Electronic digital data processing and visualization system for hot working process data

Abstract

The application relates to the field of electric digital data processing, and discloses an electric digital data processing and visualization system for hot processing process data, which comprises the following modules: a data alignment module which acquires multi-source time sequence signals under hot processing conditions and completes time domain alignment and packaging; a topology control module which stores a space topology correlation matrix, maps sample values to nodes, completes feature assignment, determines the algebraic operation relationship of the product of a state transition matrix and a historical feature vector plus an input excitation vector by calculating a topology state feature vector of a current period; and an incremental rendering module which calculates a differential feature vector of adjacent periods, extracts element coordinates whose absolute values exceed an amplitude threshold, and completes local data coverage of a display memory storage module. The application converts space correlation reconstruction into matrix state evolution, avoids global redundant recalculation and large-area pixel flushing, eliminates data bus occupation pressure, and improves the real-time performance of process situation reconstruction.

Description

Technical Field

[0001] This invention belongs to the field of electrical digital data processing technology, and particularly relates to an electrical digital data processing and visualization system for thermal processing process data. Background Technology

[0002] Currently, in the real-time monitoring of continuous thermal processing physical processes, the use of sensor arrays to collect physical quantities such as temperature, pressure, and acoustic emission, and the reconstruction of the above time-series signals into a thermal field visualization image through an electro-digital data processing system, is a fundamental means of process situation analysis. Existing processing architectures are generally based on the principle of global clock synchronization, aligning heterogeneous acquisition signals by timestamp and encapsulating them into a linear data array. In each display refresh cycle, the system extracts the global sampling values ​​of the current time section, calls the spatial interpolation algorithm to reconstruct the gridded spatial correlation, and sends instructions to the display memory unit to update the screen.

[0003] The heat conduction process has strong topological correlation in physical space, and within the sampling window, the evolution of the physical field usually exhibits a locally non-equilibrium distribution. The linear data organization logic of traditional systems severs the spatial constraints between nodes, forcing the processor to repeatedly perform interpolation calculations and memory flushing on a massive number of redundant nodes in steady state or background noise range in each cycle. This processing method causes the computing power overhead to increase exponentially with the scale of sampling points, leading to data bus congestion and rendering latency, and making it difficult to capture transient physical distortions in the thermal processing process. Increasing the number of computing cores or adopting downsampling strategies can alleviate response latency, but they do not eliminate the dimensional mismatch between the data structure and the evolution law of the physical field, and are respectively limited by the surge in hardware costs or the loss of key process features.

[0004] The aforementioned limitations are not only constrained by the objective limitations of hardware form. At the underlying data flow and rendering logic level, existing control methods also struggle to simultaneously meet the real-time requirements of high-frequency acquisition and continuous evolution of the physical field. For example, Chinese invention patent application CN120879458A discloses a method for local configuration of a primary main wiring diagram based on a microcomputer protection system. Although the scheme utilizes a dirty rectangle merging algorithm to optimize local refresh efficiency, its core logic is anchored to the discrete state UI presentation of switch quantities and other parameters in the power system. The secondary mapping structure of grid position and primitive identifier essentially treats the acquisition points as independent visual units. When faced with thermal processing scenarios with strong thermal conduction coupling attributes, the incremental refresh mechanism based on the UI level lacks a mathematical representation of the energy diffusion weight between spatial nodes, making it unable to complete numerical deduction based on the laws of physical evolution. As a result, when the system deals with non-equilibrium transient thermal distortion, it still faces the challenges of dimensional mismatch leading to computational power dissipation and situation reconstruction delay.

[0005] Therefore, how to reconstruct the topology of electronic digital data flow and realize asynchronous evolution based on local event driving, thereby reducing system resource consumption while improving the real-time performance of process status reconstruction, has become the technical problem to be solved by this invention. Summary of the Invention

[0006] This invention aims to solve the problems of computing power bottleneck and rendering latency caused by global synchronous reconstruction and spatial interpolation redundancy in the processing of thermal processing time sequence data.

[0007] In this technical solution, an electro-digital data processing and visualization system for thermal processing process data includes: The data alignment module is used to acquire multi-source timing signals during the hot processing process, and to complete the time domain alignment of each source timing signal and the encapsulation of synchronous data blocks based on the global system clock; The topology control module, connected to the data alignment module, stores a preset spatial topology correlation matrix. This spatial topology correlation matrix consists of multiple nodes representing the correlation between sampling points and weight coefficients between nodes. The weight coefficients are used to represent the correlation strength between adjacent nodes at the data level. After receiving the synchronization data block, the topology control module maps the values ​​of each sampling point in the synchronization data block to the corresponding node address in the spatial topology correlation matrix, completes the feature assignment of the node, and calculates the topology state feature vector of the current period. The topology state feature vector of the current period is equal to the product of the spatial topology correlation matrix and the topology state feature vector of the previous period, plus the input excitation vector generated by the synchronization data block image. The incremental rendering module, connected to the topology control module, is used to calculate the difference feature vector between the topology state feature vectors of two consecutive cycles, and extract the coordinates of elements whose absolute values ​​exceed a preset amplitude threshold in the difference feature vector. Based on the element coordinates, the module addresses the target pixel region in the video memory storage module to complete local data coverage.

[0008] Preferably, the weight coefficient ranges from 0.05 to 0.85, and the topology control module dynamically corrects the weight coefficient of the corresponding node by using a linear interpolation algorithm based on the rate of change of the values ​​of each sampling point in the synchronous data block with the sampling time.

[0009] Preferably, the operation logic inside the data alignment module includes: step S31, extracting the sampled values ​​and channel identifiers from the multi-source timing signals through the signal parsing unit; step S32, encapsulating the sampled values ​​within the same sampling window into the corresponding address bits of the synchronization data block according to the system clock reference through the clock alignment unit.

[0010] Preferably, the incremental rendering module includes an addressing unit and a video memory control unit; the addressing unit is used to convert the coordinates of elements in the differential feature vector whose absolute value exceeds a preset amplitude threshold into physical address indices of the video memory storage module; the video memory control unit is used to overwrite the target pixel data block based on the physical address indices and keep the data state of non-indexed addresses in the video memory storage module unchanged.

[0011] Preferably, it also includes an asynchronous snapshot module, which is connected to the topology control module and is used to extract the complete topology state feature vector at a preset time interval and write it as a full reference frame to the backup storage module; when the incremental rendering module determines that there is a data verification anomaly in the video memory storage module, it calls the most recent full reference frame to complete the full address overwrite of the video memory storage module.

[0012] Preferably, the topology control module further includes a convergence monitoring module, which is used to optimize the data processing accuracy through the following logic: Step S61, extract the set of nodes whose values ​​exceed a preset high threshold in the topology state feature vector of the current period; Step S62, calculate the numerical spatial gradient between adjacent nodes in the node set; Step S63, when the numerical spatial gradient exceeds the preset convergence threshold, increase the weight coefficient of the corresponding node in the spatial topology correlation matrix by a preset fixed step size.

[0013] Preferably, the incremental rendering module integrates a dirty rectangle refresh unit to improve rendering efficiency through the following logic: Step S71, using the minimum bounding rectangle algorithm, merge the coordinates of multiple adjacent elements with absolute values ​​exceeding a preset amplitude threshold into a single rectangle refresh area; Step S72, send the rectangle refresh area as a single addressable object to the video memory storage module to complete batch overwriting.

[0014] Preferably, the sampled values ​​in the synchronous data block include temperature data, pressure data, or acoustic emission intensity data; the node topology structure of the spatial topology correlation matrix is ​​established as a three-dimensional static mapping based on the layout coordinates of the data acquisition points in physical space.

[0015] Preferably, it also includes a situation output module, which is connected to the video memory storage module and is used to generate a visualized thermal field map based on the real-time data in the video memory storage module after receiving the local coverage completion signal from the incremental rendering module.

[0016] Compared with existing technologies, the electro-digital data processing and visualization system for thermal processing data of the present invention has the following advantages: 1. In the electrical digital data processing of thermal processing data, the topology mapping architecture transforms the spatial interpolation process of discrete data into a deterministic matrix state evolution by pre-setting the spatial correlation attributes in the topology correlation matrix in memory. Traditional processing methods require complex adjacency retrieval and interpolation operations for global sampling points in each calculation cycle, resulting in an exponential increase in computing resource consumption with the node scale. This solution uses the topology correlation matrix to solidify the physical mapping relationship between nodes. When receiving the sampling signal, the spatial state reconstruction of the physical field can be completed through matrix multiplication and addition operations. This transformation from dynamic algorithm addressing to static structure mapping makes the computational complexity return to the linear growth range, effectively eliminating the bottleneck of computing power dissipation under large-scale high-frequency sampling.

[0017] 2. The asynchronous differential propagation mechanism achieves efficient use of data bus bandwidth by adjusting the global clock synchronization drive to a local mutation event triggering logic. The system no longer requires global alignment and periodic polling of multi-source heterogeneous signals. Instead, it extracts the differential component between the sampled value of a single channel and the historical registered value, and uses this to determine the processing path. When the local physical quantity is in a stable state and the differential component does not exceed the threshold, the system directly cuts off the calculation link of redundant data. This processing method ensures that system resources are always focused on the sampling channel that produces substantial changes, improving the system's accuracy in capturing transient process fluctuations and response time without increasing the hardware clock frequency.

[0018] 3. The differentiated rendering pipeline extracts the differential feature vectors of adjacent processing cycles to accurately locate the local update area of ​​the memory unit, avoiding invalid global pixel flushing. Since the evolution of the physical field during the thermal processing has obvious spatial limitations in a short period of time, the system only sends overwrite instructions to the memory address corresponding to the non-zero element in the differential feature vector. This mechanism establishes a monotonic correlation between the memory bandwidth usage and the actual evolution frequency of the physical field, thereby removing the physical constraint of global synchronous refresh on the visualization frame rate from the underlying logic, and ensuring that the smoothness and real-time performance of image output are maintained even during peak data traffic periods. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the data logic evolution and incremental rendering process of the hot processing technology of this invention; Figure 2 This is a visual system architecture diagram of the asynchronous snapshot and data self-healing guarantee of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state (as shown in the accompanying drawings). They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0022] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0023] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0024] An electro-digital data processing and visualization system for thermal processing process data, comprising: The data alignment module is used to acquire multi-source timing signals during the hot processing process, and to complete the time domain alignment of each source timing signal and the encapsulation of synchronous data blocks based on the global system clock; The topology control module, connected to the data alignment module, stores a preset spatial topology correlation matrix. This spatial topology correlation matrix consists of multiple nodes representing the correlation between sampling points and weight coefficients between nodes. The weight coefficients are used to represent the correlation strength between adjacent nodes at the data level. After receiving the synchronization data block, the topology control module maps the values ​​of each sampling point in the synchronization data block to the corresponding node address in the spatial topology correlation matrix, completes the feature assignment of the node, and calculates the topology state feature vector of the current period. The topology state feature vector of the current period is equal to the product of the spatial topology correlation matrix and the topology state feature vector of the previous period, plus the input excitation vector generated by the mapping of the synchronization data block. The incremental rendering module, connected to the topology control module, is used to calculate the difference feature vector between the topology state feature vectors of two consecutive cycles, and extract the coordinates of elements whose absolute values ​​exceed a preset amplitude threshold in the difference feature vector. Based on the element coordinates, the module addresses the target pixel region in the video memory storage module to complete local data coverage.

[0025] Preferably, the weight coefficient ranges from 0.05 to 0.85, and the topology control module dynamically corrects the weight coefficient of the corresponding node by using a linear interpolation algorithm based on the rate of change of the values ​​of each sampling point in the synchronous data block with the sampling time.

[0026] Preferably, the operation logic inside the data alignment module includes: step S31, extracting the sampled values ​​and channel identifiers from the multi-source timing signals through the signal parsing unit; step S32, encapsulating the sampled values ​​within the same sampling window into the corresponding address bits of the synchronization data block according to the system clock reference through the clock alignment unit.

[0027] Preferably, the incremental rendering module includes an addressing unit and a video memory control unit; the addressing unit is used to convert the coordinates of elements in the differential feature vector whose absolute value exceeds a preset amplitude threshold into physical address indices of the video memory storage module; the video memory control unit is used to overwrite the target pixel data block based on the physical address indices and keep the data state of non-indexed addresses in the video memory storage module unchanged.

[0028] Preferably, it also includes an asynchronous snapshot module, which is connected to the topology control module and is used to extract the complete topology state feature vector at a preset time interval and write it as a full reference frame to the backup storage module; when the incremental rendering module determines that there is a data verification anomaly in the video memory storage module, it calls the most recent full reference frame to complete the full address overwrite of the video memory storage module.

[0029] Preferably, the topology control module further includes a convergence monitoring module, which is used to optimize the data processing accuracy through the following logic: Step S61, extract the set of nodes whose values ​​exceed a preset high threshold in the topology state feature vector of the current period; Step S62, calculate the numerical spatial gradient between adjacent nodes in the node set; Step S63, when the numerical spatial gradient exceeds the preset convergence threshold, increase the weight coefficient of the corresponding node in the spatial topology correlation matrix by a preset fixed step size.

[0030] Preferably, the incremental rendering module integrates a dirty rectangle refresh unit to improve rendering efficiency through the following logic: Step S71, using the minimum bounding rectangle algorithm, merge the coordinates of multiple adjacent elements with absolute values ​​exceeding a preset amplitude threshold into a single rectangle refresh area; Step S72, send the rectangle refresh area as a single addressable object to the video memory storage module to complete batch overwriting.

[0031] Preferably, the sampled values ​​in the synchronous data block include temperature data, pressure data, or acoustic emission intensity data; the node topology structure of the spatial topology correlation matrix is ​​established as a three-dimensional static mapping based on the layout coordinates of the data acquisition points in physical space.

[0032] Preferably, it also includes a situation output module, which is connected to the video memory storage module and is used to generate a visualized thermal field map based on the real-time data in the video memory storage module after receiving the local coverage completion signal from the incremental rendering module.

[0033] Example 1: In the application scenario of real-time monitoring of temperature field in a large continuous hot rolling production line, when the system faces the data pressure of 256 infrared temperature sensors deployed at the exit of the finishing mill receiving data in parallel at a sampling frequency of 10kHz, the traditional system uses linear timing alignment and global interpolation to reconstruct the image, resulting in a data bus occupancy rate that normally exceeds 90%, and the image reconstruction delay accumulating to more than 45ms, making it difficult to capture transient thermal distortions on the surface of the rolled piece. The specific operation of this technical solution is as follows: The data alignment module acquires multi-source timing signals and extracts each based on the global system clock. The signal pulses from the sampling channel encapsulate the temperature data within the same sampling window into the corresponding address bits of the synchronization data block according to the timestamp, achieving deterministic alignment with a time-domain error of less than 10μs. After receiving the synchronization data block, the topology control module retrieves the spatial topology correlation matrix pre-stored in memory. This matrix consists of multiple nodes and weight coefficients characterizing the spatial correlation of the sensor's physical deployment points. The weight coefficients range from 0.05 to 0.85. The system directly maps the temperature sampling values ​​in the synchronization data block to the corresponding node addresses in the spatial topology correlation matrix to complete feature assignment, and then applies the formula... Perform the topological state evolution calculation for the current cycle, where, This is the topological state feature vector for the current period. It is a spatial topological correlation matrix. This is the topological state feature vector of the previous period. The input excitation vector is generated for the current period's synchronous data block mapping. During this algebraic operation, for stationary channels that have not reached the amplitude threshold, the system does not change the global definition dimension of the feature vector, but forces the excitation component corresponding to it in the differential feature stage to zero. Through the sparse matrix-vector multiplication instruction supported by the underlying hardware, the multiply-accumulate unit automatically bypasses the calculation cycle of zero-value operands. Thus, while maintaining the complete mapping logic of the matrix dimension, the computation link of non-mutated nodes is substantially cut off. This processing method, which solidifies the physical space topology relationship to the matrix weight, eliminates the need for the system to re-search the adjacency relationship of global nodes in each display cycle, reducing the computational complexity from exponential to linear growth range, and exhibiting the characteristics of computational efficiency when processing large-scale node data.

[0034] The incremental rendering module calculates the difference feature vector between the topology state feature vectors of adjacent cycles, extracts the coordinates of elements whose absolute value exceeds a preset amplitude threshold in the difference feature vector, and converts these coordinates into physical address indexes of the video memory storage module through the addressing unit. The video memory control unit only performs overwrite operations on the target pixel area pointed to by the index, while keeping the data state of non-indexed addresses unchanged, thus avoiding invalid global pixel flushing. During continuous system operation, if an abnormal condition with a local temperature drop rate exceeding 150℃ / s is detected in the rolled piece, the topology control module extracts the set of high-threshold nodes and calculates the numerical space gradient through the built-in convergence monitoring module. When the numerical space gradient exceeds the preset convergence threshold, the weight coefficient of the corresponding node is increased by a preset fixed step size to enhance the local response sensitivity. Through the above asynchronous evolution logic based on local event driving, the system reduces the data bus occupancy rate to below 15% and stably controls the end-to-end latency of visualization reconstruction to within 2ms. Finally, the visualized thermal field map of the rolling process accurately presents the details of transient cooling caused by local oxide scale peeling while maintaining global smoothness.

[0035] During high-frequency signal synchronization alignment, the data alignment module establishes a circular buffer for each sensor signal, the buffer depth L of which is determined by the sampling frequency. The calculation logic is determined by the data bus width B. Where L is the buffer depth, Where B is the sampling frequency and B is the data bus width. To synchronize the system clock step size, the signal parsing unit extracts the original bit stream containing the channel identifier. The clock alignment unit triggers data stacking based on the pulse edge of the global system clock. Signal frames with a time offset exceeding 10μs are marked as delayed frames and compensated to the current sampling window using linear extrapolation. Based on the preset channel mapping table, the sampled values ​​from different sources are written to the fixed address bits of the synchronization data block. After the single frame data is encapsulated, it is transported to the input register of the topology control module through the bus controller.

[0036] Example 2: In a real-time monitoring test environment for high-power laser cladding technology, to verify the computational convergence and rendering timeliness of this system when processing non-equilibrium transient thermal evolution data, a test platform simulating a 1.2m x 0.8m processing area was built, and 1024 sampling nodes were deployed. The data source used by this test platform was generated based on the finite element analysis simulation model of the heat transfer control equation. Gaussian white noise with a signal-to-noise ratio of 20dB and power frequency interference harmonics with a frequency of 50Hz were actively superimposed on the original temperature time series signal to simulate the physical jitter of a real sensor array in an industrial electromagnetic environment. When setting the sampling period of the data alignment module, the core technology consideration was to achieve a balance between the real-time capture of the workpiece cooling and the bus data load, ensuring the capture of 5000°C per second at the edge of the molten pool. Temperature transitions without signal aliasing are achieved. Based on the Nyquist sampling theorem, the sampling period is set to 0.1 ms. The weighting coefficient in the topology control module is set between 0.05 and 0.85. This coefficient physically represents the energy diffusion coupling strength caused by heat conduction between adjacent nodes. It is dynamically corrected based on the rate of change of the sampled value over time using a linear interpolation algorithm. Before dynamic correction, the control unit pre-reads the sampled values ​​from three consecutive sampling periods, filtering out invalid fluctuations with amplitudes smaller than the quantization noise tolerance to determine the node temperature data as the input state. Based on the physical law that the local temperature evolution rate and energy relaxation time are negatively correlated in thermal diffusion dynamics, the topology control module performs interpolation operations through deterministic algebraic logic. Specifically, the control unit extracts the absolute value of the first derivative of the current node's temperature in adjacent periods as the rate of change R and calculates it according to the formula... Calculate the updated weight coefficients In linear mapping logic and represents the lower and upper bounds of the weight coefficient truncation, respectively, both being dimensionless constants, with values ​​of 0.05 and 0.85 to ensure the mathematical convergence of the matrix iteration. and The calibration benchmarks for the rate of change under steady-state thermal background and extreme thermal shock conditions are respectively characterized. The international standard units are all °C / s. They are calibrated based on historical operating conditions of 10 and 500 respectively. The input rate of change R is less than The system will lock the weighting coefficients. The rate of change of input R is greater than Time system lock Form a closed-loop parameter control flow that covers the entire test boundary.

[0037] The experimental process inputs raw sampling data containing the aforementioned background noise, and the observed random fluctuation amplitude of single-point temperature measurements remained at approximately 1.5℃. The topology control module performed state evolution calculations, extracting the topological state feature vector for the current period. During this process, due to the solidification of physical diffusion constraints by the spatial topological correlation matrix, the system generated differential feature vectors in the core derivation stage of the second act. Data shows that the amplitude of differential elements in the steady-state background region was suppressed to below 0.2%, while the differential feature value in the local region affected by the laser beam jumped to 12.8%. Finally, the incremental rendering module improved... The coordinates of elements whose absolute values ​​exceed a preset amplitude threshold are taken and physical addressing is completed. The actual test results show that at a sampling rate of 10kHz, the bus bandwidth utilization of the sample group of this invention is only 12.4%, and the end-to-end reconstruction latency is stable at 1.85ms. To verify the synergistic effect, a control group 1 was set up, in which the incremental rendering module was removed and full pixel flushing was adopted. The results show that its memory bandwidth utilization soared to 88.6% and the latency increased to 42.8ms, indicating that a single topology control cannot eliminate bus congestion while maintaining high-frequency sampling if it lacks the cooperation of differentiated rendering pipelines.

[0038] In the empirical verification of the parameter boundaries, a second out-of-range control group with a weighting coefficient of 0.95 was set. It was observed that when processing signals containing noise interference, the topological state eigenvector exhibited significant numerical oscillations, and the computational convergence time increased from 0.5 ms to over 15.2 ms. This indicates that when the weighting coefficient exceeds the upper limit of 0.85, the system exhibits a nonlinear amplification effect on background noise disturbances, making it unable to maintain the stability of the computational logic. Simultaneously, a third out-of-range control group with a weighting coefficient of 0.02 was set. The results showed that its sensitivity to capturing transient thermal distortion decreased by 75%, confirming the 0.95 limit specified in this invention. The weight range of 0.05 to 0.85 represents the optimal working window that balances noise immunity and response accuracy. Furthermore, by introducing a problem intensity gradient control system, in gradient experiments where laser power increases from 1000W to 5000W, the rendering latency of the proposed solution remains consistently within a stable range of less than 2.1ms, while the latency of control group 1 increases exponentially with the increase of thermal field complexity. This experiment directly demonstrates that the proposed system effectively overcomes the bottleneck of computational power dissipation under high-frequency sampling conditions by converting spatial correlation reconstruction into matrix state evolution and combining it with an incremental rendering mechanism, thereby achieving high-fidelity capture and reconstruction of transient physical features.

[0039] Example 3: This example combines Figures 1 to 2 A description of the electrical digital data processing and visualization system used for thermal processing data, such as... Figure 1As shown, multi-source timing signals and the global system clock are input to the data alignment module. The data alignment module completes the time-domain alignment of each source timing signal and the encapsulation of synchronous data blocks. The encapsulated synchronous data blocks are transmitted to the topology control module. The spatial topology correlation matrix is ​​associated with the topology control module by providing a physical image. This spatial topology correlation matrix consists of multiple nodes representing the correlation of the sampling points and the weight coefficients between the nodes. The topology control module maps the sampling point values ​​to the node addresses to complete feature assignment and calculates the topology state feature vector of the current period. The topology state feature vectors of two consecutive periods are output to the incremental rendering module. The incremental rendering module calculates the difference feature vector of two consecutive periods and extracts the coordinates of elements whose absolute values ​​exceed the amplitude threshold. Finally, based on the element coordinates, the target pixel area addressed in the video memory storage module is used to complete local data coverage.

[0040] like Figure 2 As shown, a complete visualization architecture is constructed through the hardware connections between various modules. The multi-source timing signal, data alignment module, topology control module, incremental rendering module, video memory storage module, and situation output module are connected in sequence. The auxiliary output of the topology control module is connected to the asynchronous snapshot module. The asynchronous snapshot module transmits the extracted feature data to the backup storage module for storage. The feedback output of the backup storage module points to the video memory storage module to perform a full address overwrite when data verification is abnormal. Finally, the situation output module receives the real-time data in the video memory storage module and generates a visualized thermal field map.

[0041] Example 4: This system is applied to the multi-dimensional monitoring of internal stress and thermal field evolution in large forgings. When handling the parallel access of 512 heterogeneous acquisition channels at a sampling frequency of 20kHz, the acquisition points are distributed within a hemispherical area with a radius of 2.0m. The topology control module determines the three-dimensional coordinates of the sensor layout in physical space. The data is converted into a one-dimensional node index j in the spatial topological correlation matrix. The conversion process is based on the discretization index of the data collection points along each coordinate axis. The calculation logic for the one-dimensional node index j is as follows: Where j is a one-dimensional node index, These are the axial indices of the data collection points within the discretized grid in physical space. These represent the preset maximum number of grids for each axis.

[0042] During the steady-state window before the start of hot processing, the system performs amplitude threshold calibration. The data alignment module continuously extracts 1000 frames of synchronization data blocks as background samples. The topology control module calculates the topology state feature vector of each node in steady state and calculates the standard deviation of its amplitude. ,in, The standard deviation of the amplitude at each node is expressed in quantization units; the incremental rendering module determines the preset amplitude threshold as follows: This filters out background random fluctuations with a signal-to-noise ratio below 15dB in the sampling channel, keeping the amplitude of the differential element in the steady-state background region below 0.2%.

[0043] To address signal jumps caused by sudden changes in local thermal stress during hot working, the convergence monitoring module calculates the first-order rate of change of node sampled values ​​with sampling time. and second-order rate of change When the second rate of change When the convergence threshold of 120 is exceeded, the topology control module increases the weight coefficients of the corresponding nodes in the spatial topology correlation matrix with a fixed step size of 0.02 to enhance the local response sensitivity. Under this condition, the absolute value of the differential feature vector extracted by the topology control module reaches 12.8%, which exceeds the value obtained by... The calculated dynamic threshold of 0.85% triggers the incremental rendering module to complete the local overwrite of the video memory; the reconstruction frequency of the visualized thermal field map is stable at 500Hz, and the accuracy of capturing transient thermal stress fluctuations that occur within 0.1s reaches 99.2%; during a continuous 48-hour operation test, the cumulative drift rate of the system calculation error is less than 0.05%, realizing the reconstruction of highly sparsity, high-frequency abrupt process data.

[0044] Example 5: In a newly deployed laser additive manufacturing application scenario, when the system faces a spatial correlation reconstruction situation caused by changes in the heterogeneous sensor arrangement scheme, the topology control module initiates the offline calibration process of the spatial topology correlation matrix. A controlled thermal pulse excitation source applies a constant 500W thermal shock at the geometric center point of the physical processing plane. The data alignment module simultaneously acquires the initial temperature response signal returned by the sensor array, and the topology control module calculates the response hysteresis time between adjacent nodes. and amplitude attenuation ratio ,in, Let be the response lag time between node i and node j. The amplitude attenuation ratio is calculated as follows: during the natural relaxation phase after the controlled thermal shock stops, the transient temperature extreme values ​​recorded by nodes i and j are locked through the data alignment module. The extreme value detected by the node far from the heat source center is divided by the extreme value detected by the node near the heat source center to extract an objective ratio characterizing the thermal resistance dissipation background of the physical conduction path medium. The system converts the measured values ​​into initial values ​​of weighting coefficients based on the discretized form of the heat conduction equation. Its calculation logic is determined as follows: ,in, is the weighting coefficient in the spatial topological correlation matrix, and k is the thermal diffusion constant coefficient, completing the static data filling of the spatial topological correlation matrix.

[0045] During the online self-inspection process before entering the hot processing production cycle, the convergence monitoring module performs a baseline consistency check on the filled spatial topology correlation matrix. The excitation unit sequentially simulates single-pulse input excitation vectors at the physical coordinates corresponding to the sensor points. The topology control module derives the topology state feature vector for the current period through matrix multiplication. By comparing the preset thermal field diffusion physical limit curve, when the energy diffusion gradient of a local node deviates from the theoretical calculation value by more than 2.5%, the topology control module calls the linear regression correction function to perform second-order fine-tuning of the weight coefficients of the corresponding region, so that the weight coefficients of each node in the spatial topology correlation matrix are within the range of 0.05 to 0.85. The system shows a numerical convergence trend when facing transient physical distortions in the subsequent processing, until the full reference frame is written to the backup storage module to complete the initial deployment.

[0046] Example 6: In a continuously reconstructing physical field scenario, when the system encounters transmission errors caused by on-site interference, the incremental rendering module performs spatial clustering and merging through its internal dirty rectangle refresh unit to calculate the topological state feature vector for the current period. Topological state feature vector of the previous period The difference is used to extract the set of element coordinates whose absolute value exceeds a preset amplitude threshold of 0.85%, and logical clustering is performed based on the distribution of element coordinates in the physical address space of video memory; the dirty rectangle refresh unit determines the maximum value of the horizontal coordinate by comparing the horizontal and vertical coordinate components of each point in the coordinate set. and minimum value and the maximum value of the ordinate. and minimum value ,in, and These represent the maximum and minimum values ​​of the x-axis, respectively. and These represent the maximum and minimum values ​​of the ordinate, with each coordinate component in pixels. During the establishment of the two-dimensional coordinate extrema, to prevent physical failures caused by the dimensionality reduction mapping leading to the scattering of nodes that were originally adjacent in three-dimensional space in two-dimensional memory, thus causing the bounding rectangle to encompass a large area of ​​invalid regions, the width of a single row of textures in memory is pre-electrically equal to the total number of nodes in a single layer of the three-dimensional discrete mesh. This ensures that adjacent distorted point sequences in the same layer of the three-dimensional thermal field maintain continuous memory distribution after two-dimensional mapping, allowing the minimum bounding rectangle determined by the extrema to accurately converge to the compact pixel clusters where abrupt changes actually occur. The system uses the coordinates of the top-left vertex... Coordinates of the bottom right vertex By establishing a single rectangular refresh area, scattered local addressing instructions are converted into instructions that continuously overwrite the physical address of the single rectangular refresh area, thereby improving the single data throughput of the video memory control unit and reducing the latency caused by random addressing.

[0047] The asynchronous snapshot module extracts the topology state feature vector at 500 sampling periods and writes it as a full reference frame to the backup storage module. Each full reference frame is accompanied by a 32-bit cyclic redundancy check (CRC) code for data integrity verification. In the specific execution link of reading the video memory check bits, the incremental rendering module silently maps a dedicated read-only memory page during the vertical blanking period of the graphics processing unit through the direct memory access control interface provided by the operating system. In this way, without interfering with the underlying video memory independent rendering pipeline and without initiating a high-cost synchronization mutex lock, the real-time CRC data of the hardware underlying layer is safely extracted. After receiving the topology state feature vector, the incremental rendering module reads the memory check bits corresponding to the target pixel area and calculates the real-time checksum. When the XOR logic output of the calculation result and the expected value is observed to be non-zero, it is determined that there is a data verification anomaly in the video memory storage module and the current differential rendering pipeline is interrupted. The system retrieves the full reference frame with the latest timestamp from the backup storage module and resets the data state of the video memory storage module using a full address overwrite method, so that the reconstruction accuracy of the visualized thermal field map is restored to the physical consistency level within 20ms.

[0048] In the real-time output of the visualized thermal field map, the coordinates of the differential feature vector elements extracted by the incremental rendering module are mapped to texture coordinates in the video memory storage module by the shader program. The mapping mechanism is based on the correspondence between one-dimensional linear addressing of discrete data and two-dimensional spatial rasterization in computer graphics. The shader program reads the differential feature vector elements, including the one-dimensional node index j, as the initial input parameter and completes the two-dimensional transformation of the address space through a sequence of modulo and integer division operations. Specifically, the shader program follows the formula... Calculate the horizontal coordinate u of the target pixel using the formula Calculate the vertical coordinate v, where the parameter is... The maximum texture pixel width of a single row in the current thermal field map is determined by the dimensionless positive integers u and v, which are both non-negative integers. Together, they constitute the target pixel data frame index in the video memory storage module. The hardware addressing unit aligns the continuous one-dimensional state feature data to the terminal physical display matrix. The dirty rectangle refresh unit merges multiple adjacent coordinate sets into the minimum bounding rectangle. Then, the video memory control unit locks the starting address of this rectangular region. With offset After receiving the signal that the local coverage is complete, the situation output module calls the texture mapping operator to rasterize the data blocks that have changed in the video memory storage module. Since the overwrite operation only applies to the set of pixels covered by the calculated physical address index, the display driver no longer initiates a full video memory refresh instruction. When facing a 1024×768 resolution thermal field visualization task, the system's video memory bandwidth utilization rate is less than 15% of the theoretical peak. The transient thermal conduction wavefront under rolling conditions exhibits a smooth gradient evolution characteristic on the monitoring screen.

[0049] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A digital data processing and visualization system for thermal processing technology data, characterized in that, include: The data alignment module is used to acquire multi-source timing signals during the hot processing process, and to complete the time domain alignment of each source timing signal and the encapsulation of synchronous data blocks based on the global system clock; The topology control module, connected to the data alignment module, stores a preset spatial topology correlation matrix. The spatial topology correlation matrix consists of multiple nodes representing the correlation between sampling points and weight coefficients between nodes. The weight coefficients represent the correlation strength between adjacent nodes at the data level. After receiving the synchronization data block, the topology control module maps the values ​​of each sampling point in the synchronization data block to the corresponding node address in the spatial topology correlation matrix, completes the feature assignment of the node, and calculates the topology state feature vector of the current period. The topology state feature vector of the current period is equal to the product of the spatial topology correlation matrix and the topology state feature vector of the previous period, plus the input excitation vector generated by the mapping of the synchronization data block. The incremental rendering module, connected to the topology control module, is used to calculate the difference feature vector between the topology state feature vectors of two consecutive cycles, and extract the coordinates of elements whose absolute values ​​exceed a preset amplitude threshold in the difference feature vector. Based on the element coordinates, the module addresses the target pixel region in the video memory storage module to complete local data coverage.

2. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The weight coefficient ranges from 0.05 to 0.85, and the topology control module dynamically corrects the weight coefficient of the corresponding node by using a linear interpolation algorithm based on the rate of change of the values ​​of each sampling point in the synchronization data block with the sampling time.

3. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The internal operation logic of the data alignment module includes: step S31, extracting the sampled values ​​and channel identifiers from the multi-source timing signals through the signal parsing unit; step S32, encapsulating the sampled values ​​within the same sampling window into the corresponding address bits of the synchronization data block according to the system clock reference through the clock alignment unit.

4. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The incremental rendering module includes an addressing unit and a video memory control unit. The addressing unit is used to convert the coordinates of elements in the differential feature vector whose absolute value exceeds a preset amplitude threshold into physical address indices of the video memory storage module. The video memory control unit is used to overwrite the target pixel data block based on the physical address index and keep the data state of non-indexed addresses in the video memory storage module unchanged.

5. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, It also includes an asynchronous snapshot module, which is connected to the topology control module. It is used to extract the complete topology state feature vector at a preset time interval and write it as a full reference frame to the backup storage module. When the incremental rendering module determines that there is a data verification anomaly in the video memory storage module, it calls the most recent full reference frame to complete the full address overwrite of the video memory storage module.

6. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The topology control module also includes a convergence monitoring module, which optimizes the data processing accuracy through the following logic: Step S61, extract the set of nodes whose values ​​exceed a preset high threshold in the topology state feature vector of the current period; Step S62, calculate the numerical spatial gradient between adjacent nodes in the node set; Step S63, when the numerical spatial gradient exceeds the preset convergence threshold, increase the weight coefficient of the corresponding node in the spatial topology correlation matrix by a preset fixed step size.

7. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The incremental rendering module integrates a dirty rectangle refresh unit, which is used to improve rendering efficiency through the following logic: Step S71, using the minimum bounding rectangle algorithm, the coordinates of multiple adjacent elements with absolute values ​​exceeding a preset amplitude threshold are merged into a single rectangle refresh area. Step S72: Send the rectangular refresh area as a single addressable object to the video memory storage module to complete the batch overwrite.

8. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, The sampled values ​​in the synchronous data block include temperature data, pressure data, or acoustic emission intensity data; the node topology of the spatial topology correlation matrix is ​​established as a three-dimensional static mapping based on the layout coordinates of the data acquisition points in physical space.

9. The electrical digital data processing and visualization system for thermal processing process data according to claim 1, characterized in that, It also includes a situation output module, which is connected to the video memory storage module. After receiving the local coverage completion signal from the incremental rendering module, the situation output module generates a visualized thermal field map based on the real-time data in the video memory storage module.

Images