Inspection model learning device, inspection device, inspection model learning method, inspection method, program, and recovery method

Abstract

A verification model training device 1 with which it is possible to verify, with high precision, the breakage state of a manufactured material, the verification model training device 1 comprising: a known material data input unit 11 for accepting input of the distribution of tapping sound data that includes a plurality of feature values relating to tapping sound obtained by hitting a known material for which the breakage state is known, and input of known material data indicating the breakage state of the known material; a training proportion calculation unit 13 for determining a plurality of segments to divide the distribution on the basis of the known material data and calculating a training proportion at which the tapping sound data corresponding to the breakage state is included in each of the plurality of segments; and a model training unit 14 for training a model on the basis of the training proportion, the model outputting the breakage state of an unknown material, which is a material for which the breakage state is unknown.

Description

[Technical Field]

[0001] The present disclosure relates to an inspection model learning device, an inspection device, an inspection model learning method, an inspection method, a program, a recovery method, a continuous fiber reinforced resin composite material, and a manufacturing method. [Background technology]

[0002] It has been known that physical property information of a resin composition can be estimated based on manufacturing condition information for manufacturing the resin composition. For example, Patent Document 1 describes a method of generating a learning model for estimating physical property information of a resin composition from manufacturing condition information for manufacturing the resin composition, inputting the manufacturing condition information into the learning model, and outputting physical property information. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-163783 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technique described in Patent Document 1 does not inspect materials that have actually been manufactured, and it is desired to inspect the state of destruction of manufactured materials with high accuracy.

[0005] In view of the above circumstances, the purpose of the present disclosure is to provide an inspection model learning device, an inspection device, an inspection model learning method, an inspection method, a program, a recovery method, a continuous fiber reinforced resin composite material, and a manufacturing method that can inspect the fracture state of manufactured materials with high accuracy. [Means for solving the problem]

[0006] As a result of extensive research conducted by the inventors in order to solve the above problems, they discovered that it is possible to obtain an inspection model learning device, an inspection device, an inspection model learning method, an inspection method, a program, a recovery method, a continuous fiber reinforced resin composite material, and a manufacturing method that can inspect the state of failure of manufactured materials with high accuracy, and thus completed the present invention.

[0007] That is, the embodiments of the present invention are as follows. [1] a known material data input unit that receives input of a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by hammering a known material whose fracture state is known, and known material data indicating the fracture state of the known material; a learning ratio calculation unit that determines a plurality of divisions into which the distribution is divided based on the known material data, and calculates a learning ratio that is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; a model learning unit that learns a model that outputs a failure state of an unknown material whose failure state is unknown based on the learning ratio; An inspection model learning device comprising: And, the learning ratio calculation unit determines the plurality of categories by non-hierarchical clustering. . [ 2 ] The plurality of characteristic values ​​are amplitude and duration, [1 ] The inspection model learning device described above. [ 3 ] The plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission. [1] or [2] The inspection model learning device according to claim 1. [ 4 ] An inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hitting sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that shows a distribution of hitting sound data including a plurality of characteristic values ​​related to hitting sounds obtained by hitting a known material whose failure state is known, and the failure state of the known material, an unknown material data input unit that receives input of unknown material data that indicates a distribution of hammering sound data that indicates the plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; a judgment ratio calculation unit that calculates a judgment ratio, which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of sections; a determination unit that inputs the determination ratio into the model and outputs a fracture state of the unknown material based on the determination ratio; An inspection device comprising: And, the plurality of divisions are determined by dividing the distribution of the hammering sound data by non-hierarchical clustering. . [ 5 ] The determination unit causes the model to output the failure state based on a determination ratio included in one of the plurality of divisions in the hammering sound data and a determination ratio included in another division different from the one division in the hammering sound data. 4 ] An inspection device described in [ 6 ] The plurality of characteristic values ​​are amplitude and duration; 4] or [5] The inspection device described in [ 7 ] The plurality of characteristic values ​​relating to the hammering sounds are obtained by acoustic emission, 4 ]from[ 6 ] crab The inspection device described. [ 8 ] A step of receiving input of known material data indicating a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking a known material whose fracture state is known, and the fracture state of the known material; determining a plurality of divisions into which the distribution is divided based on the known material data, and calculating a learning ratio which is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; a step of learning a model that outputs a failure state of an unknown material, the failure state of which is unknown, based on the learning ratio; A method for learning a model for inspection, including And, the plurality of divisions are determined by dividing the distribution of the hitting sound data by non-hierarchical clustering. . [ 9 ] The plurality of characteristic values ​​relating to the hammering sounds are obtained by acoustic emission, 8 ] A method for learning a test model described in [1 0 ] Multiple characteristic values ​​of impact sounds obtained by hitting a known material with a known fracture state of an inspection method performed by an inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hammering sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that indicates the failure state of the known material, receiving input of unknown material data indicating a distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; calculating a judgment ratio which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of divisions; a step of inputting the judgment ratio into the model and outputting a fracture state of the unknown material based on the judgment ratio; Inspection methods including And, the plurality of divisions are determined by dividing the distribution of the hammering sound data by non-hierarchical clustering. .

[11] the known material and the unknown material are continuous fiber reinforced resin composite materials containing continuous reinforcing fibers and a thermoplastic resin, The inspection method according to

[10] , wherein the failure state of the unknown material includes each of the states of "non-destructive (normal)", "fiber breakage", "unimpregnated", and "interface failure". [1 2 ] The plurality of characteristic values ​​relating to the hammering sound are obtained by acoustic emission, 0] or

[11] The inspection method described in [1 3 ] Select the computer from [1] to [ 3 ] crab A program for causing the device to function as the inspection model learning device described above. [1 4 ] Select your computer from the 4 ]from[ 7 ] crab A program for causing the inspection device to function as described above. [1 5 ] [1 1

[0023] receiving input of data indicating the fracture state of the unknown material determined using the inspection method described in determining a recovery condition for recovering the interface of the unknown material based on the failure state; recovering the interface of the unknown material according to the recovery condition; A method for recovering unknown materials, including 。 [Effects of the Invention]

[0008] The inspection model learning device, inspection device, inspection model learning method, inspection method, program, recovery method, continuous fiber reinforced resin composite material, and manufacturing method disclosed herein enable the fracture state of a manufactured unknown material to be inspected with high accuracy. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a schematic diagram of a testing model learning device according to an embodiment of the present invention. [Figure 2] FIG. 10 is a diagram showing an example of a hammering test. [Figure 3] 2 is a diagram showing an example of two pieces of known material data input by a known material data input unit shown in FIG. 1. FIG. [Figure 4] FIG. 4 is a diagram showing an example of a plurality of known material data sets, which are greater than the two known material data sets shown in FIG. 3. [Figure 5] FIG. 10 is a diagram illustrating an example of a learning ratio. [Figure 6] FIG. 10 is a diagram illustrating an example of a distribution of learning ratios. [Figure 7] 2 is a flowchart showing an example of the operation of the testing model learning device shown in FIG. [Figure 8]1 is a schematic diagram of an inspection device according to an embodiment of the present invention. [Figure 9] This is a diagram in which the judgment ratio of hammering sound data for an unknown material is added to the distribution of the learning ratio shown in FIG. 6. [Figure 10] 9 is a flowchart showing an example of the operation of the inspection device shown in FIG. 8. [Figure 11] FIG. 2 is a schematic diagram of a recovery control device according to the present embodiment. [Figure 12] 12 is a flowchart showing an example of the operation of the testing model learning device shown in FIG. [Figure 13] FIG. 10 is a diagram illustrating the evaluation accuracy of models according to algorithms. [Figure 14] FIG. 1 is a schematic diagram showing a method for calculating the void ratio between the continuous reinforcing fibers and the thermoplastic resin using a scanning electron microscope image (magnification: 4000 times) of a test piece (b) cut out from the continuous fiber reinforced resin composite material after the recovery step. [Figure 15] 14A and 14B are schematic diagrams each showing a part of the structure of the interface between the continuous reinforcing fibers and the thermoplastic resin, and more specifically, a schematic diagram showing a region between the dotted line showing a similarity of the shape of the interface between the continuous reinforcing fibers and the thermoplastic resin and the solid line showing the shape of the interface between the continuous reinforcing fibers and the thermoplastic resin. [Figure 16] FIG. 2 is a hardware block diagram of an inspection model learning device and an inspection device. DETAILED DESCRIPTION OF THE INVENTION

[0010] <Configuration of the inspection model learning device> The configuration of a testing model learning device 1 of this embodiment will be described with reference to FIGS.

[0011] As shown in Figure 1, the inspection model learning device 1 of this embodiment includes a known material data input unit 11, a known material data storage unit 12, a learning ratio calculation unit 13, a model learning unit 14, and a model storage unit 15.

[0012] The known material data input unit 11 may be configured with an input interface. The input interface is an interface that accepts input of information and may be a pointing device, keyboard, mouse, etc. The input interface may also be an interface that accepts input of information received via a communication interface. For example, standards such as Ethernet (registered trademark), FDDI (Fiber Distributed Data Interface), and Wi-Fi (registered trademark) may be used for the communication interface. The known material data storage unit 12 and the model storage unit 15 may be configured with memory. The memory may be configured with a hard disk drive (HDD), a solid state drive (SSD), an electrically erasable programmable read-only memory (EEPROM), a read-only memory (ROM), a random access memory (RAM), etc. The learning ratio calculation unit 13 and the model learning unit 14 may constitute a control unit (controller). The control unit may be configured with dedicated hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array), or may be configured with a processor, or may be configured with both.

[0013] The known material data input unit 11 accepts input of known material data indicating the distribution of hammering sound data, including multiple characteristic values ​​related to hammering sounds obtained by hammering a known material whose failure state is known, and the failure state of the known material. The failure state refers to the state of failure of the material and includes each of the following states in the material: "non-failure (normal)," "fiber breakage," "unimpregnated," and "interface failure." The "unimpregnated" state refers to a state in which the resin that forms the material has not sufficiently penetrated into the reinforcing fibers due to the short time required to mold the material. A material whose failure state is unknown is called an "unknown material," and is sometimes simply referred to as "material" without distinguishing between known and unknown materials.

[0014] The hammering data includes multiple characteristic values ​​related to hammering sounds obtained by hammering a material during hammering testing. Examples of materials include continuous fiber reinforced resin composites, resin materials, rubber materials, metal materials, concrete materials, wood, paper materials, fiber materials, glass, organic natural materials, ceramics, and combinations of these materials. Continuous fiber reinforced resin composites and resin materials are particularly suitable for this testing method due to their complex fracture patterns. Continuous fiber reinforced resin composites may be molded materials made by impregnating glass fiber woven fabric with a thermoplastic resin and solidifying it. Furthermore, continuous fiber reinforced resin composites may be unidirectional or may be made of random fibers, carbon fibers, thermosetting resins, short fibers, or the like.

[0015] In hammering testing, for example, as shown in FIG. 2, when a material MT is struck by a pen PN, it emits vibrations and sound waves, such as AE (acoustic emission) waves, and an AE sensor attached to the surface of the material MT detects multiple characteristic values ​​of the vibrations and sound waves. Examples of vibrations and sound waves include sound, AE, and ultrasound. AE is preferred because it is less affected by the measurement environment and enables highly accurate inspection of the state of destruction. In the example shown in FIG. 2, the material MT is struck by a pen PN, but this is not limited thereto. For example, the material MT may be struck with a tool such as a metal rod or hammer. The AE sensor is preferably installed close to the portion of the material MT to be inspected. Furthermore, it is preferable that the surface of the portion to be inspected be smooth. Therefore, the surface may be artificially smoothed by applying grease or the like in advance. Furthermore, it is preferable that the material MT be struck on the side opposite to the side where the AE sensor is located.

[0016] Multiple characteristic values ​​are amplitude (dB), duration (μs), frequency (Hz), sound energy (W / m 2) and the like. The multiple characteristic values ​​may be amplitude and duration. In such a configuration, the inspection device 2 can inspect the fracture state of the unknown material with higher accuracy using the model learned by the model learning unit 14. An example in which the multiple characteristic values ​​are amplitude and duration will be described in detail below.

[0017] In the example shown in FIG. 3, hammering sound data including amplitude and duration obtained by striking one known material whose failure state is "non-destructive" (circles in FIG. 3), and hammering sound data including amplitude and duration obtained by striking one known material whose failure state is "interfacial failure" (squares in FIG. 3) are shown.

[0018] In such an example, the known material data input unit 11 accepts input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by striking a known material whose failure state is "non-failure", and a failure state indicating that the failure state of the known material is "non-failure". Similarly, the known material data input unit 11 accepts input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by striking a known material whose failure state is "interfacial failure", and a failure state indicating that the failure state of the known material is "interfacial failure".

[0019] Note that FIG. 3 shows hammering sound data for one known material whose failure state is "non-destructive" and hammering sound data for one known material whose failure state is "interface failure", but the known material data input unit 11 accepts input of known material data for each of a plurality of known materials whose failure state is "non-destructive" and known material data for each of a plurality of known materials whose failure state is "interface failure".

[0020] The known material data input unit 11 may also accept input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by hammering a known material whose failure state is “fiber breakage”, and a failure state indicating that the failure state of the known material is “fiber breakage”. The known material data input unit 11 may also accept input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by hammering a known material whose failure state is “unimpregnated”, and a failure state indicating that the failure state of the known material is “unimpregnated”.

[0021] The known material data storage unit 12 stores known material data, the input of which is accepted by the known material data input unit 11, including a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking the known material, and a destruction state indicating a destruction state of the known material.

[0022] In the example shown in FIG. 3 , the known material data storage unit 12 stores known material data including a distribution of hammering sound data, including the amplitude and duration of hammering sounds obtained by striking a known material whose failure state is “non-destructive,” and a failure state indicating that the failure state of the known material is “non-destructive.” Similarly, the known material data storage unit 12 stores known material data including a distribution of hammering sound data, including the amplitude and duration of hammering sounds obtained by striking a known material whose failure state is “interfacial failure,” and a failure state indicating that the failure state of the known material is “interfacial failure.” Note that FIG. 3 shows hammering sound data for one known material whose failure state is “non-destructive” and hammering sound data for one known material whose failure state is “interfacial failure.” However, the known material data input unit 11 stores known material data for each of a plurality of known materials whose failure state is “non-destructive” and known material data for each of a plurality of known materials whose failure state is “interfacial failure.”

[0023] Furthermore, in a configuration in which the known material data input unit 11 accepts input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by hammering a known material whose failure state is "fiber breakage" and a failure state indicating that the failure state of the known material is "fiber breakage", the known material data storage unit 12 may store the known material data.In a configuration in which the known material data input unit 11 accepts input of known material data including a distribution of hammering sound data including amplitudes and durations of hammering sounds obtained by hammering a known material whose failure state is "unimpregnated", and a failure state indicating that the failure state of the known material is "unimpregnated", the known material data storage unit 12 may store the known material data.

[0024] The learning ratio calculation unit 13 determines a plurality of divisions into which the distribution of the hammering sound data is divided based on the known material data, and calculates a learning ratio Ra, which is the ratio at which the hammering sound data corresponding to the destroyed state is included in each of the plurality of divisions.

[0025] Specifically, first, the learning ratio calculation unit 13 determines a plurality of divisions for dividing the distribution of hammering sound data for each of a plurality of known materials included in the known material data whose failure state is "non-destructive" and hammering sound data for each of a plurality of known materials included in the known material data whose failure state is "interface failure." For example, the learning ratio calculation unit 13 may determine a plurality of divisions by non-hierarchical clustering. Specifically, the learning ratio calculation unit 13 may determine a plurality of divisions for dividing the distribution of hammering sound data by the K-means method.

[0026] 4 shows the distribution of hammering sound data for each of a plurality of known materials included in the known material data whose failure state is "non-destructive" and the distribution of hammering sound data for each of a plurality of known materials included in the known material data whose failure state is "interface failure." In the example of FIG. 4, the learning ratio calculation unit 13 determines six divisions CL, divisions CL1 to CL6, by the K-means method.

[0027] Next, the learning ratio calculation unit 13 calculates the ratio of the hammering sound data included in each section for each known material. Fig. 5 shows the number of hammering sound data and the learning ratio Ra of the hammering sound data included in each of the six sections CL from section CL1 to section CL6, corresponding to an identifier (ID: Identification) that uniquely identifies the known material.

[0028] In the example of FIG. 5 , the learning ratio calculation unit 13 calculates that, for the known material identified by ID "1", the learning ratio Ra1 of the hammering sound data included in the category CL1 is 0%, the learning ratio Ra2 of the hammering sound data included in the category CL2 is approximately 3.88%, the learning ratio Ra3 of the hammering sound data included in the category CL3 is approximately 0.72%, the learning ratio Ra4 of the hammering sound data included in the category CL4 is approximately 14.08%, the learning ratio Ra5 ​​of the hammering sound data included in the category CL5 is approximately 0.86%, and the learning ratio Ra6 of the hammering sound data included in the category CL6 is approximately 78.06%.

[0029] The model learning unit 14 learns a model that outputs the fracture state of the unknown material based on the learning ratio. The model learning unit 14 can learn the model using algorithms such as Random forest, MLP (Multilayer Perceptron), Logistic regression, SVM (Support Vector Machine), and kNN (K-Nearest Neighbor Algorithm).

[0030] Specifically, the model learning unit 14 learns a model that outputs the failure state of an unknown material, whose failure state is unknown, based on the distribution of known materials in the learning ratios Ra of two divisions CL among the multiple divisions determined by the learning ratio calculation unit 13. FIG. 6 shows the distribution of hammering sound data of known materials, with the learning ratio Ra4 shown in FIG. 5 on the horizontal axis and the learning ratio Ra6 on the vertical axis. As an example, the learning ratio Ra4 of the known material corresponding to ID "2" shown in FIG. 5 is approximately 12.77%, and the learning ratio Ra6 of the known material is approximately 72.19%. Therefore, in FIG. 6, a triangle N2 indicating that the failure state of the known material corresponding to ID "2" is "interfacial failure" is plotted at the position where the learning ratio Ra4 is approximately 12.77% and the learning ratio Ra6 is approximately 72.19%. Similarly, in Figure 6, for other known materials, a triangle mark indicating that the failure state of the known material is "interface failure" and a circle mark indicating that the failure state of the known material is "non-failure" are plotted based on the learning ratios Ra4 and Ra6.

[0031] In this way, the model learning unit 14 stores, for each known material, the learning ratio Rai of the hammering sound data included in the division CLi (i=1 to M (M is the number of divisions CL)) and the learning ratio Raj of the hammering sound data included in the division CLj (j=1 to M, where j≠i). Then, the model learning unit 14 stores the learning ratio Rai (i=1 to M, where M is the number of divisions CL) corresponding to each of the plurality of divisions and the failure state corresponding to the range of the learning ratio Raj (j=1 to M, where j≠i). In this way, a model capable of outputting the failure state according to the learning ratios Rai and Raj is learned.

[0032] In the example shown in Figure 6, in range Rg1 for learning ratios Ra4 and Ra6, there are more known materials with a failure state of "interfacial failure" than there are known materials with a failure state of "non-failure." Also, in range Rg2 for learning ratios Ra4 and Ra6, there are more known materials with a failure state of "non-failure" than there are known materials with a failure state of "interfacial failure."

[0033] 6, i=4 and j=6, and the model learning unit 14 learns a model for all combinations of the divisions CLi and CLj. In this configuration, the model can output a destruction state for each combination of the divisions CLi and CLj.

[0034] Furthermore, although the above describes examples in which the failure states are "non-destructive" and "interface failure," the model learning unit 14 may similarly train the model using known materials whose failure states are "fiber breakage" and "unimpregnated." In such a configuration, the model can output a failure state from "non-destructive," "interface failure," "fiber breakage," and "unimpregnated" for every combination of the divisions CLi and CLj. Furthermore, the model may output whether the material needs to be replaced depending on the failure state of the material, or may output repair conditions for the material depending on the failure state of the material.

[0035] Furthermore, when learning the teacher data, the model learning unit 14 may set importance levels corresponding to the combination of categories so as to most closely approximate the destruction state of the actual teacher data. Specifically, the model learning unit 14 sets the importance levels of the destruction state candidates, which will be described in detail later, for each combination of categories determined for the known material data so as to output the destruction state of the known material data with the highest accuracy.

[0036] The model storage unit 15 stores the model learned by the model learning unit 14.

[0037] <Operation of the inspection model learning device> Here, the operation of the inspection model learning device 1 according to this embodiment will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the operation of the inspection model learning device 1 according to this embodiment. The operation of the inspection model learning device 1 described with reference to Fig. 7 corresponds to an example of an inspection model learning method for the inspection model learning device 1 according to this embodiment.

[0038] In step S11, the known material data input unit 11 receives input of known material data indicating the distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking a known material whose fracture state is known, and the fracture state of the known material.

[0039] In step S12, the learning ratio calculation unit 13 determines a plurality of divisions CL into which the distribution of the hitting sound data is divided, based on the known material data.

[0040] In step S13, the learning ratio calculation unit 13 calculates a learning ratio Ra, which is the ratio at which the hammering sound data corresponding to the broken state is included in each of the plurality of sections CL.

[0041] In step S14, the model learning unit 14 learns a model that outputs the fracture state of the unknown material, the fracture state of which is unknown, based on the learning ratio Ra.

[0042] <Configuration of inspection equipment> The configuration of the inspection device 2 of this embodiment will be described with reference to Figures 8 and 9. The inspection device 2 may be configured integrally with the inspection model learning device 1, or may be configured separately.

[0043] The inspection device 2 according to this embodiment performs inspection using a model that outputs the failure state of an unknown material based on a training ratio Ra, which is the ratio of the hammering sound data included in each of multiple classifications CL in the distribution, trained using known material data indicating the failure state of the known material and a distribution of hammering sound data including multiple characteristic values ​​related to hammering sounds obtained by striking a known material whose failure state is known. The training ratio Ra is the ratio of the hammering sound data included in each of multiple classifications CL in the distribution. The inspection device 2 can be configured as a portable device capable of inspecting materials installed in products such as automobiles. As shown in FIG. 8 , the inspection device 2 includes an unknown material data input unit 21, a judgment ratio calculation unit 22, a judgment unit 23, an output unit 24, and a model storage unit 15. The inspection device 2 may further include one or more of the known material data input unit 11, the known material data storage unit 12, the learning ratio calculation unit 13, and the model training unit 14 of the inspection model training device 1.

[0044] The unknown material data input unit 21 may be configured by an input interface. The judgment ratio calculation unit 22 and the judgment unit 23 may form a control unit. The output unit 24 may be configured by an output interface. The output interface is an interface that outputs information and may be, for example, a display. The output interface may also be a communication interface for outputting information to another device.

[0045] The model storage unit 15 stores the model acquired by the inspection model learning device 1 described above. That is, the model storage unit 15 provided in the inspection device 2 may be the same as the model storage unit 15 provided in the inspection model learning device 1. The inspection device 2 may receive the model from the inspection model learning device 1 via a communication network, or may acquire the model from the inspection model learning device 1 via any medium. Furthermore, when the inspection device 2 is configured integrally with the inspection model learning device 1, the model storage unit 15 of the inspection device 2 may be the model storage unit 15 of the inspection model learning device 1 itself.

[0046] The unknown material data input unit 21 accepts input of unknown material data indicating a distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering an unknown material. The types of the plurality of characteristic values ​​indicated by the hammering sound data accepted as input by the unknown material data input unit 21 are the same as the types of the plurality of characteristic values ​​indicated by the hammering sound data accepted as input by the known material data input unit 11. As in the above example, in a configuration in which the plurality of characteristic values ​​indicated by the hammering sound data accepted as input by the known material data input unit 11 are amplitude and duration, the plurality of characteristic values ​​indicated by the hammering sound data accepted as input by the unknown material data input unit 21 are amplitude and duration.

[0047] The determination ratio calculation unit 22 calculates a determination ratio Rb, which is the ratio of the hitting sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of sections CL.

[0048] The determination unit 23 inputs the determination ratio Rb to the model, and outputs the fracture state of the unknown material based on the determination ratio Rb. Specifically, the determination unit 23 may output the fracture state based on the determination ratio Rb included in one of the multiple divisions CL in the hammering sound data and the determination ratio Rb included in another division CL different from the one division CL in the hammering sound data.

[0049] Fig. 9 is a diagram in which marks indicating the judgment ratios Rb of the hammering sound data of the unknown material are plotted on the distribution of the learning ratios Ra of the hammering sound data of the known material shown in Fig. 6. In the example of Fig. 9, the model can output the failure state of the unknown material in which the judgment ratios Rb4 and Rb6 based on the hammering sound data fall within the range Rg1 as "interface failure". In addition, the model can output the failure state of the unknown material in which the judgment ratios Rb4 and Rb6 based on the hammering sound data fall within the range Rg2 as "non-destruction".

[0050] In the example of Figure 9, for one unknown material MT1, the judgment ratio Rb4 is approximately 48% and the judgment ratio Rb6 is approximately 22%, and for another unknown material MT2, the judgment ratio Rb4 is approximately 25% and the judgment ratio Rb6 is approximately 4%. In this example, the judgment ratios Rb4 and Rb6 of the unknown material MT1 are included in the range Rg1 described with reference to Figure 6, and the judgment ratios Rb4 and Rb6 of the unknown material MT2 are included in the range Rg2 described with reference to Figure 6. Therefore, the model outputs that the failure state of the unknown material MT1 is "interface failure" and the failure state of the unknown material MT2 is "non-failure."

[0051] In this way, the determination unit 23 makes the model output a failure state for each combination of the divisions CLi and CLj based on the determination ratio Rbi of the hammering sound data included in the division CLi and the determination ratio Rbj of the hammering sound data included in the division CLj in the hammering sound data of the unknown material.

[0052] Furthermore, the determination unit 23 may treat the failure state for each combination of the divisions CLi and CLj output by the model as a failure state candidate, and determine the failure state for each of the failure state candidates based on the importance level that is preset corresponding to the combination of the divisions CLi and CLj. As described above, the importance level is set by the model learning unit 14 of the inspection model learning device 1.

[0053] For example, in an example where the number M of sections CL is 3, the preset importance levels corresponding to the combinations of sections CL1 and CL2, sections CL2 and CL3, and sections CL3 and CL1 are 70, 40, and 35, respectively. In such an example, it is assumed that the failure state is output as "non-failure" based on the judgment ratios Rb1 and Rb2 of sections CL1 and CL2, that the failure state is output as "interface failure" based on the judgment ratios Rb2 and Rb3 of sections CL2 and CL3, and that the failure state is output as "interface failure" based on the judgment ratios Rb3 and Rb1 of sections CL3 and CL1. In this case, the determination unit 23 determines the failure state candidate based on the determination ratios Rb1 and Rb2 for the divisions CL1 and CL2 to be "non-failure," the failure state candidate based on the determination ratios Rb2 and Rb3 for the divisions CL2 and CL3 to be "interface failure," and the failure state candidate based on the determination ratios Rb3 and Rb1 for the divisions CL3 and CL1 to be "interface failure." Here, the determination unit 23 calculates that the importance corresponding to the failure state candidate "non-failure" is 70, and the importance corresponding to the failure state candidate "interface failure" is 40 + 35 = 75. The determination unit 23 determines that the failure state is "interface failure" because the importance corresponding to the failure state candidate "interface failure" is higher than the importance corresponding to the failure state candidate "non-failure."

[0054] Furthermore, the determination unit 23 may determine whether the unknown material is repairable or not based on the state of destruction of the unknown material. The method by which the determination unit 23 determines whether the unknown material is repairable or not may be any method.

[0055] As an example, the determination unit 23 may determine whether an unknown material is repairable based on a destruction state and repairability information indicating whether the material is repairable, which are stored in advance in memory in association with each other. Specifically, the determination unit 23 may determine that the unknown material is repairable when the repairability information stored in association with the destruction state of the unknown material indicates that the material is repairable, and may determine that the unknown material is unrepairable when the repairability information stored in association with the destruction state of the unknown material indicates that the material is unrepairable.

[0056] As another example, the determination unit 23 may determine whether the unknown material is repairable based on a destruction state and a value indicating the degree of destruction that are previously associated and stored in memory. Specifically, the determination unit 23 may determine that the unknown material is repairable if the value indicating the degree of destruction stored in association with the destruction state is within a predetermined range, and may determine that the unknown material is not repairable if the value indicating the degree of destruction is outside the predetermined range.

[0057] The output unit 24 outputs to a display or other device the data indicating the fracture state of the unknown material that the determination unit 23 has caused the model to output. For example, the output unit 24 may output the data indicating the fracture state of the unknown material to the recovery control device 3, which will be described later.

[0058] As an example, the output unit 24 may not output data indicating that the failure state is "non-destructive" to the recovery control device 3, but may output data indicating that the failure state is determined to be repairable, among data indicating that the failure state is "interface failure," "fiber breakage," or "unimpregnated," to the recovery control device 3. In such a configuration, for example, an unknown material in a "non-destructive" state may be returned to be used as it was originally. Also, an unknown material in a non-repairable state may be recovered.

[0059] In the above description, an example has been described in which the model outputs whether the failure state of the unknown material is "non-destructive" or "interface failure." However, in a configuration in which the inspection model learning device 1 learns a model by further using known materials whose failure states are "fiber breakage" and "unimpregnated," the inspection device 2 may use the model to cause the model to output whether the failure state is "non-destructive," "interface failure," "fiber breakage," or "unimpregnated."

[0060] The unknown material data input unit 21 may also accept input of multiple pieces of unknown material data obtained by hitting one unknown material multiple times. In this configuration, the determination ratio calculation unit 22 calculates multiple determination ratios Rb based on each of the multiple pieces of unknown material data. The determination unit 23 then causes the model to output multiple failure states based on each of the multiple determination ratios Rb. The determination unit 23 may then determine that the most common failure state among the multiple failure states output by the model is the failure state of the unknown material.

[0061] <Operation of the inspection device> Here, the operation of the inspection device 2 according to this embodiment will be described with reference to FIG. 10 . FIG. 10 is a flowchart showing an example of the operation of the inspection device 2 according to this embodiment. The operation of the inspection device 2 described with reference to FIG. 10 corresponds to an example of an inspection method of the inspection device 2 according to this embodiment. This operation is performed by the inspection device 2, which performs inspection using a model that outputs the failure state of an unknown material based on a learning ratio Ra, which is the ratio of the hammering sound data included in each of multiple categories CL in the distribution, learned using known material data showing the failure state of the known material and a distribution of hammering sound data including multiple characteristic values ​​related to hammering sounds obtained by hammering a known material whose failure state is known. Prior to the inspection device 2 performing its operation, an existing (e.g., used) molded product may be disassembled into parts, and parts containing a continuous fiber reinforced plastic composite material may be recovered from the disassembled parts. In such a case, the inspection device 2 performs its operation using the continuous fiber reinforced plastic composite material included in the recovered parts as the unknown material.

[0062] In step S21, the unknown material data input unit 21 accepts input of unknown material data indicating a distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering an unknown material.

[0063] In step S22, the determination ratio calculation unit 22 calculates a determination ratio Rb, which is the ratio of the hitting sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of sections CL.

[0064] In step S23, the determination unit 23 inputs the determination ratio Rb into the model and outputs the failure state of the unknown material based on the determination ratio Rb. At this time, the determination unit 23 may use the failure state output from the model as a failure state candidate and determine the failure state based on the importance of the failure state candidate. After this, the output unit 42 may output data indicating the failure state.

[0065] <Configuration of recovery control device> 11, the recovery control device 3 according to this embodiment includes a broken state input unit 31, a recovery condition determination unit 32, and a recovery condition output unit 33. The recovery control device 3 may further include a recovery unit 34. The recovery control device 3 may also include a display unit 35.

[0066] The destruction state input unit 31 is configured by an input interface, the recovery condition determination unit 32 is configured by a controller, and the recovery condition output unit 33 is configured by an output interface. The recovery unit 34 is configured by a heater. The recovery unit 34 may also be configured by a heater and one or more of a pressurizer and a water absorber. The display unit 35 may also be configured by a display interface such as an organic EL (Electro Luminescence) or liquid crystal panel.

[0067] The recovery control device 3 may further include one or more of the unknown material data input unit 21, the judgment ratio calculation unit 22, the judgment unit 23, the output unit 24, and the model storage unit 15 of the above-mentioned inspection device 2. The recovery control device 3 may further include one or more of the known material data input unit 11, the known material data storage unit 12, the learning ratio calculation unit 13, and the model learning unit 14 of the inspection model learning device 1.

[0068] The destruction state input unit 31 receives input of data indicating the destruction state of the unknown material output by the output unit 24 of the inspection device 2.

[0069] The recovery condition determination unit 32 determines recovery conditions for recovering the interface of the unknown material based on the fracture state. When the unknown material is a composite material containing two or more materials with different chemical structures, the interface is the surface where the materials with different chemical structures meet. When the unknown material is a continuous fiber reinforced resin composite material, the interface is, for example, the interface between the continuous reinforcing fiber and the thermoplastic resin. The recovery conditions include conditions related to the treatment performed on the unknown material to recover the interface of the unknown material, and can be, for example, one or more of the temperature conditions for heating the unknown material, the pressure conditions, and the water absorption conditions. In addition, the conditions of the recovery process described below may be appropriately selected as the recovery conditions.

[0070] The recovery condition output unit 33 may output the recovery condition to the recovery unit 34. Furthermore, the recovery condition output unit 33 may transmit a control signal indicating the recovery condition to a recovery device having the same function as the recovery unit 34. In a configuration in which the recovery condition output unit 33 transmits a control signal to the recovery device, the recovery control device 3 does not need to include the recovery unit 34. Furthermore, the recovery condition output unit 33 may output the recovery condition to the display unit 35.

[0071] The recovery unit 34 recovers the interface of the unknown material according to the recovery conditions calculated by the recovery condition determination unit 32. Specifically, in a configuration in which the material is a composite material (e.g., a continuous fiber reinforced resin composite material) containing materials with two or more different chemical structures, the recovery unit 34 recovers the interface of the unknown material by executing a recovery process described below. Note that, as described above, in a configuration in which the recovery condition output unit 33 transmits a control signal indicating the recovery conditions to the recovery device, the recovery device recovers the interface of the unknown material according to the recovery conditions indicated by the control signal.

[0072] (Recovery process) Here, the recovery step performed by the recovery unit 34 will be described in detail. The recovery step in this embodiment may be a step of heating the continuous fiber reinforced resin composite material subjected to a load test in a temperature range from the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material subjected to the load test to the glass transition temperature of the thermoplastic resin plus 150°C or less, or less than the melting point of the thermoplastic resin. This can promote the repair of voids at the interface between the continuous reinforcing fibers and the thermoplastic resin. In this embodiment, the continuous fiber reinforced resin composite material subjected to a load test is preferably heated to a temperature range equal to or higher than the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material and lower than the melting point of the thermoplastic resin. In the recovery step, the heating temperature range for heating the continuous fiber reinforced resin composite material subjected to the load test is more preferably from the glass transition temperature of the thermoplastic resin +100°C to the melting point of the thermoplastic resin -50°C, even more preferably from the glass transition temperature of the thermoplastic resin +50°C to the melting point of the thermoplastic resin -30°C, and even more preferably from the glass transition temperature of the thermoplastic resin +30°C to the melting point of the thermoplastic resin -20°C. Here, the glass transition temperature refers to the temperature obtained when a dried thermoplastic resin is measured by DSC. When the heating temperature in the recovery step is within the above range, the recovery rate of the continuous fiber reinforced resin composite material is increased, making it easier to maintain the shape of the composite material before and after the recovery step. Furthermore, in the recovery step, when the continuous fiber reinforced resin composite material contains two or more thermoplastic resins, it is preferable to heat the material at a temperature equal to or higher than the highest glass transition temperature of the thermoplastic resins having two or more glass transition temperatures. In this embodiment, the heating time within the temperature range from the glass transition temperature of the thermoplastic resin or higher to the glass transition temperature of the thermoplastic resin + 150°C or lower, or lower than the melting point of the thermoplastic resin, is not particularly limited, and is preferably, for example, 15 minutes or longer. The upper limit of the heating time is preferably 10 hours or shorter, more preferably 2 hours or shorter, and even more preferably 1 hour or shorter.

[0073] The recovery step of this embodiment may be a step of heating the continuous fiber reinforced resin composite material that has been subjected to a load test within a predetermined temperature range, as described above. The heating atmosphere (or the atmosphere of the recovery step) is preferably at least one condition selected from the group consisting of a non-pressurized (atmospheric pressure) condition, a pressurized condition, and a water-absorbing condition. A preferred mode of the recovery step of this embodiment is to carry out the recovery step under at least one condition selected from the group consisting of a non-pressurized (atmospheric pressure) condition, a pressurized condition, and a water-absorbing condition. In this embodiment, "under non-pressurized (atmospheric pressure) conditions" refers to 0.1 MPa or less, and "under pressurized conditions" refers to pressures greater than 0.1 MPa and less than or equal to 25 MPa. Furthermore, "under water absorption conditions" refers to a condition in which the water absorption rate of the continuous fiber reinforced resin composite material of this embodiment is at least 0.3%, and preferably 0.3% or more and 2% or less. A water absorption treatment for controlling the water absorption rate of the continuous fiber reinforced resin composite material within the above range may include a treatment in which a continuous fiber reinforced resin composite material subjected to a load test is subjected to water absorption in water at 80°C for 18 hours. The water absorption treatment may be performed by a water absorber constituting the recovery unit 34 described above. The water absorption rate (%) is calculated using the following formula. Water absorption rate (%) = (mass of continuous fiber reinforced resin composite material after water absorption treatment - mass of continuous fiber reinforced resin composite material before water absorption treatment) / mass of continuous fiber reinforced resin composite material before water absorption treatment × 100

[0074] The heating in the recovery step of this embodiment may be performed under pressure. When the recovery step in formula (1) representing the recovery rate of this embodiment is performed on the continuous fiber reinforced resin composite material after the load test under pressure (for example, a pressure of 20 MPa), the recovery step in formula (1) representing the recovery rate of this embodiment is preferably performed by heating for 15 minutes or more in a temperature range from the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material to the glass transition temperature + 100°C. The method for applying pressure is not particularly limited, and examples thereof include a method for applying pressure using a continuous compression molding machine or a press, a method for applying pressure using an autoclave, and a method for applying pressure using a foaming material in a mold. The heating in the recovery step of this embodiment may be performed without pressure or under a pressure lower than atmospheric pressure (e.g., 0.1 MPa or less). Examples of the pressure lower than atmospheric pressure (e.g., 0.1 MPa or less) include a method in which the heating is performed in a vacuum dryer. When the recovery step of this embodiment is performed on the continuous fiber reinforced resin composite material after the load test without applying pressure (for example, under atmospheric pressure), the recovery step in formula (1) representing the recovery rate of this embodiment includes heating for 15 minutes or more within a temperature range from the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material to the glass transition temperature + 100°C, inclusive. This can further promote the repair of voids at the interface between the continuous reinforcing fibers and the thermoplastic resin. The recovery step in this embodiment may include a cooling step, which may involve cooling at room temperature in the case of a recovery step under no pressure, or cooling the mold and the like with cooling water in the case of a recovery step under pressure. The heating in the recovery step of this embodiment may be performed while the continuous fiber reinforced resin composite material has absorbed water. The water absorption rate of the continuous fiber reinforced resin composite material is preferably 0.3% or more. In this embodiment, heating of the continuous fiber reinforced resin composite material subjected to the load test in a water-absorbed state is preferably carried out at a temperature equal to or lower than the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material + 50°C, and more preferably at a temperature equal to or lower than the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material + 30°C. The heating step in this embodiment is not particularly limited, and examples thereof include a method in which the heating step is carried out in water and a method in a humid environment.

[0075] A preferred recovery step in this embodiment is to heat the continuous fiber reinforced resin composite material that has been subjected to the load test for 15 to 60 minutes using a heating means (e.g., a heater) at a humidity of 0 to 100% RH, a pressure of 0.5 to 10 MPa, and a temperature of 80 to 265°C. The heating means used in the recovery step in this embodiment is not particularly limited as long as it is a heating means that can achieve a heating temperature of 80 to 265°C, and examples thereof include a thermostatic bath, a molding machine, a hot air circulation dryer, a thermo-humidistat bath, and a vacuum dryer. In one aspect of the recovery step of this embodiment, the continuous fiber reinforced resin composite material after the load test is heated in a temperature range of 180 to 260°C for 15 to 30 minutes without pressure (a pressure of 0.1 MPa or less). It is preferable to heat in an inert gas atmosphere, preferably nitrogen or argon. In another aspect of the recovery step of this embodiment, the continuous fiber reinforced resin composite material after the load test is heated under pressure (pressure of 0.5 to 5 MPa) at a temperature range of 80 to 150°C for 30 to 60 minutes in an atmosphere with a humidity of 90 to 100% RH. The continuous fiber reinforced resin composite material may be pressurized by a pressurizer constituting the recovery section 34 described above.

[0076] -Long-term characteristic recovery rate- The long-term property recovery rate against load, which indicates an index of recovery of the fatigue properties of the continuous fiber reinforced resin composite material of this embodiment, can be evaluated, for example, by performing a vibration fatigue test on the continuous fiber reinforced resin composite material at any intensity until the number of tests reaches an arbitrary number, and then subjecting the continuous fiber reinforced resin composite material that has undergone the vibration fatigue test to a recovery process, and then performing the vibration fatigue test again under the same conditions, and dividing the resulting number of breaks by the number of breaks in a vibration fatigue test of a new continuous fiber reinforced resin under the same conditions. The long-term property recovery rate (%) of this embodiment is expressed by the formula described in the Examples section below, and the long-term property recovery rate (%) of the continuous fiber reinforced resin composite material of this embodiment is preferably 65% ​​or more, and more preferably 80% or more.

[0077] -Bending strength recovery rate- The bending strength recovery rate, which indicates one index of recovery of fatigue properties of the continuous fiber reinforced resin composite material of this embodiment, can be evaluated by conducting a bending test, which is an example of a load test, and then subjecting the continuous fiber reinforced resin composite material that has been subjected to the bending test to a recovery process, and then conducting the bending test again in the same manner to obtain a bending strength (MPa) divided by the bending strength (MPa) of the continuous fiber reinforced resin that has been subjected to the first bending test. The bending strength recovery rate (%) of this embodiment is expressed by the formula described in the Examples section below, and the bending strength recovery rate (%) of the continuous fiber reinforced resin composite material of this embodiment is preferably 65% ​​or more, and more preferably 80% or more.

[0078] -Flexural modulus and recovery rate- The flexural modulus recovery rate, which indicates an index of recovery of the fatigue properties of the continuous fiber reinforced resin composite material of this embodiment, can be evaluated by conducting a bending test, which is an example of a load test, and then subjecting the continuous fiber reinforced resin composite material that has been subjected to the bending test to a recovery process, and then conducting the bending test again in the same manner, and dividing the obtained flexural modulus (GPa) by the flexural modulus (GPa) of the continuous fiber reinforced resin that has been subjected to the first bending test. The flexural modulus recovery rate (%) of this embodiment is expressed by the formula described in the Examples section below, and the flexural modulus recovery rate (%) of the continuous fiber reinforced resin composite material of this embodiment is preferably 65% ​​or more, and more preferably 80% or more.

[0079] The display unit 35 displays the recovery conditions. In such a configuration, the operator may refer to the recovery conditions and make various settings for recovering the interface of the unknown material according to the recovery conditions.

[0080] <Recovery Device Operation> Here, the operation of the recovery control device 3 according to this embodiment will be described with reference to Fig. 12. Fig. 12 is a flowchart showing an example of the operation of the recovery control device 3 according to this embodiment. The operation of the recovery control device 3 described with reference to Fig. 12 corresponds to an example of a recovery method of the recovery control device 3 according to this embodiment.

[0081] In step S31, the failure state input unit 31 receives input of data indicating the failure state of the unknown material output by the output unit 24 of the inspection device 2.

[0082] In step S32, the recovery conditions for recovering the interface of the unknown material are determined based on the fracture state.

[0083] In step S33, the interface of the unknown material is restored under the restoration conditions. At this time, the restoration control device 3 equipped with the restoration unit 34 or the restoration device may contain the unknown material and restore the interface of the contained unknown material. In this configuration, once the interface of the unknown material has been restored, the unknown material is removed from the restoration control device 3 or the restoration device. Furthermore, the unknown material with the restored interface may be recovered and used to mold a part.

[0084] As described above, according to this embodiment, the inspection model learning device 1 includes: a known material data input unit 11 that receives input of known material data indicating the distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by hammering a known material whose failure state is known, and the failure state of the known material; a learning ratio calculation unit 13 that determines a plurality of sections CL in the distribution based on the hammering sound data and calculates a learning ratio Ra, which is the ratio at which the hammering sound data is included in each of the plurality of sections CL; and a model learning unit 14 that learns a model that outputs the failure state of the unknown material based on the learning ratio Ra. Also, according to this embodiment, the inspection device 2 includes: an unknown material data input unit 21 that receives input of unknown material data indicating the distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering an unknown material whose failure state is unknown; a judgment ratio calculation unit 22 that calculates a judgment ratio Rb, which is the ratio at which the hammering sound data, the distribution of which is indicated by the unknown material data, is included in each of the plurality of sections CL; and a judgment unit that inputs the judgment ratio Rb to the model and outputs the failure state of the unknown material based on the judgment ratio Rb. This allows the inspection model learning device 1 and the inspection device 2 to inspect the fracture state of the manufactured unknown material with high accuracy.

[0085] Furthermore, according to this embodiment, the plurality of characteristic values ​​are amplitude and duration, which allows the fracture state of the manufactured unknown material to be inspected with higher accuracy.

[0086] Furthermore, according to this embodiment, the inspection device 2 causes the model to output a fracture state based on the judgment ratio Rb included in one of the multiple divisions CL and the judgment ratio Rb included in another division CL different from the one division CL in the hammering sound data. This allows the inspection device 2 to inspect the fracture state of the manufactured unknown material with higher accuracy than a device that outputs the fracture state based on the distribution of hammering sound data without determining the division CL.

[0087] Furthermore, according to this embodiment, when the inspection device 2 performed an inspection using a model trained by the inspection model learning device 1 using the Random forest, MLP, Logistic regression, SVM, and kNN algorithms, the coefficients of determination R2, which indicate the evaluation accuracy of the model, were 80%, 80%, 74%, 76%, and 67%, respectively, as shown in Fig. 13. In this way, by training a model using the Random forest or MLP algorithm, the inspection model learning device 1 can inspect the fracture state of a manufactured unknown material with higher accuracy.

[0088] Furthermore, according to this embodiment, materials can be inspected before product shipment, making it possible to more appropriately determine whether or not a product can be shipped.

[0089] <Continuous fiber reinforced resin composite material> As described above, a material that can be measured by the above-mentioned inspection method is, for example, a continuous fiber reinforced resin composite material containing continuous reinforcing fibers and a thermoplastic resin. Here, the continuous fiber reinforced resin composite material that can be inspected by the above-mentioned inspection method will be described in detail. The continuous fiber reinforced resin composite material of this embodiment (hereinafter also simply referred to as "composite material") contains continuous reinforcing fibers and a thermoplastic resin, and has a recovery rate of 50% or more, which is expressed by the following formula (1). Recovery rate=(B / A)×100 Equation (1) (In the above formula (1), A represents the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin, among the continuous reinforcing fibers present in a scanning electron microscope image (magnification: 4000 to 4500 times) of a test piece (a) cut out from the continuous fiber reinforced resin composite material, which are measured at any 20 points so that the approximately circular cross section of the continuous reinforcing fibers is included in the image, The continuous fiber reinforced resin composite material is subjected to a load test and a recovery step after the load test, and a test piece (b) is cut out from the recovered continuous fiber reinforced resin composite material. Measurements are made at any 20 points so that the approximately circular cross sections of the continuous reinforcing fibers appear in an image (magnification: 4000 to 4500 times) taken by a scanning electron microscope, and B is the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin, among any of the continuous reinforcing fibers present in the image. When the recovery rate is within the above range, damage to the interface between the continuous reinforcing fibers and the thermoplastic resin and the fatigue characteristics or physical properties of the continuous fiber reinforced resin composite material that have been reduced due to the damage can be easily recovered. In this embodiment, the recovery rate is preferably 60% or more, more preferably 70% or more, even more preferably 70 to 100%, and even more preferably 80 to 100%.

[0090] <Recovery rate> The recovery rate in this embodiment is one of the characteristics of the continuous fiber reinforced resin composite material of this embodiment. It is an index that represents the extent to which the interface state between the continuous reinforcing fibers and the thermoplastic resin constituting the continuous fiber reinforced resin composite material has been restored when the target continuous fiber reinforced resin composite material is subjected to a specific load test and a specific recovery process, compared to the continuous fiber reinforced resin composite material before the specific load test. Therefore, when the recovery rate is within the above range, damage to the interface between the continuous reinforcing fibers and the thermoplastic resin and the fatigue properties or physical properties of the continuous fiber reinforced resin composite material that have been reduced due to the damage can be easily recovered. This allows for repeated use of components, etc., using the continuous fiber reinforced resin composite material and improves the handleability of the composite material at high temperatures. More specifically, the recovery rate in this embodiment is a characteristic that represents the proportion of voids at the interface between the continuous reinforcing fibers and the thermoplastic resin in the continuous fiber reinforced resin composite material after the recovery process that have been repaired by subjecting a continuous fiber reinforced resin composite material containing continuous reinforcing fibers and a thermoplastic resin to a specific load test to obtain a continuous fiber reinforced resin composite material (also referred to as the continuous fiber reinforced resin composite material after the load test), and then subjecting the continuous fiber reinforced resin composite material after the load test to a specific recovery process to obtain test pieces (the continuous fiber reinforced resin composite material after the recovery process), and observing each test piece with an electron microscope. The continuous fiber reinforced resin composite material after being subjected to the load test and the recovery process after the load test is also referred to as a recovered continuous fiber reinforced resin composite material.

[0091] In the above formula (1), A refers to the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fiber and the thermoplastic resin in the continuous fiber reinforced resin composite material, out of any 20 continuous reinforcing fibers present in a scanning electron microscope image (magnification 4000 to 4500 times, for example 4000 times) of a test piece (a) cut out from the continuous fiber reinforced resin composite material (before the load test), after measuring any 20 points so that one approximately circular cross section of the continuous reinforcing fiber appears in the image. Similarly, in the above formula (1), B is the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin, out of any 20 continuous reinforcing fibers present in a scanning electron microscope image (magnification 4000 to 4500 times, for example 4000 times) of a test piece (b) cut out from the recovered continuous fiber reinforced resin composite material, which has been subjected to a load test and a recovery process after the load test. After measuring 20 arbitrary points so that one continuous reinforcing fiber appears in the image, B is the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin, out of any 20 continuous reinforcing fibers present in the image. In this specification, the method for determining whether or not voids exist between the continuous reinforcing fiber and the thermoplastic resin is that if the void ratio in the outer peripheral region of the thermoplastic resin, which is one-tenth of the radius (r1) of one continuous reinforcing fiber, is 10% or less, no voids exist between the continuous reinforcing fiber (one fiber) and the thermoplastic resin. This will be explained in detail below using Figures 14 and 15. FIG. 14 is a schematic diagram showing a method for determining the presence or absence of voids between the continuous reinforcing fibers and the thermoplastic resin using a scanning electron microscope image (magnification 4000x) of a test piece (b) cut out from the continuous fiber-reinforced resin composite material after the recovery process. The image shown in FIG. 14 shows three cross sections of the approximately circular continuous reinforcing fibers, and the interface between the continuous reinforcing fibers and the thermoplastic resin in the center of FIG. 14 is depicted by a solid line. Also, in FIG. 14, a dotted line depicts a similarity diagram of the shape of the interface between the continuous reinforcing fibers and the thermoplastic resin, along the solid line, at a location 1 / 10 of the cross-sectional radius (r1) of one continuous reinforcing fiber in the center away from the interface between the continuous reinforcing fibers and the thermoplastic resin, i.e., the solid line portion. Finally, FIG. 15 is a schematic diagram of a cut-out region between the dotted line depicting the similarity diagram of the shape of the interface between the continuous reinforcing fibers and the thermoplastic resin and the solid line depicting the shape of the interface between the continuous reinforcing fibers and the thermoplastic resin. Furthermore, the region between the dotted line and the solid line (the cut-out region shown in FIG. 15) is defined as the outer peripheral region, and when the void ratio in this outer peripheral region is 10% or less, it is defined as no void exists between the continuous reinforcing fiber (one fiber) and the thermoplastic resin. In FIGS. 14 and 15, the white region is the void portion. When the cross section of the continuous reinforcing fiber is displayed in white, for example, the void portion may be displayed in black. Therefore, the method of counting the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fiber and the thermoplastic resin, which will be described later, is based on an example in which the void portion is the white region. Therefore, in this embodiment, the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin is counted according to the following procedures (I) to (III). (I) Image analysis software (ImageJ, manufactured by the National Institutes of Health, USA) is used to perform binarization processing, and the approximately circular cross section of one continuous reinforcing fiber shown in the scanning electron microscope image of test piece (a) or test piece (b) is drawn with a solid line, and then a dotted line similar to the shape of the solid line (the interface between the continuous reinforcing fiber and the thermoplastic resin) is drawn along the shape of the solid line at a location 1 / 10 of the cross-sectional radius (r1) of the one continuous reinforcing fiber away from the solid line. (II) The area obtained by cutting out the solid line portion (cross-sectional view of one continuous reinforcing fiber) from the similar diagram is defined as the outer peripheral area, and the total area of ​​the white voids in the outer peripheral area is calculated using the following formula (x): Total area of ​​white area / total area of ​​outer peripheral area × 100 (x) When the porosity expressed as ρ is 10% or less, it is defined as no porosity. (III) The above procedures (I) and (II) are carried out on scanning electron microscope images of test piece (a) or test piece (b) taken at any 20 points, and the number of continuous reinforcing fibers surrounded by an outer peripheral region with a void ratio of 10% or less is counted to calculate A and B. The cross-sectional radius (r1) of the continuous reinforcing fiber is the cross-sectional radius (r1) of one substantially circular continuous reinforcing fiber drawn with a solid line among the cross sections of the substantially circular continuous reinforcing fibers shown in the image taken by a scanning electron microscope, and is calculated for each of the 20 arbitrary points. The cross-sectional radius (r1) of one continuous reinforcing fiber is calculated by using the image analysis software to find the circumferential length of the substantially circular cross section of the continuous reinforcing fiber drawn with a solid line (= the outer periphery of the cross section of the substantially circular continuous reinforcing fiber) and dividing it by 2π to calculate the cross-sectional radius (r1).

[0092] The recovery rate of the continuous fiber reinforced resin composite material of the present disclosure is determined by subjecting the continuous fiber reinforced resin composite material to a load test (e.g., a vibration fatigue test, a bending test, or an impact test) and then cutting the material after a predetermined recovery process using a band saw or the like. The cross section perpendicular to the longitudinal direction of the continuous reinforcing fibers is then polished using a polishing machine to produce a test piece (b). The polished surface is then observed using a scanning electron microscope (FESEM) to calculate the number of continuous reinforcing fibers (B) that do not have voids at the interface between each continuous reinforcing fiber and the thermoplastic resin. Furthermore, a test piece (a) is prepared by cutting and polishing the continuous fiber reinforced resin composite material before the load test using the same method as described above. The number of continuous reinforcing fibers (A) that do not have voids at the interface between each continuous reinforcing fiber and the thermoplastic resin in the observed test piece (a) is then calculated. The recovery rate (%) can then be calculated by dividing B by A and multiplying by 100. In one example of a method for producing the test piece (a) in this embodiment, a 10 mm x 10 mm test piece is cut out from any location of the continuous fiber reinforced resin composite material using a band saw, and the test piece is polished with a polishing machine using a polishing pressure of 416 g / cm2 on the polished surface, in the following order: waterproof paper grit #2000 for 10 minutes, silicon carbide film with a grain size of 9 μm for 5 minutes, alumina film with a grain size of 5 μm for 5 minutes, alumina film with a grain size of 3 μm for 5 minutes, alumina film with a grain size of 1 μm for 5 minutes, and colloidal silica (Baicalox 0.1CR) with a grain size of 0.1 μm using polyurethane foam buffing paper for 5 minutes, while adding water at approximately 7 mL / min. On the other hand, in this embodiment, one example of a method for producing test piece (b) is to (1) subject a continuous fiber-reinforced resin composite material to a load test and recovery process described below, and then cut the continuous fiber-reinforced resin composite material using a band saw, or (2) cut the continuous fiber-reinforced resin composite material using a band saw, then subject it to a load test and recovery process described below, and then perform the following polishing process. The polishing process involves applying a polishing pressure of 416 g / cm2 to the polished surface using a #2000 waterproof sandpaper for 10 minutes, followed by 5 minutes with silicon carbide film with a 9 μm grain size, 5 minutes with alumina film with a 3 μm grain size, 5 minutes with alumina film with a 1 μm grain size, and 5 minutes with colloidal silica (Baicalox 0.1CR) with a 0.1 μm grain size using polyurethane foam buffing paper, and polishing the surface with water at a rate of approximately 7 mL / min.

[0093] <Load test> The load test in this embodiment refers to a test in which a mechanical energy load is applied to at least a part of a test piece (a) of the continuous fiber reinforced resin composite material in order to examine the physical properties of the continuous fiber reinforced resin composite material that is the subject of the load test and the microstructure of the continuous fiber reinforced resin composite material (the interfacial state between the continuous reinforcing fibers and the thermoplastic resin), and examples of such a test include a vibration fatigue test, a bending test, and an impact test. Therefore, the load test in this embodiment may be at least one test selected from the group consisting of a vibration fatigue test, a bending test, and an impact test. The vibration fatigue test of this embodiment is carried out under the conditions described in the Examples section below. Specific vibration fatigue test conditions applicable to this embodiment include preparing a test piece for vibration fatigue testing from a continuous fiber-reinforced resin composite material produced in accordance with ASTM-D1822, and testing it using a servo-type strength testing machine at a test temperature of 20 to 30°C (23°C in the Examples described below), a frequency of 10 to 30 Hz (20 Hz in the Examples described below), a sine wave waveform, and a chuck distance of 20 to 40 mm (35 mm in the Examples described below). The bending test of this embodiment is carried out under the conditions described in the Examples section below. Specific bending test conditions applicable to this embodiment include cutting rectangular bending test specimens from the prepared continuous fiber-reinforced resin composite material, each having a length of 70 mm to 150 mm (100 mm in the Examples described below), a width of 5 to 15 mm (10 mm in the Examples described below), and a thickness of 1 to 3 mm (2 mm in the Examples described below), and drying them in a vacuum dryer at 65 to 95 ° C (80 ° C in the Examples described below) for 10 hours or more (18 hours or more in the Examples), and then bending them in a universal testing machine using a three-point bending jig, setting the span to thickness × 10 (mm) or more (thickness × 16 (mm) in the Examples), at a speed of 0.5 to 2 mm / min (1 mm / min in the Examples), at 20 to 30 ° C (23 ° C in the Examples described below), in an environment of 20 to 70% RH (50% RH in the Examples), and in an environment of 100 ° C, 50% RH. The impact test of this embodiment is carried out under the conditions described in the Examples section below. Specific impact test conditions applicable to this embodiment include cutting out an impact test specimen having a length of 100 mm to 250 mm (150 mm in the examples described later), a width of 70 to 150 mm (100 mm in the examples described later), and a thickness of 1 to 3 mm (2 mm in the examples described later) from the prepared continuous fiber reinforced resin composite material, fixing the impact test specimen to a stainless steel jig with a hole having a diameter of 15 mm to 45 mm (30 mm in the examples described later), and allowing a steel striker with a hemispherical tip (diameter 10 mm to 24 mm, 16 mm in the examples described later) to freely drop onto the center of the impact test specimen to collide with it.

[0094] <Recovery characteristics of physical properties of continuous fiber reinforced resin composite materials> The recovery characteristics of the physical properties of the continuous fiber reinforced resin composite material of this embodiment can be evaluated by, for example, conducting a physical property test such as a bending test or a tensile test, subjecting the continuous fiber reinforced resin composite material to a recovery process after it breaks, and then conducting the physical property test again under the same conditions, and dividing the obtained strength or elastic modulus by the strength or elastic modulus obtained in the physical property test of a new continuous fiber reinforced resin under the same conditions. In other words, if a combination of one load test and one recovery step is considered to be one cycle, the continuous fiber reinforced resin composite material of this embodiment can recover its physical properties to those of the original continuous fiber reinforced resin composite material even after two or more cycles of load tests and recovery steps are performed. For example, the long-term property recovery rate, flexural strength recovery rate, and flexural modulus recovery rate described above all represent the recovery characteristics of the physical properties of the continuous fiber reinforced resin composite material by calculating the ratio between the properties after two load tests and the properties after the first load test. Furthermore, in order to have the property of being able to recover to (close to) the physical property values ​​of the initial continuous fiber reinforced resin composite material even after performing such multiple cycles of load testing and recovery process, it is preferable to perform a recovery process in which the continuous fiber reinforced resin composite material subjected to the load test is heated for 15 minutes or more at a temperature equal to or higher than the glass transition temperature of the thermoplastic resin contained in the continuous fiber reinforced resin composite material + 150°C, select a polyamide resin as the thermoplastic resin, and a glass fiber substrate as the continuous reinforcing fiber, and satisfy the following: the number of layers of the polyamide resin is 1 to 10 (particularly preferably 6) and the number of layers of the glass fiber substrate is 1 to 9 (particularly preferably 5), and the order in which the polyamide resin layers and the glass fiber substrate are alternately stacked.

[0095] <Handling in high-temperature environments> The handleability of the continuous fiber reinforced resin composite material of this embodiment in a high-temperature environment can be evaluated, for example, by the change in thickness of the continuous fiber reinforced resin composite material when the continuous fiber reinforced resin composite material is heated to the melting temperature (melting point or glass transition temperature if the continuous fiber reinforced resin composite material does not have a melting point) of the thermoplastic resin contained in the continuous fiber reinforced resin composite material. More specifically, the handleability in a high-temperature environment can be evaluated by the method described in the examples. If the continuous fiber reinforced resin composite material has good handleability in a high-temperature environment, it is possible to minimize deterioration in the appearance or physical properties of the continuous fiber reinforced resin composite material due to molding. In this specification, as an index of the handleability of a continuous fiber reinforced resin composite material at high temperatures, the change in thickness and the maintenance of shape of the continuous fiber reinforced resin composite material upon heating are measured, as shown in the Examples described below. The reason for this is as follows: Continuous fiber reinforced resin composite materials are generally produced by applying pressure, and therefore residual stress remains in the continuous fiber reinforced resin composite material itself. Therefore, when the continuous fiber reinforced resin composite material is heated, this residual stress is released, and the resin component in the continuous fiber reinforced resin composite material expands, causing the shape to collapse and making the continuous fiber reinforced resin composite material difficult to handle. Therefore, the smaller the change in thickness of a continuous fiber reinforced resin composite material upon heating, the better the handleability at high temperatures tends to be.

[0096] (Preferred form of continuous fiber reinforced resin composite material) The continuous fiber reinforced resin composite material of the present embodiment includes continuous reinforcing fibers, a thermoplastic resin, and a sizing agent (1) containing a coupling agent, and is obtained by laminating the continuous reinforcing fibers and the thermoplastic resin, It is preferable that the interfacial polishing strength is 0.8 to 1.2 times the interfacial polishing strength of a reference composite material measured by the following (i) to (ii). (i) A reference composite material is prepared by laminating reference continuous reinforcing fibers obtained by surface-treating the continuous reinforcing fibers with only the same coupling agent as the coupling agent contained in the bundling agent (1) in the same amount as the coupling agent described above, and the thermoplastic resin. (ii) Measuring the interfacial abrasive strength of the reference composite material. When the interfacial polishing strength of the continuous fiber reinforced resin composite material of this embodiment is 0.8 to 1.2 times that of the reference composite material, the interfacial adhesion strength between the continuous reinforcing fibers and the thermoplastic resin is good, and the recovery rate when subjected to the specified load test and specified recovery process of this embodiment can be easily controlled to a range of 50% to 100%. The method for producing the standard continuous reinforcing fiber is to prepare an aqueous solution containing sizing agent (1) including a coupling agent, and then immerse the continuous reinforcing fiber in the aqueous solution containing sizing agent (1) (50 to 90°C) for 1 to 30 minutes, preferably 1 to 10 minutes, to surface treat the continuous reinforcing fiber. In this specification, a solution containing 0.5 mass% of γ-aminopropyltriethoxysilane and 95.5 mass% of pure water is used as the aqueous solution containing sizing agent (1).

[0097] (Another preferred embodiment of continuous fiber reinforced resin composite material) The continuous fiber reinforced resin composite material of the present embodiment is a continuous fiber reinforced resin composite material obtained by laminating continuous reinforcing fibers surface-treated with a sizing agent (1) and a thermoplastic resin, The interfacial polishing strength of the continuous fiber reinforced resin composite material is preferably at least twice as high as the interfacial polishing strength of a continuous fiber reinforced resin composite material that has not been subjected to surface treatment with the sizing agent (1) (or from which surface treatment with the sizing agent (1) has been removed). In other words, the continuous fiber reinforced resin composite material of the present embodiment contains continuous reinforcing fibers, a thermoplastic resin, and a sizing agent (1), and is a continuous fiber reinforced resin composite material obtained by laminating the continuous reinforcing fibers and the thermoplastic resin, It is preferable that the continuous fiber reinforced resin composite material contains 0.01 mass % or more of the sizing agent (1) relative to the total amount (100 mass %) of the continuous fiber reinforced resin composite material, and that the interfacial polishing strength of the continuous fiber reinforced resin composite material is at least twice the interfacial polishing strength of a continuous fiber reinforced resin composite material containing less than 0.001 mass % of the sizing agent (1). By using continuous reinforcing fibers surface-treated with sizing agent (1), the interfacial polishing strength is at least twice as high as that of continuous fiber-reinforced resin composite materials that are not substantially surface-treated with sizing agent (1). For these reasons, when emphasis is placed on interfacial polishing strength, the continuous reinforcing fibers of this embodiment are preferably surface-treated with the sizing agent (1). That is, the continuous fiber reinforced resin composite material preferably contains 0.01 mass % or more, and more preferably 0.03 mass % or more, of the sizing agent (1) relative to the total amount (100 mass %) of the continuous fiber reinforced resin composite material. The sizing agent (1) preferably contains one or more selected from the group consisting of a silane coupling agent, a lubricant, and a binder. In this embodiment, the following method can be used as an example of a method for measuring the interfacial polishing strength. Test pieces for measuring polishing strength, cut from continuous fiber-reinforced resin composites, were polished using a polishing machine under the following polishing conditions: the polishing pressure was increased from 100 to 300 g / cm² (200 g / cm² in the examples) in increments of 10 to 30 g / cm² (20 g / cm² in the examples). The polished surface was observed using a field-emission scanning electron microscope to determine the maximum polishing pressure at which no voids were formed between each continuous reinforcing fiber and the thermoplastic resin. This maximum polishing pressure was determined just before the polishing pressure at which voids were observed between each continuous reinforcing fiber and the thermoplastic resin. That is, if voids were observed after applying a polishing pressure of 520 g / cm², for example, the maximum polishing pressure was determined to be 500 g / cm². In this embodiment, the polishing conditions are preferably as follows: #2000 waterproof sandpaper for 5 to 30 minutes (10 minutes in this example), 5 to 13 μm (9 μm in this example) silicon carbide film for 1 to 30 minutes (5 minutes in this example), 2 to 8 μm (5 μm) alumina film for 1 to 30 minutes (5 minutes in this example), 1 to 6 μm (3 μm) alumina film for 1 to 30 minutes (5 minutes in this example), 0.5 to 2 μm (1 μm) alumina film for 1 to 30 minutes (5 minutes in this example), and 0.05 to 0.2 μm (0.1 μm) colloidal silica (Baicalox 0.1CR) using polyurethane foam buffing paper for 1 to 30 minutes (5 minutes in this example). Each polishing was performed with water added at a rate of approximately 1 to 15 mL / min (7 mL / min in this example).

[0098] [Form of continuous fiber reinforced resin composite material] The form of the continuous fiber reinforced resin composite material in this embodiment is not particularly limited, and may be any of the following various forms: a form in which a woven, knitted, braided, or pipe-shaped continuous reinforcing fiber is combined with a thermoplastic resin, a form in which continuous reinforcing fibers aligned in one direction are combined with a thermoplastic resin, a form in which yarns made of continuous reinforcing fibers and a thermoplastic resin are aligned in one direction and shaped, and a form in which yarns made of continuous reinforcing fibers and a thermoplastic resin are shaped into a woven, knitted, braided, or pipe-shaped continuous reinforcing fiber. The continuous fiber reinforced resin composite material of this embodiment may be a flat plate or a laminate including a layer of continuous reinforcing fibers and a layer of thermoplastic resin. For example, the length direction of the continuous reinforcing fibers may be arranged approximately parallel to the surface of the flat plate. The continuous reinforcing fiber layer may be a layer including continuous reinforcing fibers (e.g., a continuous reinforcing fiber substrate) and may be a layer in which the interior of the continuous reinforcing fibers is impregnated with a thermoplastic resin. The form of the intermediate material before shaping of the continuous fiber reinforced resin composite material is not particularly limited, and examples thereof include a mixed yarn of continuous reinforcing fibers and resin fibers, a coated yarn in which a bundle of continuous reinforcing fibers is coated with resin, continuous reinforcing fibers pre-impregnated with resin and formed into a tape, continuous reinforcing fibers sandwiched between resin films, continuous reinforcing fibers with resin powder attached, a braided cord made of a bundle of continuous reinforcing fibers surrounded by resin fibers as a core material, reinforcing fibers pre-impregnated with resin, and forms in which continuous reinforcing fibers are in contact with molten resin.

[0099] [Method of manufacturing continuous fiber reinforced resin composite material] The method for producing the continuous fiber reinforced resin composite material of this embodiment is not particularly limited, and the following various methods can be mentioned.

[0100] In one example of a method for producing the above-mentioned continuous fiber reinforced resin composite material, for example, a substrate (e.g., a substrate made of continuous reinforcing fibers, a substrate made of a thermoplastic resin) that constitutes the continuous fiber reinforced resin composite material is cut or shaped to fit the desired composite material, and the required number of pieces or sheets are stacked in consideration of the thickness of the desired product, and then set in a mold according to the mold shape.

[0101] The substrate may be cut one by one, or after stacking the desired number of sheets. From the viewpoint of productivity, it is preferable to cut the substrate in a stacked state. Any cutting method may be used, for example, a water jet, a blade press, a hot blade press, a laser, a plotter, etc. Among these, a hot blade press is preferred, as it provides an excellent cross-sectional shape and, further, improves handling by welding the end faces when cutting multiple sheets stacked together. An appropriate cut shape can be adjusted by repeated trial and error, but it is preferable to set it by performing simulations using CAE (computer-aided engineering) in accordance with the shape of the mold. The base material may be shaped by any method, for example, into a sheet shape.

[0102] A thermoplastic resin substrate and a continuous reinforcing fiber substrate are set in a laminated state (e.g., a state in which multiple continuous reinforcing fiber substrates and multiple thermoplastic resin substrates are alternately stacked), and then the mold is closed and compressed. The mold is then adjusted to a temperature above the melting point of the thermoplastic resin to melt and shape the thermoplastic resin. There are no particular restrictions on the mold clamping pressure, but it is preferably 1 MPa or more, more preferably 3 MPa or more. The mold may be clamped once for degassing, etc., and the mold clamping pressure may be released once after compression molding. From the viewpoint of strength development, the compression molding time is preferably as long as the thermoplastic resin used does not undergo thermal degradation, but from the viewpoint of productivity, it is preferably within 2 minutes, more preferably within 1 minute. Other methods include a method in which the substrate constituting the continuous fiber reinforced resin composite material is continuously supplied using a double belt press or a continuous compression molding device, heated to a temperature above the melting point of the thermoplastic resin, compression molded at an arbitrary pressure, and cooled to a temperature below the crystallization temperature or glass transition temperature of the thermoplastic resin. In the method for producing a continuous fiber-reinforced resin composite material of this embodiment, the continuous reinforcing fibers or the substrate of the continuous reinforcing fibers are preferably surface-treated with a sizing agent (1) before being laminated with the thermoplastic resin substrate, which further improves the interfacial polishing strength.

[0103] The interfacial polishing strength of the continuous fiber reinforced resin composite material obtained by the method for producing a continuous fiber reinforced resin composite material of this embodiment is preferably 0.80 to 1.20 times, and more preferably 0.85 to 1.15 times, the interfacial polishing strength of a continuous fiber reinforced resin composite material in which the sizing agent (1) used for the surface treatment of the continuous reinforcing fibers is only a coupling agent. The interfacial polishing strength of the continuous fiber reinforced resin composite material obtained by the method for producing a continuous fiber reinforced resin composite material of this embodiment is preferably at least two times, more preferably at least three times, and even more preferably at least four times the interfacial polishing strength of the continuous fiber reinforced resin composite material when the surface treatment area of ​​the continuous reinforcing fibers using the sizing agent (1) is removed. As shown in the examples described above and below, the interfacial polishing strength can be determined by polishing a continuous fiber-reinforced resin material with a band saw while varying the polishing pressure applied to a cross section perpendicular to the longitudinal direction of the continuous reinforcing fibers, observing the polished surface with an FESEM, and determining the maximum polishing pressure at which no voids are formed between a single continuous reinforcing fiber and the resin. More specifically, it can be determined by the method described in the examples described below. A continuous fiber-reinforced resin composite material in which the sizing agent (1) for the continuous reinforcing fibers is a coupling agent, preferably an aminosilane, can be prepared by treating a continuous reinforcing fiber substrate in an electric furnace or the like to remove the sizing agent (1), and then using a continuous reinforcing fiber substrate to which a coupling agent has been added to produce a continuous fiber-reinforced resin composite material. Alternatively, a continuous reinforcing fiber substrate can be prepared using continuous reinforcing fibers treated with only a coupling agent, and then using the resulting continuous fiber-reinforced resin composite material. The continuous fiber reinforced resin composite material obtained by removing the sizing agent (1) from the continuous reinforcing fibers can be produced by treating a continuous reinforcing fiber substrate in an electric furnace or the like and using the continuous reinforcing fiber substrate from which the sizing agent (1) has been removed, or by preparing a continuous reinforcing fiber substrate using continuous reinforcing fibers that have not been treated with the sizing agent (1) and producing a continuous fiber reinforced resin composite material. In this embodiment, a specific method for removing the surface treatment portion of the continuous reinforcing fiber with the sizing agent (1) is to place the glass fiber in an electric furnace and heat it at 600°C for 3 hours to remove the surface treatment portion with the sizing agent (1) when the continuous reinforcing fiber is glass fiber. Alternatively, if the continuous reinforcing fiber is carbon fiber, the method can be to hold it in an argon atmosphere at 350°C for 30 minutes to remove the surface treatment portion with the sizing agent (1). When the interfacial strength of the continuous fiber reinforced resin composite material is within this range, the recovery rate of the continuous fiber reinforced resin composite material tends to be large.

[0104] The method for producing a continuous fiber reinforced resin composite material of the present embodiment is a method for producing a continuous fiber reinforced resin composite material obtained by laminating continuous reinforcing fibers that have been surface-treated with a sizing agent (1) containing a coupling agent and a thermoplastic resin, The interfacial polishing strength of the continuous fiber reinforced resin composite material is preferably 0.8 to 1.2 times (more preferably 0.85 to 1.15 times) the interfacial polishing strength of a continuous fiber reinforced resin composite material obtained by laminating continuous reinforcing fibers that have been surface-treated with a sizing agent (2) that is greater than the coupling agent content in the sizing agent (1) (for example, a coupling agent content of 50 mass% or more of the total amount, preferably only coupling agent), and the thermoplastic resin. The sizing agent (2) in this embodiment preferably contains a higher amount of coupling agent than the coupling agent contained in the sizing agent (1). For example, the sizing agent (2) may be a mixed solution of a coupling agent (γ-aminopropyltriethoxysilane) and pure water. The amount of coupling agent preferably accounts for 50 mass % or more of the total amount of the sizing agent (2). In this embodiment, the surface treatment agent (e.g., sizing agent (1)) used when performing surface treatment using only a coupling agent is an aqueous solution of a coupling agent and water, preferably a solution of γ-aminopropyltriethoxysilane mixed with pure water. The sizing agent (1) in this embodiment is preferably a solution or dispersion in which a coupling agent, preferably γ-aminopropyltriethoxysilane, pure water, a lubricant, and a binder are mixed. The method for producing a continuous fiber reinforced resin composite material of the present embodiment is a method for producing a continuous fiber reinforced resin composite material containing a sizing agent (1), continuous reinforcing fibers, and a thermoplastic resin, and obtained by laminating the continuous reinforcing fibers and the thermoplastic resin, It is preferable that the interfacial polishing strength of the continuous fiber reinforced resin composite material is at least two times (preferably at least three times, more preferably at least four times) the interfacial polishing strength of a continuous fiber reinforced resin composite material obtained by laminating the continuous reinforcing fibers from which the surface treatment with the bundling agent (1) has been removed and the thermoplastic resin. For the above reasons, when importance is attached to the interfacial polishing strength, it is preferable that the continuous reinforcing fibers of this embodiment are surface-treated with a sizing agent.

[0105] (continuous reinforcing fiber) The continuous fiber reinforced resin composite material of this embodiment contains continuous reinforcing fibers as an essential component. The continuous reinforcing fibers of this embodiment may be those used in ordinary continuous fiber reinforced resin composite materials. Examples of the continuous reinforcing fibers include, but are not limited to, glass fibers, carbon fibers, plant fibers, aramid fibers, ultra-high strength polyethylene fibers, polybenzazole fibers, liquid crystal polyester fibers, polyketone fibers, metal fibers, and ceramic fibers. From the viewpoints of mechanical properties, thermal properties, and versatility, glass fibers, carbon fibers, plant fibers, and aramid fibers are preferred, and from the viewpoint of productivity, glass fibers are preferred. The above continuous reinforcing fibers may be used alone or in combination of two or more. The continuous reinforcing fibers are preferably surface-treated with a surface treatment agent such as a bundling agent (1). The content of the continuous reinforcing fibers in this embodiment is preferably 90 to 525 mass%, more preferably 150 to 340 mass%, and even more preferably 180 to 300 mass%, relative to 100 mass% of the continuous fiber reinforced resin composite material. The average single fiber diameter of the continuous reinforcing fibers in this embodiment is preferably 8 to 23 μm, and more preferably 14 to 19 μm.

[0106] <Sizing agent> When glass fibers are selected as the continuous reinforcing fibers in this embodiment, the glass fibers may be treated with a surface treatment agent, and a sizing agent (1) may be used as the surface treatment agent. The sizing agent (1) (also referred to as a sizing agent) preferably contains one or more selected from the group consisting of a silane coupling agent, a lubricant, and a binder, and more preferably contains at least a binder or a silane coupling agent. The sizing agent (1) may be a mixture of a silane coupling agent and a binder, or a mixture of a silane coupling agent, a lubricant, and a binder. By using a sizing agent (1) that creates a strong bond between the glass fibers and the resin coating around them, it is possible to obtain a continuous fiber reinforced resin composite material with a low void ratio. In this embodiment, the sizing agent (1) may be added externally to the material to be used, or may be contained internally in the material to be used. For example, a lubricant may be contained in a commercially available product of the thermoplastic resin to be used. Similarly, the sizing agent (2) preferably contains one or more selected from the group consisting of a silane coupling agent, a lubricant, and a binder, and more preferably contains at least a binder or a silane coupling agent.

[0107] -Silane coupling agent-- The silane coupling agent is generally used as a surface treatment agent for glass fibers, and contributes to improving the interfacial adhesive strength. Examples of the silane coupling agent include, but are not limited to, aminosilanes such as γ-aminopropyltrimethoxysilane, γ-aminopropyltriethoxysilane, and N-β-(aminoethyl)-γ-aminopropylmethyldimethoxysilane; mercaptosilanes such as γ-mercaptopropyltrimethoxysilane and γ-mercaptopropyltriethoxysilane; epoxysilanes; vinylsilanes; maleic acids; etc. When polyamide is used as the thermoplastic resin, it is preferable to select one that easily bonds with the carboxyl group or amino group that is the terminal group of the polyamide resin, and aminosilanes are preferred.

[0108] -Lubricant- The lubricant contributes to improving the openability of the glass fibers. As the lubricant, any ordinary liquid or solid lubricating material can be used depending on the purpose, as long as it does not interfere with the silane coupling agent and the binder. Examples of the lubricant include, but are not limited to, animal, vegetable, or mineral waxes such as carnauba wax and lanolin wax; surfactants such as fatty acid amides, fatty acid esters, fatty acid ethers, aromatic esters, and aromatic ethers; and the like.

[0109] -Binding agent- The binder contributes to improving the bundling property of the glass fibers and improving the interfacial adhesive strength. As the material for the binder, polymers according to the purpose, and other thermoplastic resins (c) other than the thermoplastic resins as the main material of the continuous fiber reinforced resin composite material can be used. Examples of polymers that can be used as binders include, but are not limited to, homopolymers of acrylic acid, copolymers of acrylic acid with other copolymerizable monomers, and salts of these with primary, secondary, and tertiary amines. Polyurethane resins synthesized from isocyanates such as m-xylylene diisocyanate, 4,4'-methylenebis(cyclohexyl isocyanate), and isophorone diisocyanate, and polyester or polyether diols are also suitable. The homopolymer of acrylic acid preferably has a weight average molecular weight of 1,000 to 90,000, more preferably 1,000 to 25,000. The copolymerizable monomer constituting the copolymer of acrylic acid and other copolymerizable monomers is not limited to the following, but examples thereof include, among monomers having a hydroxyl group and / or a carboxyl group, one or more selected from the group consisting of acrylic acid, maleic acid, methacrylic acid, vinylacetic acid, crotonic acid, isocrotonic acid, fumaric acid, itaconic acid, citraconic acid, and mesaconic acid (excluding the case where only acrylic acid is used). It is preferable to use one or more ester-based monomers as the copolymerizable monomer. Examples of the salts of the above-mentioned acrylic acid homopolymers and copolymers with primary, secondary, and tertiary amines include, but are not limited to, triethylamine salts, triethanolamine salts, glycine salts, etc. The degree of neutralization is preferably 20 to 90%, more preferably 40 to 60%, from the viewpoints of improving the stability of a mixed solution with other concomitant chemicals (such as a silane coupling agent) and reducing the amine odor. The weight average molecular weight of the acrylic acid polymer that forms the salt is not particularly limited, but is preferably in the range of 3,000 to 50,000. From the viewpoint of improving the bundling ability of the glass fibers, it is preferably 3,000 or more, and from the viewpoint of improving the properties of a composite molded article, it is preferably 50,000 or less. In this embodiment, when polyamide is used as the thermoplastic resin, it is preferable to use a resin as the binder that has good wettability or a surface tension similar to that of the polyamide resin. Specifically, for example, an emulsion of a polyurethane resin, an emulsion of a polyamide resin, or a modified product thereof can be selected.

[0110] The thermoplastic resin (c) used as the binder is not limited to the following, but examples thereof include polyolefin resins, polyamide resins, polyurethane resins, polyacetal resins, polycarbonate resins, polyester resins, polyether ketone, polyether ether ketone, polyether sulfone, polyphenylene sulfide, thermoplastic polyetherimide, thermoplastic fluorine-based resins, and modified thermoplastic resins obtained by modifying these. If the thermoplastic resin (c) used as the binder is the same type of thermoplastic resin (c) and / or modified thermoplastic resin (c) as the resin that coats the continuous reinforcing fibers, the adhesion between the glass fibers and the thermoplastic resin is improved after the composite material is formed, which is preferable.

[0111] Furthermore, from the viewpoint of further improving the adhesion between the continuous reinforcing fibers and the thermoplastic resin coating them, and reducing the proportion of the emulsifier component or eliminating the need for an emulsifier when the bundling agent is attached to the glass fibers as an aqueous dispersion, a modified thermoplastic resin (d) is preferred as the thermoplastic resin (c) used as the bundling agent. Here, the modified thermoplastic resin (d) means a resin obtained by copolymerizing a different monomer component other than the monomer component that can form the main chain of the thermoplastic resin (c) with the purpose of changing the properties of the thermoplastic resin (c), thereby modifying the hydrophilicity, crystallinity, thermodynamic properties, etc. The modified thermoplastic resin (d) used as the binder is not limited to the following, but examples thereof include modified polyolefin resins, modified polyamide resins, modified polyester resins, and the like.

[0112] The modified polyolefin resin used as the binder is a copolymer of an olefin monomer such as ethylene or propylene with a monomer copolymerizable with the olefin monomer, such as an unsaturated carboxylic acid and / or its ester, or a homopolymer of a monomer copolymerizable with the olefin monomer, such as an unsaturated carboxylic acid and / or its ester, and can be produced by a known method. It may be a random copolymer in which an olefin monomer is copolymerized with an unsaturated carboxylic acid and / or its ester, or a graft copolymer in which an unsaturated carboxylic acid is grafted onto an olefin. Examples of the olefin-based monomer include, but are not limited to, ethylene, propylene, 1-butene, etc. These may be used alone or in combination of two or more. Examples of monomers copolymerizable with the olefin-based monomer include unsaturated carboxylic acids such as acrylic acid, maleic acid, maleic anhydride, methacrylic acid, vinylacetic acid, crotonic acid, isocrotonic acid, fumaric acid, itaconic acid, citraconic acid, and mesaconic acid, and esters of these unsaturated carboxylic acids (methyl esters, ethyl esters, and the like). These may be used alone or in combination of two or more. When the modified polyolefin resin is a copolymer of an olefin monomer and a monomer copolymerizable with the olefin monomer, the monomer ratio is preferably 60 to 95% by mass of the olefin monomer and 5 to 40% by mass of the monomer copolymerizable with the olefin monomer, and more preferably 70 to 85% by mass of the olefin monomer and 15 to 30% by mass of the monomer copolymerizable with the olefin monomer, with the total mass of the copolymer being 100% by mass. If the olefin monomer is 60% by mass or more, the affinity with the matrix is ​​good, and if the mass% of the olefin monomer is 95% by mass or less, the water dispersibility of the modified polyolefin resin is good and it is easy to apply it uniformly to the continuous reinforcing fibers.

[0113] The modified polyolefin resin used as the binder may have modified groups, such as carboxyl groups, introduced by copolymerization and neutralized with a basic compound. Examples of basic compounds include, but are not limited to, alkalis such as sodium hydroxide and potassium hydroxide; ammonia; and amines such as monoethanolamine and diethanolamine. The weight-average molecular weight of the modified polyolefin resin used as the binder is not particularly limited, but is preferably 5,000 to 200,000, and more preferably 50,000 to 150,000. From the viewpoint of improving the bundling ability of glass fibers, a molecular weight of 5,000 or more is preferred, and from the viewpoint of emulsion stability when the resin is made water-dispersible, a molecular weight of 200,000 or less is preferred.

[0114] The modified polyamide resin used as the binder is a modified polyamide compound having hydrophilic groups such as polyalkylene oxide chains and tertiary amine components introduced into the molecular chain, and can be produced by known methods. When a polyalkylene oxide chain is introduced into the molecular chain, for example, it is produced by copolymerizing a polyethylene glycol, a polypropylene glycol, or the like, which is partially or completely modified with a diamine or a dicarboxylic acid. When a tertiary amine component is introduced, it is produced by copolymerizing, for example, aminoethylpiperazine, bisaminopropylpiperazine, α-dimethylamino ε-caprolactam, or the like.

[0115] The modified polyester resin used as the binder is a copolymer of a polycarboxylic acid or an anhydride thereof with a polyol, and has hydrophilic groups in the molecular skeleton including the terminals, and can be produced by a known method. Examples of the hydrophilic group include a polyalkylene oxide group, a sulfonate, a carboxyl group, and neutralized salts thereof. Examples of the polycarboxylic acid or anhydride thereof include an aromatic dicarboxylic acid, a sulfonate-containing aromatic dicarboxylic acid, an aliphatic dicarboxylic acid, an alicyclic dicarboxylic acid, and a tri- or higher functional polycarboxylic acid. Examples of the aromatic dicarboxylic acid include, but are not limited to, phthalic acid, terephthalic acid, isophthalic acid, orthophthalic acid, 1,5-naphthalenedicarboxylic acid, 2,6-naphthalenedicarboxylic acid, and phthalic anhydride. Examples of the sulfonate-containing aromatic dicarboxylic acid include, but are not limited to, sulfoterephthalate, 5-sulfoisophthalate, and 5-sulfoorthophthalate. Examples of the aliphatic dicarboxylic acid or alicyclic dicarboxylic acid include, but are not limited to, fumaric acid, maleic acid, itaconic acid, succinic acid, adipic acid, azelaic acid, sebacic acid, dimer acid, 1,4-cyclohexanedicarboxylic acid, succinic anhydride, and maleic anhydride. Examples of the tri- or higher functional polycarboxylic acid include, but are not limited to, trimellitic acid, pyromellitic acid, trimellitic anhydride, and pyromellitic anhydride. Among these, from the viewpoint of improving the heat resistance of the modified polyester resin, it is preferable that 40 to 99 mol% of the total polycarboxylic acid components be aromatic dicarboxylic acids, and from the viewpoint of emulsion stability when the modified polyester resin is made into an aqueous dispersion, it is preferable that 1 to 10 mol% of the total polycarboxylic acid components be sulfonate-containing aromatic dicarboxylic acids.

[0116] Examples of the polyol constituting the modified polyester resin include diols and tri- or higher functional polyols. Examples of the diol include, but are not limited to, ethylene glycol, diethylene glycol, polyethylene glycol, propylene glycol, polypropylene glycol, polybutylene glycol, 1,3-propanediol, 1,4-butanediol, 1,6-hexanediol, neopentyl glycol, polytetramethylene glycol, 1,4-cyclohexanediol, 1,4-cyclohexanedimethanol, bisphenol A or an alkylene oxide adduct thereof, etc. Examples of the tri- or higher functional polyol include trimethylolpropane, glycerin, pentaerythritol, etc.

[0117] The copolymerization ratio of the polycarboxylic acid or its anhydride and the polyol constituting the modified polyester resin is preferably 40 to 60 mass% of the polycarboxylic acid or its anhydride and 40 to 60 mass% of the polyol, and more preferably 45 to 55 mass% of the polycarboxylic acid or its anhydride and 45 to 55 mass% of the polyol, where the total mass of the copolymerization components is 100 mass%. The weight average molecular weight of the modified polyester resin is preferably 3,000 to 100,000, and more preferably 10,000 to 30,000. From the viewpoint of improving the bundling ability of the glass fibers, it is preferably 3,000 or more, and from the viewpoint of emulsion stability when it is made water-dispersible, it is preferably 100,000 or less.

[0118] The polymer or thermoplastic resin used as the binder may be used alone or in combination of two or more kinds. In this embodiment, the total amount of the binder is taken as 100% by mass, and it is preferable to use at least 50% by mass, and more preferably at least 60% by mass, of one or more polymers selected from homopolymers of acrylic acid, copolymers of acrylic acid and other copolymerizable monomers, and salts of these with primary, secondary, and tertiary amines.

[0119] The sizing agent (1) contained in the continuous fiber reinforced resin composite material of this embodiment is preferably contained in an amount of 0.1 to 3 mass %, more preferably 0.2 to 2 mass %, relative to 100 mass % of the continuous reinforcing fibers. When the content of the sizing agent (1) is within the above range, the surface properties of the continuous reinforcing fibers can be effectively modified, thereby improving the adhesion between the continuous reinforcing fibers and the thermoplastic resin. In this embodiment, when the sizing agent (1) is composed of a silane coupling agent and a binder, the sizing agent (1) is applied and attached in an amount of preferably 0.1 to 3 mass%, more preferably 0.2 to 2 mass%, and even more preferably 0.2 to 1 mass%, based on 100 mass% of the glass fibers, as the total mass of the silane coupling agent and the binder. From the viewpoint of controlling the bundling ability of the glass fibers and improving the interfacial adhesive strength, the amount of the sizing agent (1) attached is preferably 0.1 mass% or more, based on 100 mass% of the glass fibers, as the total mass of the silane coupling agent and the binder, and from the viewpoint of yarn handleability, it is preferably 3 mass% or less. In this embodiment, when the sizing agent (1) is composed of a silane coupling agent, a lubricant, and a binder, the sizing agent (1) is applied and attached in an amount of preferably 0.1 to 3 mass%, more preferably 0.2 to 2 mass%, and even more preferably 0.2 to 1 mass%, based on 100 mass% of the glass fibers, as the total mass of the silane coupling agent, lubricant, and binder. From the viewpoint of controlling the bundling ability of the glass fibers and improving the interfacial adhesive strength, the amount of the sizing agent (1) attached is preferably 0.1 mass% or more, based on 100 mass% of the glass fibers, as the total mass of the silane coupling agent, lubricant, and binder, and from the viewpoint of yarn handleability, it is preferably 3 mass% or less.

[0120] <Composition of glass fiber sizing agent (1)> In this embodiment, when glass fibers are selected as the continuous reinforcing fibers, the amount of the silane coupling agent in the sizing agent (1) (hereinafter also referred to as a glass fiber sizing agent) used for the glass fibers is preferably 0.1 to 2 mass %, more preferably 0.1 to 1 mass %, and even more preferably 0.2 to 0.5 mass %, relative to 100 mass % of the glass fibers, from the viewpoints of improving the bundling ability of the glass fibers, improving the interfacial adhesive strength, and improving the mechanical strength of the composite molding. The amount of lubricant in the glass fiber sizing agent is preferably 0.01% by mass or more, more preferably 0.02% by mass or more, relative to 100% by mass of glass fiber, from the viewpoint of providing sufficient lubrication, and is preferably 1% by mass or less, more preferably 0.5% by mass or less, from the viewpoint of improving the interfacial adhesive strength and the mechanical strength of the composite molded product. The amount of binder in the sizing agent for glass fibers is preferably 1 to 25 mass %, more preferably 3 to 15 mass %, and even more preferably 3 to 10 mass %, relative to 100 mass % of glass fibers, from the viewpoints of controlling the bundling properties of the glass fibers, improving the interfacial adhesive strength, and improving the mechanical strength of the composite molded body.

[0121] In the present embodiment, when glass fibers are used as the continuous reinforcing fibers and the glass fiber sizing agent is composed of a silane coupling agent, a lubricant, and a binder, the glass fiber sizing agent preferably contains 0.1 to 2 mass% of the silane coupling agent, 0.01 to 1 mass% of the lubricant, and 1 to 25 mass% of the binder, respectively, and these components are preferably diluted with water to adjust the total mass of the glass fiber sizing agent to 100 mass%.

[0122] <Usage of sizing agent> In the present embodiment, the sizing agent for glass fibers may be prepared in any form, such as an aqueous solution, a colloidal dispersion, or an emulsion using an emulsifier, depending on the mode of use. From the viewpoint of improving the dispersion stability and heat resistance of the sizing agent, however, it is preferably prepared in the form of an aqueous solution. The glass fibers as the continuous reinforcing fibers constituting the continuous fiber reinforced resin composite material of this embodiment can be continuously obtained by applying the above-mentioned sizing agent (1) to glass fibers using a known method such as a roller-type applicator in a known glass fiber manufacturing process, and drying the produced glass fibers.

[0123] Similarly, when carbon fibers are selected as the continuous reinforcing fibers, a sizing agent (1) may be used, and the sizing agent (1) is preferably composed of a coupling agent, a lubricant, and a binder. The coupling agent can be selected to have good compatibility with the hydroxyl groups present on the surface of the carbon fibers, the binder can be selected to have good wettability with the selected thermoplastic resin or a surface tension similar to that of the selected thermoplastic resin, and the lubricant can be selected to not interfere with the coupling agent and the binder. The type of sizing agent used for the carbon fibers is not particularly limited, and any known sizing agent can be used. Specifically, for example, the one described in JP 2015-101794 A can be used.

[0124] When other continuous reinforcing fibers are used, the type and amount of sizing agent that can be used for glass fibers and carbon fibers may be appropriately selected depending on the characteristics of the continuous reinforcing fibers, and it is preferable to use the same type and amount of sizing agent as those used for carbon fibers.

[0125] <Shape of continuous reinforcing fibers> The continuous reinforcing fiber of this embodiment is a multifilament consisting of a plurality of filaments, and the number of single fibers is preferably 30 to 15,000 from the viewpoint of handling. From the viewpoints of strength and ease of handling, the single filament diameter R of the continuous reinforcing fibers is preferably 2 to 30 μm, more preferably 4 to 25 μm, even more preferably 6 to 20 μm, and most preferably 8 to 18 μm. In this embodiment, the product RD of the single filament diameter R (μm) and density D (g / cm3) of the continuous reinforcing fibers is preferably 5 to 100 μm g / cm3, more preferably 10 to 50 μm g / cm3, even more preferably 15 to 45 μm g / cm3, and still more preferably 20 to 45 μm g / cm3, from the viewpoints of ease of handling of the continuous reinforcing fibers and strength of the composite material.

[0126] The density D of the continuous reinforcing fibers of this embodiment can be measured with a specific gravity meter. On the other hand, the single filament diameter R (μm) of the continuous reinforcing fiber of this embodiment is calculated from the density D (g / cm), the fineness (dtex), and the number of single filaments (fibers) by the following formula (2):

number

[0127] In the above formula (2), to set the RD product (=single filament diameter × density) of the continuous reinforcing fiber within a predetermined range, the fineness (dtex) and the number of single filaments (fibers) of commercially available continuous reinforcing fibers can be appropriately selected according to the density of the continuous reinforcing fiber. For example, when glass fiber is used as the continuous reinforcing fiber, the density is approximately 2.5 g / cm3, so a single filament diameter of 2 to 40 μm can be selected. Specifically, when the single filament diameter of the glass fiber is 9 μm, selecting a glass fiber with a fineness of 660 dtex and 400 single filaments results in a RD product of 23. Furthermore, when the single filament diameter of the glass fiber is 17 μm, selecting a glass fiber with a fineness of 11,500 dtex and 2,000 single filaments results in a RD product of 43. When carbon fiber is used as the continuous reinforcing fiber, the density is approximately 1.8 g / cm3, so a single filament diameter of 2.8 to 55 μm can be selected. Specifically, when the carbon fiber has a single filament diameter of 7 μm, selecting a carbon fiber having a fineness of 2,000 dtex and 3,000 filaments results in a product RD of 13. When using aramid fiber as the continuous reinforcing fiber, the density is approximately 1.45 g / cm3, so a single filament diameter of 3.4 to 68 μm should be selected. Specifically, when the aramid fiber has a single filament diameter of 12 μm, selecting an aramid fiber having a fineness of 1,670 dtex and 1,000 filaments results in a product RD of 17.

[0128] Continuous reinforcing fibers, for example, glass fibers, are produced by measuring and mixing raw glass materials, molten glass in a melting furnace, spinning the molten glass into glass filaments, applying a sizing agent (1), passing the sizing agent through a spinning machine, and winding the sizing agent into direct wound rovings (DWR), cakes, twisted yarns, and other wound forms. The continuous reinforcing fibers may be in any form, but are preferably wound into yarn, cake, or DWR, as this increases productivity and production stability in the resin coating process. From the viewpoint of productivity, DWR is the most preferred.

[0129] The form of the continuous reinforcing fibers is not particularly limited, and various forms such as woven fabric, knitted fabric, braided cord, pipe-shaped material, non-crimp fabric, and unidirectional material are possible, with woven fabric, non-crimp fabric, and unidirectional material being preferred.

[0130] (thermoplastic resin) The continuous fiber-reinforced resin composite material of this embodiment contains a thermoplastic resin as an essential component. Examples of the thermoplastic resin include, but are not limited to, polyolefin resins such as polyethylene and polypropylene; polyamide resins such as polyamide 6, polyamide 66, polyamide 46, polyamide 612, polyamide 6I, polyamide 1010, polyamide 12, polyamide 610, polyamide 410, and polyamide 12; polyester resins such as polyethylene terephthalate, polybutylene terephthalate, and polytrimethylene terephthalate; polyacetal resins such as polyoxymethylene; polycarbonate resins; polyether resins such as polyether ketone, polyether ether ketone, polyether glycol, polypropylene glycol, and polytetramethylene ether glycol; polyethersulfone; polyphenylene sulfide; thermoplastic polyetherimide; thermoplastic fluorine-based resins such as tetrafluoroethylene-ethylene copolymer; polyurethane resins; acrylic resins; and modified thermoplastic resins obtained by modifying these resins. In addition, the thermoplastic resin in this embodiment may be used alone or as a mixture of two or more kinds. In this specification, the term "thermoplastic resin" refers to a thermoplastic resin that is a matrix resin (a resin that constitutes the 1 to 200 thermoplastic resin layers) in a laminate of 1 to 100 continuous reinforcing fibers and 1 to 200 thermoplastic resin layers, which constitutes a continuous fiber reinforced resin composite material known as a prepreg. On the other hand, in this specification, the term "thermoplastic resin (c)" refers to a resin used in a sizing agent (1), which is an optional component used for surface modification of continuous reinforcing fibers, etc.

[0131] Among the thermoplastic resins listed above, preferred thermoplastic resins in this embodiment include polyolefin resins, polyamide resins, polyester resins, polyether resins, polyethersulfone, polyphenylene sulfide, thermoplastic polyetherimides, and thermoplastic fluorine-based resins. Polyolefin resins, modified polyolefin resins, polyamide resins, polyester resins, polyurethane resins, and acrylic resins are more preferred from the viewpoints of mechanical properties and versatility, and polyamide resins and polyester resins are even more preferred from the viewpoint of thermal properties. Furthermore, polyamide resins are even more preferred from the viewpoint of durability against repeated loads.

[0132] -Polyester resin- The polyester resin in the thermoplastic resin of this embodiment means a polymer compound having an --CO--O-- (ester) bond in the main chain. Examples of the polyester resin include, but are not limited to, polyethylene terephthalate, polybutylene terephthalate, polytetramethylene terephthalate, poly-1,4-cyclohexylene dimethylene terephthalate, and polyethylene-2,6-naphthalenedicarboxylate. The polyester resin may be a homopolyester or a copolymer polyester, and these polyester resins may be used alone or as a mixture of two or more types. When a copolymer polyester is used as the thermoplastic resin of this embodiment, it is preferable to use a homopolyester copolymerized with an appropriate third component. Examples of the third component include, but are not limited to, diol components such as diethylene glycol, neopentyl glycol, and polyalkylene glycol, and dicarboxylic acid components such as adipic acid, sebacic acid, phthalic acid, isophthalic acid, and 5-sodium sulfoisophthalic acid. In addition, polyester-based resins using raw materials derived from biomass resources can also be used, and examples thereof include, but are not limited to, aliphatic polyester-based resins such as polylactic acid, polybutylene succinate, and polybutylene succinate adipate, and aromatic polyester-based resins such as polybutylene adipate terephthalate.

[0133] -Polyamide resin- The polyamide resin in the thermoplastic resin of this embodiment refers to a polymer compound having an —CO—NH— (amide) bond in the main chain, such as an aliphatic polyamide, an aromatic polyamide, or a wholly aromatic polyamide.

[0134] Examples of the polyamide resin include, but are not limited to, polyamides obtained by ring-opening polymerization of lactams, polyamides obtained by self-condensation of ω-aminocarboxylic acids, polyamides obtained by condensing diamines and dicarboxylic acids, and copolymers thereof. These polyamide resins may be used alone or in combination of two or more. Examples of the lactam include, but are not limited to, pyrrolidone, caprolactam, undecane lactam, and dodecalactam. Examples of the ω-aminocarboxylic acid include, but are not limited to, ω-amino fatty acids, which are compounds obtained by ring-opening lactams with water. Two or more types of lactam or ω-aminocarboxylic acid may be condensed together. Examples of the diamine (monomer) include, but are not limited to, linear aliphatic diamines such as hexamethylenediamine and pentamethylenediamine; branched aliphatic diamines such as 2-methylpentanediamine and 2-ethylhexamethylenediamine; aromatic diamines such as p-phenylenediamine and m-phenylenediamine; and alicyclic diamines such as cyclohexanediamine, cyclopentanediamine, and cyclooctanediamine. Examples of the dicarboxylic acid (monomer) include, but are not limited to, aliphatic dicarboxylic acids such as adipic acid, pimelic acid, and sebacic acid; aromatic dicarboxylic acids such as phthalic acid and isophthalic acid; and alicyclic dicarboxylic acids such as cyclohexanedicarboxylic acid. The diamine and dicarboxylic acid monomers may be condensed either individually or in combination.

[0135] Examples of polyamide-based resins of this embodiment include, but are not limited to, aliphatic polyamides such as polyamide 4 (poly-α-pyrrolidone), polyamide 6 (polycaproamide), polyamide 11 (polyundecaneamide), polyamide 12 (polydodecanamide), polyamide 46 (polytetramethylene adipamide), polyamide 66 (polyhexamethylene adipamide), polyamide 610, and polyamide 612; semi-aromatic polyamides such as polyamide 6T (polyhexamethylene terephthalamide), polyamide 9T (polynonamethylene terephthalamide), and polyamide 6I (polyhexamethylene isophthalamide); and copolymer polyamides containing these as constituent components. Examples of the copolymerized polyamide include, but are not limited to, a copolymer of hexamethylene adipamide and hexamethylene terephthalamide, a copolymer of hexamethylene adipamide and hexamethylene isophthalamide, and a copolymer of hexamethylene terephthalamide and 2-methylpentanediamine terephthalamide.

[0136] In the present embodiment, when a polyamide resin is used as the thermoplastic resin, the thermoplastic resin may contain 50 to 99 parts by mass of (C) an aliphatic polyamide and 1 to 50 parts by mass of (D) a semi-aromatic polyamide, and in this case, the semi-aromatic polyamide (D) preferably contains dicarboxylic acid units containing at least 75 mol % isophthalic acid units and diamine units containing at least 50 mol % diamine units having 4 to 10 carbon atoms. When the thermoplastic resin of this embodiment contains the aliphatic polyamide (C) and the semi-aromatic polyamide (D) within the above ranges, the physical properties (strength, rigidity, high-temperature properties, water absorption properties, impact properties, appearance, etc.) of the continuous fiber reinforced resin composite material tend to be improved compared to when the thermoplastic resin contains only the aliphatic polyamide (C) as the polyamide. In this embodiment, the total content of the (C) aliphatic polyamide and the (D) semi-aromatic polyamide is preferably 70 to 100 mass%, more preferably 80 to 100 mass%, and even more preferably 90 to 100 mass%, relative to 100 mass% of the thermoplastic resin.

[0137] Examples of the aliphatic polyamide (C) include, but are not limited to, polyamide 4, polyamide 6, polyamide 11, polyamide 12, polyamide 46, polyamide 66, polyamide 610, polyamide 612, and the like. Furthermore, the content of the (C) aliphatic polyamide in the thermoplastic resin of this embodiment is preferably 50 to 99 mass%, more preferably 60 to 90 mass%, and even more preferably 70 to 80 mass%, relative to the total amount (100 mass%) of polyamide-based resins contained in the thermoplastic resin.

[0138] Examples of the semi-aromatic polyamide (D) include, but are not limited to, polyamide 6I, polyamide 9I, polyamide 10I, and the like. The total amount of the isophthalic acid units and the diamine units having 4 to 10 carbon atoms is preferably 80 to 100 mol %, more preferably 90 to 100 mol %, and even more preferably 95 to 100 mol %, relative to 100 mol % of all structural units of the (D) semi-aromatic polyamide. The proportion of the monomer units constituting the (D) semi-aromatic polyamide can be measured, for example, by 13C nuclear magnetic resonance spectroscopy (NMR).

[0139] In the (D) semi-aromatic polyamide, the proportion of isophthalic acid units in the dicarboxylic acid units is at least 75 mol%, preferably 85 mol% or more, and more preferably 90 mol% or more. When the proportion of isophthalic acid units in the dicarboxylic acid units is within the above range, high-temperature properties and water absorption properties tend to be improved.

[0140] In the (D) semi-aromatic polyamide, the proportion of diamine units having 4 to 10 carbon atoms in the diamine units is at least 50 mol %, preferably 60 mol % or more, and more preferably 70 mol % or more. When the proportion of diamine units having 4 to 10 carbon atoms in the diamine units is within the above range, high-temperature properties and water absorption properties tend to be improved.

[0141] The content of the (D) semi-aromatic polyamide in 100% by mass of polyamide in the thermoplastic resin is preferably 1 to 50% by mass, more preferably 10 to 40% by mass, and even more preferably 20 to 30% by mass.

[0142] The (C) aliphatic polyamide and (D) semi-aromatic polyamide may be end-capped with a known end-capping agent, and the total amount of blocked ends of the (C) aliphatic polyamide and (D) semi-aromatic polyamide, expressed as an equivalent weight per gram of polyamide (C) and (D) combined, is preferably 5 to 180 μequivalents / g, more preferably 10 to 170 μequivalents / g, even more preferably 20 to 160 μequivalents / g, particularly preferably 30 to 140 μequivalents / g, and most preferably 40 to 140 μequivalents / g. When the amount of blocked ends is within the above range, physical properties (strength, rigidity, high-temperature properties, water absorption properties, impact properties, appearance, etc.) tend to be improved. Here, the amount of blocked terminals refers to the total amount of amino terminals and carboxyl terminals blocked with a blocking agent. The amount of blocked terminals can be measured using H-NMR, specifically, by the method described in the Examples below.

[0143] The terminal group concentration of the aliphatic polyamide (C) is preferably 1 / 2 or less, more preferably 2 / 5 or less, of the terminal group concentration of the semi-aromatic polyamide (D). When the terminal group concentration of the aliphatic polyamide (C) is 1 / 2 or less of the terminal group concentration of the semi-aromatic polyamide (D), physical properties (strength, rigidity, high-temperature properties, water absorption properties, impact properties, appearance, etc.) tend to be improved. The terminal group concentrations of the (C) aliphatic polyamide and the (D) semi-aromatic polyamide can be measured using 1H-NMR, specifically by the method described in the examples below.

[0144] The difference in peak temperature of tan δ between the aliphatic polyamide (C) and the semi-aromatic polyamide (D) is preferably 45 to 100° C., more preferably 50 to 90° C., and even more preferably 60 to 90° C. When the difference in peak temperature of tan δ between the aliphatic polyamide (C) and the semi-aromatic polyamide (D) is within the above range, high-temperature properties and water absorption properties tend to be improved. The peak temperatures of tan δ of the (C) aliphatic polyamide and the (D) semi-aromatic polyamide can be measured, for example, using a viscoelasticity measurement analyzer, specifically, by the method described in the examples below.

[0145] From the viewpoints of strength, rigidity, moldability, and appearance, the difference in viscosity between (C) the aliphatic polyamide and (D) the semi-aromatic polyamide is preferably 3 times or more, and more preferably 4 times or more. The viscosity of a thermoplastic resin can be determined by MFR measurement (in accordance with ISO1133), and specifically, can be observed by the method described in the examples below. The content of the thermoplastic resin in this embodiment is preferably 16 to 53 mass %, more preferably 23 to 40 mass %, and even more preferably 25 to 36 mass %, relative to 100 mass % of the continuous fiber reinforced resin composite material.

[0146] [Additives] The continuous fiber-reinforced resin composite material of this embodiment may contain additives as needed, such as colorants, antioxidants, antioxidants, weathering agents, metal deactivators, light stabilizers, heat stabilizers, UV absorbers, antibacterial and antifungal agents, deodorizers, conductivity-imparting agents, dispersants, softeners, plasticizers, crosslinking agents, co-crosslinking agents, vulcanizing agents, vulcanization aids, foaming agents, foaming aids, flame retardants, vibration dampers, nucleating agents, neutralizing agents, lubricants, antiblocking agents, dispersants, flow improvers, and mold release agents. The content of the additive in this embodiment may be 3% by mass or less relative to 100% by mass of the continuous fiber reinforced resin composite material.

[0147] <Coloring agent> Examples of the colorant include carbon black, nigrosine, aluminum pigment, titanium dioxide, ultramarine, cyanine blue, cyanine green, quinacridone, diatomaceous earth, monoazo salt, perylene, disazo, condensed azo, isoindoline, red iron oxide, nickel titanium yellow, diketone pyrrolopyrrole, metal salt, perylene red, metal oxide, bismuth vanadate, cobalt green, cobalt blue, anthraquinone, phthalocyanine green, phthalocyanine blue, etc. Among these, black colorants are preferred, and carbon black and nigrosine are more preferred.

[0148] The continuous fiber reinforced resin composite material of this embodiment preferably has a continuous reinforcing fiber content of 90 to 525 parts by mass and a content of other components (e.g., sizing agent (1), additives) of 0 to 2 parts by mass per 100 parts by mass of thermoplastic resin, and more preferably has a continuous reinforcing fiber content of 150 to 340 parts by mass and a content of other components of 0 to 1 part by mass per 100 parts by mass of thermoplastic resin.

[0149] [Applications of continuous fiber reinforced resin composite materials] The continuous fiber reinforced resin composite material of this embodiment can be suitably used as a structural material for aircraft, cars, construction materials, robots, and the like. In automotive applications, the material can be used for, but not limited to, chassis / frames, undercarriage, drivetrain parts, interior parts, exterior parts, functional parts, and other parts. Specifically, steering shafts, mounts, sunroofs, steps, suspension trim, door trim, trunks, boot lids, bonnets, seat frames, seat backs, retractors, retractor support brackets, clutches, gears, pulleys, cams, argon, elastic beams, baffling, lamps, reflectors, glazing, front end modules, back door inners, brake pedals, steering wheels, electrical materials, sound absorbing materials, door exteriors, interior panels, instrument panels, rear gates, ceiling sills, seats, seat frames, wiper posts, EPS (Electric Power Steering), small motors, heat sinks, ECU (Engine Control Unit) boxes, ECU housings, steering gear box housings, plastic housings, EV (Electric Vehicle motor housings, wire harnesses, on-board meters, combination switches, small motors, springs, dampers, wheels, wheel covers, frames, subframes, side frames, motorcycle frames, fuel tanks, oil pans, intake manifolds, propeller shafts, drive motors, monocoques, hydrogen tanks, fuel cell electrodes, panels, floor panels, exterior panels, doors, cabins, roofs, hoods, valves, EGR (Exhaust GasRecirculation valves, variable valve timing units, connecting rods, cylinder bores, members (engine mountings, front floor cloth, footwell cloth, seat cloth, inner side, rear cloth, suspension, pillar reinforcement, front side, front panel, upper, dash panel cloth, steering), tunnels, fastening inserts, crash boxes, crash rails, corrugated roof rails, upper body, side rails, braiding, door surround assemblies, airbag components, body pillars, dash-to-pillar gussets, suspension towers, bumpers, body pillar lowers, front body pillars, reinforcements (instrument panels, rails, roofs, front body pillars, roof rails, roof side rails, rockers, door belt lines, front floor unders, front body pillar uppers, front body pillar lowers, center pillars, center pillar hinges, door outside panels), side outer panels, front door window frames, MICS (Minimum Intrusion CabinSystem bulk, torque box, radiator support, radiator fan, water pump, fuel pump, electronically controlled throttle body, engine control ECU, starter, alternator, manifold, transmission, clutch, dash panel, dash panel insulator pad, door side impact protection beam, bumper beam, door beam, bulkhead, outer pad, inner pad, rear seat rod, door panel, door trim board sub-assembly, energy absorber (bumper, impact absorption), impact absorber, impact absorption garnish, pillar garnish, roof side inner garnish, resin rib, side rail front spacer, side rail rear spacer, seat belt pretensioner, airbag They can be suitably used as parts such as sensors, arms (suspension, lower, hood hinge), suspension links, shock absorbing brackets, fender brackets, inverter brackets, inverter modules, hood inner panels, hood panels, cowl louvers, cowl top outer front panels, cowl top outer panels, floor silencers, dump seats, hood insulators, fender side panel protectors, cowl insulators, cowl top ventilator loopers, cylinder head covers, tire deflectors, fender supports, strut tower bars, transmission center tunnels, floor tunnels, radio core supports, luggage panels, luggage floors, accelerator pedals, and accelerator pedal bases.

[0150] [Molding of continuous fiber reinforced plastic composite materials] The continuous fiber reinforced resin composite material of this embodiment can be further molded. Examples of the molding method include a method in which the continuous fiber reinforced resin composite material of this embodiment is cut into a predetermined size, heated with an infrared heater, and then heated and compressed in a press molding machine.

[0151] [Hybrid composite materials] The continuous fiber reinforced resin composite material of this embodiment may be used as a hybrid composite material. The hybrid composite material is preferably configured by injecting a discontinuous reinforcing material into the continuous fiber reinforced resin composite material of this embodiment to integrate it. More specifically, a hybrid composite material may be produced by further injecting and filling a hybrid thermoplastic resin composition into a continuous fiber reinforced resin composite material containing continuous reinforcing fibers and a thermoplastic resin. In the manufacturing process for the hybrid composite material, a substrate used for the above-described continuous fiber reinforced resin composite material (e.g., a substrate made of continuous reinforcing fibers, a substrate made of a thermoplastic resin, or a substrate made of a continuous fiber reinforced resin composite material) is placed in a mold, the mold is closed, pressurized, and after a predetermined time, a predetermined hybrid thermoplastic resin composition is further injected and filled to mold the substrate, thereby bonding the thermoplastic resin of the substrate and the predetermined hybrid thermoplastic resin composition, thereby producing a hybrid composite material.

[0152] The timing of injection and filling of a predetermined thermoplastic resin composition for hybrid use greatly affects the interfacial strength between the two thermoplastic resins (between the thermoplastic resin in the continuous fiber reinforced resin composite material and the thermoplastic resin in the thermoplastic resin composition for hybrid use). The timing of injection and filling of a predetermined thermoplastic resin composition for hybrid use is preferably within 30 seconds after the substrate used in the continuous fiber reinforced resin composite material is set in a mold, the mold is closed, and the mold temperature is raised to above the melting point or glass transition temperature of the thermoplastic resin constituting the substrate. The mold temperature when injecting and filling a predetermined thermoplastic resin composition for hybrid use is preferably equal to or higher than the melting point or glass transition temperature of the thermoplastic resin constituting the substrate used in the continuous fiber reinforced resin composite material to be bonded with the thermoplastic resin composition for hybrid use, more preferably equal to or higher than the melting point or glass transition temperature of the thermoplastic resin constituting the substrate used in the continuous fiber reinforced resin composite material to be bonded with the thermoplastic resin composition for hybrid use +10°C, even more preferably equal to or higher than the melting point or glass transition temperature +20°C, and even more preferably equal to or higher than the melting point or glass transition temperature +30°C.

[0153] In the hybrid composite material of this embodiment, it is preferable that the joint between the thermoplastic resin constituting the substrate used in the continuous fiber reinforced resin composite material and the thermoplastic resin composition for hybrid formed by injection molding has an uneven structure in which they are mixed with each other. In order to enhance the interfacial strength, it is effective to set the mold temperature at or above the melting point of the thermoplastic resin composition for hybrid to be injected and to set the resin dwell pressure during injection molding at a high level, for example, at 1 MPa or higher. To enhance the interfacial strength, the dwell pressure is preferably set at 5 MPa or higher, and more preferably at 10 MPa or higher. Furthermore, from the viewpoint of increasing interfacial strength, it is preferable to maintain the pressure for a long time, for example, 5 seconds or more, preferably 10 seconds or more, and more preferably for a time until the mold temperature falls below the melting point of the thermoplastic resin composition for hybrid use.

[0154] (Thermoplastic resin composition for hybrids) The thermoplastic resin composition for hybrid use for injection molding used to produce the above hybrid composite material is not particularly limited as long as it is a thermoplastic resin composition that is generally used for injection molding. In this embodiment, the thermoplastic resin contained in the thermoplastic resin composition for hybrids is not limited to the following, but examples include polyethylene, polypropylene, polyvinyl chloride, acrylic resin, styrene-based resin, polyethylene terephthalate, polybutylene terephthalate, polyarylate, polyphenylene ether, modified polyphenylene ether resin, wholly aromatic polyester, polyacetal, polycarbonate, polyetherimide, polyethersulfone, polyamide-based resin, polysulfone, polyetheretherketone, and a mixture of two or more thermoplastic resins such as polyetherketone.

[0155] The thermoplastic resin composition for hybrid use may contain various fillers. The thermoplastic resin composition for hybrid use may be a black resin composition containing a colorant. Examples of various fillers include short fiber and long fiber materials, which are discontinuous reinforcing materials of the same type as the above-mentioned continuous reinforcing fibers. In this specification, discontinuous reinforcing materials refer to fibers having a number average fiber length of 3 mm or less as measured by the following measurement method, while continuous reinforcing fibers refer to fibers having a number average fiber length of more than 3 mm as measured by the following measurement method. Test pieces to be measured were cut out from the thermoplastic resin composition for hybrids or the hybrid composite material, and heated in air at 650°C for 180 minutes in an electric furnace to thoroughly burn off the resin components and separate only the discontinuous reinforcing material. After that, 10 fibers were randomly selected, and the lengths of the 10 fibers were measured using an optical microscope. The number-average fiber length was calculated using the following formula. Number average fiber length = (ΣLi) / 10 Li: length of each fiber (mm) When short glass fibers or long glass fibers are used as the discontinuous reinforcing material, the same sizing agent (1) as that applied to the continuous reinforcing fibers constituting the continuous fiber-reinforced resin composite material of this embodiment may be used. The sizing agent (1) preferably comprises a silane coupling agent, a lubricant, and a binder. The types of silane coupling agent, lubricant, and binder that can be used are the same as those used for the sizing agent (1) for the continuous reinforcing fibers.

[0156] From the viewpoint of interfacial strength with the thermoplastic resin to be joined, the thermoplastic resin contained in the thermoplastic resin composition for hybrid use used in injection molding is preferably similar to, and more preferably the same type as, the thermoplastic resin at the joining surfaces constituting the continuous fiber reinforced resin composite material. Specifically, when polyamide 66 is used as the thermoplastic resin at the joining surfaces, the resin material of the thermoplastic resin composition for hybrid use used in injection molding is preferably polyamide 66.

[0157] Other methods include a molding method in which the substrate used for the continuous fiber reinforced resin composite material is placed in a mold and compressed using a double belt press; a method in which a mold frame is placed so as to surround all four sides of the placed substrate used for the placed continuous fiber reinforced resin composite material, and then pressurized and molded using a double belt press; and a molding method in which a heating compression molding machine set to one or more temperatures and a cooling compression molding machine set to one or more temperatures are prepared, and molds in which the substrate used for the continuous fiber reinforced resin composite material is placed are placed in the compression molding machine in order to mold the material. [Example]

[0158] The present invention will be specifically explained below with reference to examples and comparative examples, but it goes without saying that the present invention is not limited to these examples and can be practiced in various modified forms within the scope of the gist of the present invention. [Bending strength recovery rate and bending modulus recovery rate, recovery rate of continuous fiber reinforced resin composite material] (Measurement of flexural strength recovery rate and flexural modulus recovery rate) Before conducting the bending test, rectangular test pieces (A1) measuring 100 mm in length, 10 mm in width, and 2 mm in thickness were cut from the continuous fiber-reinforced resin composite materials prepared in the Examples and Comparative Examples and dried in a vacuum dryer at 80°C for at least 18 hours. Then, a bending test, an example of a load test, was conducted using an Instron universal testing machine with a three-point bending jig, with a span set to thickness x 16 (mm) at a speed of 1 mm / min under conditions of 23°C and 50% RH and 100°C and 50% RH, and the bending strength (MPa) and bending modulus (GPa) were measured. Then, the test piece (A1) after the bending test was recovered by the recovery process described in each example to produce a test piece (B1), and then a bending test was performed on the test piece (B1) of the recovered continuous fiber reinforced resin composite material. The obtained bending strength or bending modulus was divided by the bending strength or bending modulus of the test piece (A1) of the continuous fiber reinforced resin composite material before the load test, and the values ​​obtained were defined as the bending strength recovery rate and the bending modulus recovery rate. Bending strength recovery rate (%) = bending strength of test piece (B1) / bending strength of test piece (A1) × 100 Flexural modulus recovery rate (%) = flexural modulus of test piece (B1) / flexural modulus of test piece (A1) × 100

[0159] (Calculation method for recovery rate) The continuous fiber reinforced resin molded body before and after the load test was cut with a band saw, and the cut pieces were polished with a polishing machine (small precision specimen preparation system IS-POLISHER ISPP-1000 (Ikegami Seiki Co., Ltd.)) with a polishing pressure of 416 g / cm2 on the polished surface, and approximately 7 mL of water was polished in the following order: 10 minutes with waterproof sandpaper #2000, 5 minutes with silicon carbide film with a grain size of 9 μm, 5 minutes with alumina film with a grain size of 5 μm, 5 minutes with alumina film with a grain size of 3 μm, 5 minutes with alumina film with a grain size of 1 μm, and 5 minutes with colloidal silica (Baicalox 0.1CR) with a grain size of 0.1 μm using polyurethane foam buffing paper for 5 minutes. The test pieces were polished at a rate of 1 / min while adding water to prepare test pieces (A2) before the load test and test pieces (B2) after the load test. The polished surfaces of the test pieces (A2) and (B2) were then observed under the following conditions using a field emission scanning electron microscope (FESEM (S-4700, Hitachi High-Technologies Corporation)). The number of continuous reinforcing fibers out of any 20 continuous reinforcing fibers that did not have a gap between one continuous reinforcing fiber and the thermoplastic resin coating that one continuous reinforcing fiber was determined. The method for determining whether or not a gap exists between a single continuous reinforcing fiber and the thermoplastic resin coating that single continuous reinforcing fiber is as follows: according to the procedures shown in Figures 14 and 15 described above, if the void ratio in the outer peripheral region of the thermoplastic resin, located one-tenth of the radius (r1) of one continuous reinforcing fiber away from the interface between the continuous reinforcing fiber and the thermoplastic resin, is 10% or less, it is determined that no gap exists between a single continuous reinforcing fiber and the thermoplastic resin coating that single continuous reinforcing fiber. The void ratio (%) was measured by using the obtained field emission scanning electron microscope image to determine which areas appeared black in the outer peripheral region of the thermoplastic resin, which was one-tenth the radius (r1) of one continuous reinforcing fiber away from the interface between the continuous reinforcing fiber and the thermoplastic resin, as voids, and then analyzing the image to calculate the void ratio (%) as the ratio of the void area to the total area of ​​the outer peripheral region of the resin. More specifically, a test piece (Aa) was cut out from a continuous fiber reinforced resin composite material before the load test described below and polished under the above conditions. Of 20 of the continuous reinforcing fibers present in an image (magnification: 4000 to 4500 times) of 50 fields photographed with a field emission scanning electron microscope, the number A of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin was measured. Next, the continuous fiber reinforced resin composite material after the load test described below is subjected to a recovery process described below to produce the recovered continuous fiber reinforced resin composite material, and then a test piece (Bb) is cut out from the recovered continuous fiber reinforced resin composite material and polished under the conditions described above. Of 20 arbitrary continuous reinforcing fibers present in an image (magnification 4000 to 4500 times) taken from 50 fields of view using a field emission scanning electron microscope, the number B of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin is measured. And the following equation (1): Recovery rate (%)=(B / A)×100 The recovery rate (%) of the continuous fiber reinforced resin composite material is calculated by substituting the measured numbers A and B into the above. The conditions for calculating the recovery rate for each load test are explained in detail below.

[0160] <Calculation of recovery rate by bending test> A test piece (A1) was cut out from the continuous fiber reinforced resin composite material before the bending test and polished under the conditions described above. Of 20 of the continuous reinforcing fibers present in an image (magnification: 4000 to 4500 times) of 50 fields of view photographed with a field emission scanning electron microscope, the number Ai of continuous reinforcing fibers that did not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin was measured. Next, the continuous fiber reinforced resin composite material after the bending test was subjected to the recovery process described in Examples and Comparative Examples below to produce the recovered continuous fiber reinforced resin composite material, and then a thickness test piece (B1) was cut out from the portion where the bending test was performed. Images (magnification: 4000 to 4500 times) of 50 fields of view of the thickness test piece (B1) were taken with a field emission scanning electron microscope. Of 20 arbitrary continuous reinforcing fibers present in the images, the number Bi of continuous reinforcing fibers that did not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin was measured, and Bi / Ai × 100 was calculated from the above formula (1) to calculate the recovery rate (%) after the bending test.

[0161] <Calculation of recovery rate by impact test> A test piece (A2) was cut out from the continuous fiber reinforced resin composite material before the impact test and polished under the above conditions. 50 fields of view were photographed using a field emission scanning electron microscope. Of the 20 continuous reinforcing fibers present in the image (magnification: 4000 to 4500 times), the number Aii of continuous reinforcing fibers that did not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin was measured. Next, a test piece (A2) measuring 150 mm in length, 100 mm in width, and 2 mm in thickness was cut from the continuous fiber-reinforced resin composite. The cut test piece (A2) was fixed in a stainless steel jig with a 30 mm diameter hole, and a steel striker with a hemispherical tip (16 mm in diameter) was dropped freely onto the center of the test piece to perform an impact test. After the impact test, the test piece (A2) was subjected to the recovery process described in each Example and Comparative Example below. The impact test site was then cut out and polished using a polishing machine (a small precision specimen preparation system, IS-POLISHER ISPP-1000, Ikegami Seiki Co., Ltd.) under the above conditions to produce test piece (B2). The polished surface of test piece (B2) after the impact test was then observed using a field emission scanning electron microscope (FESEM (S-4700, Hitachi High-Technologies Corporation)). Thus, among any 20 continuous reinforcing fibers present in the image of the test piece (B2) (magnification 4000 to 4500 times), the number Bii of continuous reinforcing fibers that had no voids at the interface between the continuous reinforcing fibers and the thermoplastic resin was measured, and the recovery rate (%) from the impact test was calculated by calculating Bii / Aii × 100 from the above formula (1).

[0162] [Long-term characteristic recovery rate] In accordance with ASTM-D1822, tensile impact dumbbell Type S specimens (C) were prepared from the continuous fiber-reinforced resin composites produced in the Examples and Comparative Examples. A first vibration fatigue test was conducted using an EHF-EB50kN-40L(RV) (Shimadzu Corporation) at a test temperature of 23°C, a frequency of 20 Hz, a sinusoidal waveform, and a chuck distance of 35 mm. The first vibration fatigue test was stopped when the number of cycles to failure reached 10,000 at a load stress of 250 MPa. The specimens (C) were then recovered using the recovery method described in each Example. A second vibration fatigue test was then conducted under the same conditions as the first vibration fatigue test, and the number of cycles to failure (NB2) at a load stress of 250 MPa was calculated. A separate specimen (C) was also prepared, and the number of cycles to failure (NB1) at 250 MPa was calculated for this specimen (C) under the same conditions as the second vibration fatigue test. Thereafter, the number of times to break (NB2) was divided by the number of times to break (NB1) × 100, and the long-term property recovery rate (%) was calculated. Long-term characteristic recovery rate (%) = (number of breaks (NB2) / number of breaks (NB1)) x 100

[0163] [Interfacial polishing strength of continuous fiber reinforced resin composites] The continuous fiber-reinforced resin molded articles prepared in the Examples and Comparative Examples were cut using a band saw. The cut test pieces (D) were polished using a polishing machine (small precision specimen preparation system IS-POLISHER ISPP-1000 (Ikegami Seiki Co., Ltd.)) with a polishing pressure increasing in 20 g / cm2 increments from 200 g / cm2 under the following conditions. The polished surface was observed using a field emission scanning electron microscope (FESEM (S-4700, Hitachi High-Technologies Corporation)) to determine the maximum polishing pressure required to prevent voids from forming between each continuous reinforcing fiber and the thermoplastic resin. This maximum polishing pressure was determined just before the polishing pressure required to detect voids between each continuous reinforcing fiber and the thermoplastic resin. For example, if voids were observed after applying a polishing pressure of 520 g / cm2, the maximum polishing pressure was 500 g / cm2. The polishing conditions were: 10 minutes with #2000 waterproof paper, 5 minutes with 9 μm silicon carbide film, 5 minutes with 5 μm alumina film, 5 minutes with 3 μm alumina film, 5 minutes with 1 μm alumina film, and 5 minutes with 0.1 μm colloidal silica (Baicalox 0.1CR) using polyurethane foam buffing paper. Each polishing was performed with water added at approximately 7 mL / min.

[0164] [Handling of continuous fiber reinforced resin composite materials at high temperatures] Test pieces (E) of 50 mm x 50 mm were cut out from the continuous fiber reinforced resin composite materials prepared in the Examples and Comparative Examples using a band saw. These were then heated to the melting temperature of the thermoplastic resin used in each Example using an infrared heater (Infrastein H7GS-71298NGK, NGK Insulators, wavelength 3 to 7 μm). Test pieces with a thickness change before and after heating ((thickness of continuous fiber reinforced resin composite material after heating) / (thickness of continuous fiber reinforced resin composite material before heating)) of 1.2 or less were evaluated as "good," and those with a change of more than 1.2 were evaluated as "poor." The reason why the change in thickness of the continuous fiber reinforced resin composite material was measured as an indicator of the handleability of the continuous fiber reinforced resin composite material at high temperatures is that continuous fiber reinforced resin composite materials are generally produced by applying pressure, and therefore residual stress remains in the continuous fiber reinforced resin composite material itself. Therefore, when the continuous fiber reinforced resin composite material is heated, the residual stress is released, causing the resin component in the continuous fiber reinforced resin composite material to expand and lose its shape, making the continuous fiber reinforced resin composite material difficult to handle. Furthermore, as one of the handleability characteristics of the continuous fiber reinforced resin composite material at high temperatures, the maintenance of the shape of the test piece (E) was confirmed according to the following criteria.

[0165] <Criteria for maintaining shape> Test pieces (E) measuring 300 mm x 300 mm were cut from the continuous fiber-reinforced resin composite materials prepared in the Examples and Comparative Examples using a band saw. These were heated to the melting point of the thermoplastic resin used in each Example and Comparative Example + 30°C using an infrared heater (Infrastein H7GS-71298NGK, NGK, wavelength 3-7 μm). They were then placed in a box-shaped mold with a temperature controlled to the glass transition temperature of the resin + 100°C, with a long side of 200 mm, a bottom of 200 mm, and sides of 50 mm, with a 60° angle between the top and sides. They were pressed under a load of 15 MPa for 1 minute. The resin was then removed from the resulting molded product in an electric furnace. If the angle between the bottom and side of the continuous reinforcing fiber was 45° to 70°, the product was evaluated as "good," and if not, it was evaluated as "bad."

[0166] The materials used in the examples and comparative examples are as follows. [Continuous reinforcing fiber] Glass fiber 1 (GF1): 100% by mass of glass fiber with a fineness of 1.15 g / m and 2,000 single fibers was produced by adhering 0.30% by mass of sizing agent (1). The winding method was DWR, and the average single fiber diameter was approximately 18 μm. The sizing agent (1) was prepared by mixing with deionized water a copolymer compound having a weight-average molecular weight of 20,000, which was obtained by copolymerizing 2.0% by mass of γ-aminopropyltriethoxysilane (Tokyo Chemical Industry Co., Ltd.), 1% by mass of polyethylene wax, 2% by mass of non-yellowing urethane emulsion (Toyo Polymer Co., Ltd.), and 10% by mass of methyl methacrylate. Glass fiber 2 (GF2): GF1 was treated in an electric furnace at 600°C for 3 hours to produce glass fiber 2 from which the surface treatment agent (sizing agent (1)) had been removed. Glass fiber 3 (GF3): Glass fiber 2 was immersed in a 0.5% by mass aqueous solution of a coupling agent (γ-aminopropyltriethoxysilane (KBE-903, manufactured by Shin-Etsu Chemical Co., Ltd.)) for 30 minutes and then dried at 110°C for 3 hours to prepare glass fiber 3 (GF3) treated only with the coupling agent.

[0167] [Preparation of continuous reinforcing fiber substrate] Glass cloth: A continuous fiber-reinforced substrate was produced by weaving the above glass fibers 1 to 3 as warp and weft yarns using a rapier loom (weaving width 1 m). The weaving form of the obtained glass cloth was (plain weave, weaving density 6.5 threads / 25 mm, basis weight 640 g / m2). The glass cloth produced using glass fiber 1 (GF1) was called glass cloth 1 (or GC1), the glass cloth produced using glass fiber 2 (GF2) was called glass cloth 2 (or GC2), and the glass cloth produced using glass fiber 3 (GF3) was called glass cloth 3 (or GC3).

[0168] [Thermoplastic resin] Resin 1: Polyamide 66 The polymerization reaction of polyamide was carried out by the "hot melt polymerization method" as follows. 1500 g of an equimolar salt of adipic acid (Wako Pure Chemical Industries) and hexamethylenediamine (Tokyo Chemical Industry Co., Ltd.) and 200 g of adipic acid were dissolved in 1500 g of distilled water to prepare a homogeneous aqueous solution containing 50% by weight of the raw material monomers. This aqueous solution was charged into a 6.2 L autoclave and purged with nitrogen. The solution was concentrated to a solution concentration of 65% by weight by gradually removing water vapor while stirring at a temperature of 110°C to 140°C. The internal temperature was then raised to 230°C. The autoclave was then pressurized to 1.9 MPa. The reaction was continued for 1 hour while maintaining the pressure at 1.9 MPa by gradually removing water vapor until the internal temperature reached 235°C. The pressure was then reduced over 1.5 hours. The autoclave was then maintained at a reduced pressure of 650 torr for 13 minutes using a vacuum device. The final internal temperature of the polymerization was 265°C. The mixture was then pressurized with nitrogen and formed into strands from the lower spinneret (nozzle), cooled with water, cut, and discharged as pellets. These were then dried at 100°C in a nitrogen atmosphere for 12 hours to obtain Resin 1 (Polyamide 66). Mw = 35,000, Mw / Mn = 2.0, melting point (Tm) = 265°C, and glass transition temperature (Tg) = 50°C.

[0169] Resin 2: Polyamide 6I The polymerization reaction of polyamide was carried out by the "hot melt polymerization method" as follows. 1500 g of an equimolar salt of isophthalic acid (Wako Pure Chemical Industries) and hexamethylenediamine, and 4.0 mol % of isophthalic acid based on the total equimolar salt components, were dissolved in 1500 g of distilled water. The solution was concentrated to a solution concentration of 65% by mass by gradually removing water vapor while stirring at a temperature of 110°C to 140°C. The internal temperature was then raised to 230°C. The autoclave was then pressurized to 1.9 MPa. The reaction was continued for 1.2 hours, with water vapor gradually removed until the internal temperature reached 235°C. The pressure was then gradually removed over 30 minutes. The autoclave was then maintained at a reduced pressure of 650 torr for 12 minutes using a vacuum device. The final internal temperature of the polymerization was 265°C. The mixture was then pressurized with nitrogen and formed into strands from the lower spinneret (nozzle), cooled with water, cut, and discharged as pellets. The pellets were dried at 100°C for 12 hours in a nitrogen atmosphere to obtain Resin 2 (Polyamide 6I). Mw = 20,000, Mw / Mn = 2.0, Tg = 130°C.

[0170] Resin 3: Dry blend of Resin 1 and Resin 2 (PA66:PA6I=2:1) ​​Tg=110℃

[0171] Resin 4: Polyamide 6 (Tm = 225°C, Tg = 50°C) A mixed solution (4) was prepared by dissolving 1,500 g of ε-caprolactam and 15 g of hexamethylenediamine in 1,875 g of distilled water. The mixed solution (4) was sealed in an autoclave equipped with a pressure control valve and a stirring blade. After purging with nitrogen, the autoclave's internal pressure was increased to 0.05 MPa with nitrogen. While stirring the mixed solution (4) at a temperature between 110°C and 140°C, water vapor was gradually removed to concentrate the solution to a concentration of 65% by mass. The internal temperature of the autoclave was then increased to 230°C. The autoclave's pressure was then increased to 1.8 MPa. The reaction was continued for 1 hour while gradually removing water vapor to maintain the pressure at 1.8 MPa until the internal temperature reached 235°C. The pressure was then reduced over 1.5 hours. The autoclave was then maintained at a reduced pressure of 650 torr for 13 minutes using a vacuum device. The final internal temperature of the polymerization was 265°C. The mixture was then pressurized with nitrogen and formed into strands from the lower spinneret (nozzle), cooled with water, cut, and discharged as pellets. These were then dried at 100°C in a nitrogen atmosphere for 12 hours to obtain Resin 4 (Polyamide 6). Mw = 24,000, Mw / Mn = 2.1, melting point (Tm) = 225°C, and glass transition temperature (Tg) = 50°C.

[0172] Resin 5: Polyamide 9T (Tm = 300°C, Tg = 125°C) A mixed solution (5) was prepared by dissolving 1500 g of an equimolar salt of terephthalic acid and 1,9-nonanediamine in 1500 g of distilled water. While stirring the mixed solution (5) at a temperature between 110°C and 140°C, the water vapor was gradually removed to concentrate the solution to a concentration of 65% by mass. The internal temperature was then raised to 260°C. The autoclave was then pressurized to 2.1 MPa. The reaction was continued for 1.2 hours, with the pressure maintained at 2.1 MPa by gradually removing water vapor, until the internal temperature reached 265°C. The pressure was then reduced over 40 minutes. The autoclave was then maintained at a reduced pressure of 650 torr for 20 minutes using a vacuum device. The final internal temperature of the polymerization was 300°C. The mixture was then pressurized with nitrogen and formed into strands from the lower spinneret (nozzle), cooled with water, cut, and discharged as pellets. The pellets were dried at 100°C for 12 hours in a nitrogen atmosphere to obtain Resin 5 (Polyamide 9T). Mw = 26,000, Mw / Mn = 2.0, Tg = 125°C.

[0173] Resin 6: Polyamide 12 (Tm = 175°C, Tg = 50°C) Resin 7: Polyamide 1010 (Daicel-Evonik Co., Ltd., Tm = 202 °C, Tg = 37 °C)

[0174] [Preparation of thermoplastic resin film] Using a T-die extrusion molding machine (manufactured by Soken Co., Ltd.), each of the above resins 1 to 7 was molded to obtain a thermoplastic resin film, each of which had a thickness of 200 μm.

[0175] [Example 1] Thermoplastic resin film 1 was obtained by the above method using resin 1. Next, five sheets of glass cloth 1 (GC1) and six sheets of thermoplastic resin film 1 were prepared, and the glass cloth 1 and thermoplastic resin film 1 were alternately stacked so that the thermoplastic resin film 1 was on the surface, and molding was performed under the following molding conditions to obtain continuous fiber reinforced resin composite material (1). At this time, the charged volume ratio of the thermoplastic resin was 50%. The molding conditions used in Example 1 are as follows. The molding machine used for this molding was a continuous compression molding machine. The laminate obtained by stacking the glass cloth 1 and the thermoplastic resin film 1 as described above was placed in the molding machine, and the temperature of the heating zone in the continuous compression molding machine was adjusted to 330°C, the temperature of the cooling zone was adjusted by water cooling, and compression molding was performed at a pressure of 3 MPa and a belt speed of 0.5 m / min to produce a continuous fiber reinforced resin composite material (1). The resulting continuous fiber reinforced resin composite material (1) was then subjected to a load test under the conditions shown in Table 1 and the following recovery step (1). The flexural strength recovery rate, flexural modulus recovery rate, recovery rate of the continuous fiber reinforced resin composite material, recovery rate by bending test, recovery rate by impact test, long-term property recovery rate, interfacial polishing strength of the continuous fiber reinforced resin composite material, and handleability at high temperatures of the continuous fiber reinforced resin composite material were evaluated. The results are shown in Table 1. The recovery step (1) in Example 1 is as follows. Recovery process (1): The test piece after the load test was placed in a mold with a spigot structure, and a hydraulic molding machine (Shoji Co., Ltd.) with a maximum mold clamping force of 50 tons was used to heat the temperature inside the molding machine to 150°C, then the mold was clamped with a mold clamping force of 5 MPa and hot-pressed for 15 minutes. After heating and cooling, the mold was opened and the recovered test piece was removed. In addition, the following experiment was conducted to confirm the change in interface polishing strength depending on the type of glass cloth. Continuous fiber reinforced resin composite material (1(GF2)) and continuous fiber reinforced resin composite material (1(GC3)) were also produced in the same manner as in Example 1, except that GF2 or GF3 was used instead of the glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (1). The interface polishing strength when using glass cloth 2 (GC2) or glass cloth 3 (GC3) was measured according to the procedure described in the above section [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0176] [Example 2] A continuous fiber reinforced resin composite material (2) was obtained in the same manner as in Example 1, except that the recovery step was changed to the following recovery step (2), and the items shown in Table 1 were evaluated in the same manner as in Example 1. Recovery step (2): After the load test, the test piece was placed in a thermostatic water bath set at 80°C for 1 hour. After being removed from the thermostatic water bath, it was dried in a vacuum dryer set at 80°C to obtain a recovered test piece. In addition, similarly to Example 1, continuous fiber reinforced resin composite material (2(GC2)) and continuous fiber reinforced resin composite material (2(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (2), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above column [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0177] [Example 3] A continuous fiber reinforced resin composite material (3) was obtained in the same manner as in Example 1, except that the recovery step was changed to the following recovery step (3), and the items shown in Table 1 were evaluated in the same manner as in Example 1. Recovery step (3): The test piece after the load test was placed in a hot air circulation dryer set at 150°C, treated for 15 minutes, and then the recovered test piece was taken out. In addition, similar to Example 1, continuous fiber reinforced resin composite material (3(GF2)) and continuous fiber reinforced resin composite material (3(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GF2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (3), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above column [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0178] [Example 4] A continuous fiber reinforced resin composite material (4) was obtained in the same manner as in Example 1 except that Resin 2 was used as the thermoplastic resin, and the recovery process (1) was carried out. The items shown in Table 1 were evaluated in the same manner as in Example 1. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (4(GC2)) and continuous fiber reinforced resin composite materials (4(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (4). The interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interface polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0179] [Example 5] A continuous fiber reinforced resin composite material (5) was obtained in the same manner as in Example 1, except that Resin 3 was used as the thermoplastic resin and the following recovery process (4) was performed. The material was evaluated for the items shown in Table 1 in the same manner as in Example 1. Recovery step (4): The recovery step (4) was carried out in the same manner as in Example 1, except that the heating temperature was set to 110°C. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (5(GC2)) and continuous fiber reinforced resin composite materials (5(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (5). The interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above [Interface polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0180] [Example 6] A continuous fiber reinforced resin composite material (6) was obtained in the same manner as in Example 2, except that Resin 4 was used as the thermoplastic resin, and the recovery process (2) was carried out. The items shown in Table 1 were evaluated in the same manner as in Example 1. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (6(GC2)) and continuous fiber reinforced resin composite materials (6(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (6), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above column [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0181] [Example 7] A continuous fiber reinforced resin composite material (7) was obtained in the same manner as in Example 3, except that Resin 5 was used as the thermoplastic resin and the following recovery process (5) was performed. The material was evaluated for the items shown in Table 1 in the same manner as in Example 1. Recovery step (5): The recovery step (5) was carried out in the same manner as in Example 3, except that the heating temperature was 160°C. In addition, similar to Example 1, continuous fiber reinforced resin composite materials (7(GC2)) and continuous fiber reinforced resin composite materials (7(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (7). The interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interface polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0182] [Example 8] A continuous fiber reinforced resin composite material (8) was obtained in the same manner as in Example 1, except that the recovery step was the following recovery step (6), and was evaluated for the items shown in Table 1 in the same manner as in Example 1. Recovery step (6): The recovery step (6) was carried out in the same manner as in Example 1, except that the heating temperature was set to 265°C. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (8(GC2)) and continuous fiber reinforced resin composite materials (8(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (8), and the interfacial polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interfacial polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0183] [Example 9] A continuous fiber reinforced resin composite material (9) was obtained in the same manner as in Example 1, except that the recovery step was performed as follows, and the recovery step (7) was performed. The items shown in Table 1 were evaluated in the same manner as in Example 1. Recovery step (7): After the load test, the test piece was placed in a constant temperature and humidity chamber set at 80°C and 57% humidity for 24 hours, and then the recovered test piece was taken out. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (9(GC2)) and continuous fiber reinforced resin composite materials (9(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (9). The interfacial polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interfacial polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0184] [Example 10] A continuous fiber reinforced resin composite material (10) was obtained in the same manner as in Example 5, except that the recovery step was carried out as follows, and the recovery step (8) was carried out. The items shown in Table 1 were evaluated in the same manner as in Example 1. Recovery step (8): The recovery step (8) was carried out in the same manner as in Example 1, except that the heating temperature was set to 110°C. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (10(GC2)) and continuous fiber reinforced resin composite materials (10(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (10). The interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interface polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0185] [Example 11] A continuous fiber reinforced resin composite material (11) was obtained in the same manner as in Example 1, except that the recovery step was performed as follows, and the recovery step (9) was performed. The items shown in Table 1 were evaluated in the same manner as in Example 1. Recovery step (9): The recovery step (9) was carried out in the same manner as in Example 1, except that the heating temperature was 80°C. In addition, similarly to Example 1, continuous fiber reinforced resin composite materials (11(GC2)) and continuous fiber reinforced resin composite materials (11(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (11). The interfacial polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above section [Interfacial polishing strength of continuous fiber reinforced resin composite materials]. The results are shown in Table 1.

[0186] [Comparative Example 1] A commercially available prepreg ("Tepex dynalite 101" manufactured by Bond Laminate) in which polyamide 66 was impregnated into glass cloth was used as a continuous fiber reinforced resin composite material (c1). The same recovery process (1) as in Example 1 was carried out, and the items shown in Table 1 were evaluated. In addition, similar to Example 1, continuous fiber reinforced resin composite materials (c1(GC2)) and continuous fiber reinforced resin composite materials (c1(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (c1). The interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0187] Comparative Example 2 A continuous fiber reinforced resin composite material (c2) was obtained in the same manner as in Example 1 except that Resin 6 was used as the thermoplastic resin, and the recovery step (1) was carried out, followed by evaluation of the items shown in Table 1. In addition, similar to Example 1, continuous fiber reinforced resin composite material (c2(GC2)) and continuous fiber reinforced resin composite material (c2(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (c2), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0188] Comparative Example 3 A continuous fiber reinforced resin composite material (c3) was obtained in the same manner as in Example 1 except that Resin 7 was used as the thermoplastic resin, and the recovery step (1) was carried out, followed by evaluation of the items shown in Table 1. In addition, similar to Example 1, continuous fiber reinforced resin composite material (c3(GC2)) and continuous fiber reinforced resin composite material (c3(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (c3), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0189] [Reference example 1] A continuous fiber reinforced resin composite material (1) obtained in the same manner as in Example 1 was evaluated for the items shown in Table 1 without undergoing the recovery step. In addition, similar to Example 1, continuous fiber reinforced resin composite material (1(GC2)) and continuous fiber reinforced resin composite material (1(GC3)) were also produced in the same manner as in Example 1, except that glass cloth 2 (GC2) or glass cloth 3 (GC3) was used instead of glass cloth 1 (GC1) used in the production of the continuous fiber reinforced resin composite material (1), and the interface polishing strength when GF2 or GF3 was used was measured according to the procedure described in the above [Interface polishing strength of continuous fiber reinforced resin composite material]. The results are shown in Table 1.

[0190] [Table 1]

[0191] As can be seen from Table 1 above, Examples 1 to 11 had recovery rates of 50% or more, and exhibited very high bending recovery properties, impact recovery properties, and long-term recovery properties.

[0192] (Variation) In the above-described embodiment, the learning ratio calculation unit 13 determines multiple divisions CL based on the hammering sound data in the distribution. However, this is not limited thereto, and the learning ratio calculation unit 13 may determine multiple provisional divisions by non-hierarchical clustering or the like. In such a configuration, the learning ratio calculation unit 13 may determine, as multiple divisions CL, divisions corresponding to the largest rectangles included in each of the regions representing the multiple provisional divisions in a graph showing the distribution of hammering sound data, with multiple characteristic values ​​included in the hammering sound data plotted on each axis. This allows the inspection device 2 to reduce the processing load required to determine which of the multiple divisions CL the hammering sound data of the unknown material is included in. Therefore, the inspection device 2 can reduce the processing load required to inspect the unknown material.

[0193] In the above-described embodiment, the inspection model learning device 1 may further include an output unit. In such a configuration, the output unit may output the model learned by the model learning unit 14 to a plurality of other computers via a communication network, and the model may be managed by a blockchain formed by the plurality of other computers. Furthermore, the inspection device 2 can perform the inspection as described above using the model managed by the blockchain.

[0194] <Program> The above-described inspection model learning device 1 and inspection device 2 can be realized by a computer 101. A program for causing the inspection model learning device 1 and inspection device 2 to function may be provided. The program may be stored on a storage medium or provided via a network. FIG. 16 is a block diagram showing a schematic configuration of a computer 101 that functions as the inspection model learning device 1 and inspection device 2, respectively. Here, the computer 101 may be a general-purpose computer, a dedicated computer, a workstation, a PC (Personal Computer), an electronic notepad, or the like. The program instructions may be program code, code segments, or the like for performing required tasks.

[0195] The broken state input unit 31 and recovery condition determination unit 32 of the recovery control device 3 can be realized by a computer 101. A program for causing the broken state input unit 31 and recovery condition determination unit 32 of the recovery control device 3 to function may be provided. The program may be stored in a storage medium or provided via a network. The computer that functions as the broken state input unit 31 and recovery condition determination unit 32 of the recovery control device 3 may also be configured similarly to the computer 101 shown in FIG. 16.

[0196] 16, the computer 101 includes a processor 110, a ROM 120, a RAM 130, a storage 140, an input interface (I / F) 150, an output interface (I / F) 160, and a communication interface (I / F) 170. Each component is connected to each other via a bus 180 so as to be able to communicate with each other. The processor 110 is specifically a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an SoC (System on a Chip), or the like, and may be configured by a plurality of processors of the same type or different types.

[0197] The processor 110 controls each component and performs various arithmetic processing. That is, the processor 110 reads a program from the ROM 120 or the storage 140 and executes the program using the RAM 130 as a work area. The processor 110 controls each component and performs various arithmetic processing in accordance with the program stored in the ROM 120 or the storage 140. In the above-described embodiment, the program according to the present disclosure is stored in the ROM 120 or the storage 140.

[0198] The program may be stored in a storage medium readable by the computer 101. Using such a storage medium, the program can be installed in the computer 101. Here, the storage medium on which the program is stored may be a non-transitory storage medium. The non-transitory storage medium is not particularly limited, and may be, for example, a CD-ROM, a DVD-ROM, or a USB (Universal Serial Bus) memory. Furthermore, the program may be downloaded from an external device via a network.

[0199] The ROM 120 stores various programs and various data. The RAM 130 temporarily stores programs or data as a working area. The storage 140 is configured with an HDD or SSD and stores various programs including an operating system and various data.

[0200] The input interface 150 includes one or more input interfaces that receive a user's input operation and acquire information based on the user's operation. For example, the input interface 150 may be, but is not limited to, a pointing device, a keyboard, a mouse, etc.

[0201] The output interface 160 is a display that outputs information in the form of video, but is not limited to this. If the output interface 160 is a touch panel display, it also functions as the input interface 150.

[0202] The communication interface (I / F) 170 is an interface for communicating with an external device.

[0203] The following additional notes are provided regarding the above-described embodiments.

[0204] (Additional note 1) an input interface that receives input of a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking a known material whose failure state is known, and known material data indicating the failure state of the known material; a controller; The controller determining a plurality of divisions into which the distribution is divided based on the known material data, and calculating a learning ratio which is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; learning a model that outputs a failure state of the unknown material, the failure state of which is unknown, based on the learning ratio; Model learning device for inspection. (Additional note 2) 2. The inspection model learning device according to claim 1, wherein the controller determines the plurality of divisions by non-hierarchical clustering. (Additional note 3) 3. The testing model learning device according to claim 1, wherein the plurality of characteristic values ​​are amplitude and duration. (Additional note 4) 4. The inspection model learning device according to claim 1, wherein the plurality of characteristic values ​​related to the hitting sounds are obtained by acoustic emission. (Additional note 5) An inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hitting sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that shows a distribution of hitting sound data including a plurality of characteristic values ​​related to hitting sounds obtained by hitting a known material whose failure state is known, and the failure state of the known material, an input interface that accepts input of unknown material data that indicates a distribution of hammering sound data that indicates the plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; a controller; The controller calculating a judgment ratio which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of divisions; The judgment ratio is input to the model, and the fracture state of the unknown material is output based on the judgment ratio. Inspection equipment. (Additional note 6) The inspection device according to Appendix 5, wherein the controller causes the model to output the fractured state based on a judgment ratio included in one of the plurality of divisions in the hammering sound data and a judgment ratio included in another division different from the one division in the hammering sound data. (Additional note 7) 7. The inspection device according to claim 5, wherein the plurality of divisions are determined by dividing a distribution of the hammering sound data by non-hierarchical clustering. (Additional note 8) 8. The inspection device according to any one of appendixes 5 to 7, wherein the plurality of characteristic values ​​are amplitude and duration. (Additional note 9) 9. The inspection device according to any one of appended items 5 to 8, wherein the plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission. (Additional note 10) A step of receiving input of known material data indicating a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking a known material whose fracture state is known, and the fracture state of the known material; determining a plurality of divisions into which the distribution is divided based on the known material data, and calculating a learning ratio which is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; a step of learning a model that outputs a failure state of an unknown material, the failure state of which is unknown, based on the learning ratio; A method for learning a model for testing, including: (Additional note 11) 11. The inspection model learning method according to claim 10, wherein the plurality of characteristic values ​​related to the hitting sounds are obtained by acoustic emission. (Additional note 12) An inspection method performed by an inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hammering sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that indicates a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by hammering a known material whose failure state is known, and the failure state of the known material, receiving input of unknown material data indicating a distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; calculating a judgment ratio which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of divisions; a step of inputting the judgment ratio into the model and outputting a fracture state of the unknown material based on the judgment ratio; An inspection method including: (Additional note 13) Item 13. The inspection method according to item 12, wherein the plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission. (Additional note 14) A non-transitory storage medium storing a program executable by a computer, the non-transitory storage medium storing the program causing the computer to function as the testing model learning device described in any one of appendix 1 to 4. (Additional note 15) A non-transitory storage medium storing a program executable by a computer, the non-transitory storage medium storing the program causing the computer to function as the inspection device described in any one of appendixes 5 to 9. (Additional note 16) receiving input of data indicating the fracture state of the unknown material determined using the inspection method according to claim 12 or 13; determining a recovery condition for recovering the interface of the unknown material based on the failure state; recovering the interface of the unknown material according to the recovery condition; A method for recovering unknown materials, comprising: (Additional note 17) A continuous fiber reinforced resin composite material containing continuous reinforcing fibers and a thermoplastic resin, which can be measured by the inspection method described in Supplementary Item 12 or 13, The following formula (1): Recovery rate (%)=(B / A)×100 Formula (1) (In the above formula (1), A represents the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin, among the continuous reinforcing fibers present in a scanning electron microscope image (magnification: 4000 to 4500 times) of a test piece (a) cut out from the continuous fiber reinforced resin composite material, which are measured at any 20 points so that the cross section of the continuous fiber reinforcing fibers is included in the image, The continuous fiber reinforced resin composite material is subjected to a load test and a recovery step after the load test, and a test piece (b) is cut out from the recovered continuous fiber reinforced resin composite material. Measurements are made at 20 arbitrary points so that cross sections of the continuous reinforcing fibers in the recovered continuous fiber reinforced resin composite material are included in an image (magnification: 4000 to 4500 times) taken by a scanning electron microscope, and B is the number of continuous reinforcing fibers that do not have voids at the interface between the continuous reinforcing fibers and the thermoplastic resin among the continuous reinforcing fibers present in the image. A continuous fiber reinforced resin composite material characterized in that the recovery rate (%) represented by the following formula is within the range of 50 to 100%. (Additional note 18) Item 18. The continuous fiber reinforced resin composite material according to Addendum 17, wherein the recovery step in formula (1) is a step of heating the continuous fiber reinforced resin composite material after the load test in a temperature range of not less than the glass transition temperature of the thermoplastic resin in the continuous fiber reinforced resin composite material but not more than the glass transition temperature + 150°C. (Additional note 19) 19. The continuous fiber reinforced resin composite material according to Appendix 17 or 18, wherein the recovery step in formula (1) includes heating the continuous fiber reinforced resin composite material after the load test at a temperature equal to or higher than the glass transition temperature of the thermoplastic resin and lower than the melting point of the thermoplastic resin. (Additional note 20) 20. The continuous fiber reinforced resin composite material according to any one of appended items 17 to 19, wherein the recovery step in formula (1) includes heating the continuous fiber reinforced resin composite material after the load test in a state where the water absorption of the continuous fiber reinforced resin composite material after the load test is set to a range of at least 0.3 mass%. (Additional note 21) 21. The continuous fiber reinforced resin composite material according to any one of appended items 17 to 20, wherein the recovery step in formula (1) includes heating the continuous fiber reinforced resin composite material after the load test under pressure. (Additional note 22) 21. The continuous fiber reinforced resin composite material according to any one of appended items 17 to 20, wherein the recovery step in formula (1) includes heating the continuous fiber reinforced resin composite material after the load test under no pressure. (Additional note 23) A method for producing a continuous fiber reinforced resin composite material obtained by laminating continuous reinforcing fibers surface-treated with a sizing agent (1) containing a coupling agent and a thermoplastic resin, comprising: 23. The method for producing a continuous fiber reinforced resin composite material according to any one of appended items 17 to 22, wherein the interfacial polishing strength of the continuous fiber reinforced resin composite material is 0.8 to 1.2 times the interfacial polishing strength of a continuous fiber reinforced resin composite material obtained by laminating continuous reinforcing fibers surface-treated with only the coupling agent and the thermoplastic resin. (Additional note 24) A method for producing a continuous fiber reinforced resin composite material comprising a sizing agent (1) containing a coupling agent, continuous reinforcing fibers, and a thermoplastic resin, and obtained by laminating the continuous reinforcing fibers and the thermoplastic resin, comprising: 23. The method for producing a continuous fiber reinforced resin composite material according to any one of appended items 17 to 22, wherein the interfacial polishing strength of the continuous fiber reinforced resin composite material is at least twice as high as the interfacial polishing strength of a continuous fiber reinforced resin composite material obtained by laminating the continuous reinforcing fibers and the thermoplastic resin when the surface treatment using the sizing agent (1) is removed.

[0205] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, and technical standard was specifically and individually indicated to be incorporated by reference.

[0206] Although the above-described embodiments have been described as typical examples, it will be apparent to those skilled in the art that many modifications and substitutions can be made within the spirit and scope of the present disclosure. Therefore, the present invention should not be interpreted as being limited by the above-described embodiments, and various modifications or alterations are possible without departing from the scope of the claims. For example, multiple building blocks shown in the block diagrams of the embodiments can be combined into one, or one building block can be divided. [Industrial Applicability]

[0207] The continuous fiber reinforced resin composite material of the present embodiment can be used industrially as a recoverable reinforcing material for materials that are subject to destruction due to fatigue, impact, etc., such as structural parts for various machines, automobiles, etc., and as a composite material for thermoplastic resin compositions. [Explanation of symbols]

[0208] 1. Test model learning device 2. Inspection equipment 3 Recovery Control Device 11 Known material data input section 12 Known materials data storage section 13 Learning ratio calculation unit 14 Model Learning Section 15 Model memory section 21 Unknown material data input section 22 Judgment ratio calculation section 23 Judgment section 24 Output section 31 Destruction state input section 32 Recovery condition determination section 33 Recovery condition output section 34 Recovery Department 35 Display section 101 Computer 110 processors 120 ROM 130 RAM 140 Storage 150 Input Interface 160 output interface 170 Communication Interface 180 Bus

Claims

1. a known material data input unit that receives input of a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by hammering a known material whose fracture state is known, and known material data indicating the fracture state of the known material; a learning ratio calculation unit that determines a plurality of divisions into which the distribution is divided based on the known material data, and calculates a learning ratio that is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; a model learning unit that learns a model that outputs a failure state of an unknown material whose failure state is unknown based on the learning ratio; A testing model learning device comprising: The learning ratio calculation unit determines the plurality of categories by non-hierarchical clustering.

2. The testing model learning device according to claim 1 , wherein the plurality of characteristic values ​​are amplitude and duration.

3. 3. The inspection model learning device according to claim 1, wherein the plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission.

4. An inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hitting sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that shows a distribution of hitting sound data including a plurality of characteristic values ​​related to hitting sounds obtained by hitting a known material whose failure state is known, and the failure state of the known material, an unknown material data input unit that receives input of unknown material data that indicates a distribution of hammering sound data that indicates the plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; a judgment ratio calculation unit that calculates a judgment ratio, which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of sections; a determination unit that inputs the determination ratio into the model and outputs a fracture state of the unknown material based on the determination ratio; An inspection device comprising: The plurality of divisions are determined by dividing the distribution of the hammering sound data by non-hierarchical clustering.

5. 5. The inspection device according to claim 4, wherein the determination unit causes the model to output the fractured state based on a determination ratio included in one of the plurality of divisions in the hammering sound data and a determination ratio included in another division different from the one division in the hammering sound data.

6. 6. The inspection apparatus according to claim 4, wherein the plurality of characteristic values ​​are amplitude and duration.

7. 6. The inspection device according to claim 4, wherein the plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission.

8. A step of receiving input of known material data indicating a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by striking a known material whose fracture state is known, and the fracture state of the known material; determining a plurality of divisions into which the distribution is divided based on the known material data, and calculating a learning ratio which is a ratio at which the hammering sound data corresponding to the fracture state is included in each of the plurality of divisions; a step of learning a model that outputs a failure state of an unknown material, the failure state of which is unknown, based on the learning ratio; A testing model learning method comprising: The inspection model learning method, wherein the plurality of divisions are determined by dividing the distribution of the hitting sound data by non-hierarchical clustering.

9. 9. The inspection model learning method according to claim 8, wherein the plurality of characteristic values ​​related to the hitting sounds are obtained by acoustic emission.

10. An inspection method performed by an inspection device that performs an inspection using a model that outputs a failure state of an unknown material whose failure state is unknown based on a learning ratio that is a ratio of the hammering sound data included in each of a plurality of divisions in the distribution, the learning ratio being learned using known material data that indicates a distribution of hammering sound data including a plurality of characteristic values ​​related to hammering sounds obtained by hammering a known material whose failure state is known, and the failure state of the known material, receiving input of unknown material data indicating a distribution of hammering sound data indicating a plurality of characteristic values ​​related to hammering sounds obtained by hammering the unknown material; calculating a judgment ratio which is a ratio of the hammering sound data whose distribution is indicated by the unknown material data to be included in each of the plurality of divisions; a step of inputting the judgment ratio into the model and outputting a fracture state of the unknown material based on the judgment ratio; An inspection method comprising: The inspection method, wherein the plurality of categories are determined by dividing the distribution of the hammering sound data by non-hierarchical clustering.

11. The known material and the unknown material are continuous fiber reinforced resin composite materials containing continuous reinforcing fibers and a thermoplastic resin, The inspection method according to claim 10 , wherein the failure state of the unknown material includes each of the states of “non-destructive (normal)”, “fiber breakage”, “unimpregnated”, and “interface failure”.

12. The inspection method according to claim 10 or 11, wherein the plurality of characteristic values ​​related to the hammering sounds are obtained by acoustic emission.

13. A program for causing a computer to function as the inspection model learning device according to claim 1 or 2.

14. A program for causing a computer to function as the inspection device according to claim 4 or 5.

15. receiving input of data indicating the fracture state of the unknown material determined using the inspection method according to claim 11; determining a recovery condition for recovering the interface of the unknown material based on the failure state; recovering the interface of the unknown material according to the recovery condition; A method for recovering unknown materials, comprising:

Images