Temperature Anomaly Monitoring System for the Processing Area of ​​Aluminum-based Silicon Carbide Ultrathin Reflectors

Abstract

This invention relates to the field of anomaly monitoring technology, specifically to a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin reflector. It includes: a thermal imaging sensing and processing unit that continuously acquires thermal images of the processing area of ​​the aluminum-based silicon carbide ultrathin reflector and outputs a temperature field; and a processing area adaptive positioning unit that calculates the temperature spatial gradient based on the temperature field, and, combined with a temperature spatial gradient threshold and a minimum connected area threshold, identifies the processing area where the grinding tool and workpiece interact thermally, while simultaneously outputting a processing area spatial mask. This invention introduces a curvature-sensing directional radiation correction mechanism based on second-order temperature field changes and a dual-channel modeling method for heat load transfer modes, accurately identifying the actual processing area and constructing a reference temperature response model that conforms to physical mechanisms without relying on external geometric models or cooling parameters.

Description

Technical Field

[0001] This invention relates to the field of anomaly monitoring technology, and more specifically, to a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin mirror. Background Technology

[0002] In the precision grinding process of aluminum-based silicon carbide (Al / SiC) ultrathin mirrors, due to the material's high hardness, low thermal conductivity, and non-uniform composite structure, the grinding area is prone to transient high temperatures or thermal imbalances caused by localized frictional heat concentration, uneven material removal, or abnormal tool conditions. These thermal anomalies can induce microcracks, residual stress accumulation, and geometric deformation on the workpiece surface, and in severe cases, even lead to the failure of the mirror's optical surface accuracy, directly affecting its performance in high-end optical systems such as aerospace and astronomy. Traditional temperature monitoring methods mostly rely on contact thermocouples or fixed-threshold infrared temperature measurement, which are difficult to achieve continuous, non-contact, and high spatial resolution temperature field sensing in high-speed dynamic processing areas. Furthermore, existing thermal imaging systems are susceptible to interference from surface curvature, changes in observation angle, and environmental thermal drift when dealing with the complex curved surfaces, strong directional radiation characteristics, and small thermally affected areas of ultrathin mirrors. This can lead to inaccurate positioning of the processing area, distortion of thermal response feature extraction, and consequently, misjudgments or missed detections. Furthermore, conventional alarm mechanisms lack the ability to model the physical evolution mechanism of processing thermal loads, cannot distinguish between normal process fluctuations and actual abnormal thermal behavior, and are insufficient to support highly reliable real-time anomaly alarms. Therefore, a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin mirror is designed. Summary of the Invention

[0003] The purpose of this invention is to provide a temperature anomaly monitoring system for the processing area of ​​aluminum-based silicon carbide ultrathin reflective mirrors, in order to solve the problem of uncontrolled processing quality and false alarms caused by the difficulty in accurately sensing, locating and judging thermal anomalies during the precision grinding of aluminum-based silicon carbide ultrathin reflective mirrors as mentioned in the background art.

[0004] To achieve the above objectives, the present invention aims to provide a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin mirror, comprising:

[0005] A thermal imaging sensing, acquisition, and processing unit performs continuous thermal imaging acquisition of the processing area of ​​the aluminum-based silicon carbide ultrathin reflector and outputs the temperature field.

[0006] The adaptive positioning unit for the processing area calculates the temperature spatial gradient based on the temperature field and combines the temperature spatial gradient threshold and the minimum connected area threshold to identify the processing area where the grinding tool and the workpiece are thermally interacting, and outputs a processing area spatial mask.

[0007] Among them, the process of identifying the processing area is optimized by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field;

[0008] A temperature spatiotemporal feature construction unit, which is based on a processing area spatial mask, extracts the thermal response feature vector of processing heat generated by the contact friction between the grinding tool and the workpiece surface, material removal and local plastic deformation through joint analysis of the time and spatial dimensions of the temperature field.

[0009] The material thermal response modeling unit establishes a mapping model between the processing thermal load and the temperature evolution behavior of the aluminum-based silicon carbide ultrathin mirror based on the thermal response feature vector, which is used to output the reference temperature response.

[0010] The temperature anomaly detection and alarm unit compares the real-time acquired thermal response feature vector with the reference temperature response, calculates the real-time temperature deviation index based on the processing area spatial mask, and performs weighted fusion of the temperature deviation index to form a comprehensive temperature anomaly index, while simultaneously issuing an anomaly alarm for anomaly monitoring.

[0011] As a further improvement to this technical solution, the thermal imaging sensing acquisition and processing unit includes a thermal imaging sensing acquisition module and a thermal imaging data preprocessing module.

[0012] The thermal imaging sensing and acquisition module performs continuous infrared thermal imaging acquisition on the processing area of ​​the aluminum-based silicon carbide ultrathin reflector to obtain a time-series thermal image sequence.

[0013] The thermal imaging data preprocessing module preprocesses the time-series thermal image sequence to form a two-dimensional temperature field in a unified coordinate system. .

[0014] As a further improvement to this technical solution, the adaptive positioning unit for the machining area calculates the spatial temperature gradient based on the temperature field and identifies the machining area where the tool and workpiece interact thermally, including the following steps:

[0015] S1.1, Based on two-dimensional temperature field At every moment Below, the first-order spatial gradient components of temperature are calculated along two orthogonal directions in space to characterize the local variation trend of temperature in the horizontal and vertical directions;

[0016] S1.2 Calculation of temperature gradient amplitude based on first-order spatial gradient components To characterize the intensity of temperature change within a unit spatial scale;

[0017] Specifically, by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field, the temperature gradient amplitude is adjusted. Directional radiation correction is performed to generate the corrected temperature gradient amplitude. This is used to reduce the interference of curvature and observation angle changes on the identification of the processing area;

[0018] S1.3, Temperature gradient amplitude With the preset gradient threshold The comparisons are performed, and a binary map of the candidate thermal impact region is constructed as the candidate thermal impact region image.

[0019] S1.4 Perform connected component analysis on the candidate thermal effect region image to extract candidate regions that are spatially continuous.

[0020] S1.5 Calculate the spatial area of ​​each candidate region, based on the preset minimum connected area threshold. Filter out spaces with a spatial area greater than the minimum connected area threshold. Candidate regions are formed to create a set of effective thermal action regions;

[0021] S1.6 From the set of effective heat-affected areas, determine and identify the processing areas where the tool and workpiece experience thermal interaction based on the area temperature characteristics and temporal continuity, and output them in the form of a processing area spatial mask.

[0022] As a further improvement to this technical solution, in S1.2, a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field is introduced to correct the temperature gradient amplitude. Directional radiation correction includes the following steps:

[0023] S1.21, Based on two-dimensional temperature field The second-order partial derivatives are calculated along two orthogonal directions in space, and a pixel-level temperature curvature surrogate is constructed. ;

[0024] S1.22. Based on the coupling relationship between thermal radiation directionality and observation angle on temperature measurement results, pixel-level temperature curvature proxy quantity is used. Normalized and mapped to equivalent observation angle bias factor ;

[0025] S1.23. Combining the equivalent observation angle deviation, construct the directional radiation compensation weight function and generate the directional radiation compensation weight. ;

[0026] S1.24, Using directional radiation to compensate for weights For the original temperature gradient magnitude Perform directional radiation correction to obtain the corrected temperature gradient amplitude. .

[0027] As a further improvement to this technical solution, in S1.22, based on the coupling relationship between the thermal radiation directionality and the observation angle on the temperature measurement results, the pixel-level temperature curvature proxy quantity is... Normalized and mapped to equivalent observation angle bias factor This includes the following steps:

[0028] Pixel-level temperature curvature proxy Steady-state constraint normalization is performed within its local neighborhood to generate a normalized pixel-level temperature curvature surrogate quantity. ;

[0029] A directional response function is constructed based on the directionality of thermal radiation and the observation angle;

[0030] Mapping the directional response function to an equivalent observation angle deviation factor. .

[0031] As a further improvement to this technical solution, the temperature spatiotemporal feature construction unit extracts a thermal response feature vector characterizing the processing thermal behavior through joint analysis of the temporal and spatial dimensions of the temperature field, including the following steps:

[0032] S2.1 Receive the processing area spatial mask and apply the processing area spatial mask to the temperature field. This yields a local temperature field that includes only the processing area. ;

[0033] S2.2, Based on local temperature field For each time period, the spatial temperature characteristics are calculated based on the spatial mask of the processing area, including at least: the average temperature of the processing area, the peak temperature of the processing area, the standard deviation of the temperature of the processing area, and the mean temperature gradient of the processing area.

[0034] S2.3. Based on the extraction of temporal dimension features from the spatial mask of the processing area, at least the following should be included: temperature rise rate, peak temperature change rate, temperature fluctuation amplitude, and duration of local temperature change.

[0035] S2.4. The extracted spatial temperature features and temporal features are normalized and combined in chronological order to form a thermal response feature vector characterizing the processing thermal behavior. Among them, the thermal response feature vector is used to characterize the equivalent time sequence of the processing thermal load on the material surface formed by the combined effects of contact friction between the grinding tool and the workpiece, material removal, and local heat accumulation during the processing.

[0036] As a further improvement to this technical solution, the material thermal response modeling unit establishes a mapping model between the processing thermal load and the temperature evolution behavior of the aluminum-based silicon carbide ultrathin mirror based on the thermal response feature vector, including the following steps:

[0037] S3.1. Time-align the thermal response feature vector and use a fixed time window length. and sliding step size The thermal response feature vectors at consecutive time points are organized into a window to form a thermal response feature sequence;

[0038] S3.2 Based on the heat load transfer modes in the material, the thermal response feature sequence is structurally reconstructed by a dual-channel modeling method for processing temperature evolution based on heat load transfer mode decomposition, and used as the input variable of the mapping model.

[0039] S3.3 Collect n sets of complete processing data, obtain the corresponding thermal response characteristic sequence and the actual measured temperature evolution sequence of the processing area for each set of processing data, and construct a normal processing thermal response sample set according to the time window.

[0040] S3.4. Based on the normal processing thermal response sample set, a mapping model between the thermal response feature sequence and the temperature evolution behavior of the processing area is established by using the channel-weighted regression method, which is used to output the reference temperature response under a given processing thermal load condition.

[0041] As a further improvement to this technical solution, in step S3.2, the thermal response feature sequence is structurally reconstructed using a dual-channel modeling method for processing temperature evolution based on thermal load transfer mode decomposition, including the following steps:

[0042] S3.21. In each thermal response characteristic sequence, calculate the spatial diffusion consistency index and the time decay dominance index;

[0043] S3.22. Based on the spatial diffusion consistency index and the time decay dominance index, construct the modal proportion vector corresponding to the heat load transfer path. ;

[0044] S3.23. For the same thermal response characteristic sequence, based on the modal proportion vector Two parallel characteristic sequences of heat load transfer channels were constructed: a diffusion-dominant channel characteristic sequence and a heat transfer-dominant channel characteristic sequence.

[0045] S3.24. Use the characteristic sequences of diffusion-dominant channels and heat transfer-dominant channels as input variables for the mapping model.

[0046] As a further improvement to this technical solution, in S3.22, a modal proportion vector corresponding to the heat load transfer path is constructed based on the spatial diffusion consistency index and the time decay dominance index. This includes the following steps:

[0047] Regarding the first The thermal response characteristic sequences, respectively, affect the spatial diffusion consistency index. and time decay dominant index Perform dimensionless normalization;

[0048] Based on the normalized spatial diffusion consistency index and time decay dominance index, diffusion dominance tendency function and heat transfer dominance tendency function are constructed to reflect the dominance of different heat load transfer mechanisms.

[0049] Normalizing the diffusion-dominant and heat transfer-dominant tendency functions yields the modal weights corresponding to the heat load transfer paths: the proportion of the thermal diffusion-dominant mode within the material. The proportion of surface heat transfer dominant modes ;

[0050] The proportion of the dominant mode of internal thermal diffusion in the material The proportion of surface heat transfer dominant modes The modal proportion vector of the heat load transfer path is formed by combining these modal proportions. .

[0051] As a further improvement to this technical solution, the temperature anomaly detection and alarm unit includes a temperature anomaly detection module and a temperature anomaly alarm module;

[0052] The temperature anomaly detection module calculates multiple temperature deviation indices based on the processing area spatial mask, real-time thermal response feature vector, and reference temperature response, including regional average temperature deviation, regional peak temperature deviation, and temperature change rate deviation. It normalizes the multiple temperature deviation indices and constructs a comprehensive temperature anomaly index using a weighted fusion method.

[0053] The temperature anomaly alarm module compares the comprehensive temperature anomaly index with a preset threshold to classify the processing temperature anomaly and issues an alarm based on the classification result.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0055] 1. This invention relates to a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin reflector. By introducing a "curvature-sensing directional radiation correction mechanism based on second-order temperature field changes" and a "dual-channel modeling method for heat load transfer modes," the system accurately identifies the actual processing area and constructs a reference temperature response model that conforms to physical mechanisms without relying on external geometric models or cooling parameters. Based on this, the temperature anomaly discrimination module integrates multi-dimensional deviation indicators to generate a comprehensive anomaly index and implements graded alarms according to preset thresholds. This mechanism avoids false alarms caused by surface curvature, observation angle, or heat transfer mode switching, ensuring that the alarm signal truly reflects the abnormal processing state.

[0056] 2. This invention relates to a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin reflector. The system classifies the severity of anomalies into four levels (normal, mild, moderate, and severe) through a temperature anomaly alarm module, and triggers different priority audible and visual, interface, and remote communication alarms. Especially for high-value, heat-deformable workpieces like aluminum-based silicon carbide ultrathin reflectors, this graded alarm mechanism can trigger low-disturbance alerts in the early stages of thermal anomalies and initiate emergency intervention when there is a severe risk of overheating. This ensures processing quality while preventing equipment damage, achieving an intelligent monitoring upgrade from "passive response" to "proactive prevention." Attached Figure Description

[0057] Figure 1 This is an overall flowchart of the present invention;

[0058] The meanings of the labels in the diagram are as follows:

[0059] 1. Thermal imaging sensing and acquisition unit; 2. Adaptive positioning unit for processing area; 3. Temperature spatiotemporal feature construction unit; 4. Material thermal response modeling unit; 5. Temperature anomaly detection and alarm unit. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example: Please refer to Figure 1 As shown, a temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin mirror is provided, comprising:

[0062] The thermal imaging sensing, acquisition and processing unit 1 performs continuous thermal imaging acquisition of the processing area of ​​the aluminum-based silicon carbide ultrathin mirror and outputs the temperature field.

[0063] In this embodiment, the thermal imaging sensing acquisition and processing unit 1 includes a thermal imaging sensing acquisition module and a thermal imaging data preprocessing module.

[0064] The thermal imaging sensing and acquisition module performs continuous infrared thermal imaging acquisition on the processing area of ​​the aluminum-based silicon carbide ultrathin reflector to obtain a time-series thermal image sequence (the processing area here is around the nominal processing coverage area (observation area, ROI initial area) of the aluminum-based silicon carbide ultrathin reflector, which is determined in advance by the equipment installation position, camera field of view, and processing table size, and is essentially the observation domain of thermal imaging).

[0065] The thermal imaging data preprocessing module preprocesses the time-series thermal image sequence. This preprocessing first performs radiometric correction on the continuously acquired raw infrared thermal images, converting pixel grayscale values ​​into corresponding apparent radiance based on the thermal imaging camera's calibration parameters and radiative transfer model, and further converting them into absolute temperature values. Then, combining ambient temperature sensing information and a background reference area, environmental temperature drift compensation is performed on the thermal image sequence to eliminate the influence of slow ambient temperature changes and equipment thermal drift on the temperature measurement results. Based on this, spatial registration processing is performed on the thermal images at each time point, eliminating pixel shifts caused by camera micro-vibration or workpiece micro-displacement through sub-pixel-level geometric alignment, ensuring consistent correspondence between different time frames in a unified spatial coordinate system. Finally, the registered temperature data undergoes temporal consistency verification and outlier suppression processing to remove instantaneous noise and isolated abnormal pixels, outputting a two-dimensional temperature field in a unified coordinate system, forming a two-dimensional temperature field in a unified coordinate system. .

[0066] The adaptive positioning unit 2 for the processing area calculates the temperature spatial gradient based on the temperature field, and combines the temperature spatial gradient threshold and the minimum connected area threshold to identify the processing area where the grinding tool and the workpiece are thermally interacting, while outputting a spatial mask of the processing area.

[0067] Among them, the process of identifying the processing area is optimized by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field;

[0068] In this embodiment, the adaptive positioning unit (2) of the machining area calculates the spatial temperature gradient based on the temperature field and identifies the machining area where the tool and the workpiece interact thermally, including the following steps:

[0069] S1.1, Based on two-dimensional temperature field At every moment Below, the first-order spatial gradient components of temperature are calculated along two orthogonal directions in space: (temperature field in) First-order partial derivatives in the direction), (temperature field in) The first-order partial derivatives in the direction are used to characterize the local variation trend of temperature in the horizontal and vertical directions;

[0070] S1.2 Calculation of temperature gradient amplitude based on first-order spatial gradient components To characterize the intensity of temperature change within a unit spatial scale;

[0071] Specifically, by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field, the temperature gradient amplitude is adjusted. Directional radiation correction is performed to generate the corrected temperature gradient amplitude. This is used to reduce the interference of curvature and observation angle changes on the identification of the machining area and improve the positioning accuracy of the actual thermal action area between the tool and the workpiece.

[0072] Furthermore, the curvature-sensing directional radiation correction mechanism based on second-order temperature field changes addresses the apparent temperature deviation caused by variations in the micro-curvature of the aluminum-based silicon carbide ultrathin mirror surface and the observation angle during thermal imaging monitoring. In precision grinding, the mirror surface exhibits minute geometric undulations, and the viewing angle of the infrared imaging sensor changes with the position of the machining table or camera. This results in a systematic deviation in the pixel temperature measured at the same real temperature due to directional radiation, which interferes with the accurate positioning of the machining area and the extraction of thermal response features. This correction mechanism utilizes the curvature information reflected by the second-order temperature field change, combined with the coupling relationship between directional radiation and the observation angle, to correct the original temperature gradient. This reduces the interference of surface micro-geometric undulations and changes in the observation angle on the identification of the machining area and subsequent temperature anomaly detection, improving the positioning accuracy and monitoring reliability of the temperature anomaly monitoring system for the actual thermal interaction area between the tool and the workpiece.

[0073] The core innovation of this method lies in the curvature proxy mechanism without an external geometric model, namely, the curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field: by directly utilizing the second-order change characteristics of the temperature field, the influence of the curvature of the processed surface on the infrared radiation temperature measurement results is implicitly characterized; specifically, the second-order change of the temperature field is regarded as the radiation deviation response caused by the micro-geometric undulations of the surface, and the deviation is expressed by equivalent angle through a nonlinear mapping function, and then a directional compensation weight based on the cosine power function is constructed to directly correct the temperature gradient characteristics, thereby achieving adaptive suppression of the curvature and directional radiation coupling error without explicitly obtaining the workpiece geometry or observation posture information;

[0074] By introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field, the temperature gradient amplitude is adjusted. Directional radiation correction includes the following steps:

[0075] S1.21, Based on two-dimensional temperature field Calculate the second-order partial derivatives along two orthogonal directions in space: , , (In the formula, For the temperature field in The second partial derivative in the horizontal direction represents the degree of change in the rate of temperature change along the horizontal direction. For the temperature field in The second partial derivative in the vertical direction represents the degree of change in the rate of temperature change along the vertical direction. For the temperature field about and The mixed second-order partial derivative is used to describe the coupled variation characteristics of temperature in two orthogonal spatial directions, reflecting the oblique curvature or asymmetric variation of temperature distribution in space. For time, This represents the spatial coordinate variable of the thermal imaging image in the horizontal direction, corresponding to the column direction of the image or the horizontal position in actual physical space. For the spatial coordinates of the thermal imaging image in the vertical direction (corresponding to the row direction of the image or the vertical position in actual physical space), a pixel-level temperature curvature surrogate is constructed. ;

[0076] in, Used to reflect the second-order response intensity of the surface micro-geometric undulations of the processing area to the temperature field distribution;

[0077] S1.22, The coupling relationship between thermal radiation directionality and observation angle on temperature measurement results (i.e. (Item), will be used to measure pixel-level temperature curvature. Normalized and mapped to equivalent observation angle bias factor ;

[0078] In infrared thermal imaging temperature measurement, the radiance measured by the pixel is not only related to the actual temperature of the measured surface, but also closely related to the directional radiation characteristics of the material surface and the observation angle of the camera relative to the surface. For the processing area of ​​the aluminum-based silicon carbide ultrathin mirror, its surface exhibits non-Lambertian radiation characteristics during precision grinding, resulting in differences in effective emissivity in different emission directions. When the observation angle changes, the angle between the surface normal and the imaging optical axis changes, thereby causing a change in the effective radiation energy received by the pixel, resulting in a systematic deviation in the apparent temperature measured when the actual temperature remains constant. Therefore, the temperature measurement result can be regarded as the coupled result of the actual temperature, directional emissivity, and observation angle, and this coupling relationship constitutes the physical basis for subsequent radiation correction and temperature compensation.

[0079] Furthermore, based on the coupling relationship between the directionality of thermal radiation and the observation angle on the temperature measurement results, pixel-level temperature curvature proxy is used. Normalized and mapped to equivalent observation angle bias factor This includes the following steps:

[0080] Pixel-level temperature curvature proxy Steady-state constraint normalization is performed within its local neighborhood to generate a normalized pixel-level temperature curvature surrogate quantity. (This step is used to suppress the interference of thermal noise and curvature fluctuations in the non-processed area on subsequent angle mapping, where,) Within the neighborhood The mean, Within the neighborhood standard deviation To prevent stable terms with a denominator of zero;

[0081] Based on the directionality of thermal radiation and the observation angle, a curvature-driven directional response function is constructed (this function ensures weak angular perturbation in the low curvature region and gradually enhances directional offset in the high curvature region, which is used to characterize the comprehensive influence of the equivalent observation angle change caused by the geometric undulation of the surface of the processing area on the infrared radiation directionality temperature measurement results, thereby establishing an implicit coupling relationship between the second-order change of the temperature field and the temperature measurement angle deviation):

[0082] ;

[0083] in, The curvature-directivity coupling sensitivity coefficient is used to adjust the response intensity of curvature changes to radiation directionality shifts. Its value ranges from 0.5 to 3.0 and is determined through expert experience.

[0084] Mapping the directional response function to an equivalent observation angle deviation factor. :

[0085] ;

[0086] In the formula, The maximum allowable equivalent angle deviation for observations is used to limit the physical rationality of the mapping results. Its value ranges from 45° to 60° and is determined by expert experience.

[0087] S1.23. Combining the equivalent observation angle deviation, construct the directional radiation compensation weight function and generate the directional radiation compensation weight. Here, the directional radiation thermometry deviation caused by the curvature of the processing surface and the change of the observation angle is explicitly transformed into a weighting factor that can be used for feature correction. By giving stronger or weaker compensation intensity to pixels that are more affected by the angle, the original temperature gradient features are directionally adjusted.

[0088] in, This is the directional radiation sensitivity adjustment coefficient, used to control the influence of curvature on the radiation compensation intensity. Its value ranges from 0.5 to 2.5 and is determined through expert experience.

[0089] S1.24, Using directional radiation to compensate for weights For the original temperature gradient magnitude Perform directional radiation correction to obtain the corrected temperature gradient amplitude. ;

[0090] S1.3, Temperature gradient amplitude With the preset gradient threshold Compare (when satisfied) When the corresponding pixel is marked as a gradient saliency point, a binary map of candidate thermal regions is constructed (for regions that satisfy the following conditions). The pixels marked as potential thermally affected pixels, i.e. gradient salient points, and the remaining pixels marked as non-thermally affected pixels, form a binary map of candidate thermally affected regions, which serves as the candidate thermally affected region image;

[0091] S1.4. Perform connected component analysis on the candidate thermal impact region image to extract spatially continuous candidate regions. Specifically, at each time step, select the binary image of the candidate thermal impact region as input. Based on the preset pixel connectivity rules (using the eight-neighbor connectivity criterion to ensure the integrity of the diagonally continuous thermal distribution), scan and label all points marked as thermal impact pixels pixel by pixel. Then, through the connectivity search algorithm, spatially adjacent thermal impact pixels are automatically merged into the same connected component, and a unique region identifier is assigned to each connected component. During this process, the pixel set contained in each connected component and its spatial boundary information are recorded simultaneously, thereby decomposing the original binary candidate thermal impact region image into several spatially continuous and mutually distinguishable candidate regions, providing structured input for subsequent region area calculation, feature evaluation, and effective processing region selection.

[0092] S1.5 Calculate the spatial area of ​​each candidate region, based on the preset minimum connected area threshold. (10~50 pixels (corresponding to an actual area of ​​approximately)) ), filtering out spaces with an area greater than the minimum connected area threshold. Candidate regions are selected, and only connected regions that satisfy the area constraint are retained to form a set of effective thermal action regions. Specifically, for each candidate connected region, the number of effective pixels contained within it is first counted, and then, combined with the spatial resolution of the thermal imaging system or the physical size of the pixels, the number of pixels is converted into the corresponding spatial area to characterize the actual coverage range of the candidate region within the processing area. Subsequently, the calculated spatial area of ​​the candidate region is compared with a preset minimum connected area threshold. For comparison, for spaces with an area greater than The connected regions are identified as invalid regions caused by thermal noise, local transient disturbances, or non-processing heat sources and are removed. Only connected regions with a spatial area greater than the minimum connected area threshold are retained. Through the above area constraint screening process, a set of effective thermal action regions consisting of multiple connected regions that meet the requirements of spatial continuity and minimum area is finally formed, which serves as the candidate input for subsequent processing region uniqueness determination and temporal consistency analysis.

[0093] S1.6 From the set of effective thermal impact regions, based on the regional temperature characteristics and temporal continuity, determine and identify the processing regions where the tool and workpiece undergo thermal interaction, and output them in the form of a processing region spatial mask (used to limit the calculation range of thermal response characteristics in the temperature spatiotemporal feature construction unit); specifically: for the set of effective thermal impact regions obtained in step S1.5, calculate the regional average temperature, peak temperature and temperature change rate of each connected region, and perform a comprehensive evaluation in combination with the spatial overlap of processing regions at adjacent times; when there is only one connected region in the set of effective thermal impact regions that meets the area and temperature constraints, directly determine the connected region as the processing region where the tool and workpiece undergo thermal interaction; when there are multiple connected regions that meet the conditions, select the connected region with the largest comprehensive evaluation index as the processing region where the tool and workpiece undergo thermal interaction, thereby completing the unique convergence of the processing region.

[0094] Temperature spatiotemporal feature construction unit 3 is based on the processing area spatial mask. Through joint analysis of the time and spatial dimensions of the temperature field, it extracts the thermal response feature vector of the processing heat generated by the contact friction between the grinding tool and the workpiece surface, material removal and local plastic deformation.

[0095] In this embodiment, the temperature spatiotemporal feature construction unit 3 extracts a thermal response feature vector characterizing the processing thermal behavior through joint analysis of the temporal and spatial dimensions of the temperature field, including the following steps:

[0096] S2.1 Receive the processing area spatial mask and apply the processing area spatial mask to the temperature field. (At each moment, for the two-dimensional temperature field) A pixel-by-pixel mask mapping process is performed, setting the temperature values ​​corresponding to invalid areas in the mask to zero, empty, or set to a preset invalid flag, retaining only the temperature data within the valid areas of the mask, thereby obtaining the local temperature field. This is used to mask pixels in non-processed areas, resulting in a localized temperature field that contains only the processed areas. ;

[0097] In the formula, This indicates the machining area where the tool and workpiece experience thermal interaction (i.e., the machining heat area formed by the contact friction between the grinding tool and the workpiece surface and the removal of material), used to eliminate the interference of the background area on the extraction of thermal features;

[0098] S2.2, Based on local temperature field For each time The spatial temperature characteristics calculated based on the spatial mask of the machining area (used to characterize the instantaneous spatial distribution of machining heat within the machining area during precision grinding) include at least: the average temperature of the machining area. (The average temperature of the processing area at the current moment is obtained by summing the temperature values ​​of all effective pixels within the processing area and dividing by the total number of effective pixels. This average temperature reflects the overall thermal level of the processing area during the grinding process.) Peak temperature of the processing area. (Regarding the local temperature field) All pixels marked as valid by the spatial mask of the machining area are compared point by point, and the pixel with the highest temperature value is selected as the peak temperature at that moment; this peak temperature is used to characterize the temperature response characteristics of the location with the strongest instantaneous heat input in the local contact area between the grinding tool and the workpiece, and the standard deviation of the machining area temperature is calculated. (The average temperature of the processing area is calculated based on the spatial mask of the processing area) Subsequently, the deviation between the temperature value of each effective pixel in the processing area and the average temperature is squared, and the mean of all squared deviations is calculated. Finally, the square root of this mean is taken to obtain the standard deviation of the temperature in the processing area, which is used to characterize the dispersion of the temperature distribution. The mean temperature gradient in the processing area (based on the local temperature field) is also calculated. Calculate the temperature gradient magnitude corresponding to each effective pixel within the processing area (the gradient magnitude can be obtained by calculating the first-order spatial partial derivative), and under the constraint of the spatial mask of the processing area, statistically average the temperature gradient magnitudes of all effective pixels; this average gradient is used to characterize the overall steepness of temperature change within the processing area, reflecting the degree of concentration of processing heat during grinding and its uniformity of distribution within the processing area from a spatial variation perspective.

[0099] S2.3. Extracting temporal dimension features based on the spatial mask of the machining area (for the dynamic changes in machining heat input during precision grinding), including at least: temperature rise rate. Peak temperature change rate Temperature fluctuation amplitude (i.e., the difference between the maximum and minimum temperature values ​​within the calculated time window, used to reflect the stability of heat input during grinding), and duration of local temperature changes (used to describe the heat accumulation effect of processing heat on the material surface and inside).

[0100] S2.4. The extracted spatial temperature features and temporal features are normalized and combined in chronological order to form a thermal response feature vector characterizing the processing thermal behavior. Among them, the thermal response feature vector is used to characterize the equivalent time sequence of the machining thermal load on the material surface formed by the combined effects of contact friction between the grinding tool and the workpiece, material removal, and local heat accumulation during the machining process (machining thermal load refers to the equivalent heat input intensity and its time variation generated by contact friction between the grinding tool and the workpiece surface, material removal, and the resulting local heat accumulation during the precision grinding of aluminum-based silicon carbide ultrathin reflectors).

[0101] Material thermal response modeling unit 4 establishes a mapping model between processing thermal load and temperature evolution behavior of aluminum-based silicon carbide ultrathin mirror based on thermal response feature vector, which is used to output reference temperature response;

[0102] In this embodiment, the material thermal response modeling unit 4 establishes a mapping model between the processing thermal load and the temperature evolution behavior of the aluminum-based silicon carbide ultrathin mirror based on the thermal response feature vector, including the following steps:

[0103] S3.1. Time-align the thermal response feature vector and use a fixed time window length. and sliding step size The thermal response feature vectors at consecutive time points are organized into a window to form a thermal response feature sequence, which is used to describe the change state of the processing thermal load over a period of time. Each thermal response feature sequence is used to comprehensively characterize the change state of the processing thermal load and its effect on the material surface within the corresponding time window.

[0104] S3.2 Based on the heat load transfer mode in the material, the thermal response feature sequence is structurally reconstructed by the dual-channel modeling method of processing temperature evolution based on heat load transfer mode decomposition, and used as the input variable of the mapping model. In the thermal response feature sequence, each feature component implicitly represents the intensity and stability of the processing heat input from aspects such as the rate of temperature change, the temperature amplitude of the processing area and its time fluctuation characteristics, thereby constituting the equivalent input description of the processing heat load.

[0105] In the precision machining of aluminum-based silicon carbide ultrathin mirrors, the same machining heat load acts on the workpiece through different dominant heat transfer paths at different working conditions, resulting in implicit switching of the temperature evolution mechanism. A single mapping model is difficult to reliably distinguish between "mechanism switching under normal working conditions" and "real abnormal thermal behavior". In the machining area of ​​aluminum-based silicon carbide ultrathin mirrors, the machining heat load is mainly transferred outward through the internal heat diffusion path of the material (such as the continuous grinding section) or the surface heat transfer path (such as intermittent tool feed / lift). The temperature evolution laws corresponding to these two types of transfer paths are essentially different, but in the thermal imaging sequence, they often appear as: the same temperature rise amplitude, similar instantaneous gradient or rate of change. If the thermal response feature sequence is directly used as the whole input to establish a single mapping model, the model cannot distinguish between "mechanism change" and "abnormal deviation", which is very easy to produce false alarms or false alarms.

[0106] The dual-channel modeling method for processing temperature evolution based on heat load transfer mode decomposition identifies the switching behavior between the "heat diffusion-dominated mode" and the "forced heat transfer-dominated mode" in the temperature evolution of the processing area. Without introducing any cooling parameters, it decomposes the processing heat load and temperature mapping into two sub-mappings with different structures and adaptively fuses the output according to the mode ratio.

[0107] This invention uses a dual-channel method to decompose the processing heat load through internal thermal diffusion paths and surface heat transfer paths, enabling the model to distinguish different heat transfer mechanisms and making the temperature evolution law more physically interpretable. By calculating the spatial diffusion consistency index and the time decay dominance index, the proportion of diffusion-dominant and heat transfer-dominant modes is dynamically determined, and the outputs of the two sub-mappings are adaptively fused accordingly. This allows the model to reflect the changes in the dominant heat transfer mechanism under the current processing state in real time, improving prediction accuracy. The structurally reconstructed dual-channel feature sequence implicitly represents the intensity, stability, and instantaneous change characteristics of the processing heat input, enabling the model to stably output the reference temperature response under different operating conditions, thereby reducing false alarms and false negatives in temperature anomaly detection.

[0108] The thermal response feature sequence is structurally reconstructed using a dual-channel modeling method for processing temperature evolution based on thermal load transfer mode decomposition, including the following steps:

[0109] S3.21. Calculate the spatial diffusion consistency index in each thermal response characteristic sequence. (Used to characterize the diffusion continuity and uniform propagation characteristics of the temperature field in the spatial dimension) and the time decay dominant index (Used to characterize the evolutionary features of rapid temperature decay over time and surface heat transfer dominating).

[0110] Among them, the spatial diffusion consistency index for:

[0111] ;

[0112] Time decay dominant index for:

[0113] ;

[0114] In the formula, The set of pixels corresponding to the spatial mask of the processing area. For a moment The spatial gradient vector of the temperature field. To prevent stable terms with a denominator of zero, This represents the spatial mean of the temperature gradient amplitude within the processing area. For a moment Average temperature in the lower processing area For a moment The rate of change of the lower average temperature over time This is the time decay weighting coefficient, used to balance temperature level and rate of change, with a value ranging from 0.01 to 0.1. Determined through expert experience. These are the spatial pixel coordinates in a two-dimensional temperature field;

[0115] In this embodiment, the spatial diffusion consistency index and time decay dominant index The calculation is based on the local temperature field. and from the thermal response feature vector The average temperature of the processing area extracted and its rate of change The process was carried out; specifically: By masking the processing area Acting on the temperature field Obtain the local temperature field Calculate the spatial gradient distribution and second-order curvature field within the local temperature field; analyze the spatial orientation consistency and local smoothness of the gradient vector, and substitute them into the formula ( This is obtained to characterize the proportion of heat conducted within the material; By extracting the average temperature of the processing area from the thermal response feature vector The evolutionary sequence was analyzed, and its instantaneous rate of change and energy dissipation trend within the sliding window were examined. The fitting residuals between the real-time temperature drop curve and the forced heat transfer physical model were calculated and substituted into the formula (…). The result is used to characterize the dominance of surface convection heat transfer under the current operating conditions;

[0116] S3.22. Based on the spatial diffusion consistency index and the time decay dominance index, construct the modal proportion vector corresponding to the heat load transfer path. (This weight does not describe the cooling intensity, but rather whether the current processing heat load is primarily transferred through the "material diffusion path" or the "surface heat transfer path").

[0117] Furthermore, based on the spatial diffusion consistency index and the time decay dominance index, a modal proportion vector corresponding to the heat load transfer path is constructed. This includes the following steps:

[0118] Regarding the first The thermal response characteristic sequences, respectively, affect the spatial diffusion consistency index. and time decay dominant index Dimensionless normalization is performed to eliminate the influence of different dimensions and numerical ranges on modal discrimination;

[0119] Based on the normalized spatial diffusion consistency index and time decay dominance index, diffusion dominance tendency function and heat transfer dominance tendency function are constructed to reflect the dominance of different heat load transfer mechanisms. (The dominance of different heat load transfer mechanisms refers to the relative strength of the effects of various physical transfer paths (such as internal heat diffusion and surface heat transfer) on the overall temperature change when external processing heat loads propagate in the material and cause temperature evolution at a specific processing moment or time window. This concept does not determine whether a mechanism exists, but rather describes which type of transfer mechanism mainly controls the spatial diffusion pattern and time decay characteristics of the temperature field under the current operating conditions. That is, which mechanism contributes more to the temperature evolution and has a stronger explanatory power for the response characteristics, thus providing a basis for subsequent modeling and analysis using the corresponding dominant modes.)

[0120] The diffusion-dominant tendency function is:

[0121] ;

[0122] The dominant heat transfer tendency function is:

[0123] ;

[0124] In the formula, This is the normalized spatial diffusion consistency index. The normalized time decay dominant index;

[0125] Normalizing the diffusion-dominant and heat transfer-dominant tendency functions (to describe the relative proportions of different heat load transfer paths under the current processing state) yields the modal proportion weights corresponding to the heat load transfer paths: the proportion of the heat diffusion-dominant mode within the material. The proportion of surface heat transfer dominant modes ;

[0126] ;

[0127] ;

[0128] The proportion of the dominant mode of internal thermal diffusion in the material The proportion of surface heat transfer dominant modes The modal proportion vector of the heat load transfer path is formed by combining these modal proportions. ;

[0129] S3.23. For the same thermal response characteristic sequence, based on the modal proportion vector Two parallel characteristic sequences of heat load transfer channels were constructed: a diffusion-dominant channel characteristic sequence and a heat transfer-dominant channel characteristic sequence.

[0130] Among them, the diffusion-dominant channel characteristic sequence for:

[0131] ;

[0132] Heat exchange dominant channel characteristic sequence for:

[0133] ;

[0134] In the formula, This represents the proportion of the dominant mode of thermal diffusion within the material, with a value ranging from [0,1]. It is determined through expert experience and reflects the degree to which the current processing thermal load dominates the diffusion path within the material during this time window. The percentage of the dominant surface heat transfer mode, with a value range of [0,1], is determined by expert experience and reflects the degree to which the current processing heat load dominates the surface heat transfer path within this time window;

[0135] Through the above structural reconstruction, the same original thermal response feature sequence is decoupled into two sub-feature sequences that are different in both physical meaning and transmission path;

[0136] S3.24. The feature sequences of diffusion-dominant channels and heat transfer-dominant channels are used as input variables of the mapping model to learn: the temperature evolution mapping relationship under the action of processing heat load through the diffusion path inside the material; and the temperature evolution mapping relationship under the action of processing heat load through the heat transfer path on the workpiece surface. In the subsequent modeling stage (step S3.4), the two types of sub-mapping outputs are adaptively fused according to the modal proportion vector to obtain the final processing temperature evolution prediction result.

[0137] S3.3 Under normal processing conditions, collect n sets of complete processing data. For each set of processing data, obtain the corresponding thermal response characteristic sequence and the actual measured temperature evolution sequence of the processing area. Then, according to the time window, construct a normal processing thermal response sample set to reflect the temperature evolution law of aluminum-based silicon carbide ultrathin mirror under normal thermal load conditions. When constructing the normal processing thermal response sample set, simultaneously retain two types of characteristic sequences: diffusion-dominated and heat transfer-dominated, for subsequent channel-by-channel modeling of the mapping model.

[0138] S3.4 Based on the normal processing thermal response sample set, a mapping model between the thermal response feature sequence and the temperature evolution behavior of the processing area is established by using the channel-weighted regression method, which is used to output the reference temperature response under a given processing thermal load condition;

[0139] During the machining process, the mapping model is corrected or updated online using real-time acquired thermal response feature vectors to compensate for the effects of tool wear, environmental changes, and individual material differences on thermal response behavior, and to maintain the model's adaptability to the current machining state. Based on the updated mapping model, the machining thermal load characteristics at the current moment are input, and the corresponding reference temperature evolution results are output as the benchmark input for the temperature anomaly discrimination and classification evaluation unit, which is used for subsequent real-time temperature deviation calculation.

[0140] Furthermore, the mapping model is as follows:

[0141] ;

[0142] ;

[0143] In the formula, The temperature evolution prediction value of the processing area is calculated by the mapping model based on the thermal response characteristic sequence of the diffusion-dominant channel and the heat transfer-dominant channel. It is used to characterize the reference temperature response of the aluminum-based silicon carbide ultrathin mirror under the corresponding processing thermal load conditions.

[0144] in, and All are multiple linear regression models:

[0145] ;

[0146] ;

[0147] in, This is the regression coefficient vector (column vector) corresponding to the feature sequence of the diffusion channel, used to linearly combine the features of each dimension and characterize the weight of each feature's influence on the temperature output. This is the bias term for the diffusion channel. Let be the regression coefficient vector of the heat exchange channel. This is the bias term for the heat exchange channel.

[0148] The temperature anomaly detection and alarm unit 5 compares the real-time acquired thermal response feature vector with the reference temperature response, calculates the real-time temperature deviation index based on the processing area space mask, and performs weighted fusion of the temperature deviation index to form a comprehensive temperature anomaly index, while simultaneously issuing an anomaly alarm for anomaly monitoring.

[0149] In this embodiment, the temperature anomaly detection and alarm unit 5 includes a temperature anomaly detection module and a temperature anomaly alarm module;

[0150] The temperature anomaly detection module calculates multiple temperature deviation indices based on the processing area spatial mask, real-time thermal response feature vector, and reference temperature response. These indices include regional average temperature deviation, regional peak temperature deviation, and temperature change rate deviation. The module then normalizes these temperature deviation indices and constructs a comprehensive temperature anomaly index using a weighted fusion method. (In the formula, This represents the normalized regional average temperature deviation. This represents the normalized regional peak temperature deviation. This represents the normalized deviation in the rate of temperature change. The weighting for the regional average temperature deviation. The weighting for the regional peak temperature deviation. As the weight of the temperature change rate deviation, );

[0151] The temperature anomaly alarm module compares a comprehensive temperature anomaly index with a preset threshold to classify the processing temperature anomaly into four levels: normal, mild, moderate, and severe. An alarm is then triggered based on the classification result. Specifically, the current processing area temperature level is determined through a step-by-step assessment: when… When it is judged as a normal state; when It was initially determined to be a mild abnormality; when When it is judged as moderately abnormal; when If an anomaly is detected, a corresponding alarm signal is triggered based on the severity level: a minor anomaly triggers a low-priority alarm, such as a text prompt on the interface or a flashing indicator light; a moderate anomaly triggers a medium-priority alarm, and may also include an audible prompt or a short buzzer to attract the operator's attention; a severe anomaly triggers a high-priority alarm, including continuous audible and visual alarms, system pop-up warnings, and sending real-time alarm messages to the higher-level monitoring platform. This can be combined with automatic recording of the anomaly time, processing parameters, and temperature data for subsequent analysis and tracing, thereby achieving graded response and operable intervention for different degrees of processing temperature anomalies.

[0152] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A temperature anomaly monitoring system for the processing area of ​​an aluminum-based silicon carbide ultrathin reflector, characterized in that, include: Thermal imaging sensing acquisition and processing unit (1) performs continuous thermal imaging acquisition of the processing area of ​​aluminum-based silicon carbide ultrathin reflector and outputs the temperature field. The adaptive positioning unit (2) of the processing area calculates the temperature spatial gradient based on the temperature field, and combines the temperature spatial gradient threshold and the minimum connected area threshold to identify the processing area of ​​the grinding tool and the workpiece thermal interaction, and outputs the processing area spatial mask at the same time. Among them, the process of identifying the processing area is optimized by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field; Temperature spatiotemporal feature construction unit (3), the temperature spatiotemporal feature construction unit (3) is based on the processing area space mask, and extracts the thermal response feature vector of the processing heat generated by the contact friction between the grinding tool and the workpiece surface, material removal and local plastic deformation through joint analysis of the time and space dimensions of the temperature field; Material thermal response modeling unit (4), the material thermal response modeling unit (4) establishes a mapping model between processing thermal load and temperature evolution behavior of aluminum-based silicon carbide ultrathin mirror based on thermal response feature vector, and is used to output reference temperature response; Temperature anomaly detection and alarm unit (5) compares the real-time thermal response feature vector with the reference temperature response, calculates the real-time temperature deviation index based on the processing area space mask, and performs weighted fusion of the temperature deviation index to form a comprehensive temperature anomaly index, and at the same time performs anomaly alarm for anomaly monitoring. The material thermal response modeling unit (4) establishes a mapping model between the processing thermal load and the temperature evolution behavior of the aluminum-based silicon carbide ultrathin mirror based on the thermal response feature vector, including the following steps: S3.

1. Time-align the thermal response feature vector and use a fixed time window length. and sliding step size The thermal response feature vectors at consecutive time points are organized into a window to form a thermal response feature sequence; S3.2 Based on the heat load transfer modes in the material, the thermal response feature sequence is structurally reconstructed by a dual-channel modeling method for processing temperature evolution based on heat load transfer mode decomposition, and used as the input variable of the mapping model. S3.3 Collect n sets of complete processing data, obtain the corresponding thermal response characteristic sequence and the actual measured temperature evolution sequence of the processing area for each set of processing data, and construct a normal processing thermal response sample set according to the time window. S3.4 Based on the normal processing thermal response sample set, a mapping model between the thermal response feature sequence and the temperature evolution behavior of the processing area is established by using the channel-weighted regression method, which is used to output the reference temperature response under a given processing thermal load condition; In step S3.2, the thermal response feature sequence is structurally reconstructed using a dual-channel modeling method for processing temperature evolution based on thermal load transfer mode decomposition, including the following steps: S3.

21. In each thermal response characteristic sequence, calculate the spatial diffusion consistency index and the time decay dominance index; S3.

22. Based on the spatial diffusion consistency index and the time decay dominance index, construct the modal proportion vector corresponding to the heat load transfer path. ; S3.

23. For the same thermal response characteristic sequence, based on the modal proportion vector Two parallel characteristic sequences of heat load transfer channels were constructed: a diffusion-dominant channel characteristic sequence and a heat transfer-dominant channel characteristic sequence. S3.

24. Use the characteristic sequences of diffusion-dominant channels and heat transfer-dominant channels as input variables for the mapping model.

2. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin mirror according to claim 1, characterized in that: The thermal imaging sensing acquisition and processing unit (1) includes a thermal imaging sensing acquisition module and a thermal imaging data preprocessing module; The thermal imaging sensing and acquisition module performs continuous infrared thermal imaging acquisition on the processing area of ​​the aluminum-based silicon carbide ultrathin reflector to obtain a time-series thermal image sequence. The thermal imaging data preprocessing module preprocesses the time-series thermal image sequence to form a two-dimensional temperature field in a unified coordinate system. .

3. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin reflector according to claim 2, characterized in that: The adaptive positioning unit (2) of the machining area calculates the spatial temperature gradient based on the temperature field and identifies the machining area where the tool and workpiece interact thermally, including the following steps: S1.1, Based on two-dimensional temperature field At every moment Below, the first-order spatial gradient components of temperature are calculated along two orthogonal directions in space to characterize the local variation trend of temperature in the horizontal and vertical directions; S1.2 Calculation of temperature gradient amplitude based on first-order spatial gradient components To characterize the intensity of temperature change within a unit spatial scale; Specifically, by introducing a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field, the temperature gradient amplitude is adjusted. Directional radiation correction is performed to generate the corrected temperature gradient amplitude. This is used to reduce the interference of curvature and observation angle changes on the identification of the processing area; S1.3, Temperature gradient amplitude With the preset gradient threshold The comparisons are performed, and a binary map of the candidate thermal impact region is constructed as the candidate thermal impact region image. S1.4 Perform connected component analysis on the candidate thermal effect region image to extract candidate regions that are spatially continuous. S1.5 Calculate the spatial area of ​​each candidate region, based on the preset minimum connected area threshold. Filter out spaces with a spatial area greater than the minimum connected area threshold. Candidate regions are formed to create a set of effective thermal action regions; S1.6 From the set of effective heat-affected areas, determine and identify the processing areas where the tool and workpiece experience thermal interaction based on the area temperature characteristics and temporal continuity, and output them in the form of a processing area spatial mask.

4. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin reflector according to claim 3, characterized in that: In S1.2, a curvature-sensing directional radiation correction mechanism based on the second-order change of the temperature field is introduced to correct the temperature gradient amplitude. Directional radiation correction includes the following steps: S1.21, Based on two-dimensional temperature field The second-order partial derivatives are calculated along two orthogonal directions in space, and a pixel-level temperature curvature surrogate is constructed. ; S1.

22. Based on the coupling relationship between thermal radiation directionality and observation angle on temperature measurement results, pixel-level temperature curvature proxy quantity is used. Normalized and mapped to equivalent observation angle bias factor ; S1.

23. Combining the equivalent observation angle deviation, construct the directional radiation compensation weight function and generate the directional radiation compensation weight. ; S1.24, Using directional radiation to compensate for weights For the original temperature gradient magnitude Perform directional radiation correction to obtain the corrected temperature gradient amplitude. .

5. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin reflector according to claim 4, characterized in that: In S1.22, based on the coupling relationship between the thermal radiation directionality and the observation angle on the temperature measurement results, the pixel-level temperature curvature proxy is used. Normalized and mapped to equivalent observation angle bias factor This includes the following steps: Pixel-level temperature curvature proxy Steady-state constraint normalization is performed within its local neighborhood to generate a normalized pixel-level temperature curvature surrogate quantity. ; A directional response function is constructed based on the directionality of thermal radiation and the observation angle; Mapping the directional response function to an equivalent observation angle deviation factor. .

6. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin mirror according to claim 2, characterized in that: The temperature spatiotemporal feature construction unit (3) extracts a thermal response feature vector characterizing the processing thermal behavior through joint analysis of the time and spatial dimensions of the temperature field, including the following steps: S2.1 Receive the processing area spatial mask and apply the processing area spatial mask to the temperature field. This yields a local temperature field that includes only the processing area. ; S2.2, Based on local temperature field For each time period, the spatial temperature characteristics are calculated based on the spatial mask of the processing area, including at least: the average temperature of the processing area, the peak temperature of the processing area, the standard deviation of the temperature of the processing area, and the mean temperature gradient of the processing area. S2.

3. Based on the spatial mask extraction of the processing area, the temporal dimension features include at least: temperature rise rate, peak temperature change rate, temperature fluctuation amplitude, and local temperature change duration. S2.

4. The extracted spatial temperature features and temporal features are normalized and combined in chronological order to form a thermal response feature vector characterizing the processing thermal behavior. Among them, the thermal response feature vector is used to characterize the equivalent time sequence of the processing thermal load on the material surface formed by the combined effects of contact friction between the grinding tool and the workpiece, material removal, and local heat accumulation during the processing.

7. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin mirror according to claim 1, characterized in that: In step S3.22, a modal proportion vector corresponding to the heat load transfer path is constructed based on the spatial diffusion consistency index and the time decay dominance index. This includes the following steps: Regarding the first The thermal response characteristic sequences, respectively, affect the spatial diffusion consistency index. and time decay dominant index Perform dimensionless normalization; Based on the normalized spatial diffusion consistency index and time decay dominance index, diffusion dominance tendency function and heat transfer dominance tendency function are constructed to reflect the dominance of different heat load transfer mechanisms. Normalizing the diffusion-dominant and heat transfer-dominant tendency functions yields the modal weights corresponding to the heat load transfer paths: the proportion of the thermal diffusion-dominant mode within the material. The proportion of surface heat transfer dominant modes ; The proportion of the dominant mode of internal thermal diffusion in the material The proportion of surface heat transfer dominant modes The modal proportion vector of the heat load transfer path is formed by combining these modal proportions. .

8. The temperature anomaly monitoring system for the processing area of ​​the aluminum-based silicon carbide ultrathin mirror according to claim 1, characterized in that: The temperature anomaly detection and alarm unit (5) includes a temperature anomaly detection module and a temperature anomaly alarm module; The temperature anomaly detection module calculates multiple temperature deviation indices based on the processing area spatial mask, real-time thermal response feature vector, and reference temperature response, including regional average temperature deviation, regional peak temperature deviation, and temperature change rate deviation. It normalizes the multiple temperature deviation indices and constructs a comprehensive temperature anomaly index using a weighted fusion method. The temperature anomaly alarm module compares the comprehensive temperature anomaly index with a preset threshold to classify the processing temperature anomaly and issues an alarm based on the classification result.

Images