Method for calculating short-time flicker and computer device

Abstract

The invention discloses a method for calculating short-time flicker and a computer device, and belongs to the technical field of power grid power. The method comprises the following steps: 1) sampling an instantaneous flicker signal to obtain a standard signal, and performing compensation calculation on the standard signal by using a correction filter to obtain a compensated signal; wherein the correction filter adopts an IIR (Infinite Impulse Response) structure as a prototype filter structure, and the coefficient of the correction filter is calculated by taking the sum of squares or the minimum sum of absolute values of deviations of all test frequency points in a flicker passband range as an objective function; and 2) calculating a cumulative probability density curve of the compensated signal under the cumulative duration of the short-time flicker, and determining the short-time flicker by using the cumulative probability density curve. Compensation calculation is carried out on the standard signal through the correction filter so as to compensate calculation result precision deviation caused by a calculation method of an IEC recommendation scheme and digitization, the accuracy of the flicker value is effectively improved, and the error between the short-time flicker value and the actual flicker value is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid electricity, and particularly relates to a method and a computer device for calculating short-term flicker. Background Art

[0002] With the development of industrial technology, a large number of high-power variable loads, such as large electric arcs and electric locomotives, have entered the power grid, causing rapid changes in the power load, thus resulting in voltage fluctuations and flicker, which seriously affect various electrical equipment operating in the power grid. To solve the problem of power grid voltage flicker, it is necessary to detect voltage flicker in a timely manner. There are many methods for detecting flicker, such as the square detection method in the time domain, the effective value detection method, and the envelope detection method in the frequency domain. The method recommended in the national standard is the square detection method recommended by IEC. This method is convenient for calculation, but it ignores some small signals in the flicker passband, resulting in an inherent error in the test results. The flicker measurement scheme recommended by IEC only gives the principle block diagram and the data output standard, without giving a specific design scheme. The traditional method for calculating short-term flicker is to store the instantaneous flicker sampling values for 10 minutes, perform probability distribution statistics on their grading, calculate the corresponding flicker values for a specific probability, substitute them into the formula to obtain the short-term flicker value, and calculate the long-term flicker value based on this. The 10-minute sampling data required to be stored during the calculation process will occupy a large amount of storage space. When the sampling rate is 200 Hz, 120,000 data need to be stored. To improve the calculation accuracy, a higher sampling rate may be required, which means a larger storage space.

[0003] In the prior art, there are inevitable inherent errors in calculating short-term flicker using the square detection method. To improve the calculation accuracy, it is necessary to increase the sampling rate, obtain more data, and occupy a large amount of storage space. For example, the invention patent application document with publication number CN103713223A discloses an adaptive range low-storage data flicker measurement method. This invention divides the actual change range of the flicker value into several ranges using the maximum value of the instantaneous flicker, and then adaptively selects a suitable range for data statistics according to the actually measured instantaneous flicker value. However, this scheme directly obtains the short-term flicker value by statistically calculating the actually measured instantaneous flicker value, and there is an inherent error between the actually measured instantaneous flicker value and the true instantaneous flicker value, resulting in an unavoidable error between the short-term flicker value calculated subsequently and the actual value. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and a computer device for calculating short-term flicker to solve the problem of low calculation accuracy of short-term flicker caused by inherent errors in the calculation of instantaneous flicker values in the prior art.

[0005] The present invention provides a method for calculating short-term flicker to solve the above technical problems. The method includes:

[0006] 1) Sample the instantaneous flicker signal to obtain a standard signal, and use a calibration filter to perform compensation calculation on the standard signal to obtain a compensated signal; wherein the calibration filter uses an IIR structure as the prototype filter structure, and the coefficients of the calibration filter are calculated with the objective function of minimizing the sum of squares or the sum of absolute values of the deviations of all test frequency points within the flicker passband range.

[0007] 2) Calculate the cumulative probability density curve of the compensated signal under the cumulative duration of a single short-term flicker, and use this cumulative probability density curve to determine the short-term flicker.

[0008] Furthermore, the coefficients of the calibration filter are obtained by performing differential evolution on the objective function and iterative calculation.

[0009] Furthermore, the calculation process of the cumulative probability density curve of the compensated signal under the cumulative duration of a single short-term flicker in step 2) is as follows: Real-time collect and cache the compensated signal. When the cached compensated signal reaches the preset cache length, use the signal of this preset cache length as a segment of signal; Compare the range of this segment of cached signal with the range of the previous segment of cached signal. If the range of this segment of cached signal exceeds the range of the previous segment of cached signal, resample the cumulative probability density curve of the previous segment of cached signal, and cumulatively add this segment of cached signal to the resampled cumulative probability density curve, and use this curve as the cumulative probability density curve of this segment. Otherwise, use the cumulative probability density curve of the previous segment of cached signal as the cumulative probability density curve of this segment of cached signal;

[0010] Repeat the above process until the cumulative duration of a single short-term flicker, and calculate the short-term flicker according to the cumulative probability density curve obtained from the last segment.

[0011] Furthermore, if the range of this segment of signal exceeds the range of the previous segment of signal when the maximum value in this segment of signal is greater than the maximum value in the previous segment of signal, shift the cumulative probability density curve corresponding to the previous segment of signal to the range of this segment of cached signal, calculate the positions of the new sampling points based on the fractional sampling rate conversion, calculate the deviations of the new sampling points, interpolate the sampling values of the new sampling points through the deviations of the new sampling points, and traverse the sampling points on the cumulative probability density curve corresponding to this segment of cached data to obtain a new cumulative probability density curve.

[0012] Furthermore, on the new cumulative probability density curve, the interpolation formula used for the resampled sampling points is:

[0013] fcdf′(k) = delta * (fcdf(m) - fcdf(m - 1)) + fcdf(m - 1);

[0014] Wherein, m is the number of sampling points, k is the position of the sampling point after resampling calculation, fcdf() is the cumulative probability density curve of the previous segment, fcdf’() is the new cumulative probability density curve, and delta is the decimal deviation.

[0015] Further, if the signal range of this segment exceeds the signal range of the previous segment when the minimum value in the signal of this segment is less than the minimum value in the signal of the previous segment, shift the cumulative probability density curve corresponding to the previous segment of the signal to the range of the buffer signal of this segment, add sampling points on the left side of the cumulative probability density curve corresponding to the previous segment of the signal to obtain the sampling point deviation, and interpolate the sampling value of the new sampling point through the new sampling point deviation. Traverse the sampling points on the cumulative probability density curve corresponding to the buffer data of this segment to obtain the new cumulative probability density curve.

[0016] Further, the number of added sampling points is NK = [(xmin - xmin’) / T], where xmin’ is the minimum value in the signal of this segment, xmin is the minimum value in the signal of the previous segment, and T is the sampling interval of the previous segment of the signal.

[0017] Further, when the signal range R’ of this segment and the signal range R of the previous segment satisfy N*R < (N - 1)*R′, and Let fcdf’(k + t) = fcdf(N), where N is the number of levels of buffer data, k is the position of the sampling point after resampling calculation, t is the number of sampling points of the probability density curve that have not been updated when the above conditions are met, t ∈ [0, N - k], fcdf() is the cumulative probability density curve of the previous segment, and fcdf’() is the new cumulative probability density curve.

[0018] A computer device, which includes a processor, and the processor is used to run a computer program to implement the method for calculating short-term flicker as described in any one of the above.

[0019] The beneficial effect of the present invention is that as an improved invention, after obtaining the instantaneous flicker value, the present invention uses a correction filter to perform compensation calculation on the instantaneous flicker value to compensate for the calculation method of the IEC recommended scheme and the accuracy deviation of the calculation result of the instantaneous flicker value caused by digitization. Without additionally increasing the quantity of collected data, the accuracy of the flicker value is effectively improved. Subsequently, the present invention uses the compensated flicker value to update the cumulative probability density curve in real time to ensure the accuracy of the cumulative probability density curve and reduce the error between the short-term flicker value and the actual flicker value. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is the flowchart of the short-term flicker calculation described in the present invention;

[0021] Figure 2 is the amplitude-frequency characteristic curve of the correction filter;

[0022] Figure 3 Schematic diagram of resampling for cumulative probability density curve. Detailed implementation manners

[0023] The following further describes the detailed implementation manners of the present invention with reference to the accompanying drawings.

[0024] The present invention first performs square detection, band-pass filtering, visual sensitivity weighted filtering, squarer calculation, and first-order low-pass filtering on the sampled instantaneous flicker information according to the IEC recommended scheme to obtain the instantaneous flicker value s(n). The optimal coefficients of the correction filter are calculated by the differential evolution algorithm, and the correction filter compensates the instantaneous flicker value s(n), compensating for some small signals ignored in the processing of the squarer and the first-order low-pass filter, avoiding the inherent errors in the prior art, and improving the accuracy of flicker calculation. Subsequently, the compensated signal is collected in real time, and the cumulative probability density curve is calculated. The maximum and minimum values of the compensated signal collected subsequently are compared with the maximum and minimum values of the signal in the previous cache respectively to determine whether the instantaneous flicker data range becomes larger. The cumulative probability density curve is updated in real time according to the change of the instantaneous flicker data range until the sampling of the cumulative duration of a short-term flicker is completed, and the cumulative probability density curve required by the short-term flicker calculation formula is obtained.

[0025] Method embodiment

[0026] As Figure 1 shown, the specific steps of the method for calculating short-term flicker proposed in this application include:

[0027] S1: Sample the instantaneous flicker. According to the IEC recommended scheme, perform square detection, band-pass filtering, visual sensitivity weighted filtering, squarer calculation, and first-order low-pass filtering on the instantaneous flicker to calculate the instantaneous flicker value s(n).

[0028] Specifically, after visual sensitivity weighted filtering, the signal d(n) is:

[0029]

[0030] where U represents the rated voltage of the power frequency signal, k f represents the gain of the visual weighting filter for the frequency f component, m f represents the modulation coefficient of the voltage component with frequency f, f represents the modulation voltage signal frequency, w f represents the angular frequency corresponding to f, n represents the sampling point number of the amplitude modulation signal, m i represents the modulation coefficient of the voltage component with frequency i, m j represents the modulation coefficient of the voltage component with frequency j, w i represents the angular frequency corresponding to frequency i, wj It represents the angular frequency corresponding to frequency j, i represents frequency i, and j represents frequency j.

[0031] Subsequently, the signal d(n) is processed by a squarer and a first-order low-pass filter to obtain the standard signal s(n):

[0032]

[0033] As can be seen from the above formula, s(n) only retains the first term in the signal d(n), and the omitted three terms result in an inherent deviation in the calculation.

[0034] S2: Iteratively obtain the optimal coefficients of the correction filter through crossover operations and perform compensation calculations on the standard signal.

[0035] For the convenience of calculation, the IEC recommended scheme discards some frequencies in the signal d(n) when obtaining the standard signal s(n). Since some small signals in these flicker passbands are ignored, there are inherent errors in the test results. Therefore, the present invention uses a correction filter to compensate for the calculation method of the IEC recommended scheme and the accuracy deviation of the calculation results caused by digitization.

[0036] The deviation of all test frequency points within the flicker passband range is the main source of error. Taking the sum of the squares of the deviations of all test frequency points within the flicker passband range as the objective function, when the value of the objective function reaches the preset error range, that is, the deviations of all test frequency points within the flicker passband range are small, the parameters obtained at this time are the optimal coefficients of the correction filter.

[0037] The differential evolution algorithm is used to solve the above objective function, and the obtained result is recorded as the optimal coefficients of the correction filter. In the present invention, the correction filter is based on an IIR filter as a prototype, and the selected order is 6. Then the control parameters to be obtained are 13. Let the control parameter xcoeff

[13] = [b0 b1 b2 b3 b4 b5 b6 a1 a2 a3 a4 a5 a6], then the transfer function is The amplitude-frequency characteristic curve of the correction filter is as Figure 2 shown. The design steps are as follows:

[0038] S21: Initialize the control parameters and take the initial value xcoeff(i).

[0039] Specifically, the expression for taking the initial value xcoeff(i) is:

[0040] xcoeff(i) = xcoeffmin + (xcoeffmax - xcoeffmin) * rand(0,1);

[0041] Let xcoeffmin = 0, xcoeffmax = 1; xcoeffmin is the minimum value of the control parameter, xcoeffmin is the maximum value of the control parameter, and i = 1, 2... 13. When the random number takes 1, xcoeff(i) = 1, and when the random number takes 0, xcoeff(i) = 0.

[0042] S22: Select the mutation strategy.

[0043] The selected mutation strategy is:

[0044] V i (g + 1) = X r1 (g) + F * (X r2 (g) - X r3 (g));

[0045] Among them, r1, r2, and r3 are all random numbers. The value range of the random number is [1, NP], NP is the population size, the scaling factor F takes 0.85, and g represents the number of iterations.

[0046] S23: Perform a crossover operation on the variant strategy and the initial value.

[0047] The crossover operation is performed according to the following rules:

[0048]

[0049] Among them, the crossover probability CR takes 0.5. When the random number takes 0 and rand(0, 1) ≤ CR, then U ij (g + 1) = V i,j (g + 1); when rand(0, 1) > CR, U ij (g + 1) = xcoeff i,j (g).

[0050] S24: Use the greedy selection strategy to select the parameters as the iterative parameters for the next round of loop.

[0051] The formula of the greedy selection strategy is:

[0052]

[0053] Among them, y is the objective fitness function.

[0054] S25: Repeat steps S22 - S24 until the objective function value reaches within the preset error range or reaches the preset number of iterations to obtain the optimal coefficients of the calibration filter.

[0055] As another implementation, the objective function can also be the sum of the absolute values of the deviations of all test frequency points within the flash passband range.

[0056] For example, when the number of iterations is selected as 100, the hierarchical connection coefficients of the 6th-order IIR filter are obtained as follows:

[0057] [-1.4142 2.3006 -0.5063 1.0000 0.9730 0.6288 1.0000 1.9188 0.8478 1.0000 1.2825 0.7404

[0059] 1.0000 -0.1307 1.0356 1.0000 -1.2489 0.9376].

[0060] S3: Real-time collect the compensated signal and calculate the cumulative probability density curve.

[0061] Use the optimal coefficients of the calibration filter to compensate the standard signal, real-time collect the compensated signal x(n), and cache the collected compensated signal x(n). When the cache length reaches the preset cache length L, perform probability statistics on the L collected compensated signals x(n). Based on the maximum value xmax and the minimum value xmin in the L collected compensated signals x(n), divide the cached data into N levels, and accumulate them in levels to obtain the cumulative probability density curve fcdf. The segment length L is determined according to the storage and processing speed.

[0062] S4: Dynamically adjust the cumulative probability density curve according to the relationship between the maximum and minimum values of the current cached data and the maximum and minimum values of the previous cached data.

[0063] Continue to collect the compensated signal x(n). When the cache length reaches L again, dynamically adjust the cumulative probability density curve fcdf according to the currently collected data segment.

[0064] Specifically, find the maximum value xmax' and the minimum value xmin' in the currently collected data segment, compare xmax' with xmax, and compare xmin' with xmin. If xmax' > xmax or xmin' < xmin, the instantaneous flicker data range R becomes larger. Keep the number of sampling times unchanged, and the hierarchical sampling interval becomes larger, which is equivalent to a change in the sampling efficiency. Resample the cumulative probability density curve fcdf to generate a new cumulative probability density curve fcdf'.

[0065] Since the range R is dynamically changing and unpredictable, the sampling rate after resampling is not fixed, and the new sampling point deviation delta needs to be calculated in real time. Since the cumulative probability density curve fcdf is a monotonically increasing curve, linear interpolation can be used according to the new sampling point deviation delta by fractional sampling rate conversion.

[0066] Specifically, the specific steps for fractional sampling rate conversion are as follows:

[0067] 1) If xmax’ > xmax, the cumulative probability density curve fcdf needs to be shifted to the right. The original range R becomes R1. Traverse the sampling points m of the cumulative probability density curve fcdf, calculate the position k of the resampled sampling points. The original sampling interval T = R / N, then the resampling interval T1 = R1 / N, k = [m * T / T1], where [] is the floor function. The deviation of the new sampling point delta = k * T1 / T - (m - 1). The sampling value at point k is obtained by interpolation according to the deviation delta of the new sampling point:

[0068] fcdf′(k) = delta * (fcdf(m) - fcdf(m - 1)) + fcdf(m - 1);

[0069] Here, the cumulative probability density curve fcdf is a monotonically increasing curve, and linear interpolation is used.

[0070] 2) If xmin’ < xmin, the cumulative probability density curve fcdf needs to be shifted to the left. The original range R becomes R2. Add NK = [(xmin - xmin’) / T] sampling points to the left of the cumulative probability density curve fcdf. It is easy to know that these sampling values are 0. Start traversing from the newly added sampling points using the IEC recommended scheme. The deviation of the new sampling point is:

[0071] delta = m * T1 / T - k + (Nk + 1 - (xmin - xmin’) / T).

[0072] Since the change of the range R’ is unpredictable, it may occur that N * R < (N - 1) * R′. At this time, when m traversal ends, k < N. At this time, processing is required. Let fcdf’(k + t) = fcdf(N), where t is the number of sampling points of the probability density curve that have not been updated, and t ∈ [0, N - k].

[0073] The schematic diagram of the curves before and after the sampling rate conversion is as Figure 3 , the red line is the sampling points of the cumulative probability density curve fcdf, and the blue "×" represents the new sampling points. After generating the new cumulative probability density curve fcdf’, the data of this segment is hierarchically accumulated to the fcdf’ curve, and fcdf’ is iterated to become the final cumulative probability density curve.

[0074] In the patent application documents mentioned in the background art, a method for dynamically adjusting the cumulative probability density curve is also disclosed. However, in this method, the instantaneous flicker maximum value is a preset value that does not change, and the minimum value is fixed at zero. When the instantaneous flicker value changes, only the grading range close to the instantaneous flicker value is selected, resulting in poor tracking ability for changes. When the value of the instantaneous flicker maximum is too large or too small, it cannot correctly reflect the probability distribution of the instantaneous flicker value, and large errors are likely to occur. At the same time, when the grading range becomes larger in this application document, the new grading value is updated by the way of two-stage merging and accumulation. In contrast, the present invention calculates new sampling points of the probability density sampling curve according to the range change, and linearly interpolates according to the non-decreasing characteristic of the probability density curve to obtain a new probability density curve, with higher calculation resolution.

[0075] S5: Repeat steps S3 - S4 until the sampling of the short-term flicker cumulative duration is completed, and the cumulative probability density curve required by the short-term flicker calculation formula is obtained.

[0076] When the short-term flicker cumulative duration is completed, the cumulative probability density curve pcdf = 1 - fcdf required by the short-term flicker calculation formula is obtained by slightly processing the cumulative probability density curve fcdf at this time, and P 0.1 、P 1 、P 3 、P 10 、P 50 are interpolated and substituted into the following formula to obtain the short-term flicker Pst.

[0077]

[0078] Among them, P 0.1 、P 1 、P 3 、P 10 、P 50 are the flicker visibility values corresponding to the probabilities exceeding 0.1%, 1%, 3%, 10%, and 50% respectively during the detection time. K 0.1 = 0.0314, K 1 = 0.0525, K 3 = 0.0657, K 10 = 0.28, K 50 = 0.08.

[0079] The short-term flicker is calculated according to the test standard in Table 1. The comparison of the results obtained by the traditional method and the method of the present invention is shown in Table 1. Table 1 lists the calculation results with and without correction filtering. Here, N = 64 and L = 256. Comparing the short-term flicker Pst of the traditional statistical method and the case without correction filtering, it can be seen that the calculation results of the two methods are basically the same, and the scheme of the present invention does not cause loss of accuracy.

[0080] Table 1 Short-term flicker test standard

[0081]

[0082] As can be seen from the table, although the error of the calculation result is also within the allowable range of the national standard without adding the correction filter, the inherent error caused by the calculation method makes the calculation result deviate from the theoretical value to a certain extent. At this time, the sum of the deviations of all test items is 0.0945. The test results after adding the correction filter as described in step S2 are shown in Table 1, and the sum of the deviations of all test items within the frequency band range drops to 0.0389, without the need for segmentation or frequency division point correction, and it is applicable to the entire passband range.

[0083] When calculating the short-term flicker by the traditional statistical method, it is necessary to cache the sampling data for 10 minutes. The scheme described in the patent updates the instantaneous flicker cumulative probability density distribution in real time, and only needs to cache the fcdf curve data (i.e., 64 points) obtained by each hierarchical update. At the cost of adding a small amount of calculation, it greatly saves the storage space and improves the practicability of the algorithm.

[0084] Embodiment of the computer device

[0085] The computer device includes a processor, and the processor is used to run a computer program to implement the method for calculating the short-term flicker as described above. The specific implementation process has been described in detail in the method embodiment and will not be elaborated here.

Claims

1. A method for calculating short-time flicker, characterized in that: The method includes: 1) The instantaneous flicker signal is sampled and processed to obtain a standard signal, and the standard signal is compensated by a correction filter to obtain a compensated signal; wherein the correction filter adopts an IIR structure as a prototype filter structure, and the coefficient of the correction filter is calculated by taking the minimum sum of squares or absolute values ​​of all test frequency deviations within the flicker passband as the objective function; 2) Calculate the cumulative probability density curve of the compensated signal under the cumulative duration of a short-term flicker, and use the cumulative probability density curve to determine the short-term flicker.

2. The method for calculating short-time flicker according to claim 1, characterized in that: The coefficients of the correction filter are obtained by differential evolution of the objective function and iterative calculation.

3. The method for calculating short-time flicker according to claim 1, characterized in that: The calculation process of the cumulative probability density curve of the compensated signal under the cumulative duration of a short-time flicker in the step 2) is as follows: the compensated signal is collected and cached in real time, and when the cached compensated signal reaches a preset cache length, the signal of the preset cache length is used as a segment of the signal; the range of the current segment of the cached signal is compared with the range of the previous segment of the cached signal, if the range of the current segment of the cached signal exceeds the range of the previous segment of the cached signal, the cumulative probability density curve of the previous segment of the cached signal is resampled, and the current segment of the cached signal is hierarchically accumulated to the resampled cumulative probability density curve, and the curve is used as the current segment of the cumulative probability density curve, otherwise, the cumulative probability density curve of the previous segment of the cached signal is used as the cumulative probability density curve of the current segment of the cached signal; The above process is repeated until the cumulative duration of a short-time flicker is reached, and the short-time flicker is calculated based on the cumulative probability density curve obtained in the last section.

4. The method for calculating short-time flicker according to claim 3, characterized in that: If the signal range of this segment exceeds the signal range of the previous segment, that is, when the maximum value in this segment is greater than the maximum value in the previous segment, the cumulative probability density curve corresponding to the previous segment is shifted right to the range of the cached signal of this segment. Based on the decimal multiple sampling rate transformation, the position of the new sampling point after resampling is calculated, and the deviation of the new sampling point is calculated. The sampling value of the new sampling point is obtained by interpolating the deviation of the new sampling point, and the sampling points on the cumulative probability density curve corresponding to the cached data of this segment are traversed to obtain a new cumulative probability density curve.

5. The method for calculating short-time flicker according to claim 4, characterized in that: On the new cumulative probability density curve, the interpolation formula used for the resampled sampling points is: ； Where m is the number of sampling points, k is the location of the sampling points after resampling, fcdf() is the previous cumulative probability density curve, fcdf'() is the new cumulative probability density curve, and delta is the decimal deviation.

6. The method for calculating short-time flicker according to claim 3, characterized in that: If the signal range of this segment exceeds the signal range of the previous segment, that is, when the minimum value in this segment is smaller than the minimum value in the previous segment, the cumulative probability density curve corresponding to the previous segment is moved left to the range of this segment of cached signal, and a sampling point is added to the left side of the cumulative probability density curve corresponding to the previous segment to obtain the sampling point deviation, and the sampling value of the new sampling point is obtained by interpolating the new sampling point deviation, and the sampling points on the cumulative probability density curve corresponding to this segment of cached data are traversed to obtain a new cumulative probability density curve.

7. The method for calculating short-time flicker according to claim 6, characterized in that: The number of added sampling points is NK=[(xmin-xmin') / T], where xmin' is the minimum value in the current signal segment, xmin is the minimum value in the previous signal segment, and T is the sampling interval of the previous signal segment.

8. The method for calculating short-time flicker according to claim 3, characterized in that: When the current signal range R' and the previous signal range R meet ,and When , let fcdf'(k+t)=fcdf(N), N is the number of cache data levels, k is the sampling point position after resampling, t is the number of sampling points of the probability density curve that are not updated when the above conditions are met, t∈[0,Nk], fcdf() is the previous cumulative probability density curve, and fcdf'() is the new cumulative probability density curve.

9. A computer device, characterized in that: The device comprises a processor, wherein the processor is used to run a computer program to implement the method for calculating short-time flicker as described in any one of claims 1 to 8.

Images