DE102016102477B4 - Method for producing a fiber composite component - Google Patents

Abstract

Method for producing a fiber composite component by means of an injection method, in which a matrix material (17) is injected into a fiber material (13) and the injected matrix material (17) is cured, with a) providing a plurality of simulated injection paths (20) for respectively different process parameters of the injection process in a data memory (19), wherein each simulated injection course (20) simulated sensor values of at least one, the injection process characterizing measurement parameters are assigned, which were determined in simulation of the injection process with those of the respective simulation of the injection process underlying process parameters, wherein during the Injecting the matrix material (17) into the fiber material (13) while monitoring the injection process, real sensor values of at least one measurement parameter characterizing the injection process, which corresponds to one of the measurement parameters of the simulated sensor values, mi c) to monitor the real injection process, that simulated injection course (20) is determined by an evaluation unit (18) whose simulated sensor values correspond to the detected real sensor values or correlate with the detected real sensor values.

Description

The invention relates to a method for producing a fiber composite component by means of an injection or infusion method, in which a matrix material is introduced into a fiber material and the introduced matrix material is cured.

In the production of molded components made of fiber-reinforced plastics, short fiber composite components, a fiber material, for example in the form of a semi-finished fiber, is usually introduced into a mold, which at least partially has the later component form. Subsequently, the fiber material introduced into the mold is infiltrated with a matrix material, for example a resin, in order to impregnate the entire fiber material with the matrix material and to obtain the fiber-reinforced plastic component after curing or polymerization of the matrix material.

An essential quality feature in such an injection or infusion process is the complete impregnation or impregnation of the fiber material with the injected or infused matrix material. Because remain after the injection process still unoccupied sites within the fiber material whose dimensions or volume are greater than a maximum limit, the component must be taken out of production incorrectly. This is because unimpregnated residues represent a weak point in the later fiber composite component. In the further course, the injection process is understood to mean the course of an injection or infusion process.

It is therefore desirable to monitor the state of injection during the injection process so as to be able to determine, for example, whether the fiber material has been completely saturated or whether dry areas have possibly arisen in the fiber material. The earlier there is corresponding information, the earlier it may possibly be possible to intervene accordingly by adjusting the process parameters of the injection process, such as, for example, the injection pressure. It is also conceivable to take the component early out of the process when determining unimpacted points, whereby the effectiveness can be increased, especially in mass production.

It is therefore an object of the present invention to specify an improved method for producing a fiber composite component by means of an injection method in that during the injection of the matrix material into the fiber material, the injection process can be monitored and if necessary predicted.

From the DE 10 2011 112 141 A1 a method and an apparatus for producing a fiber composite component is known, wherein during the introduction of the matrix material into the cavity in discrete zones of the cavity, the temperature is raised compared to the average temperature of the cavity. In this case, the choice of the temperature profiles over time is preferably determined in advance by means of a simulation.

From the DE 10 2010 037 849 A1 discloses a method and an apparatus for the production of fiber composite components in the infusion process, in which process parameters are recorded by means of ultrasound signals and a transit time measurement. The ultrasound signals are compared with pattern signals.

From the EP 2 913 180 A1 A system and method for monitoring the injection process is known, in which the course of injection is recorded with a camera and then compared with an idealized course of injection.

From the DE 10 2013 100 092 A1 discloses a method and an apparatus for impregnation test, in which the degree of impregnation of the component is detected by means of ultrasonic signals.

The object is achieved by the method according to claim 1 according to the invention.

According to claim 1, a method for producing a fiber composite component by means of an injection or infusion method is claimed, in which a matrix material is introduced into a fiber material and the introduced matrix material is then cured, for example under temperature control and pressurization. This accordingly includes all the process steps required for producing a fiber composite component by means of an injection or infusion method in which the matrix material is introduced into the fiber material, even if these are not specified in detail, in particular the introduction of the fiber material into a tool and, if appropriate, the insertion of the tool with the fiber material and matrix material into an autoclave.

First, according to the invention, a plurality of simulated injection profiles of a simulated injection process are provided, wherein the simulated injection profiles were simulated for respectively different values of process parameters of the injection process. The simulated injection curves result from the fact that at least one value of a process parameter is varied in each case, so that for a process parameter with each varying parameter values in each case at least one simulated injection course is present. Simulated sensor values of at least one measuring parameter characterizing the injection process, which are derived from the simulation of the injection process with the process parameters on which the respective simulation of the injection process is based, are assigned in each simulated injection course. Such sensor values of measurement parameters are values of measurement parameters which can also be determined by corresponding real sensors in a real injection process, such that for each simulated injection course, the values of the process parameters underlying this simulated injection course vary over the multiplicity of simulated injection progressions and the simulated sensor values simulated in each simulated injection run, all provided together in a data store.

Thus, the respective simulated injection profiles including their simulated sensor values are available for one or more varying process parameters, so that a parameter space can be set up in which at least one simulated injection profile including its simulated sensor values can be determined for specific values of the process parameters. In the broadest sense, a simulated injection process is defined by its sensor values or sensor characteristics that are derived from the simulation.

According to the invention, the injection method for injecting / infusing the matrix material into the fiber material is now carried out, the injection process being monitored during the injection / infusion of the matrix material into the fiber material. This is achieved by real sensor values of one or more measurement parameters being determined during the injecting / infusing of the matrix material into the fiber material with the aid of real sensors, wherein the measurement parameters correspond to at least one of the measurement parameters of the simulated sensor values. As a result, a concrete, real sensor value is now available for the previously simulated measurement parameter, so that a bridge between the sensor system and the simulation can be struck thereby.

Subsequently, to monitor the real injection process, that simulated injection course is determined by an evaluation unit whose simulated sensor values correspond to the detected real sensor values or correlate with the detected real sensor values.

From the detected, real sensor values, it is now possible by comparison with the data in the data memory to determine at least one simulated injection course whose simulated sensor values correspond to the sensor values actually acquired during the injection procedure or at least correlate with these, so that at least with a very high probability thereof It can be assumed that the simulated injection course corresponds to the actual injection course within the tool.

This makes it possible for the first time to obtain more accurate knowledge of the actual state of the injection process, as well as knowledge about the flow front and possible problems during the injection of the matrix material, without the need to install an enormously complex and complex sensor within the tool. Rather, it was recognized that with a few sensors can be determined with very high probability, which injection state in the current injection process is present, which otherwise would not be possible due to the completed overall system.

This makes it possible to monitor almost completely the entire injection process, which also makes it possible to detect at an early stage whether the injection process actually present will possibly lead to a defective component. For if the determined simulated injection course, whose simulated sensor values correlate with or correspond to the real sensor values, leads to a component having dry spots or is not completely soaked at the end of the injection process, then it can in all probability also be assumed that the real injection process being performed in the present case leads to the same result. The injection process can thus be influenced at an early stage or even canceled, which promises greater efficiency, especially in mass production.

An injection course in the sense of the present invention is understood in particular to be a discrete or continuous indication of the flow front of the injected matrix material, this flow front generally being wetted and infiltrated accordingly due to a pressure gradient between the gate on the tool and the non-saturated area.

For the purposes of the present invention, an injection state, which can advantageously be derived from the course of the injection, is understood to mean, in particular, the injection progress, filling state or drenching state, and thus transmits information about soaked or infiltrated areas and unimpregnated or non-infiltrated areas of the fiber material. Under an injection condition can thus in particular the position of the current flow front in temporal relation to the current time in the injection process are understood.

In an advantageous embodiment, the simulated injection courses are provided by a machine learning system, for example a neural network, in which the parameters of the simulated injection courses together with their respective simulated sensor values form the training data with which the machine learning system assigns sensor values as input and simulation parameters of the injection course is trained as an output with respect to the input. The simulated injection profiles are thus provided together with their sensor values in the form of a machine learning system.

The real sensor values of the real injection process are thus input as input to the machine learning system, and due to the plurality of training data resulting from the varying simulations of the simulated injection processes, the machine learning system based on the real sensor values then those or those parameters for a simulated Injection history is determined whose simulated sensor values correlate with the real sensor values as input with respect to the machine learning system. In particular, values between individual sensor values, which for example do not originate from the simulation, can be interpolated or extrapolated and assigned to corresponding simulated injection progressions.

In a further advantageous embodiment, the one measurement parameter which is used both in the simulation of the injection courses on the one hand and in the real injection process on the other hand, the time since the beginning of the introduction of the matrix material until the matrix material reaches a predetermined sensor position. If several of these sensors are installed in the tool, the position of the flow front at a certain point in time can be derived with a very high degree of accuracy based on the time until the matrix material reaches the corresponding sensor position. In addition, this measurement parameter can also be determined relatively easily in the simulation so that calculation complexity can be reduced when determining the simulated sensor values.

In addition, one of the measurement parameters may also be the transit time of an ultrasound signal which has been emitted into the component by means of an ultrasound transmitter and received by means of an ultrasound receiver. It is possible to derive conclusions about the filling state and the state of polymerization from the transit time of an ultrasonic signal. It is also conceivable that a measurement parameter is the cavity pressure during the introduction of the matrix material and / or a measurement parameter is the cavity temperature during the introduction.

Advantageously, corresponding value-time courses are determined for the measurement parameters, so that a corresponding sensor value can be derived at any time during the introduction of the matrix material.

In a further advantageous embodiment, the transit time of an ultrasonic signal within the fiber or matrix material is determined with the aid of an ultrasonic sensor as a real sensor, whereby, for example, changes within the component can be determined. In the simplest case, this can be used to determine when the flow front has arrived at the corresponding sensor position.

It is also expedient to determine the cavity pressure with the aid of a pressure sensor as a real sensor, as a result of which the achievement of the flow front at the corresponding sensor position can likewise be determined.

It is also conceivable that with the aid of a timer as a real sensor, the time since the beginning of the introduction of the matrix material and with the aid of a temperature sensor as a real sensor, the tool internal temperature during the introduction of the matrix material is determined.

In this context, it is particularly advantageous if, with the aid of the timer, the time duration since the beginning of the introduction of the matrix material until the matrix material has reached a predetermined sensor position is determined as a function of the determined sensor values of other real sensors, whereby statements about the flow front course are made can be derived from the discrete sensor positions. For if the flow front reaches the corresponding sensor position, for example of an ultrasound or pressure sensor, then the signal response changes, from which it can then be clearly deduced that the matrix material has infiltrated the region around the sensor position.

In a further advantageous embodiment, the simulated injection course determined for monitoring the real injection process is visualized and displayed on a display, so that during the injection process itself the real processes running inside and out can be displayed on a display or screen.

In a further advantageous embodiment, as a function of the determined simulated course of injection by the evaluation unit, an injection state of the real injection process determined, which can be derived directly, for example, the position of the flow front and the flow front. Thus, during the injection process from the simulated injection curves at any time of the injection state of the current real injection process can be determined.

In a further advantageous embodiment, the evaluation unit determines whether or not an injection end state of the current injection process corresponds to predetermined quality parameters as a function of the determined simulated injection course. Thus, it can be predicted from the determined simulated injection course whether complete impregnation of the fiber material corresponding to the quality requirements will take place or whether the component will have a possibly defective location, for example in the form of dry spots. In addition, regions can be identified in which defects can be expected and regions in which the process was ideal. The effort of the component testing can be significantly reduced as a result, since the downstream quality assurance then only needs to examine those regions in which an impregnation was not ideal.

In a further advantageous embodiment, the evaluation unit is set up to determine the time from which the deviation of the infusion / injection with a certain probability leads to a rejection and thus the deviation is irreversible. If the process is aborted at this time, the component can be filled by a second, manual injection / infusion process with modified sprue strategies and thus possibly be rescued. As a result, the scrap can be significantly reduced with minor manual rework.

In a further advantageous embodiment, the simulated injection profiles are determined with the aid of a parameterizable model, wherein the values of the process parameters of the individual simulations are varied. The predetermined process parameters enter as an input into the parameterizable model, so that a simulated injection course, for example in the form of a set of sensor values, is determined for each set of concrete values of process parameters.

The variable process parameters may be, for example, the permeability of the fiber material, the viscosity of the matrix material, the fiber volume content of the fiber material, the component thickness, the injection pressure and / or the position of leading edges in or on the edges of the fiber material. The leading edges in the fiber material are areas in which the permeability of the fiber material relative to the rest of the fiber material increases significantly, so that the flow resistance in these areas is significantly lower than in the other areas of the fiber material, resulting in faster flow of the injected matrix material. The flow front thus does not develop uniformly along the pressure gradient, but deviates from it. In particular, leading edges are causes of gas inclusions or other defects.

For the simulation of the injection profiles, the entire component is likewise subdivided into a finite number of small segments, which may mean, for example, that the boundaries of fiber layers are depicted as elements or expected critical areas are resolved more finely. The segmented component is now used as the basis for the parameterizable model because the corresponding process parameters are then varied accordingly for each simulation process.

In a further advantageous embodiment, the boundary conditions of the injection process, which can be changed during the injection process, are adapted in dependence on the simulated injection process determined and, depending on the injection state derived therefrom, so as to take account of the current injection state. Thus, for example, in the case of a determined simulated course of injection which predicts a low permeability in certain areas, the injection pressure can be raised in order to achieve complete saturation in the available pot life.

In addition, it can be derived from the simulated injection progressions, which are possible causes for the deviation from an ideal course, if the corresponding process parameters with which this simulated injection course was simulated are stored for each simulated injection course. With knowledge of the corresponding process parameters, the possible cause for a non-ideal injection course can be determined.

The invention will be explained by way of example with reference to the attached figures.

• 1 - schematische Darstellung einer Anlage zum Injizieren von Matrixmaterial;
• 2a, 2b - Darstellung eines idealen Injektionsverlaufs;
• 3a, 3b - Darstellung eines Injektionsverlaufes mit variierten Prozessparametern.

Show it:

• 1 - Schematic representation of a system for injecting matrix material;
• 2a . 2 B - representation of an ideal injection course;
• 3a . 3b - Presentation of an injection course with varied process parameters.

1 shows a plant 10 , can be made with the fiber composite components in an injection process. The attachment 10 has a tool for this 11 which is a shaping tool surface 12 has on a fiber material 13 can be stored. That on the shaping tool surface 12 deposited fiber material 13 is then using a vacuum film 14 sealed vacuum-tight.

1 is just one embodiment of a possible plant configuration. It is also conceivable that, for example, two mold halves are provided, between which the fiber material can be introduced.

The attachment 10 has a plant control 15 in particular for controlling the injection of the in the reservoir 16 located matrix material 17 into the fiber material 13 is trained. The plant control 15 It is designed in such a way that it in particular the essential boundary conditions and process parameters in the injection of the matrix material 17 into the fiber material 13 adjust and adjust if necessary.

In the embodiment of 1 is an evaluation unit 18 Part of the system control 15 which does not have to be mandatory.

Furthermore, the attachment points 10 a data store 19 in which a variety of simulated injection progressions 20 stored together with the respective simulated sensor values. The storage of these simulated injection profiles with the simulated sensor values can be done by means of a machine learning system, in which the sensor values and the associated injection profiles form the corresponding training data. The data store 19 is with the evaluation unit 18 connected by signal technology.

Furthermore, the evaluation unit 18 with a display 21 connected to represent appropriate information regarding the real injection process can.

The attachment 10 further includes a plurality of ultrasonic sensors 22 on, with which the duration of an ultrasonic signal can be measured in the direction of component thickness. The ultrasonic sensors 22 are there with the plant control 15 or directly with the evaluation unit 18 connected, so that the evaluation unit 18 corresponding information about the speed of sound within the component at the corresponding sensor positions of the ultrasonic sensors 22 receives.

The use of ultrasonic sensors 22 in 1 is to be understood as an example only. Conceivable here is any type of sensor that can give information about the flow front within the tool and fiber material.

The matrix material 17 is doing over the sprue 23 into the fiber material 13 injected, wherein the flow direction of the matrix material in the direction of the marked arrow RF is defined. That in the fiber material 13 injected matrix material 17 Thus, successively, the individual sensor positions of the ultrasonic sensors 22 happen, which is detectable by changing the duration of the ultrasonic signals. It can thus be determined as a measurement parameter with respect to the respective sensor position, how long the matrix material 17 since the beginning of the injection of the matrix material up to the respective sensor position needs.

The evaluation unit 18 is now formed from these data, based on the corresponding sensor values and the time associated with it until the matrix material has reached the corresponding sensor positions, the respective one in the data memory 19 stored simulated injection course 20 whose simulated sensor values correspond to or correlate to the real sensor values. This way from the evaluation unit 18 determined simulated injection course 20 will then be on a display 21 so that you can immediately see how the flow front is within the fiber material 13 spreads.

2a shows schematically greatly simplified the simulation of an injection process. The manufactured component is greatly simplified, which serves only illustrative purposes. The component consists of four segments, where at the marked points five sensors S ₁ to S ₅ are arranged. The sprue of the matrix material is at the position 23 intended.

The sensors are designed so that they can detect the presence or absence of matrix material, which is possible for example with the aid of the already mentioned ultrasonic sensors. In this case, it is possible to determine the time that the matrix material has needed since the beginning of the injection process in order to reach the corresponding sensor position.

As can be seen, the matrix material spreads 17 initially circular around the gate 23 out. In the middle picture it can be seen that the matrix material 17 after a while the sensors S ₁ . S ₂ such as S ₃ has reached. In the further course then also the sensors S ₄ and S ₅ reached (right picture).

The time periods or duration of each sensor until the matrix material has reached this sensor is in 2 B shown. There it can be seen that the sensors S ₁ and S ₂ already reached after about 500 seconds, while the middle sensor S ₃ was reached after about 1,000 seconds. The sensors S ₄ and S ₅ located at the maximum distance to Angus 23 are reached after about 3,500 seconds.

For this, in 2a In the course of the simulation of this injection course, the sensor values can be determined in this case during the simulation of this injection course S ₁ to S ₅ in the form of the time duration until passing through the matrix material at the respective sensor position and store as simulated sensor values. For the injection course there is thus a sensor tuple S ₁ to S ₅ in front. The in 2a and 2 B illustrated injection curve with the sensor values can be considered as an ideal injection process.

In the 3a and 3b is shown a further simulated injection course and related sensor values, in which case a process parameter was varied with respect to the ideal course of injection, namely that between the sensors S ₂ and S ₅ an advance edge has been specified, in which the permeability is significantly higher than in relation to the rest of the fiber material, so that at a constant injection pressure, a lead of the matrix material is to be expected here. This is the second picture of the 3a to recognize where in the upper league area 30 the matrix material already the sensor S ₅ has reached, while the rest of the flow front lies clearly behind it.

Finally, this can also be seen in the sensor values, which in 3b are set out. It can be seen that sensor S ₅ indicates the presence of matrix material after only 1,500 seconds, whereas in the idealized injection course, shown in 2 B , there the sensor only after about 3,500 seconds struck.

If a real component is now manufactured in a real injection process, the sensor values are continuously recorded here as well. The real sensors on the tool or component correspond to the sensor positions in the simulation in order to obtain comparable sensor values. If, in the real injection process, a similar or identical combination of the sensor values is obtained S1 to S5 , as in 3b is shown, it can be assumed with high probability that the course of injection is as it is in 3a was shown and simulated. This allows conclusions to be drawn during the injection process as to how the current actual injection state is, ie the position of the flow front can be determined. In addition, predictions can be made as to how the flow front will possibly be in the further course of injection.

LIST OF REFERENCE NUMBERS

10

investment

11

Tool

12

shaping tool surface

13

fiber material

14

vacuum film

15

system control

16

reservoir

17

matrix material

18

evaluation

19

data storage

20

simulated injection courses

21

display

22

ultrasonic sensors

23

sprue

30

Voreilbereich

S ₁ to S ₅

sensors

Claims (10)

Method for producing a fiber composite component by means of an injection method, in which a matrix material (17) is injected into a fiber material (13) and the injected matrix material (17) is cured, with a) providing a plurality of simulated injection curves (20) for respectively different process parameters of the injection process in a data memory (19), wherein each simulated injection curve (20) simulated sensor values of at least one, the injection process characterizing measurement parameters are associated with the simulation of the injection process with the process parameters underlying the respective simulation of the injection process were determined, wherein during the injection of the matrix material (17) into the fiber material (13) the injection process is monitored by b) real sensor values of at least one measurement parameter indicative of the injection process, which corresponds to one of the measurement parameters of the simulated sensor values, are detected with the aid of real sensors, and c) for monitoring the real injection process, that simulated injection course (20) is determined by an evaluation unit (18) whose simulated sensor values correspond to the detected real sensor values or correlate with the detected real sensor values. Method according to Claim 1 , characterized in that the simulated injection courses (20) are provided by a machine learning system, in particular by a neural network, in which the simulated injection courses (20) together with their respective simulated sensor values form the training data with which the machine learning system allocates of sensor values as input and simulated injection history (20) trained as output with respect to the input. Method according to Claim 1 or 2 , characterized in that a measurement parameter is the time since the beginning of the injection of the matrix material (17) until the matrix material (17) has reached a predetermined sensor position. Method according to one of the preceding claims, characterized in that the transit time of an ultrasonic signal is determined with the aid of an ultrasonic sensor (22) as a real sensor and / or the cavity pressure is determined by means of a pressure sensor as a real sensor. Method according to one of the preceding claims, characterized in that the simulated injection profile (20) determined for monitoring the real injection process is visualized on a display (21). Method according to one of the preceding claims, characterized in that the injection state of the real injection process is determined by the evaluation unit (18) as a function of the determined simulated injection course (20). Method according to one of the preceding claims, characterized in that it is determined in dependence on the determined simulated injection course (20) by the evaluation unit (18) whether an injection end state of the current injection process will correspond to predetermined quality parameters. Method according to one of the preceding claims, characterized in that a plurality of injection curves (20) are simulated by means of a parameterizable model, wherein the values of the process parameters of the individual simulations are varied. Method according to Claim 8 , characterized in that the variable process parameters are the permeability of the fiber material (13), the viscosity of the matrix material (17), the fiber volume content of the fiber material (13), the component thickness, the injection pressure and / or positions of leading edges in the fiber material (13) , Method according to one of the preceding claims, characterized in that depending on the determined simulated injection course (20) and possibly the injection state, the parameters of the injection process, which are variable during the injection process, set or adjusted.

Images