Artificial joint automatic production line system

Abstract

The invention relates to an artificial joint automatic production line system, which belongs to the field of automatic control and manufacturing, and comprises a data acquisition module used for acquiring a material hardness prediction distribution diagram, ideal processing time and a candidate strategy parameter set of a target workpiece; the prediction processing module is used for determining the predicted actual processing time and the predicted process accumulated risk index under each candidate strategy based on the material hardness prediction distribution diagram and the candidate strategy parameter set; the decision optimization module is used for combining ideal processing time, predicting actual processing time and predicting a process accumulated risk index, and calculating a predicted benefit score of each candidate strategy; an optimal processing and quality inspection strategy is determined based on the predicted benefit score; and the self-adaptive execution module is used for actively and prospectively calculating and determining an optimal balance point which maximizes the overall benefit according to the optimal processing and quality inspection strategy and the material hardness prediction distribution diagram.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automation control and manufacturing, in particular to an artificial joint automatic production line system. BACKGROUND

[0002] In the field of high-end medical device manufacturing, especially in the automatic production scene of downstream hospitals providing personalized custom artificial joints, there is a long-term and core technical contradiction. On the one hand, in order to meet the instant demand of emergency surgery and effectively control the cost, the production system must pursue the extreme production efficiency. On the other hand, as a precision component implanted in the human body, each artificial joint must achieve micron-level geometric accuracy to ensure its biocompatibility, functionality and long-term reliability.

[0003] The existing automatic production line, such as the production line using CNC machine tools, usually operates based on fixed, offline programmed machining programs. This mode exposes its inherent defects when dealing with personalized customization tasks, especially when dynamic external interference factors are introduced. The main interference comes from the physical property differences between batches of raw materials. Medical-grade metals or high-molecular polymers, even if they meet the same medical standards, still have unpredictable small fluctuations in microscopic hardness and other physical properties between batches or even different positions within the same batch. Such fluctuations are fatal to micron-level precision. For example, when a CNC cutter cuts at a high speed with preset parameters, if it encounters an area with slightly higher hardness than expected, the cutter will produce a small physical pop phenomenon, which is enough to cause the entire high-value custom part to be directly scrapped.

[0004] To avoid risks, existing technologies usually adopt conservative worst-case strategies, such as using extremely low cutting speeds and significantly increasing the number of production process downtime inspections. These measures guarantee precision to some extent, but severely sacrifice production efficiency, making the production cycle of personalized customization extremely long and the unit cost extremely high, which cannot effectively respond to the urgent needs of the clinic.

[0005] Therefore, there is an urgent need in the art for a new technical solution to solve the core technical problem of dynamically adjusting machining and quality inspection strategies in real time and adaptively under the core interference of continuous dynamic fluctuations in raw material batch characteristics, so as to find and continuously maintain a dynamic optimal balance point between the two key targets of production efficiency and micron-level precision.

[0006] The above information disclosed in the above BACKGROUND section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0007] The present application aims to provide an artificial joint automatic production line system to solve the problems in the background art.

[0008] The technical solution of the present application comprises: a data acquisition module for obtaining a material hardness prediction distribution map of a target workpiece, an ideal machining time, and a candidate strategy parameter set;

[0009] A prediction processing module for determining a predicted actual machining time and a predicted process cumulative risk index under each candidate strategy based on the material hardness prediction distribution map and the candidate strategy parameter set;

[0010] A decision optimization module for calculating a predicted benefit score of each candidate strategy by combining the ideal machining time, the predicted actual machining time, and the predicted process cumulative risk index, and determining an optimal machining and quality inspection strategy based on the predicted benefit score;

[0011] An adaptive execution module for generating adaptive machining instructions and quality inspection instructions according to the optimal machining and quality inspection strategy and in combination with real-time sensed material hardness fluctuations.

[0012] Preferably, the prediction processing module determines the predicted actual machining time and the predicted process cumulative risk index by:

[0013] determining a predictive adjustment feed rate curve based on candidate machining parameters in the candidate strategy parameter set and in combination with the material hardness prediction distribution map, calculating the predicted actual machining time according to the predictive adjustment feed rate curve, and determining the predicted process cumulative risk index based on the predictive adjustment feed rate curve.

[0014] Preferably, the prediction processing module determines the predicted process cumulative risk index based on the predictive adjustment feed rate curve by:

[0015] calculating the instantaneous feed acceleration in the predicted process based on the predictive adjustment feed rate curve, integrating the acceleration over the entire predicted machining time, and finally performing normalization processing to determine the cumulative process uncertainty of each machining feature, mapping the cumulative process uncertainty through a risk conversion function to determine the failure risk probability of each machining feature, and determining the maximum value among the failure risk probabilities of all machining features as the predicted process cumulative risk index.

[0016] Preferably, the decision optimization module calculates the predicted benefit score of each candidate strategy by:

[0017] determining an efficiency benefit item based on the ratio of the ideal machining time to the predicted actual machining time, determining a risk cost item based on the predicted process cumulative risk index, and determining the predicted benefit score according to the difference between the efficiency benefit item and the risk cost item.

[0018] Preferably, the adaptive execution module generates the adaptive processing instructions, including:

[0019] An online sensing unit is used to acquire material hardness fluctuation values ​​in real time during processing;

[0020] The adjustment unit is used to calculate the feed rate after real-time adjustment based on the optimal processing parameters in the optimal processing and quality inspection strategy and combined with the material hardness fluctuation value obtained in real time, so as to generate processing instructions.

[0021] Preferably, the adaptive execution module generates the adaptive quality inspection instruction, including:

[0022] The uncertainty calculation unit is used to determine the real-time cumulative process uncertainty of the processed features based on the actual executed adjusted feed rate history.

[0023] The decision-making unit is used to compare the real-time accumulated process uncertainty with the quality inspection trigger threshold in the optimal processing and quality inspection strategy.

[0024] The instruction generation unit is used to generate targeted quality inspection instructions when the uncertainty of the real-time cumulative process is greater than or equal to the quality inspection trigger threshold, and to generate instructions to skip the inspection when the uncertainty of the real-time cumulative process is less than the quality inspection trigger threshold.

[0025] Preferably, the uncertainty calculation unit determines the uncertainty of the real-time cumulative process, including:

[0026] Obtain the absolute value of the first derivative of the adjusted feed rate with respect to time as the instantaneous feed acceleration; integrate the instantaneous feed acceleration over the entire machining path and normalize the integration result to determine the real-time cumulative process uncertainty.

[0027] Preferably, the risk transformation function is a logistic function, using a preset risk sensitivity coefficient and a candidate quality inspection trigger threshold from the candidate strategy parameter set as model parameters.

[0028] This invention provides an automated production line system for artificial joints, which has the following improvements and advantages compared with the prior art:

[0029] 1. This changes the operational paradigm; the decision optimization module performs a forward-looking evaluation of multiple candidate strategy parameter sets before the processing task begins; by calculating the predicted benefit score of each candidate strategy, this module achieves a quantitative trade-off between efficiency and risk, and transforms a complex multi-objective optimization problem into a solvable single-objective optimization problem by calculating the difference between the efficiency benefit item represented by the processing time ratio and the risk cost item represented by the cumulative risk index of the prediction process; the system no longer passively makes an either-or choice between efficiency and accuracy, but can proactively and forward-lookingly calculate and determine an optimal balance point that maximizes overall benefits for each specific target workpiece based on its material hardness prediction distribution map;

[0030] 2. The adaptive execution module of this invention generates adaptive processing instructions in real time based on the processing parameters in the optimal strategy using an adaptive adjustment model. The practical significance of this model is that it only reduces the feed rate on demand and non-linearly when the material hardness fluctuation is positive in real time, while maintaining a high-efficiency baseline feed rate on most paths where the material hardness is normal or slightly soft. This module calculates the real-time accumulated process uncertainty and compares it with the quality inspection trigger threshold in the optimal strategy to determine whether to generate targeted quality inspection instructions. This means that the system abandons blind, fixed-interval inspections and instead only detects key features where the stability index of the processing process exceeds the safety threshold, greatly reducing unnecessary downtime for inspection. The combined effect of these two improvements significantly shortens the production cycle of personalized artificial joints, directly reducing the manufacturing cost per unit product.

[0031] 3. The application of the adaptive adjustment model in this invention enables the system to proactively suppress machining errors caused by sudden changes in cutting force based on real-time perceived material hardness fluctuations. This is an active suppression mechanism at the source of error. A deeper level of protection comes from the risk prediction capability of the predictive processing module. This module calculates the accumulated process uncertainty and uses a risk transformation function to accurately map the instability index of the physical process into a probabilistic failure risk probability. It can predict which geometric features have the highest failure risk under specific machining parameters. This predictive capability allows the decision optimization module to avoid high-risk machining strategies before machining begins, ensuring the accuracy and reliability of the final product at the strategy level.

[0032] 4. The solution of this invention significantly enhances the system's environmental adaptability and autonomous decision-making ability. Because the system can perceive and adaptively adjust in real time to cope with the uncertainty of material hardness, it has stronger tolerance to fluctuations in the physical properties of incoming materials, that is, the system's robustness is improved. In addition, the entire system's operating logic, from data acquisition and predictive processing to decision optimization and adaptive execution, is entirely driven by data and mathematical models. This complete closed-loop decision and control process reduces the reliance on the operator's personal skills and experience, raising the automation level of the production process to a new level of intelligence and autonomous optimization. Attached Figure Description

[0033] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0034] Figure 1 This is a flowchart of the system of the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0036] Example 1:

[0037] Please see Figure 1 The present invention provides a technical solution for an automated production line system for artificial joints, including: a data acquisition module for acquiring a material hardness prediction distribution map of the target workpiece, an ideal processing time, and a set of candidate strategy parameters;

[0038] The prediction processing module is used to determine the predicted actual processing time and the cumulative risk index of the prediction process under each candidate strategy based on the material hardness prediction distribution map and the candidate strategy parameter set.

[0039] The decision optimization module is used to combine the ideal processing time, predict the actual processing time and the cumulative risk index of the prediction process, calculate the prediction benefit score of each candidate strategy, and determine the optimal processing and quality inspection strategy based on the prediction benefit score.

[0040] The adaptive execution module is used to generate adaptive processing and quality inspection instructions based on the optimal processing and quality inspection strategy and in combination with real-time sensing of material hardness fluctuations.

[0041] The automated production line system for artificial joints provided in this embodiment aims to resolve the inherent contradiction between production efficiency and processing precision in the manufacturing process of personalized medical implants. The system constructs a closed-loop control method from prediction and decision-making to adaptive execution through the logical coordination of four core modules. During the initialization phase of the processing task, the data acquisition module pre-scans the target workpiece to obtain a predicted distribution map of material hardness along the predetermined processing path, and simultaneously loads the ideal processing time generated by CAM software and a set of candidate strategy parameters containing different risk control parameters. Based on the above data, the prediction processing module deduces for each candidate strategy and calculates... This strategy predicts the actual processing time and the cumulative risk index of the process. The decision optimization module substitutes the predicted processing time and risk index into a preset benefit model to calculate the predicted benefit score for each candidate strategy, and selects a processing and quality inspection strategy that can achieve the optimal balance between efficiency and risk. During the actual processing stage, the adaptive execution module strictly follows this optimal strategy and uses online sensors to sense small fluctuations in material hardness in real time to dynamically generate adaptive processing and quality inspection instructions. This complete technical solution transforms production control from traditional offline programming and fixed execution into an integrated technical paradigm of online sensing, predictive decision-making, and adaptive control.

[0042] The pre-scan can be achieved using non-contact ultrasonic phased array scanning technology. The data acquisition module first obtains the ultrasonic surface wave propagation velocity c(l) and material density ρ of the workpiece on the processing path as known parameters, and calculates the acoustic impedance Z(l) = ρ·c(l) distributed along the path, where c(l) refers to the ultrasonic surface wave propagation velocity of the workpiece at position l on the processing path; ρ refers to the material density of the workpiece, which is usually input as a known parameter; and Z(l) refers to the acoustic impedance of the workpiece at position l on the processing path. The system has pre-calibrated an empirical relationship model between acoustic impedance and Vickers hardness by measuring a series of standard material test blocks with known different hardnesses. This model can be expressed as the following quadratic polynomial:

[0043] H pred (l)=a·Z(l) 2 +b·Z(l)+c

[0044] Where a, b, and c are calibration coefficients obtained through regression analysis of the test data of the standard test block; H pred (l) refers to the predicted absolute hardness of the material at path position l; the data acquisition module substitutes the calculated acoustic impedance Z(l) into this formula to obtain the predicted absolute hardness H along the processing path. pred (l); By comparing with the ideal hardness H of the reference material ideal By comparison, the final material hardness prediction distribution map, i.e., the hardness fluctuation prediction map ΔH(l)=H, is obtained.pred (l)-H ideal ; where ΔH(l) refers to the predicted distribution map of material hardness, that is, the predicted fluctuation value of material hardness at processing path position l relative to the ideal hardness. This distribution map will be passed to the prediction processing module as the core input.

[0045] Example 2

[0046] The forecasting module determines the predicted actual processing time and the cumulative risk index of the forecasting process, including:

[0047] Based on the candidate processing parameters in the candidate strategy parameter set and combined with the material hardness prediction distribution map, a predictive adjustment feed rate curve is determined through an adaptive adjustment model; based on the predictive adjustment feed rate curve, the predicted actual processing time is calculated; based on the predictive adjustment feed rate curve, the cumulative risk index of the prediction process is determined.

[0048] The predictive processing module determines the cumulative risk index of the prediction process based on the predictive adjustment feed rate curve, including:

[0049] Based on the predictive adjustment feed rate curve, the instantaneous feed acceleration during the prediction process is calculated, and the acceleration is integrated over the entire prediction processing time and then normalized to determine the cumulative process uncertainty of each processing feature. The cumulative process uncertainty is mapped through a risk transformation function to determine the failure risk probability of each processing feature. The maximum value among the failure risk probabilities of all processing features is determined as the cumulative risk index of the prediction process.

[0050] The risk transformation function is a logistic function, using a preset risk sensitivity coefficient and a candidate quality inspection trigger threshold from the candidate strategy parameter set as model parameters.

[0051] The prediction processing module in this embodiment is used to perform quantitative prediction modeling of future processing. This module calls an adaptive adjustment model and compares the predicted material hardness distribution map ΔH(l) obtained from the pre-scan with candidate processing parameters in the candidate strategy, such as the feed rate adjustment coefficient k. H Combined, a predictive feed rate adjustment curve f is generated. pred_adj (l);

[0052] The specific mathematical expression of this adaptive adjustment model in the prediction phase remains consistent with the model structure in the execution phase. It aims to simulate possible feed rate adjustments during future processing. The formula is as follows:

[0053]

[0054] Among them, f pred_adj (l) is the predicted and adjusted feed rate at position l on the machining path; f baseIt is the reference feed rate derived from the CAM system, corresponding to the machining speed on materials of ideal hardness; k H This is the feed rate adjustment coefficient in the current candidate strategy; while ΔH(l) is the predicted fluctuation value of material hardness at path position l provided by the aforementioned data acquisition module. By substituting the ΔH(l) value at each point on the path into this formula, a complete and continuous predictive feed rate adjustment curve that reflects hardness changes can be generated;

[0055] Based on this prediction curve, the actual processing time T is predicted. pred_actual It is obtained by integrating the reciprocal of the predictive feed rate along the entire machining path, as shown in the formula.

[0056] Among them, T pred_actual This refers to the predicted actual processing time; ∫path is the reciprocal of the predictive feed rate adjustment; the calculation method ensures that the predicted processing time can accurately reflect the time changes caused by the dynamic adjustment of the feed rate due to changes in material hardness;

[0057] The predictive processing module determines the cumulative risk index R of the predictive process based on the same predictive adjustment feed rate curve. pred_proc The determination process includes the following steps: Following the method described in Example 2, the determination process first calculates the predictive cumulative process uncertainty U for each processing feature. feat This is used to quantify the physical instability of the processing in the prediction; to establish an accurate mapping between the uncertainty of this process and the failure risk, the system uses the logistic function as the risk transformation function, with the expression:

[0058]

[0059] The purpose of introducing this formula is to construct a nonlinear probabilistic model from process uncertainty to product failure risk. Compared with the linear model, the logistic function can more accurately characterize the rapid change characteristics of risk near the critical threshold. That is, when the uncertainty index approaches a certain critical point, the risk increases rapidly, while the change is gradual when it is far away from that point. This nonlinear characteristic is more in line with the failure law in engineering practice.

[0060] Among them, P fail The failure risk probability of a single processing feature is a dimensionless scalar in the interval (0,1); U feat The predictive cumulative process uncertainty derived from the prediction curve in the aforementioned steps is a dimensionless index calculated in the previous step. Its specific calculation method has been detailed in the relevant paragraphs of this specification; that is, it is determined by integrating and normalizing the predicted instantaneous feed acceleration. η riskθ is a dimensionless risk sensitivity coefficient used to control the steepness of the risk curve near a threshold, determining the system's sensitivity to changes in uncertainty; U The candidate quality inspection trigger threshold, defined for the candidate strategy parameter set, represents the inflection point where risk changes drastically; parameter η risk With θ U All of these are achieved through analysis of historical production data, including U values ​​for various characteristics. feat The calculated values ​​and their corresponding final product inspection results are calibrated using logistic regression analysis to ensure the accuracy of the model's predictions.

[0061] During the prediction phase, the prediction processing module processes each key processing feature by substituting its U... feat The value is used to calculate the corresponding failure risk probability P. fail,i This function effectively transforms the measurement indicators of a physical process into probabilistic indicators required by the decision-making layer, enabling the system to predict which processing feature is most likely to become the link with the highest failure risk under a specific processing strategy. This module takes the maximum value among the failure risk probabilities of all processing features as the cumulative risk index of the predicted process under the current candidate strategy. This risk index, along with the predicted processing time, will be used as input for subsequent decision optimization; where R... pred_proc The cumulative risk index for the prediction process is the maximum value among all processing feature failure probability; i is an index representing each key processing feature.

[0062] Example 3

[0063] The decision optimization module calculates the predicted benefit score for each candidate strategy, including:

[0064] The efficiency benefit item is determined based on the ratio of ideal processing time to predicted actual processing time; the risk cost item is determined based on the cumulative risk index of the prediction process; and the prediction benefit score is determined based on the difference between the efficiency benefit item and the risk cost item.

[0065] The decision optimization module in this embodiment is used to select the optimal solution from all candidate strategies. This module achieves this function by calculating a quantified prediction benefit score, the formula for which is defined as:

[0066]

[0067] This formula is designed to unify and reduce the mutually constraining efficiency and quality risk objectives in the production process; the technical motivation lies in transforming a complex multi-objective optimization problem into a single-objective problem, namely S. pred The mathematical problem of finding the maximum value enables the system to make a forward-looking prediction of the comprehensive benefits of different strategies based on data before processing;

[0068] Among them, S pred To predict the benefit score, a dimensionless scalar is used as the final basis for evaluating the merits of each candidate strategy; T ideal The ideal machining time, derived from the theoretical shortest processing time calculated by CAM software based on standard material properties and optimal cutting parameters, serves as a benchmark for measuring efficiency; T pred_actual To predict the actual processing time, the aforementioned prediction processing module calculates the time based on the current candidate strategy; R pred_proc The cumulative risk index for the forecasting process is also calculated by the forecasting processing module; w E with w R These are dimensionless weighting coefficients for efficiency and risk, respectively. Their values ​​can be preset according to the specific production task requirements, such as the urgency of the task or cost sensitivity, to reflect the flexibility of decision-making.

[0069] At the application level, the decision optimization module traverses all candidate policy parameter sets, for example, different k H and θ U Combination; for each candidate strategy, the module calls the prediction processing module to obtain the corresponding T. pred_actual and S pred_proc Substitute into the above formula to calculate S. pred After completing the calculation of all candidate strategies, the module will... pred The strategy with the highest score was determined as the optimal processing and quality control strategy. This mechanism ensures that the system has completed a comprehensive, data-driven quantitative trade-off before processing, and proactively determines the combination of operating parameters that maximizes the overall output benefits, fundamentally avoiding the problem of efficiency and quality imbalance caused by relying on experience or adopting a single conservative strategy.

[0070] Example 4

[0071] The adaptive execution module generates adaptive processing instructions, including:

[0072] An online sensing unit is used to acquire material hardness fluctuation values ​​in real time during processing;

[0073] The adjustment unit is used to calculate the feed rate after real-time adjustment based on the optimal processing parameters in the optimal processing and quality inspection strategy and combined with the material hardness fluctuation value obtained in real time, so as to generate processing instructions.

[0074] In this embodiment, the adaptive execution module is responsible for converting the optimal strategy determined by the decision optimization module into precise physical actions of the machine tool. When generating adaptive machining instructions, this module completes this process through the collaboration of its internal online sensing unit and adjustment unit. During machining, the online sensing unit, such as a torque or acoustic emission sensor integrated on the spindle, continuously monitors the cutting state and calculates the material fluctuation value ΔH(t) relative to the reference hardness in real time. This calculation process is based on a cutting mechanics model, which assumes that when parameters such as cutting depth and width are constant, the spindle cutting torque is proportional to the material's cutting resistance, and the cutting resistance is strongly correlated with the material hardness. When machining materials with reference hardness, the system records a stable reference spindle torque M. base During actual processing, the online sensing unit and spindle torque sensor measure the current torque M(t) in real time; the real-time hardness fluctuation value ΔH(t) is calculated using the following linear transformation model:

[0075] ΔH(t)=K T ·(M(t)-M base )

[0076] Among them, K T M(t) is a pre-calibrated torque-stiffness conversion coefficient; M(t) is the current torque measured in real time by the spindle torque sensor at time t; M base It is the stable reference spindle torque recorded when machining materials with reference hardness; K T It is a pre-calibrated torque-hardness conversion coefficient; this coefficient was determined through preliminary experiments: a series of material samples with known hardness differences were machined, and their corresponding stable torque differences were recorded. K was then calculated through linear regression analysis. T The numerical value. This method enables the system to reliably convert easily measurable torque physical quantities into material hardness fluctuation indicators required for adaptive feed rate adjustment in real time; this real-time fluctuation value is immediately transmitted to the adjustment unit;

[0077] The adjustment unit has a built-in adaptive adjustment model, the mathematical expression of which is:

[0078]

[0079] This model aims to endow the system with the ability to actively suppress machining interference. The slight increase in material hardness is the direct cause of the tool to generate slight elastic deformation and bounce, which leads to micron-level geometric errors. By establishing an exponential decay relationship between hardness fluctuation and feed rate, the formula can nonlinearly reduce the cutting speed in a feedforward manner at the moment when the hardness shows an increasing trend, thereby actively suppressing the generation of errors at the root.

[0080] Among them, f adj (t) represents the feed rate command output to the CNC controller after real-time adjustment; kH It is the feed rate adjustment coefficient, the optimal processing parameter selected by the decision optimization module; f base The reference feed rate is the maximum processing rate set under ideal material conditions (ΔH(t)≤0); ΔH(t) is the material hardness fluctuation value measured in real time by the online sensing unit at time t; the function structure of max(0,ΔH(t)) ensures that the deceleration mechanism is activated only when the material hardness is higher than the reference value; k H This is the feed rate adjustment coefficient, a key optimal processing parameter. Its value is not fixed, but rather the optimal value calculated and selected by the aforementioned decision optimization module from numerous candidate values. To ensure dimensional consistency in the formula, the coefficient k... H The dimension of is set to be the reciprocal of the dimension of the material hardness fluctuation value ΔH(t), thus ensuring that the exponent part is a dimensionless pure number.

[0081] Feed rate adjustment coefficient k in the candidate strategy parameter set H The candidate value range is an empirical range pre-set based on historical processing data and cutting process manuals for commonly used medical materials, such as titanium alloys and polymers.

[0082] During the execution phase, the adjustment unit performs this calculation continuously at a high frequency, for example, on a millisecond-level basis, taking the real-time sensed ΔH(t) and the already optimized selected k. H Substitute the values ​​into the formula to calculate the instantaneous f. adj (t) and sends it as a machining command to the machine tool servo system; this mechanism enables the machining process to have high dynamic response capability: when the tool encounters a hard area of ​​the material, the system can instantly reduce the feed rate to maintain the cutting force stability and ensure machining accuracy; after passing through the area, the feed rate can be quickly restored to catch up with the machining time; its final technical effect is to significantly reduce the time waste caused by adopting a global conservative machining strategy while ensuring micron-level accuracy, thereby improving production efficiency.

[0083] Example 5

[0084] The adaptive execution module generates adaptive quality inspection instructions, including:

[0085] The uncertainty calculation unit is used to determine the real-time cumulative process uncertainty of the processed features based on the actual executed adjusted feed rate history.

[0086] The decision-making unit is used to compare the real-time accumulated process uncertainty with the quality inspection trigger threshold in the optimal processing and quality inspection strategy.

[0087] The instruction generation unit is used to generate targeted quality inspection instructions when the uncertainty of the real-time cumulative process is greater than or equal to the quality inspection trigger threshold, and to generate instructions to skip the inspection when the uncertainty of the real-time cumulative process is less than the quality inspection trigger threshold.

[0088] The uncertainty calculation unit determines the real-time cumulative process uncertainty, including:

[0089] Obtain the absolute value of the first derivative of the adjusted feed rate with respect to time as the instantaneous feed acceleration; integrate the instantaneous feed acceleration over the entire machining path and normalize the integration result to determine the real-time cumulative process uncertainty.

[0090] In this embodiment, in addition to controlling the processing, the adaptive execution module also performs intelligent management of the quality inspection process; its function of generating adaptive quality inspection instructions aims to replace the traditional fixed or random inspection mode with on-demand targeted inspection, thereby reducing ineffective downtime inspection time; after the processing of a key geometric feature is completed, the uncertainty calculation unit in the module is activated.

[0091] This unit indirectly assesses quality risk by analyzing the physical smoothness of the processing. Its core task is to calculate the real-time cumulative process uncertainty of the processed feature; the calculation formula is:

[0092]

[0093] The technical motivation behind this formula is to establish an index that can quantify the relationship between process instability and final product error; drastic changes in feed rate, i.e., high feed acceleration. It is the physical manifestation of the servo system making drastic adjustments in response to external interference; this adjustment itself will cause micro-vibration and dynamic error of the machine tool, which is a strong predictor of the final dimensional deviation; therefore, by integrating the feed acceleration magnitude over time on the entire machining path, the instability of the entire process can be effectively quantified, thereby indirectly reflecting the potential quality risk of this feature.

[0094] Among them, U feat_actual It is the real-time cumulative process uncertainty calculated after specific geometric features have been processed, and is a dimensionless index; f adj (t) is the historical record of the real-time adjusted feed rate output by the adaptive adjustment unit; It is the absolute value of the first derivative of the instantaneous feed rate with respect to time, i.e., the magnitude of the instantaneous feed acceleration; ∫ path ...dt represents the integration over the entire machining toolpath of the current feature in the time domain; L path It represents the total toolpath length for the current feature, provided by the CAM software; CU is a calibration coefficient with time dimensions, its function is to convert physical quantities... This physical quantity has the dimension of 1 / T, which can be converted to the dimensionless index U. feat This facilitates subsequent threshold comparisons; the coefficient C UThe numerical values ​​were obtained through statistical calibration during the system debugging phase: a series of samples were processed, the original uncertainty calculation values ​​of each feature and their final measured geometric errors were recorded, and a dimensionless U was determined through regression analysis. feat The value has the best correlation with the final risk of deviation. U value;

[0095] After calculating U feat_actual Then, the decision-making unit immediately compares it with the "quality inspection trigger threshold" θ included in the optimal processing and quality inspection strategy. U Compare; this θ U The value is also the optimal parameter selected by the top-level decision optimization module; if U feat_actual ≥θ U If the instruction generation unit determines that the processing of this feature is highly volatile and poses a high quality risk, it will generate a targeted quality inspection instruction and call upon the in-sequence inspection equipment to measure this specific feature; conversely, if U feat_actual <θ U The system then determines that the process is stable and the risk is controllable, and generates an instruction to skip the inspection. This mechanism, which assesses risk in real time based on process data and decides whether to inspect, precisely applies inspection resources to the highest-risk links, avoiding unnecessary downtime inspections of qualified characteristics, thereby directly improving the overall operating efficiency of the production line.

[0096] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An automated production line system for artificial joints, characterized in that, include: The data acquisition module is used to obtain the predicted distribution map of the material hardness of the target workpiece, the ideal processing time, and the set of candidate strategy parameters; The prediction processing module is used to determine the predicted actual processing time and the cumulative risk index of the prediction process under each candidate strategy based on the material hardness prediction distribution map and the candidate strategy parameter set. The decision optimization module is used to combine the ideal processing time, predict the actual processing time and the cumulative risk index of the prediction process, calculate the prediction benefit score of each candidate strategy, and determine the optimal processing and quality inspection strategy based on the prediction benefit score. The adaptive execution module is used to generate adaptive processing and quality inspection instructions based on the optimal processing and quality inspection strategy and in combination with real-time sensing of material hardness fluctuations.

2. The automated production line system for artificial joints according to claim 1, characterized in that, The prediction processing module determines the predicted actual processing time and the cumulative risk index of the prediction process, including: Based on the candidate processing parameters in the candidate strategy parameter set and combined with the material hardness prediction distribution map, a predictive adjustment feed rate curve is determined through an adaptive adjustment model; based on the predictive adjustment feed rate curve, the predicted actual processing time is calculated; and based on the predictive adjustment feed rate curve, the cumulative risk index of the prediction process is determined.

3. The automated production line system for artificial joints according to claim 2, characterized in that, The prediction processing module determines the cumulative risk index of the prediction process based on the predictive adjustment feed rate curve, including: Based on the predictive adjustment feed rate curve, the instantaneous feed acceleration during the prediction process is calculated, and the acceleration is integrated over the entire prediction processing time and then normalized to determine the cumulative process uncertainty of each processing feature. The cumulative process uncertainty is mapped through a risk transformation function to determine the failure risk probability of each processing feature. The maximum value among the failure risk probabilities of all processing features is determined as the cumulative risk index of the prediction process.

4. The automated production line system for artificial joints according to claim 1, characterized in that, The decision optimization module calculates the predicted benefit score for each candidate strategy, including: The efficiency benefit item is determined based on the ratio of ideal processing time to predicted actual processing time; the risk cost item is determined based on the cumulative risk index of the prediction process; and the prediction benefit score is determined based on the difference between the efficiency benefit item and the risk cost item.

5. The automated production line system for artificial joints according to claim 1, characterized in that, The adaptive execution module generates the adaptive processing instructions, including: An online sensing unit is used to acquire material hardness fluctuation values ​​in real time during processing; The adjustment unit is used to calculate the feed rate after real-time adjustment based on the optimal processing parameters in the optimal processing and quality inspection strategy and combined with the material hardness fluctuation value obtained in real time, so as to generate processing instructions.

6. The automated production line system for artificial joints according to claim 1, characterized in that, The adaptive execution module generates the adaptive quality inspection instruction, including: The uncertainty calculation unit is used to determine the real-time cumulative process uncertainty of the processed features based on the actual executed adjusted feed rate history. The decision-making unit is used to compare the real-time accumulated process uncertainty with the quality inspection trigger threshold in the optimal processing and quality inspection strategy. The instruction generation unit is used to generate targeted quality inspection instructions when the uncertainty of the real-time cumulative process is greater than or equal to the quality inspection trigger threshold, and to generate instructions to skip the inspection when the uncertainty of the real-time cumulative process is less than the quality inspection trigger threshold.

7. The automated production line system for artificial joints according to claim 6, characterized in that, The uncertainty calculation unit determines the uncertainty of the real-time cumulative process, including: Obtain the absolute value of the first derivative of the adjusted feed rate with respect to time as the instantaneous feed acceleration; integrate the instantaneous feed acceleration over the entire machining path and normalize the integration result to determine the real-time cumulative process uncertainty.

8. The automated production line system for artificial joints according to claim 3, characterized in that, The risk transformation function is a logistic function, using a preset risk sensitivity coefficient and a candidate quality inspection trigger threshold from the candidate strategy parameter set as model parameters.

Images