JP2002519538A - Implementation of bubble method using fuzzy controller - Google Patents

Abstract

(57) Abstract By using a fuzzy controller, a non-woven fabric is made from cellulosic fibers, synthetic fibers or glass fibers utilizing a foam laid process. By using a fuzzy controller, the foam laid process can be precisely controlled to enable the production of non-woven fabrics of various different types of fibers, and this web ensures that the bubbles are always stable It has great uniformity and high predictability by allowing it to be handled as well as being substantially uniform. By using the fuzzy controller, at least the level at the wire pit, the level of the mixer / pulper, the pressure of the manifold to the molding machine, the bubble density, the outflow ratio (bubble speed divided by the wire speed), surface activity The agent supply, the overall design weight of the web to be manufactured, and the level of the binder tank can be controlled, especially when a glass web is manufactured. Special input parameters are used for each fuzzy controller. For example, fuzzy controls for automatically controlling the level in the mixer / pulper tank include the density and flow rate of bubbles circulating from the wire pit to this tank, the level of bubbles in the wire pit, and the At least some (two to all) of the fibers to be obtained as input parameters. Neural net control can be utilized to perform quality control of substantially all methods and equipment for making nonwoven fabrics, and experiments have been performed to facilitate neural net control to function. Laboratory test information is provided to the neural network control.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND AND SUMMARY OF THE INVENTION As disclosed in US Pat. Nos. 3,716,449 and 3,871,952, the disclosures of which are incorporated herein by reference. ,
A foam laid process (foam-la) for producing nonwovens from fibers.
There are many advantages to implementing an effective id process over a water-laid process. However, it has heretofore been difficult to commercialize the foam laid process for many different types of fibers. Although there are several facilities for producing nonwoven fabrics of polypropylene or glass fibers, there can be difficulties in controlling such methods and also in foam laid processes using cellulosic or synthetic fibers. There has been no effective commercialization (apart from the polypropylene equipment described above).

To date, all foam laid processes for producing nonwovens have been
It is controlled manually or using a PID controller. These methods can be operated by manual control, but long-term training, complete know-how of the process, and a large concentration of individual workers so that the operators can perform the required controlled operations in the correct order and manipulative volume. Power is needed. In a steady state operation when the process is undisturbed, manual control or PID control is deemed acceptable, as it usually passes product specifications set by the customer. However, some customers have set specifications to a higher level (perhaps due to the stringent demands of the end user), and as a result, inferior goods, that is, products that do not meet customer standards and must be discarded. Extreme increases in volume easily occur. In addition, any disturbances, such as starting the machine, changing the grade, etc., create other problems and require more competent operators to enable a quick and smooth grade change or start-up.

When comparing the steps performed by using the first test version of the present invention with those operated by a combination of manual control and PID control, the time required for starting is reduced by half and the time required for changing the grade is reduced. At least in half, in some special cases the time is reduced to almost zero, the amount of inferior goods is at least halved, the variability of controllable process variables is halved, and the variability of physical variables of the web It was quickly found that it was halved. Since this result was obtained from a "beta" version of the present invention, better understanding of the present invention, and fine-tuning the fuzzy control algorithms and equipment, can be expected to yield better results. .

[0004] In accordance with the present invention, virtually any fiber and filler can be effectively used to produce a wide variety of types and weights of nonwovens that can take advantage of the foam laid process. In this way, the form raid process can be effectively controlled. A key aspect of the present invention that allows for this effective control is the use of fuzzy controllers for the many different stages used in forming the web. Neural net control is also preferably utilized to take data from quality measurements (taken off-line) and process data to provide set points for long-term preparation and prediction. Multi-variable controls for measuring the cross-section of the web and for controlling dilution or dividing distribution pipes can also be used to provide set points for various fuzzy controllers. The fuzzy controller, neural net control, and multivariable control according to the present invention are provided by Honeywell
Any conventional off-the-shelf product available from 1-Alcont.

According to one aspect of the present invention, there is provided an apparatus for producing a nonwoven fabric from cellulosic fibers, synthetic fibers or glass fibers. The equipment includes the following components: cellulose / synthetic or glass fibers, water / air, circulating air bubbles, and a mixer / pulper tank, which mixes surfactants to produce a fiber-cell slurry, from the slurry. A fiber-bubble slurry from a molding machine, mixer / pulper tank for forming nonwovens at a web speed equal to the forming speed by withdrawing liquids and bubbles and collecting at least some of the withdrawn liquids and bubbles in wire pits For pumping the product to the forming machine, means for further processing the web produced in the forming machine to obtain the final nonwoven fabric, and for automatically controlling the density of bubbles in the mixer / pulper tank At least one fuzzy controller and slurry in the mixer / pulper tank A plurality of fuzzy controllers, including at least one fuzzy controller for automatically controlling the bell.

[0006] Fuzzy controllers for automatically controlling the level in the mixer / pulper tank include the density and flow rate of bubbles circulated from the wire pit to the mixer / pulper tank, the pH of the bubbles in the tank, the pH in the wire pit. And at least some (at least two, and preferably all) of the amount of fiber added to the tank as input parameters. A fuzzy controller is provided to at least control the level of wire pits, the manifold pressure to the molding machine, and the outflow ratio, and also to at least control the surfactant supply and the total design weight of the nonwoven fabric produced. Preferably. Binders may also be added in making nonwoven fabrics containing at least 10% glass or aramid fibers, and the binder is provided in a binder tank. Under these circumstances, the device further includes a fuzzy controller for controlling the binder level tank.

[0007] The molding machine typically includes a moving wire and a headbox. One of the fuzzy controllers preferably comprises a fuzzy controller for automatically controlling the air / bubble ratio to the molding machine, including the wire speed and the pressure in the headbox, wherein the fuzzy controller is Molded web design weight, headbox pressure,
It has as input parameters at least some of the level of bubbles in the wire pit, the density of circulating bubbles, and the amount or flow of bubbles removed from the headbox.

Means for further processing the foamed web include means for cleaning the web and removing liquid from the web during or in connection with the cleaning (typically any conventional scrubber And / or a suction device for treating the nonwoven fabric). In this case, one of the fuzzy controllers automatically controls the means of cleaning and liquid removal, the fuzzy controller controlling the web forming speed, web design weight, cleaning liquid temperature, suction bubble velocity, and cleaning means. It has at least some of the pressures as input parameters.

[0009] The means for further processing the formed web may comprise a conventional dryer, wherein one of the fuzzy controllers described above automatically controls the dryer, the fuzzy controller comprising: Enter at least some of the drying set points, web travel speed, energy input to the dryer, humidity levels in the dryer, and pressure differences above and below the web at various points along the dryer. Has as a parameter.

[0010] The apparatus may be used to control the forming of a web and / or to produce a nonwoven fabric.
The apparatus may further include a neural net control to cooperate at least partially with a fuzzy controller to perform substantially all quality control of the device.

According to another aspect of the present invention, (a) cellulose, synthetic fiber or glass fiber,
Mixing water, air, circulating air bubbles, and surfactant in a mixer / pulper tank to produce a fiber-air slurry, (b) pumping the fiber-air slurry to the molding machine, (c) molding Controlling the operation of the machine; (d) web speed equal to the forming speed by extracting liquid and air bubbles from the slurry in the forming machine and collecting at least some of the extracted liquid and air bubbles in wire pits. Forming a nonwoven fabric in a molding machine, (e) further processing a web produced in the molding machine to obtain a final nonwoven fabric, and (f) at least step (a).
A method for producing a nonwoven fabric from cellulose, synthetic fibers or glass fibers, comprising the steps of:

Step (a) may be performed in part by controlling the level of the slurry in the mixer / pulper tank, and step (f) involves removing air bubbles circulated from the wire pit to the mixer / pulper tank. Density and flow rate, pH of bubbles in tank
Automatically control the level in the mixer / pulper tank by using a fuzzy controller having as input parameters the level of air bubbles in the wire pit and at least some of the amount of fiber added to the tank May be partially implemented. Further, step (a) may be performed by automatically controlling the amount of surfactant added and by circulating some of the water removed from the web and separated from air during molding, then step (F) shows the flow rate of the surfactant, the manifold pressure on the molding machine, the level of bubbles in the wire pit,
By using a fuzzy controller having as input parameters the flow rate of the added fiber and at least some of the flow rate of the circulated water, a partial control is made to automatically control the amount of surfactant added. Will be implemented.

Step (c) may be performed at least in part to automatically control the air / bubble ratio to the molding machine, including the wire speed in the molding machine and the pressure in the headbox, Step (f) inputs at least some of the design weight of the formed web, the headbox pressure, the level of bubbles in the wire pits, the density of circulating bubbles, and the amount or flow of bubbles removed from the headbox. Partially implemented by using a fuzzy controller with parameters. Step (e) is carried out to wash the web and to remove liquid from the web during or in connection with the washing, then the web forming speed, the washer pressure, the web design weight, the washing liquid Step (f) is partially controlled to automatically control step (e) by using a fuzzy controller having at least some of the temperature, suction bubble velocity, and washer pressure as input parameters. Will be implemented.

[0014] The method further includes using neural net control to perform substantially all quality control of the method of making the nonwoven fabric.

According to another aspect of the present invention, (a) cellulose-synthetic or glass fiber, water, air, circulating air bubbles, and a surfactant are mixed in a mixer / pulper tank to produce fiber-air bubbles. Making the slurry, (b) pumping the fiber-bubble slurry to the molding machine, (c) controlling the operation of the molding machine, (d) extracting and extracting liquid and air bubbles from the slurry in the molding machine. Forming a nonwoven fabric in a forming machine at a web speed equal to the forming speed by collecting at least some of the applied liquid and bubbles in a wire pit, and (e) further manipulating the web produced in the forming machine. Obtaining a final nonwoven fabric, (f) performing at least one of steps (a) to (e) using a fuzzy controller, and (g) a method of manufacturing the nonwoven fabric. A method is provided for making a nonwoven fabric from cellulose, synthetic fibers or glass fibers, comprising using neural network control to perform substantially all quality controls.

Step (c) may be performed to dry the web, and the majority of the fibers added in step (a) may be glass fibers to which a binder has been added.
In this case, step (f) is partially performed to control web drying and binder addition.

Step (f) may be performed in part to precisely control the pH in the mixer / pulper tank by using a plurality of pH meters to detect the pH, and then the pH meter Step (f) is partially performed by using a fuzzy controller to control and link

According to yet another aspect of the invention, (a) a fiber-cell slurry comprising mixing cellulose, synthetic or glass fibers, water, air, circulating cells, and a surfactant in a mixer / pulper tank. (B) pumping the fiber-bubble slurry to the molding machine; (c) controlling the operation of the molding machine; and (d) extracting and extracting liquid and air bubbles from the slurry in the molding machine. Forming at least some of the liquid and air bubbles in the wire pit, forming a nonwoven fabric in the forming machine at a web speed equal to the forming speed, and (e) further manipulating the web produced in the forming machine to a final (F) wire pit level, mixer / pulper tank level, manifold pressure on the molding machine,
A method is provided for making a nonwoven fabric from cellulose, synthetic fibers or glass fibers, comprising using fuzzy control to at least control the density of cells and the outflow ratio.

Step (f) may further be performed to control the supply of surfactant and the overall design weight of the nonwoven fabric to be produced. To produce a nonwoven fabric comprising at least 10% glass or aramid fibers, a binder provided in a binder tank may also be added, and then step (f) is controlled to control the level in the binder tank. May be.

It is a primary object of the present invention to provide effective control of the foam laid process for producing fibrous nonwovens. This and other objects of the invention will become apparent from a review of the detailed description of the invention and from the appended claims.

DETAILED DESCRIPTION OF THE DRAWINGS An example of an apparatus for producing a mat or web of cellulosic and synthetic fibers in accordance with the cell process of the present invention is shown schematically at 10 in FIG. The device includes an inlet 12 for fibers, an inlet 13 for surfactants, and an inlet 14 for other additives such as pH adjusting chemicals such as calcium carbonate or acid, stabilizers, and the like. A mixing tank or pulper (fiber unraveling machine) 11 is included. The particular properties of the fibers, surfactants, and additives are not critical, and they will vary widely depending on the exact details of the product being manufactured, including its design weight. Surfactants are preferred to use surfactants that can be washed off fairly easily, since they will reduce the surface tension of the final web if it is still present, and in particular have the undesirable properties described below. Weyerhaeuser
This is the case for exclusive products.

The tank 11 itself is entirely conventional and is of the same type as the tank used as a pulper in conventional papermaking equipment using a water laid process.
The only difference is that the side walls of the mixer / pulper 11 extend to about three times their height in the water laid process, since the bubbles have a density of about one third of the density of water. The rpm of the conventional mechanical mixer in the tank 11 and the shape of the blades will vary depending on the particular properties of the product to be produced, but are not particularly critical,
Also, a wide variety of different components and variables may be used. The wall may also be provided with a breaker. There is a vortex at the bottom of the tank 11 from which air bubbles are discharged, but once the operation is started, the vortex cannot be seen once the operation is started since the tank 11 is filled with air bubbles and fibers.

To measure the pH at a number of different locations, the tank 11 preferably also includes a number of pH meters 15. Since pH affects surface tension, it is necessary to know this accurately. The pH meter 15 is calibrated daily.

At the first start-up, water is added along with the fibers from tube 12, the surfactant from tube 13, and other additives in tube 14. However, once the operation has begun, no additional water is required, and in the tank 11 not only the generation of bubbles but also the maintenance of the bubbles is made.

The bubbles flow out of the bottom of the tank 11 through the vortex and into the tube 16 under the action of the pump 17. Pump 17 is preferably a degassing centrifugal pump, like all other pumps in device 10. The air bubbles discharged from the pump 7 pass through the pipe 18 to another component.

FIG. 1 shows an optional operational holding or buffer tank 19 in dashed lines. A holding or buffer tank 19 is not required, but if some variation is introduced into the mixer 11, it is desirable to ensure a relatively uniform distribution of the fibers in the bubbles. That is, the holding tank 19 (small, typically on the order of 5 cubic meters) functions somewhat like a "surge tank" to average out fiber dispersion. Since the total time from the mixer 11 to the headbox in practicing the method of the present invention is typically only about 45 seconds, the use of the holding tank 19 will allow time for fluctuations to be smoothed out. Can be

When the holding tank 19 is used, air bubbles are supplied from the pump 17 through the tube 20 to the top of the tank 19 and through the tube 21 under the action of a pump, preferably a centrifugal pump 22, into the tank. It flows out from the bottom and then to the tube 18. That is, when the holding tank 19 is used, the pump 17 is not directly connected to the pipe 18 and the tank 1
9 only.

The tube 18 extends to a wire pit 23. The wire pit 23 is itself a conventional tank, again as in conventional water laid papermaking equipment, but with higher sidewalls. It is important to make the wire pits 23 so that there are no dead corners and therefore the tank 23 is not too large. Tube 18
A mixture of air bubbles and fibers in the pump 25 (a wire pit 23 so that it can be operated)
A conventional structure 24 that allows it to be introduced into (close to the bottom of) is further described in connection with FIG. In any case, the pump 25 is
Pump the bubble / fiber mixture in the tube 18 and additional bubbles from the wire pits 23 into the tube 26. Because a considerable amount of air bubbles are drawn from the wire pit 23 into the pump 25, the consistency in the tube 26 is significantly smaller than the consistency in the tube 18. The consistency in tube 18 is typically 2-5% solids (fiber), while the consistency in tube 26 is typically about 0.5-2.5%.

In the wire pits 23 there is no significant separation of bubbles into layers of different densities. There is a minimal increase towards the bottom, but the magnitude of the increase is usually small and does not affect the operation of the device.

Bubble / fiber enters tube 26 into a manifold 27 with an associated bubbling nozzle 28. A nozzle 28, which is a conventional bubble generating nozzle (violently disrupting bubbles) as used in patents 3,716,449 and 3,871,952, is attached to manifold 27; Many nozzles 2
Preferably, 8 is attached to the manifold 27. Each nozzle 28
From the conduit 29, which leads to the machine's headbox 30, through which one or more conventional papermaking wires (foraminas elements).
Nous elements are passed.

The headbox 30 has a plurality of (typically 3 to 5) suction boxes 31 for extracting air bubbles from the opposite side of the wire as viewed from the air / fiber mixture introduction, and a final separation box 32. Is at the discharge end of the formed web 33 from the headbox 30. In order to control the drainage, the number of suction boxes 31 provided in the suction table is increased for higher density products or for faster operation. The formed web 33, which typically has a consistency of solids of about 40-60% (e.g., about 50%), is subjected to a cleaning action as shown schematically by a cleaning step 34 in FIG. It is preferred that it is removed. The washing step 34 is to remove the surfactant. The high consistency of the web 33 means that the use of a minimum scale drying device is required.

The web 33 passes from a washer 34 through one or more optional coaters 35 and enters a conventional drying section 36. When a synthetic sheath / core fiber (such as Cellbond) is part of the web, in a conventional drying section 36, the temperature of the web is higher than the melting temperature of the sheath (typically polypropylene), Core (typically PE
In T), the dryer 34 is operated so as to heat the web to a temperature at which it does not melt. For example, if Cellbond fibers are used in web 33, the temperature in the dryer is typically about 130 ° C. or slightly higher, which is the melting temperature of the sheath fibers or slightly higher, Well below the melting temperature of the core fibers of about 250 ° C. In this way, the binding action is provided by the sheath, but the integrity of the product (provided by the core) is not compromised.

Although not always necessary, the method of the present invention contemplates pure air bubbles at or near the headbox 30 for many advantageous purposes. As can be seen in FIG. 1, the pump, which is a centrifugal pump 41 if desired, is
Aspirate air bubbles during zero. The air bubbles in the tube 40 are pumped to the header 42, which in turn distributes it to a number of different conduits 43 and heads for the headbox 30. Bubbles are introduced directly below the roof of the headbox 30, as shown by tube 44, and / or via tube 45 (or nozzle 28) via conduit 45 to introduce the bubble / fiber mixture into headbox 30. ) May be introduced. Details of the introduction of bubbles will be described with reference to FIGS.

The suction box 31 discharges bubbles extracted from the head box 30 in the pipe 46 to the wire pit 23. Pumps are typically not required or used for this purpose.

A considerable amount of bubbles in the wire pit 23 are circulated to the pulper 11. Air bubbles are withdrawn from tube 47 by a pump, preferably a centrifugal pump 48, and then
Within the conventional in-line density measuring means 49 and back to the tank 11 for introduction as shown schematically at 50. In addition to densifying the air bubbles in the tube 47 at 49, one or more density measuring instruments (such as a hydrometer) 49 as shown schematically in FIG.
A may be installed directly on the tank 11.

In addition to the circulation of bubbles, there is also typically a circulation of water. Last suction box 3
Air bubbles extracted from 2 enter via pipe 51 into a conventional separator 53 such as a cyclone separator. The separator 53 separates air and water from bubbles introduced into the separator 53, for example, by the action of a vortex, to obtain water containing only a very small amount of water. The separated water enters the water tank 55 from the bottom of the separator 53 in the pipe 54. The air separated by the separator 53 flows out of the top of the separator 53 into the tube 56 with the help of the blower 57 and is discharged to the atmosphere or used in a combustion process or otherwise treated.

As shown diagrammatically at 60 in FIG. 1, some of the liquid overflows into the sewage or treatment section, stabilizing the liquid level 58 in the water tank 55. Water is also withdrawn from below the level 58 of the tank 55 via the pipe 61 and under the action of a pump, preferably a centrifugal pump 62, through a conventional flow meter 63 (which controls the pump 62). Pumped by pipe 61. The circulated water is finally introduced at the top of the mixer 11 as shown diagrammatically at 64 in FIG.

Exemplary exemplary flow rates are 4000 liters of bubbles / fiber in tube 18 per minute, 40,000 liters of bubbles / fiber in tube 26 per minute, 3500 liters of bubble in tube 47 per minute and in tube 51 Bubbles are 500 liters per minute.

The device 10 also includes a number of new control elements. First fuzzy controller 7
1 controls the level of bubbles in the tank 11. The second fuzzy controller 72 is
Control the addition of surfactant within 3. A third fuzzy controller 73 controls the forming of the web in the area of the headbox 30. A fourth fuzzy controller 74 is used with the washer 34. A fifth fuzzy controller 75 controls the pH meter 15 and possibly controls the addition of other additives in the tube 14 to the mixer 11. Fuzzy control is also used to control surfactants and molding. A multivariable control system and a neural network control system (FIG.
8) is also preferably provided. Multivariable control is also used to control the outflow ratio in forming the web. The variables may be varied depending on the desired adjustment of the process and the effect of the variables on the end result.

A weight scale 76 is involved in fiber introduction 12 to accurately determine the amount of fiber added per unit time to control various components. A valve 77 and also a weigh scale 78 may be installed in the tube 13 to control the introduction of the surfactant. Tube 14
May also be provided with a valve 79.

A valve that is intentionally brought into contact with an air bubble at any point when handling the air bubble is shown in FIG.
Among the foam laid devices, device 10 is considered unique because, with the possible exception of the valves installed in tube 46, described in connection with, there is virtually no installation.

When performing the method entirely with the apparatus of FIG. 10, also the bubbles are maintained under relatively high shear conditions. It is preferred to keep the bubbles under higher shear, as the higher the shear, the lower the viscosity. The cell / fiber mixture behaves as a pseudoplastic and exhibits non-Newtonian behavior.

Using a foam laid process has a number of advantages over a water laid process, especially for highly absorbent products. Web 33
In addition to reducing the capacity of the dryer due to the greater consistency of the foam, the present foam process can be used with virtually any type of fiber as long as the specific gravity of the fiber or particle is about 0.15-13. Or particles (excessive “sinking” of dense particles
While less dense, the lower density particles "sink" to some extent-they do not all submerge in water), allowing them to be evenly dispersed in the slurry (and ultimately in the web). The aeration process also allows for the production of a wide variety of design weight webs, i.e., products with increased uniformity and bulk and greater uniformity than products of the water laid process. Multiple headboxes may be installed in series, or two in a headbox with double wires.
The layers can be made simultaneously and / or a simple coating machine 35 can be installed to apply the additional layers very simply (like coating).

Regarding the details of the components of the device of FIG. 1, except for what is entirely conventional,
This will be described with reference to FIGS.

FIG. 2 shows the introduction of the bubble / fiber mixture and bubbles into the pump 25 associated with the wire pit 23. Structure 24 is known from the Wiggins Teape method of the prior art, and allows the bubble / fiber mixture to be discharged directly into inlet 85 of pump 25 from open end 84 of the conduit, as shown by bent conduit 83. Then, the bubbles / fibres passing through the tube 18 are again guided. Bubbles from the wire pit 23 also flow into the inlet 85 as indicated by the arrow 86. The level of the wire pit 23 is controlled by operating the pump 48 under fuzzy control.

If the fibers used to produce the bubbles are particularly long, ie on the order of a few inches, instead of guiding the tube 18 to the inlet 85 of the pump 25 (as seen in FIG. 2), the tube 18 It reaches a pipe 26 downstream of the pump 25. In this case, the pipe 2 from the pump 25
The pump 17 must of course provide a higher pressure than would otherwise apply, so that the flow from 18 enters the tube 26 regardless of the pressure in 6.

FIG. 3 shows details of one form of the novel additional introduction mode of the Ahlstrom method. FIG. 3 shows the bubbles themselves from the tube 45 introduced into the bubble / fiber mixture in the conduit 29 immediately before the headbox 30. In other words, pure foam is added to the foam / fiber mixture coming from manifold 27 via nozzle 28. Bubble injection tube 4
If 5 is utilized, this tube does not require that air bubbles be injected into all of tube 29, but just enough tube 29 to achieve the desired result.

FIG. 4 shows an exemplary inclined wire forming machine and its headbox 301 utilizing two different bubble injection schemes (another one in addition to the one shown in FIG. 3). In the headbox 301 of FIG. 4, the inclined conventional forming wire 90 moves in the direction of the arrow and the bubble / fiber mixture is removed from the conduit 29 by bubbling at 45 as shown generally in FIG. Distributed within. Air bubbles are also introduced into the headbox 301 via conduit 44 such that the air bubbles flow generally as shown by arrow 92 in FIG. That is, the air bubbles flowing in the direction of the arrow 92 flow below the roof 93 of the head box 301. The headbox 30 is used to ensure that air bubbles initially flow in the direction 92 from each of the plurality of conduits 44.
1 has a baffle plate 94 installed therein.

The air bubbles introduced into the conduit 44 reduce the shearing of the fibers in the headbox 301, and the shearing between the fibers and the roof 93 of the headbox 301 causes the fibers to move in one direction, ie, the movement of the wire 90. This is to prevent heading in the direction. Under the basic principles of hydrodynamics, if the bubble / fiber mixture touches the roof 93, there will be unwanted disturbances in the fiber-oriented boundary layer. Bubbles introduced to flow in the direction of 92 eliminate this boundary layer problem. Bubbles introduced into the tube 44 and flowing in the direction 92 also keep the area under the roof 93 clean, which is again preferred.

The introduction of air bubbles into the conduit 45 (typically at an angle between about 30-90 °) as shown in both FIGS. 3 and 4 is for another purpose. FIG. 5 is a schematic top view of the headbox 30 (eg, 301) (only three conduits 29 are shown, but usually many are installed) showing the differences that pure bubble injection produces. Without injecting at 45 a bubble substantially free of fibers, the bubble / fiber mixture introduced by conduit 29 is distributed as generally indicated by the line 91 in FIGS. However, the presence of bubble injection at 45 changes the cross-sectional shape at the design weight due to the greater dispersion of the bubble / fiber mixture, as shown schematically by the line 96 in FIG. The effect on the cross-sectional shape at the design weight can be seen in the schematic representation of FIG. The normal cross-sectional shape (without bubble injection) at the design weight indicated by the 91A line has a large bulge 97. However, with the injection of air bubbles, 9
The bulge 98 is smaller, as indicated by the line 6a. That is, the design weight is more uniform. The control of the sectional shape is performed immediately before or immediately after the pipe 29 (4 in FIG. 4).
This is done by diluting bubbles in the main stream of the manifold (just before seen in FIG. 5).

If desired, the tube 29 can be moved from the bubble nozzle 28 into the headbox 301, 30 V (
Bubbles can be introduced into the explosion chamber. However, there is no reason to use the explosion chamber in the headbox to perform the Ahlstrom method. When used, the explosion chamber is for safety only.

FIG. 7 shows another form of headbox that can be utilized in the device 10. The entire forming machine and also the headbox 30V has the common attributes of the conventional water laid process dual forming wire vertical forming machine and headbox, and includes forming wires 90, 90A. In the exemplary embodiment shown in FIG. 7, a suction roller 100 is shown at the discharge end of the molding machine, and rollers 101, 101A are provided to guide the wires 90, 90A. In one aspect,
The wire 90A is also guided by the suction roller 100 as shown by the dotted line, but in a normal operation, the wire 90A moves on the upper roller 101 together with the discharged web 33. The suction table is cheaper than the suction roller and is preferred, but a suction roller may be used as shown at 100 in FIG.

The head box 30 V includes a bottom 102 and side walls 103 and 104.
Bubble / fiber spaces 105, 106 are defined by side walls 103, 104 and a central wall structure 110.
Separated by The same cell / fiber mixture may be introduced into the spaces 105, 106, but generally these are quite different mixtures, with two separate
In one layer. One bubble / fiber mixture is introduced from the manifold 27 through the nozzle 28, for example via a tube 29 through the bottom 102 of the headbox 30V as indicated by the inlet 107, while the other bubble / fiber mixture comes out of the manifold 27A, Through the nozzle 28A and the bottom 10 of the headbox 30V.
2 is introduced into the second inlet 107A. Alternatively or additionally, the bubble / fiber mixture flows in conduits 29 'and 29A' through inlets 108, 108A, respectively, in sidewalls 103, 104, respectively. In any case, the bubbles introduced /
The fiber mixture flows upward in the chambers 105, 106 and comes into contact with the wires 90, 90A. Suction is performed by the conventional suction boxes 31, 31A.

The wall structure 110 in the headbox 30 V is also shown in FIG. The wall structure 110 not only separates the spaces 105, 106, but also provides additional material for the wires 90,
It is used to introduce this material into the suspension without direct contact with 90A. This is because SAP (Super Absorbent Products)
Some materials are important because they contaminate the wires 90, 90A when they come into contact with them. By introducing using the wall structure 110,
The material to be introduced (such as SAP) is supplied just before the web is actually formed, so there is no opportunity for this material to contact the wires 90, 90A or otherwise interfere with processing.

With particular reference to FIG. 8, inside the structure 110 is a plurality of conduits,
An additive material, such as SAP, which is about 10-20% consistency of solids from 11 flows upward through these conduits until discharged through the triangularly enlarged end 114 of conduits 113. Tube 1 with flared end portion 114
Between the 13, there may be provided a plate 115 for holding the tube 113 in place. On the side facing the tube 113 is provided a plate 116 (FIGS. 7 and 8), which forms a passage for the gas / fiber mixture in the chambers 105,106. The SAP or other material comprises at least a first suction box 31, as shown at 117 in FIG.
At a point remote from 31A and within the substantial center of the cell / fiber mixture, there is little possibility that the material discharged at 117 will be in direct contact with the wires 90, 90A.

The conduit 113 is preferably circular in cross section, while the flared end 114 has flat sides and has a substantially square opening where the material 117 is discharged. . The flared end 114 extends over substantially all of the structure 116 as seen in FIG.

The product manufactured utilizing the headbox 30 V has two or more different layers that are provided together in a web 33. When material 117 is introduced, it is introduced so that it is essentially between the layers and extends partially into each layer.

FIG. 9 shows the construction of the suction box 31 of the molding machine or headbox 30 currently used (and over the years) in the production of glass fibers,
An example of a structure which is likewise used in the manufacture of the web 33 according to the trom method is shown in a cross-sectional view perpendicular to the machine direction. As can be seen in FIG.
0 extends beyond the top of the suction box 31 having side walls 118. In addition to the bubbles 120 drawn from the bubble / fiber mixture on the side facing the wire 90, openings or tubes 119 are provided in the side walls 118 to allow air to flow into the suction box 31 below the wire 90. Have been. Air flows freely through the tube 119 as a result of the suction force in the suction box 31 applied in a conventional manner. However, the pipe 119 is provided with a valve whose opening can be controlled automatically or at least manually. The bubbles then pass through conduit 46 through wire pit 2
Enter 3. However, since air is being introduced through conduit 119, the excess air being introduced is removed (but the air / liquid ratio of the bubbles does not change significantly from the original ratio in the bubble / fiber mixture). Is preferred. Therefore, the conduit 121
Is connected to the conduit 46, and air is exhausted through the conduit 121 by the blower 122.

FIG. 10 is a schematic representation of a vertical forming machine, including the headbox 30V of FIG. 7, showing a wide variety of rollers used to guide and / or drive the wires 90, 90A. , And also with other components, including a washing section 34 and a dryer 36. An important special feature of FIG. 10 is that a conduit 124 leading to a collector 125 is provided, while the collector is connected to the conduit 46.
The conduit 124 is connected to both the suction boxes 31 and 31A. FIG. 11 schematically illustrates the connection of the plurality of conduits 124 to the collector 125 and the connection of the collector 125 to the wire pit 23.

FIG. 11 shows one way in which the level 128 of bubbles in the collector 125 can be controlled. In each of the conduits 46 extending from the collector 125 toward the wire pit 23, a remotely operated (eg, solenoid) valve controlled by a controller 129 is installed. If the valve 127 is closed or partially closed, air bubbles can rise into the collector 125 as shown in FIG. This allows the level 128 in the collector 125 to be controlled. When the valve 127 is fully open, air bubbles freely flow through the conduit 46 into the wire pit 23 below the level of the air bubbles.

In all aspects of the device 10, it is preferred that no pump be installed to extract the air bubbles, the air bubbles simply flowing freely into the wire pit 23 under gravity.

FIG. 12 schematically shows a washing unit and a covering unit that may be installed in the apparatus 10. The cleaning liquid is introduced through a cleaning box 34 above the web 33 and
After the cleaning liquid has passed through the web so as to mainly remove the surfactant, a suction force is applied to the bottom portion 130 through the blower 131 to remove the cleaning liquid. The washing box 34 removes the binder in the Ahlstrom glass fiber manufacturing method (
It may be of any conventional construction as currently used for (using chemicals instead of water).

The method of the present invention allows additional layers to be easily applied on the web 33 without the need for an additional headbox. Other headboxes may be used for this purpose, but one or more coatings downstream of the washer 34 to apply different materials, as indicated by the layers 132 applied by the simple coater 35. Machine 35
It is more convenient to use. A simple coater 35 is a completely conventional device that deposits a layer 132 of any other material (which may include other fiber mixtures) of a desired thickness on the upper surface of the web 33. Downstream of the coating machine 35, ie after the layer 132 has been applied, a dewatering device 133 is provided which comes into contact with this layer 132 for dewatering.

A perforated belt or forming wire 13 guided by rollers 135
4 moves in the same direction as the web 33 passes through the suction box 136 in a conventional manner. Box 136 withdraws excess fluid from layer 132 while web 33 is supported at the bottom by conventional rollers 137, conveyor belts, and the like.
It is important that the suction box 136 be beyond the layer 132 as viewed from the web 33 to properly remove excess fluid. Belt 134 and roller 137 (or other belt) provide pressure, which helps to dewater layer 132.

After the dewatering section 133, a blower 139 that blows air from above through the layer 132 / web 33 is preferably used as part of the conventional dryer 36, the air being used to aid the movement of air from the blower 139. Conduit 1 which may be connected to a suction source
Outflow through 40. Dryer 36 conventionally has other attributes as well.

Any number of coaters 35, 35 ′ may be installed, and each coater 35, 35 ′ is accompanied by a dehydrator 133, depending on the particular layer to be coated on the web 33, or Many coating machines are installed before the dewatering section 133.

FIGS. 13 to 16 show fuzzy controllers 71 to 74 for precisely controlling the device 10.
, And FIG. 17 shows the relationship between fuzzy control and other controls. This precise control of the apparatus 10 is critical to the success of the method of the present invention when other methods fail to produce commercial webs of cellulosic and synthetic fibers and large-scale production of glass or aramid fibers. Is a key element that enables

As shown in FIG. 13, a fuzzy controller 71 controls the level of bubbles in the mixer / pulper 11. The input to the fuzzy controller 71 is measured by the bubble density (either from the in-line density meter 49 or the density meter 49A in the mixer / pulper 11, but not from both), and the pH meter 15. Tube 47 determined by the pH of the centrifugal pump 48 and the rpm (measured by conventional means)
Flow rate of the bubbles circulated in the level, the level 128 of the wire pit 23 and the pipe 12
From the fiber to the mixer / pulper 11, or other variables. Tube 12
The flow of fibers within is determined accurately by utilizing a weigh scale 76 which measures the amount of fibers per unit time put into the pulper 11.

FIG. 14 illustrates a second fuzzy controller 72 used to control the damping of valve 77 and / or weigh scale 78, or other mechanisms that control the addition of surfactant to pulper 11. Indicates the input of. The inputs to the fuzzy controller 72 are the surfactant flow rate as determined by the weigh scale 78, the pressure in the manifold 27 (typically 1-1.8 bar, depending on the product being manufactured), wire The level 128 of the bubbles in the pit 23, the pH determined by the pH meter 15, the flow rate of the fiber determined by the weigh scale 76, and the flow rate of the water circulated in the pipe 61 determined by the flow meter 63.

FIG. 15 controls the air / bubble ratio to form the web in the headbox 30 (like controlling the wire speed or pressure in the forming machine headbox).
Is the input to the third fuzzy controller 73 used for Inputs to the third fuzzy controller 73 include the pressure in the headbox 30, the level 128 of the bubbles in the wire pit 23, the volume of bubbles removed from the headbox by the suction box 31, and measured by the densitometer 49 or 49A. There is the bubble density to be applied, the design weight of the web 33 (after forming the web or after the dryer 36), and the degree of suction in each suction box 31 (31A). The pressure of the head box 30 is the pump 2
5 is controlled by controlling the rpm.

FIG. 16 shows inputs for a fourth fuzzy controller 74 that controls the washer 34, ie, controls the flow rate and suction of the cleaning solution. Inputs to the fourth fuzzy controller 74 include the design weight of the web 33, the speed of the suction blower 131, the pressure at the washer 34, the temperature of the cleaning solution, and the speed of forming the web (speed of the wires 90, 90A). .

Preferably, the foam / fiber mixture is maintained at a high degree of agitation / shear for a short transit time from pulper 11 to headbox 30 (about 45 seconds). Shearing is accomplished by controlling the level of air bubbles in the pulper 11 where air bubbles are agitated by conventional rotating blades, the pressure drop through the air bubble generation nozzle 28, the location (position) of the headbox 30, the vertical and inclined headboxes 30V and 30I. It is primarily controlled by controlling the vacuum on the notch for both and by controlling the primary drainage, such as by the speed of centrifugal pumps 17, 25 and 48. The valve 127 (
Apart from (if used) the device 11 is totally free of valves and no valves are intentionally contacted by bubbles. Similar to the pressure drop through the nozzle 28, the amperage and rpm of the circulation pump 48 are measured. The amperage of the circulating pump 25 changes, but for the same density, the distribution of bubble dimensions changes. In that case, it is necessary to vary the surfactant addition through the tube 13 (by adding more or decreasing the amount of surfactant) to return the foam size to the desired distribution. .

As seen in FIG. 17, the multivariable controller provides computer set points to all of the fuzzy controllers 72-75, and the neural network control 145 of FIG. 17 obtains data from quality measurements and process parameters 149. Performs long-term adjustments and forecasts, and also gives set points.

FIG. 17 illustrates a conventional neural network operably coupled to provide data and controls to conventional multivariable controls 146 and fuzzy logic controls 147, 148 and to receive them from these controls. The control 145 is shown schematically.
A quality parameter from a laboratory test at 149 (typically performed off-line, as for bubble stability) is an input to a neural network control 145 that can make long-term adjustments and predictions. One example of such a measurement is bubble stability, discussed below.

The gas bubbles must remain stable and substantially uniform throughout the process. Bubble stability is typically measured by a simple test performed off-line. Fill a 1 liter container, graduated along the sides, with bubbles to the top and scrape any bubbles that rise above the top surface of the container. The timer is started as soon as bubbles are put into the container. Measure the net weight of the bubbles in the container (in g),
Divide it by two. A timer is activated until enough water is drained from the bubbles to reach a level (in millimeters) corresponding to the weight of the bubbles divided by 2 (in millimeters). (To perform this test, all weights are assumed to be derived from water, ie, the weight of air is zero.) As an example, one liter of air bubbles would have a weight of 320 g. Dividing 320 by 2 gives 160. As soon as the level of water in the container reaches 160 ml, the timer is stopped. The bubble stability is optimal when half of the water takes about 7 minutes to drain. If the test time is outside the range of 4-10 minutes, bubble stability is unacceptable.

The foam laid process of the present invention is performed utilizing the parameters in Table 1 below. Many of these parameters, such as pH and manifold pressure, are product dependent, but the values given are unitary strats.
ifified Composite (USC) and Reticulated S
Initial values proposed for manufacturing two proprietary products, known as storage core (RSC). These proprietary products of Weyerhaeuser are a combination of synthetic and cellulosic fibers. Other parameters may be used to make the glass web.

[0077]

Table 1 Example of typical parameters of bubbles / process (better range of parameters could be wider if product range is wider) Parameter pH (virtually all equipment) approx. 5 Temperature about 20-40 ° C Manifold pressure 1-1.8 bar Consistency in mixer 2.5% Consistency in headbox 0.5-2.5% Consistency of SAP additive about 10-20% Molding Web consistency about 40-60% change in web design weight less than 1/2% foam density 250-450 g / l at 1 bar foam size of foam average diameter (Gaussian distribution). 3-. 5 mm air content of bubbles 25-75% (varies with process pressure) Viscosity No "target" viscosity, but bubbles are on the order of 2-5 centipoise under high shear conditions and 200-300k under low shear conditions Molding speed of webs typically having a viscosity of centipoise Initially about 200 m / min, target is 500 m / min Specific gravity of fiber or additive. Values in the range of 15-13 Surfactant concentration Depends on many factors such as water hardness, pH, fiber type and the like. Normally 0.1 to 0.3% with circulating water Forming wire tension 2 to 10 N / cm Example of flow rate ・ 4000 liter / min from mixer to wire pit ・ 40,000 liter / min from wire pit to head box ・Bubble circulation conduit 3500 l / min ・ Suction and extraction to water circulation 500 l / min

In a complete and complex device according to the invention (eg for the production of nonwovens of glass fibers) items controlled by fuzzy logic control (and / or multivariable control and / or neural net control) include: Includes:

The total design weight (fibre mass flow, fiber mass humidity, binder suction, binder flow, suction before binder supply, binder content, binder viscosity, binder pH, binder temperature, and machine speed At least some (eg 2
And preferably all have as input parameters).

The level of the binder tank (binder supply, binder formulation, binder dry content, suction before binder supply, wire speed, binder pH, and at least some of the binder air content (eg, 2) and preferably all have as input parameters). Binders can be supplied and controlled during the washing and chemical addition stages.

The level of the wire pit (23) (level control pump, suction in suction box (molding), operating pump (rpm), operating pump energy, manifold pressure, head box pressure, suction pipe flow rate, and bubble density) At least some (
E.g. two) and preferably all have as input parameters).

The level of the mixing tank (11) (manifold pressure, bubble density in the mixing tank, pH of the bubbles, feed returned from wire pits or short circuit, feed of surfactant, flow rate of returned water, Short circulation bubble density, agglomerate feed, buffer tank level, wire pit level, stirrer energy, and at least some (eg, two) and preferably all of bubble temperature as input parameters).

Manifold (27) pressure (pump (25) rpm, manifold outlet valve, molding machine suction pressure, bubble density, bubble stability, surfactant supply, mixing tank level, buffer tank level, wire pit level, bubble pH, lump feed, returned water flow rate, mixing tank density, wire shower pressure, wire water control suction, dry suction box suction, molding machine exit suction, molding machine overflow, and bubble temperature At least some (eg two) and preferably all as input parameters).

Bubble density (surfactant supply, all tank levels, temperature, pH, returned water flow, lump supply, wire water controlled suction, manifold overflow, line speed, and pump and blender At least some (eg, two) and preferably all of the energy as input parameters).

• Outflow ratios (manifold pressure, bubble density, headbox pressure, all suction of the molding machine, mass feed, wire pit level, wire speed, temperature, and at least some of the bubble stability (eg, 2 And preferably all have as input parameters).

• Supply of surfactant (cell density, cell temperature, fiber mass supply, and at least some (eg two) and preferably all of the cell stability as input parameters).

The fuzzy controller, neural network control, and multivariable control utilized in accordance with the present invention are all conventional off-the-shelf products such as those available from Honeywell-Alcount.

Multivariable controls typically measure the cross-sectional shape of the web and control dilution in or to separate distribution pipes and provide set points to various fuzzy controls. Neural net control captures data from quality measurement and process data and also provides set points for long-term adjustments and predictions. All variables may be changed depending on which of these has greater weight to make the appropriate adjustments and which are important in producing the final product.

FIG. 18 schematically illustrates various interrelationships between control elements according to the present invention using a neural network 145 (accepting laboratory values from 149 of FIG. 17).
The three different parts controlled by the neural net 145, namely FIG.
150, a web forming section, a binder device 151 (typically used only when the majority of the fibers making up the web are glass fibers or aramid fibers, etc.), and a drying device 152. . Neural net 145
There are three basic subsystems that are coupled to the system: optimization control 153, horizontal multivariable predictive control (HMPC) 154 (a conventional multivariable controller), and statistical process control (SPC) 155. . The control of the molding machine is shown schematically at 156,
The various inputs and self-controls associated therewith are shown diagrammatically under reference numeral 156 in FIG. The same applies to the binder 157 and the dryer 158.

Thus, a first level of control of the bubble process of the present invention includes a neural net model 145 that operates to control the quality of substantially the entire manufacturing process. Mode
Any of the first to third editions of the l-CC, the PROP (proportion) -algorithm model, the evolution algorithm (ENZO), or any combination thereof, is the core of the neural net model 145, the education algorithm, the prediction code, the simulation code and Can be used as optimization code. New versions of the above and also a whole new core of neural net models, teaching algorithms, prediction codes, simulation codes and optimization codes can be used.

The input of the neural network model is: design weight, glass weight, binder percentage, thickness, porosity, tear strength, strength, fiber orientation, high temperature tensile strength, oil porosity, opacity, wet tensile strength , An offline measurement (eg, 149 in FIG. 17) such as bubble stability, or a process quality measurement from an online determination.

The OUTPUT of the model 145 is a control value or a set value of a process parameter. These include, among others, fiber feed 145, manifold (2
7) pressure, pulper (11) level, buffer tank (19) level, circulating pressure (
For example, pumps 25, 48 and / or 62), and molding machine (31) suction.

In the test run, the combination of the Evolution Algorithm (ENZO) with PROP gave <1.4 g / m ^{2 with} a 95% confidence level for the design weight.

The machine or transverse cross-sectional shape of the final web 33 from the manufacturing process is based on an HMPC (Horizon Mul) based on online measurement.
variable Predictive Control) Controller 154
, Or a predictive multivariable controller
e controller). These controls can also be used to control any other part of the process where the control problem is too complicated for conventional control methods (PID controllers).

Preferably, the HPMC control 154 is used to control the cross-sectional shape of the web 33 in the machine direction. The controller for machine direction includes a model-based predictive multivariable algorithm. The HMPC controller 154 has multiple inputs / multiple outputs (multi
A control algorithm in the form of an i-input / multi-output (matrix) matrix.
It is used to predict the Ying state. The HMPC controller 154 also takes into account situations where the optimization functions of the actuators and adjustable variables are constrained.

In the multi-variable control device (HPMC 164), the variables to be controlled (design weight, speed,
Interference between process variables (actuator variables such as speed and pulp flow) and measurable variables such as humidity are kept at their set values. Table 2 shows a control matrix model for HMPC control in the machine direction.

[0097]

Table 2 Control matrix model for HMPC control in the machine direction Filler supply Binder suction speed Glass amount Binder amount Speed

The control 154 adjusts the outputs simultaneously and keeps the control variables at the desired values.
The HMPC controller 154 also takes into account disturbance variables. Such a disturbance variable is
It is taken into account when the machine (entire process) is started or when the grade is changed (for example, the control weight is changed). The disturbance variable is a measurable variable that affects the control variable, but is not controlled by the control 154. Disturbance variables can also be used for feedforward control. Control 154 predicts how the disturbance variable will affect the control variable. The prediction is then used to make any necessary modifications to the output of control 154.

One of the advantages of controller 154 is that it predicts the steady state of the process. These predictions inform the user in more detail about future situations. The final status of the control is also displayed to the user. The HMPC controller 154 can also predict that the control will be driven to the operating limit of the actuator and that the control strategy can be adjusted depending on the situation. Predictability allows weighted factors to be set for control and prioritization of functions (fun) to optimize machine conditions.
ction prioritization) is used. For example, if the percentage of binder needs to be increased to increase the thickness and the design weight is already at the upper limit of operation, the control can automatically reduce the setpoint of the design weight.

In one example of the bubble process of the present invention, a 3 × 3 matrix is used for control in the machine direction. However, depending on the number of variables, other types of matrices can be used, for example a 10 × 10 matrix (10 inputs and 10 outputs). Controllable variables are the amount of glass (or other fibers), the amount of binder (if used), and the speed. Actuator variables are fiber feed, binder suction, and speed. By using the control according to the invention for the production of glass webs, the variation in design weight, glass weight and binder percentage could be reduced to 50% in the machine direction in the test operation.

The purpose of the optimization control (153 in FIG. 18) is to minimize cost, maximize yield,
Or to eliminate manufacturing bottlenecks that are part of the process. Examples of optimization include material flow, chemical supply, energy consumption, manufacturing quality goals,
And to optimize the production capacity in each case in a correlated manner. The best possible way to operate a process has been established by optimizing the process according to both set goals and process constraints.

An SPC (Statistical Process Control Method) 155 may also be used in some cases.

[0103] Process control is implemented using fuzzy logic, neural nets, PID controllers, or a combination thereof. That is, the bubble process of the present invention utilizes fuzzy logic, a neural net, a PID controller, or a combination thereof to control the machine part 156, binder part 157, and dryer part 158 of the process. You. The molding machine portion 156 can be controlled by fuzzy logic, neural nets, PID controllers, or a combination thereof. The following items can be controlled in the molding machine part: bubble density in wire pit 23, bubble density in pulper 11, fiber orientation, level in pulper 11, level in wire pit 23, head box 27 pressure,
The height of the slice opening in the headbox 30 (the thickness of the web at the outlet of the headbox 30), the weight of the glass (weight flow of glass fibers or other fibers—12, 76 in FIG. 1). , PH, level in buffer tank 19, surfactant flow 13, cross-section in molding (cross-section in suction or discharge using 31 or 32), speed of wire 30, design weight of fiber feed (k
g / min-weight flow), pressure in the suction box 31, pressure in the manifold 27, overall suction of the molded part (flow, suction level, pressure in the suction box 31), suction at the trailing edge (after web forming) , Dry suction (after high pressure suction), change in suction pressure difference variable during molding (top-bottom dP), web 33 thickness, and web 33 porosity.

In further describing some of the parameters described above, it is assumed that the fiber orientation corresponds to the strength ratio, ie the ratio of the strength of the web 33 in the machine direction to the strength of the web 33 in the cross machine direction. be able to. This is controlled by the speed of the wire 30, the flow into the headbox 30, the pressure, the bubble density, and the suction profile from the suction box 31. Pump pressure, bubble discharge time, bubble density, p
H, and other factors, will also play a role. Finally, the outflow ratio, which is the bubble velocity compared to the wire velocity, is calculated. The speed of the wire is usually kept constant for any particular process. Control is performed at the positions of the suction boxes 31 and 32, respectively.

Suction pressure (tra) at the trailing edge of the web forming and trailing edge of the normal suction box 31
Illusion edge suction is a pressure above the suction box pressure. This sucks air with bubbles as shown at 32 in FIG. 1, and a separator 53 is utilized for the air.

The dry suction is typically after a liquid ring vacuum pump, such as a Nash pump, or other high pressure suction means. Typically this is drying (see 36 in FIG. 1)
Water removal performed immediately before.

The binder part 157 of the process includes fuzzy control, neural network control, P
It can be controlled by using ID control or a combination thereof. At least the following items are controlled: web 33 (binder molding suction after binder addition control at 157 with binder additive control), binder pH, design weight and binder molding suction, binder circulation tank level , Binder temperature, and suction speed.

The dryer portion 158 of the process can be controlled by using a fuzzy controller, a neural net, PID control, or a combination thereof. At least the following items can be controlled: the drying temperature at various points along the dryer, the speed of the web 33, the energy supplied to the dryer 36, the humidity in the dryer, and the pressure differential (to the dryer). (Above and below web 33 at various points along it).

An example of using fuzzy logic according to the present invention for bubble density control is shown in FIGS.
Seen at 0.

FIG. 19 shows a bubble density controller 160, which is graphically coupled to a fuzzy control 161.
And process 162. The bubble density set point is an input to controller 160, while other parameters are inputs to fuzzy controller 161 and process 162, which ultimately provide the measured density of the bubble. Subtracting the measured density from the set point is the deviation between them. This data is utilized as shown in the schematic control diagram of bubble density in FIG.

As shown in FIG. 20, various inputs 163 are provided to a fuzzification 164 that applies a rule base, and a rule base 165 provides a defuzzification output to control process metrology. The operations of 164-166 together provide a conventional fuzzy controller 167. Inputs provided at 163 include, for example, bubble density readings from 49, overflow of holding tank 19, pH of bubbles (eg, as measured by a pH meter), fiber velocity (at 12, 76 in FIG. 1). Such as the bubble temperature, the bubble density in the pulper 11, the difference between the bubble density set point and the measured density,
Returned water flow, bubble viscosity, fiber quality, bubble elimination time (determined by the test described above), water quality (such as pressure and hardness), wire cleaned (water removed from wire Suction after suction, surfactant chemistry (Z potential), which may depend on surfactant, pH of surfactant, specific fibers used, and especially water hardness, among others , And the feed rate of the surfactant (13, 77, 78 in FIG. 1).

The bubble density (static pressure + level) in the pulper 11 or the density of short-range circulation (4
9) is used as the actual value for the fuzzy controller (164-166). In this example, the density of the short circulation (from 49) is used as the actual value. The density at the pulper is used as a reference value, and its change in the deviation variable compared to the bubble density is used as one of the values at control input 163.

The input variables and disturbance variables of the fuzzy control can be improved by SPC control or neural network control or PID control. In this case, the process measurement value is set as a constant by using the PID control, or the input value is improved by using a neural network as the input.

In the first stage, the input to the fuzzy control is fuzzified (164). Fuzzification can be performed in three or five stages. During fuzzification, the numerical values of variables are converted into membership in a set, ie, dimensionless comparable values. In the example, the application is such that it is easy to transition from three-stage fuzzification to five-stage fuzzification or vice versa. FIG. 21 illustrates the principle for fuzzifying process measurements into membership of five fuzzy groups. Fuzzification 164 is a membership level function (memb) that indicates the membership of each fuzzy set as a function of the numerical value of the variable.
error level function).

FIG. 22 shows the process increment of the bubble density.
fuzzification of the values, for example, a membership level function for bubble density measurements.

VE is the deviation variable of the bubble density. That is, VE = SET-MET (set value-measured value in FIG. 19). The output of the membership level is as follows.

BPO is a big plus. POS is a plus. ZER is zero. NEG is negative. BNE is a big minus.

The tuning variables for adjustment are as follows. FBPO is a tuning variable for the BPO group. FPOS is a tuning variable for the POS group. FZER is a tuning variable for the ZER group. FNEG is a tuning variable for the NEG group. FBNE is a tuning variable for the BNE group.

In FIG. 22, the horizontal axis represents a deviation variable VE (g / l) of the bubble density (VE = SET-MES;
-30 g / l to +30 g / l) and the vertical axis represents the weighting factor (0 to 1.0) for the membership level function. At the zero point VE = 0 of the bubble density deviation variable on the horizontal axis, the FZER function represented by the two sloping lines starting from FZER = 1.0 is located. These lines have FNEG = 1 (
It crosses the horizontal axis at the point where VEPOS = 0 (left as viewed from VE = 0) and FPOS = 0 (right as viewed from VE = 0). Similarly, the function FPOS is represented by two sloped lines starting from FPOS = 1. The left line crosses the horizontal line (FPOS = 0) at FZER = 0 and the right line crosses at FBPO = 0. Further, the function FBPO is FBPO = 1
Are represented by two lines starting from. The left line is inclined and FPOS
= 1 and intersects the horizontal axis (FBPO = 0), and the line on the right is horizontal at the level of FBPO = 1. Similar statements apply to functions FNEG and FBNE. In other words, there are five different functions and their graphical representation in FIG. According to FIG. 22, a vertical line (dotted line) is drawn at the point where VE = 20 g / l on the horizontal axis, and the vertical line is horizontal through the point where it intersects the graphical representation of the functions FBPO and FPOS. The interpretation of the deviation variable for the bubble density VE = 20 g / l is defined so that the line (dotted line) is drawn to the vertical axis. The intersection of the horizontal line and the vertical axis indicates that the deviation variable of the cell density of VE = 20 g / l is a large plus at the level of BPO = 0.2 and POS =
Indicates that a level of 0.8 is interpreted as positive. The vertical dotted line is VE = 2
Since at 0 g / l only intersects the lines indicating the functions FBPO and FPOS, the value of the remaining levels of membership is zero.

In short, the block of FIG. 22 is a membership level function of a triangle, whose vertices and corner points are defined by tuning variables as shown in FIG. Thus, only two levels can simultaneously accept non-zero values (the sum of the two levels must still be 1). The tuning variable may be either a set variable or a variable that can be controlled by the user. Dynamic variables may be hidden inside the tuning variables to implement the adaptation function.

The function of the fuzzy logic is defined by creating a rule base (165 in FIG. 20), that is, the function logic. FIG. 23 schematically shows the rule-based operation principle. The period of the control parameters and other tuning parameters is determined from the step function response test and by analyzing the measured data. The functional speed of the control is defined by the period of the control, the filtering of measurements, and the gain of the control. Control gain is defined based on both fuzzification and defuzzification parameters. These factors are specified in the step function response test.

Each rule is applied with a weighting factor that is the same as the input membership level mentioned in the condition part of the rule. For example, if "VE (the deviation variable of the bubble density) is NEG" zero "and VDE (the change of the deviation variable of the bubble density is NEG)
) Is NEG “minus” and DPY (change in buffer tank overflow) POS “
If "plus", then the rule that O1 (control to dispersant supply) is ZER "zero""is based on using the measured membership value 0 as the weighted value of several values. Applied. All other rules are each applied by using their own weighted values. The combined effect of a plurality of rules is calculated by an algorithm called interference. The output has five values that vary between 0 and 1, FDN, D
N, ZER, UP, and FUP, which are the membership levels, rapid decrease (FDN), decrease (DN), zero (ZER), increase (U
P) and the control output value as a fast increase (FUP).

The concept of fuzzy logic is based on direct numerical va
The purpose of the present invention is to define the behavior of the control as a desired response in various situations, rather than as a function operated on in terms. In the following example, the input variables of the control are the deviation variable of the bubble density and its rate of change, and the rate of change of the buffer tank overflow.
The apostrophe associated with the variable indicates the value measured in the following control round.

Bubble Density Deviation Variable: VE = MES-SET Here, VE is the bubble density deviation variable (g / l), SET is the set value (g / l), and MES is the measured value (g / l). l).

Change of deviation variable of bubble density: DVE = VE−VE ′ where DVE is change of deviation variable of bubble density [((g / l) / hour)], and VE is deviation variable of bubble density [ ((G / l) / hr)] and VE 'is the deviation variable of the bubble density measured in the following control round [((g / l) / hr)].

Change in buffer tank overflow: DPY = PY−PY ′ where DPY is the change in buffer tank overflow [(l / s) / min], and PY is the buffer tank overflow [(l / S) / min)] and PY 'is the buffer tank overflow [(l / s) / min] measured in the following control round.

Table 3 shows some examples of fuzzy control rules. The table also shows the fuzzification of the change in buffer tank overflow, which is not used in the first stage of the rule logic.

[0128]

[Table 3]

FIG. 23 shows Table 3 used to assist design, and this rule table shows the rule-based operation principle. FIG. 23 Helps to better understand the functional logic of the control.
In FIG. 23, the X-axis represents time (time), and the Y-axis represents a deviation variable (g / l) of the bubble density increasing or decreasing from the target value. On the X-axis, the deviation variable for bubble density is 0 g / l, and when moving up, the deviation variable for bubble density is positive, ie, above the target value, and when moving down, the deviation variable is It is minus. For example, when following rule A and FIG. 23 (point A), the deviation variable (VE) of the bubble density
Is negative (NEG, less than the target value) and if the rate of change of the deviation variable of bubble density (DVE) and the rate of change of overflow of the buffer tank (DPY) are not taken into account at all, the dispensing circuit of the dispersant is controlled. It has a value, O1 reduction (DN), which means that the supply of dispersant is reduced, and thus the bubble density starts to increase towards the target value. According to the rule B ″ in Table 3 and the point B ″ in FIG. 23, the deviation variable (VE) of the bubble density is positive (POS, exceeding the target value), and the change rate of the overflow of the buffer tank (DVE) ) Is positive (POS, which means that the direction of change is towards heavier bubbles), and the rate of change of buffer tank overflow (
If DPY) is positive (POS, more fresh water constantly flowing into the device, causing an increase in bubble density), the dispersant feed dosing circuit will increase rapidly (FUP).
)), Meaning that the supply of dispersant should be increased considerably. The period of the control parameters and other tuning parameters can be determined from the step function response test and by analyzing the measured data. Filtering of the measurement data related to fuzzification is defined based on the measurement data. The functional speed of the control is defined by the period of the control, the filtering of the measurements and the gain of the control. The control gain is defined based on both the fuzzification parameter and the defuzzification parameter. These factors are specified in the step function response test.

Table 4 shows a slightly different form of the rule base of Table 3 and this table consists of several line examples of a total of 25 rules. The rule base in Table 4 does not include buffer tank overflow changes, which have been modified to the first stage. The first phase variables are the bubble density deviation variable (VE) and the bubble density deviation variable change (DVE). These conditions form the control value.

[0131]

Table 4 Some examples of the first stage rule base. 1. If (VE is BPO) and (DVE is BNE), then (O1 ZER) 2. If (VE is BPO) and (DVE is NEG) (O1 is UP) 3. If (VE is POS) and (DVE is BNE) (O1 is DN) If (VE is POS) and (DVE is NEG) (O1 is ZER)

Table 4 can be presented in a much simpler and easier to interpret form. A simplified form of the rule base of Table 4 is shown in Table 5.

[0133]

[Table 5]

Table 5 shows the deviation variable of the cell density (VE) in the horizontal column and the change (DVE) of the deviation variable of the cell density in the vertical column. The direction of the control (O1) is indicated by the intersection of both axes. For example, the measured value of the deviation variable (VE) of the bubble density is plus (PO
S), it is interpreted as a large plus (BPO) at the level of 0.8, and the measured value of the change rate (DVE) of the deviation variable of the bubble density is minus (NEG) at the level of 0.9,
If interpreted as a large negative (BNE) at the level of 0.1, the following output is given as control (O1) of the dispersant supply according to the rule base in Table 5. ZER = 0.2 UP = 0.8 Dn = 0.1 Zer = 0.1

When defining an output value, the output value will always be a smaller value. For example, when the change speed (DVE) of the bubble density deviation variable is NEG = 0.9 and the bubble density deviation variable (VE) is POS = 0.2, the control value is Zero = 0.2. is there.

Thereafter, the final output value of the rule base is defined for defuzzification.
The following output values are obtained based on the rule base. FUP = 0 UP = 0.8 ZER = 0.2 DN = 0.1 FDN = 0

In this example, the result of the logic is that Zer has two values, Zer = 0.2 and Ze
With r = 0.1, the larger of these will always be valid. That is, Ze
r = 0.2.

In addition, the rate of change of buffer tank overflow (DPY) will be considered. An example of a rule table modified into two stages is shown in Table 6.

[0139]

[Table 6]

According to rule 22, if O1 = FUP (bubble density control increases rapidly) and DPY = POS (change in overflow of holding tank is positive), then O2 = F
UP (bubble density control increases rapidly). In this case, the set of output values (O2) to be fuzzified is the control (O1) and the rate of change of the overflow of the buffer tank (DPY).
). Other rules can also be added to the basic rules of Tables 3-6, including the cell temperature, pH, fiber feed, or cell viscosity, if needed. "Condition" in Table 6 is replaced with "IF" or "OR" or "
And other conventional logical operations such as "AND / OR".

In the last stage of the fuzzy control, exact numerical values are assigned to the control that can be used as inputs to the actuator. This is performed according to defuzzification (166 in FIG. 20) shown in more detail in FIG.

The defuzzification algorithm may be an algorithm similar to that in FIG. There is a column of weights for each fuzzy group of output values. The height of the column is 1, and the position on the X axis can be adjusted. When calculating the instantaneous output value, the height of each weight column is determined according to each membership level, which is represented by the bold line below the column. The numerical output value is the projection of the compounded center of gravity of the determined weight column onto the X-axis. For example, in FIG.

Output variables: O1 or O2 is the numerical output value for the dispersant. Entering the membership level is the same as described above. That is, BPO is "big plus", POS is "plus", ZER is "zero", NEG is "minus", and BNE is "big minus".

FBPO is a tuning variable for the set “large plus”, FPSO is a tuning variable for the set “plus”, FZER is a tuning variable for the set “zero”, and FNEG is a tuning variable for the set “minus”. Yes, FBNE is a tuning variable for the set "big minus".

To adjust the control circuit, feed-forward control for adjusting the control output is used. The feed forward control is changed in increments or decrements by 0.1% or a desired change width. The fuzzy logic considers the direction and rate of change of the disturbance variables and adjusts the bubble density control using the fuzzy logic rule base. The control diagram of FIG. 20 shows a main operation of feedforward control of a disturbance variable in fuzzy control. In controlling bubble density, the following disturbance variables are used as feedforward control factors: bubble density, measured deviation variable, buffer tank overflow, bubble pH, fiber supply, bubble temperature, pulper density and bubble density. Changes in deviation variables during, return water flow, bubble viscosity, system activity, fiber quality, bubble half-life, water quality (pressure, hardness), wire aspiration, surface chemistry (Z
-Potential), and the amount of dispersant.

In the bubble process of the present invention, control of the molding machine 156 is preferably performed by using neural nets 170, 145 to control fiber orientation. The principle of the neural network controls 170, 145 is to achieve a more stable fiber orientation, which can be easily duplicated. Neural net 170,
145 is used to control the longitudinal cross-sectional shape (degree of vacuum) at the time of molding suction, and is used to shape the orientation of the fibers of the web 33. In the measurement, the upward suction of the molding machine is also used as a reference value. The extent of the longitudinal cross-sectional shape of the suction is adjusted by using the height of the trailing edge. Control of the degree of longitudinal cross-sectional shape of the suction is implemented as either a fuzzy controller 167 (FIG. 20) or a PID controller 171.

By the fuzzy control, it was possible to reduce the fluctuation of the bubble density by half as compared with the conventionally used PID control. Furthermore, the control of the process at the start and stop of the process is considerably improved. Today, the process can reach equilibrium one hour earlier than before.

Thus, in accordance with the present invention, control of the aeration process for producing a fibrous nonwoven can be achieved by using different types of fibers or fiber blends, if necessary or desirable, including fillers or binders. Controlling the foaming process to produce a fibrous nonwoven so that a wide variety of different types of nonwovens can be produced with optimal cell stability and with web uniformity and strength Has been found to be implemented in an advantageous manner. While the invention has been shown above and described in what is presently considered to be most practical and in its preferred embodiments, many modifications may be made to the invention within the scope of the invention and It will be apparent to one skilled in the art that the scope should be given the broadest interpretation in the appended claims so that the scope encompasses all equivalent methods and devices.

[Brief description of the drawings]

FIG. 1 is a general schematic of an example of an apparatus for performing the bubble process of the present invention.

FIG. 2 schematically shows in detail, partially in section and partially in elevation, the supply of bubbles / fibers from the mixer to the pump supplying the manifold and the headbox.

FIG. 3 is a perspective view, partially in section and partially in elevation, schematically showing that the bubbles themselves can be added into the conduit between the manifold and the headbox.

FIG. 4 is a side view diagrammatically showing details of an example of a tilted wire forming machine that can be used in a bubble process, partially in section and partially in elevation.

FIG. 5 is a schematic representation showing the effect of adding air bubbles to the conduit from the manifold to the headbox.

FIG. 6 is a schematic representation of a design weight profile of the headbox of FIGS. 4 and 5 with or without the addition of air bubbles.

7 is an end view diagrammatically showing an example of a vertical molding machine that may be used in a bubble process instead of the inclined molding machine of FIG. 4, partially in section and partially in elevation.

FIG. 8 is an end view with parts of the components cut away for clarity of illustration, and shows in section a conduit of the centrally located structure of FIG. 7 for introducing other materials; .

FIG. 9 is an end view diagrammatically showing a suction box used with the headbox / former of either FIG. 4 or 7 partially in cross section and partially in elevation;

FIG. 10 is a side view showing the molding machine of FIG. 7 in connection with other components of an apparatus for performing a bubble process.

FIG. 11 schematically shows an embodiment of the components of the apparatus of FIG. 10 together with a mechanism for returning air bubbles from the suction box to the wire pit.

FIG. 12 is a schematic side view showing an example of processing a web formed with the apparatus of FIG. 1 after its formation, including cleaning the web and depositing layers using a simple coating machine.

FIG. 13 shows diagrammatically various input and control functions in the apparatus of FIG.

FIG. 14 shows diagrammatically various input and control functions in the device of FIG.

FIG. 15 schematically shows various input and control functions in the apparatus of FIG. 1;

FIG. 16 schematically shows various input and control functions in the apparatus of FIG.

FIG. 17 schematically illustrates the interconnection between fuzzy logic control, neural net control, and multivariable control that can be utilized in accordance with the present invention.

FIG. 18 is a more detailed control scheme than the control of FIG. 17, showing various devices and parameters controlled in accordance with the present invention, and inputs to the control.

FIG. 19 illustrates the use of fuzzy control to determine the deviation between a desired density and a measured density of bubbles utilized in a foam-laid process according to the present invention.

FIG. 20 is another diagram illustrating bubble density control utilizing a fuzzy controller.

FIG. 21 schematically illustrates fuzzifying process measurements into membership in a set.

FIG. 22 graphically illustrates fuzzification of process measurements of bubble density.

FIG. 23 schematically shows the operation principle of “rule base” used in fuzzification control.

FIG. 24 is a diagram similar to FIG. 21 but for defuzzification only.

FIG. 25 is a schematic representation of a defuzzification algorithm.

[Procedural Amendment] Submission of translation of Article 34 Amendment of the Patent Cooperation Treaty

[Submission date] August 17, 2000 (2000.8.17)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Lochman, Kai Finland, Karfura, Pilcon Maenkatu 4 Ass, 15F Term (Reference) 4L055 AF04 AF09 AF13 AG99 AH29 CA10 CD30 DA29 DA30 DA32 DA40 FA22 GA39 [Continued Summary] Density It has as input parameters the flow rate, the level of air bubbles in the wire pit, and at least some (two to all) of the fibers placed in this tank. Neural net control can be used to perform quality control of virtually all methods and equipment for producing nonwoven fabrics, and to facilitate the functioning of neural net control. Next, laboratory test information is provided to the neural network control.

Claims (30)

[Claims] (A) cellulose, synthetic fiber or glass fiber, water, air,
The circulating foam and surfactant are mixed in a mixer / pulpper tank to produce a fiber-bubble slurry, (b) the fiber-bubble slurry is pumped to the molding machine, and (c) the operation of the molding machine (D) extracting the liquid and air bubbles from the slurry in the molding machine and collecting at least some of the extracted liquid and air bubbles in a wire pit, whereby the web speed equal to the molding speed Forming the nonwoven fabric in a forming machine at: (e) further manipulating the web produced in the forming machine to obtain a final nonwoven fabric; and (f) performing at least step (a) using fuzzy control A method for producing a nonwoven fabric from cellulose, synthetic fiber or glass fiber, comprising: 2. The step (a) is partially performed by controlling the level of slurry in the mixer / pulper tank and the density and flow rate of bubbles circulating from the wire pit to the mixer / pulper tank, the tank PH of bubbles in the
Automatically control the level in the mixer / pulper tank by using a fuzzy controller having as input parameters at least two of the level of air bubbles in the wire pit and the amount of fiber added to the tank The method of claim 1, wherein step (f) is performed partially. 3. The step (a) is further performed by automatically controlling the amount of surfactant added and circulating some of the water removed from the web and separated from air during molding. Also, use a fuzzy controller having at least two of surfactant flow rate, manifold pressure on the molding machine, level of bubbles in wire pits, flow rate of added fiber and flow rate of circulated water as input parameters. 3. The method of claim 2, wherein step (f) is thereby partially performed to automatically control the amount of surfactant added. 4. The molding machine includes a moving wire and a headbox, and automatically controls the air / bubble ratio to the molding machine, including the wire speed in the molding machine and the pressure in the headbox. Step (c) is at least partially performed, and the formed web's basis weight, headbox pressure, level of bubbles in the wire pits, density of circulating bubbles, and removal from the headbox 4. The method of claim 3, wherein step (f) is partially performed by using a fuzzy controller having at least two of the amount or flow rate of the generated bubbles as input parameters. 5. Step (e) is performed to clean the web and to remove liquid from the web during or in connection with the cleaning, and the web forming speed, cleaning machine pressure, web design. By using a fuzzy controller having as input parameters at least two of weight, temperature of the washing liquid, suction bubble velocity, and pressure of the washing machine, step (f) is automatically controlled to step (e). 5. The method of claim 4, which is partially performed. 6. The method of claim 5, wherein at least two of each of the input parameters comprises all of the input parameters. 7. The step (a) is further performed by automatically controlling the amount of surfactant added and circulating some of the water removed from the web and separated from air during molding. The flow rate of the surfactant, the manifold pressure on the molding machine, the level of air bubbles in the wire pits, the flow rate of the added fiber,
Step (f) is partially controlled so as to automatically control the amount of added surfactant by using a fuzzy controller having at least two of the flow rates of the circulated water as input parameters. 2. The method of claim 1, wherein the method is performed. 8. The molding machine includes a moving wire and a headbox, and automatically controls the air / bubble ratio to the molding machine, including wire speed in the molding machine and pressure in the headbox. Step (c) is at least partially performed and the formed web design weight, headbox pressure, level of bubbles in the wire pits, density of circulating bubbles, and amount of bubbles removed from the headbox or The method of claim 1, wherein step (f) is partially performed by using a fuzzy controller having at least two of the flow rates as input parameters. 9. Step (e) is performed to clean the web and to remove liquid from the web during or in connection with the cleaning, and the forming speed of the web,
Step (f) is partially performed by using a fuzzy controller having at least two of the following parameters as input parameters: washer pressure, web design weight, washing liquid temperature, suction bubble velocity, and washer pressure. The method of claim 1, wherein the method is performed. 10. Step (e) is performed to dry the web, and the majority of the fibers added in step (a) are glass fibers with a binder added, and a fuzzy controller is used. The method of claim 1, wherein step (f) is partially performed so as to control web drying and binder addition by use. 11. The method of claim 1, further comprising using neural net control to perform substantially all quality control of the method of making the nonwoven. 12. A fiber-cell slurry is prepared by mixing (a) cellulose, synthetic fiber or glass fiber, water, air, circulating air bubbles, and a surfactant in a mixer / pulper tank; Pumping the foam slurry to the molding machine; (c) controlling the operation of the molding machine; (d) extracting liquid and air bubbles from the slurry in the molding machine and wire-pitting at least some of the extracted liquid and air bubbles. (E) further processing the web produced in the molding machine to obtain the final nonwoven fabric, (f) fuzzy control Performing at least one of steps (a) to (e) using
And (g) a method of producing a nonwoven from cellulose, synthetic fibers or glass fibers, comprising using neural net control to perform substantially all quality controls of the method of producing the nonwoven. 13. Step (e) is performed to dry the web, and the majority of the fibers added in step (a) are glass fibers with added binder and use fuzzy control. 13. The method of claim 12, wherein step (f) is partially performed to thereby control web drying and binder addition. 14. The molding machine includes a moving wire and a headbox, and automatically controls the air / bubble ratio to the molding machine, including the wire speed in the molding machine and the pressure in the headbox. Step (c) is at least partially performed and the formed web design weight, headbox pressure, level of bubbles in the wire pits, density of circulating bubbles, and amount of bubbles removed from the headbox or 13. The method of claim 12, wherein step (f) is partially performed by using a fuzzy controller having at least two of the flow rates as input parameters. 15. By controlling the slurry in the mixer / pulper tank, step (a) is partially performed and the mixing of the mixer / pulper from the wire pit is performed.
At least two of the density and flow rate of the bubbles circulated through the pulper tank, the density of the bubbles in the mixer / pulper tank, the pH of the bubbles in the tank, the level of bubbles in the wire pit, and the amount of fiber added to the tank. 13. The method of claim 12, wherein step (f) is partially performed to automatically control the level in the mixer / pulper tank by using a fuzzy controller having as an input parameter. 16. The method of claim 15, wherein the at least two input parameters comprise all of the input parameters. 17. Step (a) is performed in part to precisely control the pH in the mixer / pulper tank by using a plurality of pH meters to sense the pH, and controlling the pH meter 13. The method of claim 12, wherein step (f) is partially performed by using a fuzzy controller to perform and communicate. 18. A fiber-cell slurry is prepared by mixing (a) cellulose, synthetic or glass fibers, water, air, circulating cells, and a surfactant in a mixer / pulper tank; Pumping the foam slurry to the molding machine; (c) controlling the operation of the molding machine; (d) extracting liquid and air bubbles from the slurry in the molding machine and wire-pitting at least some of the extracted liquid and air bubbles. Forming the nonwoven in a forming machine at a web speed equal to the forming speed by collecting in a (wire pit); (e) further manipulating the web produced in the forming machine to obtain a final nonwoven; (F) at least wire pit level, mixer / pulper tank level, manifold pressure on the molding machine, bubble density, and outflow ratio A method for producing a nonwoven fabric from cellulose, synthetic fibers or glass fibers, comprising using fuzzy control to control. 19. The method according to claim 18, wherein step (f) is further performed to control the surfactant supply and the overall design weight of the nonwoven fabric to be produced. 20. To produce a nonwoven fabric comprising at least 10% glass fibers or aramid fibers, a binder provided in a binder tank is also added, and step (f) also reduces the level of the binder tank. 20. The method of claim 19, which is implemented to control. 21. A mixer / pulper tank for mixing cellulose, synthetic or glass fibers, water, air, circulating air bubbles, and surfactants, removing liquids and air bubbles from the slurry, and removing the liquid and air bubbles from the slurry. A forming machine for forming the nonwoven at a web speed equal to the forming speed by collecting at least some in the wire pit, a pump for feeding the fiber-cell slurry from the mixer / pulper tank to the forming machine, manufactured in the forming machine A plurality of fuzzy controllers, including at least one fuzzy controller for automatically controlling the level of slurry in the mixer / pulper tank, to further manipulate the web to obtain a final nonwoven fabric. Manufacture nonwovens from included cellulose, synthetic fibers or glass fibers Because of the device. 22. A fuzzy controller for automatically controlling the level in the mixer / pulper tank, the density and flow rate of bubbles circulated from the wire pit to the mixer / pulper tank, the pH of bubbles in the tank, the wire 22. Apparatus according to claim 21 having at least two as input parameters: the level of air bubbles in the pit and the amount of fiber added to the tank. 23. The apparatus of claim 21, wherein a fuzzy controller is provided to at least control the level of the wire pit, the manifold pressure on the molding machine, the density of the bubbles, and the outflow ratio. 24. The apparatus of claim 23, wherein a fuzzy controller is also provided to control the surfactant supply and the overall design weight of the nonwoven fabric to be produced. 25. A fuzzy controller for producing a nonwoven fabric comprising at least 10% glass fiber or aramid fiber, wherein a binder provided in a binder tank is also added and a level of the binder tank is controlled. 22. The device of claim 21, further comprising: 26. A molding machine including a moving wire and a headbox,
One of the fuzzy controllers also comprises a fuzzy controller for automatically controlling the air / bubble ratio to the molding machine, including the wire speed in the molding machine, and the pressure in the headbox. The apparatus has as input parameters at least two of the design weight of the formed web, the headbox pressure, the level of bubbles in the wire pits, the density of the circulated bubbles, and the amount or flow of bubbles removed from the headbox. 22. The device of claim 21, comprising: 27. The means for further processing the formed web comprises means for cleaning the web and removing liquid from the web during or in connection with cleaning, and the cleaning and liquid removal. Is automatically controlled by one of the fuzzy controllers, which controls at least two of the web forming speed, web design weight, cleaning liquid temperature, suction bubble velocity, and cleaning means pressure. 22. The device according to claim 21 having as input parameters. 28. The means for further processing the formed web comprises a dryer for drying the web, wherein the dryer is automatically controlled by one of the fuzzy controllers, the fuzzy controller comprising: The dryer has at least two of a set point for drying, the speed of web travel, the energy input to the dryer, the level of humidity in the dryer, and the pressure difference above and below the web at various points along the dryer. 22. The device according to claim 21, having one as an input parameter. 29. The method of claim 21, further comprising a neural net control to cooperate at least partially with a fuzzy controller for controlling web shaping.
An apparatus according to claim 1. 30. The apparatus of claim 21, further comprising a neural net control to perform quality control of substantially all of the devices for producing the nonwoven.