An AI-based intelligent analysis system for workpiece processing data

Abstract

The present invention relates to the field of industrial intelligent manufacturing technology, and in particular to an AI-based workpiece processing data intelligent analysis system, comprising: a workpiece data acquisition module for collecting workpiece design data; a multi-source data acquisition module for collecting workpiece processing history data, including workpiece processing data and environmental data within a processing cycle; a historical data analysis module for setting the workpiece processing cycle and the processing parameters of each processing cycle within the processing cycle; a surface error detection module for constructing a workpiece surface quality index; an adaptive analysis module for optimizing the analysis results of the surface quality status; a workpiece property detection module for constructing a workpiece optical property index; a quality prediction module for predicting the offset status of the workpiece and analyzing abnormal processing status of the workpiece; and a processing management module for optimizing the processing parameters for the next processing cycle. The present invention effectively improves the efficiency of workpiece processing.

Description

Technical Field

[0001] The present invention relates to the field of industrial intelligent manufacturing technology, and in particular to an AI-based workpiece processing data intelligent analysis system. Background Art

[0002] As manufacturing evolves toward intelligent manufacturing, effectively utilizing the vast amounts of data generated during production becomes crucial for improving manufacturing processes. Traditional methods, which rely on manual experience or simple statistical analysis, struggle to meet the precision, efficiency, and flexibility demands of modern manufacturing. Therefore, developing intelligent systems that can automatically identify patterns, predict outcomes, and provide optimization recommendations is crucial.

[0003] Chinese patent publication number CN118485351A discloses a quality management system for irregular workpieces based on multi-source data, including a high-speed dynamic capture module, a vibration pattern analysis module, a real-time adjustment execution module, and a quality control update module. In this invention, through high-speed dynamic capture and vibration pattern analysis methods, the morphological changes and vibration frequencies of the workpiece are accurately captured, and real-time identification and adjustment of abnormal vibration patterns are achieved. Refined monitoring improves the accuracy of workpiece processing, significantly reduces the scrap rate, and further improves production efficiency and product reliability. Combined with the comparison of dynamic morphological data with historical vibration data, it allows for more flexible and precise control of the production process, making the processing of irregular workpieces more in line with quality requirements. Real-time monitoring and automatic adjustment of workpiece processing parameters ensure that each workpiece can be processed under optimal production conditions. It can be seen that when performing quality management on the workpiece, this invention does not analyze the properties of the workpiece itself, and there is a problem of low management efficiency when performing quality management on special workpieces. Summary of the Invention

[0004] The object of the present invention is to provide an AI-based intelligent analysis system for workpiece processing data to solve at least one of the problems existing in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] An AI-based intelligent analysis system for workpiece processing data, characterized by comprising:

[0007] Workpiece data acquisition module, used to collect workpiece design data;

[0008] Multi-source data acquisition module, used to collect workpiece processing history data, as well as workpiece processing data and environmental data within the processing cycle;

[0009] A historical data analysis module is used to set the processing cycle of the workpiece and the processing parameters of each processing cycle within the processing cycle according to the workpiece processing history data, and to obtain the standard surface shape and standard optical property parameters of each processing cycle;

[0010] A surface error detection module is used to construct a workpiece surface quality index based on the processing parameters of each processing cycle within the processing cycle and the workpiece processing data within the processing cycle;

[0011] a workpiece property detection module for constructing a workpiece optical property index based on optical property parameters within a processing cycle and standard optical property parameters;

[0012] A quality prediction module is used to predict the deviation state of the workpiece based on the workpiece surface quality index and the workpiece optical property index during the processing cycle, and analyze the abnormal processing state of the workpiece based on the prediction results;

[0013] The processing management module is used to optimize the processing parameters of the next processing cycle according to the analysis results of the abnormal processing status of the workpiece in the processing cycle.

[0014] Furthermore, the historical data analysis module includes a processing cycle analysis unit, and the processing cycle analysis unit is used to set the processing cycle of the workpiece according to the processing history data of the workpiece;

[0015] The processing cycle analysis unit counts the processing cycle length of historical workpieces and calculates the weighted processing cycle T of historical workpieces. ;

[0016] Where T(i) is the processing cycle time of the i-th historical workpiece, Lv(i) is the detection scattering loss ratio of the i-th historical workpiece, Lt(i) is the detection transmittance of the i-th historical workpiece, LV is the target scattering loss ratio, LT is the target transmittance, a1 is the scattering weight, a2 is the transmittance weight, a1+a2=1, and N is the number of historical workpieces;

[0017] The processing cycle analysis unit uses the weighted processing cycle T of the historical workpiece as the processing cycle of the workpiece.

[0018] Furthermore, the historical data analysis module further includes a processing parameter analysis unit, and the processing parameter analysis unit is used to set the processing parameters of the workpiece in each processing cycle according to the processing data of the workpiece;

[0019] The processing parameter analysis unit sets the rated polishing pressure P(j) of each processing cycle according to the polishing pressure of each historical workpiece, and sets ;

[0020] The processing parameter analysis unit sets the rated spindle speed R(j) of each processing cycle according to the spindle speed of each historical workpiece, and sets ;

[0021] The processing parameter analysis unit sets the rated polishing liquid flow rate Q(j) for each processing cycle according to the polishing liquid flow rate of each historical workpiece, and sets ;

[0022] Wherein, P(j), R(j) and Q(j) are the rated polishing pressure, rated spindle speed and rated polishing fluid flow rate of the jth processing cycle within the machining cycle, respectively; P(j,i) is the polishing pressure of the i-th historical workpiece in the jth processing cycle; R(j,i) is the spindle speed of the i-th historical workpiece in the jth processing cycle; Q(j,i) is the polishing fluid flow rate of the i-th historical workpiece in the jth processing cycle; PV(j,i) is the peak-to-valley value of the surface shape of the i-th historical workpiece in the jth processing cycle compared with the standard surface shape;

[0023] The processing parameter analysis unit is further used to store the standard surface shape and standard optical property parameters of each processing cycle.

[0024] Furthermore, the surface error detection module extracts the processing parameters in the current processing cycle in the processing cycle, and calculates the equipment deviation index α(j), setting α(j)=ln{1+[PP(j)] / P(j)×{[RR(j)] / R(j)}×{[QQ(j)] / Q(j)}}; wherein P is the polishing pressure of the processing equipment in the processing cycle, R is the spindle speed of the processing equipment in the processing cycle, and Q is the polishing liquid flow rate of the processing equipment in the processing cycle;

[0025] The surface error detection module analyzes the surface quality state of the workpiece within the processing cycle according to the equipment offset index α(j) and the peak-to-valley value of the workpiece within the processing cycle: if α(j)×PV<λ / η×(Nj) / N, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is normal; otherwise, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is abnormal, and constructs the workpiece surface quality index β(j), setting β(j)=exp{[α(j)×PV-λ / η] / (λ / η×(Nj) / N)}; wherein PV is the peak-to-valley value of the surface shape of the workpiece within the processing cycle and the standard surface shape, λ is the detection wavelength, and η is the error proportional constant.

[0026] Furthermore, it also includes an adaptive analysis module, which is used to analyze the thermal conductivity state according to the environmental data in the processing cycle and optimize the analysis result of the surface quality state according to the analysis result;

[0027] The adaptive analysis module further includes a thermal conductivity anomaly analysis unit, which is used to analyze the thermal conductivity state according to the ambient temperature within the processing cycle to determine whether the thermal conductivity state is normal, and adjust the device offset index to α(j)' when the thermal conductivity state is abnormal.

[0028] Furthermore, the adaptive analysis module also includes a bubble defect analysis unit, which is used to analyze the bubble risk according to the ambient humidity and equipment vibration frequency during the processing cycle to determine whether the bubble risk is normal, and correct the preset ambient temperature to K' when the bubble risk is abnormal.

[0029] Furthermore, the workpiece property detection module analyzes the optical state of the workpiece according to the optical property parameters and the standard optical property parameters within the processing cycle, and constructs the workpiece optical property index according to the analysis results: if a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)<G×(Nj) / N, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is normal; otherwise, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is abnormal, and constructs the workpiece optical property index γ(j), and sets γ(j )=ln{e+{a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)-G×(Nj) / N} / (G×(Nj) / N)}; where Lm(j) is the rated scattering loss ratio within the j-th processing cycle, Lk(j) is the rated target transmittance within the j-th processing cycle, Lk(j) is the rated target transmittance within the j-th processing cycle, Lv(j) is the scattering loss ratio within the j-th processing cycle, Lt(j) is the detected transmittance within the j-th processing cycle, G is the optical property threshold, and e is the natural logarithm.

[0030] Furthermore, the quality prediction module includes an overall offset analysis unit, which is used to predict the offset state of the workpiece based on the workpiece surface quality index construction results and the workpiece optical property index construction results within the processing cycle;

[0031] The overall offset analysis unit counts the workpiece surface quality index and the workpiece optical property index of each processing cycle within the processing cycle, and predicts the offset state of the workpiece based on the statistical results. The specific process is as follows: If <D1, the overall offset analysis unit determines that the offset state of the workpiece is normal; if D1≤ <D2, the overall offset analysis unit determines that the offset state of the workpiece is orderly offset; if ≥D2, the overall offset analysis unit determines that the offset state of the workpiece is an over-limit offset;

[0032] Wherein, D1 is the first preset offset coefficient, D2 is the second preset offset coefficient, D1<D2, c1 is the surface shape weight, c2 is the optical property weight, and c1+c2=1.

[0033] Furthermore, the quality prediction module also includes an alarm unit, which is used to analyze the abnormal processing state of the workpiece based on the prediction result of the offset state of the workpiece, and to alarm the user based on the analysis result: if the offset state of the workpiece within the processing cycle is normal, the alarm unit determines that the processing of the workpiece is normal, and does not alarm the user; if the offset state of the workpiece within the processing cycle is an ordered offset, the alarm unit determines that the abnormal processing state of the workpiece is a fine-tunable abnormality, and sends a second-level alarm to the user; if the offset state of the workpiece within the processing cycle is an out-of-limit offset, the alarm unit determines that the abnormal processing state of the workpiece is a serious offset abnormality, and sends a first-level alarm to the user.

[0034] Furthermore, the processing management module optimizes the processing parameters of the next processing cycle based on the analysis results of the processing abnormality state of the workpiece within the processing cycle: if the processing abnormality state of the workpiece is normal, the processing management module does not change the processing parameters of the next processing cycle; if the processing abnormality state of the workpiece is an ordered offset, the processing management module optimizes the processing parameters of the next processing cycle to P(j+1)', R(j+1)' and Q(j+1)'; if the processing abnormality state of the workpiece is an out-of-limit offset, the processing management module treats the workpiece as an unqualified workpiece.

[0035] Compared with the existing technology, the beneficial effect of the present invention is that: this system realizes precise control of the processing process, rapid response to exceptions and dynamic optimization of parameters through AI-driven multi-module collaboration, effectively solving the problem of traditional methods relying on manual experience and low efficiency, and providing efficient and reliable technical support for industrial intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 This is a structural diagram of the AI-based workpiece processing data intelligent analysis system in this embodiment.

[0038] Figure 2 This is a schematic diagram of the structure of the historical data analysis module in this embodiment.

[0039] Figure 3Schematic diagram of the structure of the adaptive analysis module in this embodiment.

[0040] Figure 4 Schematic diagram of the structure of the quality prediction module of this embodiment. DETAILED DESCRIPTION

[0041] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the following detailed description is illustrative rather than restrictive and should not be used to limit the scope of protection of the present invention.

[0042] It should be noted that, although the terms "first," "second," and "third" may be used to describe the embodiments of the present application, the description should not be limited to these terms. These terms are merely used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, "first" may also be referred to as "second," and similarly, "second" may also be referred to as "first."

[0043] Specifically, the AI-based workpiece processing data intelligent analysis system described in this embodiment is applied to a scenario in which a workpiece is processed using a CNC machine tool; the workpiece described in this embodiment is specifically an optical lens; the AI-based workpiece processing data intelligent analysis system described in this embodiment is installed in a CNC machine tool (CNC); the processing of the workpiece in this embodiment is specifically a polishing process for an optical lens.

[0044] See also Figure 1 As shown in FIG, it is a structural diagram of the AI-based workpiece processing data intelligent analysis system according to this embodiment, including:

[0045] The workpiece data acquisition module is used to collect workpiece design data; the workpiece design data includes the workpiece's target morphological parameters and optical property parameters, including the workpiece's three-dimensional design drawing, and the optical property parameters including the target scattering loss ratio and target transmittance. The workpiece design data is acquired through user interactive input. By collecting the workpiece's target morphological parameters and optical property parameters (such as the three-dimensional design drawing, target scattering loss ratio, and transmittance), accurate benchmark data is provided for subsequent analysis, ensuring that the processing process is consistent with the design objectives and reducing human input errors.

[0046] A multi-source data acquisition module is used to collect workpiece processing data and environmental data within the processing cycle; the workpiece processing data includes polishing pressure, spindle speed, polishing liquid flow, lens surface temperature and vibration spectrum data; it can be understood that the collection process of workpiece processing data in this embodiment is fixed-frequency collection, and the workpiece processing data within the processing cycle includes workpiece processing data collected multiple times; real-time collection of processing data (polishing pressure, spindle speed, etc.) and environmental data (temperature, humidity, vibration frequency) realizes multi-dimensional monitoring of the entire process, provides a data basis for dynamic optimization, and improves the system's adaptability to complex working conditions.

[0047] Specifically, this embodiment does not impose any specific limitation on the value of the processing cycle duration. Those skilled in the art can set it freely as long as the value requirement of the processing cycle duration is met. In this embodiment, the value of the processing cycle duration can be set to 5 seconds. It is worth noting that in this embodiment, the processing cycle is to process the workpiece for a duration of the processing cycle, stop processing, and collect the workpiece processing data within this processing cycle.

[0048] Please continue reading Figure 1 As shown, the system also includes a historical data analysis module, which is connected to the multi-source data acquisition module and the workpiece data acquisition module. The historical data analysis module is used to set the processing cycle of the workpiece and the processing parameters of each processing cycle within the processing cycle according to the workpiece processing history data; based on the weighted calculation of the historical data, the processing cycle and rated parameters (such as polishing pressure, spindle speed) are combined with the scattering loss and transmittance weights to dynamically optimize the processing strategy, reduce dependence on experience, and improve the scientificity and consistency of parameter setting.

[0049] Specifically, the processing cycle is the time period from the start to the end of processing of a workpiece, and a processing cycle is composed of multiple processing cycles; the processing parameters specifically include rated polishing pressure, rated spindle speed, and rated polishing liquid flow; it can be understood that the data processing process is simplified in this embodiment, and the duration of the processing cycle is set to a fixed value. Those skilled in the art can freely set the duration of each processing cycle according to actual conditions to achieve the purpose of improving efficiency.

[0050] See also Figure 2 As shown, the historical data analysis module includes a processing cycle analysis unit, and the processing cycle analysis unit is used to set the processing cycle of the workpiece according to the processing history data of the workpiece;

[0051] The processing cycle analysis unit counts the processing cycle length of historical workpieces and calculates the weighted processing cycle T of historical workpieces. ;

[0052] Among them, T(i) is the processing cycle length of the i-th historical workpiece, i is the number of the historical workpiece, i∈N + , Lv(i) is the detection scattering loss ratio of the i-th historical artifact, Lt(i) is the detection transmittance of the i-th historical artifact, LV is the target scattering loss ratio, LT is the target transmittance, a1 is the scattering weight, a2 is the transmittance weight, a1+a2=1, and N is the number of historical artifacts;

[0053] The processing cycle analysis unit uses the weighted processing cycle T of the historical workpiece as the processing cycle of the workpiece.

[0054] Specifically, in this embodiment, to ensure the accuracy of the historical analysis results, the minimum value of the number N of historically processed workpieces should be 500.

[0055] Please continue reading Figure 2 As shown, the historical data analysis module further includes a processing parameter analysis unit, which is connected to the processing cycle analysis unit, and is used to set the processing parameters of the workpiece in each processing cycle according to the processing data of the workpiece;

[0056] The processing parameter analysis unit sets the rated polishing pressure P(j) of each processing cycle according to the polishing pressure of each historical workpiece, and sets ;

[0057] The processing parameter analysis unit sets the rated spindle speed R(j) of each processing cycle according to the spindle speed of each historical workpiece, and sets ;

[0058] The processing parameter analysis unit sets the rated polishing liquid flow rate Q(j) for each processing cycle according to the polishing liquid flow rate of each historical workpiece, and sets ;

[0059] Where P(j), R(j) and Q(j) are the rated polishing pressure, rated spindle speed and rated polishing fluid flow rate of the jth processing cycle within the machining cycle, respectively; P(j,i) is the polishing pressure of the i-th historical workpiece in the jth processing cycle; R(j,i) is the spindle speed of the i-th historical workpiece in the jth processing cycle; Q(j,i) is the polishing fluid flow rate of the i-th historical workpiece in the jth processing cycle; PV(j,i) is the peak-to-valley value of the surface shape of the i-th historical workpiece in the jth processing cycle compared with the standard surface shape; j is the number of the processing cycle, j∈N + ;

[0060] The processing parameter analysis unit is further used to store the standard surface shape and standard optical property parameters of each processing cycle.

[0061] Specifically, the standard surface shape and standard optical property parameters of each processing cycle described in this embodiment are the standard surface shape and standard optical property parameters obtained by reasoning using reasoning software when processing is performed based on the processing parameters within each processing cycle. The standard surface shape specifically includes a three-dimensional data map of the workpiece, and the standard optical properties include a rated scattering loss ratio and a rated target transmittance. It can be understood that the process of obtaining the standard surface shape and standard optical property parameters in this embodiment is fully disclosed in the prior art and will not be repeated here in this embodiment.

[0062] Please continue reading Figure 1 As shown, the system further includes a surface error detection module, which is connected to the historical data analysis module. The surface error detection module is used to analyze the surface quality status of the workpiece according to the processing parameters of each processing cycle within the processing cycle and the workpiece processing data within the processing cycle, and to construct a workpiece surface quality index based on the analysis results;

[0063] The surface error detection module extracts the processing parameters in the current processing cycle in the processing cycle, and calculates the equipment deviation index α(j), setting α(j)=ln{1+[PP(j)] / P(j)×{[RR(j)] / R(j)}×{[QQ(j)] / Q(j)}}; wherein P is the polishing pressure of the processing equipment in the processing cycle, R is the spindle speed of the processing equipment in the processing cycle, and Q is the polishing liquid flow rate of the processing equipment in the processing cycle;

[0064] The surface error detection module analyzes the surface quality state of the workpiece within the processing cycle according to the equipment offset index α(j) and the peak-valley value of the workpiece within the processing cycle: if α(j)×PV<λ / η×(Nj) / N, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is normal; otherwise, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is abnormal, and constructs the workpiece surface quality index β(j), setting β(j)=exp{[α(j)×PV-λ / η] / (λ / η×(Nj) / N)}; wherein PV is the peak-valley value of the surface shape of the workpiece within the processing cycle and the standard surface shape, λ is the detection wavelength, and η is the error proportional constant; through the joint analysis of the equipment offset index (α(j)) and the peak-valley value (PV), the surface quality abnormality is accurately judged, and the surface quality index (β(j)) is constructed to achieve rapid defect positioning and reduce the defective rate.

[0065] Specifically, this embodiment does not impose any specific limitation on the value of the error proportional constant η, and those skilled in the art can freely set it as long as the value requirement of the error proportional constant η is met. In this embodiment, the optimal value of the error proportional constant η is 20; at the same time, the detection wavelength λ is the measurement wavelength for surface detection of the workpiece. In this embodiment, a helium-neon laser is used to measure the surface of the workpiece, and its detection wavelength is 632.8 nm; it can be understood that in this embodiment, when the surface quality state of the workpiece is normal, the default workpiece surface quality index is 0.

[0066] Please continue reading Figure 1 As shown, the system also includes an adaptive analysis module, which is connected to the surface error detection module. The adaptive analysis module is used to analyze the thermal conductivity state according to the environmental data within the processing cycle and optimize the analysis results of the surface quality state according to the analysis results.

[0067] See also Figure 3 As shown, the adaptive analysis module also includes a thermal conductivity anomaly analysis unit, which is used to analyze the thermal conductivity state according to the ambient temperature within the processing cycle: if t(j) < K, the thermal conductivity anomaly analysis unit determines that the thermal conductivity state of the workpiece within the processing cycle is normal and does not perform optimization; if t(j) ≥ K, the thermal conductivity anomaly analysis unit determines that the thermal conductivity state of the workpiece within the processing cycle is abnormal, and optimizes the device offset index to α(j)', setting α(j)'=α(j)×{1+[t(j)-K] / K}; where K is the preset ambient temperature; the device offset index (α(j)') is dynamically adjusted according to the ambient temperature to avoid processing errors caused by temperature fluctuations and enhance system stability.

[0068] Specifically, this embodiment does not impose any specific limitation on the value of the preset ambient temperature K. Those skilled in the art can freely set it as long as the value requirement of the preset ambient temperature K is met. In this embodiment, the optimal value of the preset ambient temperature K is 25°C.

[0069] Please continue reading Figure 3As shown, the adaptive analysis module also includes a bubble defect analysis unit, which is connected to the thermal conduction abnormality analysis unit. The bubble defect analysis unit is used to analyze the bubble risk according to the ambient humidity and the equipment vibration frequency during the processing cycle, and correct the analysis process of the thermal conduction state according to the analysis results: if s(j) < S and Zp(j) < YP, the bubble defect analysis unit determines that the bubble risk during the processing cycle is normal and does not perform correction; if s(j) ≥ S or Zp(j) ≥ YP, the bubble defect analysis unit determines that the bubble risk during the processing cycle is abnormal and corrects the preset ambient temperature to K', setting K'=K×{b1×[s(j)-S] / S+b2×[Zp(j)-YP] / YP};

[0070] Where b1 is the humidity weight, b2 is the vibration weight, b1+b2+1, s(j) is the ambient humidity during the j-th processing cycle, Zp(j) is the device vibration frequency during the j-th processing cycle, S is the ambient humidity threshold, and YP is the device vibration frequency threshold. The preset ambient temperature (K') is corrected in combination with humidity and vibration frequency to effectively suppress the risk of bubble defects and improve the optical uniformity of optical lenses.

[0071] It is worth noting that in this embodiment “K′=K×{b1×[s(j)-S] / S+b2×[Zp(j)-YP] / YP}”, when the value of “[s(j)-S] / S” or “[Zp(j)-YP] / YP” is less than 0, the value is 0.

[0072] Specifically, this embodiment does not impose specific restrictions on the values ​​of the ambient humidity threshold S and the device vibration frequency threshold YP. Those skilled in the art can set them freely as long as the value requirements of the ambient humidity threshold S and the device vibration frequency threshold YP are met. In this embodiment, the optimal value of the ambient humidity threshold S is 50%, and the optimal value of the device vibration frequency threshold YP is 5HZ.

[0073] Please continue reading Figure 1As shown, the system also includes a workpiece property detection module, which is connected to the historical data analysis module. The workpiece property detection module is used to analyze the optical state of the workpiece according to the optical property parameters and standard optical property parameters within the processing cycle, and construct a workpiece optical property index according to the analysis results: if a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)<G×(Nj) / N, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is normal; otherwise, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is abnormal, and constructs the workpiece optical property index γ(j). Let Define γ(j)=ln{e+{a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)-G×(Nj) / N} / (G×(Nj) / N)}; where Lm(j) is the rated scattering loss ratio in the j-th processing cycle, Lk(j) is the rated target transmittance in the j-th processing cycle, Lv(j) is the scattering loss ratio in the j-th processing cycle, Lt(j) is the detected transmittance in the j-th processing cycle, G is the optical property threshold, and e is the natural logarithm. The optical property index (γ(j)) is used to quantify the scattering loss and transmittance deviation, and the optical performance of the workpiece is evaluated in real time to ensure that the processing process meets the optical design standards.

[0074] Specifically, this embodiment does not impose any specific restrictions on the value of the optical property threshold G. Those skilled in the art can freely set it as long as the value requirements of the optical property threshold G are met. In this embodiment, the optimal value of the optical property threshold G is 0.6; it can be understood that in this embodiment, when the optical state is normal, the default workpiece optical property index is 0.

[0075] Please continue reading Figure 1 As shown, the system also includes a quality prediction module, which is connected to the workpiece property detection module and the surface error detection module. The quality prediction module is used to predict the offset state of the workpiece based on the workpiece surface quality index construction results and the workpiece optical property index construction results within the processing cycle, and to analyze the abnormal processing state of the workpiece based on the prediction results, and to alarm the user based on the analysis results of the abnormal processing state of the workpiece.

[0076] Please continue reading Figure 4 As shown, the quality prediction module includes an overall offset analysis unit, which is used to predict the offset state of the workpiece according to the workpiece surface quality index construction results and the workpiece optical property index construction results within the processing cycle;

[0077] The overall offset analysis unit counts the workpiece surface quality index and the workpiece optical property index of each processing cycle within the processing cycle, and predicts the offset state of the workpiece based on the statistical results. The specific process is as follows: If <D1, the overall offset analysis unit determines that the offset state of the workpiece is normal; if D1≤ <D2, the overall offset analysis unit determines that the offset state of the workpiece is orderly offset; if ≥D2, the overall offset analysis unit determines that the offset state of the workpiece is an over-limit offset;

[0078] Among them, D1 is the first preset offset coefficient, D2 is the second preset offset coefficient, D1<D2, c1 is the surface weight, c2 is the optical property weight, c1+c2=1; the surface quality index and the optical property index are combined to predict the offset status (normal, orderly offset, out-of-limit offset) and realize early warning of processing abnormalities.

[0079] Specifically, this embodiment does not impose specific limitations on the values ​​of the first preset offset coefficient D1 and the second preset offset coefficient D2. Those skilled in the art can set them freely, as long as the value requirements of the first preset offset coefficient D1 and the second preset offset coefficient D2 are met. In this embodiment, the optimal value of the first preset offset coefficient D1 is 0.1, and the optimal value of the second preset offset coefficient D2 is 0.3. It can be understood that in this embodiment, the slopes of the offsets are summed to determine whether the offsets are orderly throughout the entire processing cycle; at the same time, in this embodiment, the surface shape weight c1 and the optical property weight c2 are of equal weight, that is, c1=0.5, c2=0.5, and their values ​​can be freely set by those skilled in the art according to actual conditions; it is worth noting that the surface shape weight determines the ability of the optical lens to refract light, and the optical property weight determines the ability to transmit light and scatter light.

[0080] It can be understood that in this embodiment, "statisticing the workpiece surface quality index and workpiece optical property index of each processing cycle within the processing cycle" means: counting the workpiece surface quality index and workpiece optical property index of the processing cycles that have been specifically passed within the processing cycle, and statistics are not performed on the processing cycles that have not yet been processed.

[0081] Please continue reading Figure 4As shown, the quality prediction module also includes an alarm unit, which is connected to the overall offset analysis unit. The alarm unit is used to analyze the abnormal processing state of the workpiece according to the offset state prediction result of the workpiece, and to alarm the user according to the analysis result: if the offset state of the workpiece within the processing cycle is normal, the alarm unit determines that the processing of the workpiece is normal, and does not alarm the user; if the offset state of the workpiece within the processing cycle is an ordered offset, the alarm unit determines that the abnormal processing state of the workpiece is a fine-tunable abnormality, and sends a secondary alarm to the user; if the offset state of the workpiece within the processing cycle is an over-limit offset, the alarm unit determines that the abnormal processing state of the workpiece is a serious offset abnormality, and sends a primary alarm to the user; the hierarchical alarm mechanism (primary and secondary) distinguishes the severity of the abnormality, guides the operator to respond quickly, and avoids waste of resources.

[0082] Specifically, the second level alarm and the first level alarm described in this embodiment indicate the severity level of the processing error, and the first level indicates an error that is so serious that processing cannot be continued.

[0083] Please continue reading Figure 1 As shown, the system further includes a processing management module, which is connected to the quality prediction module. The processing management module is used to optimize the processing parameters of the next processing cycle according to the analysis result of the processing abnormality state of the workpiece within the processing cycle: if the processing abnormality state of the workpiece is normal, the processing management module does not change the processing parameters of the next processing cycle; if the processing abnormality state of the workpiece is an ordered offset, the processing management module optimizes the processing parameters of the next processing cycle to P(j+1)', R(j+1)' and Q(j+1)'; if the processing abnormality state of the workpiece is an over-limit offset, the processing management module regards the workpiece as an unqualified workpiece;

[0084] Among them, set:

[0085] P(j+1)'=P(j+1)+ΔP; ΔP=P(j+1) / PV;

[0086] R(j+1)'=R(j+1)+ΔR; ΔR=R(j+1)×γ(j)-R(j+1);

[0087] Q(j+1)'=Q(j+1)+ΔQ; ΔQ=0.05×[Kt(j)];

[0088] The quality prediction module outputs the optimization results to the user; dynamically optimizes the processing parameters (such as ΔP, ΔQ) of the next processing cycle based on the abnormal state, or marks unqualified workpieces to achieve closed-loop control, significantly improving processing efficiency and yield rate.

[0089] It is worth noting that the prerequisite for optimizing the processing parameters of the next processing cycle in this embodiment is that the current processing cycle is not the last processing cycle.

[0090] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. An AI-based workpiece processing data intelligent analysis system, characterized in that: include: Workpiece Data acquisition module, used to collect workpiece design data; Multi-source data acquisition module, used to collect workpiece processing history data, as well as workpiece processing data and environmental data within the processing cycle; A historical data analysis module is used to set the processing cycle of the workpiece and the processing parameters of each processing cycle within the processing cycle according to the workpiece processing history data, and to obtain the standard surface shape and standard optical property parameters of each processing cycle; A surface error detection module is used to construct a workpiece surface quality index based on the processing parameters of each processing cycle within the processing cycle and the workpiece processing data within the processing cycle; a workpiece property detection module for constructing a workpiece optical property index based on optical property parameters within a processing cycle and standard optical property parameters; A quality prediction module is used to predict the deviation state of the workpiece based on the workpiece surface quality index and the workpiece optical property index during the processing cycle, and analyze the abnormal processing state of the workpiece based on the prediction results; A processing management module is used to optimize the processing parameters of the next processing cycle based on the analysis results of the abnormal processing status of the workpiece during the processing cycle; Also included is an adaptive analysis module, the adaptive analysis module is used to analyze the thermal conductivity state according to the environmental data in the processing cycle, and to adjust the analysis result of the surface quality state according to the analysis result; The adaptive analysis module further includes a thermal conductivity anomaly analysis unit, which is configured to analyze the thermal conductivity state according to the ambient temperature during the processing cycle to determine whether the thermal conductivity state is normal, and adjust the device offset index to α(j)' when the thermal conductivity state is abnormal, setting α(j)'=α(j)×{1+[t(j)-K] / K}; wherein K is a preset ambient temperature; The adaptive analysis module further includes a bubble defect analysis unit, which is used to analyze the bubble risk according to the ambient humidity and the equipment vibration frequency during the processing cycle to determine whether the bubble risk is normal, and correct the preset ambient temperature to K' when the bubble risk is abnormal; If s(j) < S and Zp(j) < YP, the bubble defect analysis unit determines that the bubble risk within the processing period is normal and does not perform correction; if s(j) ≥ S or Zp(j) ≥ YP, the bubble defect analysis unit determines that the bubble risk within the processing period is abnormal and corrects the preset ambient temperature to K', setting K'=K×{b1×[s(j)-S] / S+b2×[Zp(j)-YP] / YP}; Where b1 is the humidity weight, b2 is the vibration weight, b1+b2+1, s(j) is the ambient humidity during the j-th processing cycle, Zp(j) is the device vibration frequency during the j-th processing cycle, S is the ambient humidity threshold, and YP is the device vibration frequency threshold.

2. The AI-based workpiece processing data intelligent analysis system according to claim 1 is characterized in that: The historical data analysis module includes a processing cycle analysis unit, and the processing cycle analysis unit is used to set the processing cycle of the workpiece according to the processing history data of the workpiece; The processing cycle analysis unit counts the processing cycle length of historical workpieces and calculates the weighted processing cycle T of historical workpieces. ; Where T(i) is the processing cycle time of the i-th historical workpiece, Lv(i) is the detection scattering loss ratio of the i-th historical workpiece, Lt(i) is the detection transmittance of the i-th historical workpiece, LV is the target scattering loss ratio, LT is the target transmittance, a1 is the scattering weight, a2 is the transmittance weight, a1+a2=1, and N is the number of historical workpieces; The processing cycle analysis unit uses the weighted processing cycle T of the historical workpiece as the processing cycle of the workpiece.

3. The AI-based workpiece processing data intelligent analysis system according to claim 2 is characterized in that: The historical data analysis module further includes a processing parameter analysis unit, which is used to set the processing parameters of the workpiece in each processing cycle according to the processing data of the workpiece; The processing parameter analysis unit sets the rated polishing pressure P(j) of each processing cycle according to the polishing pressure of each historical workpiece, and sets ; The processing parameter analysis unit sets the rated spindle speed R(j) of each processing cycle according to the spindle speed of each historical workpiece, and sets ; The processing parameter analysis unit sets the rated polishing liquid flow rate Q(j) for each processing cycle according to the polishing liquid flow rate of each historical workpiece, and sets ; Wherein, P(j), R(j) and Q(j) are the rated polishing pressure, rated spindle speed and rated polishing fluid flow rate of the jth processing cycle within the machining cycle, respectively; P(j,i) is the polishing pressure of the i-th historical workpiece in the jth processing cycle; R(j,i) is the spindle speed of the i-th historical workpiece in the jth processing cycle; Q(j,i) is the polishing fluid flow rate of the i-th historical workpiece in the jth processing cycle; PV(j,i) is the peak-to-valley value of the surface shape of the i-th historical workpiece in the jth processing cycle compared with the standard surface shape; The processing parameter analysis unit is further used to store the standard surface shape and standard optical property parameters of each processing cycle.

4. The AI-based workpiece processing data intelligent analysis system according to claim 3 is characterized in that: The surface error detection module extracts the processing parameters in the current processing cycle in the processing cycle, and calculates the equipment deviation index α(j), setting α(j)=ln{1+[PP(j)] / P(j)×{[RR(j)] / R(j)}×{[QQ(j)] / Q(j)}}; wherein P is the polishing pressure of the processing equipment in the processing cycle, R is the spindle speed of the processing equipment in the processing cycle, and Q is the polishing liquid flow rate of the processing equipment in the processing cycle; The surface error detection module analyzes the surface quality state of the workpiece within the processing cycle according to the equipment offset index α(j) and the peak-to-valley value of the workpiece within the processing cycle: if α(j)×PV<λ / η×(Nj) / N, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is normal; otherwise, the surface error detection module determines that the surface quality state of the workpiece within the processing cycle is abnormal, and constructs the workpiece surface quality index β(j), setting β(j)=exp{[α(j)×PV-λ / η] / (λ / η×(Nj) / N)}; wherein PV is the peak-to-valley value of the surface shape of the workpiece within the processing cycle and the standard surface shape, λ is the detection wavelength, and η is the error proportional constant.

5. The AI-based workpiece processing data intelligent analysis system according to claim 4 is characterized in that: The workpiece property detection module analyzes the optical state of the workpiece according to the optical property parameters within the processing cycle and the standard optical property parameters, and constructs the workpiece optical property index according to the analysis results: if a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)<G×(Nj) / N, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is normal; otherwise, the workpiece property detection module determines that the optical state of the workpiece within the processing cycle is abnormal, and constructs the workpiece optical property index γ (j), set γ(j)=ln{e+{a1×[Lv(j)-Lm(j)] / Lm(j)+a2×[Lt(j)-Lk(j)] / Lk(j)-G×(Nj) / N} / (G×(Nj) / N)}; where Lm(j) is the rated scattering loss ratio in the j-th processing cycle, Lk(j) is the rated target transmittance in the j-th processing cycle, Lv(j) is the scattering loss ratio in the j-th processing cycle, Lt(j) is the detected transmittance in the j-th processing cycle, G is the optical property threshold, and e is the natural logarithm.

6. The AI-based workpiece processing data intelligent analysis system according to claim 5 is characterized in that: The quality prediction module includes an overall offset analysis unit, which is used to predict the offset state of the workpiece based on the workpiece surface quality index construction results and the workpiece optical property index construction results within the processing cycle; The overall offset analysis unit counts the workpiece surface quality index and the workpiece optical property index of each processing cycle within the processing cycle, and predicts the offset state of the workpiece based on the statistical results. The specific process is as follows: If <D1, the overall offset analysis unit determines that the offset state of the workpiece is normal; if D1≤ <D2, the overall offset analysis unit determines that the offset state of the workpiece is orderly offset; if ≥D2, the overall offset analysis unit determines that the offset state of the workpiece is an over-limit offset; Wherein, D1 is the first preset offset coefficient, D2 is the second preset offset coefficient, D1<D2, c1 is the surface shape weight, c2 is the optical property weight, and c1+c2=1.

7. The AI-based workpiece processing data intelligent analysis system according to claim 6 is characterized in that: The quality prediction module also includes an alarm unit, which is used to analyze the abnormal processing state of the workpiece based on the prediction result of the offset state of the workpiece, and to alarm the user based on the analysis result: if the offset state of the workpiece within the processing cycle is normal, the alarm unit determines that the processing of the workpiece is normal and does not alarm the user; if the offset state of the workpiece within the processing cycle is an ordered offset, the alarm unit determines that the abnormal processing state of the workpiece is a fine-tunable abnormality and sends a second-level alarm to the user; if the offset state of the workpiece within the processing cycle is an out-of-limit offset, the alarm unit determines that the abnormal processing state of the workpiece is a serious offset abnormality and sends a first-level alarm to the user.

8. The AI-based workpiece processing data intelligent analysis system according to claim 7 is characterized in that: The processing management module optimizes the processing parameters of the next processing cycle according to the analysis results of the processing abnormality state of the workpiece within the processing cycle: if the processing abnormality state of the workpiece is normal, the processing management module does not change the processing parameters of the next processing cycle; if the processing abnormality state of the workpiece is an ordered offset, the processing management module optimizes the processing parameters of the next processing cycle to P(j+1)', R(j+1)' and Q(j+1)'; if the processing abnormality state of the workpiece is an over-limit offset, the processing management module regards the workpiece as an unqualified workpiece; Among them, set: P(j+1)'=P(j+1)+ΔP; ΔP=P(j+1) / PV; R(j+1)'=R(j+1)+ΔR; ΔR=R(j+1)×γ(j)-R(j+1); Q(j+1)’=Q(j+1)+ΔQ;ΔQ=0.05×[K-t(j)]。

Images