JP2000121424A - Decision device for metering condition of dynamic metering apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a decision device by which a metering condition including the starting point of time of the acquisition of measured data to be computed can be decided so that a most stable metered value can be computed by installing a dispersion- degree calculation means or the like which calculates the dispersion degree of every metered value. SOLUTION: When a product 6 which is conveyed by an in-feed conveyor 1 is placed on a metering conveyor 2, a load cell 14 outputs a metering signal, and the product 6 is sent out to a let-off conveyor 3. Then, a storage means stores every metered value in which a plurality of metered values by metering the product 6 in a plurality of numbers times are classified into metered sets whose readout starting point of time is common. In a dispersion-degree calculation means, the dispersion degree of the respective metered values stored in the storage means is calculated for the respective sets. Then, a readout starting-point-of-time decision means selects a comparatively small dispersion degree out of a plurality of dispersion degrees calculated by the dispersion-degree calculation means. A readout starting point of time corresponding to the sets of the dispersion degrees is decided as a readout starting point of time in a formal operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a scale having a limited time range for obtaining weight data for obtaining a weighing value with high accuracy, for example, an automatic weight sorter. When weighing, or when weighing the weight of articles put into a weighing hopper with a fixed-quantity filling weigher or combination weigher, a dynamic section that determines the reading section of measurement data for calculation so as to obtain the most accurate weighing value The present invention relates to a weighing condition determining device for a weighing device.

[0002]

2. Description of the Related Art An example of a conventional weighing condition determining device of a dynamic weighing device (hereinafter, may be simply referred to as a "weighing condition determining device") is disclosed in JP-A-6-43011. As shown in FIG. 11, this dynamic weighing device is configured to start the operation immediately before the articles 6 on the weighing conveyor 2 start contacting the unloading conveyor 3 from the point P when the articles 6 are completely fed from the infeed conveyor 1 onto the weighing conveyor 2. up to T _I "time is the interval that is capable of measuring the weight of the article 6.
According to this dynamic weighing device, in the test stage before the actual operation, the tip of the article 6 is about １ ／ of the entire length of the weighing conveyor 2.
Between T _I ′ and T _I ″ at the point where
When passing through (the time width of _W ′), the weighing signal output by the weight detector 4 in time series (every time τ elapses) is acquired as measurement data (w _{1,1 to} w _{n, 1} ). Figure 12
The measurement is repeated m times as shown in FIG.
_{1 ', ··, t k'} , ··, standard deviation sigma ₁ measurement data for each _{t n ', ··, σ k} , ··, and calculates the sigma _n, the standard deviation sigma (variation degree) is most A small time is selected and the selected time is determined as the weighing time.

This weighing condition determining apparatus can be mainly applied to a high-order low-pass filter, for example, an A / D conversion output value of an analog filter output signal having a large response delay or an output value of a recursive filter. And
Each time point t ₁ ′ to t _n ′ at which the output value is measured is recognized by a timer started by the photosensor 5, and the standard deviation σ of the output value at each measurement time point t ₁ ′ to t _n ′ is σ Is determined as the timing for determining the weight value of the article 6.

Another example of a conventional weighing condition determining device for a dynamic weighing device is disclosed in Japanese Patent Application Laid-Open No. 8-62032. This dynamic metering device, as shown in FIG. 13, because of the relatively measurement data metric determines the time-series measurement data of the section T _F which is stable in the weight capable of measuring a period T _W of the article 6 The measurement measurement data reading (acquisition) section is performed by calculating the time series measurement data of the section _TF using a single or multiple moving average filter, thereby obtaining one or more periodic oscillations included in the time series measurement data. It seeks to eliminate noise and thereby accurately weigh the article. In other words, this single or multiple moving average filter represents the noise periods included in the measurement data, each period of N representative noises having an amplitude that causes a measurement error, as T ₁ , T ₂ , ..., T _N
Assuming that the time width that can be measured is T _W , it is set so as to satisfy T _W ≧ T ₁ + T ₂ +... + T _N − (N−1) × τ (1) Therefore, even if there is still a noise of the period T _{(N + 1)} to be further removed, if a filter to which the period T _{(N + 1)} of the noise is added is used, T _W <T ₁ + T ₂ +... + _TN + T _{(N + 1)} -N × τ When (2) holds, the period T _{(N + 1)} is not added. The reason is that the measurement data in the range exceeding the measurable section T _W is data in a state where the article 6 is not completely on the weighing conveyor 2 or in a state immediately after the article 6 is on the weighing conveyor 2. This is because it contains an extremely large error. For this reason, the multiple moving average filter is set so that the expression (1) is satisfied.

Accordingly, (1) whether the number of representative periodic vibration noise that causes the metering error is small, or, if sufficiently long time measurable interval T _W on the sum of these noise cycle However, when another long-period noise is added, the interval T _W
When exceeds _{the, T W >> T 1 + T} 2 + ··· + T N - (N-1) × τ moving average filter of N-fold as the (3) is determined. (2) If the amount of noise that needs to be removed is relatively large with respect to the length of the section T _W , T _W = T ₁ + T ₂ +... + T _N − (N−1) × τ. (4) Or, T _W > T ₁ + T ₂ +... + _TN − (N−1) × τ (5) That is, the sum of the noise periods to be removed by the multiple moving average filter is the measurable section T
_It is determined that if the time is substantially equal to the time of _W , or if another noise period is added, the time period exceeds the section T _W. Basically, it is considered that the larger the number of moving average filters, the larger the number of fixed-period vibration noises that can be removed, which is advantageous for reducing the measurement error.

[0006]

However, Japanese Patent Application Laid-Open No. 6-4 / 1994
For the purpose of calculating the most stable weighing value, the former weighing condition determination device disclosed in Japanese Patent No. 3011 discloses a timing t at which the measured data is most stable in the section T _W ′ shown in FIG. _K 'is determined to be the final timing of measurement data reading, and this is intended to improve weighing accuracy.However, measurement data is read so that the most stable weighing value can be calculated. It does not determine when to start. This reading start time T ₁ ′ is about 1/1 of T ₁ ″.
It is just set as 4.

[0007] Then, the latter weighing condition determining apparatus disclosed in Japanese Patent Laid-Open No. 8-62032, for the purpose of improving the metering accuracy, are included in the measurement data in the T _W section 13 Although one or two or more periodic vibration noises that cause a weighing error are removed, the reading start time of the measurement data to be subjected to the moving average calculation is determined so that the most stable weighing value can be calculated. It does not do. That is, (1) even when determining the noise cycle so as to satisfy the equation, may join the measurement data around the start point P of the measurable period T _W filter operation,
In this case, the measurement data including the impact amplitude immediately after the article 6 gets on the weighing conveyor 2 participates in the calculation, and as a result, high weighing accuracy may not be obtained.

The present invention provides a weighing condition determining device of a dynamic weighing device that determines weighing conditions including a start time of acquisition of measured data to be calculated so that a most stable weighing value can be calculated. The purpose is to:

[0009]

According to a first aspect of the present invention, there is provided a weighing signal output from a weight detector from a time point at which a predetermined weighing signal is read to a time point at which a predetermined weighing signal is read out. In a dynamic weighing device that calculates a weighing signal of a reading section to determine a weighing value of an article,
The weight detector reads a weighing signal for each of a plurality of reading sections at different reading start times when weighing the article by the weight detector, calculates the read weighing signal for each section, and for each of these reading sections. A reading means for reading the calculated weighed value, and a plurality of weighed values read by the reading means when the article is weighed a plurality of times, each of which has a common reading start time and is divided into these weighed values. , A variability calculating means for calculating the variability of each weighing value stored in the storing means for each set, and a comparative variability calculated by the variability calculating means. At the start of reading to select a small degree of variation and determine the reading start time corresponding to the set of the degree of variation as the reading start time in actual operation It is characterized in that it comprises a determining means.

According to a second aspect of the present invention, in the weighing condition determining apparatus for a dynamic weighing apparatus according to the first aspect, the reading means sets a plurality of different reading start points when the articles are weighed by the weight detector. A weighing signal is read for each of a plurality of reading sections, and a single or multiple moving average calculation is performed on the weighing signal with a predetermined sampling number as a filter constant for each reading section, and the calculation is performed for each of these reading sections. It is a configuration for reading a weighed value, characterized in that it comprises a changing means for changing the sampling number or the multiple of the multiple moving average calculation according to each reading section.

According to a third aspect, in the weighing condition determining apparatus for a dynamic weighing apparatus according to the first or second aspect, the storage means reads the article when the article is weighed a plurality of times during the actual operation. A plurality of weighed values are grouped by those having a common reading start time, and the weighed values thus grouped are stored.

According to the weighing condition determining device of the dynamic weighing device according to the first invention (hereinafter, referred to as a weighing condition determining device), it is possible to determine a reading start time of the weighing signal. The dynamic weighing device reads the weighing signal from the determined reading start time to the reading end time, and can calculate the read weighing signal to determine the weighing value of the article. According to this weighing condition determination device, when the articles are weighed at the stage before the actual operation, or at the stage of the actual operation,
A weighing signal read for each of a plurality of reading sections at different reading start times is calculated for each section, and the weighing value calculated for each of these reading sections is read by the reading means. The storage unit divides a plurality of weighed values read by the reading unit when the article is weighed a plurality of times by those having a common reading start time, and stores the weighed values thus grouped. Next, the variation degree calculating means calculates the variation degree of the metric value stored in the storage means for each set. Then, the reading start time determination means selects a relatively small variation degree from among the plurality of variation degrees calculated by the variation degree calculation means,
The reading start time corresponding to the set of the degree of variation can be determined as the reading start time in the actual operation.

According to the second aspect of the present invention, the weighing value is obtained by performing a single or multiple moving average operation on the weighing signal using a predetermined number of samplings as a filter constant. Vibration noise can be removed from the weighing signal. Since the weighing signal is read for each of a plurality of reading sections at different reading start times and the number of samplings or the number of multiple moving average calculations is changed according to each reading section, the degree of variation in the weighing value is changed. By selecting a reading section where is relatively small, it is possible to select the number of samplings or the multiple of the multiple moving average calculation in which the degree of dispersion of the weighing values is relatively small.

According to the third aspect of the present invention, a new reading start time is determined by using a plurality of weighing values read by the reading means when articles are sequentially weighed during the actual operation, and the reading start time is changed to a new one. The weighing value of the article can be determined by changing the weighing signal of the reading section corresponding to the reading start time after the change.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION One embodiment of a weighing condition determining device (hereinafter, also simply referred to as a "weighing condition determining device") of a dynamic weighing device according to the present invention implements a weighing condition determining device in a weight sorter. It was done. As shown in FIG. 11, the weight sorter includes an infeed conveyor 1, a weighing conveyor 2, and an outfeed conveyor 3, and the weighing conveyor 2 is supported by a weight detector 4 such as a load cell. Also,
An article detector such as a photo sensor 5 for detecting the presence or absence of the article 6 is provided between the feeding conveyor 1 and the weighing conveyor 2. The article 6 is the feed conveyor 1,
It is conveyed from the left side to the right side in FIG.

FIG. 10 is a block diagram showing an electric circuit of the weight sorter and the weighing condition determining device. 4 is a weight detector, 7 is an amplifier, 8 is a filter, 9 is an A / D converter,
Reference numeral 10 denotes an input / output circuit of the microcomputer;
Denotes a microprocessor (hereinafter, referred to as a CPU), 12
Denotes a storage unit having a memory element such as a ROM, a RAM, and an E ² ROM, 5 denotes a photosensor, and 13 denotes an input unit.

When the weight of an article is to be sorted by the weight sorter, the article is supplied to a feeding conveyor 1 shown in FIG. The article 6 conveyed by the infeed conveyor 1 shields the photosensor 5 at a position a immediately before the weighing conveyor 2, and at this time, an output pulse of the photosensor 5 shown in FIG. When the article 6 is put on the weighing conveyor 2, the weight detector 4 outputs a weighing signal. The weighing signal is amplified by the amplifier 7, the high frequency component is removed by the filter 8, and the weight shown in FIG. It is processed into a weighing signal represented by a waveform (output signal of the weight detector 4). Then, this weighing signal is converted into a digital weighing signal by the A / D converter 9. The A / D converter 9 has an extremely short time interval τmsec, for example, 0.1 msec to 10 msec.
This is a type in which A / D conversion is performed once for c. Then, as shown in FIG. 7, when the predetermined time _TaI has elapsed from the time point a at which the output pulse of the photosensor 5 is input to the CPU 11, the reading time point I has elapsed from the time point I (I is the reading start time point of the weighing signal). The weighing signals are read to calculate the weighing value of the article 6 by calculating the read weighing signals, the weighing value is compared with a preset allowable weight value, and the comparison result is sent out as a sorting signal to the conveyor 3.
To send to. As a result, the articles outside the allowable weight are sent to a different transport path from the articles within the allowable weight.

The weight sorter has an operation mode for selecting the articles 6 to be sequentially conveyed as described above, and a test mode for determining and setting weighing conditions such as a reading start time of a weighing signal. And setting modes (1) and (2) for setting various weighing conditions. In simple terms, this weighing condition determining apparatus is used for a plurality of reading sections in which the reading start time of the weighing signal is different from each other in a test mode (a stage before the actual operation) and an operation mode (a stage during the actual operation). The weighing value of the article is obtained for each reading section by performing a single or multiple moving average operation differently in the number of samplings or the number of weights different from each other, and the reading section in which the variance of the weighing value is relatively small is determined. By selecting this, select the number of samplings or the number of moving average calculations that make the degree of dispersion of the weighing value relatively small, and use the selected reading section, sampling number, and other weighing conditions during actual operation to calculate the moving average. , High-precision weighing can be performed.
In this way, the weighing conditions including the weighing signal reading start point I, the weighing signal reading section I to R, the sampling number of each moving average calculation, and the moving average calculation weight, which provide high weighing accuracy, are automatically set. Can be determined.

This weighing condition determination device is constituted by a CPU 11, a storage unit 12, and a program previously written in the storage unit 12. That is, the CPU 11
The weighing signal supplied from the weight detector 4 is processed as measurement data in accordance with a program written in the storage unit 12, and based on the processing result, a weighing condition including a reading start time of the weighing signal at the time of actual operation and the like. Has the function of determining In other words, the apparatus includes a reading unit, a storage unit, a variation calculation unit, and a reading start time determination unit for achieving this function.

The reading means includes a test stage before the actual operation (test mode) and a stage during the actual operation (operation mode).
When the article 6 is sent to the weighing conveyor 2 and the article 6 is weighed, the weighing signal is read for each of a plurality of reading sections having a plurality of different reading start times when the weight detector 4 weighs the article 6, A predetermined single or multiple moving average calculation is performed on the weighing signal with the sampling number determined for each reading section and the preset sampling number as a filter constant, and the calculation is performed for each of these reading sections. And a changing means for changing the number of samplings or the weight of the multiple moving average calculation in accordance with each reading section.

The storage means is a RAM constituting the storage unit 12
In the test mode and the operation mode, when articles 6 having substantially the same weight or different weights are weighed L times a preset number of times, reading of a plurality of weight value data read by the reading means is started. This is a means for grouping data items having a common time point and storing these grouped weighing value data.

The variance calculating means calculates a standard deviation σ (variation) of each weighing data stored in the storage unit 12.
Is calculated for each set. The reading start time determination means selects a relatively small standard deviation σ from the plurality of standard deviations σ calculated by the variation degree calculating means, and reads the reading start time corresponding to the set of the standard deviation σ during the actual operation. This is a means for determining the start time. The reading start time I determined by the reading start time determining means is
Is stored (set) in the storage unit 12 and is used as the reading start time I of the weighing signal during the actual operation. Then, the weighing conditions including the reading section I to R of the weighing signal corresponding to the reading start time I, the number of samples of each moving average calculation, and the weight of the moving average calculation can be automatically determined and set. .

Next, FIG.
The procedure for determining the reading start time I of the weighing signal shown in FIG. 9 will be described with reference to the flowcharts shown in FIGS. A program corresponding to this flowchart is stored in the PROM of the storage unit 12. First, as shown in FIG. 1A, a worker operates a keyboard (input unit 1).
By operating 3), the program is set to the setting mode (1) (S100). Next, the length L ₁ of the weighing conveyor 2,
Article transport speed V, setting the length L ₂ of the objects to be weighed article 6 (S102). Then, the CPU 11 sets L ₁ , V and L ₂
Measurable section T _W (= (L ₁ -L
₂ ) / V) and the evaluation interval T _X (= j ×
T _W ) is calculated (S104, S106). Where j
Is a value of 1 or less, for example, 0.3. The longer the measurement accuracy evaluation section T _X is, the wider the range of selection of the measurement signal reading section can be. As a result, it is possible to relatively accurately select a reading section in which the measurement accuracy is high. If it is too long, the calculation time will be long, so j (= 0.3) is set so as to be an appropriate time. Next, a time T _aP (= L ₂₎ from a when the photo sensor 5 detects the article 6 to a start point P of the measurable section T _W.
/ V) is set in the timer in the CPU 11 (S10).
8).

Next, as shown in FIG. 1B, an N-fold moving average is obtained in the same manner as the latter one of the conventional weighing condition determining device (the device disclosed in Japanese Patent Application Laid-Open No. 8-62032). the numbers of taps of the filter D _1, D _2, ···, sets the D _N. That is, first, the operator operates the input unit 13 and
The program is set to the setting mode (2) (S200),
Number of taps D ₁ , D ₂ ,... Of the N-fold moving average filter
, _DN are set (S202). N is 1, 2, ...
Is a natural number. Each tap number D ₁ of the the N-heavy moving average filter, D _2, · · ·, functions and methods to set the D _N is a detailed description thereof will be omitted because it is equivalent to the conventional latter metering condition determining apparatus.

The N-fold moving average filter is a CPU 1
1 are included in the digital weighing signals w ₁ , w ₂ ,... Shown in FIG.
This is for attenuating various types of periodic vibration noise. Approximately 1 of each of the N types of periodic vibration noise
N types of sampling numbers corresponding to the period are used as filter constants of the N-fold moving average filter, that is, the number of taps. The weighing signal W _{K + 1, N} filtered by the N-fold moving average filter becomes an average value of sampling data corresponding to approximately one cycle of each of N types of fixed-period vibration noises. This means that N types of periodic vibration noise components have been removed. However, the number of the taps of the N-fold running average filter D _1, D _2, · · ·, when setting the D _N is set to satisfy the expression (1). As shown in FIG. 6, the first, second,.
The sampling number corresponding to approximately one cycle of the N-th periodic oscillation noise is D ₁ , D ₂ ,..., D _(N−1) (= 4
) And D _N (= 5). The N-th order moving average filter included in the N-fold moving average filter converts the measurement data w _{K + 1} , w _{K + 2} ,... ) Is calculated as moving averages w _{K + 1,1} , w
_{K + 2,1} ,... _Are calculated. The (N-1) th-order moving average filter sets the moving average value w _{K + 1,1} , w _{K + 2,1} ,... Calculate the moving average value w _{K + 1,2} , w
Calculate _{K + 2,2} , ... In this way, the first
By performing the moving average calculation N times up to the next moving average filter, the weight value W _{K + 1, N} of the article 6 can be calculated.
As a result, the temporary reading start point Q of the weighing signal shown in FIG. 6 is determined, so that the time points P, Q, the time interval T _W ,
T _M and T _F are determined, and each is stored in the storage unit 12.

Next, although not shown in the flow chart, in the test stage before the actual operation, the same as the former one of the conventional weighing condition determination device (the device disclosed in Japanese Patent Laid-Open No. 6-43011) is used. Then, the CPU 11 is caused to perform an operation for obtaining the reading end time t _K ′ (see FIG. 12) of the weighing signal. That is, as shown in FIG. 6, R (= T) immediately before the article 6 on the weighing conveyor 2 starts to contact the sending-out conveyor 3 from the time point P when the article 6 is completely sent from the sending-in conveyor 1 onto the weighing conveyor 2. _I ") is a section where the weight of the article 6 can be measured. Therefore, in the test stage before the actual operation, the tip of the article 6 is positioned at about 1/3 of the total length of the weighing conveyor 2. When passing through a period from Q (= T _I ′) to R (time width of T _F (= T _W ′)), the weight detector 4 chronologically changes (τ
The reading of the output weighing signal as measurement data (w _{K + 1} , w _{K + 2} ,..., W _M ) is repeated m times as shown in FIG. Measurement time t _{K + 1} ,..., T for reading the measurement data
Standard deviation σ ₁ of measurement data for each _M (not shown)
, Σ _M is calculated, the time with the smallest standard deviation σ (variation degree) is selected, and the selected time is determined as the weighing time. Here, for the sake of simplicity, the selected time is assumed to be R, and this R is determined as the end point of reading the weighing signal.

Next, a method of determining the reading start time I of the weighing signal, which is a feature of the present invention, in this test stage will be described. In other words, the current state is a state in which an N-fold moving average filter for attenuating N kinds of typical periodic vibration noises is set. According to the present invention, it is possible to determine which one of the average filter and the multiple moving average filter having a different number of taps can be used to measure the weight of the article 6 with the highest accuracy. This is the feature of

First, as shown in FIG. 2, the operator operates the input unit 13 to set the program to the test mode (S300). Next, the reference article is carried into the weighing conveyor 2 by the feed-in conveyor 1. Then, loading the start of the article is detected by the photo sensor 5, between time R that has elapsed T _W time than P time has elapsed T _aP time from when a which detects the article (see FIG. 6), CPU 11 is A /
Digital weighing signals w _{11 to} w _{12 to} w output by the D converter 9
_1M is read and stored in the storage unit 12 as measurement data in the reading order. Note that the A / D converter 9 converts the analog weighing signal into a digital weighing signal every τ time, so that every time an article passes through the weighing conveyor 2 once, (T _W / τ) + 1 = M Are stored in the storage unit 12. Then, when the article is detected by the photo sensor 5 and reaches the point of time R, the count value TC1 of the number of times the reference article has passed through the weighing conveyor 2 is incremented by _1, and the number of times of weighing of the reference article is counted. In this way, the measurement of the reference article is repeated L times, and the measurement data (w ₁₁ , w ₁₂ ,.
_{w 1M), (w 21,} w 22 ~w 2M) ~ (w L1, w L2 ~w LM)
Is stored in the memory area of the storage unit 12 (S302).
This measurement data is read by the reading means.

Next, the evaluation section T for measuring accuracy is_XIs section T_M
The processing method differs between longer and shorter cases
Therefore, although not shown in the figure, first, T_X≤T_MWhether or not
Is determined. Note that the evaluation section T_XIs, for example, T_X≤
T_MIn the case of, as shown in FIG. _XEach total in the section of
Measurement data w₁~ W_KIs obtained at each measurement time t₁~ T
_KIs the start time of reading the measurement data
(N + 1) -th (N +) included in the multiple moving average filter
1) Set the number of K taps of the next moving average filter
It is a section for. Then, the (N + 1) th order moving average
The filter tap number and sampling data are read and opened.
Start time is t₁Is (K + 1) and w₁~ W
_{K + 1}And the reading start time is t_TwoIf K
And w_Two~ W_{K + 1}It is. Start reading in the same way
Time is t_KIs two and w_K, W_{K + 1}In
You.

Then, T_X> T_MIn the case of
Like T_MT within the section of _X≤T_MSame as
(N + 1) th moving average fill with K taps
This is a section for setting data, and
The measurement data w_{K + 1}~ W_{K + 3}Is obtained at each measurement
Point t_{K + 1}~ T_{K + 3}Open the reading of the measurement data
Nth moving average filter included in the N-fold moving average filter
To set three types of tap numbers for the next moving average filter
It is a section of. However, evaluation section T_XWhen the end point S is measured
Point t_{K + 3}If the time is later than
Data w_{K + 1}~ W_{K + 5}Is obtained at each measurement time t_{K + 1}~ T
_{K + 5}Is the start time of reading the measurement data
Nth moving average filter included in the N-fold moving average filter
It is a section for setting the number of taps of four types of filters,
Each measurement data w_{K + 5}~ W_{K + 8}Is obtained at each measurement time t
_{K + 5}~ T_{K + 8}At the start of measurement data reading
The point included in the (N-1) multiple moving average filter
(N-1) Set three types of tap numbers for the next moving average filter.
It is a section to determine. The same applies hereinafter. The (N +
1) Tap number and sampling data of next moving average filter
Ta is T_X≤T_MIs the same as And N double transfer
N-th order moving average filter
The number of taps and sampling data
t_{K + 1}Is 5 and w_{K + 1}~ W_{K + 5}And
In the same manner, the reading start time is t_{K + 4}If it is
Is two and w_{K + 4}, W_{K + 5}It is. Similarly, (N
-1) (N-1) th order in the case of using a heavy moving average filter
The number of taps and sampling data of the moving average filter are
Reading start time is t_{K + 5}Is 4 and w
_{K + 5}~ W_{K + 8}, And the reading start time is t
_{K + 7}Is two and w_{K + 7}, W_{K + 8}It is. So
And evaluation section T_XIs the measurement time t_{K + 8}After
In the case where the current time is the same as above, (N-
2) Set multiple moving average filters with weights less than
Perform a moving average calculation on the measured data in the same way as
Calculate the value. In this way, start reading measurement data
The time can be changed, or the (N + 1) th, Nth, (N−
1) Next, change the number of taps of the moving average filter,
And changing the order of the multiple moving average filter
It is a step.

[0031] Therefore, if it is determined that the T _X ≦ T _M as shown in FIG. 6, as shown in FIG. 7, further the (N + 1) th to N heavy moving average filter by adding the following moving average filter ( N + 1) A multiple moving average filter is set. When it is determined that T _X > T _M as shown in FIG. 8, an (N + 1) -fold moving average filter can be set, and as shown in FIG. It is possible to set an N-fold moving average filter in which the number of samplings is smaller than the original number. And N
It is also possible to set the (Nu) heavy moving average filter by subtracting the number of heavy moving average filters. Here, u is a natural number smaller than N.

First, T_X≤T_MAnd T_X= T_MIn
Is determined, the N-fold moving average filter has the (N +
1) Describe the processing procedure when increasing the next moving average filter
I will tell. That is, as shown in step S302 of FIG.
When the L measurements have been completed, the total
Measurement data (w₁₁, W₁₂~ W_1M) To T shown in FIG.
_XEach measurement data w in the section₁₁~ W_1KIs obtained at each measurement
Point t₁~ T_KAt the start of measurement data reading
Divided into K sets of measurement data as points (read K sets)
Divided into sections), and the measured data of each group (section)
Perform calculation using (N + 1) multiple moving average filter, K
Weighing data string W₁₁~ W_1KIs calculated and read (S3
04). However, each of the first to Nth moving average filters
Number of taps D₁~ D_NIs the communication set in step S202.
It is. Number of taps of the (N + 1) th order moving average filter
D_{N + 1}Is the measurement data w_{1 (K + 1)}To
Read time t_{1 (K + 1)}To the number of measurement data up to
You. That is, the measurement data at the reading start point is w₁₁In the case of
(K + 1), and the measurement data at the reading start point is w
₁₂In the case of, set to K. Then, in the same manner,
Is set, and the measurement data at the reading start point is_1K
In the case of, set to 2. Similarly, the first to L-th
Measurement data obtained by weighing (w_{twenty one}, W_{twenty two}~ W_2M) ~ (W
_L1, W_L2~ W_LM) For each of the K weighing values
Data string (W_{twenty one}~ W_2K) ~ (W_L1~ W_LK) Calculate and read
Remove (S304). These weighing data strings (W₁₁~ W
_1K) ~ (W_L1~ W_LKIs the reading means.
You.

Next, these weighing value data (W₁₁~ W_1K)
~ (W_L1~ W_LK) Is changed to the reading start time t of the measurement data.
₁~ T_KAre grouped by common ones (W₁₁~
W_L1), (W₁₂~ W_L2) ~ (W_1i~ W_Li) ~ (W_1K~ W
_LK), And stores the grouped weighing value data in the storage unit 12.
Remember (S306). Then, the reading start time t₁of
Weighing value data (W₁₁~ W_L1)), For example,
Standard deviation σ₁Is calculated by the CPU 11. Similarly, CP
U11 is the reading start time t_Two, T_Three, ..., t_kof
Each pair (W₁₂~ W_L2) ~ (W_1K~ W _LK) Standard deviation σ
_Two, Σ_Three, ..., σ_kIs calculated (S306). Soshi
And these σ₁~ Σ_kAre arranged in the order of the number of times of weighing.
It is stored in the storage unit 12. It is a rose to calculate this standard deviation
It is an attachment degree calculating means.

Next, the CPU 11 sets these σ₁~ Σ_kBaby
The smallest one, for example σ_i(S30
8) At the start of reading measurement data with the best weighing accuracy
Point t _iTo determine. i is an integer from 1 to K. This
Reading start time t of measured data_iRead temporary
Instead of the starting point Q, it is determined to be the reading start point in the actual operation
This is the reading start time determining means.

Next, the article 6 shields the photosensor 5 from light.
Reads the weighing signal of the article after a predetermined time has passed since
Start time t_iThat is, the reading start time I shown in FIG.
(= T_i= T_Four) The setting of the timer time
You. Measurable section T_WFrom the time point P at the beginning of (w_1i~ W
_Li), The time T of the section PI up to the point I_PITo T _PI
= (I−1) × τ (S310).
Next, measurement data starts when the photo sensor 5 detects an article.
Ta (w_1i~ W_Li) Until the start of acquisition_aITo T
_aI= T_aP+ T_PIAnd the calculated time
T_aIIs set in the timer of the CPU 11 (S312).
This is the reading start time t_iThe timer time setting
Complete.

Further, the (N + 1) th order moving average filter
Number of taps D_{(N + 1)}To D_{(N + 1)}= K−i + 2
And stores it in the storage unit 12 (S314). this
And T _X≤T_MStart reading of weighing signal in case of
I, reading end time R, weight of multiple moving average filter
(N + 1), (N + 1) th Moving Average Filter Tap
Number D_{(N + 1)}Exit test mode
I do. In FIG. 7, the (N + 1) th order moving average fill
Tap number D_{(N + 1)}For the Nth order moving average filter
Number of taps D_NIs the original 5, (N-1) th moving average fill
Tap number D_(N-1)Is the original 4, ...
Is shown. w_4,1, W_5,1, ... are the (N + 1) th
Data obtained by the calculation of the next moving average filter, w
_4,2, W_5,2, ... are the performance of the Nth moving average filter.
Data obtained by arithmetic, w_4,3, W_5,3...
Is obtained by the calculation of the (N-1) th order moving average filter.
Data, W_{4, N + 1}Is the first-order moving average filter
This is data obtained by calculation, and this W_{4, N + 1}Object
Output as the weighing value of the product.

Next, when the CPU 11_X> T_MIt is determined that
The processing procedure when the setting is performed will be described. That is, the step shown in FIG.
As shown in step S302, when the L weighing is completed,
T shown in 8_XT in section_MIn the section, T_X≤T_M
In the same manner as in the case of
Calculates the filter and calculates the weighing data (W₁₁~ W_1K) ~
(W_L1~ W_LK). Next, T_XTime point of Q in section
Beyond S (= t_{(K + 3)}) Until the (N + 1) th
Nth moving average without calculating the next moving average filter
The reading start time of the measurement data in the filter
From the time point by τ time
Number of taps D of the Nth moving average filter_N= 5 or less
Three Nth moving average fills with 5, 4, and 3 taps
Data for each of the three N-th moving average filters.
The calculation of the N-fold moving average filter is performed for each data. This
Corresponding to the weighing value data described in step S304.
(W_{1 (K + 1)}, W_{1 (K + 2)}, W _{1 (K + 3)}), (W_{2 (K + 1)}, W
_{2 (K + 2)}, W_{2 (K + 3)}) ~ (W_{L (K + 1)}, W_{L (K + 2)},
W_{L (K + 3)}).

Here, when the time point S at the end point of the T _X section is the time point when the measurement data w _{K + 5} is obtained, the four N-th moving averages of 5, 4, 3, and 2 tap numbers are used. A filter operation is performed, and an N-fold moving average filter operation is performed for each of the four types of Nth moving average filters. by this,
(W) corresponding to the weighing value data described in step S304.
_{1 (K + 1)} , W1 _{(K + 2)} , W1 _{(K + 3)} , W1 _{(K + 4)} ),
(W2 _{(K + 1)} , W2 _{(K + 2)} , W2 _{(K + 3)} , W2 _{(K + 4)} )-(W
_{L (K + 1)} , WL _{(K + 2)} , WL _{(K + 3)} , WL _{(K + 4)} ).

Further, the time point of the end point of S of T _X section acquires the time t _{K + 5} (containing this point) and the measured data w _{K + 7} for acquiring measurement data w _{K + 5} t _{K + 7} (Including this time point), that is, if it is later than the reading start time t _{K + 5} of the measurement data of the (N-1) th order moving average filter, the (N + 1) th order, And without performing the calculation of the N-th order moving average filter, the reading start time of the measurement data in the (N-1) th order moving average filter is determined by the time τ time from the time t _{K + 5} at which the measurement data w _{K + 5} is read. The tap number D _(N-1) of the original (N-1) th order moving average filter is sequentially shifted in the direction of progress, and the number of taps whose number of taps is equal to or less than four is four, three, and two. ) Next moving average filter operation is performed, and (N-1) weights are calculated for each of the three (N-1) th moving average filters. Performing the calculation of the dynamic average filter. This corresponds to the weighing value data described in step S304 (W1 _{(K + 5)} , W1 _{(K + 6)} , W1 _{(K + 7)} ),
(W2 _{(K + 5)} , W2 _{(K + 6)} , W2 _{(K + 7)} )-(WL _{(K + 5)} , W
_{L (K + 6)} and WL _{(K + 7)} are obtained.

However, if the time point S at the end point of the T _X section is after the time point t _{K + 8} at which the measurement data w _{K + 8} is obtained, the subsequent processing is not performed. However, depending on the use of the dynamic weighing device, the order of the (N−2) th order moving average filter and the (N−th) order smaller than this order may be used.
u) Change the number of taps of the next moving average filter to obtain (N−
u) The weight value data may be obtained by performing a calculation of a heavy moving average filter.

In this manner, each section (w _{1 to} w _K ) to which the reading start time of the weighing signal belongs, (w _{K + 1} to
_{w K + 4), (w} K + 5 ~w K + 7) weighing for each data _{[(W 11 ~W 1K) ~ (} W L1 ~W LK) ], _{[(W 1 (K + 1)} ~
W1 _{(K + 4)} )-(WL _{(K + 1)} -WL _{(K + 4)} )], [(W1 _{(K + 5)
~W 1 (K + 7))} ~ (W L (K + 5) ~W L (K + 7) ] If is obtained, as in step S306, the read start point of the measurement data t ₁ ~t K _{+ 7} are divided into groups common to each other (W ₁₁ -W _L1 )-(W _1i -W _Li )-(W _1K -W _LK )-
(W1 _{( K + 7) to} WL _{(K + 7)} ), and the weighed value data thus grouped is stored in the storage unit 12 (see S306). And
As in step S306, the start time t ₁ read _{CPU11, t 2, t 3,} ···, t (K + 7) of each pair _{(W 11 ~W L1) ~ (} W 1 (K + 7 ₎ ~W L _{(K + 7)} standard deviation sigma ₁ of), σ _2, stores ···, σ _{(K + 7)} calculated in the storage unit 12.

Next, as in step S308, C
The PU 11 selects the smallest one of these σ _{1 to} σ _{(K + 7)} , for example, σ _i , and determines the reading start time t _i of the measurement data with the best weighing accuracy. i is 1 to
It is an integer up to (K + 7). Subsequent processing is equivalent to steps S310 to S314, and a description thereof will be omitted. Thus, when T _x > T _M , the reading start time I of the weighing signal, the reading end time R, the weight of the multiple moving average filter (N + 1), N or (N−u), and the (N + 1) th The number of taps D of the Nth or (Nu) th order moving average filter can be determined, and the test mode ends. In FIG. 9, N-th moving average tap number D _N 3 of the filter, the (N-1) number of taps of the next moving average filter D _N-1 is the original 4, shows an example in which a ... ing. w _{K + 3,1} , w _{K + 4,1} ,... represent data obtained by the calculation of the N-th moving average filter, w
_{K + 3,2} , w _{K + 4,2} ,... _Are the data obtained by the calculation of the (N−1) th order moving average filter, and W _{K + 3, N} is
This is data obtained by the calculation of the first-order moving average filter, and this W _{K + 3, N} is output as the weighing value of the article.

Next, the procedure for weight sorting of articles in the actual operation will be described with reference to the flowchart of FIG. After the weighing conditions such as the reading start time I of the weighing signal are automatically set as described above, the operation mode is set (S400). Thereafter, the articles 6 are transferred to the weighing conveyor 2.
When fed to the photo sensor 5 is shielded by the article 6 (S402), T _aI mse from the shaded point
When c has elapsed and the reading start time I of the weighing signal is reached, the CPU 11 sequentially reads the weighing signal in the IR section and stores it in the storage unit 12 (S406). Then, using the measurement data of the weighing signal stored in the storage unit 12, the data (N +
1), N, or (Nu) calculation is performed by a heavy moving average filter, and the weight value data W4 _{, (N + 1)} , or W of the article 6 is calculated.
_{(K + 3), N,} etc. are read (S408). Next, the read weighing value data is compared with a preset allowable weight value, and the comparison result is sent out to the conveyor 3 as a selection signal. As a result, the articles outside the allowable weight are sent to a different transport path from the articles within the allowable weight. In this way, it is possible to sort the weight of articles sequentially sent to the weighing conveyor 2.

In addition, the weighing condition determining apparatus is configured such that, during the actual operation, the weight sorter divides the articles 6 by a predetermined number of times.
Every time (≧ L) weighing is performed, the same processing method as that in which the weighing conditions (the weighing signal reading start point I, the number of taps, and the number of single or multiple moving average filters) are determined in the test mode is used. Then, calculation processing of the measurement data of the article 6 weighed during the actual operation is performed, a new weighing condition is automatically determined, and the new weighing condition set in the test mode is replaced with the new weighing condition during the actual operation. Can be changed to the weighing conditions determined in (1). Further, the weighing conditions set during the actual operation can be changed to the weighing conditions newly determined during the actual operation. That is, in the actual operation mode, the first to E-th weighings by the weight sorter are performed by weighing the weight of the article 6 based on the weighing conditions such as the reading start time point I set in the test mode. I do. Then, from the (E + 1) th weighing to the 2Eth weighing, the weight of the article is weighed based on the weighing conditions determined from the first weighing to the Lth weighing in the actual operation mode. Hereinafter, in the same manner, the (2E + 1)
From the first weighing to the third E weighing, the weight of the article is weighed based on the weighing conditions determined by the (E + 1) th weighing to the (E + L) th weighing in the actual operation mode.

Next, the processing contents of the CPU 11 will be described more specifically. A program for performing this processing is stored in the storage unit 12. (1) The CPU 11 adds 1 to the count value of the weighing counter every time the weighing is performed once in the actual operation. (2) When the count value of the weighing counter reaches E, the CPU 11 resets the count value and counts the number of weighings from the next time onward. (3) when the count value of the metering counter reaches E, CPU 11 is to set the first flag, all stores time-series measurement data between T _W metered L batch weighing signals thereafter The information is stored in the unit 12. The data acquisition counter counts L times. When the count value of the data acquisition counter becomes L, the count value is reset and the first flag is reset. (4) When the count value of the data acquisition counter reaches L, the CPU 11 sets the second flag. (5) When the second flag is set, the CPU 11 sets the weighing condition including the reading start time I of the weighing signal, the number of taps, and the number of single or multiple moving average filters in parallel with the process of weighing the article. The above calculation for determining the optimum weighing condition is performed, and when the count value of the weighing counter reaches E, the new weighing condition is set. Perform based on. Note that the second flag is reset when the new weighing condition is set.

As described above, a new reading start time I is determined by using a plurality of weighing values read by the reading means when the articles 6 are sequentially weighed during the actual operation, and the newly determined reading start time I is determined. And reading section IR corresponding to
The weight value of the article 6 can be obtained by calculating the weighing signal of the article 6 at any time during the actual operation. That is, for example, during the actual operation, the frictional resistance of the conveying surface of the weighing conveyor 2 with respect to the article 6 changes with the elapse of the operation time, and the shock (transient response vibration) generated when the article 6 gets on the weighing conveyor The weight of the article 6 can be accurately weighed even when the reading start time of the weighing signal for obtaining a relatively high weighing accuracy changes due to the change in the waveform.

Next, objects having different weights obtained during the actual operation
The best weighing conditions using measured data
Explain why you can do it. Now, the weight will be different during production
Assuming that L articles 6 have been measured, step S306 in FIG.
Weight data (W₁₁~ W_L1), (W₁₂~ W_L2) ~
(W_1i~ W_Li) ~ (W_1K~ W_LK) Is obtained. This
Here, the factors that cause the dispersion of the weighing data of each set
Is the variation due to differences in the weight of
Changes in the characteristics of the scale of conveyor 2 (for example,
Ft) and variations due to differences in transient response
And are duplicates. This variation degree is standard
As represented by the deviation σ, the first sequence (W₁₁~ W_L1)of
Standard deviation is σ₁, Standard deviation σ ₁Weight of goods contained in
Σ is the standard deviation based on_1a, Standard deviation σ₁To
Standard deviation based on the change in the characteristics of the included weighing conveyor 2
The difference is σ_1b, Standard deviation σ ₁In transient response included in
Standard deviation based on the variation due to_1cAnd As well
, The second sequence (W₁₂~ W_L2) Corresponding respective markers
The quasi-deviation is σ_Two, Σ _2a, Σ_2b, Σ_2cAnd These relationships
Is

[0048] a _{^{σ 1 2 = σ 1a 2 +}} σ 1b 2 + σ 1c 2 ··· (6) σ 2 2 = σ 2a 2 + σ 2b 2 + σ 2c 2 ··· (7).

Σ_1aAnd σ_2aIs compared to the first column
(W₁₁~ W_L1) And the second sequence (W₁₂~ W_L2) Weight
Standard deviation σ_1aAnd σ_2aAre the same L items
Since it represents the standard deviation of the weight, σ_1a= Σ_2aSay
And therefore σ_1a= Σ _2a= ... = σ_KaHolds
You. Therefore, the difference in the weight of the goods is
Deviation σ₁, Σ_TwoIt must be the cause of the difference
I can say. Therefore, during operation, articles with different weights
Weighing conditions newly obtained based on the measurement data of
Also, the weight of the article can be measured with high accuracy.

When σ _1b and σ _2b are compared, the metric value data (W ₁₁ -W _L1 ) and (W
₁₂ time intervals to _W-L2) is obtained, since it is very short τ time, the time interval for weighing data of the first row and the second row of the sequence is obtained, thus first K columns from the first column Time interval (K-1) × τ ≒ in which the metric data of the sequence of eyes is obtained
The change in the characteristics of the balance at 100 ms is very small and can be neglected. In addition, although there is variation in the characteristic change of the balance for each number of times of weighing, the standard deviations σ _1b and σ _2b of the characteristic change (drift) of the weigher represent the standard deviation of the characteristic change in the same L times of weighing. Therefore, it can be said that σ _1b = σ _2b . Therefore, σ _1b = σ _2b =... = Σ _Kb holds. Therefore, it can be said that the difference in the characteristic change of the balance does not cause the difference in the standard deviation σ ₁ , σ ₂ ,. When σ _1c and σ _2c are compared, the difference in the degree of impact disturbance and the degree of convergence in the transient response waveform can be almost ignored because the time difference is very short in the metric value data in the first and second columns. However, comparing σ _1c to σ _Kc from the first column to the Kth column, there is a difference in the degree of shock disturbance in the transient response waveform due to the time difference. Therefore, the difference between σ _1c and σ _Kc means a difference due to the transient response waveform, and the present invention detects a change in the degree of impact disturbance or the like in the transient response waveform and reduces an error based on the transient response waveform. The purpose is to select and set weighing conditions that can be removed.

However, in the above embodiment, FIGS.
It cited as an example metering condition determining apparatus applied to a dynamic weighing device for determining the metric value by performing an operation by single or multiple moving average filter measurement data read in I~R interval (T _IR hours) shown in Instead of this, weighing conditions applied to a dynamic weighing device that reads measurement data in a predetermined time zone for weighing an article and performs a calculation other than the above moving average filter to obtain a weighing value The determining device may be used.

In the above embodiment, in the test mode, the weighing conditions at the time of the actual operation are determined by weighing the reference articles having substantially the same weight and using the measurement data obtained by the weighing. Instead of the articles, articles having different weights may be weighed, and measurement data obtained by the weighing may be used to determine the weighing conditions in the actual operation.

In the above embodiment, the weighing condition determining device is applied to a weight sorter, but may be applied to other dynamic weighing devices such as a weight type filling weigher and a truck scale. In the case of application to a weight-type filling weigher, the time point P shown in FIGS. 6 to 9 corresponds to a time T from the timing when the feeding machine cuts out a predetermined weight of the article when the filling machine has delivered the article.
_This is the time point delayed by _d . The _Td time is an estimated time until the falling difference article existing between the article placed on the balance and the filling machine completely falls on the balance. In addition,
It is not necessary to accurately estimate the time _Td , and as described in the above embodiment, it is possible to select the reading start time I of the measurement data at which high-precision weighing can be performed.

Step S2 in FIG. 1B of the above embodiment.
In 02, the number of taps of the N-fold moving average filter is set in order to remove a typical periodic vibration noise included in the weighing signal of the dynamic weighing device, and the number of taps is determined by a conventionally known method. can do. For example, a representative system noise is extracted by performing a Fourier transform operation (FFT operation) on the measurement data to determine the number of taps so that Expression (1) is satisfied in consideration of the length of the time interval T _W. May be. Furthermore, a representative system noise may be calculated from the design conditions and operating conditions of the dynamic weighing device to determine the number of taps.

In the above embodiment, among the standard deviations σ calculated for each pair, the smallest one is selected. However, the second or third smallest standard deviation is selected, and the standard deviation σ of the selected standard deviation is selected. At the start of reading the set of weighing signals,
Digital filters can be used during production.
In short, a desired standard deviation can be selected according to the weighing accuracy required for the dynamic weighing device.

In the above embodiment, as shown in FIG. 6, the first time point of the section of T _X is P, but may be a time point after the P time point in the direction of elapse of time. In other words, since the measurement data around the point P includes a large shock disturbance in the transient response waveform, it may be inappropriate to read the measurement data as the measurement data for calculating the weighing data.
This can be omitted. Thereby, the time for calculating such measurement data can be omitted.

[0057] In the above embodiment, than when a lapse T _aI time from when the article is shielded photosensor 5 A /
Although the D converter 9 is started to read and store the weighing signal at each measurement time point, a comparator can be used instead of the photosensor 5. That is, a threshold value can be set, and the time point when the analog weighing signal first exceeds the threshold value can be set as the start time point of the A / D converter 9. Then, a period during which the analog weighing signal exceeds the threshold may be set as each measurement time point. Further, the A / D converter 9 is always operated, and the digital signal output from the A / D converter 9 is compared with a predetermined comparison value set in advance.
A digital signal exceeding this comparison value may be stored in the storage unit 12 as a weighing signal at each measurement point. Therefore, the weighing signal stored in the storage unit 12 can be used as measurement data.

In the above embodiment, the standard deviation σ is used as an evaluation criterion for the degree of dispersion of the weighing data.
² may be used to compare the degree of dispersion of the weighing data of each group, or the deviation R between the maximum value and the minimum value of the weighing data of each group may be used to compare the variance of the weighing data of each group. The degrees may be compared.

In the above embodiment, the number of taps of a filter such as the N-th or (N-1) -th moving average filter is changed so that high-precision weighing can be performed.
Although the number of filters is reduced in order from the N-th moving average filter to reduce the order of the multiple moving average filter,
As the moving average filter for changing the number of taps, for example, a filter having a relatively large number of taps is preferable. The reason is that a slight decrease in the number of taps does not significantly affect the noise removal characteristics of this filter. Then, it is preferable that a filter for removing noise having relatively little influence on the weighing accuracy be selected as the moving average filter to be removed. However, as a filter for changing the number of taps and a filter for removing, all the filters of the N-fold moving average filter are sequentially applied to obtain metric data for each filter, and a filter with the most stable metric data is determined as the filter. Is also good.

[0060]

According to the first aspect of the present invention, when an article is weighed, a weighing signal read for each of a plurality of reading sections at a plurality of different reading start times is calculated for each section, and each of these reading sections is calculated. Determine the weighing value for each.
Then, the process of sequentially weighing a plurality of articles and determining the weighing value for each reading section is repeated. Next, the plurality of weighed values are grouped and stored for those having a common reading start time. Then, the variation degree calculating means,
Calculate the degree of dispersion of the weighing values for each set, select a relatively small degree of dispersion among the plurality of degrees of dispersion, and determine the reading start time corresponding to the set of the degree of dispersion as the reading start time in the actual operation. can do. As a result, there is an effect that the reading start time of the weighing signal, which can weigh the weight of the article relatively accurately (for example, most accurately), can be automatically determined in a short time. Therefore, even an inexperienced operator can determine an optimal reading start time of a weighing signal, similarly to a skilled worker.

According to the second aspect of the present invention, each of the plurality of reading sections in which the reading start time of the weighing signal is different from each other is performed a plurality of times by the multiple moving average calculation of the different sampling numbers or multiples, thereby performing each reading. By obtaining the weighing value of the article for each section and selecting a reading section in which the variance of the weighing value is relatively small, the sampling number or the multiple of the multiple moving average operation in which the variance of the weighing value is relatively small is selected. be able to. Accordingly, high-precision weighing can be performed by performing multiple moving average calculation of the corresponding sampling number or multiple for the weighing signal in the selected reading section. That is, for example, in order to remove at least one type of periodic vibration noise from the weighing signal, the reading start time Q of the weighing signal shown in FIG. 13 is made earlier (moved in a direction approaching P). It is necessary to lengthen the reading interval T _F of the weighing signal. However, if the reading start time Q of the weighing signal is set too early, the weighing signal including the shock (transient response vibration waveform) generated when the article 6 gets on the weighing conveyor 2 will be read, and the weighing accuracy will be reduced. In some cases. Therefore, the present invention calculates a weighing value by calculating for each of a plurality of reading sections in which the reading start time of the weighing signal is different from each other, and selects a reading section in which the degree of variation in the weighing value is relatively small, In addition, it is possible to determine a reading start point (reading section) of a weighing signal at which relatively high weighing accuracy is obtained.

According to the third invention, a new reading start time is determined by using a plurality of weighing values read by the reading means when articles are sequentially weighed during the actual operation, and the newly determined reading start time is determined. The weighing signal of the article can be obtained by calculating the weighing signal of the reading section corresponding to, and at any time during the actual operation, the article is weighed based on the weighing signal from the new reading start time sequentially changed. The weight can be accurately measured. That is, for example, during the actual operation, the frictional resistance of the conveying surface of the weighing conveyor with respect to the article changes with the elapse of the operation time, and the impact (transient response vibration waveform) generated when the article gets on the weighing conveyor is generated. Even when the reading start time of the weighing signal that provides relatively high weighing accuracy changes due to the change, the weight of the article can be accurately weighed based on the weighing signal read from the new reading start time after the change. it can.

[Brief description of the drawings]

FIGS. 1A and 1B are flowcharts showing a procedure for setting various data before actual operation by a weighing condition determination device of a dynamic weighing device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a procedure for determining a weighing condition such as a reading start time of a weighing signal in a test mode before the actual operation by the weighing condition determining apparatus according to the embodiment.

FIG. 3 is a flowchart showing a procedure for determining a weighing condition such as a start time of reading a weighing signal in a test mode before the actual operation by the weighing condition determining apparatus according to the embodiment.

FIG. 4 is a flowchart showing a procedure for determining a weighing condition such as a reading start time of a weighing signal in a test mode before the actual operation by the weighing condition determining apparatus according to the embodiment.

FIG. 5 is a flowchart showing a procedure in which the weight sorter weighs the article during the actual operation according to the weighing conditions set by the weighing condition determining device according to the embodiment.

FIG. 6 is a waveform diagram showing a weighing signal processing state and a weighing signal before the weighing condition is determined by the weighing condition determining apparatus according to the embodiment.

FIG. 7 is a waveform diagram showing a weighing signal processing state and a weighing signal after a weighing condition is determined by the weighing condition determining apparatus according to the embodiment.

FIG. 8 is a waveform diagram showing a weighing signal processing state and a weighing signal before the weighing condition is determined by the weighing condition determining apparatus according to the embodiment.

FIG. 9 is a waveform diagram showing a processing state of a weighing signal and a weighing signal after the weighing condition is determined by the weighing condition determining apparatus according to the embodiment.

FIG. 10 is a block diagram showing an electric circuit of the weight sorter and the weighing condition determination device of the embodiment.

FIG. 11 is a diagram showing an outline of a conventional weight sorter, a measurement timing, and the like.

FIG. 12 is a diagram showing a matrix of measurement data obtained by a conventional weighing condition determination device.

FIG. 13 is a waveform diagram showing a processing state of a weighing signal and a weighing signal after determination of a weighing condition by another conventional weighing condition determining apparatus.

[Explanation of symbols]

 4 Weight detector 9 A / D converter 11 CPU 12 Storage unit

Claims (3)

[Claims] An article is calculated by calculating a weighing signal in a reading section from a reading start time of a preset weighing signal to a reading end time of a preset weighing signal among the weighing signals output from the weight detector. In a dynamic weighing device for determining a weighing value, a weighing signal is read for each of a plurality of reading sections at a plurality of different reading start times when weighing an article by the weight detector, and the read weighing signal is A reading means for calculating each section and reading the weighed value calculated for each reading section, and a plurality of weighed values read by the reading means when the article is weighed a plurality of times, having a common reading start time A storage unit for storing the weighed values that are grouped together and storing the weighed values, and a degree of variation of each weighed value stored in the storage unit. Means for calculating the degree of variation calculated for each set, and selecting a relatively small degree of variation from among the plurality of degrees of variation calculated by the degree of variation calculation, and setting a read start time corresponding to the set of degrees of variation in the actual operation. A weighing condition determining device for a dynamic weighing device, comprising: a reading start time determining means for determining a reading start time. 2. The weighing condition determination device for a dynamic weighing device according to claim 1, wherein the reading means sets a plurality of different reading start times when a plurality of different reading start times are obtained when the weight detector weighs the articles. A weighing signal is read for each section, a single or multiple moving average calculation is performed on the weighing signal using a predetermined sampling number as a filter constant for each reading section, and the weighing value calculated for each of these reading sections is read. A weighing condition determination device for a dynamic weighing device, characterized by comprising changing means for changing the sampling number or the weight of the multiple moving average calculation according to each of the reading sections. 3. The weighing condition determining device for a dynamic weighing device according to claim 1, wherein the storage means reads a plurality of weighing values read by the reading means when the articles are weighed a plurality of times during the actual operation. Characterized in that the reading start times are grouped in common, and each of the grouped weighing values is stored.