JPH0449660B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a digital frequency detection method for calculating the frequency of an input signal through arithmetic processing using sampling values of the input signal.

[Technical background of the invention]

A pulse count type frequency detector is widely known as a frequency detector for detecting the frequency of an input signal. This method prepares a high-precision, high-frequency pulse oscillator, counts the number of pulses from the pulse oscillator during a half period or one period of an input signal, and converts the counted value into a frequency.

[Problems with background technology]

However, such conventional pulse counting methods require a high-precision pulse oscillator to increase frequency detection accuracy, and a means to accurately detect the zero point of the input signal in order to accurately detect a half cycle or one cycle. required. Therefore, it has the drawbacks that the circuit becomes complicated and the detection device becomes expensive.

[Purpose of the invention]

The purpose of this invention is to detect frequencies ranging from direct current to the rated frequency without the need for a high-precision pulse oscillator, and without requiring a complex circuit for detection. It is an object of the present invention to provide a frequency detection method that can detect frequencies with high precision by simple arithmetic processing of data obtained by sampling.

[Summary of the invention]

In the frequency detection method according to the present invention, an input signal varying from direct current to a rated frequency of ₀ is sampled with a sampling signal having a frequency _S higher than the rated frequency of ₀ , and when detecting the frequency of the input signal, The number of samplings in one cycle period 2n (n: integer) is a predetermined value
When D is greater than or equal to ₀ , the frequency is determined to be zero,
The number of samplings 2n is smaller than the value D ₀ and n _{0 0} / (2n + 1)・2n≦k (k is a predetermined detection error value)
When satisfying the following, calculate according to = n ₀ / n ₀ (n ₀ : integer) (where n ₀ is the number of samplings during a half period of the rated frequency ₀ ), and if the number of samplings 2n is smaller than the value D ₀ and n When _{0 0} /(2n+1)・2n>k is satisfied, the initial phase value θ ₁ is obtained as θ ₁ = n _{0 0} /n _S ×360°, and then a predetermined value is determined according to this initial phase value θ ₁ . Using the selection data table, correct the set of adjacent sampling values (x ₀ , x ₁ ) and (x _o , x _o+1 ) before and after the polarity changes, and use this corrected sampling value. θ=180°−θ ₁ (|X ₁ |/|X ₀ |+|X ₁ |+|X _o |/|
X _{o | + |} It is a feature.

[Embodiments of the invention]

FIG. 1 shows a block diagram of an apparatus for carrying out the frequency detection method according to the invention. The voltage of system 1 is reduced to signal level via potential transformer 2 and becomes the input signal. The frequency detection device 10 is composed of an analog input circuit 3, a microprocessor 4, a digital output circuit 5, and a display 6. The input signal input to the analog input circuit 3 is sampled at a predetermined sampling frequency _S and The signal is converted from analog to digital and input to the microprocessor 4. The sampling data is processed by the microprocessor 4 to determine the frequency of the input signal, and after being converted from binary to decimal by the digital output circuit 5, it is displayed numerically on the display 5.

Next, the operation of the frequency detection device 10 will be explained in more detail. FIG. 2 is a diagram for explaining the sampling of the input signal, and FIG. 3 is a flowchart for explaining the contents of the arithmetic processing of the frequency detection device 10. The input signal is sampled by the analog input card 3 at a sampling frequency _S ,
Sampling values x ₀ , x ₁ ... x _o ... are obtained. This sampling is performed over a period of time slightly longer than one cycle of the input signal, and the sampled values are temporarily stored in the memory within the microprocessor 4 (step 101).

Next stored sampling value x ₀ , x ₁ ... x _o+1
It is determined whether the polarity of the first data is positive (step 102). If the judgment result is true, that is, positive, a pair of consecutive positive and negative sampling values x ₀ and x ₁ when the polarity of the sampling value changes from positive to negative.
is detected (step 103), and a pair of consecutive sampling values x _o and x _{o+1 whose} polarity changes from negative to positive is further detected (step 104). If the judgment result is false, conversely to the above, a pair of consecutive sampling values x 0 and x 1 that change from negative to positive are detected (step 106), and then a pair of consecutive sampling values x ₀ and x ₁ that change from positive to negative are detected. Detect a pair of sampling values x _o and x _o+1 (step
107). Based on the above detection results, the sampling number n of sampling values x ₁ to x _o of the same negative or positive polarity is determined (step 105 or 108).

That is, the number of samplings n represents the number of samplings during a half cycle.

Next, the frequency of the input signal is calculated in three cases depending on whether the sampling number n thus obtained satisfies the following conditions (step 109).

Case 1 When 2n≧D ₀ Here, D ₀ is a preset value, and if _D is the frequency that can be considered as direct current at frequencies below it, it is calculated as D ₀ = _S / _D. If this condition is satisfied, the input signal is considered to be DC, and its frequency is determined to be zero (step
110).

On the other hand, if the frequency of the input signal is in a region considerably lower than the rated frequency ₀ , the number of samplings n in a half cycle by the sampling frequency _S is also large, so using this number of samplings n, the frequency of the input signal can be calculated using the proportional relationship formula. can be found.

However, if the input signal frequency is close to the rated frequency ₀ , then the sampling frequency
If the frequency of the input signal is determined only by the number n of samplings in a half cycle of _S , the error will become large.

Therefore, in order to determine which region the frequency of the input signal belongs to, the resolution of _one point n ₀ f ₀ /2n and the resolution of the next point n ₀ f ₀ /(2n+1) due to the sampling frequency S are calculated. It is determined whether the difference, that is, n _{0 0} /(2n+1)·2n is greater than a constant k. and,
Case 2 is a case where the value is not larger than the constant k, and Case 3 is a case where the value is larger than the constant k.

Case 2 When 2n<D ₀ and n ₀ f ₀ /(2n+1)・2n≦k (2) Here, k is a preset value, and if the detection error is _E , then k = _E. . If this condition is satisfied, the frequency of the input signal is calculated by the following equation (step 111).

=n ₀ /n ₀ (2) Here, ₀ is the rated frequency, and this value is usually the maximum value of the frequency detection range, and when measuring a commercial power source, the value of the commercial frequency is used.

Further, N ₀ represents the number of samples of the same polarity over a half cycle period at the sampling frequency _S at the rated frequency.

Case 3 2n<D ₀ and n ₀ f ₀ /(2n+1)・2n>k (3) In this case, the above-mentioned sampling values x ₀ , x ₁ ,
Find the frequency by calculation using ......

First, one sampling period corresponds to approximately how many electrical degrees relative to one period of the input signal, that is,
The initial phase value θ ₁ is determined by θ ₁ =n _{0 0} /n _S ×360° (4) (step 112).

Next, the sampling values before and after the polarity change mentioned above.
Correct x ₀ , x ₁ , x _o , x _o+1 using a selection data table prepared in advance within a range that maintains linearity near the polarity change. This selection data table is predetermined according to the value of the initial phase value θ ₁ and stored in the memory within the microprocessor 4. In this way, the sampling values x ₀ , x ₁ , x _o ,
The reason for correcting x _o+1 according to the initial phase value θ ₁ is that when the frequency is low, the adjacent sampling values before and after the polarity change are small near the zero point, and the difference when A-D conversion becomes large. This is because accuracy deteriorates.
To prevent this deterioration of accuracy, correction is performed by replacing the sampling value with a sampling value before x ₀ or after x ₁ within a range where linearity is maintained. Similar corrections are made for x _o and x _o+1 (step 113).

Next, using the sampling values corrected in this manner, the correct electrical angle of one sampling period for one period of the input signal, that is, the phase value θ, is determined by the following equation.

θ=180°−θ ₁ × (｜X ₁ ｜/｜X ₀ ｜+｜X ₁ ｜+｜X _o ｜/
|X _o |+|X _o+1 )/n−1...(5) θ ₁ ×(|X ₁ |/|X _{0 |} + _| ／｜X _o ｜＋｜X
_o+1 ) is the electrical angle from the first zero-crossing point of the input signal until the sampling value x ₁ is obtained, and the electrical angle from the obtaining of the sampling value X _o to the next zero-crossing point, as shown in Figure 2. represents the sum of

Note that the calculation using equation (5) is repeated until |θ−θ ₁ | is less than a certain tolerance δ (step
114, 115, 116). When |θ−θ ₁ | becomes less than or equal to the allowable deviation δ, the frequency of the input signal is calculated using the following formula using the phase value θ calculated at that time (step
117).

= θ / θ ₀・₀ (θ ₀ = (θ ₀ = ₀ / _S × 360°) =
_S θ/360° (6) The frequency determined as above is displayed on the display 6 via the digital output circuit 5.

〔Effect of the invention〕

As explained above based on the embodiments, the frequency detection method according to the present invention changes the detection method depending on the number of samplings of the input signal and the frequency range. With a sampling frequency _S on the order of about 10 times that of , it is possible to detect frequencies from direct current to about 1/10 of the sampling frequency with high precision.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a device for realizing the frequency detection method of the present invention, FIG. 2 is a diagram for explaining sampling of an input signal, and FIG. 3 is a flowchart showing an example of arithmetic processing according to the present invention. It is. 3...Analog input circuit, 4...Microprocessor, 5...Digital output circuit, 6...Display device, _x0 , _x1 ,... _xo , _xo+1 ...sampling value,
₀ ...Rated frequency, _S ...Sampling frequency,
f...Input frequency.

Claims (1)

[Claims] 1. An input signal varying from direct current to a rated frequency _{of 0} is sampled with a sampling signal having a frequency _S higher than the rated frequency of ₀ , and when detecting the frequency of the input signal, one cycle of the input signal is detected. When the number of samplings 2n (n: integer) in the period is greater than or equal to the predetermined value D ₀ , the frequency is determined to be zero, and the number of samplings 2n is equal to the value D ₀ .
When it is smaller and satisfies n _{0 0} / (2n + 1)・2n≦k (k is a predetermined detection error value), = n ₀ / n ₀ (n ₀ : integer) (where n ₀ is the half-cycle period of the rated frequency ₀ ) When the number of samplings 2n is smaller than the value D ₀ and satisfies n _{0 0} / (2n + 1)・2n>k, the initial phase value θ ₁ is calculated according to the following: θ ₁ = n _{0 0} /n _S ×360°, and then using a selection data table predetermined according to this initial phase value θ ₁ , a set of adjacent sampling values before and after the polarity changes (x ₀ , x ₁ ) and (x _o , x _o+1 ), and using this corrected sampling value, the phase value θ is calculated as θ=180°−θ ₁ (|X ₁ |/|X ₀ |+|X ₁ ｜＋｜X _o ｜／｜
Calculate the frequency according to = _S θ / 360° by repeating X _o | + | X o _{+ 1} | A frequency detection method characterized by: