Calculation method for quenching hardness J15 of 20CrMnTiH steel

Abstract

The invention relates to the technical field of heat treatment, in particular to a method for calculating the quenching hardness J15 of 20CrMnTiH steel, which comprises the following steps of: collecting data, controlling heat treatment process conditions, controlling the chemical components of the 20CrMnTiH steel to comprise the following components in percentage by mass: 0.17-0.21% of C, 0.81-0.87% of Mn, 0.21-0.27% of Si, 1.01-1.07% of Cr and 0.05-0.07% of Ti, determining model parameters through an iterative optimization algorithm, establishing a nonlinear regression model of the quenching hardness of the 20CrMnTiH steel J15, and calculating the quenching hardness J15 of the 20CrMnTiH steel J15. An artificial intelligence algorithm is adopted to train the model and optimize model parameters through fitting interaction and quadratic terms, and a quenching hardness multivariate nonlinear regression model is obtained; according to the method, the problems of low prediction precision, poor adaptability, lack of flexibility and the like in the prior art are solved, the high-precision nonlinear regression model is established, and accurate prediction and control on the J15 quenching hardness of the 20CrMnTiH steel are realized.

Description

Technical Field

[0001] The present invention relates to the field of heat treatment technology, in particular to a method for calculating the quenching hardness J15 of 20CrMnTiH steel. Background Art

[0002] 20CrMnTiH steel is a commonly used structural steel with guaranteed hardenability and excellent overall performance, making it widely used in machinery manufacturing. Quenching hardness, a key indicator of steel heat treatment effectiveness, has a decisive impact on the steel's ultimate performance. However, the prediction and control of quenching hardness of 20CrMnTiH steel J15 still presents many challenges.

[0003] In the existing technology, the prediction of quenching hardness of 20CrMnTiH steel mainly adopts empirical formulas or linear regression models. However, since quenching hardness is affected by multiple factors (such as steel composition and heat treatment process parameters), and these factors often have complex nonlinear relationships, the prediction accuracy of empirical formulas and linear regression models is limited, which makes it difficult to meet the requirements of high-precision heat treatment processes.

[0004] Specifically, the shortcomings of the existing technology are mainly reflected in the following aspects:

[0005] Low prediction accuracy: Since the nonlinear relationship between various factors is not fully considered, the prediction results deviate greatly from the actual values.

[0006] Poor adaptability: Steels from different batches and compositions need to be modeled separately, resulting in poor model versatility.

[0007] Lack of flexibility: The model cannot be quickly adjusted and optimized according to actual needs.

[0008] Therefore, it is urgent to develop a more accurate, efficient and flexible 20CrMnTiH steel J15 quenching hardness prediction model to meet the needs of high-precision composition control and heat treatment process. Summary of the Invention

[0009] The present invention provides a calculation method for the quenching hardness J15 of 20CrMnTiH steel, which solves the problems of low prediction accuracy, poor adaptability and lack of flexibility in the prior art. A high-precision nonlinear regression model is established to accurately predict and control the quenching hardness J15 of 20CrMnTiH steel.

[0010] In order to achieve the above object, the present invention adopts the following technical solutions:

[0011] A method for calculating the quenching hardness J15 of 20CrMnTiH steel comprises the following steps:

[0012] S1. Data collection: Collect quenching hardness data of 20CrMnTiH steel with different chemical compositions under the same heat treatment process conditions;

[0013] S2. Control heat treatment process conditions: normalizing temperature 910℃±10℃, holding temperature 30-35min, quenching temperature 880℃±5℃, cooling water temperature 20±5℃, the interval between sample removal and quenching should not exceed 5s;

[0014] S3. Control the chemical composition of 20CrMnTiH steel to be as follows: C: 0.17-0.21%, Mn: 0.81-0.87%, Si: 0.21-0.27%, Cr: 1.01-1.07%, Ti: 0.05-0.07% by mass;

[0015] S4. Data preprocessing and model establishment: A nonlinear regression analysis of chemical composition and J15 quenching hardness was established for the preprocessed data. The model parameters were determined through an iterative optimization algorithm. A nonlinear regression model of 20CrMnTiH steel J15 quenching hardness was established. The influence of each element on quenching hardness was determined through this model. The model was trained and the model parameters were optimized by fitting the interaction and quadratic terms using an artificial intelligence algorithm. The two-sided confidence intervals of the regression model were set, and the confidence level of all intervals was 90%. The fitting degree was greater than 80% after fitting, resulting in a multivariate nonlinear regression model for quenching hardness:

[0016] Quenching hardness J15 = -55.6 + 4.113C - 0.954Si + 0.993Mn + 158.4Cr - 73.8Cr 2

[0017] Among them, C, Si, Mn, and Cr are the mass percentages of their respective components. According to the verification results, the model parameters and algorithms are optimized and adjusted until the set quenching hardness control accuracy is achieved.

[0018] Furthermore, the data preprocessing includes data cleaning, outlier detection and processing.

[0019] Furthermore, the artificial intelligence algorithms include Minitab regression analysis, DeepSeek regression analysis and random forest.

[0020] Furthermore, the quenching hardness control accuracy is that the difference between the quenching hardness calculation prediction result and the actual value is within 2.4HRC.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] 1) It can accurately predict the quenching hardness under specific heat treatment process conditions, providing technical support for optimizing the composition design of 20CrMnTiH;

[0023] 2) Through model optimization and application, the accuracy of quenching hardness prediction for 20CrMnTiH steel can be improved, the scrap rate and rework rate due to unqualified hardness can be reduced, and production efficiency and economic benefits can be improved;

[0024] 3) The present invention has broad application prospects and can be applied to the prediction and optimization of quenching hardness of other metal materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a relationship diagram between quenching hardness and C content according to the present invention.

[0026] Figure 2 This is a relationship diagram between quenching hardness and Si composition according to the present invention.

[0027] Figure 3 This is a relationship diagram between quenching hardness and Mn composition according to the present invention.

[0028] Figure 4 This is a relationship diagram between quenching hardness and Ti composition according to the present invention.

[0029] Figure 5 This is a relationship diagram between the quenching hardness and the Cr composition of the present invention. DETAILED DESCRIPTION

[0030] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0031] The present invention provides a method for calculating the quenching hardness J15 of 20CrMnTiH steel, comprising the following steps:

[0032] S1. Data collection: 1000 sets of quenching hardness data of 20CrMnTiH steel with different main chemical compositions of C, Mn, Si, Cr, and Ti under the same heat treatment process conditions were collected.

[0033] S2. Control the heat treatment process conditions: normalizing temperature 910℃±10℃ for 30min, quenching temperature 880℃±5℃, cooling water temperature 20℃, and the interval between sample removal and quenching should not exceed 5s.

[0034] S3. Control the chemical composition mass percentage of 20CrMnTiH steel to be: C: 0.17-0.21%, Mn: 0.81-0.87%, Si: 0.21-0.27%, Cr: 1.01-1.07%, Ti: 0.05-0.07%.

[0035] S4. Data preprocessing and model building:

[0036] Data preprocessing: preprocessing the collected data, including data cleaning, outlier detection and processing, to ensure that the data quality meets the modeling requirements;

[0037] Model establishment: Using artificial intelligence to perform nonlinear regression analysis on preprocessed data, analyze the relationship between each element and quenching hardness J15, determine the model parameters through iterative optimization algorithm, and establish a high-precision nonlinear regression model of 20CrMnTiH steel J15 quenching hardness;

[0038] Study the effect of each element composition on hardenability, average the quenching hardness value J15 of all groups of a single component at each mass fraction, and analyze the effect of each element on the quenching hardness J15, such as Figure 1-5 As shown;

[0039] It can be seen from the scatter plot that there is an obvious linear correlation between C, Mn, Cr and quenching hardness J15. However, in order to avoid missing significant factors, the chemical compositions C, Mn, Si, Cr, and Ti are selected as inputs, and the quenching hardness J15 is selected as output. Considering the fitting of two-way interactions and quadratic terms, in order to capture the synergistic effect between elements, the product terms of all pairwise combinations are generated: C×Mn, C×Si, C×Cr, C×Ti, Mn×Si, Mn×Cr, Mn×Ti, Si×Cr, Si×Ti, Cr×Ti, a total of 10 interaction features, reflecting the influence of synchronous changes in element content on hardness; Minitab regression analysis is used to provide stepwise regression and ANOVA analysis under the traditional statistical framework to ensure the interpretability of the model and significance test; DeepSeek regression analysis, regression forest and other artificial intelligence algorithms automatically adjust parameters through Bayesian optimization to maximize the validation set R 2 The model is trained and its parameters are optimized. A two-sided confidence interval is set in the regression model, with a confidence level of 90% for all intervals. This ensures that the predicted value has a 90% probability of falling within the interval, thus avoiding the potential error direction being missed by a one-sided interval.

[0040] After fitting, the model's degree of fit = 81.26%, and the multivariate nonlinear regression equation is obtained:

[0041] Quenching hardness J15 = -55.6 + 4.113C - 0.954Si + 0.993Mn + 158.4Cr - 73.8Cr 2

[0042] Among them, C, Si, Mn, and Cr are the mass percentages of their respective components. Based on the verification results, the model parameters and algorithms are optimized and adjusted until the set quenching hardness control accuracy is achieved;

[0043] 4.113C: Indicates the effect of carbon content on quenching hardness. Carbon is an important element for increasing the hardness of steel. Its coefficient is positive, indicating that as the carbon content increases, the predicted quenching hardness will also increase.

[0044] -0.954Si: represents the effect of silicon content on quenching hardness. Silicon in steel usually plays a role in improving toughness and reducing brittleness. Its coefficient is negative, indicating that as the silicon content increases, the predicted quenching hardness will decrease;

[0045] 0.993Mn: Indicates the effect of manganese content on quenching hardness. Manganese can improve the strength and hardness of steel. Its coefficient is positive, indicating that as the manganese content increases, the predicted quenching hardness will also increase;

[0046] 158.4Cr: Indicates the effect of chromium content on quenching hardness. Chromium is an important element that improves the hardenability and corrosion resistance of steel. Its coefficient is positive, indicating that as the chromium content increases, the predicted quenching hardness increases. However, it should be noted that there is also a quadratic term for chromium in the model.

[0047] -73.8Cr 2: represents the effect of the square of the chromium content on the quenching hardness. This is a quadratic term that takes into account the nonlinear effect of the change in chromium content on the quenching hardness. Since the coefficient is negative, it means that as the chromium content increases, its effect on the improvement of the quenching hardness will gradually weaken;

[0048] In summary, this regression model comprehensively considers the effects of the four chemical components of carbon, silicon, manganese and chromium on the quenching hardness, and predicts the quenching hardness value through linear combination and nonlinear terms (quadratic term of chromium).

[0049] S4. Model Validation: 50 independent validation data sets were used to validate the model and evaluate its prediction accuracy. The data are shown in Table 1. From the data, it can be seen that the difference between the predicted and actual quenching hardness J15 fluctuates within 2.4HRC. The data fit is very high and can be applied.

[0050] Table 1

[0051]

[0052]

[0053]

[0054] Based on the verification results, the model parameters and algorithms are optimized and adjusted until satisfactory prediction results are achieved; the optimized model is re-verified to ensure its stability and reliability on different data sets.

[0055] S5. Model Application

[0056] The target hardness of 20CrMnTiH steel J15 in actual production is set at 30.5HRC. The composition design of 10 heats of steel is carried out using the regression model and the error is verified.

[0057] According to the regression model, the chemical composition mass percentage of 20CrMnTiH steel is set as: C: 0.18%, Mn: 0.81%, Si: 0.22%, Cr: 1.02%, and Ti is not limited. The results of the quenching hardness J15 of 10 batches of 20CrMnTiH produced are shown in Table 2. The fluctuation range between the predicted value and the actual value is 0.1 to -2.2HRC.

[0058] Table 2

[0059] Serial number Actual value of quenching hardness J15 Predicted value of quenching hardness J15 Difference 1 30.4 30.5 0.1 2 30.6 30.5 0.1 3 31.6 30.5 1.1 4 30.1 30.5 0.4 5 28.3 30.5 2.2 6 30.4 30.5 0.1 7 32.1 30.5 1.6 8 31.0 30.5 0.5 9 32.6 30.5 2.1 10 30.3 30.5 0.2

[0060] Update data and retrain models regularly to adapt to changes and demands in production.

[0061] Through the above implementation, it can be ensured that the nonlinear regression model of quenching hardness of 20CrMnTiH steel J15 of the present invention is effectively applied in actual production, and significant economic and social benefits can be achieved by rationally controlling the main product components.

[0062] The above embodiments are implemented under the premise of the technical solution of the present invention, and detailed implementation methods and specific operation processes are given, but the protection scope of the present invention is not limited to the above embodiments. The methods used in the above embodiments are conventional methods unless otherwise specified.

Claims

1. A method for calculating the quenching hardness J15 of 20CrMnTiH steel, characterized in that: The steps include: S1. Data collection: Collect quenching hardness data of 20CrMnTiH steel with different chemical compositions under the same heat treatment process conditions; S2. Control heat treatment process conditions: normalizing temperature 910℃±10℃, holding temperature 30-35min, quenching temperature 880℃±5℃, cooling water temperature 20±5℃, the interval between sample removal and quenching should not exceed 5s; S3. Control the chemical composition of 20CrMnTiH steel to be as follows: C: 0.17-0.21%, Mn: 0.81-0.87%, Si: 0.21-0.27%, Cr: 1.01-1.07%, Ti: 0.05-0.07% by mass; S4. Data preprocessing and model establishment: A nonlinear regression analysis of chemical composition and J15 quenching hardness was established for the preprocessed data. The model parameters were determined through an iterative optimization algorithm. A nonlinear regression model of 20CrMnTiH steel J15 quenching hardness was established. The influence of each element on quenching hardness was determined through this model. The model was trained and the model parameters were optimized by fitting the interaction and quadratic terms using an artificial intelligence algorithm. The two-sided confidence intervals of the regression model were set, and the confidence level of all intervals was 90%. The fitting degree was greater than 80% after fitting, resulting in a multivariate nonlinear regression model for quenching hardness: Quenching hardness J15 = -55.6 + 4.113C - 0.954Si + 0.993Mn + 158.4Cr - 73.8Cr 2 Among them, C, Si, Mn, and Cr are the mass percentages of their respective components. According to the verification results, the model parameters and algorithms are optimized and adjusted until the set quenching hardness control accuracy is achieved.

2. The method for calculating the quenching hardness J15 of 20CrMnTiH steel according to claim 1, characterized in that: The data preprocessing includes data cleaning, outlier detection and processing.

3. The method for calculating the quenching hardness J15 of 20CrMnTiH steel according to claim 1, characterized in that: The artificial intelligence algorithms include Minitab regression analysis, DeepSeek regression analysis and random forest.

4. The method for calculating the quenching hardness J15 of 20CrMnTiH steel according to claim 1, characterized in that: The quenching hardness control accuracy is that the difference between the quenching hardness calculation prediction result and the actual value is within 2.4HRC.

Images