Release film quality detection method, 3D printing equipment and readable storage medium

Abstract

The present invention is a method for detecting the quality of a release film, wherein the release film is installed at the bottom of a resin tank of a printing device, and the method comprises the following steps: A), establishing a neural network model, wherein the neural network model reflects the structural relationship between a slice image of a printed model and a time-series curve of a peeling force of a printing platform; B), during the printing process, collecting a time-series curve of a peeling force of a printing platform when each layer of resin material is cured; C), inputting the time-series curve of the peeling force into the neural network model, and outputting a simulated image; D), comparing the simulated image with the corresponding slice image, and judging the quality of the release film according to the comparison result. The present application collects the time-series curve of the peeling force of a printing platform, outputs a simulated image through a neural network model, and then judges the quality of the release film by comparing the simulated image with the slice image actually printed, thereby reminding the staff to replace the release film in time.

Description

Technical Field

[0001] The present application relates to the field of additive manufacturing technology, and in particular to a release film quality detection method, a 3D printing device, and a readable storage medium. Background Art

[0002] DLP (Digital Light Processing) digital light processing technology has a relatively simple hardware system structure and low cost. It has a greater "cost-effectiveness" advantage among similar light-curing technologies. Therefore, it has become the main research direction of 3D continuous printing light-curing printing technology.

[0003] The characteristic size of the molded parts printed by the current DLP3D continuous molding printing technology is relatively small. When printing larger-sized molded parts, there are problems such as untimely replenishment of liquid resin flow and tight adhesion between the molded parts and the bottom of the resin tank. The solution adopted in the prior art is to add a release film made of polydimethylsiloxane (PDMS) at the bottom of the resin tank, so that the photosensitive resin is solidified on the surface of the release film to separate the molded part from the bottom of the resin tank. However, due to the problems of tight adhesion between the release film and the molded part, rupture of the release film, and decreased transmittance during the long-term large-scale continuous molding 3D printing process of the PDMS material release film, the printing success rate and the quality of model molding are affected. If the release film ruptures or the transmittance decreases in the middle of printing, it is easy to waste printing materials and printing time. Summary of the invention

[0004] The purpose of this application is to provide a release film quality detection method, which can promptly detect problems such as release film rupture and damage during the printing process, and remind the staff to replace the new release film in time to avoid material waste.

[0005] In order to achieve the above objectives, this application adopts the following technical solutions:

[0006] In a first aspect, a method for detecting the quality of a release film is provided, wherein the release film is installed at the bottom of a resin tank of a printing device, and the printing device further comprises a printing platform located above the resin tank and a light source assembly located below the resin tank, respectively, wherein the method comprises the following steps:

[0007] A), establishing a neural network model, wherein the neural network model reflects the structural relationship between the slice image of the printed model and the time series variation curve of the peeling force of the printing platform;

[0008] B) During the printing process, collect the peeling force time series curve of the printing platform when each layer of resin material is solidified;

[0009] C), inputting the peeling force time series variation curve into the neural network model and outputting a simulation image;

[0010] D) Compare the simulated image with the corresponding slice image and judge the quality of the release film based on the comparison result.

[0011] In a possible implementation, the neural network model is established by the following method:

[0012] Given a slice image;

[0013] Construct a model, collect the peeling force time series variation curve of the printing platform when the printing platform is peeled off from the release film, mark the printing results, and establish a mapping set between the slice image and the peeling force time series variation curve and the printing results;

[0014] A training model is provided, wherein a neural network model is trained using a set of slice images, wherein the slice image set includes a plurality of slice images.

[0015] In a possible implementation manner, the printing result includes: at least one of successful printing, failed printing, and rupture of the release film.

[0016] In a possible implementation, the peeling force is collected by a sensor disposed on the printing platform.

[0017] In a possible implementation, the resolution of the simulation image is lower than that of the slice image.

[0018] In a possible implementation, in step D, the simulated image and the slice image are compared based on their similarity, and the comparison result includes:

[0019] When the similarity is less than or equal to the first threshold, it is determined that the release film is broken and printing fails;

[0020] When the similarity is greater than the first threshold and less than or equal to the second threshold, it is determined that the release film is damaged and a prompt message for replacing the release film is issued;

[0021] When the similarity is greater than the second threshold, the quality of the release film is judged to be qualified and printing continues;

[0022] The second threshold is greater than the first threshold.

[0023] In a possible implementation, when the similarity is lower than a first threshold, a printing procedure termination message is issued.

[0024] In a second aspect, the present application provides a 3D printing device, comprising a resin tank, a printing platform located above the resin tank, and a light source assembly below the resin tank, wherein a release film is fixedly installed at the bottom of the resin tank, and characterized in that the 3D printing device also includes a processor and a memory, wherein a program or instruction is stored in the memory, and when the processor executes the program or instruction, the steps of the release film quality detection method are implemented.

[0025] In a possible implementation, the resin tank is configured to be rotatably disposed on a frame so as to be gradually separated from the printing platform from one side to the other.

[0026] In a third aspect, the present application provides a readable storage medium having a program or instruction stored thereon, wherein the program or instruction, when executed by a processor, implements the steps of the release film quality detection method.

[0027] This application collects the peeling force time-series change curve of the printing platform, outputs a simulated image through a neural network model, and then compares the simulated image with the actual printed slice image to judge the quality of the release film, thereby reminding the staff to replace the release film in time.

[0028] It can be understood that the 3D printing device described in the second aspect and the readable storage medium described in the third aspect can both include the release film quality detection method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the release film quality detection method provided above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic flow chart of a release film quality detection method provided in an embodiment of the present application.

[0030] Figure 2 A schematic diagram of a 3D printing device after a printing platform and a resin tank are separated provided in an embodiment of the present application.

[0031] Figure 3 A schematic diagram of a 3D printing device provided in an embodiment of the present application when a printing platform and a resin tank are peeled off.

[0032] Figure 4 This is a slice image given in an embodiment provided in this application.

[0033] Figure 5 For printing Figure 4 When the slice image in the figure is taken, the peeling force time series curve is collected when the release film quality is qualified;

[0034] Figure 6 The neural network model is based on Figure 5 The simulated image output from the peeling force timing curve in the figure has a similarity of 90% with the original slice image.

[0035] Figure 7 For printing Figure 4 The peeling force time series curve collected when the release film is partially broken when the slice image in the figure is shown;

[0036] Figure 8 The neural network model is based on Figure 7 The simulated image output from the peeling force time series curve in the figure has a similarity of 53% with the original slice image.

[0037] Fig. 9 For printing Figure 4 The peel force time series curve collected when the release film needs to be replaced when the slice image in the image is taken;

[0038] Fig.10 The neural network model is based on Fig. 9 The simulated image output from the peeling force timing curve in the figure has a similarity of 74% with the original slice image.

[0039] Among them: 1. Frame; 2. Resin tank; 3. Vertical guide rail; 4. Lifting platform; 5. Printing platform; 6. Light source assembly; 7. Printing model. DETAILED DESCRIPTION

[0040] To describe the technical content, structural features, achieved purpose and effect of the invention in detail, the technical scheme in the embodiment of the present application will be described below in conjunction with the drawings in the embodiment of the present application. Obviously, the described embodiment is only a part of the embodiment of the present application, rather than all the embodiments. In the following description, for the purpose of explanation, many specific details are set forth to provide a detailed description of various exemplary embodiments or implementations of the invention. However, various exemplary embodiments may also be implemented without these specific details or in the case of one or more equivalent arrangements. In addition, various exemplary embodiments may be different, but not necessarily exclusive. For example, without departing from the inventive concept, the specific shape, structure and characteristics of the exemplary embodiment may be used or implemented in another exemplary embodiment.

[0041] The embodiments of the present application provide a release film quality detection method, a 3D printing device using the method, and a readable storage medium.

[0042] See also Figure 2 , 3As shown, a schematic diagram of the structure of a 3D printing device provided in an embodiment of the present application is shown. The 3D printing device includes a frame 1, a vertical guide rail 3 fixedly connected to the frame 1, a lifting platform 4 slidably connected to the vertical guide rail 3, a printing platform 5 suspended on the lifting platform 4, a plurality of sensors (not shown in the figure), a resin tank 2 liftably arranged on the frame 1, a light source assembly 6, a processor and a memory.

[0043] The 3D printing device belongs to a DLP light curing device. The resin tank 2 contains photosensitive resin, and a transparent release film is fixed at the bottom of the resin tank 2. The light source component 6 sequentially projects the slice image of the model to be printed on the upper surface of the release film to cure the photosensitive resin. Each time a layer of material is cured, the lifting platform 4 drives the printing platform 5 to rise a layer height h. Each layer of cured material corresponds to a slice image, and after being stacked layer by layer, a 3D printed model 7 is formed.

[0044] After one layer is cured, the printing platform rises and the cured material is separated from the surface of the release film. In previous 3D printing equipment, in order to ensure printing quality, the light source assembly, resin tank and printing platform need to be kept level, and the resin tank is usually fixed on the frame 1. When the printing platform and the resin tank are separated, the two remain parallel, which makes the peeling force between the two very large, making the release film easy to tear or damage. Figure 3 As shown, in the present application, in order to make the release film and the printing platform 5 separate more smoothly and the peeling force fluctuation more stable, it is preferred to start separation from one side of the printing platform. Figure 3 The resin tank 2 can be rotatably connected to the frame 1. After each layer is cured, the resin tank 2 descends, and the end close to the rotation connection is peeled off first, and then the other end is gradually peeled off.

[0045] In this embodiment, the resin tank 2 rotates around a rotation axis extending in the left-right direction. In other embodiments, the resin tank can also rotate around a rotation axis extending in the front-back direction, or can also be rotated by the printing platform, thereby achieving the effect of peeling off one side first.

[0046] In order to accurately obtain the size and real-time changes of the peeling force between the printing platform and the release film, a plurality of sensors, such as 2-6, are used in the preferred embodiment of the present application. More sensors can provide more training data for the machine learning model. In this embodiment, 4 sensors are used. They are distributed in a rectangular shape on two opposite sides of the printing platform. The sensor is located between the lifting platform 4 and the printing platform 5, and is connected to the processor signal. It is configured to collect the real-time data of the peeling force of the printing platform and feed it back to the processor. The peeling force signals fed back by the two sensors on the same side should be close and have the same trend of change. If the signal difference between the two sensors on the same side is found to be large, it is very likely that the release film has broken.

[0047] The embodiment of the present application does not impose any special restrictions on the specific types of the above-mentioned sensors. For example, it can be a stress sensor, a weight sensor, etc.

[0048] The memory of the 3D printing device is used to store instructions, and when the instructions are executed by the processor, the 3D printer executes the steps of the following release film quality detection method.

[0049] The present application also discloses a computer-readable storage medium storing at least one program, which can implement the steps of the following release film quality detection method when called.

[0050] Specifically, the present application discloses a method for detecting the quality of a release film, see Figure 1 As shown, the method comprises the following steps:

[0051] Step A), establishing a neural network model, wherein the neural network model reflects the structural relationship between the slice image of the printed model and the time series variation curve of the peeling force of the printing platform.

[0052] The neural network model is established by the following method:

[0053] Given a training slice image, print it;

[0054] Construct a model, collect the peeling force time series variation curve of the printing platform when the printing platform is peeled off from the release film, mark the printing results, and establish a mapping set between the slice image and the peeling force time series variation curve and the printing results;

[0055] The training model uses the peeling force time series curve as input and the slice image as output to train the encoding-decoding neural network, and the output is a low-resolution predicted printing image, and the predicted printing image is required to be as close as possible to the slice image morphology. The neural network model is trained using a set of slice images, and the slice image set includes a plurality of slice images.

[0056] The printing result includes at least one of successful printing, failed printing and rupture of the release film.

[0057] The neural network here is an autoencoder, which is a neural network model for unsupervised learning. The purpose of the autoencoder is to convert the input signal into a lower-dimensional representation and reconstruct an output that matches the original input signal as closely as possible from the low-dimensional encoding. The learning process of the autoencoder is to calculate the error between the input and the reconstructed signal, and learn to automatically extract appropriate features by minimizing the error.

[0058] Step B), during the printing process, collect the peeling force time series curve of the printing platform when each layer of resin material is cured.

[0059] When the printed model is separated from the release film, the printing platform is subject to suction and adhesion. The suction is caused by atmospheric pressure. The thicker the material, the harder it is to separate, and the greater the suction. Therefore, there will be great resistance when the printing platform is lifted and lowered. Peeling force = adhesion force + atmospheric pressure. 1 square centimeter is about 12 Newtons, and an area the size of a palm can reach 1000 Newtons, which is about the weight of an adult. Hard pulling can easily damage the release film, glass and printed model, and even cause damage to the lifting motor. Therefore, the platform should be driven up at a slow speed. Once a gap appears between the printed model and the release film, the suction will disappear very quickly. At this time, it can be pulled up at a faster speed to replenish the resin material. Therefore, when separating, the rising action of the printing platform is slow at first and then fast.

[0060] Step C), input the peeling force time series variation curve into the neural network model, and output a simulated image.

[0061] The resolution of the simulated image is lower than that of the slice image.

[0062] This process is a decoding process, that is, using the neural network to decode the structural features between the slice image and the peeling force time-series variation curve and output a simulated image.

[0063] Step D), compare the simulated image with the corresponding slice image for similarity, and judge the quality of the release film according to the comparison result.

[0064] The similarity comparison of pictures can be done by using methods such as feature extraction algorithm, direct comparison algorithm, structural similarity index or perceptual hashing algorithm. After the comparison, the simulated images can be scored for similarity, such as 80% similarity, 99% similarity, etc.

[0065] By manually setting the first threshold and the second threshold, the printing quality can be judged according to the quality of the image output, and then the quality of the release film can be judged. The first threshold and the second threshold are empirical values ​​and can be set manually, and the second threshold is greater than the first threshold.

[0066] Figure 4-10 This is an embodiment given in the present application. In this embodiment, the first threshold is 60%, and the second threshold is 75%. Figure 4 To print slice images, Figure 5 , 7 9 and 9 are respectively the peeling force time series curves collected by the sensor under different conditions of the release film. Figure 6 , 8 Figures 1 and 2 are the simulated images obtained after different peeling force time series curves are input into the neural network model. Figure 6 The similarity of the simulated images in is 90%, Figure 8 The similarity is 53%. Fig.10 The similarity is 74%. Figure 5 , 7 9, the horizontal axis is the sampling interval, which is 100 milliseconds in this embodiment, and the vertical axis is the peeling force, in kilograms.

[0067] Will Figure 5 The peeling force time series variation curve in the neural network is input into the neural network to obtain Figure 6 After comparing the simulated images, the similarity of the simulated images is 90%, which is higher than the second threshold value of 75%. It is determined that the quality of the release film is qualified and printing can continue.

[0068] Will Figure 7 The peeling force time series variation curve in the neural network is input into the neural network to obtain Figure 8 After comparison, the similarity of the simulated images is only 53%, which is less than the first threshold of 60%. It is judged that the release film is broken and the printing fails. At the same time, a print program termination message is issued. The termination program here can be regarded as a pause. After replacing the release film, the printing is continued. Figure 7 It can also be seen that the force conditions of the four sensors are quite different, reflecting that the force conditions of the two sensors on the same edge are not consistent and the difference is large.

[0069] Will Fig. 9 The peeling force time series variation curve in the neural network is input into the neural network to obtain Fig.10 After comparison, the similarity of the simulated images is 74%, which is greater than the first threshold of 60% and less than the second threshold of 75%. This indicates that there are tears or scratches on the surface of the release film, the transmittance is reduced, and the correct molding quality has been affected, which is not conducive to continued printing. At this time, it is judged that the release film is damaged, and a prompt message (such as an alarm message) for replacing the release film is issued.

[0070] Through the printing method of the present application, the result of printing failure can be discovered in time, and a prompt alarm can be issued in time to remind the staff to intervene. It can also automatically return to the previous position and re-expose the layer that failed to print, thereby improving the automation rate.

[0071] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and the scope of protection of the present invention is defined by the attached claims, description and their equivalents.

Claims

1. A method for detecting the quality of a release film, wherein the release film is installed at the bottom of a resin tank of a printing device, and the printing device further comprises a printing platform located above the resin tank and a light source assembly located below the resin tank, wherein: The method comprises the following steps: A), establishing a neural network model, wherein the neural network model reflects the structural relationship between the slice image of the printed model and the time series variation curve of the peeling force of the printing platform; B) During the printing process, collect the peeling force time series curve of the printing platform when each layer of resin material is solidified; C), inputting the peeling force time series variation curve into the neural network model and outputting a simulation image; D) Compare the simulated image with the corresponding slice image, and judge the quality of the release film according to the comparison result; In step D, the simulated image and the slice image are compared based on their similarity, and the comparison results include: When the similarity is less than or equal to the first threshold, it is determined that the release film is broken and printing fails; When the similarity is greater than the first threshold and less than or equal to the second threshold, it is determined that the release film is damaged and a prompt message for replacing the release film is issued; When the similarity is greater than the second threshold, the quality of the release film is judged to be qualified and printing continues; The second threshold is greater than the first threshold.

2. The release film quality detection method according to claim 1, characterized in that: The neural network model is established by the following method: Given a slice image; Construct a model, collect the peeling force time series variation curve of the printing platform when the printing platform is peeled off from the release film, mark the printing results, and establish a mapping set between the slice image and the peeling force time series variation curve and the printing results; A training model is provided, wherein a neural network model is trained using a set of slice images, wherein the slice image set includes a plurality of slice images.

3. The release film quality detection method according to claim 2, characterized in that: The printing result includes at least one of successful printing, failed printing and rupture of the release film.

4. The release film quality detection method according to claim 1, characterized in that: The peeling force is collected by a sensor arranged on the printing platform.

5. The method for detecting the quality of release film according to claim 1, characterized in that: The resolution of the simulated image is lower than that of the slice image.

6. The method for detecting the quality of release film according to claim 1, characterized in that: When the similarity is lower than the first threshold, a printing procedure termination message is issued.

7. A 3D printing device, comprising a resin tank, a printing platform located above the resin tank, and a light source assembly located below the resin tank, wherein a release film is fixedly installed at the bottom of the resin tank, characterized in that: The 3D printing device also includes a processor and a memory, wherein the memory stores a program or instruction, and when the processor executes the program or instruction, the steps of the release film quality detection method as described in any one of claims 1 to 6 are implemented.

8. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by the processor, the steps of the release film quality detection method as described in any one of claims 1 to 6 are implemented.

Images