FDM-3D printing multi-target process optimization method driven by fusion of point cloud and response curved surface

Abstract

The invention provides a point cloud and response curved surface fusion driven FDM-3D printing multi-target process optimization method, which comprises the steps of building a whole-process monitoring data acquisition platform, acquiring multi-modal sensor data in a 3D printing process and extracting time domain and frequency domain features; acquiring high-precision point cloud data of a printing sample based on a three-dimensional scanning technology, and extracting roughness parameters of each surface area; constructing a regression model of the process parameters and the multi-region roughness parameters, and analyzing the main effect and interaction of the process parameters; integrating a composite satisfaction function and an entropy weight method, and dynamically distributing objective weights of the multi-region quality indexes; an optimal combination of process parameters is determined through a multi-objective optimization method, and the surface roughness of multiple areas is reduced; the surface roughness of all the side faces of the 3D printing part is influenced, and optimal optimization process parameters of all the side faces are obtained; according to the method, the comprehensive surface roughness can be reduced through parameter coordinated regulation and control, and meanwhile, the surface uniformity of the top area, the periphery area and the bottom area is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of 3D printing surface quality optimization, and in particular to a multi-objective process optimization method for FDM-3D printing driven by the fusion of point cloud and response surface. Background Art

[0002] Additive Manufacturing (AM), a discrete-accumulation molding technology based on digital models, demonstrates significant potential in fields such as medicine, architecture, and aerospace through its layer-by-layer processing strategy. 3D printing technologies, such as Fused Deposition Modeling (FDM), support free-form manufacturing and mass customization, offering both flexibility and sophistication, making them key enabling technologies for the clinical implementation of precision biomaterials. However, the mainstream 3D printing process, FDM, still suffers from issues such as weak interlayer bonding and rough surfaces.

[0003] Existing process optimization methods have the following limitations: First, surface morphology characterization mostly relies on contact measurement (such as stylus roughness testers), which has low spatial resolution and is easy to damage the sample, making it difficult to achieve high-precision multi-region morphology quantification; second, existing optimization methods mostly focus on a single roughness index (such as Ra), ignoring the collaborative optimization of surface quality in different areas such as the top, sides and bottom of the printed part, resulting in limited overall performance improvement; third, the multi-objective weight allocation relies on manual experience, lacks objectivity, and is prone to introduce subjective bias, affecting the credibility of the optimization results. In addition, the traditional response surface method does not fully combine high-resolution point cloud data, making it difficult to accurately analyze the impact mechanism of process parameter interactions on multi-region surface quality. Therefore, there is an urgent need for a process optimization method that integrates high-precision point cloud analysis, multi-region quality collaborative optimization and objective weight decision-making to break through the existing technical bottleneck and improve the comprehensive surface quality of FDM printed parts. Summary of the Invention

[0004] The purpose of the present invention is to provide a multi-objective process optimization method for FDM-3D printing driven by the fusion of point cloud and response surface. The proposed optimization framework can reduce the overall surface roughness through coordinated parameter control, while significantly improving the surface uniformity of the top, surrounding and bottom areas, providing reliable theoretical support and process guidance for high-precision FDM 3D printing manufacturing.

[0005] In order to achieve the above-mentioned object of the invention, the present invention provides a multi-objective process optimization method for FDM-3D printing driven by the fusion of point cloud and response surface, comprising the following steps:

[0006] S1. Build a full-process monitoring data acquisition platform to collect multimodal sensor data in real time during the 3D printing process and extract time and frequency domain features;

[0007] S2. Obtain high-precision point cloud data of the printed sample based on 3D scanning technology and extract the roughness parameters of each surface area;

[0008] S3. Based on the process parameters and multi-region roughness parameters, the regression equations of each surface are obtained and the Pareto chart of the standardized effects is drawn to analyze the main effects and interactions of the process parameters.

[0009] S4, integrating the composite satisfaction function and the entropy weight method to dynamically assign objective weights to multi-region quality indicators;

[0010] S5. Determine the optimal combination of process parameters through a multi-objective optimization method to achieve reduction of surface roughness in multiple regions;

[0011] S6. Use response surface analysis to analyze the effect of process parameters on the surface roughness of each side of the 3D printed part and obtain the best optimized process parameters for each side.

[0012] Preferably, in step S1, the multimodal sensor data includes vibration and temperature data, and the time domain and frequency domain characteristics of the nozzle and the hot bed are collected respectively by a vibration sensor and a thermocouple sensor installed on the nozzle and the hot bed, and the frequency domain characteristics are extracted from the vibration sensor using a fast Fourier transform algorithm.

[0013] Preferably, in step S2, the roughness parameters include the surface arithmetic mean deviation Sa and the root mean square deviation Sq, and the Sa and Sq values ​​of the top, surrounding and bottom areas of the printed part are calculated by a non-contact measurement method.

[0014] Preferably, in steps S1-S2, data collection is performed through an FDM 3D printer, a vibration sensor, a thermocouple sensor, a three-dimensional scanning device, and a terminal computer, wherein the three-dimensional scanning device is used to generate an STL file of the printed sample and perform point cloud comparison with the standard model, and finally calculate the surface arithmetic mean deviation Sa and root mean square deviation Sq roughness parameters of each surface of the 3D printed part.

[0015] Preferably, in step S3, a quadratic polynomial regression model is constructed based on a Box-Behnken experimental design with process parameters as inputs and Sa and Sq values ​​as responses, wherein the process parameters include extruder temperature, filling density, extrusion rate, Z-axis height, and hot bed temperature, and the quadratic polynomial regression model covers the Sa and Sq responses of each surface area;

[0016] Regression equations were established for the arithmetic mean roughness Sa and root mean square roughness Sq respectively. A Pareto chart was drawn based on the effect values ​​of the interactions between the process parameters in the quadratic polynomial regression model to obtain the process parameters with a greater impact on surface roughness.

[0017] Preferably, in step S4, the entropy weight method calculates the weight of each roughness parameter based on information entropy, the composite satisfaction function is used to normalize the multi-objective response value, and finally a weighted comprehensive defect function WCDF is generated as the optimization index.

[0018] Preferably, in step S5, the optimal process parameter combination is: extruder temperature 180°C, filling density 52.5%, extrusion rate 95%, Z-axis height -1.565mm, and hot bed temperature 62.5°C.

[0019] Preferably, the Box-Behnken experimental design includes three-level combinations of five process parameters, and a total of 46 groups of experiments are designed. The Pareto chart of the standardized effect in the quadratic polynomial regression model is used to identify parameters that significantly affect surface roughness.

[0020] Preferably, the method further comprises analyzing the influence trend of process parameters on the roughness of each surface area by response surface method, and generating a dynamic adjustment strategy to achieve real-time optimization.

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

[0022] 1. This invention uses 3D scanning technology to obtain high-resolution point cloud data. Combined with non-contact measurement methods (such as calculating the surface arithmetic mean deviation Sa and root mean square deviation Sq), it breaks through the spatial resolution limitations of traditional contact measurement and can accurately quantify the 3D topographic features of multiple areas such as the top, sides, and bottom of the printed part, providing a reliable data basis for process optimization.

[0023] 2. A multi-region roughness regression model covering the top, sides, and bottom is established. The Box-Behnken experimental design is used to analyze the differential effects of process parameters (such as extrusion temperature, filling density, and Z-axis height) on each region. This solves the one-sidedness of the optimization of a single roughness index in the existing technology and significantly improves the overall surface uniformity of the printed part.

[0024] 3. Innovatively integrate the entropy weight method and the composite satisfaction function, dynamically allocate multi-objective weights based on information entropy, eliminating the defects of traditional methods that rely on subjective experience or equal assumptions.

[0025] 4. Through the quadratic polynomial regression model and response surface methodology, the nonlinear coupling relationship between process parameters (such as the synergistic effect of extrusion temperature and Z-axis height on bottom surface roughness) is quantified, and the standardized effect Pareto chart is used to identify key parameters, significantly improving the scientificity and pertinence of parameter optimization.

[0026] 5. The generated dynamic adjustment strategies (such as adjusting the extrusion rate based on real-time vibration data and controlling the hot bed cooling rate based on temperature feedback) lay the foundation for future intelligent printing systems and support real-time quality optimization of the entire surface of complex structural parts.

[0027] 6. The optimal parameter combination determined through multi-objective optimization greatly reduces the overall surface roughness while ensuring structural strength, greatly improving the accuracy compared to traditional processes, and the printing efficiency is not significantly affected.

[0028] In summary, the present invention realizes the intelligent coordinated control of FDM 3D printing process parameters through data-driven and model fusion, provides theoretical support and technical guarantee for high-precision additive manufacturing, and has significant engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0030] Figure 1 Flowchart of the multi-objective process optimization method for FDM-3D printing driven by point cloud and response surface fusion provided in an embodiment of the present invention;

[0031] Figure 2 A diagram of vibration extraction data provided by an embodiment of the present invention;

[0032] Figure 3 This is a temperature extraction data diagram provided in an embodiment of the present invention;

[0033] Figure 4 This is a distribution diagram of the average values ​​of the surface roughness parameters of each side surface provided in the embodiment of the present invention;

[0034] Figure 5 Normal distribution diagram of residuals on each side provided in the embodiment of the present invention;

[0035] Figure 6 A Pareto diagram of standardized effects on various aspects provided in the embodiments of the present invention;

[0036] Figure 7 This is a surface diagram showing the influence of the process parameters on surface roughness provided in the embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. Of course, the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0038] This embodiment provides a multi-objective process optimization method for FDM-3D printing driven by the fusion of point cloud and response surface, including the following steps:

[0039] 1. Realize data collection through 3D printing system. Figure 1 As shown, the 3D printing system primarily consists of an FDM 3D printer, a monitoring and scanning device, and a terminal computer. The printing equipment used is a Prusa i3MK3S+ printer manufactured by Czech company Prusa Research. It has an extruder diameter of 0.4mm and a build volume of 25×21×21cm. It is equipped with an MK52 magnetic heated bed and a removable spring steel print plate. It supports three surface finishes: smooth, textured, and satin, meeting diverse printing needs.

[0040] The monitoring equipment mainly includes vibration sensors, thermocouple sensors and microscope cameras, and the scanning device is EinScan SE&SP. The computer serves as the terminal control device of the experiment, mainly responsible for controlling the operation of the FDM 3D printer and collecting process data collected by the monitoring equipment.

[0041] Second, extract multimodal sensor data such as vibration and temperature (such as Figure 2-Figure 3 The specific experimental steps are as follows:

[0042] (1) Bind the measuring end to the nozzle and the bottom of the hot bed respectively.

[0043] (2) Use ADXL and SSCOM host computers to record the vibration data of the hot bed and nozzle respectively, and the baud rate is selected as 9600.

[0044] (3) Save the recorded data file and name it.

[0045] (4) Based on the obtained sensor data, the statistical characteristics of each modal data are extracted from the time domain as shown in Formula 1 to Formula 9. Formula 1 to Formula 9 are as follows:

[0046] Max=max(x i ) (1)

[0048] Median=Median(x i ) (2)

[0050]

[0051] Min=min(x i ) (4)

[0053]

[0054] Peak=max(|x i |) (8)

[0056]

[0057] In the formula, Max represents the maximum value of the data sequence, x i represents the i-th data point, and uses the max function to get the maximum value of all data points; Median represents the median of the data sequence, x i represents the i-th data point; Mean represents the arithmetic mean of the data sequence, N represents the total number of data points, x i represents the i-th data point; Min represents the minimum value in the data sequence, x i represents the i-th data point, and uses the min function to get the minimum value among the data points; RMS represents the root mean square value, N represents the total number of data points, and x i represents the i-th data point; Skewness is used to measure the asymmetry of data distribution, where the Mean value is calculated by formula (3), N value is the total number of data points, x i represents the i-th data point; Kurtosis is used to measure the steepness or tail thickness of the data distribution, where the Mean value is still calculated by formula (3), N value is the total number of data points, x i represents the i-th data point; Peak represents the peak value of the data sequence, that is, the maximum absolute value of all data points, x i represents the i-th data point, |x i |Take the absolute value of the i-th data point and use the max function to get the maximum value among all absolute values; Variance represents the degree of dispersion of data distribution, where N represents the total number of data points, x i represents the i-th data point, and the Mean value is obtained by formula (3), x i -Mean value is the difference between each data point and the mean, reflecting the degree of deviation of a single point. (x i -Mean) 2 Squaring the difference eliminates the effect of positive and negative offsets and amplifies the weight of large deviations.

[0058] Furthermore, four frequency domain features are extracted from the vibration sensor using a fast Fourier transform algorithm, as shown in Formula 10 to Formula 13. Formula 10 to Formula 13 are as follows:

[0059]

[0060] Max=max(A k ) (11)

[0062] Min=min(A k ) (12)

[0064] Median=Median(A k ) (13)

[0066] In the formula, Mean value represents the average amplitude of frequency domain data, reflecting the overall energy level of the signal in the frequency domain, A k Represents the amplitude of the kth frequency component, N represents the total number of frequency components. This formula is mainly used to measure the average energy distribution of frequency domain signals; Max represents the maximum amplitude in the frequency domain data, corresponding to the strongest frequency component in the signal, A k Represents the amplitude of the kth frequency component, and uses the max function to take the maximum amplitude among all frequency components. This formula is mainly used to identify the main frequency in the signal; the Min value represents the minimum amplitude in the frequency domain data, corresponding to the weakest frequency component in the signal, A k A represents the amplitude of the kth frequency component. The min function is used to take the minimum amplitude among all frequency components. This formula is used to analyze the intensity of background noise or secondary frequency components. The Median value represents the median amplitude of the frequency domain data, that is, the value in the middle after the amplitude is sorted by size. k Represents the amplitude of the kth frequency component. This formula provides robust statistics of the amplitude distribution and avoids the influence of extreme values.

[0067] 3. Quantification and verification of surface roughness, such as Figure 4 As shown, the average values ​​of the surface roughness parameters of each side were obtained. Analysis of the surface roughness distribution characteristics revealed that the average value of the top area (Sa-top) reached 0.5, and the degree of dispersion (Sq-top) was relatively high, which may be related to the uneven shrinkage of the material under high temperature; while the Sq value of the bottom surface (Sq-bottom) suddenly rose to 1.75. It is speculated that the warpage was caused by improper setting of the Z-axis height, resulting in a significant increase in the dispersion of the bottom surface material distribution. This result highlights the importance of coordinated optimization of process parameters, especially the matching of the Z-axis height and the extrusion rate, which plays a key role in suppressing bottom surface defects. The specific experimental steps are as follows:

[0068] (1) Use scanning equipment to scan the 3D printed parts in all directions to obtain the STL files corresponding to each part.

[0069] (2) The STL file generated by scanning is compared with the SolidWorks standard model in the CloudCompare software.

[0070] (3) The surface arithmetic mean deviation (Sa) and root mean square deviation (Sq) roughness parameters of each surface of the 3D printed part were calculated using CloudCompare software. The results are shown in Table 1.

[0071] Table 1 Point cloud results

[0072]

[0073]

[0074] 4. Regression model construction and residual analysis. The specific steps are as follows:

[0075] (1) Using process parameters as input and surface arithmetic mean deviation (Sa) and root mean square deviation (Sq) as response values, the regression equations for each surface are obtained, such as Formula 14-Formula 19;

[0076]

[0077] In the formula, Sa-top represents the Sa value of the top, Sq-top represents the Sq value of the top, Sa-wall represents the Sa value of the four sides, Sq-wall represents the Sq value of the four sides, Sa-bottom represents the Sa value of the bottom, Sq-bottom represents the Sq value of the bottom, parameter A represents the extruder temperature, parameter B represents the filling density, parameter C represents the extrusion rate, parameter D represents the Z-axis height, and parameter E represents the hot bed temperature.

[0078] (2) Verify the validity of the regression model based on residual analysis, such as Figure 5 As shown in the figure, the residuals of each surface parameter closely fit the normal distribution reference line, without systematic deviation or heteroscedasticity.

[0079] (3) Use the effect values ​​of each parameter in the regression equation to draw a Pareto chart of the standardized effect, such as Figure 6 shown.

[0080] 5. Multi-response result optimization and parameter combination verification

[0081] Through a multi-objective optimization method, the effects of different process parameter combinations on the surface roughness (Sa, Sq) of each region (top, sides, and bottom) of the 3D printed part were comprehensively evaluated, and the weighted composite defect function (WCDF) was introduced as an optimization indicator. Table 2 shows the WCDF values ​​of the 46 experimental groups and their corresponding regional roughness scores.

[0082] Table 2 Expectation function (DF) and weighted composite expectation function (WCDF)

[0083]

[0084]

[0085] Among them, the entropy method is a multi-index weighting method based on information theory, which is used to determine the weight of each indicator in a multi-index evaluation system. Its core idea is to reflect the degree of variation and importance of the indicator by calculating the information entropy of each indicator. The smaller the entropy value, the greater the amount of information and the higher the weight; conversely, the larger the entropy value, the smaller the amount of information and the lower the weight. This study calculates the weight of the response based on the entropy method, as shown in Formula 20-Formula 23

[0086]

[0087] d j =1-e j (twenty three)

[0089] Formula (20) first calculates the data x ij After normalization, μ ij , and then calculate the entropy value e by formula (21) and (22) j Entropy is a concept in information theory and is often used to measure the uncertainty or discreteness of data in satisfaction function. Finally, we can get d by formula (23) j , which is related to the entropy value, reflects certain characteristics of the data and is used for subsequent weight calculation or indicator evaluation.

[0090]

[0091] The formula (24) is used to calculate the weight ω j By adding d j Divide by all d j The sum of the factors is used to obtain the weight of each factor, thereby determining the relative importance of each factor in the calculation of overall satisfaction. The calculation results are shown in Table 3.

[0092] Table 3 Entropy-based response weights

[0093]

[0094] Table 4 shows the optimal levels of each process parameter: extruder temperature (ET) is -1 level (180°C), filling density (FD) is 0 level (52.5%), extrusion rate (ER) is -1 level (95%), Z-axis height (ZH) is -1 level (-1.565mm), and hot bed temperature (BT) is 0 level (62.5°C). The optimization logic of this combination is:

[0095] (1) Low temperature extrusion (180°C): Reduce surface fluctuations caused by excessive melting of the material and avoid the accumulation of thermal stress caused by high temperature.

[0096] (2) Medium filling density (52.5%): Balances material filling efficiency and internal porosity to avoid stress concentration at high density (100%) or loose structure at low density (5%).

[0097] (3) Low extrusion rate (95%): It inhibits accumulation or poor interlayer bonding caused by excessive material flow rate, especially in the bottom area, which can effectively reduce sagging defects.

[0098] (4) Low Z-axis height (-1.565mm): Enhances inter-layer bonding strength and reduces the step effect caused by layer height deviation.

[0099] (5) Moderate hot bed temperature (62.5°C): Ensure the balance between material adhesion and cooling rate to avoid warping caused by too high temperature or delamination caused by too low temperature.

[0100] Table 4 Optimal level based on the average value of weighted composite satisfaction function (WCDF)

[0101]

[0102]

[0103] 5. Surface Plot Analysis and 3D Printing Process Parameter Optimization

[0104] (1) With the process parameters as input, the surface arithmetic mean deviation (Sa) and root mean square deviation (Sq) values ​​as responses, the surface graph is obtained. The effects of different process parameters and the interactions between process parameters on the surface roughness are analyzed for the top, sides and bottom surfaces, such as Figure 7 shown.

[0105] (2) Using the response surface methodology, the optimal process parameter combinations for each side were obtained, as shown in Table 5. While current 3D printers struggle to adjust multiple side parameters in real time, the results of this study lay the foundation for the development of future intelligent systems. For example, through real-time sensor feedback, the temperature, extrusion rate, and Z-axis height can be dynamically adjusted to optimize overall surface quality. This approach will significantly improve the printing accuracy and consistency of complex structural parts.

[0106] Table 5 Optimal parameter values ​​for each side

[0107]

[0108] 6. To optimize the surface quality of 3D-printed parts, this paper proposes a point cloud data-driven, high-resolution quantitative surface topography assessment method, breaking through the spatial resolution limitations of traditional contact measurement. It also establishes a collaborative optimization model for multi-surface quality to address the one-sidedness of single-objective optimization. Finally, it designs an objective decision-making mechanism based on the entropy weight method to eliminate subjective bias in multi-objective weight allocation. This provides solid theoretical support for optimizing intelligent processes in additive manufacturing and offers valuable technical references for practical applications. It is expected to drive additive manufacturing technology toward greater efficiency and precision, helping related industries achieve a qualitative leap in product quality and production efficiency.

[0109] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-objective process optimization method for FDM-3D printing driven by the fusion of point cloud and response surface, characterized in that: The following steps are involved: S1. Build a full-process monitoring data acquisition platform to collect multimodal sensor data in real time during the 3D printing process and extract time and frequency domain features; S2. Obtain high-precision point cloud data of the printed sample based on 3D scanning technology and extract the roughness parameters of each surface area; S3. Based on the process parameters and multi-region roughness parameters, the regression equations of each surface are obtained and the Pareto chart of the standardized effects is drawn to analyze the main effects and interactions of the process parameters. S4, integrating the composite satisfaction function and the entropy weight method to dynamically assign objective weights to multi-region quality indicators; S5. Determine the optimal combination of process parameters through a multi-objective optimization method to achieve reduction of surface roughness in multiple regions; S6. Use response surface analysis to analyze the effect of process parameters on the surface roughness of each side of the 3D printed part and obtain the best optimized process parameters for each side.

2. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 1 is characterized in that: In step S1, the multimodal sensor data includes vibration and temperature data. The time domain and frequency domain characteristics of the nozzle and the hot bed are collected respectively by the vibration sensor and thermocouple sensor installed on the nozzle and the hot bed, and the frequency domain characteristics are extracted from the vibration sensor using the fast Fourier transform algorithm.

3. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 1 is characterized in that: In step S2, the roughness parameters include the surface arithmetic mean deviation Sa and the root mean square deviation Sq, and the Sa and Sq values ​​of the top, surrounding and bottom areas of the printed part are calculated by a non-contact measurement method.

4. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 3 is characterized in that: In steps S1-S2, data is collected through an FDM 3D printer, a vibration sensor, a thermocouple sensor, a 3D scanning device, and a terminal computer. The 3D scanning device is used to generate an STL file of the printed sample and compare the point cloud with the standard model. Finally, the surface arithmetic mean deviation Sa and root mean square deviation Sq roughness parameters of each surface of the 3D printed part are calculated.

5. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 3 is characterized in that: In step S3, a quadratic polynomial regression model is constructed based on the Box-Behnken experimental design with process parameters as input and Sa and Sq values ​​as responses. The process parameters include extruder temperature, filling density, extrusion rate, Z-axis height, and hot bed temperature. The quadratic polynomial regression model covers the Sa and Sq responses of each surface area. Regression equations were established for the arithmetic mean roughness Sa and root mean square roughness Sq respectively. A Pareto chart was drawn based on the effect values ​​of the interactions between the process parameters in the quadratic polynomial regression model to obtain the process parameters with a greater impact on surface roughness.

6. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 1, characterized in that: In step S4, the entropy weight method calculates the weight of each roughness parameter based on information entropy, and the composite satisfaction function is used to normalize the multi-objective response value, and finally generates the weighted comprehensive defect function WCDF as the optimization index.

7. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 1, characterized in that: In step S5, the optimal process parameter combination is: extruder temperature 180°C, filling density 52.5%, extrusion rate 95%, Z-axis height -1.565mm, and hot bed temperature 62.5°C.

8. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 5, characterized in that: The Box-Behnken experimental design includes three-level combinations of five process parameters, and a total of 46 groups of experiments are designed. The Pareto chart of the standardized effects in the quadratic polynomial regression model is used to identify the parameters that significantly affect the surface roughness.

9. The FDM-3D printing multi-objective process optimization method driven by point cloud and response surface fusion according to claim 1, characterized in that: The method also includes analyzing the influence trend of process parameters on the roughness of each surface area through response surface method, and generating a dynamic adjustment strategy to achieve real-time optimization.

Images