Aluminum hydroxide production quality detection method and system based on data analysis

Abstract

The present invention discloses a method and system for detecting the production quality of aluminum hydroxide based on data analysis, which mainly relates to the field of quality detection technology. Among them, the method includes obtaining preset standard quality parameters and preset quality tolerance intervals; obtaining M storage grids; obtaining M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets; based on the preset standard quality parameters, centralized identification of quality deviations is performed to obtain M storage grid production quality deviation values, and the mean value is calculated in combination with the obtained Q incremental storage grid production quality deviation values ​​to determine the target production quality detection result. The method solves the technical problem in the prior art that when aluminum hydroxide is subjected to batch quality inspection, the inspection quality has a low fit with the actual situation and the inspection results are inaccurate, thereby achieving the technical effect of improving the reliability of the inspection.

Description

Technical Field

[0001] The present invention relates to the technical field of quality detection, and in particular to a method and system for detecting the production quality of aluminum hydroxide based on data analysis. Background Art

[0002] Aluminum hydroxide, a key chemical raw material, is widely used in bauxite processing and the production of various chemical products. Ensuring product quality is crucial in the production of aluminum hydroxide. However, traditional production quality testing methods often overlook the overall picture when analyzing multiple quality parameters and cannot guarantee the rationality of the sampling source distribution of test results. This can lead to poor alignment between test quality and actual conditions, resulting in inaccurate test results. Summary of the Invention

[0003] The present invention provides a method and system for detecting the production quality of aluminum hydroxide based on data analysis, which is used to solve the technical problem in the prior art that, when performing quality detection of aluminum hydroxide in batches, the detection quality has a low fit with the actual situation and the detection results are inaccurate.

[0004] In view of the above problems, the present invention provides a method and system for detecting the production quality of aluminum hydroxide based on data analysis.

[0005] The first aspect of the present invention provides a method for detecting the production quality of aluminum hydroxide based on data analysis, the method comprising: obtaining preset standard quality parameters and preset quality tolerance intervals; dividing the target batch of aluminum hydroxide into storage grids in the order of production time according to a preset division scale to obtain M storage grids, wherein M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packaging in the storage grid has a timestamp identifier; according to a preset sample extraction ratio, respectively extracting samples from the M storage grids to obtain M storage grid sample sets, and using an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer and an infrared moisture meter to traverse the M storage grid sample sets for production quality detection to obtain a chemical purity set of M storage grid samples. a set of M storage grid sample particle size distributions, a set of M storage grid sample crystal structures, and a set of M storage grid sample water contents; based on the preset standard quality parameters, the M storage grid sample chemical purity sets, the M storage grid sample particle size distribution sets, the M storage grid sample crystal structure sets, and the M storage grid sample water contents sets are centrally identified for quality deviations to obtain M storage grid production quality deviation values; according to the M storage grid production quality deviation values ​​and the time interval identifiers, the M storage grids are incrementally extracted and tested for time-series associated samples to obtain Q incremental storage grid production quality deviation values, the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values ​​are averaged to determine the target production quality test result.

[0006] In a possible embodiment, the preset standard quality parameters include a standard value for chemical purity, a standard value for particle size distribution, a standard value for crystal structure, and a standard value for water content.

[0007] In a possible embodiment, based on the preset standard quality parameters, the quality deviations of the M storage grid sample chemical purity sets, the M storage grid sample particle size distribution sets, the M storage grid sample crystal structure sets and the M storage grid sample water content sets are centrally identified to obtain M storage grid production quality deviation values, including: according to the chemical purity standard value, the particle size distribution standard value, the crystal structure standard value and the water content standard value, the quality deviations of the M storage grid sample chemical purity sets, the M storage grid sample particle size distribution sets, the M storage grid sample crystal structure sets and the M storage grid sample water content sets are identified to obtain M storage grid sample chemical purity deviation value sets, M storage grid sample particle size distribution deviation value sets, M storage grid sample crystal structure deviation value sets and M sets of moisture content deviation values ​​of storage grid samples; performing linear fitting centralized identification based on the chemical purity deviation value sets of the M storage grid samples, the particle size distribution deviation value sets of the M storage grid samples, the crystal structure deviation value sets of the M storage grid samples and the moisture content deviation value sets of the M storage grid samples to determine the concentrated chemical purity deviation values ​​of the M storage grid samples, the concentrated particle size distribution deviation values ​​of the M storage grid samples, the concentrated crystal structure deviation values ​​of the M storage grid samples and the moisture content deviation values ​​of the M storage grid samples; performing weighted calculation on the chemical purity concentrated deviation values ​​of the M storage grid samples, the particle size distribution concentrated deviation values ​​of the M storage grid samples, the crystal structure concentrated deviation values ​​of the M storage grid samples and the moisture content concentrated deviation values ​​of the M storage grid samples, respectively, to obtain M storage grid production quality deviation values.

[0008] In one possible implementation, linear fitting centralized identification is performed based on the chemical purity deviation value sets of the M storage grid samples to determine the concentrated chemical purity deviation values ​​of the M storage grid samples, including: in combination with a timestamp identifier, two-dimensionally mapping the chemical purity deviation value sets of the M storage grid samples in a two-dimensional mapping coordinate system to determine M two-dimensional mapping coordinate point sets; performing interior point discrimination straight line iteration on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines; pre-constructing a diffusion bandwidth function, using the diffusion bandwidth function to generate diffusion bandwidths multiple times to perform fitting neighborhood diffusion on the M initial linear fitting reference lines to determine M target linear fitting neighborhoods; traversing and calculating the average of the absolute values ​​of the horizontal coordinates of the two-dimensional mapping coordinate points in the M target linear fitting neighborhoods to determine the concentrated chemical purity deviation values ​​of the M storage grid samples.

[0009] In one possible implementation, the M two-dimensional mapping coordinate point sets are iterated to determine the inlier discriminant line, including: randomly extracting two points from the M two-dimensional mapping coordinate point sets to construct M first discriminant lines; according to a preset distance threshold, the two-dimensional mapping coordinate points whose distances to the M first discriminant lines are within the preset distance threshold are taken as inliers and the number of inliers is counted to obtain the M first discriminant line inlier quantities; randomly extracting two points from the M two-dimensional mapping coordinate point sets multiple times to construct discriminant lines until the preset number of constructions is met, to obtain M discriminant line sets and M discriminant line inlier quantity sets; and taking the discriminant lines corresponding to the maximum values ​​in the M discriminant line inlier quantity sets as the M initial linear fitting reference lines.

[0010] In a possible implementation, the diffusion bandwidth function is: ;in, is the diffusion bandwidth, , , , is the Gamma function, .

[0011] In one possible implementation, a diffusion bandwidth function is pre-constructed, and diffusion bandwidths are generated multiple times using the diffusion bandwidth function to perform fitting neighborhood diffusion on the M initial linear fitting reference lines to determine M target linear fitting neighborhoods. The method includes: extracting a first initial linear fitting reference line from the M initial linear fitting reference lines; randomly generating a first diffusion bandwidth based on the diffusion bandwidth function, performing neighborhood construction on the first initial linear fitting reference line according to the first diffusion bandwidth, and determining the first initial fitting neighborhood; randomly generating a second diffusion bandwidth again based on the diffusion bandwidth function, and performing upward and downward diffusion on the edge of the first initial fitting neighborhood according to the second diffusion bandwidth to obtain a first diffusion fitting neighborhood; determining whether the data volume of the first diffusion fitting neighborhood is greater than or equal to the data volume of the first initial fitting neighborhood; if so, randomly generating a third diffusion bandwidth again based on the diffusion bandwidth function, and performing upward and downward diffusion on the edge of the first diffusion fitting neighborhood according to the third diffusion bandwidth until a preset number of diffusions is met, and using the diffusion fitting neighborhood obtained from the last diffusion as the first target linear fitting neighborhood; and traversing the M initial linear fitting reference lines and performing fitting neighborhood diffusion in combination with the diffusion bandwidth function to obtain the M target linear fitting neighborhoods.

[0012] In one possible implementation, if the amount of data in the first diffusion fitting neighborhood is smaller than the amount of data in the first initial fitting neighborhood, the difference between the amount of data in the first diffusion fitting neighborhood and the amount of data in the first initial fitting neighborhood is calculated; if the difference is less than or equal to a preset difference, the first initial fitting neighborhood is used as the first target linear fitting neighborhood.

[0013] In one possible implementation, based on the production quality deviation values ​​of M storage grids and the time interval identifiers, the M storage grids are subjected to time-series correlation sample incremental extraction and detection to obtain Q incremental storage grid production quality deviation values, including: respectively judging whether the production quality deviation values ​​of the M storage grids exceed the preset quality tolerance interval, and if so, identifying the corresponding storage grids as abnormal storage grids to obtain an abnormal storage grid set; according to the preset pre- and post-correlated time intervals and the abnormal storage grid set, in combination with the time interval identifier corresponding to each storage grid, determining the associated storage grid set, wherein the associated storage grids are storage grids whose time interval intervals with the corresponding abnormal storage grids in the abnormal storage grid set are within the range of the preset pre- and post-correlated time intervals; according to the preset sample increments, the associated storage grid set and the abnormal storage grid set are subjected to time-series correlation sample incremental extraction and detection, and according to the detection results, Q incremental storage grid production quality deviation values ​​are obtained, wherein Q is a positive integer less than or equal to M.

[0014] The second aspect of the present invention provides an aluminum hydroxide production quality detection system based on data analysis, which is used to implement the aforementioned aluminum hydroxide production quality detection method based on data analysis, and the system includes: a tolerance interval acquisition module, which is used to obtain preset standard quality parameters and preset quality tolerance intervals; a storage grid acquisition module, which is used to divide the target batch of aluminum hydroxide into storage grids in the order of production time according to a preset division scale, and obtain M storage grids, wherein M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packaging in the storage grid has a timestamp identifier; a production quality detection module, which is used to extract samples from the M storage grids according to a preset sample extraction ratio, obtain M storage grid sample sets, and use an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer and an infrared moisture meter to traverse the M storage grid sample sets for production quality Detection, obtain M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets; a production quality deviation value acquisition module, used to centrally identify the quality deviations of the M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets based on the preset standard quality parameters, and obtain M storage grid production quality deviation values; a target production quality detection result determination module, used to perform time-series associated sample incremental extraction and detection on the M storage grids according to the M storage grid production quality deviation values ​​and the time interval identifier, obtain Q incremental storage grid production quality deviation values, perform mean calculation on the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values, and determine the target production quality detection result.

[0015] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0016] The present invention obtains preset standard quality parameters and preset quality tolerance intervals, divides the target batch of aluminum hydroxide into storage grids according to the production time sequence according to the preset division scale, and obtains M storage grids, wherein M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packaging in the storage grid has a timestamp identifier, and then according to the preset sample extraction ratio, samples are extracted from the M storage grids respectively to obtain M storage grid sample sets, and an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer and an infrared moisture meter are used to traverse the M storage grid sample sets to perform production quality inspection, and obtain M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets, then, based on preset standard quality parameters, the M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets are centrally identified for quality deviations to obtain M storage grid production quality deviation values, and then, based on the M storage grid production quality deviation values ​​and time interval identifiers, the M storage grids are incrementally extracted and tested for time-series correlation samples to obtain Q incremental storage grid production quality deviation values, and the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values ​​are averaged to determine the target production quality test results. Through the above method, the technical effect of improving the reliability of the aluminum hydroxide production quality test results is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Attachment Figure 1 This is a flow chart of a method for detecting the quality of aluminum hydroxide production based on data analysis provided by the present invention;

[0018] Attachment Figure 2 This is a structural diagram of an aluminum hydroxide production quality detection system based on data analysis provided by the present invention;

[0019] Reference numerals shown in the accompanying drawings:

[0020] 11 is a tolerance interval acquisition module, 12 is a storage grid acquisition module, 13 is a production quality inspection module, 14 is a production quality deviation value acquisition module, and 15 is a target production quality inspection result determination module. DETAILED DESCRIPTION

[0021] The present invention will be further described below in conjunction with specific examples. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the claims attached to the present invention. It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] Example 1, as shown in the attached Figure 1 As shown, the present invention provides a method for detecting the production quality of aluminum hydroxide based on data analysis, the method comprising:

[0023] S1: Obtaining preset standard quality parameters and preset quality tolerance range;

[0024] Furthermore, the preset standard quality parameters include a standard value for chemical purity, a standard value for particle size distribution, a standard value for crystal structure, and a standard value for water content.

[0025] In one possible embodiment, the preset standard quality parameters are standard data predefined by those skilled in the art for evaluating and comparing the quality of aluminum hydroxide products, including at least a standard value for the chemical purity, particle size distribution, crystal structure, and water content of aluminum hydroxide. The preset quality tolerance range is a quality deviation range set by those skilled in the art to ensure product qualification rate and production efficiency, taking into account quality fluctuations of aluminum hydroxide.

[0026] S2: Divide the target batch of aluminum hydroxide into storage grids according to the production time sequence according to a preset division scale to obtain M storage grids, where M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packages in the storage grid have a time stamp identifier;

[0027] In one embodiment of the present application, a person skilled in the art determines the maximum amount of aluminum hydroxide that can be stored in each storage grid in the storage warehouse to obtain the preset division scale. Then, in chronological order, the aluminum hydroxide of the target batch is placed in the storage grid in sequence according to the production time to obtain the M storage grids. In order to facilitate subsequent tracing, each aluminum hydroxide package has a timestamp mark, and the M storage grids are marked with time intervals based on the time interval between the timestamp mark on the first aluminum hydroxide package placed in each storage grid and the timestamp mark on the last aluminum hydroxide package placed. By time-marking the storage grids, the technical effect of providing a basis for distinguishing for subsequent effective extraction of test samples is achieved.

[0028] S3: Samples are extracted from the M storage grids according to a preset sample extraction ratio to obtain M storage grid sample sets, and production quality inspection is performed on the M storage grid sample sets using an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer, and an infrared moisture meter to obtain a chemical purity set of the M storage grid samples, a particle size distribution set of the M storage grid samples, a crystal structure set of the M storage grid samples, and a moisture content set of the M storage grid samples;

[0029] In one possible embodiment, the preset sample extraction ratio refers to the ratio of samples extracted from each storage grid, such as 1% to 5%. Samples are extracted from each of the M storage grids according to the preset sample extraction ratio, with each sample representing the quality of the aluminum hydroxide within its corresponding grid. After extracting the M storage grid sample sets, each storage grid sample set is divided into four categories and tested for chemical purity, particle size distribution, crystal structure, and water content, respectively.

[0030] Exemplarily, the storage grid sample is dissolved, and the melted sample solution is introduced into a burner through a sprayer and mixed with a gas (such as air or acetylene) to form an atomized solution. The metal elements in the solution are evaporated and atomized by heating with a burner. Then, the vaporized sample is irradiated with a light source of a specific wavelength of an atomic absorption spectrometer. Preferably, the wavelength of the light source is consistent with the characteristic absorption line of the element to be measured. When aluminum is present in the sample, a change in absorbance is detected. By comparing the measured absorbance with the standard curve, the purity of various elements in aluminum hydroxide is calculated to obtain the chemical purity set of the M storage grid samples.

[0031] Preferably, the storage grid sample is dispersed in a solvent, and a laser particle size distribution analyzer is used to emit a laser beam to irradiate the sample to obtain the angle and intensity of light scattering, and the Mie theory is used for inversion to determine the particle size distribution set of M storage grid samples. Optionally, the storage grid sample is pressed into a thin sheet and placed on the sample stage of an X-ray diffractometer. The diffraction intensity at different angles is detected by rotating the sample and the detector. According to different diffraction angles, the diffraction pattern of the crystal in the sample can be obtained. According to the diffraction pattern, the interplanar spacing d of the crystal is analyzed using Bragg's law (nλ=2dsinθ). By comparing the standard diffraction data, the crystal structure of the aluminum hydroxide sample is determined, and the crystal structure set of the M storage grid samples is obtained.

[0032] Preferably, the storage grid samples are placed on the sample tray of an infrared moisture meter, which is then activated to heat the samples using infrared radiation. A high-precision sensor monitors changes in sample mass. As water evaporates, the sample mass gradually decreases. The sample mass change is determined, and the water content is calculated, which is then output as a set of moisture contents for the M storage grid samples.

[0033] Through multi-dimensional quality testing, the key quality indicators of aluminum hydroxide samples are comprehensively evaluated, providing comprehensive data support for subsequent quality deviation analysis, thereby effectively reflecting the technical effect of the quality status of aluminum hydroxide in each storage grid.

[0034] S4: Based on the preset standard quality parameters, performing centralized quality deviation identification on the chemical purity set of the M storage grid samples, the particle size distribution set of the M storage grid samples, the crystal structure set of the M storage grid samples, and the water content set of the M storage grid samples to obtain the production quality deviation values ​​of the M storage grids;

[0035] Furthermore, in step S4 of the embodiment of the present invention, based on the preset standard quality parameters, the quality deviation of the M storage grid sample chemical purity set, the M storage grid sample particle size distribution set, the M storage grid sample crystal structure set, and the M storage grid sample water content set is centrally identified to obtain the M storage grid production quality deviation values, specifically including:

[0036] According to the chemical purity standard value, the particle size distribution standard value, the crystal structure standard value and the water content standard value, quality deviation identification is performed on the chemical purity set of M storage grid samples, the particle size distribution set of M storage grid samples, the crystal structure set of M storage grid samples and the water content set of M storage grid samples, to obtain the chemical purity deviation value set of M storage grid samples, the particle size distribution deviation value set of M storage grid samples, the crystal structure deviation value set of M storage grid samples and the water content deviation value set of M storage grid samples;

[0037] Performing linear fitting centralized identification based on the M storage grid sample chemical purity deviation value sets, the M storage grid sample particle size distribution deviation value sets, the M storage grid sample crystal structure deviation value sets, and the M storage grid sample water content deviation value sets to determine the M storage grid sample chemical purity centralized deviation values, the M storage grid sample particle size distribution centralized deviation values, the M storage grid sample crystal structure centralized deviation values, and the M storage grid sample water content centralized deviation values;

[0038] The chemical purity centralized deviation values ​​of the M storage grid samples, the particle size distribution centralized deviation values ​​of the M storage grid samples, the crystal structure centralized deviation values ​​of the M storage grid samples and the water content centralized deviation values ​​of the M storage grid samples are weightedly calculated respectively to obtain the M storage grid production quality deviation values.

[0039] In one possible embodiment, the chemical purity standard value is calculated by difference with the chemical purity sets of the M storage grid samples to determine the chemical purity deviation value sets of the M storage grid samples. The chemical purity deviation value of the storage grid sample reflects the purity of the aluminum hydroxide in the corresponding sample. Based on the same principle, the particle size distribution standard value, crystal structure standard value and water content standard value are calculated by difference with the particle size distribution sets of the M storage grid samples, the crystal structure sets of the M storage grid samples and the water content sets of the M storage grid samples to determine the corresponding particle size distribution deviation value sets of the M storage grid samples, the crystal structure deviation value sets of the M storage grid samples and the water content deviation value sets of the M storage grid samples.

[0040] Preferably, the linear fitting centralized identification process is explained by taking the chemical purity deviation value set of the M storage grid samples as an example. By performing linear fitting centralized identification on the particle size distribution deviation value set of the M storage grid samples, the crystal structure deviation value set of the M storage grid samples and the water content deviation value set of the M storage grid samples, the principle of obtaining the concentrated deviation values ​​of the particle size distribution of the M storage grid samples, the concentrated deviation values ​​of the crystal structure of the M storage grid samples and the concentrated deviation values ​​of the water content of the M storage grid samples is the same as that of obtaining the concentrated deviation values ​​of the chemical purity of the M storage grid samples.

[0041] Then, according to a weight ratio pre-set by a person skilled in the art, a corresponding weighted calculation is performed on the chemical purity concentrated deviation values ​​of the M storage grid samples, the particle size distribution concentrated deviation values ​​of the M storage grid samples, the crystal structure concentrated deviation values ​​of the M storage grid samples, and the water content concentrated deviation values ​​of the M storage grid samples, respectively, to determine the production quality deviation values ​​of the M storage grids. The production quality deviation values ​​of the M storage grids reflect the general production quality of aluminum hydroxide in the M storage grids.

[0042] Furthermore, performing linear fitting centralized identification based on the chemical purity deviation value set of the M storage grid samples to determine the chemical purity centralized deviation value of the M storage grid samples includes:

[0043] In combination with the timestamp identifier, two-dimensionally mapping the M sets of chemical purity deviation values ​​of the storage grid samples in a two-dimensional mapping coordinate system is performed to determine M sets of two-dimensional mapping coordinate points;

[0044] Performing interior point discrimination line iteration on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines;

[0045] Pre-constructing a diffusion bandwidth function, using the diffusion bandwidth function to generate diffusion bandwidths multiple times to perform fitting neighborhood diffusion on the M initial linear fitting reference lines, and determining M target linear fitting neighborhoods;

[0046] The mean of the absolute values ​​of the horizontal coordinates of the two-dimensional mapping coordinate points in the M target linear fitting neighborhoods is traversed and calculated to determine the concentrated deviation values ​​of the chemical purity of the M storage grid samples.

[0047] Furthermore, the interior point discrimination line iteration is performed on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines. In the embodiment of the present application, step S4 further includes:

[0048] Randomly extract two points from the M two-dimensional mapping coordinate point sets to construct M first discriminant lines;

[0049] According to a preset distance threshold, the two-dimensional mapping coordinate points whose distances to the M first discriminant lines are within the preset distance threshold are regarded as inliers and the number of inliers is counted to obtain the number of inliers of the M first discriminant lines;

[0050] Randomly extracting two points from the M two-dimensional mapping coordinate point sets for multiple times to construct a discriminant line until a preset number of constructions is met, thereby obtaining M discriminant line sets and M sets of points within the discriminant line;

[0051] The discriminant lines corresponding to the maximum values ​​in the set of point quantities within the M discriminant lines are respectively used as M initial linear fitting reference lines.

[0052] In an embodiment of the present invention, a linear fitting process is performed on a set of chemical purity deviation values ​​of M storage grid samples, thereby obtaining the chemical purity deviation values ​​of the M storage grid samples that best represent the chemical degree deviation of the M storage grids. First, a two-dimensional mapping coordinate system is constructed with time as the horizontal axis and the chemical purity deviation value as the vertical axis. The production time of the sample is determined based on the timestamp identifier corresponding to the storage grid sample. Then, combined with the chemical purity deviation value of the storage grid sample, the M storage grid sample chemical purity deviation value set is mapped to the two-dimensional mapping coordinate system to obtain the M two-dimensional mapping coordinate point set. Optionally, each storage grid sample chemical purity deviation value has a positive or negative identifier. When the chemical purity of the storage grid sample is greater than the chemical purity standard value, the storage grid sample chemical purity deviation value is positive; when the chemical purity of the storage grid sample is less than the chemical purity standard value, the storage grid sample chemical purity deviation value is negative.

[0053] After obtaining the M sets of two-dimensional mapping coordinate points, the data is screened to obtain the data that best represents the chemical purity deviation within each storage grid, i.e., the concentrated chemical purity deviation values ​​for the M storage grid samples. First, a line with a relatively dense distribution of coordinate points within the M sets of two-dimensional mapping coordinate points is identified to obtain the M initial linear fit reference lines. Second, after obtaining the initial linear fit reference lines, the boundaries corresponding to the areas with relatively dense distribution are determined, i.e., the M target linear fit neighborhoods. Finally, the mean of the absolute values ​​of the horizontal coordinates of the two-dimensional mapping coordinate points within the M target linear fit neighborhoods is calculated to determine the concentrated chemical purity deviation values ​​for the M storage grid samples.

[0054] Preferably, when determining the boundary, in order to avoid falling into a local optimum, a diffusion bandwidth function is used to randomly generate the amplitude of diffusion of the boundary of the neighborhood, that is, the diffusion bandwidth, so as to achieve the goal of improving the screening efficiency while avoiding missing more representative two-dimensional mapping coordinate points.

[0055] In one possible embodiment, two points are randomly selected from the set of M two-dimensional mapping coordinate points, and a straight line passing through the two points is used as a first discriminant line, thereby obtaining M first discriminant lines. A preset distance threshold is obtained, using a distance preset by a person skilled in the art that allows a two-dimensional mapping coordinate point to be used as an inlier of a discriminant line as a constraint. Then, two-dimensional mapping coordinate points whose distances to the M first discriminant lines are within the preset distance threshold are used as inliers, and the number of inliers is counted to obtain the number of inliers of the M first discriminant lines. The number of inliers of the M first discriminant lines reflects the reliability of the M first discriminant lines as reference lines for linear fitting. A larger number of inliers indicates that there are more two-dimensional mapping coordinate points distributed around the discriminant line, and in this case, the discriminant line is more reliable.

[0056] Based on the principle of constructing the first discriminant line described above, two points are randomly extracted from the M sets of two-dimensional mapping coordinate points multiple times to construct the discriminant line until a maximum number of constructions, i.e., a preset number of constructions, predetermined by a person skilled in the art, is reached. This results in M ​​sets of discriminant lines and M sets of points within the discriminant line. The discriminant line corresponding to the maximum value among the M sets of points within the discriminant line is then used as the M initial linear fitting reference lines.

[0057] Furthermore, the diffusion bandwidth function is:

[0058] ;

[0059] in, is the diffusion bandwidth, , , , is the Gamma function, .

[0060] Furthermore, a diffusion bandwidth function is pre-constructed, and the diffusion bandwidth is generated multiple times using the diffusion bandwidth function to perform fitting neighborhood diffusion on the M initial linear fitting reference lines to determine M target linear fitting neighborhoods. In this embodiment of the application, step S4 further includes:

[0061] Extracting a first initial linear fitting reference line from the M initial linear fitting reference lines;

[0062] randomly generating a first diffusion bandwidth based on the diffusion bandwidth function, constructing a neighborhood for a first initial linear fitting reference line according to the first diffusion bandwidth, and determining a first initial fitting neighborhood;

[0063] randomly generating a second diffusion bandwidth based on the diffusion bandwidth function again, and diffusing the edge of the first initial fitting neighborhood upward and downward respectively according to the second diffusion bandwidth to obtain a first diffusion fitting neighborhood;

[0064] Determining whether the amount of data in the first diffusion fitting neighborhood is greater than or equal to the amount of data in the first initial fitting neighborhood; if so, randomly generating a third diffusion bandwidth based on the diffusion bandwidth function again; and diffusing the edge of the first diffusion fitting neighborhood upward and downward according to the third diffusion bandwidth until a preset number of diffusions is met, and using the diffusion fitting neighborhood obtained from the last diffusion as the first target linear fitting neighborhood;

[0065] The M initial linear fitting reference lines are traversed and fitted neighborhood diffusion is performed in combination with the diffusion bandwidth function to obtain the M target linear fitting neighborhoods.

[0066] Furthermore, if the amount of data in the first diffusion fitting neighborhood is smaller than the amount of data in the first initial fitting neighborhood, the difference between the amount of data in the first diffusion fitting neighborhood and the amount of data in the first initial fitting neighborhood is calculated. If the difference is less than or equal to a preset difference, the first initial fitting neighborhood is used as the first target linear fitting neighborhood.

[0067] In a possible embodiment, the diffusion bandwidth generated by the diffusion bandwidth function is a random value, mainly determined by the Gamma function. in To control the range of diffusion, thus adjusting the distribution scale of diffusion bandwidth. , ,in, Obey the normal distribution, The variance is variable and can be set by those skilled in the art according to actual conditions. The variance is 1, and the fluctuation range is fixed, so The value of will fluctuate within a certain range, so uncertainty is introduced in the diffusion analysis process, thereby ensuring the fluctuation of bandwidth, and A random value between (0, 2), The randomness of will directly affect the change of bandwidth, so the randomness of the generated diffusion bandwidth can be guaranteed.

[0068] Exemplarily, a first initial linear fitting reference line is extracted from the M initial linear fitting reference lines, and a random one is selected from (0, 2). Value, such as 0.5, enter 0.5 into the diffusion bandwidth function and generate a random , the first diffusion bandwidth output is 0.3. Preferably, The value range of is mostly between [-0.5, 0.5], and there will be extreme values. When extreme values ​​appear, the regeneration instruction is triggered and random sampling is performed again. The diffusion bandwidth function is then used to generate the diffusion bandwidth.

[0069] After obtaining the first diffusion bandwidth, the first initial linear fitting reference line is shifted upward and downward by the first diffusion bandwidth to construct a first initial fitting neighborhood. A second diffusion bandwidth is randomly generated based on the diffusion bandwidth function, and the edges of the first initial fitting neighborhood are diffused upward and downward according to the second diffusion bandwidth to obtain a first diffusion fitting neighborhood. Furthermore, the data volume of the first diffusion fitting neighborhood and the first initial fitting neighborhood is counted. When the data volume of the first diffusion fitting neighborhood is greater than or equal to the data volume of the first initial fitting neighborhood, a third diffusion bandwidth is randomly generated based on the diffusion bandwidth function again, and the edges of the first diffusion fitting neighborhood are diffused upward and downward according to the third diffusion bandwidth until a predetermined number of diffusions is reached. The diffusion fitting neighborhood obtained from the last diffusion is then used as the first target linear fitting neighborhood. When the data volume of the first diffusion fitting neighborhood is less than that of the first initial fitting neighborhood, the difference between the data volume of the first diffusion fitting neighborhood and the data volume of the first initial fitting neighborhood is calculated. If the difference is less than or equal to the predetermined difference, indicating that the data volume difference between the two neighborhoods is not significant, diffusion is stopped, and the first initial fitting neighborhood is used as the first target linear fitting neighborhood.

[0070] Based on the same principle as obtaining the first target linear fitting neighborhood, the diffusion bandwidth function is used to randomly generate diffusion bandwidths multiple times, and the neighborhood is constructed and diffused on the M initial linear fitting reference lines to obtain the target linear fitting neighborhood.

[0071] S5: Based on the production quality deviation values ​​of the M storage grids and the time interval identifier, perform time-series associated sample incremental extraction and detection on the M storage grids to obtain Q incremental storage grid production quality deviation values, perform mean calculation on the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values, and determine the target production quality detection result.

[0072] Furthermore, based on the production quality deviation values ​​of the M storage grids and the time interval identifier, incremental extraction and detection of time-series associated samples are performed on the M storage grids to obtain Q incremental production quality deviation values ​​of the storage grids. In this embodiment of the application, step S5 further includes:

[0073] Determine whether the production quality deviation values ​​of the M storage grids exceed a preset quality tolerance range respectively; if so, mark the corresponding storage grid as an abnormal storage grid to obtain an abnormal storage grid set;

[0074] According to the preset pre- and post-association time intervals and the abnormal storage grid set, in combination with the time interval identifier corresponding to each storage grid, the associated storage grid set is determined, wherein the associated storage grid is a storage grid whose time interval with the corresponding abnormal storage grid in the abnormal storage grid set is within the range of the preset pre- and post-association time intervals;

[0075] According to the preset sample increment, the associated storage grid set and the abnormal storage grid set are subjected to time-series associated sample increment extraction and detection, and Q incremental storage grid production quality deviation values ​​are obtained according to the detection results, where Q is a positive integer less than or equal to M.

[0076] In one possible embodiment, when any one of the quality deviation values ​​of the M storage grids exceeds the preset quality tolerance interval, it indicates that the quality pass rate of the samples produced by the storage grid is low. At this time, more samples are needed for verification. At the same time, the storage grids produced adjacent to it may also have a low pass rate problem. However, due to the randomness of the sample extraction, the low-quality aluminum hydroxide is not detected. Therefore, it is necessary to combine the time interval identifier to perform incremental detection on the abnormal storage grid and the associated storage grid set associated with it. Then, the results of the fusion incremental detection are used to calculate the mean of the production quality deviation values ​​of the M storage grids to obtain the target production quality deviation value. Among them, the target production quality deviation value reflects the overall production quality of the target batch of aluminum hydroxide.

[0077] First, determine whether the production quality deviation values ​​of the M storage grids exceed the preset quality tolerance interval. If so, it indicates that the corresponding storage grid is abnormal, and the corresponding storage grid is identified as an abnormal storage grid to obtain an abnormal storage grid set. Then, according to the pre-set before and after associated time intervals and the abnormal storage grid set by those skilled in the art, combined with the time interval identifier corresponding to each storage grid, an associated storage grid set is determined. The associated storage grids are storage grids whose time intervals with the corresponding abnormal storage grids in the abnormal storage grid set are within the preset before and after associated time interval range. In other words, storage grids whose time intervals with the time interval identifiers of the abnormal storage grids are within the preset before and after associated time interval range are added to the associated storage grids.

[0078] By performing time-series correlation sample incremental testing on the associated storage grid set and the abnormal storage grid set based on a preset sample increment (i.e., the number of samples to be additionally tested), and following the same principles and procedures as for obtaining the production quality deviation values ​​for M storage grids, Q incremental storage grid production quality deviation values ​​are obtained. This achieves the technical effect of improving the reliability of production quality testing results.

[0079] In summary, the embodiments of the present invention have at least the following technical effects:

[0080] The present invention divides the target batch of aluminum hydroxide into storage grids by combining preset standard quality parameters, quality tolerance ranges and production time sequence, and performs sample extraction and multi-dimensional quality detection in each storage grid. It uses an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer and an infrared moisture meter to comprehensively analyze the chemical purity, particle size distribution, crystal structure and water content of the samples, and performs time-series associated sample increment extraction in combination with the quality deviation value. Finally, the production quality detection result is determined by mean calculation, thereby achieving the technical effect of efficient and reliable production quality detection of aluminum hydroxide.

[0081] Example 2, based on the same inventive concept as the method for detecting the quality of aluminum hydroxide production based on data analysis in the above embodiment, as shown in the attached Figure 2 As shown, the present application provides an aluminum hydroxide production quality detection system based on data analysis. The system and method embodiments in the embodiments of the present invention are based on the same inventive concept. The system includes:

[0082] The tolerance interval acquisition module 11 is used to obtain the preset standard quality parameters and the preset quality tolerance interval;

[0083] A storage grid obtaining module 12 is configured to divide the target batch of aluminum hydroxide into storage grids according to a preset division scale and in the order of production time, to obtain M storage grids, where M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packages in the storage grid have a timestamp identifier;

[0084] The production quality inspection module 13 is configured to extract samples from the M storage grids according to a preset sample extraction ratio to obtain M storage grid sample sets, and to perform production quality inspection on the M storage grid sample sets using an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer, and an infrared moisture meter to obtain a chemical purity set of the M storage grid samples, a particle size distribution set of the M storage grid samples, a crystal structure set of the M storage grid samples, and a moisture content set of the M storage grid samples;

[0085] The production quality deviation value obtaining module 14 is configured to perform a centralized quality deviation identification on the chemical purity set of the M storage grid samples, the particle size distribution set of the M storage grid samples, the crystal structure set of the M storage grid samples, and the water content set of the M storage grid samples based on the preset standard quality parameters, and obtain the production quality deviation values ​​of the M storage grids;

[0086] The target production quality detection result determination module 15 is used to perform time-series associated sample incremental extraction and detection on the M storage grids based on the production quality deviation values ​​of the M storage grids and the time interval identifier, obtain Q incremental storage grid production quality deviation values, calculate the mean of the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values, and determine the target production quality detection result.

[0087] Furthermore, the preset standard quality parameters include a standard value for chemical purity, a standard value for particle size distribution, a standard value for crystal structure, and a standard value for water content.

[0088] Furthermore, the production quality deviation value obtaining module 14 includes:

[0089] Deviation value set obtaining unit: used for performing quality deviation identification on M storage grid sample chemical purity sets, M storage grid sample particle size distribution sets, M storage grid sample crystal structure sets and M storage grid sample water content sets according to chemical purity standard values, particle size distribution standard values, crystal structure standard values ​​and water content standard values, and obtaining M storage grid sample chemical purity deviation value sets, M storage grid sample particle size distribution deviation value sets, M storage grid sample crystal structure deviation value sets and M storage grid sample water content deviation value sets;

[0090] A centralized deviation value determination unit is configured to perform linear fitting centralized identification based on the M storage grid sample chemical purity deviation value sets, the M storage grid sample particle size distribution deviation value sets, the M storage grid sample crystal structure deviation value sets, and the M storage grid sample water content deviation value sets, to determine the M storage grid sample chemical purity centralized deviation value, the M storage grid sample particle size distribution centralized deviation value, the M storage grid sample crystal structure centralized deviation value, and the M storage grid sample water content centralized deviation value;

[0091] Production quality deviation value obtaining unit: used to perform weighted calculations on the chemical purity concentrated deviation values ​​of the M storage grid samples, the particle size distribution concentrated deviation values ​​of the M storage grid samples, the crystal structure concentrated deviation values ​​of the M storage grid samples, and the water content concentrated deviation values ​​of the M storage grid samples, respectively, to obtain the M storage grid production quality deviation values.

[0092] Furthermore, the centralized deviation value determination unit includes a chemical purity centralized deviation value determination unit, a particle size distribution centralized deviation value determination unit, a crystal structure centralized deviation value determination unit and a water content centralized deviation value determination unit.

[0093] Furthermore, the chemical purity centralized deviation value obtaining unit specifically includes:

[0094] A first determining unit is configured to perform two-dimensional mapping on the M sets of chemical purity deviation values ​​of the storage grid samples in a two-dimensional mapping coordinate system in combination with a timestamp identifier, and determine M sets of two-dimensional mapping coordinate points;

[0095] A second determining unit is configured to perform interior point discrimination line iteration on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines;

[0096] A third determining unit is configured to pre-construct a diffusion bandwidth function, use the diffusion bandwidth function to generate diffusion bandwidths multiple times, perform fitting neighborhood diffusion on the M initial linear fitting reference lines, and determine M target linear fitting neighborhoods;

[0097] The fourth determining unit is used to traverse and calculate the mean of the absolute values ​​of the horizontal coordinates of the two-dimensional mapping coordinate points in the M target linear fitting neighborhoods, and determine the concentrated deviation values ​​of the chemical purity of the M storage grid samples.

[0098] Furthermore, the second determining unit specifically includes:

[0099] A first discriminant line construction unit is configured to randomly extract two points from the M two-dimensional mapping coordinate point sets to construct M first discriminant lines;

[0100] A first discrimination line inlier quantity obtaining unit is configured to, according to a preset distance threshold, take the two-dimensional mapping coordinate points whose distances to the M first discrimination lines are within the preset distance threshold as inliers and count the number of inliers to obtain the inlier quantities of the M first discrimination lines;

[0101] A set obtaining unit is configured to randomly extract two points from the M two-dimensional mapping coordinate point sets for multiple times to construct a discriminant line until a preset number of construction times is met, thereby obtaining M discriminant line sets and M sets of points within the discriminant line;

[0102] The initial linear fitting reference line acquisition unit is used to use the discriminant lines corresponding to the maximum values ​​in the set of point quantities within the M discriminant lines as M initial linear fitting reference lines.

[0103] Furthermore, the diffusion bandwidth function is:

[0104] ;

[0105] in, is the diffusion bandwidth, , , , is the Gamma function, which is used to adjust the scale of the distribution. .

[0106] Furthermore, the third determining unit specifically includes:

[0107] A first initial linear fitting reference line extraction unit: configured to extract a first initial linear fitting reference line from the M initial linear fitting reference lines;

[0108] A first initial fitting neighborhood determining unit is configured to randomly generate a first diffusion bandwidth based on the diffusion bandwidth function, construct a neighborhood for a first initial linear fitting reference line according to the first diffusion bandwidth, and determine a first initial fitting neighborhood;

[0109] A first diffusion fitting neighborhood obtaining unit is configured to randomly generate a second diffusion bandwidth based on the diffusion bandwidth function again, and diffuse the edge of the first initial fitting neighborhood upward and downward respectively according to the second diffusion bandwidth to obtain a first diffusion fitting neighborhood;

[0110] A judging unit is used to judge the size relationship between the data volume of the first diffusion fitting neighborhood and the data volume of the first initial fitting neighborhood;

[0111] A first processing unit is configured to, when the amount of data in the first diffusion fitting neighborhood is greater than or equal to the amount of data in the first initial fitting neighborhood, randomly generate a third diffusion bandwidth based on the diffusion bandwidth function again, diffuse the edge of the first diffusion fitting neighborhood upward and downward respectively according to the third diffusion bandwidth until a preset number of diffusions is met, and use the diffusion fitting neighborhood obtained from the last diffusion as the first target linear fitting neighborhood;

[0112] The second processing unit is configured to calculate a difference between the data volume of the first diffusion fitting neighborhood and the data volume of the first initial fitting neighborhood when the data volume of the first diffusion fitting neighborhood is less than the data volume of the first initial fitting neighborhood, and if the difference is less than or equal to a preset difference, use the first initial fitting neighborhood as a first target linear fitting neighborhood;

[0113] The target linear fitting neighborhood obtaining unit is configured to traverse the M initial linear fitting reference lines and perform fitting neighborhood diffusion in combination with the diffusion bandwidth function to obtain the M target linear fitting neighborhoods.

[0114] Furthermore, the target production quality test result determination module 15 includes:

[0115] Abnormal storage grid set obtaining unit: used to determine whether the production quality deviation values ​​of the M storage grids exceed the preset quality tolerance range respectively, and if so, mark the corresponding storage grid as an abnormal storage grid to obtain the abnormal storage grid set;

[0116] An associated storage grid set determining unit is configured to determine an associated storage grid set according to the preset pre- and post-associated time intervals and the abnormal storage grid set, in combination with the time interval identifier corresponding to each storage grid, wherein the associated storage grids are storage grids whose time intervals with the corresponding abnormal storage grids in the abnormal storage grid set are within the range of the preset pre- and post-associated time intervals;

[0117] An incremental storage grid production quality deviation value obtaining unit is configured to perform time-series associated sample incremental extraction and detection on the associated storage grid set and the abnormal storage grid set according to a preset sample increment, and obtain Q incremental storage grid production quality deviation values ​​according to the detection results, where Q is a positive integer less than or equal to M;

[0118] Mean calculation unit: used for performing mean calculation on the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values;

[0119] Target production quality test result determination unit: used to determine the target production quality test result based on the mean calculation result.

[0120] It should be noted that the order in which the embodiments of the present invention are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0121] The above description is only a preferred embodiment of the present invention and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present application.

[0122] This specification and drawings are merely illustrative of the present invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the present invention. Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the scope of the present invention. Thus, the present invention is intended to include such modifications and variations as fall within the scope of the present invention and its equivalents.

Claims

1. A method for detecting the production quality of aluminum hydroxide based on data analysis, characterized in that: The method comprises: Obtaining preset standard quality parameters and preset quality tolerance ranges; Divide the target batch of aluminum hydroxide into storage grids according to a preset division scale in chronological order of production to obtain M storage grids, where M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packages in the storage grid have a timestamp identifier; According to a preset sample extraction ratio, samples are extracted from the M storage grids respectively to obtain M storage grid sample sets, and an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer, and an infrared moisture meter are used to traverse the M storage grid sample sets to perform production quality inspections to obtain the M storage grid sample chemical purity sets, the M storage grid sample particle size distribution sets, the M storage grid sample crystal structure sets, and the M storage grid sample water content sets; Based on the preset standard quality parameters, performing centralized quality deviation identification on the chemical purity set of the M storage grid samples, the particle size distribution set of the M storage grid samples, the crystal structure set of the M storage grid samples, and the water content set of the M storage grid samples to obtain the production quality deviation values ​​of the M storage grids; According to the production quality deviation values ​​of M storage grids and the time interval identifier, the M storage grids are subjected to time-series correlation sample incremental extraction and detection to obtain Q incremental storage grid production quality deviation values, and the mean of the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values ​​is calculated to determine the target production quality detection result.

2. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 1, characterized in that, The preset standard quality parameters include a standard value for chemical purity, a standard value for particle size distribution, a standard value for crystal structure, and a standard value for water content.

3. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 2, characterized in that, Based on the preset standard quality parameters, the quality deviation of the M storage grid sample chemical purity set, the M storage grid sample particle size distribution set, the M storage grid sample crystal structure set, and the M storage grid sample water content set is centrally identified to obtain the M storage grid production quality deviation values, including: According to the chemical purity standard value, the particle size distribution standard value, the crystal structure standard value and the water content standard value, quality deviation identification is performed on the chemical purity set of M storage grid samples, the particle size distribution set of M storage grid samples, the crystal structure set of M storage grid samples and the water content set of M storage grid samples, to obtain the chemical purity deviation value set of M storage grid samples, the particle size distribution deviation value set of M storage grid samples, the crystal structure deviation value set of M storage grid samples and the water content deviation value set of M storage grid samples; Performing linear fitting centralized identification based on the M storage grid sample chemical purity deviation value sets, the M storage grid sample particle size distribution deviation value sets, the M storage grid sample crystal structure deviation value sets, and the M storage grid sample water content deviation value sets to determine the M storage grid sample chemical purity centralized deviation values, the M storage grid sample particle size distribution centralized deviation values, the M storage grid sample crystal structure centralized deviation values, and the M storage grid sample water content centralized deviation values; The chemical purity centralized deviation values ​​of the M storage grid samples, the particle size distribution centralized deviation values ​​of the M storage grid samples, the crystal structure centralized deviation values ​​of the M storage grid samples and the water content centralized deviation values ​​of the M storage grid samples are weightedly calculated respectively to obtain the M storage grid production quality deviation values.

4. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 3, characterized in that, Performing linear fitting centralized identification based on the M storage grid sample chemical purity deviation value sets to determine the M storage grid sample chemical purity centralized deviation values ​​includes: In combination with the timestamp identifier, two-dimensionally mapping the M sets of chemical purity deviation values ​​of the storage grid samples in a two-dimensional mapping coordinate system is performed to determine M sets of two-dimensional mapping coordinate points; Performing interior point discrimination line iteration on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines; Pre-constructing a diffusion bandwidth function, using the diffusion bandwidth function to generate diffusion bandwidths multiple times to perform fitting neighborhood diffusion on the M initial linear fitting reference lines, and determining M target linear fitting neighborhoods; The mean of the absolute values ​​of the horizontal coordinates of the two-dimensional mapping coordinate points in the M target linear fitting neighborhoods is traversed and calculated to determine the concentrated deviation values ​​of the chemical purity of the M storage grid samples.

5. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 4, characterized in that, Performing interior point discrimination line iteration on the M two-dimensional mapping coordinate point sets to determine M initial linear fitting reference lines includes: Randomly extract two points from the M two-dimensional mapping coordinate point sets to construct M first discriminant lines; According to a preset distance threshold, the two-dimensional mapping coordinate points whose distances to the M first discriminant lines are within the preset distance threshold are regarded as inliers and the number of inliers is counted to obtain the number of inliers of the M first discriminant lines; Randomly extracting two points from the M two-dimensional mapping coordinate point sets for multiple times to construct a discriminant line until a preset number of constructions is met, thereby obtaining M discriminant line sets and M sets of points within the discriminant line; The discriminant lines corresponding to the maximum values ​​in the set of point quantities within the M discriminant lines are respectively used as M initial linear fitting reference lines.

6. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 4, characterized in that, The diffusion bandwidth function is: ； in, is the diffusion bandwidth, , , , is the Gamma function, .

7. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 4, characterized in that, Pre-constructing a diffusion bandwidth function, using the diffusion bandwidth function to generate diffusion bandwidths multiple times to perform fitting neighborhood diffusion on the M initial linear fitting reference lines, and determining M target linear fitting neighborhoods, including: Extracting a first initial linear fitting reference line from the M initial linear fitting reference lines; randomly generating a first diffusion bandwidth based on the diffusion bandwidth function, constructing a neighborhood for a first initial linear fitting reference line according to the first diffusion bandwidth, and determining a first initial fitting neighborhood; randomly generating a second diffusion bandwidth based on the diffusion bandwidth function again, and diffusing the edge of the first initial fitting neighborhood upward and downward respectively according to the second diffusion bandwidth to obtain a first diffusion fitting neighborhood; Determining whether the amount of data in the first diffusion fitting neighborhood is greater than or equal to the amount of data in the first initial fitting neighborhood; if so, randomly generating a third diffusion bandwidth based on the diffusion bandwidth function again; and diffusing the edge of the first diffusion fitting neighborhood upward and downward according to the third diffusion bandwidth until a preset number of diffusions is met, and using the diffusion fitting neighborhood obtained from the last diffusion as the first target linear fitting neighborhood; The M initial linear fitting reference lines are traversed and fitted neighborhood diffusion is performed in combination with the diffusion bandwidth function to obtain the M target linear fitting neighborhoods.

8. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 7, characterized in that, If the amount of data in the first diffusion fitting neighborhood is less than the amount of data in the first initial fitting neighborhood, the difference between the amount of data in the first diffusion fitting neighborhood and the amount of data in the first initial fitting neighborhood is calculated. If the difference is less than or equal to a preset difference, the first initial fitting neighborhood is used as the first target linear fitting neighborhood.

9. A method for detecting the quality of aluminum hydroxide production based on data analysis as claimed in claim 1, characterized in that, According to the production quality deviation values ​​of the M storage grids and the time interval identifier, incremental extraction and detection of time-series associated samples are performed on the M storage grids to obtain Q incremental production quality deviation values ​​of the storage grids, including: Determine whether the production quality deviation values ​​of the M storage grids exceed a preset quality tolerance range respectively; if so, mark the corresponding storage grid as an abnormal storage grid to obtain an abnormal storage grid set; According to the preset pre- and post-association time intervals and the abnormal storage grid set, in combination with the time interval identifier corresponding to each storage grid, the associated storage grid set is determined, wherein the associated storage grid is a storage grid whose time interval with the corresponding abnormal storage grid in the abnormal storage grid set is within the range of the preset pre- and post-association time intervals; According to the preset sample increment, the associated storage grid set and the abnormal storage grid set are subjected to time-series associated sample increment extraction and detection, and Q incremental storage grid production quality deviation values ​​are obtained according to the detection results, where Q is a positive integer less than or equal to M.

10. A data analysis-based aluminum hydroxide production quality detection system, characterized in that: The system is used to implement the aluminum hydroxide production quality detection method based on data analysis according to any one of claims 1 to 9, and the system comprises: A tolerance interval acquisition module is used to obtain preset standard quality parameters and preset quality tolerance intervals; A storage grid acquisition module is used to divide the target batch of aluminum hydroxide into storage grids according to the production time sequence according to a preset division scale to obtain M storage grids, where M is a positive integer, each storage grid has a time interval identifier, and the aluminum hydroxide packaging in the storage grid has a time stamp identifier; a production quality inspection module, configured to extract samples from the M storage grids according to a preset sample extraction ratio to obtain M storage grid sample sets, and to traverse the M storage grid sample sets to perform production quality inspection using an atomic absorption spectrometer, a laser particle size distribution analyzer, an X-ray diffractometer, and an infrared moisture meter to obtain a chemical purity set of the M storage grid samples, a particle size distribution set of the M storage grid samples, a crystal structure set of the M storage grid samples, and a moisture content set of the M storage grid samples; a production quality deviation value obtaining module, configured to perform centralized quality deviation identification on the chemical purity set of the M storage grid samples, the particle size distribution set of the M storage grid samples, the crystal structure set of the M storage grid samples, and the water content set of the M storage grid samples based on the preset standard quality parameters, and obtain production quality deviation values ​​of the M storage grids; The target production quality test result determination module is used to perform time-series associated sample incremental extraction and detection on the M storage grids based on the production quality deviation values ​​of the M storage grids and the time interval identifier, obtain Q incremental storage grid production quality deviation values, calculate the mean of the Q incremental storage grid production quality deviation values ​​and the M storage grid production quality deviation values, and determine the target production quality test result.

Images