Integral box bottom process part model construction method based on process data

Through the method based on process data, the bottom process part model of the rocket storage tank is automatically constructed, which solves the deformation and quality control problems during the processing of the rocket storage tank, improves the development efficiency and quality, and realizes the automation and intelligent construction of the process part model.

CN119047146APending Publication Date: 2024-11-29SHANGHAI SPACE PRECISION MACHINERY RES INST
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411032603.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The entire bottom parts of the rocket storage tank have problems such as deformation, overcut or undercut during the processing process, resulting in increased difficulty in CNC processing technology planning and quality control, and the degree of automation and intelligence of existing process component models is low, which affects the development efficiency.

Method used

The integrated box bottom process component model construction method is adopted based on process data. The contour height and wall thickness data are collected through the data measurement equipment, and data processing and screening are generated to generate a cloud diagram of wall thickness error changes of the process component model. The process model is automatically constructed using an algorithm. Replace the manual interaction construction model to improve compilation efficiency and quality.

Benefits of technology

The automation and intelligent construction of process part models has been realized, the processing process preparation cycle has been shortened, the parts development efficiency has been improved, the processing quality has been ensured, and the quality problems caused by data errors have been avoided.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119047146A_ABST
    Figure CN119047146A_ABST
Patent Text Reader

Abstract

The invention relates to an integral box bottom process part model construction method based on process data, and belongs to the technical field of numerical control machining process design. The method comprises the following steps: firstly, acquiring contour height and wall thickness data of a box bottom structural member by using data measurement equipment; carrying out data processing on the collected two types of data; according to the processed data, a process part model is constructed in the forward direction; and finally, generating a wall thickness error change cloud picture of the workpiece model. According to the method, the problem of low automation and intelligence degree of process model construction is effectively solved, the preparation period of the box bottom machining process is shortened, and the development efficiency of key and important parts is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of numerical control machining process design, and relates to a method for constructing an integral box bottom process part model based on process data. Background Art

[0002] Complex structural parts of aerospace models are characterized by large size, complex appearance, large number of processing features, and high manufacturing precision requirements. With the improvement of overall performance requirements, the design and manufacturing of structural parts show a trend of integration and thin-walled development, which makes the parts very easy to deform, overcut or undercut during the processing, and puts higher requirements on the CNC machining process planning of parts and quality control during the processing.

[0003] Rocket tanks are the main load-bearing components of launch vehicles. They are large-sized thin-walled structural parts formed from high-specific-strength aluminum alloy sheets, accounting for 80% to 90% of the total weight of the rocket. The tank bottom is located at both ends of the rocket fuel tank and is a key component of the tank. The integral tank bottom is a typical large-scale thin-walled ellipsoidal structure with large geometric dimensions (diameter Φ3.35m), complex shape (local profile error after forming is up to 3mm), thin wall (blank wall thickness is about 12mm, remaining wall thickness is 1.5 to 4mm), high material removal ratio (>70%), long processing cycle, etc., poor processing technology, and difficulty in ensuring the consistency of wall thickness.

[0004] As the output of the three-dimensional process design link and the basis for later processing and manufacturing, the process part model can intuitively reflect the process information such as the geometric shape and processing quality of the parts during the processing, and can be used as a process information carrier to drive the optimization of machining process design. In addition, by constructing a dynamically evolving complete process part model sequence, the process design intent can be accurately reproduced, which makes it easier for process personnel to clarify the subsequent process planning goals and optimize process parameters, and drive the process route and CNC program compilation and optimization process. Therefore, the accurate and efficient construction of the integral bottom of the box process part model is of great significance to improving the quality and efficiency of the three-dimensional process design of the bottom of the box structure.

[0005] At present, the manufacturing method of "on-machine measurement-process part model construction-CNC machining" is usually used for the integral bottom parts of rocket tanks, and good results have been achieved. However, the construction of the process part model of the integral bottom of the tank is low in automation and intelligence, a large amount of on-machine measurement data needs to be processed manually, and the manual interactive construction of the process part surface is cumbersome, resulting in a long CNC machining programming cycle, which seriously affects the development efficiency of key and important parts. Summary of the invention

[0006] The technical problem solved by the present invention is to overcome the deficiencies of the prior art and propose a method for constructing an integral box bottom process part model based on process data.

[0007] The solution of the present invention is:

[0008] A method for constructing an integral box bottom process part model based on process data, comprising:

[0009] Use data measuring equipment to collect profile height and wall thickness data of the bottom structural parts of the box;

[0010] Perform data processing on the two types of collected data;

[0011] Based on the processed data, the process part model is constructed forward;

[0012] Generate a cloud diagram of wall thickness error changes for the process part model.

[0013] Preferably, the data measuring device collects data by point measurement, and the measurement points are a series of sequential points generated by software system planning.

[0014] Preferably, the two types of collected data are processed, including:

[0015] Perform data cleaning on the two types of collected data to remove duplicate and meaningless null values;

[0016] Aggregate the cleaned data into data groups according to the collection rules;

[0017] The profile height and wall thickness in each set of data are screened based on the normal distribution according to the "3σ principle", the true values ​​of the profile height and wall thickness corresponding to the measuring points are screened out, and the data are structured.

[0018] Preferably, the cleaned data is aggregated into N data groups according to the collection rules, and the collection rules are the following four conditions:

[0019] The range of blank profile height values ​​measured under condition 1 is [h min ,h max ];

[0020] The process wall thickness value measured under condition 2 is [t-δ-θ / 2, t+δ+θ / 2];

[0021] Condition 3: The measurement value in each data group corresponds to a time ≤ T seconds;

[0022] Condition 4: The true value of the measurement is continuous;

[0023] Among them, h min ,h max are the minimum and maximum values ​​of the blank contour height respectively; t is the theoretical wall thickness value of the process; δ is the maximum deformation of the processing; θ is the accuracy threshold, which is determined by the measurement accuracy of the data measurement equipment; the range of T is [3,10].

[0024] Preferably, the profile height and wall thickness in each set of data are screened based on the normal distribution according to the "3σ principle" to screen out the true values ​​of the profile height and wall thickness corresponding to the measuring point. The specific implementation method is as follows:

[0025] Calculate the mean μ and standard deviation σ of the contour height in each set of data;

[0026] Identify and remove outliers, which are defined as measurements less than μ-3σ or greater than μ+3σ;

[0027] Re-solve the mean value μ′ of the profile height in each set of data, and use μ′ as the true value of the profile height corresponding to the measurement point;

[0028] Calculate the mean μ1 and standard deviation σ1 of the wall thickness in each set of data;

[0029] Identify and remove outliers, which are defined as measurements less than μ1-3σ1 or greater than μ1+3σ1;

[0030] The mean value μ1′ of the wall thickness in each set of data is re-solved, and μ1′ is used as the true value of the wall thickness corresponding to the measuring point.

[0031] Preferably, the data is structured, and the specific implementation method is as follows:

[0032] The process data of each measuring point is organized into a list structure, and the list structure is defined as P1,…,P i ,…,P N , where i = 1, 2, ..., N, the i-th data P i Represented as P i (x i ,y i ,z i ,a i ,b i ,c i ,h i ,t i ), where (x i ,y i ,z i ) is the coordinate of the i-th measurement point, (a i ,b i ,c i ) is the normal vector at the i-th measurement point, h i is the profile height value at the i-th measuring point, t i is the wall thickness value at the i-th measuring point.

[0033] Preferably, the process of constructing the process part model forward is as follows:

[0034] c1) constructing the inner profile of the blank according to the contour height and wall thickness data obtained for the first time, and generating the outer profile of the blank by offsetting the inner profile to the outer edge of the normal wall thickness;

[0035] c2) Based on the inner surface of the blank model, the current intermediate process part model is generated by the current process wall thickness and its accumulated wall thickness deviation data, and the boss features are generated according to the process requirements;

[0036] c3) The wall thickness measurement data is used as the final inspection data, and the resulting three-dimensional part process piece is a digital mirror image of the actual processed part, which is used in the subsequent digital manufacturing inspection link.

[0037] Preferably, the implementation method of step c2) is as follows:

[0038] (1) Calculate the cumulative wall thickness deviation. The cumulative wall thickness deviation of the intermediate process parts is solved by accumulating the wall thickness deviation data of the processed processes;

[0039] (2) According to the accumulated wall thickness deviation value, the interpolation points of the inner profile of the blank are adjusted, and the inner profile of the current process is generated by interpolation of points;

[0040] (3) The interpolation points of the inner surface are offset outward along the normal direction by the actual measured wall thickness distance of the current process to generate the interpolation points of the outer surface, and the outer surface of the current process is interpolated in a point-by-point manner;

[0041] (4) Create a boss feature. If the machining area of ​​the process contains a boss, first use the machining feature recognition method to extract the boss area boundary of the original model, then use the feature boundary to crop the current process surface, use the cropping relationship tree to automatically determine the boss area to be offset, and finally offset the boss area outward by the boss height to generate the process surface with the boss area.

[0042] Preferably, a wall thickness error variation cloud map of the process part model is generated, and the implementation method is as follows:

[0043] Construct a geometric model, superimpose the inner surface of the process part model with the inner surface of the theoretical model, use a set of parallel and equal-height theoretical inner surfaces to segment and intercept the inner surface of the process part model, form contour lines of wall thickness error variation, set the contour lines to different colors, and generate a cloud map of the wall thickness error variation of the process part model.

[0044] Preferably, the method for constructing the geometric model is the same as the method for constructing the process part model.

[0045] The beneficial effects of the present invention compared with the prior art are:

[0046] (1) The present invention proposes a universal and simplified data collection method. By using a point-measurement data measurement method, there is no need for machine tool communication. Only through data processing algorithms, data can be quickly cleaned, grouped, and screened, thereby improving technical adaptability and enabling rapid deployment of general machine tools.

[0047] (2) The present invention ingeniously uses data error accumulation to characterize the processing status between processes, unifies the inner surface of the blank model as a benchmark, and derives the full process model through forward evolution, thereby simplifying data processing.

[0048] (3) The present invention proposes a method for constructing an integral box bottom process part model based on process data, which replaces the manual interactive construction mode. The outer surface boss structure uses an algorithm to automatically determine the cutting relationship, avoid process model construction errors, and improve program compilation efficiency and quality.

[0049] (4) The present invention proposes a technology for quickly generating wall thickness cloud maps online based on process part models and contour lines, which allows interactive viewing of wall thickness changes in a CAD / CAM three-dimensional environment, facilitates manual verification of CNC machining processes and tool paths, and prevents quality problems caused by data errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Schematic diagram of the process flow of building an integrated bottom process parts model based on process data;

[0051] Figure 2 Process data acquisition hardware connection diagram;

[0052] Figure 3 Schematic diagram of the point measurement process;

[0053] Figure 4 Schematic diagram of blank model construction;

[0054] Figure 5 Schematic diagram of boss cutting relationship;

[0055] Figure 6 Schematic diagram of wall thickness error variation contour generation;

[0056] Figure 7 Schematic diagram of the test parts at the bottom of the box;

[0057] Figure 8 Schematic diagram of the process part model of the box bottom, where (a) is the process part model without bosses, and (b) is the process part model with bosses;

[0058] Fig. 9 Schematic diagram of 3D wall thickness contour cloud map. DETAILED DESCRIPTION

[0059] The present invention will be further described below in conjunction with the accompanying drawings.

[0060] By referring to the exemplary embodiments, the objects and functions of the present invention and the methods for achieving these objects and functions will be explained. However, the present invention is not limited to the exemplary embodiments disclosed below; it can be implemented in different forms. The essence of the description is only to help those skilled in the relevant art to comprehensively understand the specific details of the present invention.

[0061] The present invention is oriented to measurement-machining integration, carries out automatic modeling of CNC machining process part models driven by models and data, and provides automated and intelligent support for CNC machining process design of structural parts.

[0062] The present invention can be implemented according to the following steps:

[0063] a) Process data acquisition.

[0064] Data collection is carried out for large box bottom structural parts. The collection system is connected to the host computer by the profile height and wall thickness data measuring equipment. The measurement data is transmitted and stored in the on-site workstation. The data collection includes two types of data: the box bottom profile height data and the wall thickness data. The measurement method is point-by-point measurement (point measurement), and the measurement points are a series of sequential points planned and generated by the software system.

[0065] On-machine measuring equipment, including but not limited to profile height measuring instruments, wall thickness measuring instruments, controllers, host computers, workstations, and connecting network cables, hardware connections such as Figure 2 As shown, the profile height measuring instrument and the wall thickness measuring instrument are connected to the machine tool spindle through the tool holder. The machine tool drives the measuring instrument probe to move to achieve point measurement. The measuring instrument probe is connected to the controller to transmit data to the host computer or host computer software, and the collected data is stored in the workstation.

[0066] During the on-machine measurement data collection process, the profile height measuring instrument and the wall thickness measuring instrument obtain the box bottom profile height data and wall thickness data respectively. The real-time data measured by the two are transmitted to the host computer through the network cable. The host computer records the real-time data measured at a fixed frequency, and the data is stored in the workstation in the form of a file for subsequent data processing. Among them, the profile height measuring instrument can use an electronic clock-type dial indicator, laser scanner, etc., and use the theoretical inner surface of the box bottom as the origin to perform height measurement. The wall thickness measuring instrument can use an ultrasonic non-contact or contact probe to perform wall thickness measurement on the outer surface of the box bottom.

[0067] b) Process data processing.

[0068] First, the collected contour height and wall thickness data are processed and data cleaning is performed to remove repeated and meaningless null values. Then, data grouping is performed to aggregate the data into data groups according to the collection rules. Finally, data screening is performed, and each group of data is processed according to the "3σ (sigma) principle". The true value corresponding to the measurement point is screened out based on the normal distribution, and the data is structured.

[0069] The measurement path adopts point-to-point measurement, and the measurement is carried out point by point on the bottom surface of the box. The measurement points are orderly. When measuring point by point, the measurement is maintained at the point for a period of time. Figure 3 During the point measurement process, segment ① is the transfer segment, the measured value is empty, and it is generally displayed as the default value; segment ② is the approach segment, the measured value changes from false to true, showing the process of numerical value generation; segment ③ is the measurement segment, the probe maintains measurement at the measurement point for a period of time T to obtain a relatively stable true value; segment ④ is the lift-off segment, the measured value changes from true to false, showing the process of numerical value disappearance; segment ⑤ is the transition segment, the measured value is empty, and it is generally displayed as the default value; segment ⑥ is the same as segment ②. From the above measurement process, it can be seen that the measured true value can be maintained for T seconds, and the measurement data has regularity. Therefore, data cleaning, data grouping, and data screening steps can be taken to obtain the measured true values ​​of N measurement points.

[0070] Data cleaning, for the transfer and transition sections of the measurement process, removes redundant data of the measurement default value and retains data with true value, which can also relieve the calculation pressure of subsequent data processing. According to general data characteristics, the measurement default value is 0 and the measurement true value is greater than 0.

[0071] Data grouping: group the cleaned data through a heuristic algorithm or a loop iteration algorithm to obtain N groups of data that meet rule 1.

[0072] Rule 1: There must be N groups to separate all the measurement data. The data in each group must satisfy:

[0073] The range of blank profile height values ​​measured under condition 1 is [h min ,h max ];

[0074] The process wall thickness value measured under condition 2 is [t-δ-θ / 2, t+δ+θ / 2];

[0075] Condition 3: The corresponding time of the measurement value in each group is ≤T seconds;

[0076] Condition 4: The true value of the measurement is continuous;

[0077] Among them, h min ,h maxare the minimum and maximum values ​​of the blank contour height respectively; t is the theoretical wall thickness value of the process; δ is the maximum deformation of the processing; θ is the accuracy threshold, which is determined by the measurement accuracy of the on-machine measuring equipment.

[0078] Data screening, for N groups of data, each group of data is processed according to the "3σ (sigma) principle" and screened based on normal distribution.

[0079] The specific steps are:

[0080] Calculate the mean μ and standard deviation σ of the contour height in each set of data;

[0081] Identify and remove outliers, which are defined as measurements less than μ-3σ or greater than μ+3σ;

[0082] The mean value μ′ of the profile height in each set of data is re-solved, and μ′ is used as the true value of the profile height corresponding to the measurement point.

[0083] Calculate the mean μ1 and standard deviation σ1 of the wall thickness in each set of data;

[0084] Identify and remove outliers, which are defined as measurements less than μ1-3σ1 or greater than μ1+3σ1;

[0085] The mean value μ1′ of the wall thickness in each set of data is re-solved, and μ1′ is used as the true value of the wall thickness corresponding to the measuring point.

[0086] The generated process data is organized into a list structure, and the list structure is defined as P1,…,P i ,…,P N , where i = 1, 2, ..., N, the i-th data P i It can be expressed as P i (x i ,y i ,z i ,a i ,b i ,c i ,h i ,t i ). Where (x i ,y i ,z i ) is the coordinate of the i-th measurement point, (a i ,b i ,c i ) is the normal vector at the i-th measurement point, h i is the height of the profile at the i-th measuring point, t i is the wall thickness value at the i-th measuring point.

[0087] c) Forward construction of process part model.

[0088] According to the process measurement data, by means of the forward construction method of the process part model, the intermediate process parts are gradually generated from the blank, and finally the part delivery model is generated.

[0089] According to the technological process, the process of machining the bottom part of the box from the blank into a part can be expressed as a process sequence ρ(ε1, ε2, …, ε n ), where ε i (i = 1, 2, …, n) is the i-th process. During the part machining process, the intermediate part solid model after completing the m-th process ε m (0 < m ≤ n) is called the process part model corresponding to this process, denoted as p m . It can be seen from this that p n is the part finished product model, and p0 is the blank model.

[0090] The process of constructing the blank model p0 is as follows: according to the theoretical inner surface of the part or the upper surface of the tooling, the sampling points are offset outward by the contour height, and the blank outer surface is formed by interpolation. Then, the sampling points are offset inward by the wall thickness value, and the blank inner surface is formed by interpolation. The specific process of the two offsets is as follows: First, according to the sampling points Z i (i = 1, 2, …, N) of the point-measurement type measurement path, the previous point L a corresponding to the upper surface of the tooling is obtained; then, according to the theoretical inner surface of the part, the normal vector V a at the point L a is solved. The point L a is offset by the contour height value h a along the normal vector V i to obtain the point L b . The point L b is the corresponding point on the outer surface of the blank model; finally, according to the outer surface of the blank model, the normal vector V b at the point L b is solved. The point L b is offset by the wall thickness value t b along the normal vector V i to obtain the point L c . The point L c is the corresponding point on the inner surface of the blank model, as shown in Figure 4 .

[0091] The process of constructing the intermediate process part p m is as follows: after the outer surface of the model is machined by the theoretical tool path, the wall thickness data of the intermediate process is measured and obtained. The wall thickness machining deviation of the intermediate process part is solved by accumulating the deviation data between the wall thickness of all the machined processes and the theoretical tool path. Based on the inner surface of the blank model, the current intermediate process part model is generated by driving with the wall thickness of the current process and its accumulated wall thickness deviation data, and the boss feature is generated according to the process requirements. The specific steps are as follows:

[0092] (1) Calculate the cumulative wall thickness deviation. The cumulative wall thickness deviation of the intermediate process parts is solved by accumulating the wall thickness deviation data of the processed process. Assume that process ε m The measured wall thickness is t m , the theoretical wall thickness is T m , then the process ε m The wall thickness deviation is t m -T m . According to the process order, process ε m The cumulative wall thickness deviation is recorded as:

[0093]

[0094] (2) In-process surface generation: According to the accumulated wall thickness deviation value, the inner surface interpolation points of the blank model are adjusted to generate the inner surface of the current process by interpolation through points.

[0095] The method for adjusting the inner surface interpolation points is as follows: m >0, that is The measured wall thickness is greater than the theoretical wall thickness, process ε m The cutting amount is less than the theoretical cutting amount, and the inner surface of the blank model is adjusted in the direction D m <0, the inner surface is adjusted inward at this point; when E m <0, that is The measured wall thickness is less than the theoretical wall thickness, process ε m The cutting amount is greater than the theoretical cutting amount, and the inner surface of the blank model is adjusted in the direction D m >0, the inner surface is adjusted outward at this point; when E m =0, no adjustment.

[0096] The inner surface interpolation process is as follows: inner surface interpolation point Z j (j=1,2,…,N), adjusted surface interpolation points Through all points The current process ε is generated by interpolation m The inner surface Among them, the normal unit vector D at a point on the inner surface m The determination method is: when D m <0, D m Inward, that is, from the outer surface to the inner surface; when D m >0, D m Outward, that is, from the inner surface to the outer surface; when D m = 0, D m In any direction.

[0097] (3) The outer surface of the process is generated, and the interpolation points of the inner surface are offset to the normal direction of the current process measured wall thickness value to generate the outer surface interpolation points Generate the current process ε by interpolating through points m External surface It can be expressed as:

[0098]

[0099] Among them, d m It is the normal unit vector at the interpolation point of the inner surface, and its direction is from the inner surface to the outer surface.

[0100] (4) Create a boss feature. If the processing area of ​​the process contains a boss, use the processing feature recognition method to extract the boss area boundary of the original model and project it onto the external surface of the current process. The boss area generated by cutting is offset outward to generate an external surface with a boss area.

[0101] The process of creating boss features is as follows: The boss feature recognition method of the box bottom model can adopt general methods such as attribute adjacency graph and rule extraction. Assuming that there are k bosses on the box bottom, the boss boundary set generated by identification and extraction can be expressed as B(β1,β2,…,β k ). First, move the boss boundary set B to Projection on the top forms the projection boundary B′(β1′,β′2,…,β′ k ). Then, the boss feature projection boundary B′ is clipped For the boss feature boundary with nested relationship, the boss area is automatically determined by the clipping relationship tree. Finally, the boss area generated by clipping is offset outward by step ε m The boss height required by the process is generated to generate the outer surface with the boss area

[0102] The process of automatically constructing the clipping relationship tree is to automatically construct a clipping relationship tree with the current boss feature and the thinning area feature, and the area with an odd depth in the clipping relationship tree is judged as the boss area, such as Figure 5 For example, Figure 5 In case 1, A is a thinning area, B is a thickening area, and C is a thinning area. The constructed clipping relationship tree judges the area B with an odd depth as a boss area. Similarly, areas B and C in case 2 are judged as boss areas.

[0103] Therefore, the i-th process ε i (i=1,2,…,n) The corresponding intermediate process part model p m , by the outer surface And the inner surface constitute.

[0104] Build the finished part modeln The process is to convert the wall thickness measurement data t n As the final inspection data, the finished part model including the external surface is formed Inner surface The finished part model is a digital mapping of the part processed on site and can be used in the subsequent digital manufacturing inspection process.

[0105] d) Three-dimensional wall thickness cloud map is automatically generated.

[0106] The geometric model is constructed by using the above-mentioned integral box bottom process part model construction method, and the process part model p m The inner surface Superimpose with the inner surface of the theoretical model, and use a set of parallel and equal-height theoretical inner surfaces to Segment and intercept to form contour lines of wall thickness error change, set the contour lines in different colors, and generate process part model p m Wall thickness error variation cloud diagram.

[0107] like Figure 6 As shown in the figure, the minimum and maximum values ​​of the wall thickness deviation are used as the accuracy range [min(E m ), max(E m )], where the distance g between contour lines is set by the user. The contour line method is used to generate the wall thickness distribution contour lines, and the geometric model is integrated with the wall thickness distribution contour lines. The multi-precision three-dimensional wall thickness distribution map displayed by the color setter is used to guide CNC machining programming and process.

[0108] Example:

[0109] like Figure 7 This is a test part for the embodiment. The maximum diameter of the box bottom part is 3350mm, and there are multiple boss structures on the outer surface.

[0110] The process model is constructed by using data error accumulation technology to characterize the processing status between processes using process data, and the forward evolution derives the following Figure 8 In (a), the process part model without boss is generated by using boss feature boundary extraction and projection clipping geometry algorithm, such as Figure 8 As shown in (b).

[0111] Use contour planes to segment and intercept the inner surface, generate contour lines of wall thickness error change, and set contour lines to different color cloud maps, such as Fig. 9 Interactively viewing the wall thickness cloud map in the CAD / CAM 3D environment facilitates manual verification of CNC machining tool paths and prevents quality problems.

[0112] The present invention, firstly, performs data collection on large box bottom structural parts, the data collection includes measuring two types of data, namely, box bottom contour height data and wall thickness data. The measurement method selects a point measurement path, and the measurement points are a series of automatically planned sequential points. The contour height and wall thickness data measuring equipment is connected to a host computer to transmit the measurement data to an on-site workstation; then, the two types of data collected are processed, the data are cleaned to remove repeated and meaningless values, the data grouping is performed to aggregate the data corresponding to the measurement points into data groups one by one, the data group corresponding to each point is screened, the true value corresponding to the measurement point is screened out, and the process data is structured; secondly, according to the process measurement data, in accordance with the process part model forward construction method, the intermediate process parts are gradually generated from the blank, and finally the part delivery model is generated. The specific steps are: 1) construct a blank model, construct the blank inner surface according to the first acquired profile and wall thickness data, and generate the blank outer surface by offsetting the inner surface outward along the normal direction by the wall thickness distance; 2) construct an intermediate process part, after the outer surface of the model is processed by the theoretical tool path, measure and obtain the wall thickness data of the intermediate process, and the wall thickness processing deviation of the intermediate process part is accumulated and solved by the deviation data of the wall thickness of all processed processes and the theoretical tool path. The inner surface of the blank model is used as the reference, and the current process wall thickness and its accumulated wall thickness deviation data are driven to generate the current intermediate process part model, and the boss feature is generated according to the process requirements; 3) construct a finished part model, and the finished part model uses the wall thickness measurement data as the final inspection data. The three-dimensional part process part formed is a digital mirror image of the actual processed part, which can be used for the subsequent digital manufacturing inspection link. Finally, a three-dimensional wall thickness change cloud map is automatically generated. The geometric model is constructed using the above-mentioned integral box bottom process part model construction method. The inner surfaces of the process part models of different processes are superimposed, and a set of parallel contour surfaces are used to divide and intercept the inner surfaces to form wall thickness error change contour lines. The contour lines are set to different colors to generate a wall thickness error change cloud map.

[0113] Aiming at the problems of cumbersome manual interaction in constructing CNC machining process part models and low modeling efficiency in the development process of large-scale integral deformable box bottom parts for aerospace using integrated measurement-machining technology, the present invention proposes an integral box bottom process part model construction method based on process data, which effectively solves the problems of low automation and intelligence in process model construction, shortens the box bottom machining process preparation cycle, ensures the development efficiency of key and important parts, and the formed process part model, as the main part of the actual model, can monitor and analyze the machining status of the workpiece and establish the correlation between the physical space and the digital space of the CNC machining process.

[0114] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A method for constructing an integral box bottom process part model based on process data, characterized in that: include: Use data measuring equipment to collect profile height and wall thickness data of the bottom structural parts of the box; Perform data processing on the two types of collected data; Based on the processed data, the process part model is constructed forward; Generate a cloud diagram of wall thickness error changes for the process part model.

2. The method for constructing an integral box bottom process part model based on process data according to claim 1, characterized in that: The data measurement equipment collects data in a point measurement manner, and the measurement points are a series of sequential points generated by the software system planning.

3. The method for constructing an integral box bottom process part model based on process data according to claim 1, characterized in that: Data processing is performed on the two types of collected data, including: Perform data cleaning on the two types of collected data to remove duplicate and meaningless null values; Aggregate the cleaned data into data groups according to the collection rules; The profile height and wall thickness in each set of data are screened based on the normal distribution according to the "3σ principle", the true values ​​of the profile height and wall thickness corresponding to the measuring points are screened out, and the data are structured.

4. The method for constructing an integral box bottom process part model based on process data according to claim 3 is characterized in that: The cleaned data is aggregated into N data groups according to the collection rules. The collection rules are the following four conditions: The range of blank profile height values ​​measured under condition 1 is [h min ,h max ]; The process wall thickness value measured under condition 2 is [t-δ-θ / 2, t+δ+θ / 2]; Condition 3: The measurement value in each data group corresponds to a time ≤ T seconds; Condition 4: The true value of the measurement is continuous; Among them, h min ,h max are the minimum and maximum values ​​of the blank contour height respectively; t is the theoretical wall thickness value of the process; δ is the maximum deformation of the processing; θ is the accuracy threshold, which is determined by the measurement accuracy of the data measurement equipment; the range of T is [3,10].

5. The method for constructing an integral box bottom process part model based on process data according to claim 3 is characterized in that: The profile height and wall thickness in each set of data are screened based on the normal distribution according to the "3σ principle" to screen out the true values ​​of the profile height and wall thickness corresponding to the measurement points. The specific implementation method is as follows: Calculate the mean μ and standard deviation σ of the contour height in each set of data; Identify and remove outliers, which are defined as measurements less than μ-3σ or greater than μ+3σ; Re-solve the mean value μ′ of the profile height in each set of data, and use μ′ as the true value of the profile height corresponding to the measurement point; Calculate the mean μ1 and standard deviation σ1 of the wall thickness in each set of data; Identify and remove outliers, which are defined as measurements less than μ1-3σ1 or greater than μ1+3σ1; The mean value μ1′ of the wall thickness in each set of data is re-solved, and μ1′ is used as the true value of the wall thickness corresponding to the measuring point.

6. The method for constructing an integral box bottom process part model based on process data according to claim 3, characterized in that: The data is structured and implemented as follows: The process data of each measuring point is organized into a list structure, and the list structure is defined as P1,…,P i ,…,P N , where i = 1, 2, ..., N, the i-th data P i Represented as P i (x i ,y i ,z i ,a i ,b i ,c i ,h i ,t i ), where (x i ,y i ,z i ) is the coordinate of the i-th measurement point, (a i ,b i ,c i ) is the normal vector at the i-th measurement point, h i is the profile height value at the i-th measuring point, t i is the wall thickness value at the i-th measuring point.

7. The method for constructing an integral box bottom process part model based on process data according to claim 1, characterized in that: The process of constructing the process part model forward is as follows: c1) constructing the inner profile of the blank according to the contour height and wall thickness data obtained for the first time, and generating the outer profile of the blank by offsetting the inner profile to the outer edge of the normal wall thickness; c2) Based on the inner surface of the blank model, the current intermediate process part model is generated by the current process wall thickness and its accumulated wall thickness deviation data, and the boss features are generated according to the process requirements; c3) The wall thickness measurement data is used as the final inspection data, and the resulting three-dimensional part process piece is a digital mirror image of the actual processed part, which is used in the subsequent digital manufacturing inspection link.

8. The method for constructing an integral box bottom process part model based on process data according to claim 7, characterized in that: The implementation method of step c2) is as follows: (1) Calculate the cumulative wall thickness deviation. The cumulative wall thickness deviation of the intermediate process parts is solved by accumulating the wall thickness deviation data of the processed processes; (2) According to the accumulated wall thickness deviation value, the interpolation points of the inner profile of the blank are adjusted, and the inner profile of the current process is generated by interpolation of points; (3) The interpolation points of the inner surface are offset outward along the normal direction by the actual measured wall thickness distance of the current process to generate the interpolation points of the outer surface, and the outer surface of the current process is interpolated in a point-by-point manner; (4) Create a boss feature. If the machining area of ​​the process contains a boss, first use the machining feature recognition method to extract the boss area boundary of the original model, then use the feature boundary to crop the current process surface, use the cropping relationship tree to automatically determine the boss area to be offset, and finally offset the boss area outward by the boss height to generate the process surface with the boss area.

9. The method for constructing an integral box bottom process part model based on process data according to claim 1, characterized in that: Generate the wall thickness error variation cloud map of the process part model. The implementation method is as follows: Construct a geometric model, superimpose the inner surface of the process part model with the inner surface of the theoretical model, use a set of parallel and equal-height theoretical inner surfaces to segment and intercept the inner surface of the process part model, form contour lines of wall thickness error variation, set the contour lines to different colors, and generate a cloud map of the wall thickness error variation of the process part model.

10. The method for constructing an integral box bottom process part model based on process data according to claim 9, characterized in that: The method of constructing a geometric model is the same as that of constructing a process part model.