Aluminum powder particle size distribution analysis method and system based on big data

Abstract

The application discloses an aluminum powder particle size distribution analysis method and system based on big data, and relates to the technical field of data processing, and comprises the following steps: acquiring light transmittance attenuation data of a particle size detection device, calculating a light transmittance attenuation rate based on the light transmittance attenuation data to determine a target pollution stage; determining a target protection state of a compensation parameter based on the light transmittance attenuation rate and cumulative adjustment data of the compensation parameter; receiving reference parameters issued by a second terminal, and when the target pollution stage is a first pollution stage and the target protection state is a first protection state, the reference parameters are used to overwrite the compensation parameter; when the target pollution stage is a second pollution stage and the target protection state is a second protection state, the overwriting of the compensation parameter is limited, and particle size inversion calculation is performed in combination with the reference parameters; and the continuity and accuracy of particle size distribution analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for analyzing the particle size distribution of aluminum powder based on big data. Background Technology

[0002] Laser particle size analyzers are the mainstream equipment for online detection in aluminum powder atomization and grading production lines. The detection accuracy is highly dependent on the light transmittance of the optical window. In high-concentration dust environments, an irreversible aluminum powder layer will form on the window wall, causing nonlinear attenuation of the light signal. Furthermore, the contamination rate varies significantly between different production lines, and peripheral equipment needs to adjust compensation parameters to correct signal deviations.

[0003] Existing technologies use a cloud-based architecture to generate global baseline parameters and periodically distribute them in full to unify quality standards. However, the cloud cannot perceive the differences in contamination between production lines. A full overwrite will clear the local valid compensation parameters, causing periodic fluctuations in detection accuracy, leading to product misjudgment and process misadjustment, and affecting the stability of the production system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for analyzing aluminum powder particle size distribution based on big data.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for analyzing the particle size distribution of aluminum powder based on big data, applied to a first terminal. The method includes the following steps: Acquire transmittance decay data from particle size detection equipment, and calculate transmittance decay rate based on the transmittance decay data to determine the target contamination stage; Based on the cumulative adjustment data of the transmittance attenuation rate and the compensation parameters, the target protection state of the compensation parameters is determined; The reference parameters sent by the second terminal are received. When the target pollution stage is the first pollution stage and the target protection state is the first protection state, the reference parameters are used to overwrite the compensation parameters. When the target pollution stage is the second pollution stage and the target protection state is the second protection state, the overwriting of the compensation parameters is restricted, and granular inversion calculation is performed in conjunction with the benchmark parameters.

[0006] As a preferred embodiment of the present invention, determining the target protection state of the compensation parameter based on the cumulative adjustment data of the transmittance attenuation rate and the compensation parameter includes: When the transmittance decay rate is greater than the first threshold, the target coefficient is determined based on the ratio of the first threshold to the transmittance decay rate. The second threshold is reduced using the target coefficient to obtain the third threshold; When the cumulative adjustment data is greater than the third threshold, the target protection state of the compensation parameter is determined as the second protection state.

[0007] As a preferred embodiment of the present invention, the step of acquiring transmittance attenuation data from a particle size detection device and calculating the transmittance attenuation rate based on the transmittance attenuation data to determine the target contamination stage includes: The continuous operating time of the particle size detection equipment since the most recent cleaning is obtained; The transmittance attenuation data is fitted using a sliding window to calculate the transmittance attenuation rate. The transmittance attenuation data and the continuous running time are matched with a preset pollution feature matching library to determine the target pollution stage.

[0008] As a preferred embodiment of the present invention, the limitation on overwriting the compensation parameter includes: When the target contamination stage is the second contamination stage and the target protection state has not reached the second protection state, the reference parameter is written to the first storage area, and the compensation parameter in the second storage area is locked to intercept write operations to the second storage area. When the target protection state reaches the second protection state, the write operation of the reference parameters sent by the second terminal for the first storage area and the second storage area is intercepted, and an exemption request information carrying the compensation parameters is generated and uploaded to the second terminal.

[0009] As a preferred embodiment of the present invention, the step of performing granularity inversion calculation in conjunction with the reference parameters includes: The reference parameters are read from the first storage area, and the compensation parameters are read from the second storage area. The fusion weight ratio is determined based on the target pollution stage and the cumulative adjustment data; The baseline parameter and the compensation parameter are linearly superimposed according to the fusion weight ratio. The superposition result is input into the granularity inversion calculation model, and a granularity distribution report is output. The granularity distribution report includes frequency distribution curve, cumulative distribution curve and characteristic particle size parameter.

[0010] As a preferred embodiment of the present invention, the method further includes: Count the number of times the second terminal continuously ignores the exemption request information; When the number of ignored attempts exceeds the fourth threshold, an exception reporting operation is triggered. The compensation parameter is forcibly configured to a locked protection state, and all subsequent writes of the reference parameter are blocked while the compensation parameter is in the locked protection state.

[0011] As a preferred embodiment of the present invention, the method further includes: Perform a validity check on the transmittance sensor used to collect the transmittance attenuation data; When the duration for which the sensor data output by the transmittance sensor has not been updated exceeds the fifth threshold or the magnitude of the change exceeds the sixth threshold, the determination of the target contamination stage is paused. Freeze the migration operation of the target protection state of the compensation parameters and enforce the step of restricting the overwriting of the compensation parameters.

[0012] As a preferred embodiment of the present invention, the method further includes: When a cleaning or maintenance signal or equipment maintenance signal is detected for the particle size detection device, the target contamination stage is forcibly reset to the first contamination stage; The compensation parameter is cleared to zero, and the cumulative adjustment data is reset to the initial zero value.

[0013] As a preferred embodiment of the present invention, after the exemption request information carrying the compensation parameters is generated and uploaded to the second terminal, the method further includes: Receive the updated baseline parameters sent by the second terminal; The updated baseline parameters are generated by the second terminal after parsing the exemption request information and extracting the compensation parameters, and configuring the compensation parameters as fixed parameters to constrain its model iterative optimization process.

[0014] This invention also discloses a big data-based aluminum powder particle size distribution analysis system, comprising: Memory, used to store computer programs; A processor for implementing the above methods when executing the computer program.

[0015] The beneficial effects of this invention are: 1. This invention determines the target contamination stage by using transmittance attenuation data from a particle size detection device, determines the target protection state of the compensation parameters based on the transmittance attenuation rate and the cumulative adjustment data of the compensation parameters, and performs differentiated overwrite control on the reference parameters issued by the second terminal according to the combination relationship between the target contamination stage and the target protection state. It establishes a correspondence between the physical attenuation process of the device and the parameter update permission, which solves the problem of indiscriminate full overwrite and forced erasure of local effective compensation data in the prior art. Under the premise of ensuring the periodic synchronization of the global process reference, it effectively protects the compensation parameters of the edge device for the local contamination state.

[0016] 2. When the target contamination stage is the first contamination stage and the target protection state is the first protection state, the present invention uses benchmark parameters to overwrite compensation parameters to ensure that the global general process optimization results can be synchronized to the edge production line in a timely manner. When the target contamination stage is the second contamination stage or the target protection state is the second protection state, the overwriting of compensation parameters is restricted, and particle size inversion calculation is performed in combination with benchmark parameters. This avoids the distortion of aluminum powder particle size detection results caused by parameter mutations, prevents erroneous detection data from misleading the upstream atomization and classification process, and improves the continuity and accuracy of particle size distribution analysis. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0018] Figure 1 This is a schematic diagram of the workflow of the aluminum powder particle size distribution analysis method of the present invention. Detailed Implementation

[0019] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] like Figure 1 As shown, the main body of the aluminum powder particle size distribution analysis method based on big data is the first terminal. In the continuous atomization production scenario of aluminum powder and other metal powders, the first terminal can be an edge online detection computer deployed on the production line. Its hardware is equipped with a digital signal processor and storage area, which is used to communicate directly with on-site detection equipment such as laser particle size analyzers, and undertake real-time signal acquisition and edge computing tasks.

[0022] During the physical data acquisition stage, the first terminal acquires the transmittance attenuation data of the particle size detection device. As aluminum powder is a micron / nano-sized powder with high surface energy and easy agglomeration characteristics, it will inevitably adhere to the optical window of the detection device during the detection process, resulting in a continuous decrease in transmittance.

[0023] The first terminal calculates the transmittance decay rate based on the transmittance decay data to determine the target contamination stage.

[0024] During the atomization process of aluminum powder, fluctuations in the protective gas flow rate or abnormal increases in powder temperature can cause a sharp increase in the adhesion rate of aluminum powder particles on the optical window, resulting in a sudden increase in the rate of light transmittance decay.

[0025] Therefore, by acquiring and calculating this rate, this application is essentially sensing the degree of abrupt changes in the flow field and the physical state of aluminum powder within the production line in real time.

[0026] To address the signal distortion caused by this physical contamination, the first terminal continuously generates compensation parameters for correction during local operation.

[0027] Since aluminum powder dust pollution has a significant difference in its blocking effect on scattered light at small and large angles, window pollution often leads to severe absorption of scattered light energy at the fine powder end. The locally accumulated compensation parameters are specifically designed to increase the light energy gain at the fine powder end.

[0028] Based on the transmittance attenuation rate calculated above and the cumulative adjustment data of the compensation parameter, the first terminal determines the target protection state of the compensation parameter.

[0029] This step directly links physical parameters reflecting the hardware degradation process to the state determination logic of software parameters, assessing the maturity of the current local compensation data in adapting to local adverse operating conditions.

[0030] When receiving baseline parameters periodically generated from training based on global historical data by a second terminal (such as an enterprise cloud big data platform), the first terminal no longer performs indiscriminate full data synchronization. Instead, it performs dynamic access control based on the currently determined target pollution stage and target protection status. When the target contamination stage is the first contamination stage and the target protection state is the first protection state, it indicates that the current window contamination is relatively light and the local compensation parameters have not yet formed a stable correspondence.

[0031] At this point, the system allows the compensation parameters to be overridden using the baseline parameters in order to incorporate general process standards from a global perspective.

[0032] When the target contamination stage is the second contamination stage or the target protection state is the second protection state, it indicates that the equipment has entered the accelerated contamination period, or the local compensation parameters are highly adapted to the current extreme adhesion conditions.

[0033] If the ideal baseline parameters in the cloud are forcibly overwritten at this time, the system will lose its ability to compensate for the blocked signal, thus failing to accurately detect the fine powder content and incorrectly outputting a coarser particle size distribution (e.g., the D10 characteristic particle size is abnormally large). This will mislead the upstream control system to make incorrect atomization pressure adjustments, and may even lead to a degradation of the quality of the entire batch of aluminum powder.

[0034] Therefore, under this condition, the first terminal restricts the overwriting of the compensation parameters, retains the local compensation data, and performs the final granular inversion calculation in conjunction with the reference parameters.

[0035] Therefore, under this condition, the first terminal restricts the overwriting of the compensation parameters, retains the local compensation data, and performs the final granular inversion calculation in conjunction with the reference parameters.

[0036] Furthermore, in actual production settings where aluminum powder particle size analysis is conducted online, using fixed amplitude thresholds to control the switching of compensation parameters often leads to control failure issues. When the optical window of the detection device suffers a severe decrease in light transmittance due to the adhesion of high concentration of aluminum powder, the first terminal needs to quickly increase the local compensation parameter to maintain the light signal strength. However, due to the fixed high threshold setting, the compensation parameter is difficult to accumulate to a value sufficient to trigger the protection mechanism in a short period of time, and is very likely to be incorrectly overwritten and cleared to zero when the second terminal sends the parameter.

[0037] To address the aforementioned issues, in practice, the processor periodically or periodically determines whether the calculated transmittance attenuation rate is greater than a first threshold, which reflects the physical critical point at which the device leaves the normal stable wear period and enters an accelerated contamination state.

[0038] During engineering deployment, the first threshold is set based on the statistical distribution of attenuation rate in the historical operation records of multiple production lines of the same type of particle size detection equipment. It is usually taken as the 75th percentile value of the historical wear rate data. The specific value is usually set as 2.0% to 3.0% decrease in transmittance per thousand hours.

[0039] When the transmittance attenuation rate is greater than the first threshold, it indicates that the physical contamination of the current optical window has worsened. At this time, the processor determines the target coefficient based on the ratio of the first threshold to the transmittance attenuation rate, and the specific calculation formula is as follows: ; Among them, C tT1 represents the target coefficient; V represents the first threshold; d This represents the transmittance attenuation rate calculated in real time.

[0040] To ensure extremely high system stability, a minimal zero-bias constant ε is introduced into the denominator of the formula. This constant is typically set to 10. -6 This is to prevent the first terminal from causing processor calculation errors when it obtains an abnormal zero value during the system initialization phase or due to a momentary interruption in sensor communication.

[0041] Based on this logical architecture, satisfying V d The target coefficient C is calculated under the trigger condition that is greater than T1. t It is a scaling factor that is always between 0 and 1, and the more severe the physical attenuation of the light transmittance of the device, the smaller the value of this coefficient.

[0042] Subsequently, the processor uses the target coefficient to reduce the second threshold to obtain the third threshold. The actual linear scaling logic it executes is as follows: ; Where T3 represents the calculated third threshold; T2 represents the second threshold.

[0043] The second threshold is the minimum cumulative adjustment amount required for the compensation parameter to move from the first protection state that allows overwriting to the second protection state that restricts overwriting under standard stable operating conditions. Its value is determined according to the allowable fluctuation range of the factory-calibrated gain value of each scattering angle channel, for example, taking 15% to 30% of the reference gain value. After dynamic mapping of the target coefficient, the generated third threshold T3 is significantly compressed.

[0044] The processor extracts the cumulative adjustment data set for each scattering angle channel and compares it with the third threshold in the processor storage unit. When the cumulative adjustment data is greater than the third threshold, the target protection state of the compensation parameter is determined as the second protection state.

[0045] For example, in a continuous production batch of a certain type of aluminum powder for 3D printing, the first terminal detects that the light transmittance decay rate exceeds the first threshold, indicating that the window is being rapidly contaminated by aluminum powder. At this time, the system immediately triggers the above formula logic, greatly compressing the third threshold required for state transition, so that the local compensation parameters for the high-frequency scattering channel only need minimal adjustment to reach the second protection state.

[0046] The system thus triggered the protection mechanism in advance, intercepting the reference parameters sent by the second terminal. By retaining the local compensation parameters for the high-frequency scattering channel and combining them with the reference parameters to perform Mie scattering inversion, the system successfully restored the true fine powder content, avoiding the illusion of missing fine powder caused by forced synchronous overwriting. This avoided the distortion of abnormally large D10 feature particle size, thereby ensuring the accuracy of the evaluation of the powder spreading performance of this batch of 3D printed aluminum powder and the stability of continuous production.

[0047] Furthermore, the processor of the first terminal obtains the continuous running time of the granularity detection device since the most recent cleaning by reading the internal real-time clock and system reset interrupt log. This duration data is recorded in the local storage area as a time axis anchor point for evaluating the hardware aging life cycle of the device.

[0048] Subsequently, the processor retrieves the transmittance decay data sequence within the current preset time window from the storage area. To filter out high-frequency turbulence noise and extract the true degradation trend, the processor performs a sliding window fitting calculation on the transmittance decay data and uses the univariate linear least squares method to extract the decay slope within the time window. The calculation formula is as follows: ; Among them, V d The calculated transmittance decay rate is represented by the absolute value of the slope to indicate the severity of degradation; N represents the total number of data sampling points within the sliding window, the value of which is set according to the stability requirements of the aluminum powder production process, typically ranging from 36 to 72, corresponding to a time span of 6 to 12 hours, with a sampling interval of 10 minutes. This window capacity is sufficient to smooth out most transmittance spikes caused by process fluctuations; t i y represents the continuous runtime time stamp of the i-th sampling point; i This represents the observed true transmittance value corresponding to the i-th sampling point.

[0049] The physical meaning of the above formula is as follows: In a scatter plot of transmittance versus time, composed of N data points, the V calculated by this formula... d This is the slope of the linear fit of the scatter points (transmittance decay rate). The numerator of the formula is N multiplied by the sum of the products of the running time and transmittance of all sampling points minus the product of the sum of the running time and the sum of the transmittance. Its value reflects the covariance trend between running time and transmittance. When the transmittance continues to decrease with the increase of running time, the numerator is negative, and the larger its absolute value, the more significant the decreasing trend is. The denominator of the formula is N multiplied by the sum of the squares of the running times of all sampling points minus the square of the total running times. Its value reflects the degree of dispersion on the time axis. The ratio of the numerator to the denominator is the slope of the fitted straight line. Taking the absolute value yields the transmittance decay rate V.d .

[0050] After acquiring the continuous running time and transmittance decay rate, the processor matches the transmittance decay data and the continuous running time with a preset pollution feature matching library to determine the target pollution stage.

[0051] The internal structure of the preset contamination feature matching library is explained in detail below: In terms of data sources, the data in this matching library is based on the statistical analysis of actual test data from multiple historical aluminum powder production batches using the same type of particle size detection equipment, combined with the benchmark life loss curves provided by optical component suppliers. After aggregation and offline calculation, the data is solidified into the read-only memory of the first terminal. At the data structure level, the matching library is represented as a correspondence table in physical storage. Its input key is a joint evaluation parameter group containing the continuous running time interval, the current transmittance threshold, and the transmittance decay rate interval. Its output value is a discrete stage identifier, including the first contamination stage (representing the initial stage of stable wear) and the second contamination stage (representing the accelerated deterioration period). At the call logic level, the processor performs a conditional interval comparison based on the current real-time input continuous runtime, the current transmittance value, and the transmittance decay rate calculated by fitting, through conditional branch judgment instructions.

[0052] For example, when the system determines that the current transmittance decreases by more than 10% of the nominal value, and the transmittance decay rate V d When the value is greater than the historical median decay value, the identification result of the second pollution stage is accurately matched and output.

[0053] Similarly, for example, in a continuous production batch of a certain type of aluminum powder for 3D printing, the equipment was stable in the early stage of production, but as production progressed, slight fluctuations in the protective gas pressure caused the aluminum powder to quickly adhere to the optical window. At this point, the processor of the first terminal quickly captures V using the aforementioned sliding window formula. d The rapid increase in the concentration of aluminum powder, and the accurate determination by the preset contamination feature matching library, indicate that the device has entered the second stage of contamination, suggesting that the window is being rapidly contaminated by aluminum powder.

[0054] At this point, the system intercepts the baseline parameters sent by the second terminal based on the target contamination stage. By retaining the local compensation parameters for the high-frequency scattering channel and combining them with the baseline parameters to perform Mie scattering inversion, the system restores the large-angle scattered light signal that was severely weakened due to dust obstruction, restores the true fine powder content, and avoids the false appearance of missing fine powder caused by forced synchronization. In other words, it avoids the distortion of abnormally large D10 feature particle size in the particle size report. From hardware to algorithm, it completely ensures the accuracy of the evaluation of the powder spreading performance of this batch of 3D printed aluminum powder and the stability of the process.

[0055] Furthermore, in order to achieve precise linkage between physical decay status and data update permissions, the non-volatile storage medium of the first terminal is divided into two mutually isolated regions in terms of physical or logical addressing. The first storage area is dedicated to storing the reference parameters periodically issued by the second terminal, while the second storage area is dedicated to maintaining locally generated compensation parameters adapted to physical wear. The processor implements independent access control for the two areas through independent read / write control instructions and a data bus.

[0056] During continuous operation of the device, the processor performs dynamic permission allocation operations.

[0057] When the target contamination stage is determined to be the second contamination stage and the target protection state has not reached the second protection state, it indicates that the optical window is in a severe condition of accelerated dust adhesion and contamination. However, the cumulative adjustment of the compensation parameters for the specific scattering channel is still in the growth stage and has not reached a fully mature and stable level.

[0058] Under this condition, the processor takes a partial lock-down defensive action: Open the write enable interface of the first storage area and write the received benchmark parameters normally into the first storage area to maintain the iteration of the basic measurement matrix; Simultaneously, the write enable signal of the second storage area is canceled, and the compensation parameters in the second storage area are locked, thereby intercepting write operations to the second storage area at the hardware level.

[0059] When the target protection state reaches the second protection state, it indicates that the compensation parameter has become highly mature and has formed a close correspondence with the current physical window's pollution state.

[0060] At this point, the general reference parameters from the second terminal have completely lost their applicability on this specific signal channel. If forced writing is attempted, it will result in severe detection distortion. Therefore, the processor instead performs a complete lock action, intercepting all write operations of the reference parameters sent by the second terminal to the first and second storage areas.

[0061] After the processor performs the interception action, it generates exemption request information carrying the compensation parameters and uploads it to the second terminal. The exemption request information is manifested at the underlying level as a message data structure with a specific frame header identifier. Its data comes from a locked storage unit in the second storage area and encapsulates the hardware device identifier of the current first terminal, the channel index of the specific scattering angle where extreme attenuation occurs, and the actual value of the currently locked compensation parameters.

[0062] The reporting logic of this exemption request information enables the first terminal not only to perform passive data defense at the edge, but also to announce the local physical specific state to the second terminal, providing a high-confidence field sample for subsequent global parameter collaborative optimization.

[0063] For example, when the optical window is being rapidly contaminated by aluminum powder, the absorption and blocking of large-angle scattered light by the aluminum powder dust layer is particularly severe. The compensation parameters for the high-frequency scattering channel that the first terminal locally increases are the only reliance for maintaining the detection rate of fine powder signals. At this time, the system triggers the protection mechanism based on the target protection status of the current parameters, and rejects the overwrite operation of the second memory area by the reference parameters sent by the second terminal through the bus interception command; By fixing the local compensation parameters for the high-frequency scattering channel in memory and then combining them with the benchmark parameters for Mie scattering inversion calculation, the system successfully restored the true fine powder content from the weak scattered light, avoiding the false impression of missing fine powder caused by forced data synchronization. This eliminated the distortion problem in the report where the D10 feature particle size was misjudged as abnormally coarse, thus ensuring the accuracy of the evaluation of the powder spreading performance of this batch of 3D printed aluminum powder and preventing the production accident of qualified products being wrongly judged as downgraded products.

[0064] Furthermore, the processor reads the reference parameters from the first memory area and the compensation parameters from the second memory area through independent addressing instructions.

[0065] Subsequently, the processor enters the calculation logic for the fusion weight ratio, the core of which is: The more severe the physical degradation of the equipment and the more comprehensive the locally accumulated correction data, the higher the weight of the local compensation parameters should be in the final calculation. The processor determines the fusion weight ratio based on the target contamination stage and the accumulated adjustment data, and the algorithm formula called is as follows: ; Among them, W f D represents the calculated fusion weight ratio; a This represents the cumulative adjustment data; D m This represents the preset full-scale deflection adjustment threshold, a fixed constant reflecting the maximum allowable parameter adjustment amount before the equipment is forced to shut down for maintenance; k s This represents the stage coefficient that maps to the target pollution stage; The query logic for this coefficient is manifested as key-value pair matching: when the target pollution stage is the initial first pollution stage, k s Set to 0.2; when the target pollution stage enters the second pollution stage of the deterioration phase, k s Adjusted to 0.85.

[0066] After obtaining the fusion weight ratio, the processor linearly superimposes the baseline parameter and the compensation parameter according to the fusion weight ratio, specifically through the following superposition formula: ; Among them, P m P represents the superposition result generated after linear superposition processing; c P represents the compensation parameter; b This refers to the reference parameter.

[0067] The effect of dust adhesion on the intensity of scattered light in a specific detection channel is, in essence, a complex nonlinear absorption and diffraction superposition process. The theoretical basis of the above formula is that the gain compensation of each scattering angle channel of the laser particle size analyzer approximately satisfies the linear time-invariant assumption within a limited dynamic adjustment range.

[0068] Engineering experiments have proven that the inversion deviation caused by this linear superposition equivalent model can be strictly controlled within a system tolerance range of 3%, which is sufficient to meet the high-frequency, low-latency online quality screening requirements of the production line.

[0069] After obtaining the superposition results, the processor inputs the superposition results into the particle size inversion calculation process. The particle size inversion calculation process is embodied in the Mie scattering solution matrix library solidified in the storage area. The processor uses the superposition results containing local physical attenuation compensation characteristics to perform deconvolution on the real-time acquired original light energy scattering array and finally outputs the particle size distribution report of the batch of materials.

[0070] The particle size distribution report is presented in the system interface or uploaded to the manufacturing execution system in the form of a standard data frame. Its data structure clearly includes a frequency distribution curve that intuitively reflects the content ratio of each particle size range, a cumulative distribution curve that reflects the overall coarse and fine distribution trend, and characteristic particle size parameters used for core quality determination, such as D10 representing the fine powder content, D50 representing the median particle size, and D90 representing the coarse powder boundary.

[0071] In specific scenarios considering the characteristics of aluminum powder production processes, the aforementioned anti-overwriting mechanism based on dynamic fusion algorithms possesses strong practical defense value. Because aluminum powder itself has extremely strong light absorption and scattering characteristics, when the viewport is contaminated, the light energy signal attenuation is particularly severe at large scattering angles (physically corresponding to fine powder particles). In a continuous production batch of a certain type of 3D printing-specific aluminum powder, the first terminal detected that the transmittance attenuation rate exceeded the first threshold, indicating that the viewport was being rapidly contaminated by aluminum powder. At this point, the system intercepted the baseline parameters sent by the second terminal.

[0072] Using the weighted fusion formula described above, the system performs Mie scattering inversion in conjunction with the reference parameters. This control flow enables the system to successfully restore the true fine powder content, avoiding the illusion of missing fine powder caused by the loss of fine powder signal gain due to forced cloud synchronization. This eliminates the misjudgment of abnormally coarse D10 feature particle size in the particle size distribution report, ensuring the accuracy of the evaluation of the powder spreading performance of this batch of 3D printed aluminum powder and effectively reducing the product misjudgment and downgrade rate.

[0073] Furthermore, if the first terminal passively intercepts local writes and sends an exemption request to the cloud, and the cloud fails to process the request in a timely manner and reflect it in subsequent parameter update packages, the first terminal will be trapped in a cycle of continuous requests and continuous receipt of overwrite instructions.

[0074] To address the aforementioned issue, after each generation and upload of the exemption request information, the processor of the first terminal initiates an internal communication monitoring program to count the number of times the second terminal consecutively ignores the exemption request information. The counting method is as follows:

[0075] If, within multiple consecutive parameter transmission cycles, the received reference parameter still attempts to overwrite the protected scattering angle channel parameter without carrying a protection confirmation flag for that channel, it is determined that an omission has occurred, and the omission count is incremented in the counter.

[0076] When the number of ignored occurrences exceeds a fourth threshold, it indicates that the collaboration mechanism between the cloud and the edge has substantially failed. This fourth threshold is typically set based on the product of the production line's maximum tolerable misalignment time and the cloud synchronization cycle, usually ranging from 3 to 5 occurrences. At this point, the processor triggers an anomaly reporting operation, sending a data packet containing the current device hardware status, the index of the continuously ignored channel, and the error code as a high-priority alarm directly to the production line monitoring terminal and the enterprise manufacturing execution system to prompt maintenance personnel for manual intervention.

[0077] The processor directly sends an instruction to the memory control register to force the compensation parameter into a locked protection state. In this locked protection state, the processor intercepts all subsequent write requests for the reference parameter at the protocol stack level of the hardware communication interface. This means that no matter what update instructions are subsequently issued from the cloud, the local compensation parameter will be fixed until it is reset after manual inspection.

[0078] In the context of aluminum powder production process, during a continuous production batch of a certain type of 3D printing aluminum powder, the first terminal detected that the light transmittance decay rate exceeded the first threshold, indicating that the window was being rapidly contaminated by aluminum powder. At this time, the system triggered a protection mechanism, intercepted the baseline parameters sent by the second terminal, and reported an exemption request. However, due to a fiber optic network failure in the factory area, the cloud failed to include the exemption request in the new round of parameter calculations, resulting in several subsequent attempts to erase the local compensation baseline parameters. The first terminal detected this anomaly through the aforementioned ignore count statistics logic.

[0079] When the number of ignored attempts exceeds the fourth threshold, the system will completely lock the local compensation parameters for the high-frequency scattering channel, blocking all subsequent cloud writes. By retaining these local compensation parameters, which are crucial for analyzing fine powder particles, the system successfully restored the true fine powder content during the several days of cloud anomalies.

[0080] Furthermore, if the processor of the first terminal directly accepts the abnormal sensor signal without verification, it will cause the system to seriously misjudge the physical state, which will lead to the local compensation parameters being incorrectly exposed to the cloud overwrite instruction.

[0081] To address the aforementioned issues, the processor of the first terminal continuously monitors the operating status of the transmittance sensor used to collect the transmittance attenuation data via a driver interface. The processor performs a dual comparison of the timestamps and values ​​in the input buffer queue.

[0082] When the duration for which the sensor data output by the transmittance sensor has not been updated exceeds the fifth threshold, which is usually set to three consecutive sampling cycles, it indicates that the sensor bus communication is interrupted or the system has crashed. Alternatively, if the change in the sensor data within an adjacent sampling period exceeds the sixth threshold, which is set according to the limit rate of physical attenuation of the device, a change exceeding this limit rate is considered a jump that violates the physical law of gradual dust adhesion, usually indicating that the probe is mechanically blocked or the circuit is short-circuited.

[0083] When any of the above conditions are triggered, the processor immediately determines that the reliability of the data currently output by the sensor is insufficient and triggers an exception handling process.

[0084] After the exception handling is triggered, the system is forced to enter a conservative operation strategy. The processor suspends the determination of the target pollution stage and no longer generates a new pollution stage identifier based on the current abnormal data. Instead, it latches and maintains the historical state identifier that was last verified as reliable.

[0085] The processor freezes the transition operation of the target protection state of the compensation parameters through state control instructions, cutting off the path of protection level downgrade due to numerical anomalies.

[0086] The system enforces the step of overriding the compensation parameters under the aforementioned restrictions. The engineering basis for this mechanism is as follows: When an edge detection node is unable to acquire valid data due to sensor failure, the safest control strategy is to assume that it is already in a poor working condition and activate write protection to preserve the existing physical compensation correspondence in local storage.

[0087] Through this conservative defense during hardware failures, the first terminal retained the local compensation parameters for the high-frequency scattering channel and continued to perform Mie scattering inversion in conjunction with the benchmark parameters. While maintenance personnel were troubleshooting and repairing the sensors, the system still successfully restored the true fine powder content, avoiding the false impression of missing fine powder caused by the cancellation of local protection and forced synchronization of parameters due to erroneous data. This eliminated the risk of the D10 feature particle size being misjudged as abnormally coarse in the report, ensuring the accuracy of the evaluation of the powder spreading performance of this batch of 3D printed aluminum powder and maintaining the smooth operation of the production line.

[0088] Furthermore, when necessary manual interventions are implemented on-site, such as cleaning the optical window, replacing the aging laser, or overhauling the classifier impeller, the physical transmittance or mechanical properties of the particle size detection equipment will undergo a sudden recovery. If the control system fails to detect this sudden change in physical state and continues to use the parameter protection strategy accumulated during the high-pollution stage, it will lead to serious detection distortion.

[0089] To this end, the processor of the first terminal continuously monitors the external industrial communication bus of the equipment, or periodically polls the production line maintenance record database. When on-site maintenance personnel enter maintenance completion instructions through the operating terminal, or when the sensor detects that a specific hardware module has been re-inserted or recalibrated, the first terminal determines that it has received a cleaning maintenance signal or equipment maintenance signal for the particle size detection equipment.

[0090] Upon detecting the aforementioned signal, it indicates that the physical foundation of the current device has deviated from the previous continuous decay trajectory and recovered to a healthy or near-new calibration state. At this point, the processor directly performs a write operation on the status register in the storage area, forcibly resetting the target contamination stage to the first contamination stage. This operation logically reinitializes the hardware lifecycle, enabling subsequent transmittance decay assessments to be recalculated based on a completely new physical baseline.

[0091] Simultaneously, the processor performs data cleanup actions on the local storage: The compensation parameter in the second storage area is cleared to zero, that is, restored to the factory default no-gain calibration state, and the cumulative adjustment data bound to the parameter is reset to the initial zero value. After the physical window is cleaned, the parameter compensation that was originally used to resist dust blockage has lost its physical basis. If these historical cumulative data are not cleared to zero, the excessive signal compensation will directly cause the light energy signal collected by the detector to be erroneously amplified, causing a serious overcompensation phenomenon.

[0092] Furthermore, in a conventional industrial IoT cloud-edge collaborative architecture, if the second terminal merely passively accepts the rejection signal from the first terminal without making adaptive adjustments to the global parameter calculation, the baseline parameters generated in the cloud will always deviate from the actual physical conditions of some production lines, thus leading to inconsistencies between cloud and edge data.

[0093] To this end, the first terminal generates and uploads exemption request information carrying the compensation parameters through the industrial network. After receiving the message, the communication module of the second terminal performs protocol parsing and extracts the specific channel index and the corresponding compensation parameters encapsulated therein. At this time, the compensation parameters are no longer local adjustment amounts of edge nodes, but are transformed into high-confidence sample data representing the extreme physical wear state of a specific factory area.

[0094] When the second terminal enters the iterative optimization phase of the global parameters for the next cycle, the processor extracts the aggregated historical production detection dataset and performs routine optimization calculations. During this process, the second terminal triggers a parameter freeze command.

[0095] This means that when calculating the deviation and updating the parameter matrix, the parameter optimization for the protected channel is forcibly blocked, and its value is fixed to the actual physical compensation value uploaded by the first terminal; while for other unconstrained parameter dimensions, the second terminal continues to use big data from multiple factories to perform normal optimization calculations, and after completing a predetermined round of iterative optimization, it generates the updated benchmark parameters.

[0096] In subsequent cycles, the first terminal receives the updated reference parameters sent by the second terminal. Since the updated reference parameters have been actively adapted to the sensitive channels of the edge production line during the generation process and have incorporated the actual physical compensation values ​​of the local area in their internal structure, when the reference parameters are written by the first terminal, they can not only bring global-level general process optimization rules to the edge devices, but also will not destroy the physical correction correspondence that has been established locally.

[0097] In a continuous production batch of a certain type of 3D printing aluminum powder, the first terminal detected that the light transmittance decay rate exceeded the first threshold, indicating that the window was being rapidly contaminated by aluminum powder. At this time, the system triggered the protection mechanism, intercepted the regular reference parameters sent by the second terminal, and encapsulated the compensation parameters used locally to compensate for the signal strength of the high-frequency scattering channel into the exemption request information and reported it to the second terminal. After the second terminal parses the data, it fixes the channel parameters during the parameter iteration and optimization process, and then sends out the updated baseline parameters. The first terminal receives the parameters and performs the calculation.

[0098] This embodiment provides an aluminum powder particle size distribution analysis system based on big data. As an edge control hub deployed on the aluminum powder production line, its hardware architecture mainly includes an industrial communication bus, a memory, and a processor electrically coupled to it.

[0099] The memory specifically adopts an industrial-grade storage medium, such as a high-capacity Flash chip or EEPROM, which is used not only to store the computer program code of the above method, but also to divide the physical address or logical block into a first storage area and a second storage area that are securely isolated, which are used to store externally introduced reference parameters and locally generated compensation parameters, respectively.

[0100] The processor is configured as a digital signal processor or an industrial-grade microprocessor chip with high-frequency floating-point operation capability, and is responsible for the unified scheduling of sensor data streams and parameter matrix operations.

[0101] After the system is running, the processor calls the computer program in the memory and uses the data acquisition interface to obtain the transmittance attenuation data of the particle size detection device in real time, and detects the transmittance attenuation rate, which reflects the physical deterioration process of the optical window, in real time.

[0102] In the flow path of control instructions, the processor, as the core decision-making component, deeply links physical degradation characteristics with data protection logic, and independently judges the target protection status of compensation parameters based on the transmittance attenuation rate and cumulative adjustment data.

[0103] When the processor receives the reference parameters periodically sent by the second terminal, it implements hardware-level permission allocation and control over the reference parameters through the memory read / write enable pin or register status bit, thereby quickly completing the defensive action of allowing or restricting overwriting.

[0104] Based on the physical architecture consisting of a processor and memory, and the built-in dynamic access control strategy, the production line online analysis equipment can achieve autonomous data quality control under extreme physical conditions.

[0105] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for analyzing the particle size distribution of aluminum powder based on big data, applied to a first terminal, characterized in that, The method includes the following steps: Acquire transmittance decay data from particle size detection equipment, and calculate transmittance decay rate based on the transmittance decay data to determine the target contamination stage; Based on the cumulative adjustment data of the transmittance attenuation rate and the compensation parameters, the target protection state of the compensation parameters is determined; The reference parameters sent by the second terminal are received. When the target pollution stage is the first pollution stage and the target protection state is the first protection state, the reference parameters are used to overwrite the compensation parameters. When the target pollution stage is the second pollution stage and the target protection state is the second protection state, the overwriting of the compensation parameters is restricted, and granular inversion calculation is performed in conjunction with the benchmark parameters.

2. The method for analyzing aluminum powder particle size distribution based on big data according to claim 1, characterized in that, The determination of the target protection state of the compensation parameter based on the cumulative adjustment data of the transmittance attenuation rate and the compensation parameter includes: When the transmittance decay rate is greater than the first threshold, the target coefficient is determined based on the ratio of the first threshold to the transmittance decay rate. The second threshold is reduced using the target coefficient to obtain the third threshold; When the cumulative adjustment data is greater than the third threshold, the target protection state of the compensation parameter is determined as the second protection state.

3. The method for analyzing aluminum powder particle size distribution based on big data according to claim 1, characterized in that, The process of acquiring transmittance attenuation data from particle size analysis equipment and calculating the transmittance attenuation rate based on the transmittance attenuation data to determine the target contamination stage includes: The continuous operating time of the particle size detection equipment since the most recent cleaning is obtained; The transmittance attenuation data is fitted using a sliding window to calculate the transmittance attenuation rate. The transmittance attenuation data and the continuous running time are matched with a preset pollution feature matching library to determine the target pollution stage.

4. The method for analyzing aluminum powder particle size distribution based on big data according to claim 1, characterized in that, The limitation on overriding the compensation parameter includes: When the target contamination stage is the second contamination stage and the target protection state has not reached the second protection state, the reference parameter is written to the first storage area, and the compensation parameter in the second storage area is locked to intercept write operations to the second storage area. When the target protection state reaches the second protection state, the write operation of the reference parameters sent by the second terminal for the first storage area and the second storage area is intercepted, and an exemption request information carrying the compensation parameters is generated and uploaded to the second terminal.

5. The method for analyzing aluminum powder particle size distribution based on big data according to claim 4, characterized in that, The step of performing granular inversion calculations in conjunction with the benchmark parameters includes: The reference parameters are read from the first storage area, and the compensation parameters are read from the second storage area. The fusion weight ratio is determined based on the target pollution stage and the cumulative adjustment data; The baseline parameter and the compensation parameter are linearly superimposed according to the fusion weight ratio. The superposition result is input into the granularity inversion calculation model, and a granularity distribution report is output. The granularity distribution report includes frequency distribution curve, cumulative distribution curve and characteristic particle size parameter.

6. The method for analyzing aluminum powder particle size distribution based on big data according to claim 4, characterized in that, After the exemption request information carrying the compensation parameters is generated and uploaded to the second terminal, the method further includes: Count the number of times the second terminal continuously ignores the exemption request information; When the number of ignored attempts exceeds the fourth threshold, an exception reporting operation is triggered. The compensation parameter is forcibly configured to a locked protection state, and all subsequent writes of the reference parameter are blocked while the compensation parameter is in the locked protection state.

7. The method for analyzing aluminum powder particle size distribution based on big data according to claim 1, characterized in that, The method further includes: Perform a validity check on the transmittance sensor used to collect the transmittance attenuation data; When the duration for which the sensor data output by the transmittance sensor has not been updated exceeds the fifth threshold or the magnitude of the change exceeds the sixth threshold, the determination of the target contamination stage is paused. Freeze the migration operation of the target protection state of the compensation parameters and enforce the step of restricting the overwriting of the compensation parameters.

8. The method for analyzing aluminum powder particle size distribution based on big data according to claim 1, characterized in that, The method further includes: When a cleaning or maintenance signal or equipment maintenance signal is detected for the particle size detection device, the target contamination stage is forcibly reset to the first contamination stage; The compensation parameter is cleared to zero, and the cumulative adjustment data is reset to the initial zero value.

9. The method for analyzing aluminum powder particle size distribution based on big data according to claim 4, characterized in that, After the exemption request information carrying the compensation parameters is generated and uploaded to the second terminal, the method further includes: Receive the updated baseline parameters sent by the second terminal; The updated baseline parameters are generated by the second terminal after parsing the exemption request information and extracting the compensation parameters, and configuring the compensation parameters as fixed parameters to constrain its model iterative optimization process.

10. A big data-based aluminum powder particle size distribution analysis system, characterized in that, include: Memory, used to store computer programs; A processor for implementing the method as described in any one of claims 1 to 9 when executing the computer program.

Images