JPH06123019A - Method for computing control parameter of spinning tension of roving in roving frame - Google Patents

Abstract

PURPOSE:To relate the detection of the spinning tension of a roving in the roving frame, and to provide the method for computing the spinning tension of the roving, giving an accurate detection value as a control parameter even when the extremely limited number of sensors are used. CONSTITUTION:When the spinning tensions of rovings in the roving frame are fed back and controlled by selecting specific plural spindles, setting sensors to the selected spindles, detecting the spinning tension values of the rovings, and subsequently using the average value of the detected values as a control parameter, (1) an average value Xmean is determined from the detected values (X1, X2,...Xn), and the maximum value (Xmax) and the minimum value (Xmin) are selected; (2) the difference (Xmax-Xmin=alpha) between the maximum value and the minimum value is computed; (3) the alpha is compared with a reference value delta; (4) in the case of alpha<=delta, the Xmean is adopted as a control parameter; (5) in the case of alpha>delta, a value separated from the Xmean is removed from the Xmax and the Xmin; and the steps (1)-(5) are again repeated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention twists a roving spun from a front roller of a draft mechanism at a constant speed in a roving machine by a flyer and winds the roving on a bobbin rotating at a higher speed. It relates to the detection of the spinning tension of the roving as it is taken to form a roving package.

[0002]

2. Description of the Related Art The roving spun from a front roller is
It is preferable that the slack is always wound with a certain degree of slack, and if the slack changes, the stretching applied to the roving during the winding changes, and the weight per unit length changes, which is not preferable. Therefore, a mechanism for keeping the slack of the roving yarn during spinning constant is proposed for the roving machine.

A mechanism is known in which a pair of cone drums are used to actuate a belt shifter each time the roving layer increases to move the belt wound around the cone drum to reduce the rotational speed of the bobbin. is there. JP-A-60-146016
In addition to this mechanism, the publication discloses a second belt shifter moving mechanism capable of operating the belt shifter irrespective of the increase in the roving layer, so that the roving being spun can be spun between the front roller and the flyer top. A non-contact type sensor for detecting the position is provided, and the propriety of the roving thread tension is judged based on the output signal of this sensor. Based on the result, the second belt shifter moving mechanism is operated to change the winding amount of the roving thread. A method of performing feedback control that adjusts the amount is disclosed.

This non-contact type sensor is composed of a plurality of pairs of light emitting elements and light receiving elements which are arranged in parallel in a vertical direction in multiple stages and are opposed to each other. Since the light between the elements at the roving yarn passage position is blocked, the roving yarn passage position is detected and output as a value corresponding to the spinning tension. This sensor is provided on a plurality of specific weights (for example, 3 weights) of the roving machine,
An output signal corresponding to the position of the roving yarn emitted from each sensor is input to a control computer, an average value is calculated, and based on this, the belt shifter moving mechanism operates to spin-out tension (passing position) of the roving yarn. The foot-back control is performed to control the to a proper value.

[0005]

Considering the drive mechanism of each part of the roving machine, the spinning tension of the roving yarn in each spindle of the roving machine should not be so different if the spinning is in a steady state. is there. However, if Shinogatari is left unattended for a long time and then Shinotsugi is performed and operation is restarted,
An abnormally low spinning tension may be detected only in that weight because the bobbin has a different diameter. If such an abnormality happens to occur in the sensor-installed weight, the average value of the three weights becomes unsuitable as a parameter for controlling the entire weight of the roving machine.

The present invention solves the above problems in the prior art, and even when a very limited number of sensors are used to detect the position of the roving during spinning, a proper detection value as a control parameter is obtained. It is an object of the present invention to provide a method for calculating the spinning tension of a roving that yields

[0007]

The object of the present invention is to detect a value of the spinning tension of a roving by selecting a plurality of specific weights in the roving machine and to use the average value of the detected values as a control parameter of the roving machine. When performing the feedback control of the spinning tension of the roving, 1) The average value X _mean is calculated from the detection values (X ₁ , X ₂ , ... X _n ) and the maximum value (X _max ) and the minimum value (X _min ) is selected, 2) the difference between the maximum value and the minimum value (X _max −X _min = α) is calculated, 3) α is compared with the reference value δ, and 4) when α ≦ δ, The _{feature is} that X _mean is adopted as a control parameter, and when α> δ, the ones apart from X _{mean in} X _max and X _min are excluded, and the steps 1) to 4) are repeated again. The method for calculating the control parameter of the spinning tension of the roving, and the value of the spinning tension of the roving detected by selecting a plurality of specific weights in the roving machine, Upon performing spinning tension feedback control of the average value as a control parameter Rovers roving value, 1) the respective detected values _{(X 1, X 2, ...} , X i ..., the average value from X _n)
Calculate X _mean and standard deviation σ, and 2) Calculate the absolute value (ε = ｜ X _i −X _mean ｜) of the difference between each detected value (X ₁ , X ₂ ,… X _n ) and the mean value X _mean. , 3) Each detected value is 0 ≤ ε <k ₁ σ, k ₁ σ ≤ ε <k ₂ σ, k ₂ σ ≤
ε <k ₃ σ, ─k _m−2 σ ≦ ε <k _m−1 σ, k _m−1 σ ≦ ε are divided into m groups, and 4) w ₁ and w ₂ are detected values of the respective groups. , w ₃ , ─w _m (however, w ₁ ＞ w ₂ ＞ w ₃ ─ ＞ w _m )
(X ₁ ', X ₂ ', ... X _n '), 5) Based on the corrected detection value, a corrected average value X _mean ' is obtained, which is a control parameter for the spinning tension of the roving yarn. It is achieved by a calculation method.

[0008]

According to the first aspect of the present invention, the detected values of the spinning tension of the roving yarn from the respective sensors are first averaged as a whole to obtain the average value X _mean . Next, the difference α between the average value and each detected value is obtained, and compared with a reference value δ empirically obtained from the standard deviation at the time of normal spinning to determine whether each detected value is abnormal. To be judged. That is, when α> δ, the detected value is determined to be abnormal, excluded from the control parameter calculation, and the average value is calculated again using only the remaining detected values. This step is α ≦ δ for all detected values
Is repeated until. As a result, abnormal ones are excluded from the detected values, so that the spinning tension of the roving yarn can be properly controlled by using the finally obtained average value as the control parameter.

However, it is desirable to provide a regulation for preventing the number of data for obtaining the average value from becoming too small. According to the second aspect of the present invention, all the detection values detected by the sensor are weighted according to the difference from the simple average value, and the closer to the average value the higher weight is given to correct the correction value. The detected value is calculated. Then, the average value is calculated again by the corrected average value, and the corrected average value X _mean 'is obtained and used as the control parameter. According to this, since all the detected values can be corrected according to the degree of the abnormality and can participate in the calculation of the control parameter, a small number of data can be effectively utilized.

Based on the preferred embodiment shown in the drawings,
The present invention will be described in more detail.

[0011]

1 shows an example of a system for controlling the roving tension in a roving spinning machine. The front roller 1, the flyer 2 and the top cone drum 3 are driven by a common main motor 4 via a rotation transmission means such as a gear train and a timing belt, and the bobbin 5 is rotated by the main motor 4 and the top cone is rotated. The rotation from the bottom cone drum 7, which is rotated at a variable speed from the drum 3 via the belt 6, is the differential mechanism 8
It is designed to be combined with.

A long rack 10, to which a belt shifter 9 for moving the belt 6 is fixed, can reciprocate laterally to a position where it meshes with a pinion 13 fitted to an upright shaft 12 which rotates when the weight 11 descends. It is installed in. On the upright shaft 12,
The rotation of the rotary shaft 18 fixed to the shaft of the ratchet wheel 15 that is regularly rotated by a predetermined amount when the bobbin rail 14 is switched up and down is transmitted via a differential gear mechanism 19.
Each time the roving layer wound on the bobbin 5 increases, that is, when the bobbin rail 14 is switched up and down, the pinion 13 is rotated by a certain amount, and the long rack 10 is moved leftward in FIG. 1 at a certain pitch. There is.

The differential gear mechanism 19 has a bevel gear 20 fixed to the lower end of the upright shaft 12, a bevel gear 21 fixed to the upper end of the section rotary shaft 18, and both bevel gears 2.
It is composed of a bevel gear 22 that meshes with 0 and 21.
The bevel gear 22 has a tooth portion 25a that meshes with a worm 24 driven by a reversible motor 23, and is supported by a rotating body 25 as a belt shifter moving mechanism that rotates about the upright shaft 12 and the rotating shaft 18 as an axis. Therefore, when the reversible motor 23 is driven with the ratchet wheel 15 fixed, the bevel gear 22 revolves around the shaft center of the upright shaft 12 together with the rotating body 25 via the worm 24 to rotate the bevel gear 20. The belt shifter 9 is moved and adjusted via the pinion 13 and the long rack 10.

The belt shifter 9 is designed to move in the right direction in FIG. 1, that is, in the direction in which the rotation of the bottom cone drum 7 is slowed when the reversible motor 23 is normally rotated. The non-contact type sensor 26 for continuously detecting the position of the roving R between the front roller 1 and the flyer top 2a includes a light emitting part 27 and a light receiving part 28 which are installed opposite to each other as shown in FIG. And the roving R is located between the two. The light emitting portion 27 is composed of an infrared light emitting diode array, and the light receiving portion 28 has a light receiving element array composed of a large number of light receiving elements 28a arranged in parallel in the vertical direction at a pitch of about half (1 mm) of the roving R. There is. Each light receiving element 28a is configured to generate an electric signal according to the intensity of the received light and input the electric signal to the microcomputer M.

Since the roving thread R blocks a part of the light from the light emitting section 27, the light receiving element 28a corresponding to the position of the roving thread R does not receive the light. Therefore, the roving thread R is detected by detecting the light receiving element. The position of can be calculated. When the roving tension is low, the position of the roving is lowered downward, and when it is high, the position is moved upward. The sensor 26 is installed by selecting a plurality of suitable weights of the roving frame, for example, three weights. The detected value of the roving yarn position from each sensor 26 is input to the microcomputer M, where it is processed to determine the tension state. Then, when it is determined that the tension is low, a signal for driving the reversible motor 23 in the forward direction is output, and when it is determined that the tension is high, a signal for rotating the reversible motor 23 in the reverse direction is output for a predetermined time. This signal current is amplified by the amplifier 29, and the reversible motor 23 is normally or reversely rotated for a predetermined time via the output relay 30 to change the rotation speed of the bobbin 5 via the bottom cone drum 7 to wind the roving. Adjust
The tension is controlled so as to reach a predetermined value.

The present invention relates to a method of processing when the detected value input from each sensor 26 to the microcomputer M contains an abnormal value which is very different from the others.
As described above, when such an abnormal value is included, unless some processing is performed, another normal detected value is dragged to this abnormal value, and the entire roving machine is illegally controlled. Because there is a risk.

In the first embodiment of the present invention, the following processing is performed. It is assumed that n sensors 26 are installed in the roving machine, and the detection values X ₁ , X ₂ , ... X _n are output from the respective sensors. As the first step, the simple average X _mean = (X ₁ + X ₂ + ... + X _n ) / n of all the detected values is calculated, and the maximum value X _max and the minimum value X _min are selected.

As the second step, the selected maximum value X
_Find the difference between _max and the minimum value X _min α = X _max −X _min .
As a third step, the preset reference value δ is compared with the difference α obtained in the second step. This reference value δ is a value determined from the standard deviation of the past tension value data, etc., and if the difference α between the maximum value and the minimum value is smaller than this, it can not be said to be statistically abnormal, but this When the value exceeds the limit, the value is clearly judged to be abnormal.

As a fourth step, as a result of this comparison, α
When ≦ δ, it is considered that the data does not include any outliers, and the simple mean value X _mean obtained in the first step
Is adopted as a parameter for control. Conversely, α>
In the case of δ, one of the maximum value X _max or the minimum value X _min is considered to be an abnormal value, and the value farther from the average value X _mean , for example, X _min is eliminated and reduced by one. The average value is calculated again using the detected values, and steps 1 to 4 are repeated.

However, since the number of sensors to be installed is limited and the number of data that can be used for one calculation is small originally, it is not preferable that the number of data is further reduced by repeating the above steps. Therefore, as a result of excluding this abnormal value, when the number of data m becomes n / 2 + 1> m ≧ n / 2, the exclusion of the abnormal value is stopped here and the simple average value at this point is used. It is used as a control parameter. That is, the limit is added so that the number of data is not less than half the initial number.

This operation step is shown in the flow chart of FIG. In order to solve the problem of the decrease in the number of data, which is the drawback of the first embodiment, in the second embodiment of the present invention, all the detected value data input to the microcomputer are effectively used. There is. That is, as the first step, the detected values from the sensors X ₁ , X ₂ , ..., X _i ... _,
From X _n , the average value X _mean = (X ₁ + X ₂ +… + X _i … + X _n ) / n and the standard deviation σ are calculated.

Next, as a second step, each detected value (X ₁ ,
X ₂ , ... X _n ) and the absolute value of the _mean X _mean ε
= | X _i − X _mean | is calculated. From this result, the third
As a step, each detected value is 0 ≤ ε <σ, σ ≤ ε <2
It is divided into 4 groups of σ, 2 σ ≦ ε <3 σ, 3 σ ≦ ε. This is because each detected value is centered on the _mean value X _mean , σ, 2σ,
This means classifying into each range of 3σ and those belonging to a range outside 3σ. These groups are designated as 1st to 4th, respectively.
It is called a group.

Then, in a fourth step, w ₁ , w ₂ , w ₃ , w ₄ (where w ₁ > w ₂ > w ₃ > w ₄ ) are added to the detected values of the respective groups.
The weighting of the ratio of the corrected detection values X ₁ ', X ₂ ',… X
Calculate _n '. As a result, the data closer to the average value becomes larger, and the data farther from the center becomes smaller. Using the corrected detection value thus obtained, the mean value X _mean 'is calculated again, and this is used as the control parameter.

Specifically, if w ₁ , w ₂ , w ₃ , w ₄ are weighted with 10, ₃ , ₁ , 0, for example, a small amount of data outside the limit of 3σ, which is clearly regarded as abnormal, is used. Is completely eliminated. The data of the first group is X ₁ , X ₂ , X ₃ , the data of the second group is X ₄ , X ₅ , the data of the third group is X ₆ , and the data of the fourth group is X _7.
Then, the modified mean value X _mean 'is calculated by the following formula.

X _mean '= [10 (X ₁ + X ₂ + X ₃ ) +3 (X ₄ + X ₅ ) + X ₆ ] / (10 × 3 + 3 × 2 + 1 × 1) This second implementation According to the example, all the data detected by the sensor can be participated in the calculation of the control parameter except the apparently abnormal data, and the small amount of data can be effectively used.

[0026]

According to the present invention, it is determined whether the roving thread tension value detected by the sensor is abnormal according to a certain standard, and it is eliminated or corrected to control the roving thread tension. Therefore, the risk of incorrect control due to an abnormal value is prevented.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a configuration of a roving machine for carrying out the present invention.

FIG. 2 is a schematic view showing an example of a non-contact sensor used in the present invention.

FIG. 3 is a flowchart of the first embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Front roller 2 ... Flyer 3 ... Top cone drum 4 ... Main motor 4 5 ... Bobbin 7 ... Bottom cone drum 9 ... Belt shifter 9 23 ... Reversible motor 26 ... Non-contact type sensor M ... Microcomputer M

Claims (3)

[Claims] 1. A roving machine selects a plurality of specific weights and installs a sensor to detect the value of the spinning tension of the roving, and the average value of the detected values is used as a control parameter for the roving of the roving machine. When performing the feedback control of the spinning tension, 1) the average value _Xmean is obtained from each of the detection values (X ₁ , X ₂ , ... X _n ), and the maximum value (X _max ) and the minimum value (X _min ) are determined. Select, 2) calculate the difference between the maximum and minimum values (X _max −X _min = α), 3) compare α with the reference value δ, and 4) if α ≤ δ, use X _mean It is adopted as a control parameter. 5) When α> δ, the ones apart from X _{mean in} X _max and X _min are excluded, and the steps 1) to 4) are repeated again. Method of calculating control parameter of spinning tension of roving. 2. In the step 5), as a result of eliminating the abnormal detection value, the number (m) of remaining detection values is n / 2 + 1> m ≧ n as compared with the first number (n) of detection values. The calculation method according to claim 1, wherein when it becomes / 2, the average value X _mean at that time is used as a control parameter. 3. The roving machine selects a plurality of specific weights and installs a sensor to detect the value of the spinning tension of the roving, and the average value of the detected values is used as a control parameter for the roving of the roving machine. When performing the feedback control of the spinning tension, 1) the average value from the detected values (X ₁ , X ₂ , ..., X _i ... _, X _n )
Calculate X _mean and standard deviation σ, and 2) Calculate the absolute value (ε = ｜ X _i −X _mean ｜) of the difference between each detected value (X ₁ , X ₂ ,… X _n ) and the mean value X _mean. , 3) Each detected value is 0 ≤ ε <k ₁ σ, k ₁ σ ≤ ε <k ₂ σ, k ₂ σ ≤
ε <k ₃ σ, ─k _m−2 σ ≦ ε <k _m−1 σ, k _m−1 σ ≦ ε are divided into m groups, and 4) w ₁ and w ₂ are detected values of the respective groups. , w ₃ , ─w _m (however, w ₁ ＞ w ₂ ＞ w ₃ ─ ＞ w _m )
(X ₁ ', X ₂ ', ... X _n '), 5) Based on the corrected detection value, a corrected average value X _mean ' is obtained, which is a control parameter for the spinning tension of the roving yarn. Calculation method.