[0001] The present invention relates to a method of operating a weaving machine for inserting
weft threads into sheds formed between warp threads during weaving cycles.
[0002] For inserting weft threads into sheds that, during each weaving cycle, are formed
between the warp threads extending in the warp direction within a rapier weaving machine,
weft thread inserting means comprising an inserting rapier for inserting a weft thread
into a shed from one side of the shed and a receiving rapier for receiving the weft
thread approximately in the center of the shed in the weft direction from the inserting
rapier and moving the weft thread transferred to the receiving rapier to the other
side of the shed are provided.
[0003] WO 2019/082 222 A1 discloses a gripper assembly with a bringer gripper acting as an inserting rapier
and a taker gripper acting as a receiving rapier of a passive weft inserting means,
also known as a negative weft inserting means. The weft thread inserted by the bringer
gripper is directly transferred from the bringer gripper to the taker gripper, when
both grippers are in a mutually overlapping positioning such that, upon a reverse
movement of the bringer gripper, the taker gripper grips the weft thread and, upon
a reverse movement of the taker gripper, moves the gripped weft thread to the other
end of the shed. In active weft inserting means, also known as positive weft inserting
means, active means to operate the clamps of the rapier, that hold the weft, such
as blades, hooks or cams for example, are used for actuating the closing or the opening
of these clamps in order to transfer the weft thread from the inserting rapier to
the receiving rapier in the middle of the shed.
[0004] A lot of different weft errors can occur during a weaving process. For example, the
inserting rapier may fail to correctly grip the weft thread offered thereto and may
start moving without having a weft thread gripped. Further, the inserting rapier may
fail to correctly transfer a weft thread to the receiving rapier so that, after the
transfer failure, the receiving rapier starts moving without having a weft thread
gripped. A weft thread may break during each phase of the inserting process, and a
scissor provided for cutting an inserted weft thread may fail to cut this weft thread.
It is essential to provide information about such a weft error as soon as possible
for immediately shutting down the weaving machine and stopping the weaving process
and, for example, pulling out a broken weft thread.
[0005] Therefore, it is the object of the present invention to provide a method of operating
a weaving machine for inserting weft threads into sheds formed between warp threads
during weaving cycles with an enhanced capability of recognizing weft errors.
[0006] For achieving this object, the present invention provides a method of operating a
weaving machine for inserting weft threads into sheds formed between warp threads
during weaving cycles, the weaving machine comprising at least one weft thread inserting
means with an inserting rapier for inserting a weft thread and a receiving rapier
for receiving the weft thread inserted by the inserting rapier, the weaving machine
further comprising a weft thread tension sensor for outputting a weft thread tension
signal indicative of a weft thread tension of a weft thread inserted by the at least
one weft thread inserting means.
[0007] The method comprises the steps of:
- A) providing a plurality of consecutive weft thread tension values within a weaving
cycle based on the weft thread tension signal,
- B) providing a predetermined number of consecutive weft thread tension reference values
on the basis of the plurality of consecutive weft thread tension values in association
with each weft thread tension measurement point of a plurality of consecutive weft
thread tension measurement points within the weaving cycle, the weft thread tension
reference values of the predetermined number of consecutive weft thread tension reference
values being indicative of the weft thread tension prior to the associated weft thread
tension measurement point,
- C) determining whether a weft error has occurred within the weaving cycle based on
the plurality of consecutive weft thread tension reference values associated with
at least one of the weft thread tension measurement points of the plurality of consecutive
weft thread tension measurement points associated with the weaving cycle.
[0008] Using weft thread tension reference values associated with respective weft thread
tension measurement points in weaving cycles provides, in association with each weft
thread tension measurement point, a pattern of weft thread tension values preceding
a respective weft thread tension measurement point that is indicative of the development
of the tension of a weft thread prior to each weft thread tension measurement point.
It has recognized that, if a weft error occurs, these patterns show an immediate reaction.
Therefore, evaluating these patterns indicative of the development of the weft thread
tension prior to each weft thread tension measurement point allows providing information
about the occurrence of a weft error with a short delay after each weft thread tension
measurement point. By monitoring these patterns, the time necessary for detecting
a weft error can be reduced from some 40° of the rotation of a main shaft of a weaving
machine on average to below 10°.
[0009] For providing a clearly defined set of weft thread tension measurement points, each
weaving cycle may be associated with a predetermined angular range of rotation of
a weaving machine main shaft, and each weft thread tension measurement point may correspond
to an angle of rotation of the weaving machine main shaft within the predetermined
angular range of rotation.
[0010] The predetermined angular range of rotation may be 360°. Further, the plurality of
consecutive weft thread tension measurement points associated with the weaving cycle
may be equidistantly distributed within the weaving cycle. For example, a weft thread
tension measurement point may be provided per each degree of the angular range of
rotation, so that each weaving cycle comprises 360 weft thread tension measurement
points.
[0011] For providing the weft thread tension reference values in an easy and clearly defined
manner, at least one of the weft thread tension reference values of the plurality
of consecutive weft thread tension reference values associated with a weft thread
tension measurement point, preferably each weft thread tension reference value of
the plurality of consecutive weft thread tension reference values associated with
a weft thread tension measurement point, may be based one of the weft thread tension
values of the plurality of consecutive weft thread tension values laying within a
predetermined measurement zone preceding the associated weft thread tension measurement
point.
[0012] In an advantageous embodiment requiring a reduced scope of calculating work, at least
one of the weft thread tension reference values of the plurality of consecutive weft
thread tension reference values associated with a weft thread tension measurement
point, preferably each weft thread tension reference value of the plurality of consecutive
weft thread tension reference values associated with a weft thread tension measurement
point, may correspond to one of the weft thread tension values of the plurality of
consecutive weft thread tension values laying within the predetermined measurement
zone.
[0013] For also considering developments in previous successful weaving cycles in which
no error has occurred, at least one of the weft thread tension reference values of
the plurality of consecutive weft thread tension reference values associated with
a weft thread tension measurement point, preferably each weft thread tension reference
value of the plurality of consecutive weft thread tension reference values associated
with a weft thread tension measurement point, may correspond to a mean value of a
weft thread tension value of the plurality of consecutive weft thread tension values
laying within the predetermined measurement zone obtained in a plurality of consecutive
weaving cycles, preferably including the instantaneous weaving cycle. The mean value
of the weft thread tension values may be a weighted mean value, in which the relevance
of the weft thread tension values used for calculating the mean value decreases from
the weft thread tension value of the most recent weaving cycle of the plurality of
consecutive weaving cycles, for example, the instantaneous weaving cycle to the weft
thread tension value of the oldest weaving cycle of the plurality of consecutive weaving
cycles.
[0014] If a weaving cycling is associated with a predetermined angular range of rotation
of the main shaft of a weaving machine, the predetermined measurement zone may correspond
to a measurement angular range of rotation of the weaving machine main shaft preceding
the associated weft thread tension measurement point.
[0015] For evaluating the sets of weft thread tension reference values associated with the
weft thread tension measurement points, the method may further comprise an inputting
step for inputting the weft thread tension reference values associated with at least
one of the weft thread tension measurement points of the plurality of consecutive
weft thread tension measurement points within the weaving cycle into an artificial
neuronal network as input values, and the artificial neuronal network may be arranged
for providing an output indicating whether a weft error has occurred or not.
[0016] For a quick evaluation of the condition of a weft thread during a weaving cycle,
the weft thread tension reference values associated with a plurality of the weft thread
tension measurement points of the plurality of consecutive weft thread tension measurement
points within the weaving cycle, preferably the weft thread tension reference values
associated with each one of the plurality of consecutive weft thread tension measurement
points within the weaving cycle, may be forwarded to the artificial neuronal network
for providing, in association with each weft thread tension measurement point of the
plurality of weft thread tension measurement points, an output indicating whether
a weft error has occurred or not.
[0017] The method may further comprise a learning step for teaching to the neuronal network
to provide the output based on the input values.
[0018] This learning step may comprise:
- a) providing a plurality of learning sets of weft thread tension reference values,
each learning set corresponding to the consecutive weft thread tension reference values
associated with one of the weft thread tension measurement points of a terminated
weaving cycle,
- b) in association with each learning set, providing a weft error information indicating
whether, in association with a respective learning set, a weft error has occurred
or not,
- c) inputting one of the learning sets of the plurality of learning sets as an input
pattern into the artificial neuronal network,
- d) comparing the output of the artificial neuronal network with the weft error information
associated with the input learning set,
- e) adapting a configuration of the artificial neuronal network at least if the output
of the artificial neuronal network does not match the weft error information associated
with the input learning set.
[0019] With such a learning step, the artificial neuronal network can be trained with using
data provided in previous weaving cycles and containing information about the development
of the weft thread tension prior to each weft thread tension measurement point, on
the one hand, and about whether, in association with such a development of the weft
thread tension, a weft error has occurred or not, on the other hand.
[0020] For enhancing the capability of the artificial neuronal network of correctly deciding
whether a weft error has occurred or not, steps c) to e) may be repeatedly carried
out so that, with an increasing number of repetitions, the percentage of repetitions
providing an output of the artificial neuronal network matching the weft error information
associated with the input learning set increases.
[0021] For example, the learning step may comprise a supervised learning procedure.
[0022] The present invention will now be explained with reference to the enclosed drawings
in which:
- Fig. 1
- is a schematic representation of a weaving machine and weft insertion means associated
therewith;
- Fig. 2
- shows the position of an inserting rapier, a receiving rapier and a weft thread transferred
therebetween during one weaving cycle;
- Fig. 3
- shows the weft thread tension during one weaving cycle;
- Fig. 4
- shows a plurality of weft thread tension values indicative of the weft thread tension
for a portion of a weaving cycle.
[0023] Fig. 1 shows, in a schematic manner, a loom or weaving machine 10. The weaving machine
10 has a weft insertion means 12 for inserting a weft thread 14 into a shed formed
between non-shown warp threads extending in a warp direction of the weaving machine
10. The weft insertion means 12 comprises an inserting rapier 16 and a receiving rapier
18 that, in Fig. 1, are shown in a retracted position in which these rapiers 16, 18
are withdrawn from the shed formed between the warp threads, and are shown in an advanced
position in which respective gripper portions of these rapiers 16, 18 inserted into
the shed are positioned such as to overlap each other in the area of a center C of
the shed in a weft direction W for transferring the weft thread 14 inserted into the
shed by the inserting rapier 16 to the receiving rapier 18. In the preferred embodiment
shown in Fig. 1, the weft insertion means 12 is embodied as a passive or negative
weft insertion means arranged for directly transferring the weft thread 14 from the
inserting rapier 16 to the receiving rapier 18 without using active means to open
or to close the clamps of the rapiers for clamping the weft.
[0024] For forwarding the weft thread 14 to the inserting rapier, a bobbin 20 for providing
the weft thread 14 and a feeder 22 buffering a portion of the weft thread 14 are provided.
If required, a knot detector 24 can be provided or can be integrated into the feeder
22. Following to the feeder 22, a weft brake 26 arranged for applying a braking force
to the weft thread 14, a weft detector 28 for detecting a movement of the weft thread
14, a weft tension sensor 30 for outputting a weft thread tension signal indicative
of the tension of the weft thread 14 between the weft brake 26 and the one of the
rapiers 16, 18 by means of which the weft thread 14 is gripped, a weft thread recuperator
32 as well as a weft selector 34 are provided in the weaving machine 10. If a plurality
of different weft threads 14 are provided for generating specific weft thread patterns,
such a weft selector 34 is used for offering the one weft thread that is to be used
in a specific weaving cycle to the inserting rapier 16. Further, a weft scissor 36
is provided for cutting the weft thread 14 after a portion thereof has been inserted
into a shed and for clamping the weft thread 14 and presenting it to the inserting
rapier and before its next portion will be inserted into the next shed formed between
the warp threads. Further, a weft release actuator 38 is associated with the receiving
rapier 18 for releasing the weft thread 14 from the receiving rapier 18 after the
receiving rapier 18 has approached its retracted position and the weft thread 14 has
been inserted into the shed.
[0025] It is to be noted that the weaving machine 10 may comprise a plurality of such weft
insertion means 12 for simultaneously inserting a plurality of weft threads into a
plurality of sheds formed between the warp threads and/or for inserting a plurality
of weft threads at different levels, for example, if the weaving machine 10 is a face-to-face
carpet weaving machine. Each such weft insertion means 12 may comprise an inserting
rapier 16 and a receiving rapier 18 and, if the weft insertion means 12 is an active
or positive weft insertion means, additionally may comprise active means for transferring
the weft thread from the inserting rapier to the receiving rapier.
[0026] The braking force applied by the weft brake 26 to the weft thread 14 during each
weaving cycle of a weaving process can be controlled based on the weft tension signal
output by the weft tension sensor 30 to a weft machine controller 40. The weft machine
controller 40 that, for example, provides a sampling of the weft tension signal continuously
output by the weft tension sensor 30 with a desired sampling frequency of some hundreds
or thousands hertz, controls the weft brake 26 by outputting a control signal for
adjusting the braking force applied during each weaving cycle for adapting the braking
force applied to the weft thread 14 to the movement of the inserting rapier 16 and
the receiving rapier 18, respectively.
[0027] Fig. 2 shows the movement of the inserting rapier 16, the receiving rapier 18 and
the weft thread 14 transferred therebetween during one weaving cycle. It is to be
noted that, normally, such a weaving cycle is defined by one complete rotation of
a main shaft of the weaving machine 10. If the weaving machine 10 is a carpet weaving
machine, the weaving shaft of the weaving machine 10 makes about 130 to 200 or even
up to 250 rotations per minute, which means that, per minute, a corresponding number
of weft thread insertion operations are carried out by the weft insertion means 12.
As the width of a carpet to be woven by such a carpet weaving machine may be up to
about 5.3 meters, during each weft thread inserting operation, the weft thread 14
is moved by the inserting rapier 16 and the receiving rapier 18 through the shed having
a corresponding extension of up to more than 5 meters in the weft direction W.
[0028] Fig. 3 shows the weft thread tension signal indicative of the tension of the weft
thread 14 as detected by the weft tension sensor 30. It can be seen in Fig. 3 that,
with an increasing speed of the weft thread 14 after the inserting rapier 16 has started
moving together with the gripped weft thread 14, the tension of the weft thread 14
increases while, in a phase in which the inserting rapier 16 approaches the area of
the center C of the shed and decelerates for transferring the weft thread 14 to the
also decelerating receiving rapier 18, the tension of the weft thread 14 decreases.
[0029] Once the weft thread 14 has been transferred to the receiving rapier 18 and the receiving
rapier 18 starts accelerating for pulling the weft thread 14 toward the other end
of the shed, there is a sharp increase of the tension of the weft thread 14 that is
reflected in the weft thread tension signal in the area of a machine position between
180° and 200° of the rotation of the weaving machine main shaft within the shown weaving
cycle. After this transfer peak of the weft thread tension has occurred, there is
a corresponding sharp decrease of the weft thread tension and, thereafter, the tension
of the weft thread 14 increases with the increasing speed of the receiving rapier
18 and, upon approaching the other end of the shed, decreases with the decreasing
speed of the receiving rapier 18.
[0030] By controlling the braking force applied to the weft thread 14 by the weft brake
26 during each weaving cycle on the basis of the weft thread tension signal output
by the weft tension sensor 30, adverse effects like looping or ballooning of the weft
thread 14 prior to and during the transfer thereof from the inserting rapier 16 to
the receiving rapier 18 can be avoided. Further, an excessive braking force increasing
the risk of a failure of the weft thread 14 can be avoided. However, the risk that
weft errors, for example, a failure to grip the weft thread 14 with the inserting
rapier 16 or the receiving rapier 18 or a failure thereof during the movement of the
weft thread 14 occur, can not be eliminated completely.
[0031] With reference to Fig. 4, a method for evaluating the condition of the weft thread
14 based on the weft thread tension signal output by the weft thread tension sensor
30 and for providing information about the occurrence of a weft error as soon as possible
will be explained.
[0032] Fig. 4 shows the weft thread tension for a portion of a weaving cycle. Within each
weaving cycle, a plurality of weft thread tension values V
T is provided, for example, by sampling the weft thread tension signal output by the
weft thread tension sensor 30 by means of the weft machine controller 40. The weft
thread tension values V
T deviate from a mean value of the weft thread tension and, as can be seen in Fig.
4, this deviation that, for example, may be expressed by means of the standard deviation,
may increase with an increasing weft thread tension and/or an increasing speed of
the weft thread 14.
[0033] A plurality of weft thread tension measurement points is defined in association with
each weaving cycle. Fig. 4 shows two weft thread tension measurement points P
i, P
y. For example, one such weft thread tension measurement point may be defined in association
with each degree of the rotation of the weaving machine main shaft, so that each weaving
cycle has 360 weft thread tension measurement points associated therewith.
[0034] In association with each one of the weft thread tension measurement points P
i, P
y, a measurement zone Z
i, Z
y is provided. Each such measurement zone Z
i, Z
y may be defined as an angular range of rotation of the weaving machine main shaft
prior to each associated weft thread tension measurement point P
i, P
y. Each measurement zone Z
i, Z
y covers a plurality of weft thread tension values V
T. Depending on the number of weft thread tension values V
T covered by each measurement zone Z
i, Z
y, each one of these weft thread tension values V
T of the instantaneous weaving cycling may be used as a weft thread tension reference
value V
R associated with a respective one of the measurement zones Z
i, Z
y, or only a part of the weft thread tension values V
T, for example, fourty weft thread tension values V
T within a respective measurement zone Z
i, Z
y may be used as weft tension reference values V
R. As indicated in Fig. 4, in an alternative option for defining the weft tension reference
values, weft thread tension reference values V
T' may be defined by interpolating between adjacent weft thread tension values V
T. In an alternative embodiment, for each weft thread tension value V
T that, in association with each one of the weft thread tension measurement points
P
i, P
y will have to be considered, a mean value can be calculated with using this particular
weft thread tension value V
T of the instantaneous weaving cycle and the corresponding weft thread tension values
V
T of a plurality of previous weaving cycles in which no weft thread error has occurred.
This mean value of the weft thread tension values may be a weighted mean value, in
which the relevance of the weft thread tension values used for calculating the mean
value decreases from the weft thread tension value of the most recent weaving cycle
of the plurality of consecutive weaving cycles, for example, the instantaneous weaving
cycle to the weft thread tension value of the oldest weaving cycle of the plurality
of consecutive weaving cycles.
[0035] Each weft thread tension measurement point P
i, P
y and the associated set of weft thread tension reference values V
R, respectively, is used as a set of data for evaluating the condition of the weft
thread 14 inserted during a respective weaving cycle. This is done by teaching to
an artificial neuronal network to decide whether, in association with such a set of
data, a weft error has occurred or not.
[0036] In a learning procedure of the neuronal network that, for example, may be a supervised
learning procedure, a plurality of such sets of data comprising the weft thread tension
reference values V
R associated with respective weft thread tension measurement points P
i, P
y of previous, already completed weaving cycles are provided as learning sets and are
input into the artificial neuronal network as input patterns. In association with
each such learning set of weft thread tension reference values V
R, information indicating whether a weft error has occurred in the weaving cycle associated
with the respective learning set is input into the artificial neuronal network. The
output of the artificial neuronal network is compared to this weft error information
indicating whether a weft error has actually occurred or not and, if the output of
the artificial neuronal network does not match this information, the configuration
of the artificial neuronal network is adapted.
[0037] By repeatedly inputting such learning sets and the associated weft error information
and comparing the output of the artificial neuronal network with the weft error information
and, if there is a deviation, adapting the artificial neuronal network, the reliability
of the output is increased such that, with an increasing probability, the output of
the artificial neuronal network correctly reflects the weft error information associated
with the learning sets input as input patterns into the artificial neuronal network.
[0038] Such an appropriately trained artificial neuronal network is used during a weaving
process for providing information about the occurrence of a weft error. During each
weaving cycle of a weaving process, the weft thread tension reference values V
R associated with each one of the weft thread tension measurement points P
i, P
y are forwarded to the artificial neuronal network as input patterns. Based on these
weft thread tension reference values V
R, the trained artificial neuronal network provides an output in association with each
weft thread tension measurement point P
i, P
y indicating whether the development of the weft thread tension within the measurement
zone Z
i, Z
y prior to each weft thread tension measurement point P
i, P
y is indicative of the occurrence of a weft error.
[0039] As this evaluation of the development of the weft thread tension may be carried out
for each weft thread tension measurement point and, therefore, may be carried out
for each degree of the rotation of the weaving machine main shaft, the information
indicating whether a weft error has occurred can be provided with a high reliability
and essentially without delay after each weft thread tension measurement point. As
soon as the output of the artificial neuronal network indicates that there is the
high probability that a weft error has occurred in association with a respective weft
thread tension measurement point, the weaving machine can be stopped and the weaving
process can be interrupted for allowing a supervisor to check whether a weft error
has actually occurred and, if there was a weft error, to take the appropriate measures
for restarting the weaving machine. For example, a broken or not gripped weft thread
can be pulled out of the shed.
[0040] This method may not only be arranged such as to provide the information indicating
whether a weft error has occurred during a weaving process or not, but may further
arranged for providing information indicating which kind of a weft error has occurred.
This can be done by teaching to the artificial neuronal network to not only provide
an output indicating that a weft error has occurred or not, but to further provide
an output indicating which kind of a weft error has occurred. In the learning procedure,
the weft error information forwarded to the artificial neuronal network in association
with each learning set may comprise the information indicating that, in association
with a specific learning set, a weft error has occurred or not, and further indicating
which kind of a weft error has occurred.
[0041] While, in association with the above embodiment a weaving cycle is represented by
one complete revolution of a weaving machine main shaft and the transfer occurrence
parameter defining the position of the transfer of weft thread within such a weaving
cycle is referred to as being an angular position of the weaving machine main shaft
within the value range of 0° to 360°, such a weaving cycle can be defined in an other
manner. For example, the rotation or movement of an other component of a weaving machine
that, as is the case with the weaving machine main shaft, repeats with each weaving
cycle can be used as the basis for defining the weaving cycle and for defining the
position of the transfer of a weft thread within a value range associated with such
a repeated movement. A weaving cycle can also be determining in a time based manner.
If, for example, a weaving machine is operated with a speed of 200 weaving cycles
per minute, each weaving cycle takes 300 ms. The start of each weaving cycle may be
triggered by a start command of the weaving machine controller, and the transfer occurrence
parameter can be defined as being a particular point of time within the value range
of 0 ms corresponding to the start of the weaving cycle and 300 ms corresponding to
the end of the weaving cycle.
1. A method of operating a weaving machine (10) for inserting weft threads (14) into
sheds formed between warp threads during weaving cycles, the weaving machine (10)
comprising at least one weft thread inserting means (12) with an inserting rapier
(16) for inserting a weft thread (14) and a receiving rapier (18) for receiving the
weft thread (14) inserted by the inserting rapier (16), the weaving machine (10) further
comprising a weft thread tension sensor (30) for outputting a weft thread tension
signal indicative of a weft thread tension of a weft thread (14) inserted by the at
least one weft thread inserting means (12),
the method comprising the steps of:
A) providing a plurality of consecutive weft thread tension values (VT) within a weaving cycle on the basis of the weft thread tension signal,
B) providing a predetermined number of consecutive weft thread tension reference values
(VR) on the basis of the plurality of consecutive weft thread tension values (VT) in association with each weft thread tension measurement point (Pi, Py) of a plurality of consecutive weft thread tension measurement points (Pi, Py) within the weaving cycle, the weft thread tension reference values (VR) of the predetermined number of consecutive weft thread tension reference values
(VR) being indicative of the weft thread tension prior to the associated weft thread
tension measurement point (Pi, Py),
C) determining whether a weft error has occurred within the weaving cycle on the basis
of the plurality of consecutive weft thread tension reference values (VR) associated with at least one of the weft thread tension measurement points (Pi, Py) of the plurality of consecutive weft thread tension measurement points (Pi, Py) associated with the weaving cycle.
2. The method of claim 1, wherein each weaving cycle is associated with a predetermined
angular range of rotation of a weaving machine main shaft, and wherein each weft thread
tension measurement point (Pi, Py) corresponds to an angle of rotation of the weaving machine main shaft within the
predetermined angular range of rotation.
3. The method of claim 2, wherein the predetermined angular range of rotation is 360°,
and/or wherein the plurality of consecutive weft thread measurement tension points
(Pi, Py) associated with the weaving cycle are equidistantly distributed within the weaving
cycle.
4. The method of claims 1 to 3, wherein at least one of the weft thread tension reference
values (VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), preferably each weft thread tension reference value (VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), is based on one of the weft thread tension values (VT) of the plurality of consecutive weft thread tension values (VT) laying within a predetermined measurement zone (Zi, Zy) preceding the associated weft thread tension measurement point (Pi, Py).
5. The method of claim 4, wherein at least one of the weft thread tension reference values
(VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), preferably each weft thread tension reference value (VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), corresponds to one of the weft thread tension values (VT) of the plurality of consecutive weft thread tension values (VT) laying within the predetermined measurement zone (Zi, Zy).
6. The method of claim 5, wherein at least one of the weft thread tension reference values
(VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), preferably each weft thread tension reference value (VR) of the plurality of consecutive weft thread tension reference values (VR) associated with a weft thread tension measurement point (Pi, Py), corresponds to a mean value of a weft thread tension value (VT) of the plurality of consecutive weft thread tension values (VT) laying within the predetermined measurement zone (Zi, Zy) obtained in a plurality of consecutive weaving cycles, preferably including the
instantaneous weaving cycle.
7. The method of claims4 to 6, when referred back to one of claims 2 or 3, wherein the
predetermined measurement zone (Zi, Zy) corresponds to a measurement angular range of rotation of the weaving machine main
shaft preceding the associated weft thread tension measurement point.
8. The method of one of claims 1 to 7, further comprising an inputting step for inputting
the weft thread tension reference values (VR) associated with at least one of the weft thread tension measurement points (Pi, Py) of the plurality of consecutive weft thread tension measurement points (Pi, Py) within the weaving cycle into an artificial neuronal network as input values, wherein
the artificial neuronal network is arranged for providing an output indicating whether
a weft error has occurred or not.
9. The method of claim 8, wherein the weft thread tension reference values (VR) associated with a plurality of the weft thread tension measurement points (Pi, Py) of the plurality of consecutive weft thread tension measurement points (Pi, Py) within the weaving cycle, preferably the weft thread tension reference values (VR) associated with each one of the plurality of consecutive weft thread tension measurement
points (Pi, Py) within the weaving cycle, are forwarded to the artificial neuronal network for providing,
in association with each weft thread tension measurement point (Pi, Py) of the plurality of weft thread tension measurement points(Pi, Py), an output indicating whether a weft error has occurred or not.
10. The method of one of claims 8 or 9, further comprising a learning step for teaching
to the neuronal network to provide the output on the basis of the input values.
11. The method of claim 10, wherein the learning step comprises:
a) providing a plurality of learning sets of weft thread tension reference values
(VR), each learning set corresponding to the consecutive weft thread tension reference
values (VR) associated with one of the weft thread tension measurement points (Pi, Py) of a terminated weaving cycle,
b) in association with each learning set, providing a weft error information indicating
whether, in association with a respective learning set, a weft error has occurred
or not,
c) inputting one of the learning sets of the plurality of learning sets as an input
pattern into the artificial neuronal network,
d) comparing the output of the artificial neuronal network with the weft error information
associated with the input learning set,
e) adapting a configuration of the artificial neuronal network at least if the output
of the artificial neuronal network does not match the weft error information associated
with the input learning set.
12. The method of claim 11, wherein steps c) to e) are repeatedly carried out such that,
with an increasing number of repetitions, the percentage of repetitions providing
an output of the artificial neuronal network corresponding to the weft error information
associated with the input learning set increases.
13. The method of one of claims 10 to 12, wherein the learning step comprises a supervised
learning procedure.