CN116204998A - SLM forming performance prediction and process parameter optimization method and system - Google Patents

Abstract

The invention relates to the field of SLM forming, in particular to a method and a system for predicting the forming performance of an SLM and optimizing process parameters. S1, designing a multi-step orthogonal experiment to obtain technological parameters and performances of an SLM forming part; s2, randomly dividing a data set; s3, respectively constructing a Gaussian process regression model and a multiple stepwise regression model, constructing the Gaussian process regression model through a constant mean function and a rational secondary covariance function, and adopting a block and binary strategy to adjust super parameters; s4, adopting MAE, RMSE, R ² Three-index evaluation of two models in S2; s5, constructing a Gaussian process regression-multiple stepwise regression prediction model to accurately predict the performance of the formed part by the SLM process parameters; s6, taking the prediction model constructed in the S5 as adaptationThe degree function is optimized through teaching and learning algorithm to obtain recommended technological parameters; s7, screening and verifying recommended technological parameters step by step. The invention not only can accurately predict the performance according to the SLM process parameters, but also can efficiently give recommended process parameters.

Description

A method and system for SLM forming performance prediction and process parameter optimization

technical field

The invention belongs to the field of metal selective laser melting (Selective Laser melting, SLM) forming, and more specifically relates to a method and system for SLM forming performance prediction and process parameter optimization.

Background technique

SLM technology is under the action of high-energy laser, the metal powder is completely melted, and after heat dissipation and solidification, it is metallurgically welded with the base metal, and then accumulated layer by layer to form a three-dimensional entity, which can directly form almost fully dense metal parts with good mechanical properties. Compared with traditional metal material forming, SLM forming can directly form parts with complex geometric shapes, and after SLM parts are formed, they can be used with only a small amount of processing or no processing, and the surface quality is excellent. Therefore, SLM is widely used in aerospace, military equipment and other fields.

The main process parameters of SLM forming include laser power, powder layer thickness, scanning speed, scanning distance, etc. These process parameters have a significant impact on the performance of formed parts. The main properties of SLM forming are density, tensile strength, yield strength, elongation after fracture, surface roughness, etc. In the manufacturing process of SLM formed parts, not only there is an interaction between various process parameters, but also the SLM forming process itself has very complex changes, and the performance of SLM formed parts is difficult to predict under the joint action of various factors.

Regarding the acquisition of performance data of SLM formed parts, the existing method is to obtain the formed parts through a series of performance testing tests after the formed parts are manufactured and cooled. This method takes a long time to obtain the performance of SLM specific process parameters. When it is necessary to obtain the performance corresponding to multiple sets of process parameters, the time and materials consumed will also increase exponentially. The process parameters for the production of high-performance SLM formed parts are set through the test method, a large number of process parameter groups are set, and the SLM equipment is produced one by one, and then the performance of the obtained SLM formed parts is tested one by one, and the process parameters corresponding to the SLM formed parts with better performance are recorded as follow-up Process parameters for industrial production. This method of obtaining high-performance SLM forming process parameters is extremely inefficient, consumes a large amount of raw materials, and takes a long time for printing and testing. For example, using the comprehensive test method, for the four process parameters of SLM, if 4 experimental points are set for each set of process parameters, 4 ⁴ , that is, 256 sets of experiments are required; if 10 experimental points are set for each set of process parameters, 10 test points need to be set ⁴ , that is, 10,000 sets of experiments. This method is inefficient, consumes a lot of raw materials, and takes a long time for printing and performance testing, and these quantities cannot ensure that the obtained process parameters are optimal. All in all, the performance of formed parts cannot be quickly obtained when the SLM process parameters are known, and the existing methods for obtaining the process parameters required for industrial production of high-performance SLM formed parts are inefficient.

Therefore, it is very necessary to use other methods to accurately and quickly obtain the performance of the formed part under the condition of known SLM process parameters, and to efficiently obtain the process parameters for manufacturing high-performance SLM formed parts for industrial production.

Contents of the invention

In view of the defects of the prior art, the purpose of the present invention is to provide a method and system for SLM forming performance prediction and process parameter optimization, aiming to solve the problem of slow and time-consuming acquisition of SLM forming part performance under the condition of known SLM process parameters. The problem is related to the problem of low efficiency in obtaining process parameters of high-performance SLM formed parts.

In order to achieve the above object, in the first aspect, the present invention provides a method for SLM forming performance prediction and process parameter optimization, comprising the following steps:

Design multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacture the designed process parameters to determine the performance of the corresponding SLM formed parts, and summarize the designed SLM process parameters and the corresponding performance of the formed parts into a data set;

Based on the data set, the Gaussian process regression model is trained to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; and based on the data set, the multiple stepwise regression model is trained to obtain the SLM process parameters and the SLM formed part performance. The second mapping relationship between the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are used to predict the performance of the SLM formed part;

Combining the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model uses two performance prediction results of the Gaussian process regression model and the multiple stepwise regression model. mode fusion, the weight of the weighted mode is determined through traversal;

Optimize the SLM performance prediction model through the teaching and learning algorithm, and obtain multiple sets of recommended SLM process parameters whose performance of the SLM formed part meets the requirements;

According to the influence of the SLM process parameters on the service life of the equipment, the multiple sets of recommended SLM process parameters are screened step by step, and multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters are obtained.

In an optional example, the method further includes the following steps:

For the multiple sets of SLM process parameters after step-by-step screening, the actual manufacturing is carried out to verify the performance of the obtained SLM formed parts, and the parameters whose performance meets the expected standards are included in the process parameter group with satisfactory verification results, as the process of high-performance SLM formed parts Parameter group; the parameters whose performance does not meet the expected standard are included in the process parameter group whose verification result is not satisfactory;

Add the unsatisfactory process parameter group to the data set, update the data set, repeat the training of Gaussian process regression model, the training of multiple stepwise regression model, the recommendation and screening process of SLM process parameters, and obtain multiple sets of SLM after new screening Process parameters.

In an optional example, multiple sets of SLM process parameters are designed through multi-step orthogonal experiments, specifically:

Design multi-step orthogonal experiments, design SLM process parameter groups through orthogonal experiments on a global scale, and actually manufacture the designed process parameters to test the performance of the preliminary designed SLM formed parts;

Carry out a preliminary test on the global performance of SLM, find out the suspicious areas with better performance of SLM formed parts in the preliminary detection, and design orthogonal experiments for suspicious areas, and obtain the correspondence of suspicious areas with better performance of SLM formed parts after many times of designing orthogonal experiments Multiple sets of SLM process parameters.

In an optional example, the Gaussian process regression model is trained based on the data set, and the multiple stepwise regression model is trained based on the data set, specifically:

Build a Gaussian process regression model; train the Gaussian process regression model based on the designed multiple sets of SLM process parameters;

Construct a multiple stepwise regression model; train the multiple stepwise regression model based on the designed multiple sets of SLM process parameters; among them, taking into account the interaction between process parameters, increase high-order items and cross items to construct complete alternatives; obtain a complete After selecting the alternatives, through the significance test, the alternatives are gradually introduced and eliminated.

In an optional example, during the training process of the Gaussian process regression model and the multiple stepwise regression model, the two models are weighted and evaluated by the three indicators of mean absolute error MAE, root mean square error RMSE and coefficient of determination ^R2 , To evaluate the deviation between the model prediction results and the actual results, optimize the model parameters; the respective weights of the three indicators are: MAE:RMSE:R ² =40:40:20.

In an optional example, the SLM performance prediction model is optimized through the teaching and learning algorithm, and multiple sets of recommended SLM process parameters for which the performance of the SLM formed part meets the requirements are obtained, specifically:

Set an appropriate number of initial groups, and use the SLM performance prediction model as the fitness function of the optimization algorithm;

By calculating the individual fitness, the individual with the largest fitness value is selected as the teacher. The initial group is updated through the "teaching stage" and "learning stage". The current optimal fitness individual, that is, the current optimal performance value and the corresponding process parameters, and the process parameters corresponding to the optimal performance value are used as the recommended SLM process parameters.

In an optional example, the SLM process parameters include: laser power, powder layer thickness, scanning speed and scanning distance; laser power and scanning speed have a relatively large impact on the life of the equipment, and laser power has a relatively large impact on the process stability of the equipment. The impact is relatively large. Set multiple thresholds corresponding to the two process parameters according to the laser power and scanning speed of the actual equipment. When the laser power or scanning speed of the actual equipment is in different threshold ranges, the laser power or scanning speed is in different ratings. Different ratings have different losses to equipment;

According to the impact of the SLM process parameters on the service life of the equipment, the multiple groups of recommended SLM process parameters are screened step by step, and multiple groups of SLM process parameters with low damage to the service life of the equipment are obtained, specifically:

The multiple groups of recommended SLM process parameters are in accordance with: ①Laser power level 1, scanning speed level 1, ②Laser power level 1, scanning speed level 2, ③Laser power level 2, scanning speed level 1, ④Laser power level 2, scanning speed 2 The levels are divided into four categories; among them, the laser power level has the lowest damage to the service life of the equipment, and as the laser power level increases, the damage to the equipment life increases; the scanning speed level has the lowest damage to the equipment life, The speed level increases, and the damage to the service life of the equipment increases;

After step-by-step screening, the laser power and scanning speed among multiple sets of recommended SLM process parameters are eliminated, and then displayed in order from ① to ④ from top to bottom, and multiple sets of process parameters after screening are obtained.

In the second aspect, the present invention provides a SLM forming performance prediction and process parameter optimization system, including:

The parameter data set design unit is used to design multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacture the designed process parameters, determine the performance of the corresponding SLM formed parts, and combine the designed SLM process parameters with the corresponding formed parts performance aggregated as a dataset;

The regression model training unit is used to train the Gaussian process regression model based on the data set to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; and based on the data set, the multiple stepwise regression model is performed Training to obtain the second mapping relationship between the SLM process parameters and the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are used to predict the performance of the SLM formed part;

The regression model combination unit is used to combine the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model combines the Gaussian process regression model and the multiple stepwise regression model. The two performance prediction results are fused in a weighted manner, and the weight of the weighted manner is determined by traversal;

The prediction model optimization unit is used to optimize the SLM performance prediction model through the teaching and learning algorithm, and obtain multiple sets of recommended SLM process parameters for which the performance of the SLM formed part meets the requirements;

The process parameter screening unit is used to screen the multiple sets of recommended SLM process parameters step by step according to the influence of the SLM process parameters on the service life of the equipment, so as to obtain multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters.

In an optional example, the system also includes:

The process parameter grouping unit is used to carry out actual manufacturing of multiple sets of SLM process parameters after step-by-step screening, to verify the performance of the obtained SLM formed parts, and to list the parameters whose performance meets the expected standard into the process parameter group with satisfactory verification results, as The process parameter group of high-performance SLM formed parts; the parameters whose performance does not meet the expected standard are included in the process parameter group with unsatisfactory verification results;

The new parameter recommendation screening unit is used to add process parameter groups with unsatisfactory verification results to the data set, update the data set, repeat Gaussian process regression model training, multiple stepwise regression model training, SLM process parameter recommendation and screening process, Multiple sets of SLM process parameters after the new screening are obtained.

In an optional example, the SLM process parameters include: laser power, powder layer thickness, scanning speed and scanning distance; laser power and scanning speed have a relatively large impact on the life of the equipment, according to the actual equipment laser power and scanning The speed sets multiple thresholds corresponding to the two process parameters. When the laser power or scanning speed of the actual equipment is in different threshold ranges, the laser power or scanning speed is in different ratings, and different ratings have different losses to the equipment;

The process parameter screening unit uses multiple groups of recommended SLM process parameters according to: ① laser power level 1, scanning speed level 1, ② laser power level 1, scanning speed level 2, ③ laser power level 2, scanning speed level 1, ④ The second level of laser power and scanning speed are divided into four categories; among them, the damage to the service life of the equipment at the first level of laser power is the lowest, and as the laser power level increases, the damage to the service life of the equipment increases; The damage is the lowest, and as the scanning speed level increases, the damage to the service life of the equipment increases; after step-by-step screening, the laser power and scanning speed in multiple groups of recommended SLM process parameters are eliminated, and the parameters from ① to The order of ④ is displayed from top to bottom, and multiple sets of process parameters after screening are obtained.

In a third aspect, the present invention provides an electronic device, including: a memory and a processor; the memory is used to store a computer program; the processor is used to implement the first aspect above when executing the computer program provided method.

In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method provided in the above-mentioned first aspect is realized.

Generally speaking, compared with the prior art, the above technical solution conceived by the present invention has the following beneficial effects:

The invention provides a method and system for SLM forming performance prediction and process parameter optimization, which builds a Gaussian process regression-multiple stepwise regression model, which can accurately predict the performance of formed parts through SLM process parameters; optimize through teaching and learning algorithms, efficiently The recommended process parameters are given; without consuming a lot of material and time costs, the recommended process parameters of high-performance SLM formed parts can be provided, which is of great significance to the SLM industrial production.

Description of drawings

Fig. 1 is a flow chart of the SLM forming performance prediction and process parameter optimization method provided by the embodiment of the present invention.

Fig. 2 is a flow chart of the SLM performance prediction and process parameter optimization method provided by the embodiment of the present invention.

Fig. 3 is a flowchart of the combination of the Gaussian process regression-multiple stepwise regression model and the teaching and learning algorithm of the SLM performance prediction and process parameter optimization method provided by the embodiment of the present invention.

Fig. 4 is an architecture diagram of the SLM forming performance prediction and process parameter optimization system provided by the embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

In the description of the present invention, reference to the terms "one embodiment," "some embodiments," "exemplary embodiments," "examples," "specific examples," or "some examples" is intended to mean that the embodiments are A specific feature, structure, material, or characteristic described by or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

The invention obtains SLM process parameters and performance data through design experiments, builds an SLM performance prediction model, accurately and quickly obtains the performance of formed parts through SLM process parameters, and optimizes through optimization algorithms to efficiently obtain recommended process parameters for high-performance SLM formed parts and use them in industrial production.

Fig. 1 is a flow chart of the SLM forming performance prediction and process parameter optimization method provided by the embodiment of the present invention, as shown in Fig. 1, including the following steps:

S101, designing multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacturing the designed process parameters, determining the performance of the corresponding SLM formed parts, and summarizing the designed SLM process parameters and the corresponding performance of the formed parts into a data set ;

S102, train the Gaussian process regression model based on the data set to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; and train the multivariate stepwise regression model based on the data set to obtain the SLM process The second mapping relationship between the parameters and the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are used to predict the performance of the SLM formed part;

S103, combine the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model combines the two performance prediction results of the Gaussian process regression model and the multiple stepwise regression model A weighted method is used for fusion, and the weight of the weighted method is determined through an traversal method;

S104, optimize the SLM performance prediction model through teaching and learning algorithms, and obtain multiple sets of recommended SLM process parameters whose performance of the SLM formed part meets the requirements;

S105, performing step-by-step screening on the multiple sets of recommended SLM process parameters according to the influence of the SLM process parameters on the service life of the equipment, to obtain multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters.

In an optional example, the method further includes the following steps:

For the multiple sets of SLM process parameters after step-by-step screening, the actual manufacturing is carried out to verify the performance of the obtained SLM formed parts, and the parameters whose performance meets the expected standards are included in the process parameter group with satisfactory verification results, as the process of high-performance SLM formed parts Parameter group; the parameters whose performance does not meet the expected standard are included in the process parameter group whose verification result is not satisfactory;

Add the unsatisfactory process parameter group to the data set, update the data set, repeat the training of Gaussian process regression model, the training of multiple stepwise regression model, the recommendation and screening process of SLM process parameters, and obtain multiple sets of SLM after new screening Process parameters.

In a specific embodiment, the object of the present invention is achieved through the following technical solutions:

Obtain the performance of the process parameters corresponding to the formed parts by designing multi-step orthogonal experiments, design the main process parameters of SLM: laser power, powder layer thickness, scanning speed, scanning distance, and detect and obtain the properties of the formed parts: density, tensile strength, yield Strength, elongation after fracture, surface roughness, etc. Build a Gaussian process regression model and a multiple stepwise regression model respectively. Both models can predict performance through SLM process parameters, and then integrate the two models to build a Gaussian process regression-multiple stepwise regression model, and predict the performance of SLM formed parts by SLM process parameters value, to achieve accurate and rapid prediction of the performance of SLM formed parts from SLM process parameters. The obtained Gaussian process regression-multiple stepwise regression model is integrated with the teaching and learning optimization algorithm, and the obtained SLM performance prediction model is used as the fitness function of the teaching and learning algorithm, and the optimal adaptation is obtained through the optimization of the teaching and learning algorithm. The degree value individual, that is, the optimal performance of SLM forming and the corresponding process parameters, and the optimal performance and the corresponding process parameters are used as the recommended process parameters. Multiple sets of recommended process parameters can be obtained through multiple operations, and multiple sets of recommended process parameters can be screened step by step according to process parameters such as laser power and scanning speed, and the recommended process parameters after screening can be obtained. Finally, the actual verification of the obtained process parameters after screening.

Specifically, the SLM performance prediction and process parameter optimization method of the present invention, as shown in Figure 2, includes the following steps:

S1. Design multi-step orthogonal experiments to obtain SLM forming process parameters and performance data, and divide the data sets. The traditional orthogonal experiment design can be completed in one step. Although the orthogonal experiment designed in this way can cover the whole situation, it has no focus. In order to obtain the process parameters of high-performance SLM forming parts, it is necessary to pay more attention to the high-performance forming area of SLM than other areas, and the traditional orthogonal experiment cannot meet this requirement. For this reason, the present invention adopts multi-step orthogonal experiment to design. To design a multi-step orthogonal experiment, first design the process parameter group through the orthogonal experiment in the global scope, and carry out the actual manufacturing of the designed process parameters to test the performance of the preliminarily designed SLM formed parts. Through the preliminary detection of the global performance of SLM, find out the suspicious areas with better performance in the preliminary detection, and then design orthogonal experiments on these suspicious areas with better performance to obtain more information about the suspicious areas with better performance, for the follow-up The performance prediction model of SLM forming parts provides more support for accurate prediction of high-performance forming parts. Finally, a data set covering the whole world and with reasonable density is obtained.

S2. Randomly divide the data set. Randomly divide the data into a training set and a prediction set for subsequent Gaussian process regression model and multiple stepwise regression model training and testing.

S3. Build the Gaussian process regression model and the multiple stepwise regression model respectively, as shown in Figure 3:

①Building a Gaussian process regression model: Gaussian process regression is a kind of machine learning. Compared with the common BP neural network, accurate prediction can be realized without a large amount of data. Dozens of experimental data are designed in the SLM forming performance prediction. Can. Select the appropriate mean function and covariance function, and assign initial values to the hyperparameters. In the range of normal SLM forming, the performance value of SLM formed parts is generally stable, and the Gaussian process regression model for predicting SLM performance is suitable for using a constant mean function; while the rational quadratic covariance function has strong generalization performance and is better for unknown data. For predictive power, a Gaussian process regression model for predicting SLM performance using a rational quadratic covariance function is appropriate.

The constant mean function has the following form:

m(x)=C, where C is a constant

The form of the rational quadratic covariance function is as follows: Let x and ′ be different input vectors, let r=||x-x′||, the expression of the rational quadratic kernel function is:

其中α>0,l>0

where α>0, l>0

Its hyperparameters are the mixing parameter α and the length scale parameter l, which are required to be positive numbers, where α is mainly used to control the decay rate of the kernel function. The Gaussian process regression model constructed is f(x)~GP(m(x), k(x)). GP stands for Gaussian process. Among them, the value of the constant mean function is the average value of the performance value of the training group data, and the rational quadratic covariance function needs to adjust the hyperparameters. The constructed Gaussian process regression model adopts a block dichotomy strategy when adjusting the initial value of hyperparameters. For example, divide the hyperparameter α from 0 to 10 into 10 areas at equal intervals, and divide the hyperparameter initial value at each part of the node Substitute into the model, and calculate the root mean square error for comparison. The smaller the root mean square error, the better the predictive ability of the hyperparameter model, and intuitively grasp the influence of the overall hyperparameter on the predictive ability of the model. Then select the appropriate block area, and then adopt the binary strategy. Each test takes the midpoint of the two adjacent nodes with the largest distance as the hyperparameter test point, gradually refines the area, and performs multiple experiments until a satisfactory hyperparameter is obtained. The initial value, scoring in the manner specified in S4. Before using the binary strategy test, the overall change trend can be known through the block, and some inappropriate areas can be eliminated, the binary strategy test area can be reduced, and the test workload can be reduced.

②Construction of multiple stepwise regression model: Stepwise regression method is used to sequentially select an independent variable that contributes most significantly to the variance from the optional variables and add it to the regression model. When introducing new variables, check the introduced independent variables one by one, and eliminate them insignificantly until no new independent variable can be introduced into the regression equation, and no independent variable can be eliminated from the regression equation. Considering that there is mutual influence among the process parameters of SLM forming, the optional independent variables include laser power, powder layer thickness, scanning speed and scanning distance, and the optional independent variables include laser power, powder layer thickness, scanning speed and scanning distance. , should also include all items formed by the permutation and combination of four process parameters, and the order of the highest item is determined according to requirements. Variable selection rules need to be set manually, for example, if the variable's significant P value is less than 0.05, it will be introduced into the regression equation; when the variable's significant P value is greater than 0.10, it will be removed from the regression equation. The calculation process of the multiple stepwise regression model is complicated, and it can be solved with the help of statistical analysis software, such as using SPSS software to solve the multiple stepwise regression model for the training set data. For the constructed multiple stepwise regression model, the scoring rules in S4 were used to evaluate the model. By adjusting the variable selection rules of the model, the multiple stepwise regression model was changed, and the multiple stepwise regression model was optimized to improve the predictive ability.

S4. The test and evaluation system of Gaussian process regression model and multiple stepwise regression model. The prediction effects of the Gaussian process regression model and the multiple stepwise regression model are weighted and evaluated by predicting the test set data and calculating the mean absolute error MAE, root mean square error RMSE and coefficient of determination ^R2 . The mean absolute error (MAE) represents the average value of the absolute error between the predicted value and the observed value. The smaller the MAE, the smaller the gap between the overall forecast and the actual value.

The mean square error MSE and the root mean square error RMSE are a measure that reflects the degree of difference between the estimator and the estimated quantity. The smaller the MSE, the smaller the difference between the predicted data and the actual data. The root mean square error RMSE is determined by the mean The square root error is obtained by taking the root sign. The coefficient of determination R ² reflects the fitting degree of the regression model to the actual value. R ² is between 0 and 1. When R ² = 0, it means that the model does not fit the data at all. When R ² = 1, it means that the model is consistent with The data fit perfectly, and the larger the value of ^R2 , the better the fit of the model.

The above three kinds of evaluation data have their own emphases, and a single evaluation only evaluates the prediction results from one aspect, and cannot make a more comprehensive evaluation of the prediction results. In order to make full use of the three evaluation indicators, the present invention sets the following evaluation criteria: if the MAE is less than 0.01, then the score is 100 points; The score is 100 points, if the RMSE is less than 0.02, the score is 90 points, and the order is uniformly recursive; if the R ² is greater than 0.95, the index score is 100 points, and if the R ² is greater than 0.85, the index score is 90 points, and the order is uniform Recursion. Since the forming process and data acquisition are affected by external factors with fluctuation interference, slight deviations between the actual and predicted performance data are allowed, so each evaluation index is rated as 100 points when the error is extremely small.

After the model prediction is completed, more attention should be paid to the deviation between the model's predicted results and the actual results. Therefore, among the three evaluation indicators, it is necessary to focus on MAE and RMSE. When the three indicators are weighted to calculate the final score, the respective weights are: MAE:RMSE:R ² =40:40:20, and the final score is calculated accordingly. If the final score is above 90 points, the prediction is considered accurate; if the final score is 80-90 points, the prediction ability is considered excellent; if the final score is 70-80 points, the prediction is considered effective; if the final score is 60-70 points, the prediction ability is considered poor ; If the final score is below 60 points or a single item score is below 60 points, it is considered unpredictable.

S5. Constructing a Gaussian process regression-multiple stepwise regression model. The Gaussian process regression model built by S3 and the multiple stepwise regression model are fused in a weighted manner. For a set of process parameter inputs, the Gaussian process regression model and the multiple stepwise regression model are respectively used to predict the performance of the two performance predictions obtained. The values are fused in a weighted manner, and the prediction results of the two models are assigned appropriate weights, and the results obtained by the weighted calculation are used as the prediction values of the Gaussian process regression-multiple stepwise regression model. Both the Gaussian process regression model and the multiple stepwise regression model can be used to predict the performance of SLM process parameters, but each has its own advantages and disadvantages. The Gaussian process regression model is very accurate in predicting the data "closer" to the known point, but the prediction accuracy drops greatly when the prediction is "far away" from the known point, that is, the prediction of the local data point is very accurate, but the prediction of the global area is somewhat Insufficient; while the prediction of the multiple stepwise regression model is not as accurate as the Gaussian process regression model in the local area, the prediction ability of the multiple stepwise regression model in the global scope is stable, and its prediction accuracy in the global scope is also better. The Gaussian process regression model and the multiple stepwise regression model are fused in a weighted manner, and the prediction results of the multiple stepwise regression model are corrected by the prediction results of the Gaussian process regression model in the area "near" the known point, and the prediction result of the multiple stepwise regression model is corrected in the area "far from" the known point by The prediction results of the multiple stepwise regression model correct the prediction results of the Gaussian process regression model, and a Gaussian process regression-multiple stepwise regression model with more prominent predictive ability in the global scope is obtained.

In the scoring system of S4, when the scores of the Gaussian process regression model and the multiple stepwise regression model both reach 90 points, it is considered that the two models can be fused to construct the Gaussian process regression-multiple stepwise regression model. The advantages and disadvantages of the weighted weights of the two separate models in the Gaussian process regression-multiple stepwise regression model constructed are evaluated by the Gaussian process regression-multiple stepwise regression model to predict the results of the test set data. The scoring method adopts the evaluation system in S4. The weighting weight is required to be accurate to 0.01. Using the traversal method, the optimal weight structure can be accurately solved through the program looping 99 times. At this time, a more stable and accurate Gaussian process regression-multiple stepwise regression model for SLM performance prediction is obtained. The Gaussian process regression-multiple stepwise regression model constructed by the fusion of two discrete models has absorbed the respective advantages of the two models. Globally, the performance of SLM formed parts can be more accurately predicted by SLM process parameters.

S6. Using the teaching and learning algorithm to optimize and obtain the recommended process parameters of the SLM process parameters. Through the optimization of the teaching and learning algorithm, the individual with the optimal fitness value can obtain the process parameters with the optimal performance value. Compared with traditional optimization algorithms, such as genetic algorithm and particle swarm algorithm, the teaching and learning algorithm has a simple principle and is easy to implement. There are very few parameters that need to be tuned, and its calculation efficiency is higher than that of traditional methods. It is suitable for Optimization of the above SLM performance prediction model. First, an appropriate number of initial groups is set, and the Gaussian process regression-multiple stepwise regression prediction model obtained in S4 is used as the fitness function of the optimization algorithm. By calculating the individual fitness, the individual with the largest fitness value is selected as the teacher. The initial group is updated through the "teaching stage" and "learning stage". The current optimal fitness individual is the current optimal performance value and the corresponding process parameters, and the process parameters corresponding to the optimal performance value are used as the recommended process parameters.

S7. Step by step screening and verification of SLM recommended process parameters. During the SLM forming process, if the process parameters are not selected properly, it is easy to cause damage to the equipment and reduce the service life of the equipment. The recommended process parameters do not guarantee that all process parameters are suitable, and the process parameters need to be screened. Randomly divide the data set, run 10 times of Gaussian process regression-multiple stepwise regression model SLM performance prediction and teaching and learning algorithm process parameter optimization, 10 different recommended process parameters can be obtained, and 10 groups of process parameters can be screened. Among the four process parameters of laser power, powder layer thickness, scanning speed, and scanning distance, laser power and scanning speed have a great influence on the life of the equipment. Specifically, the impact of laser power on the service life of the equipment is greater than that of scanning speed. The thresholds of the two process parameters need to be set according to the laser power and scanning speed of the actual equipment.

The present invention sets the laser power at 40% to 60% of the maximum power of the laser equipment as the first level; at 20% to 40% or 60% to 80% of the maximum power as the second level; at less than 20% of the maximum power or greater than the maximum 80% of the power is level three, the lower the rating of the laser power, the less damage the process parameters will have on the equipment. The rating of the scan speed is consistent with the rating of the laser power, both based on the percentage of the maximum value of the equipment, the lower the rating, the less damage. According to the laser power and scanning speed, 10 groups of recommended process parameters were screened. Firstly, according to the rating standard of laser power, the process parameter groups rated as the third level of laser power were eliminated, and the process parameter groups of the first level and the second level of laser power were respectively reassessed. Continue to subdivide according to the rating of scanning speed, eliminate the three-level process parameter group of scanning speed, and finally according to ① laser power level 1, scanning speed level 1, ② laser power level 1, scanning speed level 2, ③ laser power level 2, scanning speed Speed level one, ④ laser power level two scanning speed level two are divided into four categories. After step-by-step screening, 10 groups of process parameters are removed, and displayed in order from ① to ④ from top to bottom, and the recommended process parameters after screening are obtained. The screened recommended process parameters are then actually verified, and the screened process parameters are input into the device for printing. On the one hand, the prediction performance can be verified, and on the other hand, the data set can be expanded. For the process parameter groups after step-by-step screening, the process parameter groups with satisfactory verification results are the process parameter groups for industrial production of high-performance SLM formed parts, which can be included in the industrial production list of high-performance SLM formed parts; parameter group, add the unsatisfactory process parameter group data to the data set, update the data set, and repeat steps S2 to S7 to obtain new recommended process parameters.

Furthermore, before the individual calculates the fitness according to the Gaussian process regression-multiple stepwise regression model in S5, it is necessary to consider the accuracy of the input process parameters of the equipment, such as the accuracy of the laser power and the accuracy of the scanning speed. The input of each individual is converted into the process parameter value that can be input by the corresponding equipment according to the rounding method, and then used as the input of the individual for subsequent operations.

Example 1

The SLM performance prediction and process parameter optimization method of this embodiment obtains process parameters and performance data through design experiments, and randomly divides the acquired data into training sets and test sets, constructs a Gaussian process regression model and a multiple stepwise regression model, and adjusts The initial value of the hyperparameters enables the two models constructed to accurately predict the performance parameters of the test set. Then, the above two models will be fused to construct a Gaussian process regression-multiple stepwise regression model, which will be fused with the teaching and learning algorithm, and the obtained fusion model will be used as the fitness function to obtain the optimal performance and the best performance through the optimization of the teaching and learning algorithm. Process parameters under optimal performance, that is, optimized process parameters. The process parameters input in this example include laser power, powder layer thickness, scanning speed, scanning distance, and output density. The performance of density is predicted. If other performance parameters need to be predicted, the operation is consistent with the predicted density. Re-tuning and optimizing hyperparameters.

Firstly, data is obtained by designing experiments, and multi-step orthogonal experiments are designed. Firstly, the orthogonal experiment is designed in the global scope, and the SLM formed parts with the process parameters of the preliminary design are actually manufactured, and the density is detected; then, according to the density results of the SLM formed parts obtained from the preliminary design, the suspicious areas with better detection performance are further analyzed. Design an orthogonal experiment, manufacture and test the density again, and obtain a data set of the experimental group that covers the whole world and has a reasonable density.

In this embodiment, 60 sets of data are taken, and 50 sets are randomly selected as training set data, and 10 sets are used as test set data.

Build a Gaussian process regression model and train it with the training set data. A Gaussian process regression model was constructed using a constant mean function and a rational quadratic covariance function. The predictive ability of the Gaussian process regression model is evaluated by the results of the three indicators of MAE, RMSE, and ^R2 between the predicted value and the actual value of the data density of the test set. The specific value of the constant mean function is determined by the dense mean value of the training set data; the hyperparameters in the rational quadratic covariance function are determined, and the hyperparameters α and l are adjusted through the block dichotomy strategy to continuously improve the score of the Gaussian process regression model. Improve the prediction result score of the model to the test set data to more than 90 points, and finally obtain an accurate SLM density prediction model. The calculation part of the Gaussian process regression model is solved by the computer through the existing program, but the hyperparameter test and modification part needs manual operation.

Build a multiple stepwise regression model. In order to further improve the prediction accuracy of the multivariate stepwise regression model, a multivariate third-order stepwise regression model is constructed in this embodiment. According to the existing four process parameters, the second-order square term, the second-order cross term, the third-order cubic term and the third-order cross term are calculated, and a complex including all first-order terms, second-order terms and third-order terms is constructed. Complete alternative arguments. Then adopt the method of stepwise regression. When introducing new variables, carry out the significance test on the independent variables that have been introduced one by one, and eliminate the insignificant ones until no new independent variables can be introduced into the regression equation, and at the same time, no one can be eliminated from the regression equation. As far as one independent variable is concerned, the construction of the multivariate three-order stepwise regression model is completed. The predictive ability of the multivariate third-order stepwise regression model is also evaluated by the results of the three indicators of MAE, RMSE, and ^R2 between the predicted value and the actual value of the data density of the test set. The multivariate third-order stepwise regression model optimizes the model by adjusting the variable selection rules, so that the multivariate third-order stepwise regression model can score more than 90 points on the prediction results of the test set. The calculation and stepwise regression analysis of the multivariate third-order stepwise regression model were completed by SPSS statistical analysis software.

Build a Gaussian process regression-multivariate three-order stepwise regression model. The already constructed Gaussian process regression model and the multivariate third-order stepwise regression model are fused by weighting, and the program loops through 100 times to accurately solve the weighted weight structure with an accuracy of 0.01, and obtain a Gaussian with better predictive ability and more stability Process Regression - Multivariate three-order stepwise regression model.

Integrate Gaussian process regression-multivariate third-order stepwise regression model teaching and learning algorithms, and optimize through teaching and learning algorithms. Set the initial population size to 100, cycle 200 times, and initialize the population, that is, assign initial values to the four inputs of each individual. The obtained Gaussian process regression-multivariate three-order stepwise regression model is used as the fitness function, input into the process parameter group, and the output density prediction value is used as the fitness value, and the size of the fitness value is used as the evaluation of the individual's pros and cons. The population continues to go through the "teaching" stage and the "learning" stage, and is constantly updated until the condition for the end of the cycle is met, the optimization process of the teaching and learning algorithm is completed, and the individual with the maximum fitness value is obtained, and then the process with the most usable performance and optimal performance is obtained. Parameters, and the process parameters with the best performance as the recommended process parameters. Repeat the operation 10 times, and 10 different recommended process parameters will be obtained. For 10 groups of recommended process parameters, through the step-by-step screening of laser power and scanning speed, the process parameter groups are classified according to the actual maximum parameters of the equipment, and inappropriate recommended process parameters are eliminated to prevent inappropriate process parameters from causing damage to the equipment . The screened process parameter groups are divided into four categories: ①Laser power level 1, scanning speed level 1; ②Laser power level 1, scanning speed level 2; ③Laser power level 2, scanning speed level 1; ④Laser power level 2 , Scanning speed two. The recommendation degree of the four types of recommended process parameters from ① to ④ is gradually reduced, and the four types of process parameters are arranged from ① to ④ according to the rule from top to bottom, as the screened process parameters.

Finally, the actual verification of the screened process parameters was carried out. The screened and recommended process parameters are then actually verified, and the screened process parameters are input into the device for printing and its performance is tested for verification. For the process parameter groups after step-by-step screening, the process parameter groups with satisfactory verification results are the process parameter groups for industrial production of high-performance SLM formed parts, which are included in the list of industrial production of high-performance SLM formed parts; unsatisfactory processes are verified parameter group, save the data and add it to the data set, expand the data set, and repeat the above steps to obtain new recommended process parameters.

Fig. 4 is a system architecture diagram of SLM forming performance prediction and process parameter optimization provided by the embodiment of the present invention, as shown in Fig. 4, including:

The parameter data set design unit 410 is used to design multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacture the designed process parameters, determine the performance of the corresponding SLM formed parts, and combine the designed SLM process parameters with the corresponding forming The performance of the software is summarized as a data set.

The regression model training unit 420 is used to train the Gaussian process regression model based on the data set to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; The training is carried out to obtain the second mapping relationship between the SLM process parameters and the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are both used to predict the performance of the SLM formed part.

The regression model combination unit 430 is used to combine the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model combines the Gaussian process regression model and the multiple stepwise regression model The two performance prediction results of are fused in a weighted way, and the weight of the weighted way is determined through an ergodic way.

The prediction model optimization unit 440 is used to optimize the SLM performance prediction model through the teaching and learning algorithm, and obtain multiple sets of recommended SLM process parameters whose performance of the SLM formed part meets the requirements.

The process parameter screening unit 450 is used to screen the multiple sets of recommended SLM process parameters step by step according to the influence of the SLM process parameters on the service life of the equipment, so as to obtain multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters.

The process parameter grouping unit 460 is used to perform actual manufacturing on the multiple sets of SLM process parameters after step-by-step screening, to verify the performance of the obtained SLM formed parts, and to list the parameters whose performance meets the expected standard into the process parameter groups with satisfactory verification results, As a process parameter group of high-performance SLM formed parts; the parameters whose performance does not meet the expected standard are included in the process parameter group whose verification results are not satisfactory.

The new parameter recommendation screening unit 470 is used to add the process parameter group whose verification result is not satisfactory to the data set, update the data set, repeat the training of Gaussian process regression model, the training of multiple stepwise regression model, the recommendation and screening process of SLM process parameters , to obtain multiple sets of SLM process parameters after the new screening.

It can be understood that, for the detailed function implementation of each of the above units, reference may be made to the introduction in the foregoing method embodiments, and details are not repeated here.

In addition, an embodiment of the present invention provides an electronic device, which includes: a memory and a processor;

The memory is used to store computer programs;

The processor is configured to implement the methods in the foregoing embodiments when executing the computer program.

In addition, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the methods in the above-mentioned embodiments are implemented.

Based on the methods in the foregoing embodiments, an embodiment of the present invention provides a computer program product, which causes the processor to execute the methods in the foregoing embodiments when the computer program product runs on a processor.

Based on the methods in the foregoing embodiments, an embodiment of the present invention further provides a chip, including one or more processors and an interface circuit. Optionally, the chip can also include a bus. in:

A processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software. The above-mentioned processor may be a general-purpose processor, a digital communicator (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components. Various methods and steps disclosed in the embodiments of the present application may be implemented or executed. A general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The interface circuit can be used for sending or receiving data, instructions or information, and the processor can process the data, instructions or other information received by the interface circuit, and can send the processing completion information through the interface circuit. Optionally, the chip further includes a memory, which may include a read-only memory and a random access memory, and provides operation instructions and data to the processor. A portion of the memory may also include non-volatile random access memory (NVRAM).

Optionally, the memory stores executable software modules or data structures, and the processor can execute corresponding operations by calling operation instructions stored in the memory (the operation instructions can be stored in the operating system). Optionally, the interface circuit can be used to output the execution result of the processor. It should be noted that the corresponding functions of the processor and the interface circuit can be realized by hardware design, software design, or a combination of software and hardware, which is not limited here. It should be understood that each step in the foregoing method embodiments may be implemented by logic circuits in the form of hardware or instructions in the form of software in the processor.

It can be understood that the sequence numbers of the steps in the above embodiments do not mean the order of execution, and the execution order of each process should be determined by its functions and internal logic, and should not constitute any obligation for the implementation process of the embodiment of the present application. limited. In addition, in some possible implementation manners, the steps in the foregoing embodiments may be selectively executed according to actual conditions, may be partially executed, or may be completely executed, which is not limited here.

It can be understood that the processor in the embodiment of the present application may be a central processing unit (central processing unit, CPU), and may also be other general purpose processors, digital signal processors (digital signal processor, DSP), application specific integrated circuits ( application specific integrated circuit (ASIC), field programmable gate array (field programmable gate array, FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. A general-purpose processor can be a microprocessor, or any conventional processor.

The method steps in the embodiments of the present application may be implemented by means of hardware, or may be implemented by means of a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in random access memory (random access memory, RAM), flash memory, read-only memory (read-only memory, ROM), programmable read-only memory (programmable rom) , PROM), erasable programmable read-only memory (erasable PROM, EPROM), electrically erasable programmable read-only memory (electrically EPROM, EEPROM), register, hard disk, mobile hard disk, CD-ROM or known in the art any other form of storage medium. An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. Of course, the storage medium may also be a component of the processor. The processor and storage medium can be located in the ASIC.

In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present application will be generated in whole or in part. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in or transmitted via a computer-readable storage medium. The computer instructions may be transmitted from one website site, computer, server, or data center to another website site by wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) , computer, server or data center for transmission. The computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrated with one or more available media. The available medium may be a magnetic medium (such as a floppy disk, a hard disk, or a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (solid state disk, SSD)) and the like.

It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention, All should be included within the protection scope of the present invention.

Claims (10)

1. A SLM formability prediction and process parameter optimization method, is characterized in that, comprises the following steps: Design multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacture the designed process parameters to determine the performance of the corresponding SLM formed parts, and summarize the designed SLM process parameters and the corresponding performance of the formed parts into a data set; Based on the data set, the Gaussian process regression model is trained to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; and based on the data set, the multiple stepwise regression model is trained to obtain the SLM process parameters and the SLM formed part performance. The second mapping relationship between the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are used to predict the performance of the SLM formed part; Combining the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model uses two performance prediction results of the Gaussian process regression model and the multiple stepwise regression model. mode fusion, the weight of the weighted mode is determined through traversal; Optimize the SLM performance prediction model through the teaching and learning algorithm, and obtain multiple sets of recommended SLM process parameters whose performance of the SLM formed part meets the requirements; According to the influence of the SLM process parameters on the service life of the equipment, the multiple sets of recommended SLM process parameters are screened step by step, and multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters are obtained. 2. The method according to claim 1, further comprising the steps of: For the multiple sets of SLM process parameters after step-by-step screening, the actual manufacturing is carried out to verify the performance of the obtained SLM formed parts, and the parameters whose performance meets the expected standards are included in the process parameter group with satisfactory verification results, as the process of high-performance SLM formed parts Parameter group; the parameters whose performance does not meet the expected standard are included in the process parameter group whose verification result is not satisfactory; Add the unsatisfactory process parameter group to the data set, update the data set, repeat the training of Gaussian process regression model, the training of multiple stepwise regression model, the recommendation and screening process of SLM process parameters, and obtain multiple sets of SLM after new screening Process parameters. 3. method according to claim 1, is characterized in that, design multiple groups of SLM process parameters by multi-step orthogonal experiment, specifically: Design multi-step orthogonal experiments, design SLM process parameter groups through orthogonal experiments on a global scale, and actually manufacture the designed process parameters to test the performance of the preliminary designed SLM formed parts; Carry out a preliminary test on the global performance of SLM, find out the suspicious areas with better performance of SLM formed parts in the preliminary detection, and design orthogonal experiments for suspicious areas, and obtain the correspondence of suspicious areas with better performance of SLM formed parts after many times of designing orthogonal experiments Multiple sets of SLM process parameters. 4. The method according to any one of claims 1 to 3, wherein the Gaussian process regression model is trained based on the data set, and the multiple stepwise regression model is trained based on the data set, specifically: Build a Gaussian process regression model; train the Gaussian process regression model based on the designed multiple sets of SLM process parameters; Construct a multiple stepwise regression model; train the multiple stepwise regression model based on the designed multiple sets of SLM process parameters; among them, taking into account the interaction between process parameters, increase high-order items and cross items to construct complete alternatives; obtain a complete After selecting the alternatives, through the significance test, the alternatives are gradually introduced and eliminated. 5. method according to claim 4, it is characterized in that, in Gaussian process regression model and multiple stepwise regression model training process, by mean absolute error MAE, root mean square error RMSE and coefficient of determination R ² three indexes are paired The weighted evaluation of the two models is carried out to evaluate the deviation between the model prediction results and the actual results, and to optimize the model parameters; the respective weights of the three indicators are: MAE:RMSE:R ² =40:40:20. 6. The method according to claim 1, characterized in that, the SLM performance prediction model is optimized by a teaching and learning algorithm to obtain multiple groups of recommended SLM process parameters whose performance of the SLM formed part meets the requirements, specifically: Set an appropriate number of initial groups, and use the SLM performance prediction model as the fitness function of the optimization algorithm; By calculating the individual fitness, the individual with the largest fitness value is selected as the teacher. The initial group is updated through the "teaching stage" and "learning stage". The current optimal fitness individual, that is, the current optimal performance value and the corresponding process parameters, and the process parameters corresponding to the optimal performance value are used as the recommended SLM process parameters. 7. The method according to claim 1, wherein the SLM process parameters include: laser power, powder layer thickness, scanning speed and scanning distance; laser power and scanning speed have a relatively large impact on the life of the equipment, and laser Power has a relatively large impact on the process stability of the equipment. Set multiple thresholds corresponding to the two process parameters according to the laser power and scanning speed of the actual equipment. When the laser power or scanning speed of the actual equipment is in different threshold ranges, the laser power or The scanning speed is in different ratings, and different ratings have different losses to the equipment; According to the impact of the SLM process parameters on the service life of the equipment, the multiple groups of recommended SLM process parameters are screened step by step, and multiple groups of SLM process parameters with low damage to the service life of the equipment are obtained, specifically: The multiple groups of recommended SLM process parameters are in accordance with: ①Laser power level 1, scanning speed level 1, ②Laser power level 1, scanning speed level 2, ③Laser power level 2, scanning speed level 1, ④Laser power level 2, scanning speed 2 The levels are divided into four categories; among them, the laser power level has the lowest damage to the service life of the equipment, and as the laser power level increases, the damage to the equipment life increases; the scanning speed level has the lowest damage to the equipment life, The speed level increases, and the damage to the service life of the equipment increases; After step-by-step screening, the laser power and scanning speed among multiple sets of recommended SLM process parameters are eliminated, and then displayed in order from ① to ④ from top to bottom, and multiple sets of process parameters after screening are obtained. 8. A SLM forming performance prediction and process parameter optimization system, characterized in that it comprises: The parameter data set design unit is used to design multiple sets of SLM process parameters through multi-step orthogonal experiments, and actually manufacture the designed process parameters, determine the performance of the corresponding SLM formed parts, and combine the designed SLM process parameters with the corresponding formed parts performance aggregated as a dataset; The regression model training unit is used to train the Gaussian process regression model based on the data set to obtain the first mapping relationship between the SLM process parameters and the performance of the SLM formed part; and based on the data set, the multiple stepwise regression model is performed Training to obtain the second mapping relationship between the SLM process parameters and the performance of the SLM formed part; the first mapping relationship and the second mapping relationship are used to predict the performance of the SLM formed part; The regression model combination unit is used to combine the trained Gaussian process regression model and the trained multiple stepwise regression model to obtain the SLM performance prediction model; wherein, the SLM performance prediction model combines the Gaussian process regression model and the multiple stepwise regression model. The two performance prediction results are fused in a weighted manner, and the weight of the weighted manner is determined by traversal; The prediction model optimization unit is used to optimize the SLM performance prediction model through the teaching and learning algorithm, and obtain multiple sets of recommended SLM process parameters for which the performance of the SLM formed part meets the requirements; The process parameter screening unit is used to screen the multiple sets of recommended SLM process parameters step by step according to the influence of the SLM process parameters on the service life of the equipment, so as to obtain multiple sets of SLM process parameters with low damage to the service life of the equipment and stable process parameters. 9. The system according to claim 8, further comprising: The process parameter grouping unit is used to carry out actual manufacturing of multiple sets of SLM process parameters after step-by-step screening, to verify the performance of the obtained SLM formed parts, and to list the parameters whose performance meets the expected standard into the process parameter group with satisfactory verification results, as The process parameter group of high-performance SLM formed parts; the parameters whose performance does not meet the expected standard are included in the process parameter group with unsatisfactory verification results; The new parameter recommendation screening unit is used to add process parameter groups with unsatisfactory verification results to the data set, update the data set, repeat Gaussian process regression model training, multiple stepwise regression model training, SLM process parameter recommendation and screening process, Multiple sets of SLM process parameters after the new screening are obtained. 10. The system according to claim 8 or 9, wherein the SLM process parameters include: laser power, powder layer thickness, scanning speed and scanning distance; laser power and scanning speed have a relatively large impact on the life of the equipment , the laser power has a relatively large impact on the process stability of the equipment. Set multiple thresholds corresponding to the two process parameters according to the laser power and scanning speed of the actual equipment. When the laser power or scanning speed of the actual equipment is in different threshold ranges, the laser The power or scanning speed is in different ratings, and different ratings have different losses to the equipment; The process parameter screening unit uses multiple groups of recommended SLM process parameters according to: ① laser power level 1, scanning speed level 1, ② laser power level 1, scanning speed level 2, ③ laser power level 2, scanning speed level 1, ④ The second level of laser power and the scanning speed are divided into four categories; among them, the damage to the service life of the equipment at the first level of laser power is the lowest, and the damage to the service life of the equipment increases with the increase of the laser power level; the damage to the service life of the equipment increases at the level of scanning speed; The damage is the lowest, and as the scanning speed level increases, the damage to the service life of the equipment increases; after step-by-step screening, the laser power and scanning speed in multiple groups of recommended SLM process parameters are eliminated, and the parameters from ① to The order of ④ is displayed from top to bottom, and multiple sets of process parameters after screening are obtained.

Images