Casting inspection method and inspection system
Thermal imaging and AI-driven inspection methods for castings address the challenge of detecting defects in cast pipes, ensuring accurate and automated quality assessment, reducing waste and costs by identifying defects early in the manufacturing process.
Patent Information
- Application Number
- JP2021006436
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-01-19
- Publication Date
- 2024-11-27
- Estimated Expiration
- 2041-01-19
AI Technical Summary
Existing casting inspection methods, particularly for cast pipes, struggle to accurately detect defects such as hot spots and other minute surface irregularities, which are difficult to identify and require skilled operators, leading to inconsistent judgments and potential wasteful downstream processing of defective products.
A method and system that utilizes thermal imaging to inspect castings by acquiring images before cooling from a high temperature state to room temperature, comparing with reference images, and employing artificial intelligence to learn and identify characteristic defects, enabling automated and accurate quality determination.
Enables clear identification of casting quality regardless of operator skill, automates the inspection process, improves accuracy, and prevents defective products from progressing through subsequent manufacturing stages, reducing waste and costs.
Smart Images

Figure 0007590876000001 
Figure 0007590876000002 
Figure 0007590876000003
Abstract
Description
[Technical field]
[0001] The present invention relates to a method and system for inspecting a casting. [Background technology]
[0002] For example, iron pipes used for water supply and sewage systems are generally manufactured by casting. Patent Document 1 discloses a cast pipe manufactured by centrifugal casting. After casting, the cast pipe may have defects such as cold shut. Once such a defect occurs, the same defect may occur continuously in cast pipes cast thereafter. Furthermore, if a defective cast pipe is sent to a later process, wasteful work may be performed, such as painting a cast pipe that should have been removed from the manufacturing process. For this reason, the appearance of the cast pipe that has been cooled after casting is visually inspected by an operator. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Japanese Patent Application Publication No. 8-141721 Summary of the Invention [Problem to be solved by the invention]
[0004] Some defects in cast pipes are difficult to detect, as they appear as minute spots on the surface. For this reason, visual inspection requires highly skilled workers. For these reasons, it is difficult to make a uniform pass / fail judgment that does not depend on the worker's level of skill.
[0005] In view of the above problems, an object of the present invention is to provide a casting inspection method and inspection system that can clearly distinguish whether a casting is good or bad. [Means for solving the problem]
[0006] A method for inspecting a casting according to one embodiment of the present invention includes the steps of acquiring a thermal image of the casting before the casting is cooled from a high-temperature state to room temperature, and judging the quality of the casting based on the thermal image.
[0007] In the step of casting the casting by centrifugal casting and acquiring the thermal image, the thermal image may be acquired while the casting is cooled from a high-temperature state produced by the centrifugal casting to a room temperature state.
[0008] In the step of determining the quality of the casting, a thermal image obtained from a reference casting may be used as a reference thermal image, and the quality of the casting may be determined based on a comparison result between the thermal image and the reference thermal image.
[0009] In the step of determining whether the casting is good or bad, characteristic parts which appear in the thermal image corresponding to accumulated defect modes of the casting may be learned, and the quality of the casting may be determined based on the learning results of the characteristic parts.
[0010] In the step of acquiring the thermal image, the thermal image may be acquired at a predetermined time point, the predetermined time point including a first time point and a second time point after the first time point, a first thermal image may be acquired at the first time point and a second thermal image may be acquired at the second time point, and in the step of determining the quality of the casting, the presence or absence of a first defect mode in the casting may be determined based on the first thermal image, and the presence or absence of a second defect mode in the casting may be determined based on the second thermal image.
[0011] A casting inspection system according to one embodiment of the present invention is a casting inspection system for performing the above-mentioned casting inspection method, and the inspection system comprises an inspection space in which the casting in a high-temperature state can be placed and which can be cooled, and a thermal image capturing device capable of acquiring the thermal image of the casting in a high-temperature state placed in the inspection space.
[0012] The inspection space may be located near a centrifugal casting machine capable of casting the casting.
[0013] The apparatus may further include a judgment device for judging the quality of the casting.
[0014] The judgment device may use a thermal image obtained from a reference casting as a reference thermal image, and judge the quality of the casting based on a comparison result between the thermal image and the reference thermal image.
[0015] The judgment device may be equipped with an artificial intelligence function that learns characteristic parts that appear in the thermal image corresponding to accumulated defect modes of the casting, and judges the quality of the casting based on the learning results of the characteristic parts. Effect of the Invention
[0016] According to the casting inspection method and inspection system of one embodiment of the present invention, it is possible to clearly distinguish whether a casting is good or bad. [Brief description of the drawings]
[0017] [Figure 1] 1 is a schematic perspective view showing a casting inspection system according to an embodiment of the present invention; [Diagram 2] 1 is a schematic perspective view showing an example of a casting that is an inspection target of a casting inspection system and inspection method according to an embodiment of the present invention; [Diagram 3] 1 is a thermal image acquired by a casting inspection system and method according to an embodiment of the present invention. [Figure 4] 1 is a flowchart illustrating a method for inspecting a casting according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0018] Hereinafter, a casting inspection system according to an embodiment of the present invention will be described with reference to the accompanying drawings. Note that the embodiment described below is merely an example, and the casting inspection system of the present invention is not limited to the following embodiment.
[0019] [Casting inspection system according to one embodiment of the present invention] FIG. 1 shows an inspection system D for a casting C according to an embodiment of the present invention. The inspection system D according to this embodiment is an inspection system for performing an inspection method for a casting C described later. The inspection system D according to this embodiment is particularly suitable for inspecting a casting having a thin-walled portion. In this embodiment, the casting C to be inspected is a cast pipe, and specifically, the casting C is a casting cast by centrifugal casting (see also the enlarged view of the casting C in FIG. 2). More specifically, the casting C is a casting having a cylindrical body C0 extending along a central axis CX passing through the center of the cross section, and is provided with a receiving port C1 having a diameter larger than the body C0 at one end side and a spigot C2 having approximately the same diameter as the body C0 at the other end side along the central axis CX in order to connect with other castings. However, the casting C to be inspected is not particularly limited and can be selected from castings having various shapes. The size of the casting C is not particularly limited, but when the casting C is a cast pipe, for example, the diameter is 50 mm to 2600 mm, and the length along the central axis CX is 4 m to 6 m.
[0020] In Fig. 1, the inspection system D for the casting C is disposed in a space between a casting machine A (specifically, a centrifugal casting machine) used for casting the casting C, which is a step before the inspection step, and an annealing furnace B used for annealing the casting C, which is a step after the inspection step. However, the location of the inspection system D is not particularly limited, and can be selected from various locations, such as the space around any heat treatment device.
[0021] In this embodiment, as shown in FIG. 1, the inspection system D for the casting C includes an inspection space D1 for cooling the casting C, a thermal image pickup device D2 for acquiring a thermal image P, and a judgment device D3 for judging the quality of the casting C. In this specification, the term "thermal image" refers to an image that visually displays the temperature distribution of the subject (casting C), and may be a still image or a video image. In this specification, the term "thermal image pickup device" refers to an apparatus that can visualize the temperature distribution of the subject as a thermal image by picking up an image of the subject. As shown in FIG. 1, the inspection system D for the casting C may further include a display device D4 for displaying the thermal image P, a database D5 for storing the thermal image P, and an alarm mechanism D6 for notifying the occurrence of a defect in the casting C.
[0022] The inspection space D1 is a space in which the casting C in a high temperature state can be placed and cooled. In FIG. 1, the inspection space D1 is arranged near a casting machine A (specifically, a centrifugal casting machine) capable of casting the casting C. In this embodiment, the inspection space D1 is a space in which the casting C can be naturally cooled. However, the cooling method of the casting C in the inspection space D1 is not particularly limited, and as a cooling method, for example, a forced air cooling method in which air is blown to the casting C using a cooling fan or the like may be adopted. In FIG. 1, the inspection space D1 is provided with a rotating body D11 for placing the casting C and rotating it around the central axis CX. Although the rotating body D11 is not particularly limited, in this embodiment, the rotating body D11 is arranged on both sides of the central axis CX of the casting C and is composed of a plurality of rollers that can rotate around a rotation axis substantially parallel to the central axis CX.
[0023] The thermal image pickup device D2 is not particularly limited as long as it can acquire a thermal image P showing the temperature distribution of the subject (casting C) in some form. In this embodiment, as shown in FIG. 1, it is a thermography imaging device that acquires a thermal image P (see also an example of the thermal image P shown in FIG. 3) based on infrared rays radiated from the subject. In FIG. 1, the thermal image pickup device D2 is provided so as to be able to acquire a thermal image P of the casting C in a high temperature state arranged in the inspection space D1. Specifically, the thermal image pickup device D2 is provided in the peripheral space DR of the inspection space D1. More specifically, a plurality of thermal image pickup devices D2 are provided so as to be juxtaposed approximately parallel to the central axis CX of the casting C in the peripheral space DR of the inspection space D1. In FIG. 1, a thermal image pickup device D20 for imaging the main body C0 of the casting C, a thermal image pickup device D21 for imaging the receiving port C1 of the casting C, and a thermal image pickup device D22 for imaging the insertion port C2 of the casting C are juxtaposed. However, the arrangement of the thermal image capturing device D2 is not particularly limited, and only one thermal image capturing device D2 may be provided, or the thermal image capturing device D2 may be provided movably in the surrounding space DR.
[0024] The determination device D3 assists an operator who uses the inspection system D in determining whether the casting C is defective or not, or determines whether the casting C is defective or not without the operator. The determination device D3 is not particularly limited, but in this embodiment, the determination device D3 is connected to the thermal image pickup device D2 and determines whether the casting C is defective or not based on the thermal image P acquired by the thermal image pickup device D2. The determination device D3 is also not particularly limited, but in this embodiment, the determination device D3 is connected to an alarm mechanism D6 and generates a defective signal FS and transmits it to the alarm mechanism D6 when it is determined that the casting C is defective.
[0025] The determination device D3 is configured by an information processing device including a processor such as a CPU (Central Processing Unit) and a storage unit such as a ROM (Read Only Memory) and a RAM (Random Access Memory). The CPU of the determination device D3 executes a determination program stored in the ROM or the like using the RAM or the like as a working area, thereby functioning as each of the functional units D30 to D33 described later. The function of the determination device D3 is not particularly limited, but in this embodiment, the determination device D3 uses a thermal image P acquired from a reference casting as a reference thermal image Pr and determines the quality of the casting C based on a comparison result between the thermal image P and the reference thermal image Pr. An example of the reference casting is a good casting that is the same type as the casting C to be inspected and has no defects. In order to achieve the above function, in FIG. 1, the determination device D3 includes a comparison unit D30 that compares the thermal image P with the reference thermal image Pr and a determination unit D31 that determines the quality of the casting C. The determination device D3 may be equipped with an artificial intelligence function that learns a characteristic portion Pf (see FIG. 3) that appears in the thermal image P corresponding to a failure mode F (see FIG. 2) of the casting C, and determines the quality of the casting C based on the learning result of the characteristic portion Pf. To achieve the above function, as shown in FIG. 1, the determination device D3 may have an extraction unit D32 that extracts the characteristic portion Pf of the failure mode F of the casting C in the thermal image P, and a model generation unit D33 that learns the characteristic portion Pf of the failure mode F of the casting C, and generates a failure model M for the failure mode F based on the learning result of the characteristic portion Pf.
[0026] The comparison unit D30 generates a differential thermal image Pd by taking the difference between the thermal image P and the reference thermal image Pr, for example. In other words, the comparison unit D30 takes the difference in the gradation value for each pixel data of the thermal image P and the reference thermal image Pr to generate the differential thermal image Pd. Note that the term "thermal image P" may also refer to images that have been processed based on the thermal image P acquired by the thermal image pickup device D2, such as the differential thermal image Pd.
[0027] The judgment unit D31 judges the quality of the casting C based on the presence or absence of a characteristic portion Pf (see also FIG. 3) of the temperature distribution of the casting C that appears as light and dark in the thermal image P. The judgment unit D31 may judge the quality of the casting C based on the differential thermal image Pd generated by the comparison unit D30 described above, instead of the thermal image P. For example, if the casting C to be inspected is a cast pipe (see also FIG. 2) and has grooves in the circumferential direction on the outer and inner walls of the receiving port C1 and the insertion port C2, the shape of the grooves may be reflected in the thermal image P. In such a case, if the quality of the casting C is judged based on the differential thermal image Pd, the shape of the grooves of the casting C will not appear in the differential thermal image Pd, and the characteristic portion Pf will stand out more, making it easier to judge the quality of the casting C.
[0028] The extraction unit D32 extracts a characteristic portion Pf of the thermal image P based on the thermal image P (or the differential thermal image Pd) (see also FIG. 3). The extraction form of the characteristic portion Pf by the extraction unit D32 is not particularly limited, but for example, the extraction unit D32 extracts the shape of the characteristic portion Pf by performing image recognition on the shape of the area where shading occurs in the thermal image P (or the differential thermal image Pd) (specifically, the shape of the area in the thermal image P where the gradation value is outside a predetermined range, and in the differential thermal image Pd, the shape of the area where the differential gradation value between the thermal image P and the reference thermal image Pr is outside a predetermined range). However, the extraction unit D32 may extract the degree of shading of the characteristic portion Pf (specifically, the gradation value in the data of the thermal image P, and in the differential thermal image Pd, the differential gradation value between the thermal image P and the reference thermal image Pr).
[0029] The model generating unit D33, for example, learns the shape and gradation value (or differential gradation value) of the characteristic part Pf of the stored thermal image P (or differential thermal image Pd) to generate a defect model M related to the shape and gradation value of the defect mode F (see FIG. 2). The defect model M is not particularly limited, but is generated as a probability distribution of the outline and shading of the characteristic part Pf. The learning method of the characteristic part Pf can be selected from various machine learning methods such as deep learning, random forest, SVM (Support Vector Machine), and GBDT (Gradient Boosting Decision Tree). When the model generating unit D33 is provided, the determining unit D31 may determine the quality of the casting C by comparing the characteristic part Pf of the thermal image P with the shape or gradation value (or differential gradation value) of the defect model M. The defect model M is appropriately regenerated according to the accumulation of a new thermal image P (or a new differential thermal image Pd). In this case, the defect model M is regenerated as a more appropriate defect model, and the accuracy of the quality determination of the casting C by the determination device D3 is improved.
[0030] The display device D4 can display the above-mentioned reference thermal image Pr, differential thermal image Pd, and defect model M in addition to the thermal image P. The display device D4 is also used, for example, when an operator using the inspection system D visually judges the quality of the casting C based on the thermal image P (or the differential thermal image Pd), etc. The display device D4 is not particularly limited, and is configured, for example, by a known liquid crystal display device. In the example of FIG. 1, the display device D4 is connected to the determination device D3. However, when the determination device D3 is not provided in the inspection system D, the display device D4 may be connected to the thermal image pickup device D2.
[0031] The database D5 can store the above-mentioned reference thermal image Pr, differential thermal image Pd, and defect model M, as necessary, in addition to the thermal image P. The database D5 is not particularly limited, and is composed of, for example, a known HDD (Hard Disk Drive) or SSD (Solid State Drive). In the example of FIG. 1, the database D5 is connected to the determination device D3. However, when the determination device D3 is not provided in the inspection system D, the database D5 may be connected to the thermal image capturing device D2.
[0032] The notification mechanism D6 notifies the worker or the like who uses the inspection system D of the occurrence of a defect in the casting C. In this embodiment, the notification mechanism D6 notifies the worker or the like of the occurrence of a defect in the casting C when it receives a defect signal FS transmitted from the determination device D3. The notification mechanism D6 is not particularly limited, but is composed of, for example, a speaker that notifies the occurrence of a defect by an alarm sound, or a lamp that notifies the occurrence of a defect by lighting or blinking. In addition to or instead of notifying the worker or the like of the occurrence of a defect by the notification mechanism D6, the determination device D3 may be configured to interrupt or stop the work of a step prior to the inspection step. In this case, the determination device D3 is connected to a device of the previous step (such as the casting machine A in this embodiment) and stops the device of the previous step by generating and transmitting a defect signal FS.
[0033] [Casting inspection method according to one embodiment of the present invention] Next, a method for inspecting a casting according to an embodiment of the present invention will be described with reference to the accompanying drawings. Note that the embodiment described below is merely an example, and the method for inspecting a casting according to the present invention is not limited to the embodiment described below.
[0034] In the method for inspecting a casting C according to the present embodiment, a thermal image P of the casting C is acquired before the casting C is cooled from a high temperature state to a room temperature state, and the quality of the casting C is judged based on the acquired thermal image P. Specifically, the casting C is inspected for any casting defects. In the present embodiment, the above-mentioned inspection system D for the casting C is used to inspect the casting C. As illustrated in FIG. 2, examples of defect modes F of the casting C include cold shut F1, melt spill F2, rough surface F3, neck cut F4, and broken insert F5. For example, cold shut F1 is a defect mode in which a part where molten metal joins appears as a wrinkled shape. For example, melt spill F2 is a defect mode in which molten metal splashes in the mold and the splashed part appears as a boundary with other parts. For example, rough surface F3 is a defect mode that appears as a rough surface on the surface of the cast body. For example, neck cut F4 is a defect mode that appears as a cut or wrinkle near the mouth of the casting pipe. For example, the insertion break F5 is a defect mode that appears as a cut or wrinkle near the insertion port of the cast pipe. However, the defect mode of the casting C to be inspected is not particularly limited as long as it is a defect mode whose occurrence can be determined by the thermal image P, and the occurrence of a defect mode other than the above-mentioned defect mode F may be determined.
[0035] As shown in Fig. 4, the method for inspecting a casting C according to this embodiment includes a step of acquiring a thermal image P of the casting C (hereinafter also referred to as "step S1"), and a step of determining whether the casting C is good or bad (hereinafter also referred to as "step S2"). Below, the method will be described in more detail with reference to Figs. 1 to 4, including steps SA and SB before and after the inspection step S of the casting C.
[0036] (Pre-step in the casting inspection method) In this embodiment, as shown in FIG. 4, the pre-step SA before the inspection of the casting C (see FIG. 1) includes step SA1 of melting the raw material of the casting C, step SA2 of casting the casting C, and step SA3 of preparing to obtain a thermal image P of the casting C. Specifically, in step SA1, the raw material of the casting C is charged into a melting furnace (not shown) to obtain molten raw material. The raw material of the casting C may be appropriately changed according to the properties required for the casting C, but in this embodiment, a predetermined raw material is charged into the melting furnace (not shown) to obtain the casting C made of ductile cast iron. For example, in step SA2, as shown in FIG. 1, the casting C is cast by a casting machine A (specifically, a centrifugal casting machine). For example, in step SA3, as shown in FIG. 1, the casting C cast by the casting machine A is taken out by a drawing machine (not shown) or the like, and then the casting C is moved to the inspection space D1 by a conveying device (not shown) such as a crane or a conveyor in order to execute step S1.
[0037] (Step of acquiring a thermal image of the casting) In step S1 for acquiring a thermal image P of the casting C, a thermal image P (see FIG. 3) of the casting C is acquired while the casting C is cooled from a high temperature state to a room temperature state (see FIG. 4). In this embodiment, as shown in FIG. 1, a thermal image P is acquired while the casting C is cooled from a high temperature state by centrifugal casting to a room temperature state. Specifically, the thermal image P of the casting C is acquired by imaging the casting C arranged in the inspection space D1 with the thermal image imaging device D2. When a rotating body D11 is provided in the inspection space D1, the casting C is placed on the rotating body D11, and the casting C is rotated by the rotating body D11 and the casting C is imaged by the thermal image imaging device D2 to acquire a thermal image P. In this case, the surface of the casting C, which is initially facing vertically downward and difficult to image by the thermal image imaging device D2, is rotated by the rotating body D11 to an angle that makes it easy for the thermal image imaging device D2 to image, so that the entire circumferential surface of the casting C can be imaged.
[0038] The thermal image P to be acquired is a still image in this embodiment, although it is not particularly limited. In other words, in this embodiment, the thermal image P is acquired at a predetermined time point. The predetermined time point at which the thermal image P is acquired is not particularly limited as long as it is between the time when the casting C is cooled from a high temperature state to a room temperature state. The predetermined time point may be a plurality of time points. Specifically, for example, a first thermal image P1 is acquired at a first time point, and a second thermal image P2 is acquired at a second time point after a time has elapsed from the first time point. However, the thermal image P may be a video continuously acquired in a predetermined time period. In FIG. 1, in order to image the entire casting C, the thermal images P of the main body C0, the receiving port C1, and the insertion port C2 of the casting C are each captured by a plurality of thermal image capturing devices D20 to D22. However, the thermal image P may be in another form, such as one thermal image P that captures the entire casting C.
[0039] In this embodiment, the casting C in a high temperature state in step S1 is obtained by a casting step SA2 of the casting C as shown in FIG. 3 (see also FIG. 1). When the casting C is produced by centrifugal casting as in this embodiment, the casting C is pulled out from the cylindrical mold of the casting machine A while still in a high temperature state (for example, 800° C.), so that it is suitable to take a thermal image P by the thermal image pickup device D2. However, the casting C in a high temperature state is not limited to that obtained by the casting step SA2 of the casting C, and may be obtained by any other heat treatment step. The heat treatment step by which the casting C in a high temperature state is obtained includes, for example, an annealing step SB1 described later, which is included in the manufacturing step of the casting C, as well as any heat treatment step after the manufacturing step of the casting C is completed.
[0040] The inspection space D1 where step S1 is performed is not particularly limited, but in this embodiment, as shown in FIG. 1, it is arranged near the casting machine A where the previous step SA is performed in order to shorten the moving distance of the casting C from the previous step SA to step S1. In this case, the thermal image P can be acquired by the thermal image acquisition device D2 at an appropriate time before the casting C, which has become in a high temperature state by casting, is excessively cooled. However, the position of the inspection space D1 is not particularly limited. For example, when the inspection step S of the casting C is performed not immediately after the casting step SA2 of the casting C but immediately after the annealing step SB1 of the casting C described later, the inspection space D1 may be arranged in another space, for example, between the annealing furnace B of the annealing step SB1 of the casting C and the painting device or shape processing device (not shown) of the step SB2 of the casting C, which is the step immediately after the annealing step SB1.
[0041] (Step to judge the quality of the casting) In step S2 for determining the quality of the casting C, the quality of the casting C is determined based on the thermal image P acquired in step S1 (see FIG. 4). Specifically, the quality of the casting C is determined by the determination unit D31 of the determination device D3 described above based on the thermal image P acquired by the thermal image pickup device D2. In the thermal image P shown in FIG. 3, a characteristic portion Pf of the defect mode F (specifically, the above-mentioned cold shut F1) of the casting C appears as a linear portion extending in the circumferential direction about the central axis CX of the casting C. For example, the determination unit D31 can determine the quality of the casting C by determining the presence or absence of such a characteristic portion Pf in the thermal image P. In order to make the characteristic portion Pf more prominent in the thermal image P, the thermal image P acquired from a reference casting can be used as a reference thermal image Pr, and the quality of the casting C can be determined based on a comparison result between the thermal image P acquired in step S1 and the reference thermal image Pr. In this case, the quality of the casting C is judged by the judging unit D31 based on the differential thermal image Pd generated by the comparing unit D30 of the judgment device D3 described above. However, the quality of the casting C may be judged by the thermal image P by an operator of the inspection system D by visually inspecting the casting C using the thermal image P, or by visually inspecting the casting C using the differential thermal image Pd generated by the comparing unit D30 of the judgment device D3.
[0042] In step S2, the presence or absence of a plurality of defect modes F (see FIG. 2) in the casting C may be determined based on a plurality of thermal images P acquired at a plurality of time points. Specifically, the presence or absence of a first defect mode in the casting C may be determined based on a first thermal image P1 acquired in step S1, and the presence or absence of a second defect mode in the casting C may be determined based on a second thermal image P2 acquired in step S1. For example, among the above-mentioned defect modes F (see FIG. 2), for defect modes F that tend to appear as wrinkles or roughness on the surface of the casting C (e.g., cold shut F1, melt spill F2, rough surface F3, etc., which are diagrammatically shown in FIG. 2), the corresponding characteristic portion Pf may be more prominent in a thermal image P captured when the casting C is in a high temperature state than when it is in a low temperature state. Conversely, for example, among the above-mentioned defect modes F, for defect modes F that tend to appear as grooves engraved in the casting C (e.g., neck break F4 and insert break F5, which are shown in FIG. 2 ), the characteristic portion Pf may be more prominent in a thermal image P captured when the casting C is in a low temperature state than when it is in a high temperature state. In such a case, when determining the presence or absence of the defect mode F, the presence or absence of the defect mode F, such as a cold shut F1, a melting point F2, or a rough surface F3, is determined by using a thermal image (e.g., a first thermal image P1) acquired when the casting C is in a high temperature state (e.g., a first time point). Also, the presence or absence of the defect mode F, such as a neck break F4 and insert break F5, is determined by using a thermal image (e.g., a second thermal image P2) acquired when the casting C is in a low temperature state after cooling (e.g., a second time point). In this way, the quality of the casting C can be determined using thermal images P1 and P2 acquired at an appropriate time when the defect mode F is likely to stand out as a characteristic portion Pf of the thermal image P, thereby improving the inspection accuracy of the casting C.
[0043] In step S2, a characteristic portion Pf appearing in the thermal image P corresponding to a defect mode F (see FIG. 2) of the casting C may be learned, and the quality of the casting C may be judged based on the learning result of the characteristic portion Pf. In this case, the judgment device D3 may be equipped with an artificial intelligence function as described above, and may have an extraction unit D32 and a model generation unit D33 as shown in FIG. 1. The extraction unit D32 extracts the characteristic portion Pf of the thermal image P based on the thermal image P (or the differential thermal image Pd), and the model generation unit D33 learns the shape and gradation value (or differential gradation value) of the characteristic portion Pf of the accumulated thermal image P (or the differential thermal image Pd), and generates a defect model M related to the shape and gradation value of the defect mode F. For example, the judgment unit D31 judges the quality of the casting C by comparing the characteristic portion Pf of the thermal image P with the shape or gradation value (or differential gradation value) of the defect model M. In this way, by appropriately regenerating the defect model M in response to the accumulation of new thermal images P (or new differential thermal images Pd), the accuracy of the quality determination of the casting C by the determination device D3 is improved.
[0044] The casting C determined to be defective in step S2 may not be subjected to the subsequent step SB (see FIG. 4). Specifically, the casting C determined to be defective may be, for example, crushed and fed into a melting furnace as a raw material for the casting C in step SA1.
[0045] (Later step in the casting inspection method) In this embodiment, the post-step SB after the inspection of the casting C includes a step SB1 of annealing the casting C and a step SB2 of processing the casting C (see FIG. 4). Specifically, in step SB1, the casting C determined to be a non-defective product in the above inspection is put into an annealing furnace B and subjected to a predetermined annealing treatment (for example, 800°C to 900°C). Thereafter, in step SB2, the casting C is subjected to a predetermined processing treatment such as surface painting and shaping of the receiving port C1 and the insertion port C2. In this embodiment, since there is no waste of performing the post-step SB on the casting C determined to be defective in step S2, the manufacturing cost of the casting C can be reduced.
[0046] According to the inspection method and inspection system D for a casting C according to the present embodiment, which includes the above steps, a thermal image P of the casting C is acquired before the casting C is cooled from a high temperature state to a room temperature state, and the quality of the casting C is judged based on the acquired thermal image P. In this way, the defect mode F of the casting C appears as a characteristic portion Pf in the thermal image P in a form that is easy to visually recognize. Therefore, an operator can clearly identify the quality of the casting C regardless of his / her skill level, and the inspection of the casting C can be automated using the judgment device D3 or the like. In addition, for example, by remotely operating the thermal image pickup device D2, an operator can inspect the casting C without approaching the casting C in a high temperature state, so that the inspection method and inspection system D for a casting C according to the present embodiment is also excellent in safety.
[0047] In this embodiment, in step S1 of casting a casting C by centrifugal casting and acquiring a thermal image P, the thermal image P is acquired while the casting C is cooled from the high temperature state produced by centrifugal casting to room temperature. In this case, the casting C is usually taken out of the mold of the centrifugal casting machine A while still in a high temperature state, which is suitable for taking the thermal image P. Also, acquiring and inspecting the thermal image of the casting C immediately after casting prevents a defective casting C from being sent to a subsequent step SB (e.g., annealing step SB1 or processing step SB2) or prevents defective castings C from being continuously cast.
[0048] In this embodiment, in step S2 of determining the quality of the casting C, a thermal image P acquired from a reference casting is set as a reference thermal image Pr, and the quality of the casting C is determined based on a comparison result between the thermal image P and the reference thermal image Pr. In this case, the comparison between the thermal image P and the reference thermal image Pr makes the characteristic portion Pf of the defect mode F of the casting C more prominent, thereby further improving the inspection accuracy of the casting C.
[0049] In this embodiment, in step S2 of determining the quality of the casting C, a characteristic portion Pf appearing in the thermal image P corresponding to a defect mode F of the casting C is learned, and the quality of the casting C is determined based on the learning result of the characteristic portion Pf. In this manner, the defect model M is optimized by appropriately learning the characteristic portion Pf in response to the accumulation of new thermal images P, thereby further improving the inspection accuracy of the casting C.
[0050] In this embodiment, in step S1 of acquiring a thermal image P, a first thermal image P1 is acquired at a first time point, and a second thermal image P2 is acquired at a second time point, and in step S2 of judging the quality of the casting C, the presence or absence of a first failure mode in the casting C is judged based on the first thermal image P1, and the presence or absence of a second failure mode in the casting C is judged based on the second thermal image P2. In this case, the first thermal image P1 and the second thermal image P2 can be acquired by selecting appropriate times at which the failure mode F is likely to stand out in the thermal image P as the first time point and the second time point. Therefore, the quality of the casting C can be judged using the first thermal image P1 and the second thermal image P2 at which the failure mode F stands out, and the inspection accuracy of the casting C is further improved.
[0051] In this embodiment, the inspection system D is disposed near the centrifugal casting machine A capable of casting the casting C. In this case, the casting C in a high temperature state removed from the mold of the centrifugal casting machine A can be easily disposed in the inspection system D while still in a high temperature state, which is suitable for capturing a thermal image P. [Explanation of symbols]
[0052] A (Centrifugal) Casting Machine B Annealing furnace C Casting C0 Main Unit C1 socket C2 socket CX center axis D1 Inspection space D11 Rotating body D2, D20~D22 Thermal imaging device D3 Judging device D30 comparison section D31 Judgment section D32 Extraction part D33 Model Generation Unit D4 display device D5 Database D6 Notification mechanism DR Surrounding space F Failure Mode F1 Yubakai F2 Spilled water F3 Rough skin F4 Decapitation F5 Insert broken FS bad signal M Defective Model P, P1, P2 Thermal Images Pd differential thermal image Pf Characteristic part Pr Reference Thermal Image Pre-SA Step SA1 Lysis Step SA2 Casting Step SA3 Preparation Steps S Inspection Step S1 Thermal image acquisition step S2: Pass / fail judgment step SB Rear Step SB1 Annealing step SB2 Processing Step
Claims
1. acquiring a thermal image of the cast pipe during a period in which the cast pipe, which has been heated by casting, is cooled to room temperature; determining whether the cast pipe is good or bad based on the thermal image; Including, The step of acquiring a thermal image includes: acquiring a first thermal image at a first time; acquiring a second thermal image at a second time point that is a time lapse from the first time point; The step of determining whether the product is good or bad includes: determining whether or not a first defect mode, which is likely to appear as wrinkles or roughness on the surface of the cast pipe, is present based on the first thermal image; a second thermal image showing the presence or absence of a second defect mode which is likely to appear as a groove engraved in the cast pipe;
2. The cast pipe is cast by centrifugal casting; 2. The method for inspecting a cast pipe according to claim 1, wherein in the step of acquiring the thermal image, the thermal image is acquired while the cast pipe is being cooled from the high temperature state produced by the centrifugal casting to room temperature.
3. 3. A cast pipe inspection method as claimed in claim 1 or 2, wherein in the step of determining whether the cast pipe is pass or fail, a thermal image obtained from a reference cast pipe is used as a reference thermal image, and the quality of the cast pipe is determined based on a comparison result between the thermal image and the reference thermal image.
4. 4. A cast pipe inspection method according to claim 1, wherein in the step of determining whether the cast pipe is pass or fail, characteristic parts that appear in the thermal image corresponding to accumulated defect modes of the cast pipe are learned, and the pass or fail of the cast pipe is determined based on the learning results of the characteristic parts.
5. A cast pipe inspection system for carrying out the cast pipe inspection method according to any one of claims 1 to 4, comprising: The inspection system includes: an inspection space in which the casting pipe in a high temperature state can be placed and which can be cooled; a thermal image capturing device that is provided so as to be capable of acquiring a thermal image of the cast pipe in a high temperature state that is placed in the inspection space; A cast pipe inspection system comprising:
6. The cast pipe inspection system according to claim 5 , wherein the inspection space is disposed near a centrifugal casting machine capable of casting the cast pipe.
7. The cast pipe inspection system according to claim 5 or 6, further comprising a determination device for determining whether the cast pipe is good or bad.
8. 8. The cast pipe inspection system according to claim 7, wherein the judgment device uses a thermal image acquired from a reference cast pipe as a reference thermal image, and judges whether the cast pipe is good or bad based on a comparison result between the thermal image and the reference thermal image.
9. 8. The cast pipe inspection system of claim 7, wherein the judgment device is equipped with an artificial intelligence function that learns characteristic parts that appear in the thermal image corresponding to accumulated defect modes of the cast pipe, and judges whether the cast pipe is good or bad based on the learning results of the characteristic parts.
Citation Information
Patent Citations
Apparatus for casting tube
JP1996141721A
Inspection method of cast
JP2012107938A
Method for inspecting casting article
JP2012110910A
THERMOGRAPHIC INSPECTION MEANS AND NON-DESTRUCTIVE INSPECTION METHOD FOR NEAR SURFACE STRUCTURE OF TEST SUBJECT
JP2017527813A
Product quality defect prediction system
JP2020140365A
