JPH03248024A - Measurement controlling apparatus for quantitative scale - Google Patents

Abstract

PURPOSE:To adjust delicate input time quickly by increasing and decreasing the changing weight for the input flow rate into the next stage in each stage and the input flow rate at each stage based on the time deviation between the preset input time at the input flow rate at each stage and the actual input time based on fuzzy control. CONSTITUTION:The detected weight signal from a load cell 6 provided at a measuring hopper 5 undergoes A/D conversion 8 through a low-pass filter 7. The signal is imparted into a CPU 9. The CPU 9 computes the weight of a measured material 4 which is inputted into the hopper 5 based on the detected weight signal. The actual input times in large input, medium input and small input are measured with a built-in timer counter. In a fuzzy controller 10, fuzzy operation is performed in accordance with specified rules based on the time deviation between each actual input time and each preset input time. In order that the measured material having the specified weight is filled highly accurately in a specified time, the large input flow rate at the next time, the large-input-flow-rate changing weight for cutting the large input flow rate and the medium (small) -input-flow-rate changing weight for cutting the medium (small) flow rate are set.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention changes the input flow rate of the object to be weighed in multiple stages from high to low each time the input flow rate reaches a switching weight. The present invention relates to a weighing control device for a metering scale that weighs and fills a weighing object of a predetermined weight by switching in the following manner.

(Prior Art) Conventionally, the following is disclosed in Japanese Patent Laid-Open Publication No. 63-279120 regarding a weighing control device for a metering scale of this type.

Measure the total loading time of the object to be weighed from the start of loading to the time of stopping loading, and if this total loading time is not within the preset loading time, the total loading time will be determined to be within the loading set time. The input flow rate switching weight is increased/decreased based on PID control so that the input flow rate is within the range.

(Problem to be Solved by the Invention) However, in the above-mentioned method, when the total charging time deviates from the set charging time due to changes in the density of the object to be measured, etc., the charging flow rate is determined only by information on the total charging time. Since the total charging time is adjusted by increasing or decreasing the switching weight, there is a problem in that extremely fine adjustment of the charging time cannot be made. Further, since it is PID control, there is a problem that adjustment time is long.

The present invention aims to solve the above-mentioned problems.Even if the total charging time fluctuates due to disturbances such as changes in density during operation, it is possible to make extremely fine adjustment of the charging time. It is an object of the present invention to provide a weighing control device for a quantitative scale that requires only a short time.

(Means for Solving the Problems) In order to achieve the above-mentioned object, the weighing control device for a quantitative weighing scale according to the present invention, in the above-mentioned device, sets the antecedent variable of the control rule according to the input flow rate at each stage. The consequent variable of the control rule is set to the initial value of the switching weight for the input flow rate to the next stage at each stage, or the input flow rate to the next stage at the previous stage, corresponding to the time deviation of the actual input time from the set input time. It is provided with a fuzzy control means that corresponds to the increase/decrease in the switching weight and the initial value of the input flow rate at each stage or the increase/decrease in the input flow rate at the previous stage.

(Function) Based on fuzzy control, the switching weight of the input flow rate to the next stage at each stage and the input flow rate at each stage are increased or decreased depending on the time deviation of the actual input time from the input set time due to the input flow rate at each stage.

(Effects of the Invention) Therefore, since it is adjusted by the time deviation of the actual dosing time from the set dosing time at each stage, even if the total dosing time fluctuates due to disturbances such as density changes during operation, extremely fine dosing can be achieved. You can adjust the time. Furthermore, since the overall dosing time pattern based on the set dosing time of the dosing flow rate at each stage is maintained, stable metering is guaranteed. Furthermore, since parallel inference processing is performed using fuzzy control, the adjustment time is short.

(Example) Next, as a specific example of the weighing control device for a quantitative scale according to the present invention, a case where it is applied to a servo input type packer scale will be described with reference to the drawings.

In FIG. 1, a charging gate 2 provided in a charging hopper 1 is opened and closed by an AC servo motor 3, and a weighing object 4 stored in the charging hopper 1 is weighed according to the opening degree of the charging gate 2. A large amount of investment is made sequentially in one cycle of the act.

The material is charged into the weighing hopper 5 by medium and small amounts.

A loading pattern is realized by these large loading, medium loading, and small loading, and large loading is a loading operation in which the loading gate 2 is changed from fully closed to fully open, and medium loading is a loading operation in which the opening of loading gate 2 is fully opened. The small charge is a charge operation in which the amount is gradually reduced from 1 to 1, and the small charge is a charge operation in which the opening degree of the charge gate 2 is maintained at the end of the medium charge to ensure measurement accuracy.

The weighing hopper 5 is provided with a load cell 6 that outputs a detection signal according to the weight of the weighed object 4 put therein. The weight detection signal from the load cell 6 is sent to the A/D converter 8 after vibration components are removed by a low-pass filter.
is converted into a digital value and given to the CPU 9. This CPU 9 calculates the weight of the weighing object 4 put into the weighing hopper 5 by processing the digitalized weight detection signal, and also calculates large loading, medium loading, small loading, etc. using a built-in timer counter as described later. Measure the actual time of each input. The measured actual input times are given to the fuzzy controller 10, and the fuzzy controller 1o performs fuzzy calculations according to a predetermined control rule based on the time deviation between the actual input times and the set input times. , Large input flow rate due to the next large input flow to fill the weighed object 4 of a predetermined weight within a predetermined time with high accuracy, large input flow rate switching weight to cut the large input flow rate, medium (small) input flow rate due to medium (small) input. Set the medium (small) input flow rate switching weight to cut the input flow rate. These set values for the large input flow rate and each input flow rate switching weight are returned to the CPU 9, and the CPU 9 sends command signals to the servo amplifier 11 for opening and closing the input gate 2 based on these set values and the calculated weight of the weighing object 40. give to In this servo amplifier 11, the AC servo motor 3 is controlled by a command signal for opening/closing the input gate 2 and an output signal of a potentiometer 12 that monitors the opening degree of the input gate 2, and the AC servo motor 3 is controlled to open or close the input gate 2. Adjust the temperature.

Furthermore, when the weighing object 4 in the weighing hopper 5 reaches a predetermined weight, the input gate 2 is closed by a command signal from the CPU 9 to complete the weighing, and then the discharge cylinder 13 is driven by a command signal from the CPU 9 to open the discharge gate. 14 is opened, and the weighed object 4 is put into a container (not shown) such as a bag or can, and quantitative filling is performed. Note that 15 is a display section that displays set values such as large input flow rate and the actual input times set in the fuzzy controller 10, and 16 is a setting section that inputs settings and corrections of each input set time and the like.

Next, while referring to the relationship between the weight of the weighing object 4 thrown into the weighing hopper 5 by adjusting the opening of the charging gate 2 and the time shown in FIG. 2, the so-called charging time pattern in the weighing operation, The CPU 9 input processing will be explained based on the flowchart shown in FIG. In addition, the second
In the figure, W is the predetermined weight that is the target filling weight of the weighed object 4, WI is the large input flow rate switching weight that cuts the large input flow rate U due to large input, and W2 is the small input flow rate U-+y that is cut due to small input. The small injection flow rate switching weight, W, indicates the small injection flow rate switching weight that cuts the small injection flow rate y due to small injection.

i) Large flow (Steps S-1 to 5-4) Command to fully open the flow gate 2 to the position corresponding to the large flow i1u set in the fuzzy controller 10 based on a push button or input command from the outside. A signal is issued and charging is started with a large charging flow rate U. At the same time, the timer counter T starts measuring time, and the weight is measured by the fuzzy controller lO.
The actual large injection time t required for large injection until the large injection flow rate switching weight W1 set in is exceeded is measured.

Since it takes time for the weighing object 4 to reach the bottom of the weighing hopper 5 from the input hopper 1, the weight starts to be detected after a time Te after the input gate 2 begins to open, and this Te time is a so-called delay. It's time.

fi) Small loading (Steps S-5 to 5-8) In the small loading, a fuzzy controller is used to prevent vibration from occurring in weight detection, as detailed in Japanese Patent Laid-Open No. 62-230527. Large input flow rate U set at 10
A command signal is issued to gradually reduce the opening degree of the charging gate 2 according to the charging weight from 1 to a preset small charging flow rate y, and charging is performed at a small charging flow rate u-+y. Similarly, the actual small injection time t required for the small injection until the weight exceeds the small injection flow rate switching weight W2 set in the fuzzy controller 10 after starting time measurement by the timer counter T.
! Measure.

ii) Small injection (Steps S-9 to S-13) During small injection, a command signal is sent to maintain the opening degree of the injection gate 2 at the end of the small injection at the position corresponding to the preset small injection flow Iy. and inject it with a small input flow rate y of a constant flow rate. Similarly, time measurement is started by the timer counter T, and the actual small injection time L3 required for the small injection until the weight exceeds the small injection flow rate switching type tW3 set in the fuzzy controller 1o is measured.

Small input flow rate switching type! w3 is a value obtained by subtracting the weight drop amount P of the weighing object 4 which has not yet reached the weighing hopper 5 even when the input gate 2 is fully closed from the predetermined weight W. Note that this head difference amount P also includes the weight of the object to be weighed 4 that has fallen from the time when the CPU 9 issues a command signal to fully open the input gate 2 until the input gate 2 is actually fully closed.

In this way, one cycle of input is completed.

iv) Overshoot detection (steps 5-14 to S-
16) By the way, as for the evaluation of loading, the shorter the loading time, the better, but if the loading time is too short, the weighing object 4 will be thrown violently into the weighing hopper 5, and the response will be larger than the actual weight. This is included in the weight detection signal of the load cell 6, and the weight detection signal of the load cell 6 no longer accurately reflects the weight. This state is called an overshoot, and as shown in Figure 4, the dosing time pattern in which overshoot occurs differs from the normal dosing time pattern, with large, small, and small inputs. The input time for each stage is extremely short. Therefore, this allows overshoot to be detected.

After one cycle of input, each input actual time t1, t
! , t 3 to detect the presence or absence of overshoot,
When an overshoot is detected, a warning is output, and correction processing is performed so that the operation processing is performed based on various set values such as the large input flow rate U that was set by the fuzzy controller 10 immediately before the overshoot occurred. .

Next, the fuzzy controller 10 will be explained in detail based on FIG.

First of all, optimal feeding control means that a large feeding flow rate switching weight Wl, a small weight weight Wl, and a small This means adjusting the input switching weight W2, the small input flow rate switching weight W3, the large input flow rate U, and the small input flow rate y. However, in this embodiment, the small input flow ity is the head amount P which cannot be directly adjusted.
Frequent adjustments that significantly affect shall be taken as a thing.

Therefore, the small input flow rate y and the related small input flow rate switching weight W, (=W-P) are fixed, and the large input flow rate switching weight Wl, the small input flow rate switching weight W2, and the large input flow rate U are fixed. We will adjust and set it to realize an arbitrary input time pattern.

When the filling of the weighed object 4 of the predetermined type 1w in the k-th step is completed without overshooting, the CPU 9 calculates a large-time actual time L + ( k), the actual small injection time tz(k), and the actual small injection time ts(k) are given to the deviation calculation unit 10A. This deviation calculation section 10A
In the above, the following deviation calculation is performed using the large injection setting time ttl, the small injection setting time ttz, and the small injection setting time tts as target times set by the setting section 16.

Large input time deviation Δt t+(k)= t+(k)
t Lll entrance time deviation Δt tx(k)=
t z (k) t t *Small injection time deviation Δt t
s(k)=t 5(k) t tsNext, each time deviation ΔLt+(k), Δttt(k), ΔLt! (k)
is given to the normalization unit 10B, and normalization calculation is performed using the gain constant G t; (i = 1 to 3) for normalization as follows, and the set (-1, 1) is Normalized.

Normalized large input time deviation Δt+Oc) =Gt+
xΔt t+ (k) Normalized small injection time deviation Δ
t-z(k) = Gt, xΔttz(k) normalized small injection time savings ΔLs(k) = Gts
k) These normalized input time deviations Δt+(k)
.

Δtz(k), Δti(k) are given to the fuzzy inference unit 10C, and in this fuzzy inference unit 10C, the normalized input time deviations Δt+(k), Δtz
(k) and Δt x (k), a fuzzy calculation is performed based on the control rule from the control rule section 10D.

By the way, setting the desired input time pattern is as follows:
Although it is very difficult to operate multiple parameters at the same time, it takes three people to adjust the input time (Δt
+(f), Δtg(k), Δt=(k) ) 3 outputs (W
l(k+1), Wl(k+1), u(k+1)), and if a fuzzy control method is used, the know-how of experts can be incorporated into the control rules. In order to incorporate this into the control rule, in this example, the inference method is Mamdani's method based on fuzzy relationship synthesis agents, and both the antecedent part variable and the consequent part variable of the control rule are the membership function of the fuzzy variable. Let us use the triangle shown in FIG. 6 as a function representing .

Note that each symbol in FIG. 6 has the following meaning.

nb: Negative and large. pS: Positive and small.

ns: Negative and small. pb: Positive and large.

zo; Quite small.

Therefore, the normalized time deviations Δt + (k), Δt z (k), and Δt 3 (k) are used as antecedent variables, and large input flow rate switching type IW, medium input flow rate switching type are used as consequent variables. Increase/decrease in weight W2 and large input flow rate U ΔW1. Δ
W2. If ΔU is used, the i-th control rule can be expressed by the following format.

ifΔtl(k)=A+′ andΔt z(k)=A
z” andΔ1s(k)=A−thenΔW,=B−
, ΔW,=B,', Δu=B (if Δt, l) is A
Δt at I' t(k) is A2i and Δts(k) is A
3”, then ΔW+ is Bl'+ ΔW, is B2゛,
ΔU is B,''.

Note that A-~A3'lB1''~B, 5 are membership functions of fuzzy variables.

An example of a rule based on the know-how of an expert in adjusting the injection time based on this control rule is as follows.

Nal ifΔj ,=p b thenΔW,=n
s, Δu=p bNo, 2 ifΔt ,=p s
thenΔW,=zo, Δu=p sNo, 3
ifΔj,=zo thenΔW,=zo,Δ
W,=ZO,Δu=z.

Nα4 ifΔt,=ns thenΔW,=zo
, Δu=n sN[15ifΔt,=fi b th
enΔW,=ps, Δu=nbN[16ifΔt,=
p b thenΔW, = pb, Δu=p bNo7
5fΔt2=ps thenΔW,=psNa8
ifΔl,=z o thenΔW,=zo,
ΔL=z.

No, 9 ifΔt,=ns thenΔW,=n
sNalOifΔtz=n b thenΔW,=nb
, Δu=nbNail ifΔt ,=p b th
enΔW,=psN[112ifΔt3=ps the
nΔW,=z.

No, 13 ifΔt==zo thenΔWz=z
o,ΔW,=z.

Nn14 ifΔt,=ns thenΔW2=n
sNo, 15 ifΔt 3= n b thenΔ
W,=nb,ΔW,=nbNa16 ifΔt4=z
oandΔt 3=Z OthenΔW,=z.

Nα17 ifΔtl=zoandΔt t=z
o thenΔW,=zo,Δu=z.

Next, specific processing in the fuzzy inference section 10D will be explained based on the above-mentioned i-th control rule.

The fitness degree ω5 is calculated using the following formula.

ω, -AI' (Δt +) ×A! i (Δt z)
XA3' (Δt3) In addition, when the number of antecedent variables is 2 or 4 or more, the same calculation is performed, and
When there is one antecedent variable, the fuzzy value of that antecedent variable is used as the fitness.

Membership function BI'lB2'', B, 1's centroid G
If I ' HG t ' + G 3 ', then i
The independent inference result for the th control rule is as follows.

Increase/decrease in weight for switching to large input flow rate; Δw, i=c, increase/decrease in weight for switching flow rate for middle shell; ΔW25=G, increase/decrease in weight for large input flow rate; Δu''=G3'' Next, membership function B l' ＋ B ti＊ B
-area of SI'+S! ”+S-, the estimation results for all control rules ΔW, #, ΔW!#, Δu
l is as follows.

Large input flow rate switching weight increase/decrease; (°, °Σi is the sum of all control rules including ΔW1) (°, °Σj is the sum of all control rules including ΔW2) These fuzzy inference results ΔWI #, Δ
W2#.

Since ΔU# has been normalized to (-1, 1), it is given to the inverse normalization unit 10E, and the gain constants Gw, , Gw, , G are used for the inverse normalization as follows. Inverse normalization calculation is performed.

Increase/decrease in weight when switching to large flow rate; ΔVL(k+1)=GwIXΔW1# Increase/decrease in weight when switching to medium flow rate; ΔWz(k+1)=Gw=XΔW! # Large input flow rate increase/decrease; Δu(k+1)-G, XΔuI Next, in the output calculation unit 10F, each previous setting parameter VL(k), Wz(k), u(k)
is added to the setting parameter W+ as the current setting amount.
(k+1), Wz(k+1), and u(k+1) are calculated.

Large input flow rate switching weight; W, (k+1) = ΔW+(k + 1) +Wt(k
) Middle stage input flow rate switching weight; Wt(k +1) = Δ”ILCk +1”) +
Wz(k) large input flow rate; u(k+1)=Δu(k+1)+u(k) These setting parameters W+(k+1), w, (k+1).

u(k+1) is given to the CPU 9, and the next setting parameters Wl(k+2), Wt(k+2
).

It is held in the output calculation unit 10F in order to calculate u(k+2).

The setting parameter Wl(k+1) obtained in this way.

The +1st filling is performed by Wz (k + 1), u (k + 1), and the actual charging times t + (k + 1), tz (k + 1), t, (
k+1) is obtained, and the same process as the second time is repeated in the fuzzy controller IO. In this way, the desired dosing time pattern (t t+, t tz
, t ts) are obtained. Note that the above-mentioned calculation regarding the fuzzy control is performed between the end of one input process and the start of the next input process.

Next, the fuzzy controller IO using the aforementioned fuzzy inference
An example of what was achieved is explained below.

(Example 1) Using a backer scale as shown in Figure 1, the object to be measured 4 is a polyethylene pellet, the predetermined weight W is set to 25 kg, the small input flow rate y is set to 23 kg, and the consequent variable is the subject of fuzzy inference. The initial value and input setting time were set as follows.

Large input flow rate switching weight W,; 10 kg Medium input flow rate OFF Medium input flow rate switching weight 4 kg Large input Large input flow rate switching weight Constant time Ltt; 1 second t; 1 second Lt3i 1 second Note that this input setting time L LI+ t tz, L L
3 is the shortest input time pattern that can ensure the accuracy obtained from past experience.

Under these conditions, the following data were obtained.

It can be seen that in this example 1, a satisfactory input time pattern is achieved at the 6th step.

Setting parameters at this time, (w,,w,,u) = (10,49,24,2
7, 53) are the adjusted setting parameters. It can be seen that after the 7th step when this adjustment is achieved, fine adjustments are made every time the filling is performed, and the achieved injection time pattern is maintained in response to disturbances and environmental changes.

24.00 Mu, 23 Edict, 89 24.14 24.14 A, 14 A, 27 24.27 A, 31 A, 31 24.31 10.00 10.07 10.09 10.21 10.32 10. 43 10.49 10, 54 10, 46 10.43 10.42 Gain '1ljlC111: 0.5 Gwt:
0.5 Gu: 1aCt25ctz:5
Gtz: 5 (Example 2) The following data were obtained under the same conditions as Example 1 except that the small input flow rate was 24 kg.

24.00 edict, 73 edict, 98 edict, 81 24.0I n, 91 24.03 24.01 24.01 u, 01 24.01 u, 01 24.01 24.01 24.01 Separate, 01 10. 00 10.03 10, 15 10.25 10.36 10.46 10.36 10.22 per 1O1 10.12 10.01 9.91 9.85 9.80 9.77 9.73 Gain “83!ZGbn : 0.5 Cw!:
0.5 Gu: IGt+: 5 Gt
zi 5 Gti: 5According to this embodiment, even an inexperienced operator can meet the target input time (
Lt+.

By simply setting L L2 + L t4), the optimum input setting value can be determined, which is useful for labor saving. In addition, even if the charging time changes due to fluctuations in gas density during operation, the deviation from the target time is detected every time filling is completed, and the same operation as when determining the optimal setting value is used to Find the input flow rate switching weight, large input flow rate, and correct the input time. Furthermore, even when handling different objects to be weighed, there is no need to set a new input setting value, and the optimum input setting value can be automatically determined.

In this embodiment, the servo closing method has been described, but the same effect can be obtained even if it is applied to conventional two-stage or three-stage closing. Further, in this embodiment, a natural fall type gate system using a charging gate has been described, but it can also be applied to horizontal screw, vertical screw electromagnetic feeder systems, etc.

[Brief explanation of drawings]

1 to 5 are drawings for explaining a specific embodiment of the weighing control device for a quantitative weighing scale according to the present invention. FIG. 3 is a flowchart of the input process, FIG. 4 is an overshoot input pattern, FIG. 5 is a block circuit diagram of the fuzzy controller, and FIG. 6 is an explanatory diagram of fuzzy variables. Loading hopper Loading gate AC servo motor Object weighing hopper Load cell Low-pass filter A/D converter PU Fuzzy controller Servo amplifier Potentiometer Discharge cylinder Discharge gate Display section Setting section

Claims (1)

[Scope of Claims] 1. Weighing a predetermined weight of the object by switching the input flow rate of the object to be measured in multiple stages from high to low each time each input flow rate switching weight is reached. In a weighing control device for a metering scale for filling, the antecedent variable of the control rule is made to correspond to the time deviation of the actual dosing time from the set dosing time due to the dosing flow rate at each stage, and the consequent variable of the control rule is The initial value of the input flow rate switching weight to the next stage or the increase/decrease from the previous input flow rate switching weight to the next stage at that stage, and the increase/decrease from the initial value of the input flow rate at each stage or the previous input flow rate at that stage. A weighing control device for a quantitative scale, characterized by comprising fuzzy control means corresponding to minutes. 2. The metering control device for a metering scale according to claim 1, wherein the input flow rate in the final stage of the input flow rates in the plurality of stages is a predetermined flow rate.