Background of the Invention Noil is the waste fibre produced during combing of staple fibre and is composed of short fibre, fibre entanglements, and foreign material in the form of vegetable matter in the worsted (long-staple) system or trash in the cotton (short- staple) system. Noil is an important economic parameter in combing and any inadvertent increase due to avoidable causes represents a significant economic loss, given the throughput of modern plants. For example, a reduction of 0.5% in the noil for a wool combing plant with a capacity of 15 Mkg/year is worth about A$600,000 per year at season 2000 average top and nail prices.

The conventional procedure for measuring noil is to collect and weigh a sample of combed sliver and the waste fibre generated during its production. The percentage noil is then calculated from the known masses. Although this method is simple in principle, in practice it is complicated in the production environment because of the need to interrupt production to collect the noil. In many plants, noil is removed pneumatically and manual collection is possible only if this system is deactivated.

Because the current methods for measuring noil are slow and inconvenient, the measurements are made infrequently, which means that rational control of comb performance is difficult, if not impossible. A method for the continuous on- line measurement of noil would give combers new strategies for reducing noil. For example, it could be used to identify and correct those combs that are producing

higher noil than average or to determine the impact on noil of upstream processes, such as carding and gilling in the case of long staple processing.

The problem of determining noil in commercial plants is well recognised and was the motivation for a noil measuring system for a cotton comber described in US patent 5,404, 619 and its counterpart, European patent EP0571572. This patent describes a method in which the noil percentage is calculated from simultaneous measurements of the fibre mass flow into a comb and the noil mass flow generated during combing. The input flow is determined as the product of the linear density, fibre thickness, and the speed. The noil is collected directly and the principal method described in the patent employs a filter in the noil duct to collect a fraction of the noil and extrapolates from this to the complete sample by means of a calibration. The percentage noil is calculated from the two measurements.

A problem with this method is that flow of material in the noil duct is very small so that the measurement error is potentially large. The patent attempts to overcome this by allowing time for the noil to accumulate on the filter to increase the thickness of material tested. However, this technique is fundamentally inaccurate because of the potential for errors inherent in the methods of collection and measurement. Additionally, noil cannot be measured over short time intervals because the collection requires relatively long times. A number of alternative techniques are reported for measuring noil, for example, web thickness, optical density, and weighing.

Another problem that arises from the direct collection and measurement of noil, certainly for wool combing, is that the noil stream splits into two separate flows : a main stream collected from the circular comb and representing at most 90% of the waste, the'front'noil, and a complementary stream, comprising fly from around the working elements of the comb making up the balance, the'back' noil. The contribution of the back noil cannot be ignored, however, because measurements show that repartitioning of noil between the two streams occurs during normal operation. This means that using the front noil as a measure of total noil, as in effect is done in the aforedescribed conventional procedure, may lead to errors that are greater than the variations that are the object of control.

A further shortcoming of the principal method of US patent 5,404, 619 is that the mechanical changes required to fit the system to existing combs are substantial, so that retrofitting is likely to be difficult and expensive.

The patent also briefly describes another method for measuring noil based on measurements of mass flows into and out of a comb. The mass flows are determined by measuring the linear densities of the input and output fibre streams and multiplying by the speeds of the rollers. Linear density is measured by measuring the thickness of the fibre stream, while roller speed is not measured as such but is calculated from the comb speed and the feed length. Importantly, none of this information is measured and directly recorded; instead the data is manually updated from time to time. Consequently, a practical difficulty in a combing plant with a large number of machines, for example a wool combing plant, is keeping the recorded data up to date with actual values. A fully automatic system of noil measurement would overcome such problems, but automatic measurement of roller speeds would require shaft encoders to be fitted to the appropriate rollers, significantly increasing the complexity and cost of the system.

It is an object of this invention to provide for continuous, accurate, automatic monitoring of the total noil in combing without measuring the noil directly, and without measuring roller speeds, thereby significantly reducing the cost and complexity of implementation. Additionally, it is preferred for a noil measurement system to be explicitly suited for retrofitting to existing, installed combs.

It is also preferred to provide an improved means and apparatus for the measurement of combing noil that is adaptable to the continuous monitoring of noil and the provision of improved management of comb performance.

Summary of the Invention The invention stems from a realisation arising from the following series of equations, which are alternative expressions for calculating noil : Noil/% = 100 Mass of noil collected at time t in period At Mass of fibre used at time t in period At = 100 Input mass used at time t in At-Output mass collected at time t + T in At Mass of fibre used at time t in At ~ 100 01 Output mass collected at time t + T in At Mass of fibre used at time t in At

where t is the time at which noil is combed out of the fibre, r is the time required for fibre to pass onto the output rollers after combing, and At is some convenient period of time over which materials are collected.

There are several important features of these equations that have a significant bearing on the technology of noil measurement. Firstly, the last equation shows that total noil, not some fraction, can be measured by measuring <BR> <BR> the masses of material flowing into and out of a comb, i. e. , it is not necessary to measure noil directly. Secondly, a preferred or proper definition of noil includes an allowance for transit time of fibre through a comb. This improves the accuracy of the measurement particularly for averaging times (At) that are short compared with the transit time (i).

Finally, as the definition of noil is solely in terms of mass, not mass flow, it is only necessary to measure the mass of fibre in lengths determined by the speeds of the input and output rollers. This leads, in accordance with the invention, to an important simplification that the mass of the input and output samples can be represented by two or more spatial parameters from which noil can in turn be determined. In a particular embodiment, suitable spatial parameters include the number of timing pulses generated by the appropriate rollers. In practice, this means that shaft encoders can be replaced by digital switches, significantly reducing the cost of implementation.

Accordingly, in one aspect, the invention provides a method of monitoring noil production in textile combing, comprising continuously, or at intervals,

determining spatial parameter values that are measures of total fibre input and combed fibre output, from which spatial parameter values may be determined an indication of noil production.

Textile combing apparatus fitted with means for monitoring noil production during operation of the apparatus, comprising: first means to continuously, or at intervals, determine a first set of spatial parameter values that is a measure of total fibre input to the apparatus, and second means to continuously, or at intervals, determine a second set of combed fibre output from the apparatus from which spatial parameter values may be determined an indication of noil production.

The monitoring of noil production may include or consist of the monitoring of changes in the proportion of noil production.

Preferably, said determination of spatial parameter values includes measuring, e. g., simultaneously, two spatial parameters that combine to jointly reflect the mass flow into or out of the comb. One such parameter is the displacement of a machine component by the respective mass of travelling fibre, or the width of a gap between machine components determined by the respective mass of travelling fibres. The other parameter is, preferably, the distance that the respective fibre mass travels in a given period of time, for example, a machine cycle. This distance is, in effect, a gauge length. Displacement, or distance travelled in a selected time interval, is simpler and cheaper to measure than speed because it requires only that the cycle time of the comb be monitored, which can be done with a switch, so in practice distance can be determined simply by counting cycles of appropriate rollers. In contrast, the measurement of speed requires the determination of two parameters simultaneously, for example the magnitude of rotation of a shaft and the relevant period of time, which is usually done with a shaft encoder.

Advantageously, the component displaced by the fibre mass is under return load or bias in order to optimise the reliability and consistency of the measurement.

In a preferred embodiment, the spatial parameter that is a measure of the total fibre input to the combing apparatus is the separation of the feed rollers that present the slivers into the combing section of the machine. The advantage is that at this point the fibre mass is by design under reasonable pressure, which has the effect of reducing the variability of the measurement of the thickness of the fibre mass. As the upper feed roller is supported independently on each side of the comb, two displacement sensors are needed to obtain an accurate measurement of the variation in thickness of the total fibre mass. The signals from each displacement sensor are combined linearly to give a single reading. Only one sensor is required, of course, if the feed rollers are compact and are supported by a short shaft.

The spatial parameter that is a measure of the combed fibre output may be the separation of a pair of spring-loaded rollers through which the fibre mass, now referred to as combed silver, passes after combing. Preferably, these rollers are the normal high pressure rollers used for feeding the combed sliver into the crimper box, but any suitable pair of rollers through which the combed sliver passes can be used.

Suitable devices for measuring displacements of the spring loaded rollers in response to variations in linear density are proximity sensors, but other technologies can be used, such as load cells, strain gauges, and linear voltage differential transformers. Digital switches are used, preferably, to count the machine cycles that may be used to measure the respective gauge lengths.

The outputs of the sensors are typically digitised and the data collected by a computer for analysis.

Brief Description of the Drawings The invention will now be further described by way of example only, with reference to the accompanying drawings, in which: Figure 1 is a schematic isometric internal view of the left hand side of an input roller assembly of a comb, showing a preferred design and location for a proximity sensor to measure the linear density of the input sliver ; Figure 2 is an isometric view from an external aspect of the left hand side of the input roller assembly showing the location of a digital switch for monitoring the rotation of the input roller as it feeds sliver into the comb; Figure 3 is an isometric view from a rearwards perspective of the complete system for sensing the output of combed sliver from an archetypal commercial wool comb; Figure 4 is a side view of the assembly shown in Figure 3; Figure 5 shows the calibration of an archetypal commercial comb of mass throughput against the separation of the input and output rollers ; and Figure 6 shows a typical time record of a comb operating over a twelve- hour period.

Description of Preferred Embodiments Figures 1 to 4 are pairwise fragmentary drawings depicting the adaptations by which the desired spatial parameter values are measured in a textile comb, Figures 1 and 2 illustrate the modifications of an input roller assembly, while Figures 3 and 4 illustrate one of the rollers of an output pair.

Referring first to Figures 1 and 2, the lower roller 20, which is the drive roller, is rotatably mounted at each end to outstanding bracket assembly 22, itself fixed by its base 23 to the machine frame (not shown). The upper driven roller 25

has a projecting shaft 26 by which it is supported at each end in a pivoted lever arm 28 that allows the spacing of the rollers 20,25 to vary with the thickness of the passing sliver. Lever arm 28 is maintained under pressure so as to clamp the upper roller onto the drive roller in order to ensure a positive feed of the sliver to be combed.

The natural position of one end of the upper roller 25, i. e. , its separation at that end from roller 20, is determined by a proximity sensor 30 mounted to a support plate 31 carried in turn by bracket assembly 22. More specifically, the position of roller 25 is communicated to proximity sensor 30 by a displacement sensitive arrangement comprising a tie-rod 32 and a pivot arm 34, which are in turn pivotally hinged to pressure lever arm 28 and support plate 31 at pins 36,37 respectively.

As foreshadowed earlier, the shaft 26 of the adjustably positioned input roller is supported independently on each side of the comb frame and because of its relatively long length, proximity sensors are required at each end of the roller in order to obtain accurate measurements of the variation in thickness of the total fibre mass passing through the roller nip.

The input gauge length parameter is measured by a digital switch 40 mounted on the outside of bracket assembly 22. Input drive roller 20 carries, to the exterior of bracket 22, a multi-toothed activating plate or disc 42 that is keyed to the shaft 21 of roller 20 and activates digital switch 40 at intervals that are a measure of the number of rotations of roller 20 and therefore of the distance travelled by the passing sliver.

Turning to Figures 3 and 4, the measurement arrangement at the output roller pair will now be described. In this case, the upper roller (not shown) of the roller pair carries, outside the frame component 50 on which the output roller pair is mounted, a drive gear 52 keyed to roller shaft 53. The drive gear is carried on a bearing 51 by a pivot arm assembly 54 that is rotatably mounted in turn at 55 to a frame component 56. A proximity sensor 60 measures the exact rotational position of arm assembly 54 and thereby the spacing between the rollers that represents

the thickness of the combed sliver. As at the input rollers, there is a digital switch 70 activated by the teeth of gear 52 to measure the rotation of the gear as it drives the output rollers, and to thereby determine the output gauge length.

It will be appreciated that it is necessary to calibrate the roller separation to convert roller separation, which is measured to an accuracy of micrometres, to linear density of the fibre, which is usually measured in g/m. Calibration of the system is achieved in two steps that are referred to as a primary and a secondary calibration.

The purpose of the primary calibration is to convert the digital signal from the proximity sensors to roller separation. This is achieved by placing a series of calibrated shims, or feeler gauges, between the rollers, and recording both shim thickness and digital output. The results are combined in a linear plot for both the input and output roller pairs, and the gradients and offsets are recorded and stored by the software.

The secondary calibration requires feeding fibre through the comb for a fixed time and weighing the corresponding output and noil, mout and mnoil respectively. While the comb is running, the software integrates the signals from the sensors on the input and output roller pairs to give respectively ain = Xin. dnin and Bout = sxOut. dnout, where'n'is the number of roller rotations and'x'is the roller separation obtained from the primary calibration. The secondary calibration constant k is then calculated from the calibration masses for the input and output rollers as follows : kin = (mout+ mnOjl)/ajn and kout = mout/aout, where mout + mnoil is the mass of fibre feed to the comb, muon is the mass of noil, mout is the mass of combed sliver, and ain and Bout are the integrated sensor signals as described.

Combing noil is then obtained by scaling the signals from the proximity sensors by the primary and secondary calibration constants.

Figure 5 shows the result of a primary calibration of the noil measuring system on a commercial wool comb in which the right hand curve represents the mass flow into the comb and the left hand curve that for the mass flow out of the comb. The result shows clearly the linearity of the noil monitoring system as

indicated by the high correlation coefficients, i. e. , the extremely high R2 values of 0.997 and 0.999 for the output and input respectively.

Experience has shown that the accuracy of the measurement is particularly sensitive to roller geometry and that small changes tend to introduce errors. This is particularly true of the input rollers. In a commercial wool combing plant, the input roller is removed and replaced on a regular basis in order for the operator to thread new slivers into the comb. It has been found that, due to different spacings of the locating pins, simply reversing the position of the roller relative to that used for the calibration can introduce significant errors. This problem is eliminated for retrofit applications by providing the roller with a lock that permits it to be inserted into the comb in only one way. Clearly, more sophisticated measures can be taken for application to new machines.

After calibration of the sensors, the noil monitoring system calculates the noil according to the last of the noil-defining equations outlined above.

Examples of Operation By way of example, the invention was tested on a typical commercial comb by installing analogue inductive proximity sensors on the output roller and on the comb feed roller to monitor the thickness of the fibre. Digital switches were installed to monitor roller rotations as described. These adaptations were as illustrated in Figures 1 to 4.

The resultant plot of change in noil with time over a 12 hour period is shown in Figure 6 for a typical 23 lim diameter fibre. The plot shows data logs for the input, output, and the calculated noil (%). For different fibre types, it is necessary to use separate calibration curves, which for industrial use are applied by selecting data from prior calibration runs.

The data in Figure 6 reveals important performance details of combing as revealed by the noil monitoring system. Firstly, the record of the input shows a step decrease in the input to the comb. If this is scaled to the mean level of the

input, the ratio that results is about 1: 5. The input to a wool comb usually consists of about 20 separate'ends'or slivers of very similar linear density and therefore the record obtained was consistent with about four ends being dropped from the input to the comb. The time history shows that the input sensor system was correctly measuring the mass being fed into the comb.

Secondly, during the period that the input was reduced, there was a concomitant decrease in the linear density of the output sliver. The results for noil, however, showed no change as a result of the changed input. Although this is expected on theoretical grounds because there is no change in the wool fibre being combed, the fact that the calculated noil remained constant demonstrates that the noil monitoring system functioned correctly.

Thirdly, the record shows that over a sixteen-hour period from 0700 to 2300 on the selected day, there were a number of interruptions to combing from which the process efficiency can be determined.

These examples show how automatic measurement of noil provides a comber with important new information that will assist in reducing noil.

An important preferred refinement of the noil monitoring system is to allow for the transit time of the fibre through the comb. This is desirable because the input sensors typically sense variation in the feed at a time before that section of the feed reaches the output roller. Ignoring the transit time leads to greater variability in the calculated noil because the ratio of the signals does not correspond exactly with the noil being produced. Measurements show, however, that although the long-term average calculated without allowing for the transit time through the comb is close to the long-run value, difficulties arise in the short term if slivers are dropped from, or added to, the feed. The noil record shown in Figure 6 includes an allowance for transit time through the comb.

Some electronics is needed to convert the outputs of the proximity sensors and switches into data that is appropriate for manipulation by commercially available software. The electronics is preferably arranged so that one unit is used

to take the unprocessed inputs from several combs and to transmit the data to a central computer for processing and storage using well-known procedures.

Proprietary software may be provided for the central computer to arrange the data into convenient forms for the user of the noil monitoring system. For example, one way to use the data is to group together the display from a number of combs that are processing the same fibre so that the noil rates of the combs can be easily compared. An example of such a tabulated output is shown in Table 1. Data is automatically provided on input (kg/h), output (kg/h), calculated noil (%), machine speed (Hz), and comb efficiency (%), which is run-time expressed as a percentage of total time.

Table 1 Comb No Input Output Noil Comb Speed Comb # kg/h kg/h % mm'Efficiency % 1 33. 4 29.3 12.4 223 95.2 2 33.8 30.0 11.2 224 87.7 3 30.5 27.2 11.0 224 92.6 4 33.8 29.9 11.5 224 93.8 5 31.0 27.1 12.6 223 85.5 6 21.9 19.5 11.0 225 66.8 7 30.6 27.3 10.7 224 83.2 8 30. 2 26.3 12.9 224 83.9 Average 30. 7 27. 1 11. 7 223. 9 86. 1

The software may also provide facility for inspecting the time-history of the data from a suspect comb to better identify the cause or causes of high noil.

Figure 6 is an example of such a time-history. It will be appreciated that remedies available to correct combs in need of adjustment are well known to those skilled in the art. Such remedies may extend to settings of machines used in the upstream processes.

The noil monitoring system of the invention can optionally provide measures of a number of additional parameters that greatly improve the management of the combing process. For individual combs, this includes total waste fibre produced, total production, production rate, the average linear density and evenness of the input sliver, the average linear density of the combed sliver, the total running time, and the total, or process, efficiency. By application of proprietary software, the results from individual combs can be combined to give the equivalent data for a group of combs and the identification of group means, 95% confidence limits, and control graphs etc. , thereby considerably improving management control of combing.