A point cloud-based process capability digital twin geometry model construction method

Abstract

This invention discloses a method for constructing a digital twin geometric model of process capability based on point clouds, comprising: acquiring three-dimensional point cloud data of the area covered by the target process and its corresponding single quality characteristic of an aerospace part product, thereby obtaining a three-dimensional point cloud dataset; obtaining a sequence of measured quality characteristic values ​​for digital twin modeling based on processing and measurement modes; acquiring a theoretical three-dimensional MBD model of the aerospace part product and performing lightweight processing to obtain a lightweight model; labeling the quality characteristic and its corresponding theoretical measurement points; fusing the sequence of measured quality characteristic values ​​with the lightweight model, and performing visualization differentiation and state backtracking on this basis to obtain a digital twin geometric model that integrates measured information; determining the process accuracy using the sequence of measured quality characteristic values ​​obtained under different modes; evaluating process capability according to the process accuracy levels, and providing adjustment suggestions based on the mean offset.

Description

Technical Field

[0001] This invention relates to the field of digital twin construction technology, specifically to a method for constructing a process capability digital twin geometric model based on point clouds. Background Technology

[0002] Digital twins refer to the process and methods of describing and modeling the characteristics, behaviors, formation processes, and performance of physical entities using digital technology; they are also known as digital twin technology. A digital twin is a virtual model that completely corresponds to and is identical to a physical entity in the real world, capable of simulating its own behavior and performance in a real-world environment in real time; it is also called a digital twin model. Digital twin technology not only utilizes existing human theories and knowledge to build virtual models but also leverages simulation techniques to explore and predict the unknown world. Therefore, digital twin technology provides new concepts and tools for innovation and development in the current manufacturing industry.

[0003] With the rapid popularization and application of new-generation information and communication technologies, the application of digital twins is gradually extending and developing from the product design and operation and maintenance stages to the manufacturing stage. In the area of ​​using digital twin technology to study manufacturing process capabilities and quality control, it is still in the exploratory stage, with relatively few research results and a lack of systematic approach. Especially in production models where small batches and multiple technical states are the norm, there is significant room for research and practical application in improving the consistency of product quality.

[0004] Existing process capability assessment methods require a product sample size of at least 20 for the manufacturing process being assessed. However, modern manufacturing enterprises primarily operate on a multi-variety, small-batch production model. When the product sample size is less than 20, the confidence level decreases, making it impossible to effectively assess the process capability. Furthermore, existing process capability assessment methods do not consider the continuity of process measurement or simultaneous measurement by multiple devices, leading to deviations in process capability assessment. In addition, on-site technicians must manually input measurement data into SPC analysis software for process capability analysis, resulting in poor data accuracy, low analysis efficiency, and unintuitive presentation that is difficult to understand and interpret. Summary of the Invention

[0005] The purpose of this invention is to provide a method for constructing a digital twin geometric model of process capability based on point cloud, which solves the problems of poor data accuracy and low efficiency in the existing technology.

[0006] To achieve the above objectives, the present invention employs the following technical solution: A method for constructing a digital twin geometric model of process capability based on point clouds, comprising: For the target process of aerospace parts and its corresponding single quality characteristic, obtain the three-dimensional point cloud data of the area covered by the quality characteristic to obtain a three-dimensional point cloud dataset. Based on the processing and measurement mode, the three-dimensional point cloud dataset is processed and calculated to obtain a sequence of measured quality characteristics for digital twin modeling. The theoretical 3D MBD model of the aerospace parts is obtained and then subjected to lightweight processing to obtain a lightweight model; the mass characteristics and their corresponding theoretical measurement points are marked on the lightweight model. The measured value sequences of quality characteristics corresponding to different modes are fused with the lightweight model. Based on this, visualization and state backtracking are performed to obtain a digital twin geometric model that incorporates measured information. The process accuracy is determined by using the sequence of measured quality characteristics obtained under different modes; the process capability is evaluated by level based on the process accuracy, and adjustment suggestions are given based on the mean offset.

[0007] Furthermore, based on the actual situation of the target process, determine which of the following modes it belongs to: Mode 1: The target process only processes one aerospace part, and this quality characteristic is measured only once; Mode 2: The target process batch-processes multiple aerospace parts and measures the same quality characteristic of each aerospace part. If it is determined to be Mode 1, calculate the key quality characteristic dimensions for Mode 1; extract two theoretical measurement points from the 3D model of the aerospace part product to represent the key quality characteristic dimension values; read the 3D point cloud data corresponding to the theoretical measurement points from the 3D point cloud dataset and record them as actual measurement points; calculate the Euclidean distance of the actual measurement points in 3D space as the actual quality characteristic measured value, thereby obtaining the sequence of actual quality characteristic measured values ​​for Mode 1.

[0008] Furthermore, if the judgment is pattern two, let's assume that... Individual aviation parts products The first quality characteristic measurement yields the corresponding measured quality characteristic values, forming the original measurement sequence; When the same quality characteristic of the same batch of aerospace parts is continuously measured using the same measuring equipment, the arithmetic mean method is used for filtering; the original measurement sequence is arithmetically averaged to obtain the filtered measured value of the quality characteristic. When different measuring devices are used to measure the same quality characteristic of the same batch of aerospace parts, or when the same measuring device is used to measure the same quality characteristic at different time periods, the Kalman filter method is used to fuse the measured data; let the original measurement sequence be... According to the measurement time sequence Perform recursive filtering.

[0009] Furthermore, the recursive filtering process is as follows: For the first calculation Set initial estimate ; for Take the current measurement value Calculate the number of digits using the following formula. Step optimal estimate : ; in, This is the optimal estimate from the previous round, i.e. ; Kalman gain; Calculate the Kalman gain: ; in, The variance of the previous estimate. The variance of the current measurement; Calculate the variance of the estimated values ​​for this round. As the next round ; The optimal estimates obtained from all iterations By associating the data with the corresponding aerospace parts, a filtered sequence of measured quality characteristics is obtained. ,in The total number of aerospace parts products, For the first Measured values ​​of filtered quality characteristics for each aerospace component.

[0010] Furthermore, for each measured value of a quality characteristic in the sequence of measured quality characteristic values ​​for either Mode 1 or Mode 2... Find its corresponding theoretical value in the lightweight model. Calculate measurement deviation ; Based on the lightweight model of aerospace parts products, and The numerical relationships are rendered using different shapes and colors, and based on the deviations... The size of the model is mapped to different colors to obtain a digital twin geometric model; simultaneously, the measured value sequence of quality characteristics is retrieved according to arbitrary measurement time sequences, and the corresponding values ​​are loaded based on the retrieval conditions. The model is then re-rendered to allow for the rewinding of the digital twin model's state.

[0011] Furthermore, the consistency between the actual average value of the process and the center value of the quality characteristic specification is calculated to obtain the process accuracy Ca, and the calculation formula is: ; in It is the process average, that is, the arithmetic mean of the series of measured quality characteristics; This is the center value of the specification; For specification tolerances; Based on the absolute value of Ca, process capability is evaluated according to the following levels: Grade A: Grade B: Grade C: Grade D: .

[0012] Furthermore, calculate the process average value. With specification center value mean offset Based on the magnitude of the mean deviation, suggestions for process adjustment are proposed: when If the mean deviation of the target process is within the allowable range, then the comprehensive evaluation of the target process capability is carried out in conjunction with the process capability index Cpk value. when When this occurs, it indicates an abnormal deviation in the process mean, requiring adjustment of the process conditions to achieve the desired result. The value should be as close as possible to the tolerance center value.

[0013] A terminal device includes a processor, a memory, and a computer program stored in the memory; when the processor executes the computer program, it implements the point cloud-based process capability digital twin geometric model construction method.

[0014] A computer-readable storage medium storing a computer program; when executed by a processor, the computer program implements the point cloud-based process capability digital twin geometric model construction method.

[0015] Compared with the prior art, the present invention has the following technical features: This invention, based on the spatiotemporal characteristics of process measurement, employs arithmetic averaging and Kalman filtering to fuse measurement data, improving the accuracy of digital twin modeling of process capabilities. The filtered measured data is then fused with a lightweight 3D model, differentiated by shape and color, supporting process status backtracking based on arbitrary measurement sequences. Using the filtered measured values, the consistency between the actual process average and the quality characteristic specification center value is calculated, and the process accuracy is evaluated, making it suitable for process capability analysis in small-batch production. The offset between the measured average and the specification center is calculated and visualized using shape and color. Based on the magnitude of the offset, process adjustment suggestions can be proposed, and process anomaly diagnosis can be performed in conjunction with quality control charts. Attached Figure Description

[0016] Figure 1This is a flowchart illustrating an embodiment of the present invention. Detailed Implementation

[0017] This invention provides a method for constructing a digital twin geometric model of process capability based on point clouds. Taking a single quality characteristic as the object, a digital twin is established based on a lightweight model. This method can achieve a realistic description of the process capability of manufacturing using relatively few parameters and variables, and features high accuracy, rapid modeling, and fast model running speed. The steps of this invention are as follows: Step 1: For the target process and its corresponding single quality characteristic of the aerospace part, acquire the three-dimensional point cloud data of the area covered by the quality characteristic using measurement equipment such as 3D scanning or coordinate measuring machine, and obtain a three-dimensional point cloud dataset; the three-dimensional point cloud dataset contains the three-dimensional coordinate information of the measurement points. ,in The number of three-dimensional points, Indicates the first A three-dimensional point.

[0018] The quality characteristics refer to the geometric dimensional features of aerospace parts that are selected as key control objects.

[0019] Step 2: Based on the processing and measurement mode, the 3D point cloud dataset is processed and calculated to obtain a sequence of measured quality characteristics for digital twin modeling.

[0020] Step 2.1, determine the processing and measurement mode.

[0021] Based on the actual situation of the target process, determine which of the following modes it belongs to: Mode 1: The target process only processes one aerospace part, and this quality characteristic is measured only once.

[0022] Mode 2: Batch processing of multiple aerospace parts in the target process ( Item, ), and measure the same quality characteristics for each aerospace component.

[0023] Step 2.2: If it is determined to be Mode 1, calculate the key quality characteristic dimensions for Mode 1.

[0024] Step 2.2.1: Obtain theoretical measurement points and corresponding 3D cloud data.

[0025] Two theoretical measurement points were extracted from the 3D model of the aerospace component to characterize the dimensional values ​​of key quality characteristics. and Through an IoT platform, data is read from a 3D point cloud dataset and compared with theoretical measurement points. and The corresponding 3D point cloud data is denoted as the measured points. and .

[0026] Step 2.2.2, Calculate the measured points and Euclidean distance in three-dimensional space The sum of the squares of the distance differences in the three directions is used as the measured value of the quality characteristic.

[0027] For example, a certain target process produces only one aerospace part. A certain quality characteristic of this part was measured, and the data is as follows:

[0028] Based on the above data, the Euclidean distance is calculated as follows: .

[0029] Step 2.2.3: Form a sequence of measured quality characteristics.

[0030] The calculated single measured value As a sequence of measured quality characteristics under the model And output it to step 4.

[0031] Step 2.3: Data filtering processing for Mode 2 (batch, multiple measurements).

[0032] If the determination is Mode 2, then consider the spatiotemporal characteristics of the quality characteristic measurement process, and assume that for... Individual aviation parts products For each aerospace component, the corresponding measured quality characteristic value is obtained by following the method in step 2.2.2; these measured quality characteristic values ​​together constitute an original measurement sequence. ,in Indicates the first The measured quality characteristics obtained from the measurement; then, different filtering methods are selected according to the different measurement methods: Step 2.3.1, filtering based on the arithmetic mean method.

[0033] When the same quality characteristic of the same batch of aerospace parts is continuously measured using the same measuring equipment, the arithmetic mean method is used for filtering; the original measurement sequence is arithmetically averaged to obtain the filtered measured value of the quality characteristic. The calculation formula is as follows: .

[0034] For example, 24 units of a certain bend in an aerospace product were produced. A quality characteristic of this bend was "ensuring the bushing end face protrudes from the nut by no more than 5mm". Using a digital measuring device, 24 consecutive measurements were performed, and the measurement data are as follows: 2.5; 2.4; 2.5; 3.1; 2.0; 1.9; 2.2; 2.2; 2.5; 2.3; 2.4; 2.2; 2.5; 2.6; 3.0; 2.2; 2.2; 2.3; 2.3; 2.4; 2; 2.6; 3.0; 3; The measured values ​​of quality characteristics were calculated using the arithmetic mean method. .

[0035] Step 2.3.2, data fusion processing based on Kalman filtering.

[0036] When different measuring devices are used to measure the same quality characteristic of the same batch of aerospace parts, or when the same measuring device is used to measure the same quality characteristic at different time periods, the Kalman filter method is used to fuse the measured data; let the original measurement sequence be... According to the measurement time sequence Perform recursive filtering: Step 2.3.2.1, Initialization.

[0037] For the first calculation Set initial estimate Initial estimate of variance Based on prior knowledge or historical data.

[0038] Step 2.3.2.2: Iteratively calculate the optimal estimate.

[0039] for Take the current measurement value Calculate the number of digits using the following formula. Step optimal estimate : ; in, This is the optimal estimate from the previous round (the measurement from another measuring device or another measurement from the same measuring device), i.e. ; This is the Kalman gain, with a value between 0 and 1.

[0040] Step 2.3.2.3: Calculate the Kalman gain.

[0041] ; in, The variance of the previous estimate. The variance of the current measurement value (set according to the accuracy of the measuring equipment).

[0042] Step 2.3.2.4: Update the variance for the next iteration.

[0043] Calculate the variance of the estimated values ​​for this round. As the next round .

[0044] Step 2.3.2.5: Form the filtered sequence of measured quality characteristics.

[0045] The optimal estimates obtained from all iterations By associating the data with the corresponding aerospace parts, a filtered sequence of measured quality characteristics is obtained. ,in The total number of aerospace parts products, For the first The filtered measured quality characteristics of each aerospace part are then output to step 4.

[0046] For example, for a certain quality characteristic, the mean of two measurement data is , The standard deviation is , If the measured data X follows a normal distribution, it can be denoted as: X ~ N(u, Let the data from the first measurement be... 1~N(49,2) 2 ), second measurement data 2~N(51,3) 2 Then the Kalman gain = = =0.692.

[0047] Then calculate the optimal estimate: = + ( - )=51+0.692*(49-51)=49.6 mm.

[0048] Step 3: Obtain the theoretical 3D MBD model of the aerospace parts and perform lightweight processing on it (such as feature simplification, mesh reduction, geometric compression, etc.) to obtain a lightweight model; mark the mass characteristics and their corresponding theoretical measurement points on the model to provide a benchmark for subsequent data fusion.

[0049] Step 4, sequence the measured quality characteristics corresponding to different modes (i.e., mode 1 and mode 2). or By integrating the data with a lightweight model, and then performing visualization differentiation and state backtracking, a digital twin geometric model that incorporates measured information is obtained.

[0050] Step 4.1: Calculate the measurement deviation of the quality characteristics.

[0051] For the measured value sequence of quality characteristics or For each measured value of a quality characteristic, find its corresponding theoretical value in the lightweight model. Calculate measurement deviation : ; in express or The first in Measured values ​​of each quality characteristic; For the first The theoretical value of each quality characteristic; For the first Measurement deviation of a quality characteristic.

[0052] For example: Suppose a certain quality characteristic dimensional tolerance of an aviation product is... mm, the theoretical value is 50.0 mm, and the optimal estimated measured value of this quality characteristic is obtained through filtering in step 2. Using the optimal estimated measurement value, the measurement deviation of the quality characteristic is calculated. = - = 49.6-50.0= -0.4 mm.

[0053] Step 4.2: Perform visual differentiation and state backtracking.

[0054] Based on the lightweight model of aerospace parts products, and The numerical relationships are rendered using different shapes and colors. For example, the theoretical values ​​of quality characteristics are represented by semi-transparent gray, while the measured values ​​are represented by solid colors, and the results are adjusted according to deviations. The magnitude of the deviation is mapped to different colors (e.g., green indicates a small deviation, red indicates a large deviation), thus obtaining a digital twin geometric model; simultaneously, the measured value sequence of quality characteristics is retrieved according to arbitrary measurement time sequence (i.e., different production batches or time points), and the corresponding values ​​are loaded according to the retrieval conditions. The model is then re-rendered to allow for the rewinding of the digital twin model's state.

[0055] In this scheme, the optimal estimated measurement value calculated by filtering varies depending on the number of quality characteristic measurement points received by the process capability digital twin model. The state backtracking of the digital twin model can be performed according to any measurement time sequence.

[0056] Step 5: Utilize the sequence of measured quality characteristics obtained under different modes. or Determine the process accuracy; evaluate process capability by grade based on process accuracy, and provide adjustment suggestions based on mean offset.

[0057] Step 5.1, calculate the sequential process accuracy (Ca).

[0058] The process accuracy Ca is obtained by calculating the degree of consistency between the actual average value of the process and the center value of the quality characteristic specification. The calculation formula is as follows: ; in This refers to the process average value, i.e., the sequence of measured quality characteristics. or The arithmetic mean; This refers to the specification center value (i.e., the theoretical value of the quality characteristic). ); Specification tolerances are determined by the upper specification limit (USL) and the lower specification limit (LSL), i.e. .

[0059] Step 5.2: Evaluate process capability by grade.

[0060] Based on the absolute value of Ca, process capability is evaluated according to the following levels: Grade A: Grade B: Grade C: Grade D: .

[0061] The above-mentioned levels provide users with an intuitive basis for judging the status of the target process. For example, Level A: The process accuracy is high, and the average value is highly consistent with the specification center; Level B: The process accuracy is acceptable, but there is a slight deviation; Level C: The process accuracy is insufficient, and the deviation is obvious; Level D: The process accuracy is seriously insufficient, and immediate intervention is required.

[0062] For example, the process accuracy can be calculated using the filtered measured values: ; Then |Ca|=26.7%. Referring to the process capability digital twin model evaluation mode, the accuracy level of the target process is "C".

[0063] Step 5.3: Calculate the mean offset and propose adjustment suggestions.

[0064] Calculate process average With specification center value mean offset Based on the magnitude of the mean deviation, suggestions for process adjustment are proposed: when If the mean deviation of the target process is within the allowable range, then the comprehensive evaluation of the target process capability is carried out in conjunction with the process capability index Cpk value. For example, a comparison table of different levels, different Cpk value ranges and comprehensive evaluation conclusions can be set in advance, and the corresponding comprehensive evaluation conclusion can be determined according to the level and the specific value of Cpk.

[0065] when When this occurs, it indicates an abnormal deviation in the process mean, requiring adjustment of the process conditions to achieve the desired result. The value should be as close as possible to the tolerance center value.

[0066] in, The process standard deviation is calculated from the sequence of measured quality characteristics.

[0067] For example, calculating the offset between the measured mean and the specification center. = |49.6-50.0| = 0.4, the best estimate of the standard deviation = + ( - )=3+0.692*(2-3)=2.308 mm; Since 1.5σ = 1.5 * 2.308 = 3.462 > 0.4, the deviation of the process mean is within the allowable range. The process capability evaluation needs to be carried out in conjunction with the calculation of the process capability index Cpk.

[0068] This invention establishes a digital twin of process capability based on a lightweight 3D model of aerospace products. It features fast modeling speed, high model running speed, high accuracy of process capability analysis, and visualization of process capability, thereby improving the timeliness of decision-making in online process capability simulation. If the product manufacturing is in a small-batch production mode, and the process capability index Cpk cannot be calculated due to insufficient measurement sample size, the process capability can be evaluated by combining the digital twin with process mean offset and process accuracy. Anomalies in the process can also be detected by combining model colors. If the product manufacturing is in a large-batch production mode, the process capability can be evaluated by combining the digital twin with process mean offset and process capability index Cpk. This modeling method can be widely applied to process capability analysis scenarios in the manufacturing process of parts based on digital twins.

[0069] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for constructing a digital twin geometric model of process capability based on point clouds, characterized in that, include: For the target process of aerospace parts and its corresponding single quality characteristic, obtain the three-dimensional point cloud data of the area covered by the quality characteristic to obtain a three-dimensional point cloud dataset. Based on the processing and measurement mode, the 3D point cloud dataset is processed and calculated to obtain a sequence of measured quality characteristics for digital twin modeling. The theoretical 3D MBD model of the aerospace parts is obtained and then subjected to lightweight processing to obtain a lightweight model; the mass characteristics and their corresponding theoretical measurement points are marked on the lightweight model. The measured value sequences of quality characteristics corresponding to different modes are fused with the lightweight model. Based on this, visualization and state backtracking are performed to obtain a digital twin geometric model that incorporates measured information. The process accuracy is determined by using the sequence of measured quality characteristics obtained under different modes; the process capability is evaluated by level based on the process accuracy, and adjustment suggestions are given based on the mean offset.

2. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 1, characterized in that, Based on the actual situation of the target process, determine which of the following modes it belongs to: Mode 1: The target process only processes one aerospace part, and this quality characteristic is measured only once; Mode 2: The target process batch-processes multiple aerospace parts and measures the same quality characteristic of each aerospace part. If it is determined to be Mode 1, calculate the key quality characteristic dimensions for Mode 1; extract two theoretical measurement points from the 3D model of the aerospace part product to characterize the key quality characteristic dimension values; Read the 3D point cloud data corresponding to the theoretical measurement points from the 3D point cloud dataset and record them as the measured points; The Euclidean distance of the measured point in three-dimensional space is calculated as the measured value of the quality characteristic, thereby obtaining the sequence of measured values ​​of the quality characteristic under mode one.

3. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 2, characterized in that, If the judgment is pattern two, let's assume... Individual aviation parts products The first quality characteristic measurement yields the corresponding measured quality characteristic values, forming the original measurement sequence; When the same quality characteristic of the same batch of aerospace parts is measured continuously using the same measuring equipment, the arithmetic mean method is used for filtering. The original measurement sequence is arithmetically averaged to obtain the filtered measured values ​​of the quality characteristics; When different measuring devices are used to measure the same quality characteristic of the same batch of aerospace parts, or when the same measuring device is used to measure the same quality characteristic at different time periods, the Kalman filter method is used to fuse the measured data; let the original measurement sequence be... According to the measurement time sequence Perform recursive filtering.

4. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 3, characterized in that, The recursive filtering process is as follows: For the first calculation Set initial estimate ; for Take the current measurement value Calculate the number of digits using the following formula. Step optimal estimate : ； in, This is the optimal estimate from the previous round, i.e. ; Kalman gain; Calculate the Kalman gain: ； in, The variance of the previous estimate. The variance of the current measurement; Calculate the variance of the estimated values ​​for this round. As the next round ; The optimal estimates obtained from all iterations By associating the data with the corresponding aerospace parts, a filtered sequence of measured quality characteristics is obtained. ,in The total number of aerospace parts products, For the first Measured values ​​of filtered quality characteristics for each aerospace component.

5. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 1, characterized in that, For each measured value of a quality characteristic in the sequence of measured quality characteristic values ​​for either Mode 1 or Mode 2. Find its corresponding theoretical value in the lightweight model. Calculate measurement deviation ; Based on the lightweight model of aerospace parts products, and The numerical relationships are rendered using different shapes and colors, and based on the deviations... The size of the model is mapped to different colors to obtain a digital twin geometric model; simultaneously, the measured value sequence of quality characteristics is retrieved according to arbitrary measurement time sequences, and the corresponding values ​​are loaded based on the retrieval conditions. The model is then re-rendered to allow for the rewinding of the digital twin model's state.

6. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 1, characterized in that, The process accuracy Ca is obtained by calculating the degree of consistency between the actual average value of the process and the center value of the quality characteristic specification. The calculation formula is as follows: ； in It is the process average, that is, the arithmetic mean of the series of measured quality characteristics; This is the center value of the specification; For specification tolerances; Based on the absolute value of Ca, process capability is evaluated according to the following levels: Grade A: Grade B: Grade C: Grade D: .

7. The method for constructing a digital twin geometric model of process capability based on point cloud as described in claim 1, characterized in that, Calculate process average With specification center value mean offset Based on the magnitude of the mean deviation, suggestions for process adjustment are proposed: when If the mean deviation of the target process is within the allowable range, then the comprehensive evaluation of the target process capability is carried out in conjunction with the process capability index Cpk value. when When this occurs, it indicates an abnormal deviation in the process mean, requiring adjustment of the process conditions to achieve the desired result. The value should be as close as possible to the tolerance center value.

8. A terminal device, comprising a processor, a memory, and a computer program stored in the memory; characterized in that, When the processor executes a computer program, it implements the point cloud-based digital twin geometric model construction method for process capability as described in any one of claims 1-7.

9. A computer-readable storage medium storing a computer program; characterized in that, When the computer program is executed by a processor, it implements the point cloud-based digital twin geometric model construction method for process capability as described in any one of claims 1-7.

Images