as for various types of control. Note that the control circuit 13 includes an analog filter 1
This is to provide control timing to circuits and devices other than 4.

As explained above, each time a zero point is detected, the present invention calculates the time from that zero point to the zero point of an arbitrary multiple of half cycles in the past based on previously stored information, and detects the frequency from this time. This is how it was done. Therefore, according to the present invention, it is possible to detect the frequency every half cycle, and it can be detected with high accuracy, so frequency measurement in the grid and detection of frequency changes due to changes in grid phenomena are highly reliable. It has the effect of being able to do it.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram of the case where the frequency is detected by determining the zero point by linear approximation from the sampling value of the alternating current amount, and Fig. 2 is an illustration of the frequency relay to which the frequency detection method according to the present invention is applied. FIG. 3, which is a diagram showing the processing flow, is a diagram showing the general configuration of the frequency relay. ΔT...Sampling interval, T1 to T34...
Sampling points, v ₁ to v ₃₄ ... sampling values.

Claims (1)

[Claims] 1 The existence of a zero point is detected from the polarity change of the sampling value obtained by sampling the AC input amount at a predetermined period, and the position of the point is determined by linear approximation from the sampling values obtained immediately before and after the point. At the same time, by determining the time from this point to the sampling point corresponding to the sampling value immediately before and after the above, and determining the time related to the cycle from this time and the number of sampling points included between the zero points. In the frequency detection method for detecting the frequency of the AC input amount described above, when detecting the frequency every time a zero point is detected, the time from the zero point before the arbitrary multiple half cycle to the sampling point immediately after that point, from the zero point related to current detection A time corresponding to an arbitrary multiple half cycle is determined from the time up to the sampling point immediately before the point and the number of sampling points included between the two zero points, and the frequency of the AC input amount is determined based on the time. frequency detection method.