Carburized product hardness prediction device and method

Abstract

To provide a technique for predicting hardness of a carburized product without repeatedly producing and evaluating a prototype of the carburized product.SOLUTION: A carburized product hardness prediction device comprises a computing element. The computing element is constituted to acquire shape data including information of an outer shape of a product or a product intermediate product, acquire heat treatment condition data including information of conditions of heat treatment including carburization hardening to a workpiece as an intermediate product, predict hardness date from the shape date and the heat treatment condition data using a prediction model after mechanical-learning to predict a target variable from an explanatory variable while the shape data and the heat treatment condition data are explanatory variables and the hardness data including information of hardness of the workpiece after heat treatment is a target variable, and output the predicted hardness data.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present disclosure relates to a technique for predicting the surface hardness of a carburized product whose surface is hardened by carburizing and quenching.

Background Art

[0002] Carburizing and quenching is a heat treatment in which carbon is infiltrated and diffused into a steel material from the surface and then quenching is performed. A product subjected to carburizing and quenching has both wear resistance due to a hard surface layer and toughness due to a soft internal structure. Due to such characteristics, conventionally, for products to which a high load is applied, such as gears which are mechanical parts, carburizing and quenching has been performed for the purpose of improving wear resistance by surface hardening. For example, Patent Document 1 discloses a carburized gear made of a steel material containing 0.15 - 0.25 mass% carbon and having a hardened surface layer by carburizing and quenching.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the development stage of carburized products, an accurate evaluation is required to determine whether the products meet the required surface hardness and internal hardness distribution. However, when carburizing and quenching is performed on a member with a complex surface shape such as a gear, the carburizing depth and the amount of carbon infiltrated may become non-uniform. Furthermore, among carburized products, there are those with different required surface hardnesses depending on the part. Therefore, it is difficult to accurately evaluate the surface hardness and effective case depth of the product from the design drawing. Moreover, after heat treatment including carburizing and quenching, the workpiece is finished by machining such as cutting and polishing. However, the machining allowance during finishing changes depending on the deformation behavior during heat treatment and the degree of internal hardness, and the difficulty of finishing (hereinafter referred to as manufacturability) changes. Manufacturability is not sufficiently considered in the design. For these reasons, conventionally, in the development stage of carburized products with complex surface shapes, in order to obtain a combination of design and manufacturing conditions that meet the quality requirements of the products and consider manufacturability, the design and the production and evaluation of prototype products are repeated many times. This is one of the factors that cause a huge amount of time to be spent on product development.

[0005] The present disclosure has been made in view of the above circumstances, and an object thereof is to provide a technique for predicting the hardness of carburized products without repeating the production and evaluation of prototype carburized products.

Means for Solving the Problems

[0006] In order to solve the above problems, a hardness prediction device for carburized products according to one aspect of the present disclosure is a hardness prediction device for carburized products that predicts the hardness of a product whose surface is hardened by carburizing and quenching, and includes: equipped with an arithmetic unit, wherein the arithmetic unit acquires shape data including information on the outer shape of the product and the intermediate body of the product, acquires heat treatment condition data including information on the conditions of heat treatment including the carburizing and quenching of the workpiece which is the intermediate body, Using the shape data and the heat treatment condition data as explanatory variables, and using a prediction model obtained by machine learning to predict the hardness data including the hardness information of the workpiece after heat treatment from the explanatory variables, it is configured to output the predicted hardness data.

[0007] Also, a method for predicting the hardness of a carburized product according to an aspect of the present disclosure is a method for a computer to predict the hardness of a product whose surface is hardened by carburizing quenching, acquiring shape data including information on the outer shape of the product and an intermediate of the product, acquiring heat treatment condition data including information on the conditions of heat treatment including the carburizing quenching of the workpiece as the intermediate, using the shape data and the heat treatment condition data as explanatory variables, and using a prediction model obtained by machine learning to predict the hardness data from the shape data and the heat treatment condition data, with the hardness data including the hardness information of the workpiece after heat treatment as the target variable, and outputting the predicted hardness data.

Advantages of the Invention

[0008] According to the present disclosure, it is possible to provide a technique for predicting the surface hardness of a carburized product without repeatedly manufacturing and evaluating test pieces of the carburized product.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

[0010] Next, embodiments of the present disclosure will be described with reference to the drawings. In this embodiment, the hardness prediction technique according to the present disclosure is applied to a gear, which is an example of a carburized product. Here, the carburized product means a product whose surface is hardened by carburizing and quenching. However, the hardness prediction technique according to the present disclosure is not limited to gears and can be widely applied to heavy machine parts and machine parts whose surfaces are hardened by carburizing and quenching.

[0011] FIG. 1 is a diagram for explaining a method of manufacturing a gear. As shown in FIG. 1, the method of manufacturing a gear includes a forging process (step S1), a rough shaping process (step S2), a heat treatment process (step S3), and a finishing process (step S4).

[0012] In the forging process (step S1), a bar made of a predetermined gear material is cut to obtain a short cylindrical blank, and the blank is hot forged to obtain a forged workpiece formed into a hollow disk shape. If necessary, the forged workpiece is subjected to a normalizing treatment. Examples of the gear material include carburizing steels such as carbon steel (S09CK, S15CK, S20CK, S45C), nickel-chromium-molybdenum-manganese steel (AMS6265, AMS6260, AMS6308), chromium steel (SCr415, SCr420), chromium-molybdenum steel (SCM440, SCM415, SCM420), nickel-chromium steel (SNC815), and nickel-chromium-molybdenum steel (SNCM220, SNCM420, SNCM439).

[0013] In the rough shaping process (step S2), the forged workpiece is machined by turning, broaching, rough tooth cutting, chamfering, shaving, etc. to obtain a rough shaped workpiece shaped into a gear shape leaving the machining allowance required for finishing.

[0014] In the heat treatment process (step S3), the rough-formed workpiece is subjected to treatments such as carburizing, quenching, sub-zero treatment, and tempering. Known types of carburizing methods include solid carburizing, liquid carburizing, drip-feed carburizing, vacuum carburizing, gas carburizing, plasma carburizing, etc. In this embodiment, gas carburizing is adopted. However, carburizing is not limited to gas carburizing, and other methods may also be used. Before the heat treatment, if there are parts of the rough-formed workpiece that are not to be carburized and hardened, those parts are coated with a carburizing inhibitor. The parts that are not to be carburized and hardened are, for example, the tooth tip surfaces of gears.

[0015] In carburizing, the workpiece is held in a carburizing furnace at a predetermined treatment temperature for a predetermined treatment time. Hereinafter, the treatment temperature of carburizing is referred to as the "carburizing temperature", and the treatment time of carburizing is referred to as the "carburizing time". For example, the carburizing temperature is 900°C or higher and 1050°C or lower, and the carburizing time is 60 minutes or longer and 600 minutes or shorter. Inside the carburizing furnace, by supplying carburizing gas, the carbon potential value (i.e., the carbon concentration of the atmosphere) is maintained higher than the carbon concentration of the workpiece. The carburizing gas is, for example, a gas containing carbon dioxide, hydrogen, methane, and water vapor. Carbon atoms in the atmosphere inside the carburizing furnace penetrate into the workpiece from the surface and gradually diffuse inward. Due to this penetration and diffusion of carbon, a layer with a higher carbon content ratio (i.e., a carburized and hardened layer) is formed on the surface of the workpiece compared to other parts. The carbon potential value, carburizing temperature, and carburizing time of carburizing are adjusted according to the composition of the workpiece, the shape of the workpiece, and the characteristics of the carburizing furnace.

[0016] Quenching is performed on the workpiece after carburizing. In quenching, the workpiece is rapidly cooled by oil cooling or water cooling. Tempering is performed on the workpiece after quenching. In tempering, the workpiece is held at a predetermined tempering temperature for a predetermined tempering time and then air-cooled. For example, the tempering temperature is 150°C or higher and 250°C or lower, and the tempering time is 30 minutes or longer and 180 minutes or shorter.

[0017] The workpiece after heat treatment has a carburized hardened layer on its surface, and this carburized hardened layer appears on the surface of the gear which is the product. The workpiece after heat treatment is located inside the carburized hardened layer and has a core part with a lower carbon content than the carburized hardened layer. The composition of this core part is substantially the same as the composition of the workpiece before heat treatment. The carburized hardened layer has a high hardness compared to the core part. In this gear, the carburized hardened layer contributes to wear resistance, and the core part contributes to toughness.

[0018] In the finishing process step (step S4), for the workpiece after heat treatment, the oxide film on the surface is removed and a predetermined allowance is cut off to finish the surface of the workpiece and finish the workpiece into the shape of the finished product to obtain the product.

[0019] The hardness prediction device 1 and method according to the present disclosure predict the hardness of the gear manufactured as described above at the product development stage. The hardness information includes information specifying the hardness of the surface and surface layer of the product, such as surface hardness and cross-sectional hardness distribution. The hardness information may further include information specifying the hardness inside the product. Hereinafter, the hardness prediction device 1 will be described in detail.

[0020] 《Hardness Prediction Device》 FIG. 2 is a block diagram showing a schematic hardware configuration of the hardness prediction device 1. As shown in FIG. 2, the hardness prediction device 1 can be realized by, for example, one or more computers 40. The computer 40 includes an arithmetic unit 41 such as a CPU (Central Processing Unit), a memory 42, an auxiliary storage device 43, a communication I / F 44 for connecting to a communication network by wire or wirelessly, an input device 45 such as a mouse, keyboard, touch panel, and an output device 46 such as a display and printer, a media I / F 47 for reading and writing information to and from a portable storage medium, etc. Each function of the hardness prediction device 1 described later can be realized by loading a predetermined program stored in the auxiliary storage device 43 into the memory 42 and executing it by the arithmetic unit 41. The above-mentioned predetermined program may be downloaded from the outside via a communication network connected to the communication I / F 44, or may be loaded from a storage medium connected to the media I / F 47.

[0021] Figure 3 is a functional block diagram of the hardness prediction device 1. As shown in Figure 3, the hardness prediction device 1 includes functional units such as a data acquisition unit 21, a hardness prediction unit 22, an evaluation unit 23, and a learning unit 24.

[0022] 〔Data acquisition unit 21〕 The data acquisition unit 21 acquires shape data, heat treatment condition data, finishing process data, and required quality data. The data acquisition unit 21 can read and acquire these data from a storage medium connected to the auxiliary storage device 43 or the media I / F 47, acquire data input via the input device 45, or acquire data from the outside via a communication network connected to the communication I / F 44.

[0023] The shape data includes information related to the outer shape of the product and the intermediate body of the product. In this embodiment, the product is a gear, and the shape data of the gear includes information related to the tooth shape and dimensions of the gear. As a specific example, as shown in Figure 4, the shape data of the gear includes a module m and a pressure angle α. The module m is a value representing the size of the tooth and is the value obtained by dividing the pitch p by the pi π. The pitch p is the distance to the adjacent tooth on the reference circle. Since the module m is also the value obtained by dividing the reference circle diameter d by the number of teeth z, the shape data may include the reference circle diameter d and the number of teeth z instead of the module m. The pressure angle α is the angle at which the tooth surface is inclined with respect to the normal of the reference circle. Further, the shape data includes the tooth tip width w of the intermediate body before finishing after heat treatment. The tooth tip width w is the circumferential dimension of the tooth tip surface. The tooth tip and the tooth tip surface are locations where excessive carburization is likely to occur, and the tooth tip width w is one of the important parameters in predicting the amount of infiltrated carbon and the hardened layer depth.

[0024] The shape data may further include the root fillet radius ρf of the intermediate body after heat treatment and before finishing. The root fillet radius ρf is the radius of curvature of the root fillet portion, that is, the intersection of the tooth surface and the root. The root fillet portion is a location where the amount of carbon infiltrated tends to be sparse, and the root fillet radius ρf is one of the important parameters in predicting the amount of infiltrated carbon and the hardened layer depth. Also, in the finishing process, since the root fillet portion of the workpiece after heat treatment is a relatively large portion to be machined, the root fillet radius ρf is one of the important parameters for more accurately evaluating the difficulty of finishing the root fillet portion.

[0025] The shape data may include information related to the evaluation points of the product. The evaluation points of the product are points defined on the surface of the product. For example, as shown in FIG. 4, for a gear as a product, (i) the tip surface, (ii) the flank pitch, (iii) the root fillet portion, and (iv) four evaluation points at the root are set.

[0026] The heat treatment condition data includes the carbon potential value (Cp value) of carburizing, the carburizing temperature, and the carburizing time. Although there are other parameters in the heat treatment conditions, in particular, the influence of the carbon potential value, the carburizing temperature, and the carburizing time on the surface hardness is significant. The heat treatment condition data may further include the quenching temperature.

[0027] The finishing data includes information on the machining allowance to be removed from the surface of the workpiece in the finishing process. The machining allowance information includes the maximum machining allowance and the minimum machining allowance for each part of the product. The machining allowance value varies depending on the part even in one product.

[0028] The required quality data includes information on the hardness required for the product. The information on the hardness required for the product includes at least the hardness of the surface of the product, and may also include the hardness of the interior of the product. The required quality data for a gear includes information on the hardness required for each part of the gear. In a gear, usually, the hardness required for each part is different.

[0029] [Hardness prediction unit 22, learning unit 24] The hardness prediction unit 22 has a prediction model 25 that is machine-learned to predict the objective variable from the explanatory variables, using the shape data and the heat treatment condition data as the explanatory variables and the hardness data as the objective variable. The prediction model 25 is a regression model. In the hardness prediction unit 22, the hardness data is predicted by calculation by inputting the shape data and the heat treatment condition data into the prediction model 25.

[0030] The hardness data includes information related to the hardness of the workpiece after heat treatment. For example, one or more evaluation points are preset on the surface of the product, and the hardness data includes the cross-sectional hardness distribution of the evaluation points. The cross-sectional hardness distribution of the evaluation points represents the distribution of hardness on the normal line of the evaluation point (that is, the line passing through the evaluation point and perpendicular to the surface of the evaluation point) in relation to the depth from the product surface or the workpiece surface. The cross-sectional hardness distribution illustrated in FIG. 5 is a chart in which the horizontal axis represents the distance from the surface (that is, the depth) and the vertical axis represents the hardness, and plots the relationship between the distance from the surface and the hardness. The hardness is represented by, for example, Vickers hardness [HV]. From the cross-sectional hardness distribution, the case depth, the effective hardened layer depth, and the hardness corresponding to the depth are clear, and the degree of carbon diffusion by carburizing quenching can be inferred.

[0031] The prediction model 25 is generated by the learning unit 24. The learning unit 24 uses a large number of teacher data sets composed of combinations of explanatory variables and objective variables, with the shape data and the heat treatment condition data as the explanatory variables and the hardness data as the objective variable, and constructs the prediction model 25 by machine learning based on a predetermined algorithm. The algorithm for machine learning is not particularly limited, and examples include linear regression, ridge regression, lasso regression, RBF (Radial Basis Function), InvD (Inverse Distance Weighted), etc.

[0032] The teacher dataset used by the learning unit 24 for machine learning includes the calculation results of numerical simulations (CAE: Computer Aided Engineering) and measured data. The measured data is the hardness data obtained by designing and prototyping a prototype from shape data, heat treatment condition data, and then measuring the prototype. By using such a teacher dataset, the prediction model 25 can be said to be a kind of CAE surrogate model.

[0033] 〔Evaluation unit 23〕 The evaluation unit 23 generates report data from the hardness data predicted by the hardness prediction unit 22 and the shape data, heat treatment condition data, finishing process data, and required quality data acquired by the data acquisition unit 21, and displays and outputs or prints out the report data. The evaluation unit 23 obtains the effective hardened layer depth of the workpiece after heat treatment, the effective hardened layer depth of the product, and the surface hardness of the product for preset evaluation points based on the acquired data. Note that according to Japanese Industrial Standard (JIS G 0557), the "effective hardened layer depth" is defined as the distance from the surface of the hardened layer as-quenched or tempered at a temperature not exceeding 200°C to the position of the limiting hardness of 550 HV.

[0034] The evaluation unit 23 generates report data showing the cross-sectional hardness distribution, the effective hardened layer depth of the workpiece after heat treatment, the effective hardened layer depth of the product, and the surface hardness of the product for each evaluation point. The report data is output from the output device 46. FIG. 6 is a diagram showing an example of the displayed report data. In the report shown in FIG. 6, for the four evaluation points of (i) tooth tip surface, (ii) flank pitch, (iii) tooth root fillet portion, and (iv) tooth root, the cross-sectional hardness distribution, the effective hardened layer depth of the workpiece after heat treatment, the effective hardened layer depth of the product, and the surface hardness of the product are shown. Here, the effective hardened layer depth and the surface hardness are shown as the range from the upper tolerance limit to the lower tolerance limit. The report also shows the required effective hardened layer depth and the required surface hardness for each evaluation point with reference to the required quality data. By superimposing the predicted hardness data and the required quality data, the two can be compared at a glance.

[0035] From the report, it is possible to evaluate the predicted quality of the product, such as the presence or absence of the possibility of overcarburization and whether the surface hardness of the product meets the required surface hardness.

[0036] The evaluation unit 23 may further seek the manufacturability of the product and include the manufacturability in the report data. In this case, the evaluation unit 23 calculates the manufacturability, which is an index of the difficulty level of the finishing process (for example, the process of shaving off the surplus from the workpiece after heat treatment), based on the finishing process data and the hardness data, and outputs the calculated manufacturability as part of the report data. More specifically, the evaluation unit 23 predicts the hardness distribution of the workpiece based on the hardness data of the workpiece after heat treatment, predicts the surplus of the finishing process based on the finishing process data and the hardness distribution of the workpiece, and calculates the manufacturability based on the predicted surplus. The method for predicting the hardness distribution of the workpiece based on the hardness data of the workpiece, the method for predicting the surplus of the finishing process based on the finishing process data and the hardness distribution of the workpiece, and the method for calculating the manufacturability from the surplus are given to the evaluation unit 23 in advance as arithmetic expressions or programs. By displaying or printing out the manufacturability, the manufacturability of the product can be evaluated at the development stage without actually manufacturing and evaluating the prototype in this way. The manufacturability can be calculated, for example, as the difference between the predicted effective hardened layer depth (or surface hardness) in the product shape after shaving off the surplus from the workpiece after heat treatment and the upper and lower limit values of the required range of the effective hardened layer depth (or surface hardness range). The greater the difference, the higher the manufacturability. A high manufacturability of the product means that the workpiece after heat treatment has a sufficient effective hardened layer depth considering the surplus during the finishing process. In the workpiece after heat treatment, if the surplus is insufficient, the finishing process cannot be carried out. On the other hand, if the surplus is excessive, the finishing process is possible, but the surface hardness after shaving off to the product shape (i.e., the completed shape) is insufficient. For example, when the manufacturability of the product is low, it includes the cases where the surplus during the finishing process of the workpiece after heat treatment is insufficient and excessive. When the manufacturability is low in this way, the finishing process is impossible, or the accuracy of the surplus during the finishing process is strictly required, and the difficulty level of the process increases. Conversely, when the manufacturability of the product is high, the surplus during the finishing process of the workpiece after heat treatment is appropriately ensured, and the difficulty level of the process decreases. The manufacturability of the product is also correlated with the quality stability, and the higher the manufacturability of the product, the higher the quality stability.

[0037] 〔Summary〕 The carburized product hardness prediction device 1 according to the first item of the present disclosure is a carburized product hardness prediction device 1 that predicts the hardness of a product whose surface is hardened by carburizing quenching, and is equipped with an arithmetic unit 41, the arithmetic unit 41 acquires shape data including information on the outer shape of the product and the intermediate body of the product, acquires heat treatment condition data including information on the conditions of heat treatment including carburizing quenching on the workpiece which is the intermediate body, using the shape data and the heat treatment condition data as explanatory variables and the hardness data including information on the hardness of the workpiece after heat treatment as the objective variable, and predicting the hardness data from the shape data and the heat treatment condition data using the prediction model 25 machine-learned to predict the objective variable from the explanatory variables, is configured to output the predicted hardness data.

[0038] According to the hardness prediction device 1 configured as above, it is possible to predict the hardness of a carburized product without creating a design drawing or a three-dimensional model with CAD, performing simulations with CAE, or repeatedly manufacturing and evaluating prototypes of carburized products. Therefore, it can contribute to shortening the time required for product development of carburized products.

[0039] The carburized product hardness prediction device 1 according to the second item of the present disclosure is the carburized product hardness prediction device 1 according to the first item, wherein the heat treatment condition data includes the carbon potential value of carburizing, the treatment temperature of carburizing, and the treatment time of carburizing.

[0040] The carburized product hardness prediction device 1 according to the third item of the present disclosure is the carburized product hardness prediction device 1 according to the first or second item, wherein the product is a gear, and the shape data includes the module m of the gear, the pressure angle α of the gear, and the tooth tip width w of the intermediate body.

[0041] The carburized product hardness prediction device 1 according to the fourth item of the present disclosure is the carburized product hardness prediction device 1 according to the third item, wherein the shape data further includes the fillet radius of curvature ρf of the root fillet portion of the intermediate body.

[0042] The carburized product hardness prediction device 1 according to the fifth item of the present disclosure is the carburized product hardness prediction device 1 according to any one of the first to fourth items, wherein the hardness data includes the cross-sectional hardness distribution of a predetermined evaluation point of the product or a point on the workpiece corresponding to the evaluation point.

[0043] The carburized product hardness prediction device 1 according to the sixth item of the present disclosure is the carburized product hardness prediction device 1 according to any one of the first to fifth items, wherein the hardness data includes the surface hardness of a predetermined evaluation point of the product.

[0044] The carburized product hardness prediction device 1 according to the seventh item of the present disclosure is the carburized product hardness prediction device 1 according to any one of the first to sixth items, wherein the arithmetic unit 41 acquires finishing process data including information on the replacement during the finishing process of the workpiece after heat treatment, predicts the hardness distribution of the workpiece after heat treatment based on the hardness data, predicts the replacement of the finishing process based on the finishing process data and the predicted hardness distribution, calculates the manufacturability which is an index of the difficulty of the finishing process based on the predicted replacement, and is configured to output the calculated manufacturability.

[0045] The carburized product hardness prediction method according to the eighth item of the present disclosure is a method for predicting the hardness of a product whose surface is hardened by carburizing quenching by a computer 40, acquiring shape data including information on the outer shape of the product and the intermediate body of the product, acquiring heat treatment condition data including information on the conditions of heat treatment including carburizing quenching of the workpiece which is the intermediate body, using a prediction model 25 obtained by machine learning to predict the hardness data from the shape data and the heat treatment condition data, with the shape data and the heat treatment condition data as explanatory variables and the hardness data including information on the hardness of the workpiece after heat treatment as the objective variable, and outputting the predicted hardness data.

[0046] According to the above hardness prediction method, the hardness of the carburized product can be predicted without repeatedly creating design drawings or 3D models in CAD, performing simulations in CAE, or manufacturing and evaluating prototypes of the carburized product. Therefore, it can contribute to shortening the time required for product development of carburized products.

[0047] The functions realized by the hardness prediction device 1 described in this specification may be implemented in circuitry or processing circuitry including a general-purpose processor, an application-specific processor, an integrated circuit, ASICs (Application Specific Integrated Circuits), a CPU (Central Processing Unit), a conventional circuit, and / or a combination thereof, programmed to realize the described functions. The processor includes transistors and other circuits and is regarded as circuitry or processing circuitry. The processor may be a programmed processor that executes a program stored in a memory. In this specification, circuitry, unit, and means are hardware programmed to realize the described functions or hardware that executes. The hardware may be any hardware disclosed in this specification or any hardware known to be programmed or execute to realize the described functions. When the hardware is a processor regarded as the circuitry type, the circuitry, means, or unit is a combination of hardware and software used to configure the hardware and / or the processor.

[0048] The foregoing discussion of the present disclosure has been presented for purposes of illustration and description and is not intended to limit the present disclosure to the forms disclosed herein. For example, in the foregoing detailed description, various features of the present disclosure are grouped together in one embodiment for the purpose of streamlining the present disclosure, but some of the features may be combined. Also, the various features included in the present disclosure may be combined in alternative embodiments, configurations, or aspects other than those discussed above.

Description of Reference Numerals

[0049] 1: Characteristic prediction device 25: Prediction model 40: Computer 41: Arithmetic unit

Claims

1. A hardness prediction device for carburized products that predicts the hardness of products with a hardened surface by carburizing quenching, comprising an arithmetic unit, wherein the arithmetic unit, acquires shape data including information on the outer shape of the product and the intermediate body of the product, acquires heat treatment condition data including information on the conditions of heat treatment including the carburizing quenching of the intermediate body workpiece, uses a prediction model that performs machine learning to predict the target variable from the explanatory variables, with the shape data and the heat treatment condition data as the explanatory variables and the hardness data including information on the hardness of the workpiece after the heat treatment as the target variable, to predict the hardness data from the shape data and the heat treatment condition data, and is configured to output the predicted hardness data, a hardness prediction device for carburized products.

2. The heat treatment condition data includes the carbon potential value of carburizing, the treatment temperature of carburizing, and the treatment time of carburizing, The hardness prediction device for carburized products according to Claim 1.

3. The product is a gear, and the shape data includes the module of the gear, the pressure angle of the gear, and the tooth tip width of the intermediate body, The hardness prediction device for carburized products according to Claim 1.

4. The shape data further includes the curvature radius of the tooth root fillet portion of the intermediate body, The hardness prediction device for carburized products according to Claim 3.

5. The hardness data includes the cross-sectional hardness distribution of a predetermined evaluation point of the product or a point on the workpiece corresponding to the evaluation point, The hardness prediction device for carburized products according to Claim 1.

6. The hardness data includes the surface hardness of a predetermined evaluation point of the product, The hardness prediction device for carburized products according to Claim 1.

7. The arithmetic unit, acquires finishing process data including information on the replacement during the finishing process of the workpiece after the heat treatment, predicts the hardness distribution of the workpiece after the heat treatment based on the hardness data, predicts the replacement of the finishing process based on the finishing process data and the predicted hardness distribution, calculates the manufacturability, which is an index of the difficulty of the finishing process, based on the predicted replacement, and is configured to output the calculated manufacturability, The hardness prediction device for carburized products according to Claim 1.

8. A method for predicting the hardness of a product with a hardened surface by carburizing quenching using a computer, acquiring shape data including information on the outer shape of the product and the intermediate body of the product, Obtaining heat treatment condition data including information on the conditions of heat treatment including the carburizing quenching of the workpiece as the intermediate body, Using the shape data and the heat treatment condition data as explanatory variables and hardness data including information on the hardness of the workpiece after the heat treatment as the target variable, predicting the hardness data from the shape data and the heat treatment condition data using a prediction model machine-learned to predict the target variable from the explanatory variables, and Outputting the predicted hardness data, including A method for predicting the hardness of a carburized product.

Images