Fireproof aluminum-based composite board processing technology analysis method and system based on big data

Abstract

The invention relates to the technical field of processing technology analysis, and discloses a fireproof aluminum-based composite board processing technology analysis method and system based on big data, and the system comprises a data collection module, a data preprocessing module, a feature extraction module, an abnormality judgment module, and a root cause tracing module. By collecting multi-dimensional operation data such as temperature and pressure and performing cleaning and normalization processing, the data reliability is improved; direct and indirect features are extracted, correlation between parameters is deeply mined, such as the relation between the coating thickness and the temperature and pressure and the influence of the cooling speed on the coating quality, so that the abnormity judgment is more comprehensive and accurate; the process abnormity can be quickly identified through comprehensive index calculation and threshold comparison, and batch defective products are prevented from being generated; comprehensive monitoring and quality control of the machining process are achieved, process abnormity can be found in time, specific links can be positioned, production stability and product quality are improved, data support is provided for process optimization, and production efficiency and reliability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of processing technology analysis, and in particular to a method and system for analyzing the processing technology of fireproof aluminum-based composite panels based on big data. Background Art

[0002] Fireproof aluminum-based composite panels, a composite material that combines fire resistance with structural strength, are widely used in construction, transportation, and other fields. Their processing involves multiple steps, including hot pressing, cooling, and coating. The stability of these process parameters directly impacts product quality. However, traditional processing analysis relies on manual experience and single-point testing, which has the following shortcomings:

[0003] One-sided data collection: Relying only on discrete data from key points cannot cover parameter changes throughout the entire process, and it is difficult to capture the impact of multi-link parameter coupling on product quality.

[0004] Abnormality detection lag: Manual inspections or offline testing cannot monitor process fluctuations in real time, which can easily lead to the mass production of defective products, increasing quality costs and production losses.

[0005] Ambiguous root cause location: When process anomalies occur, the lack of systematic data comparison and correlation analysis methods makes it difficult to quickly locate the specific link (such as hot pressing temperature fluctuations, cooling speed deviations, or uneven coating thickness), resulting in a long troubleshooting cycle and low efficiency.

[0006] Blindness in process optimization: Relying on trial-and-error methods to adjust process parameters, lacking historical data support and quantitative analysis, making it difficult to explore potential correlations between parameters (such as the nonlinear relationship between coating thickness and hot pressing temperature and pressure), limiting the improvement of production stability and quality consistency.

[0007] With the development of big data technology, process optimization through multi-dimensional data collection, feature modeling, and intelligent analysis has become possible. Existing technologies have introduced data monitoring systems in some industrial scenarios. However, given the specificity of fire-resistant aluminum composite panel processing, there is still a lack of full-process analysis solutions that integrate direct parameters and indirect correlation features. This is particularly true in areas such as the dynamic correlation between coating quality and process parameters, and the precise tracing of the root causes of anomalies. Summary of the Invention

[0008] The purpose of the present invention is to provide a fireproof aluminum-based composite panel processing technology analysis method and system based on big data, which solves the technical problems raised in the background technology.

[0009] The purpose of the present invention can be achieved through the following technical solutions:

[0010] The big data-based analysis method for the processing technology of fireproof aluminum-based composite panels includes the following steps:

[0011] Step 1: Data Collection:

[0012] Collect operating data corresponding to temperature, pressure, cooling rate, and coating thickness during the processing of fireproof aluminum-based composite panels;

[0013] Step 2: Data preprocessing:

[0014] Perform data cleaning and normalization on the collected operating data to remove noise and unify the data format;

[0015] Step 3: Data feature extraction:

[0016] Direct features and indirect features are extracted from the running data after data preprocessing;

[0017] Step 4: Data anomaly judgment:

[0018] Perform data anomaly judgment on characteristic data to determine whether the process is abnormal;

[0019] Step 5: Tracing the root cause of the abnormality:

[0020] When a process abnormality is detected, the root cause of the abnormality is traced to locate the specific process link of the abnormality.

[0021] As a further solution of the present invention: the data cleaning method is as follows:

[0022] The temperature and pressure values in the hot pressing stage are respectively denoted as T i and P i ; The cooling rate in the cooling stage is recorded as V i ; The coating thickness in the coating stage is recorded as D i ; i represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the operating data, and i = 1, 2, ... n, n represents the number of batch numbers;

[0023] Select the operating data sequence corresponding to the temperature value;

[0024] Obtaining a temperature fluctuation threshold TB preset according to the fireproof aluminum-based composite panel during the hot pressing stage;

[0025] Extract the temperature values T of n-1 adjacent time points i and T i-1 , and then compare the absolute value of the difference with the temperature fluctuation threshold:

[0026] When |T i -T i-1 When |>TB, the temperature value T is determined i is the noise data, and then respectively through T i The average temperature value at two adjacent time points is corrected, that is, by: The corrected T i ` to replace the temperature value T i ;

[0027] According to the data cleaning method of the temperature data series, data cleaning is performed on other operating data series.

[0028] As a further solution of the present invention: the normalization processing method is as follows:

[0029] Select the temperature value;

[0030] In the operating data sequence corresponding to the temperature value, extract the maximum and minimum temperature values T max and T min ;

[0031] pass: Calculate the normalized temperature value T1 i .

[0032] As a further solution of the present invention: direct feature extraction is to extract the temperature value T of the hot pressing stage after data cleaning and normalization. i , pressure value P i , cooling rate V in the cooling stage i , coating thickness D during coating stage i As a direct feature.

[0033] As a further solution of the present invention: the indirect feature extraction method is as follows:

[0034] Step A2.1: Indirect feature extraction based on the relationship between coating thickness, temperature and pressure:

[0035] Extract historical operation data, which covers the temperature T of different batches of products in the hot pressing stage during the production process. j , pressure P j And the final coating thickness D j The actual measured value of

[0036] Wherein, j=1, 2, ..., m, j represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the historical operation data, and m represents the number of batch numbers corresponding to the historical operation data;

[0037] With the goal of minimizing the sum of squares of the errors between the actual coating thickness and the theoretical coating thickness, the objective function is constructed:

[0038] In the formula, β1, β2, β3 are the coefficients to be solved, and T j 、P j 、D j is the normalized value;

[0039] Then, the partial derivatives of the objective function S with respect to β1, β2, and β3 are calculated and set equal to 0. The system of equations is obtained and solved, and then the specific values of β1, β2, and β3 are obtained.

[0040] System of equations:

[0041] When solving the system of equations, the Gaussian elimination method is used for calculation. The method is as follows: first, the system of equations is transformed into an augmented matrix form, and then it is transformed into an upper triangular matrix through elementary row transformation of the matrix, and then back-substitution is performed to solve it, so as to obtain the specific values of the coefficients β1, β2, and β3;

[0042] Then the temperature value T in the hot pressing stage of the operation data is i , pressure value P i Substitute:

[0043] D0 i =β1×T i +β2×P i +β3, calculate the theoretical coating thickness D0 i ;

[0044] Where, T i 、P i is the normalized value;

[0045] Then the theoretical coating thickness D0 i Subtract the coating thickness D during the coating phase from the operating data i , get the coating thickness deviation value DC i and recorded it as an indirect characteristic of the relationship between coating thickness and temperature and pressure;

[0046] Step A2.2: Indirect feature extraction based on the relationship between cooling rate and coating quality:

[0047] Extract historical operation data, which also covers the cooling speed V of different batches of products during the cooling stage of the production process j ; Subsequently, multiple fireproof aluminum-based composite panels produced at different cooling rates were extracted as samples, and multiple coating quality tests were conducted. Based on the test results, the cooling rate that resulted in the best coating quality was comprehensively evaluated and determined as the optimal cooling rate V0;

[0048] Among them, V j is the normalized value;

[0049] Among them, the coating quality inspection is for the quality inspectors to observe the internal structure of the coating through a microscope and conduct fire resistance tests to determine the coating quality;

[0050] Then extract the cooling rate V in the cooling stage from the operating dataj :

[0051] And through: Calculate the cooling rate deviation L i and recorded it as an indirect characteristic of the relationship between cooling rate and coating quality;

[0052] Where V i is the normalized value;

[0053] When V i >V0, L i Indicates the ratio of the actual cooling rate to the optimal cooling rate;

[0054] When V i ≤V0, L i Indicates the ratio by which the actual cooling rate is slower than the optimal cooling rate.

[0055] As a further solution of the present invention: β1, β2, and β3 are solved as follows:

[0056] Step A2.1.1, construct the augmented matrix form:

[0057] Write the system of equations in augmented matrix form:

[0058] Step A2.1.2, elementary row transformation into an upper triangular matrix:

[0059] By swapping rows, doubling, and multiplying, the matrix is converted to upper triangular form:

[0060] Eliminate the off-diagonal elements of the first column: Use the first row to eliminate the ∑T in the second and third rows j P j and ∑T j ;

[0061] Eliminate the off-diagonal elements of the second column: use the second row to eliminate the ∑P in the third row j ;

[0062] Among them, in the third row, only the coefficient of c is non-zero;

[0063] Finally, we get the upper triangular matrix:

[0064]

[0065] Step A2.1.3, back-substitution solution:

[0066] Starting from the last row, back-substitute step by step to find β1, β2, and β3:

[0067] Solve β3 directly from row 3;

[0068] Substitute β3 into the second row and solve for β2;

[0069] Substitute β2 and β3 into the first row and solve for β1.

[0070] As a further solution of the present invention: the data anomaly judgment method is as follows:

[0071] pass: Calculate the characteristic comprehensive index G i ;

[0072] Where γ1, γ2, γ3, γ4, and γ5 are weight coefficients pre-set according to the degree of influence of each feature on process abnormality;

[0073] Then the characteristic comprehensive index G i Compare with the pre-set normal range of characteristic comprehensive indicators:

[0074] When the characteristic comprehensive index G i If the characteristic comprehensive index is not within the normal range, it is judged that the current process is abnormal.

[0075] As a further solution of the present invention, the root cause of the abnormality is traced by comparing the characteristic data of each process link with the historical normal data, specifically in the following manner:

[0076] Step T1, hot pressing stage root cause determination:

[0077] When an abnormality occurs in the current process, the temperature characteristic sequence T corresponding to the hot pressing stage will be i and pressure characteristic sequence P i Temperature characteristic sequence T corresponding to historical normal production j and pressure characteristic sequence P j For comparison:

[0078] Then through: Calculate the temperature difference DT and pressure difference DP;

[0079] Where r = 1, 2, ..., s, s is the number of comparisons between the characteristic data and the historical normal data;

[0080] Then DT and DP are compared with the preset temperature difference threshold DTy and pressure difference threshold DPy respectively:

[0081] If DT>DTy or DP>DPy, it is judged that there is a problem in the hot pressing stage;

[0082] Step T2, Root cause determination during cooling phase:

[0083] When an abnormality occurs in the current process, the cooling rate characteristic sequence V corresponding to the cooling stage will be iThe cooling rate characteristic sequence V corresponding to historical normal production j For comparison:

[0084] Then through: Calculate the cooling rate difference DV;

[0085] Then DV is compared with the preset cooling rate difference threshold DVy:

[0086] If DV>DVy, it is determined that there is a problem in the cooling stage;

[0087] Step T3, Root cause determination during coating stage:

[0088] When an abnormality occurs in the current process, the corresponding coating thickness deviation value DC is extracted i and cooling rate deviation L i At the same time, obtain the coating thickness deviation value DC corresponding to the historical normal data j and cooling rate deviation L j , then DC i and L i Respectively with DC j and L j For comparison:

[0089] Then through: Calculate the coating thickness deviation difference DDC and cooling rate deviation difference DL;

[0090] Then DDC and DL are compared with the coating thickness deviation value difference threshold DDCy and the cooling rate deviation difference threshold DLy respectively:

[0091] If DDC>DDCy or DDC>DDCy, it is judged that there is a problem in the coating stage.

[0092] As a further solution of the present invention: the data anomaly judgment method is also as follows:

[0093] For direct features:

[0094] Extract the normal range normalization coefficient interval pre-set according to the process standards of temperature, pressure, cooling rate and coating thickness;

[0095] Among them, the normal range normalization coefficient interval refers to the normal range corresponding to the temperature value, pressure value, cooling rate, and coating thickness, which is represented by the value in [0, 1];

[0096] When the temperature T i , pressure value P i , cooling rate V i Or coating thickness D iIf it is not within the corresponding normal range normalization coefficient interval, then the corresponding direct feature is judged to be abnormal;

[0097] For indirect features:

[0098] In the historical operation data, the coating thickness deviation value and cooling rate deviation of the defect-free fireproof aluminum-based composite panel are statistically analyzed, and their distribution range is obtained. Then, the coating thickness deviation threshold and cooling rate deviation threshold are set accordingly according to their distribution range;

[0099] When the coating thickness deviation DC i When the coating thickness deviation threshold is exceeded, or the cooling speed deviation L i When the cooling rate exceeds the deviation threshold, it is judged that the corresponding indirect feature is abnormal.

[0100] A fireproof aluminum-based composite panel processing technology analysis system based on big data is implemented by a fireproof aluminum-based composite panel processing technology analysis method based on big data, and is characterized in that the system includes:

[0101] Data acquisition module: composed of various sensors, responsible for collecting operating data corresponding to temperature, pressure, cooling rate, and coating thickness during the processing of fire-resistant aluminum-based composite panels, and transmitting the operating data to the data preprocessing module;

[0102] Data preprocessing module: used to clean and normalize the collected operation data, remove noise and unify the data format, and then transmit the processed operation data to the feature extraction module;

[0103] Feature extraction module: used to extract direct features and indirect features from the processed operating data, and transmit the extracted feature data to the abnormality judgment module;

[0104] Abnormal judgment module: used to judge the abnormality of feature data. If abnormality is found, the abnormality information is transmitted to the root cause tracing module.

[0105] Root cause tracing module: After receiving abnormal information, the characteristic data of each current process link is compared and analyzed with historical normal data to locate the specific process link where the problem occurs and output the root cause analysis results.

[0106] Beneficial effects of the present invention:

[0107] Data-driven precision monitoring: Through multi-dimensional data collection and preprocessing, data quality is improved, providing a reliable basis for process analysis and achieving refined monitoring of the processing process.

[0108] Multi-feature fusion analysis: Combining direct and indirect features, such as the relationship between coating thickness and temperature and pressure, and the relationship between cooling rate and coating quality, comprehensively explores the potential relationship between process parameters and improves the accuracy of anomaly judgment.

[0109] Efficient anomaly detection: Through comprehensive indicator calculation and feature threshold comparison, process anomalies can be quickly identified to avoid mass production of defective products and reduce quality risks.

[0110] Accurate root cause location: Based on historical data comparison and multi-link difference analysis, the specific process links of the abnormality can be quickly located, such as hot pressing, cooling, and coating stages, facilitating targeted parameter adjustments and shortening troubleshooting time.

[0111] Process optimization support: Through historical data modeling and pattern mining, it provides data support for process parameter optimization, helping to improve production stability, product quality consistency and production efficiency.

[0112] System integration management: Relying on a modular system architecture, the entire process of data collection, processing, analysis, and traceability is automated, improving the intelligence level of production management and decision-making efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0113] The present invention will be further described below with reference to the accompanying drawings.

[0114] Figure 1 This is a system block diagram of the big data-based fireproof aluminum-based composite panel processing technology analysis system of the present invention.

[0115] Figure 2 It is a flow chart of the big data-based fireproof aluminum-based composite panel processing technology analysis method of the present invention. DETAILED DESCRIPTION

[0116] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0117] As embodiment 1 of the present invention:

[0118] See also Figure 1 and Figure 2 As shown, the present invention is a method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data, comprising the following steps:

[0119] Step 1: Data Collection:

[0120] During the processing of fireproof aluminum-based composite panels, the sensor module collects the operating data of multiple key process links, including the temperature value T in the hot pressing stage. i , pressure value P i ; Cooling rate V in the cooling stage i ; Coating thickness D during coating stage i ;

[0121] Wherein, i represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the operation data, and i=1, 2, ... n, and n represents the number of batch numbers;

[0122] In this embodiment, the pressure value P i Represents the pressure applied by the hot pressing equipment on the aluminum-based composite plate; the temperature value T i Reflects the temperature during hot pressing; cooling rate indicates the degree of temperature reduction per unit time;

[0123] Step 2: Data preprocessing:

[0124] Normalize the collected running data series;

[0125] The normalization process is as follows:

[0126] Take temperature value as an example;

[0127] In the operating data sequence corresponding to the temperature value, extract the maximum and minimum temperature values T max and T min ;

[0128] pass: Calculate the normalized temperature value T1 i ;

[0129] Step 3: Data feature extraction:

[0130] Step A1: Direct feature extraction:

[0131] The temperature value T of the hot pressing stage after data cleaning and normalization is i , pressure value P i , cooling rate V in the cooling stage i , coating thickness D during coating stage i As a direct feature;

[0132] Step A2: Indirect feature extraction:

[0133] Step A2.1: Indirect feature extraction based on the relationship between coating thickness, temperature and pressure:

[0134] Extract historical operation data, which covers the temperature T of different batches of products in the hot pressing stage during the production process.j , pressure P j And the final coating thickness D j The actual measured value of

[0135] Wherein, j=1, 2, ..., m, j represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the historical operation data, and m represents the number of batch numbers corresponding to the historical operation data;

[0136] With the goal of minimizing the sum of squares of the errors between the actual coating thickness and the theoretical coating thickness, the objective function is constructed:

[0137] In the formula, β1, β2, β3 are the coefficients to be solved, and T j 、P j 、D j is the normalized value;

[0138] Then, the partial derivatives of the objective function S with respect to β1, β2, and β3 are calculated and set equal to 0. The system of equations is obtained and solved, and then the specific values of β1, β2, and β3 are obtained.

[0139] System of equations:

[0140] When solving the system of equations, the Gaussian elimination method is used for calculation. The method is as follows: first, the system of equations is transformed into an augmented matrix form, and then it is transformed into an upper triangular matrix through elementary row transformation of the matrix, and then back-substitution is performed to solve it, so as to obtain the specific values of the coefficients β1, β2, and β3;

[0141] The specific method is as follows:

[0142] Step A2.1.1, construct the augmented matrix form:

[0143] Write the system of equations in augmented matrix form:

[0144] Step A2.1.2, elementary row transformation into an upper triangular matrix:

[0145] By swapping rows, doubling, and multiplying, the matrix is converted to upper triangular form:

[0146] Eliminate the off-diagonal elements of the first column: Use the first row to eliminate the ∑T in the second and third rows j P j and ∑T j ;

[0147] Eliminate the off-diagonal elements of the second column: use the second row to eliminate the ∑P in the third row j ;

[0148] Among them, in the third row, only the coefficient of c is non-zero;

[0149] Finally, we get the upper triangular matrix:

[0150]

[0151] Step A2.1.3, back-substitution solution:

[0152] Starting from the last row, back-substitute step by step to find β1, β2, and β3:

[0153] Solve β3 directly from row 3;

[0154] Substitute β3 into the second row and solve for β2;

[0155] Substitute β2 and β3 into the first row and solve for β1;

[0156] Then the temperature value T in the hot pressing stage of the operation data is i , pressure value P i Substitute:

[0157] D0 i =β1×T i +β2×P i +β3, calculate the theoretical coating thickness D0 i ;

[0158] Where, T i 、P i is the normalized value;

[0159] Then the theoretical coating thickness D0 i Subtract the coating thickness D during the coating phase from the operating data i , get the coating thickness deviation value DC i and recorded it as an indirect characteristic of the relationship between coating thickness and temperature and pressure;

[0160] Step A2.2: Indirect feature extraction based on the relationship between cooling rate and coating quality:

[0161] Extract historical operation data, which also covers the cooling speed V of different batches of products during the cooling stage of the production process j ; Subsequently, multiple fireproof aluminum-based composite panels produced at different cooling rates were extracted as samples, and multiple coating quality tests were conducted. Based on the test results, the cooling rate that resulted in the best coating quality was comprehensively evaluated and determined as the optimal cooling rate V0;

[0162] Among them, V j is the normalized value;

[0163] Among them, the coating quality inspection is for the quality inspectors to observe the internal structure of the coating through a microscope and conduct fire resistance tests to determine the coating quality;

[0164] Then extract the cooling rate V in the cooling stage from the operating data j :

[0165] And through: Calculate the cooling rate deviation L i and recorded it as an indirect characteristic of the relationship between cooling rate and coating quality;

[0166] Where V i is the normalized value;

[0167] When V i >V0, L i Indicates the ratio of the actual cooling rate to the optimal cooling rate;

[0168] When V i ≤V0, L i Indicates the ratio of the actual cooling rate to the optimal cooling rate;

[0169] Step 4: Data anomaly judgment:

[0170] For direct features:

[0171] Extract the normal range normalization coefficient interval pre-set according to the process standards of temperature, pressure, cooling rate and coating thickness;

[0172] Among them, the normal range normalization coefficient interval refers to the normal range corresponding to the temperature value, pressure value, cooling rate, and coating thickness, which is represented by the value in [0, 1];

[0173] When the temperature T i , pressure value P i , cooling rate V i Or coating thickness D i If it is not within the corresponding normal range normalization coefficient interval, then the corresponding direct feature is judged to be abnormal;

[0174] For indirect features:

[0175] In the historical operation data, the coating thickness deviation value and cooling rate deviation of the defect-free fireproof aluminum-based composite panel are statistically analyzed, and their distribution range is obtained. Then, the coating thickness deviation threshold and cooling rate deviation threshold are set accordingly according to their distribution range;

[0176] When the coating thickness deviation DC i When the coating thickness deviation threshold is exceeded, or the cooling speed deviation L i When the cooling rate exceeds the deviation threshold, it is judged that the corresponding indirect feature is abnormal.

[0177] In the first embodiment, key process data such as hot pressing temperature, pressure, cooling rate, and coating thickness are collected through the sensor module, and normalized to ensure the uniformity and comparability of the data. In the feature extraction link, parameters such as temperature and pressure that directly reflect the process status are retained, and the coefficients are solved by constructing the objective function and the Gaussian elimination method, and a mathematical relationship between the coating thickness and the temperature and pressure is established to calculate the coating thickness deviation value; at the same time, the optimal cooling rate is determined in combination with historical data, and the cooling rate deviation is quantified. This combination of direct features and indirect features can comprehensively characterize the process characteristics. In the abnormality judgment, a normalization coefficient interval is set for the direct features, and a threshold is set for the indirect features based on historical defect-free data, realizing dual abnormality detection of single process parameters and process correlation effects, providing multi-level and multi-angle data support for quality monitoring of the fireproof aluminum-based composite panel processing process, helping to timely discover parameter anomalies in a single link and potential correlation problems between processes, and improving process stability and product quality traceability.

[0178] As the second embodiment of the present invention:

[0179] See also Figure 1 and Figure 2 As shown, in the specific implementation of this application, compared with Example 1, the technical solution of this embodiment is different from that of Example 1 only in that in this embodiment, the data anomaly judgment step is further performed by:

[0180] Calculate the characteristic comprehensive index G i ;

[0181] Where γ1, γ2, γ3, γ4, and γ5 are weight coefficients pre-set according to the degree of influence of each feature on process abnormality;

[0182] Then the characteristic comprehensive index G i Compare with the pre-set normal range of characteristic comprehensive indicators:

[0183] When the characteristic comprehensive index G i If the characteristic comprehensive index is not within the normal range, it is judged that the current process is abnormal;

[0184] When a process abnormality is detected, the root cause of the abnormality is traced to locate the specific process link of the abnormality;

[0185] The root cause of abnormalities is traced by comparing the characteristic data of each process link with historical normal data. The specific method is as follows:

[0186] Step T1, hot pressing stage root cause determination:

[0187] When an abnormality occurs in the current process, the temperature characteristic sequence T corresponding to the hot pressing stage will be iand pressure characteristic sequence P i Temperature characteristic sequence T corresponding to historical normal production j and pressure characteristic sequence P j For comparison:

[0188] Then through: Calculate the temperature difference DT and pressure difference DP;

[0189] Where r = 1, 2, ..., s, s is the number of comparisons between the characteristic data and the historical normal data;

[0190] Then DT and DP are compared with the preset temperature difference threshold DTy and pressure difference threshold DPy respectively:

[0191] If DT>DTy or DP>DPy, it is judged that there is a problem in the hot pressing stage;

[0192] Step T2, Root cause determination during cooling phase:

[0193] When an abnormality occurs in the current process, the cooling rate characteristic sequence V corresponding to the cooling stage will be i The cooling rate characteristic sequence V corresponding to historical normal production j For comparison:

[0194] Then through: Calculate the cooling rate difference DV;

[0195] Then DV is compared with the preset cooling rate difference threshold DVy:

[0196] If DV>DVy, it is determined that there is a problem in the cooling stage;

[0197] Step T3, Root cause determination during coating stage:

[0198] When an abnormality occurs in the current process, the corresponding coating thickness deviation value DC is extracted i and cooling rate deviation L i At the same time, obtain the coating thickness deviation value DC corresponding to the historical normal data j and cooling rate deviation L j , then DC i and L i Respectively with DC j and L j For comparison:

[0199] Then through: Calculate the coating thickness deviation difference DDC and cooling rate deviation difference DL;

[0200] Then DDC and DL are compared with the coating thickness deviation value difference threshold DDCy and the cooling rate deviation difference threshold DLy respectively:

[0201] If DDC>DDCy or DDC>DDCy, it is judged that there is a problem in the coating stage.

[0202] Example 2, based on Example 1, introduces a feature comprehensive index Gi, which integrates multi-dimensional features such as temperature, pressure, cooling rate, and coating thickness through preset weight coefficients to form a comprehensive process status evaluation index, avoiding the limitations of single feature abnormality judgment and being able to more comprehensively reflect whether the overall process is abnormal. When judging process abnormalities, by calculating parameters such as temperature difference, pressure difference, cooling rate difference, and coating thickness deviation difference, the current process data is compared with historical normal data, and the specific differences in the hot pressing, cooling, and coating stages are traced layer by layer to accurately locate the root cause of the abnormality. This method from comprehensive evaluation to staged root cause location can not only quickly identify process abnormalities, but also clarify the problem link, provide specific direction for process adjustment and troubleshooting, improve the efficiency and pertinence of fault diagnosis in the production process, reduce troubleshooting time and cost, and further enhance the scientific nature and effectiveness of process management.

[0203] As the third embodiment of the present invention:

[0204] See also Figure 1 and Figure 2 As shown, in the specific implementation of this application, compared with Example 1 and Example 2, the technical solution of this embodiment is to combine the solutions of Example 1 and Example 2, and the difference between the technical solution of this embodiment and Example 1 and Example 2 is only in this embodiment;

[0205] The data preprocessing step also performs data cleaning on the collected running data sequence. The data cleaning method is as follows:

[0206] Take the operating data sequence corresponding to the selected temperature value as an example;

[0207] Obtaining a temperature fluctuation threshold TB preset according to the fireproof aluminum-based composite panel during the hot pressing stage;

[0208] Extract the temperature values T of n-1 adjacent time points i and T i-1 , and then compare the absolute value of the difference with the temperature fluctuation threshold:

[0209] When |T i -T i-1 When |>TB, the temperature value T is determined i is the noise data, and then respectively through T iThe average temperature value at two adjacent time points is corrected, that is, by: The corrected T i ` to replace the temperature value T i ;

[0210] According to the data cleaning method of the temperature data series, data cleaning is performed on other operating data series.

[0211] Example 3 adds a data cleaning step based on Example 1. Taking temperature data as an example, by setting the temperature fluctuation threshold TB, noise data with temperature value differences between adjacent time points that exceed a reasonable range is identified, and the average value of the adjacent data before and after is used for correction. This data cleaning method can effectively filter out abnormal data generated by factors such as sensor errors and environmental interference during the processing process, and improve the accuracy and reliability of the original data. The same cleaning logic is used for other operating data such as pressure, cooling rate, and coating thickness to ensure the integrity and consistency of the entire data set, provide a purer data source for subsequent data preprocessing, feature extraction, and anomaly judgment, avoid feature calculation deviations and anomaly misjudgments caused by noise data, thereby improving the stability and credibility of the entire process analysis method, and making data-based process evaluation and decision-making more reliable.

[0212] As the fourth embodiment of the present invention:

[0213] See also Figure 1 and Figure 2 As shown, when the present application is implemented, compared with Example 1, Example 2 and Example 3, the technical solution of this embodiment is to combine the solutions of the above-mentioned Example 1, Example 2 and Example 3 for implementation.

[0214] Example 4 combines the technical solutions of Examples 1, 2, and 3 to form a complete process analysis system covering data collection, cleaning, preprocessing, multi-dimensional feature extraction, comprehensive anomaly judgment, and stage-by-stage root cause tracing. Data cleaning is used to eliminate noise data to ensure data quality; direct and indirect features are used to comprehensively characterize the process status; the overall anomaly assessment is performed with the help of feature comprehensive indicators, and the abnormal link is accurately located through stage-by-stage difference comparison. This combined implementation method not only ensures the accuracy of the basic data, but also realizes multi-level analysis and all-round monitoring of the process. It can detect process anomalies in a timely and accurate manner and trace the root cause. It provides a systematic solution for the optimization of the processing technology of fire-resistant aluminum-based composite panels, helps to improve the intelligence level, quality control capabilities and production efficiency of the production process, reduces the product defect rate and production costs caused by process anomalies, and enhances the competitiveness of enterprises in the field of composite material processing.

[0215] See also Figure 1 and Figure 2As shown, a fireproof aluminum-based composite panel processing technology analysis system based on big data is implemented by a fireproof aluminum-based composite panel processing technology analysis method based on big data, and is characterized in that the system includes:

[0216] Data acquisition module: composed of various sensors, responsible for collecting operating data corresponding to temperature, pressure, cooling rate, and coating thickness during the processing of fire-resistant aluminum-based composite panels, and transmitting the operating data to the data preprocessing module;

[0217] Data preprocessing module: used to clean and normalize the collected operation data, remove noise and unify the data format, and then transmit the processed operation data to the feature extraction module;

[0218] Feature extraction module: used to extract direct features and indirect features from the processed operating data, and transmit the extracted feature data to the abnormality judgment module;

[0219] Abnormal judgment module: used to judge the abnormality of feature data. If abnormality is found, the abnormality information is transmitted to the root cause tracing module.

[0220] Root cause tracing module: After receiving abnormal information, the characteristic data of each current process link is compared and analyzed with historical normal data to locate the specific process link where the problem occurs and output the root cause analysis results.

[0221] It should be stated that all user data collected in this application is collected with the user's consent and authorization, and the use of user data is legal and compliant, and the use and processing of user data complies with the relevant laws, regulations and standards of the relevant regions.

[0222] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0223] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A fireproof aluminum-based composite panel processing technology analysis method based on big data, characterized in that: The following steps are involved: Data collection: Collect the corresponding operating data of temperature, pressure, cooling rate and coating thickness during the processing of fireproof aluminum-based composite panels; Data preprocessing: Clean and normalize the collected operating data to remove noise and unify the data format; Data feature extraction: extract direct features and indirect features from the running data after data preprocessing; Data anomaly judgment: Perform data anomaly judgment on characteristic data to determine whether the process is abnormal; Abnormal root cause tracing: When it is determined that a process abnormality occurs, the abnormal root cause tracing is carried out to locate the specific process link of the abnormality.

2. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 1, characterized in that: The temperature and pressure values in the hot pressing stage are respectively denoted as T i and P i ; The cooling rate in the cooling stage is recorded as V i ; The coating thickness in the coating stage is recorded as D i ; i represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the operating data, and i = 1, 2, ... n, n represents the number of batch numbers; Direct feature extraction is to extract the temperature value T of the hot pressing stage after data cleaning and normalization. i , pressure value P i , cooling rate V in the cooling stage i , coating thickness D during coating stage i As a direct feature.

3. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 2, characterized in that: Extract indirect features based on the relationship between coating thickness and temperature and pressure as follows: Extract historical operation data, which covers the temperature T of different batches of products in the hot pressing stage during the production process. j , pressure P j And the final coating thickness D j The actual measured value of Wherein, j=1, 2, ..., m, j represents the batch number of the fireproof aluminum-based composite panel processing corresponding to the historical operation data, and m represents the number of batch numbers corresponding to the historical operation data; With the goal of minimizing the sum of squares of the errors between the actual coating thickness and the theoretical coating thickness, the objective function is constructed: In the formula, β1, β2, β3 are the coefficients to be solved, and T j 、P j 、D j is the normalized value; Then, the partial derivatives of the objective function S with respect to β1, β2, and β3 are calculated and set equal to 0 to obtain a system of equations and solve them, and then the specific values of β1, β2, and β3 are obtained. When solving the system of equations, the Gaussian elimination method is used for calculation, which is as follows: first, the system of equations is transformed into an augmented matrix form, and then it is transformed into an upper triangular matrix through elementary row transformation of the matrix, and then back-substitution is performed to solve it, so as to obtain the specific values of the coefficients β1, β2, and β3. Then the temperature value T in the hot pressing stage of the operation data is i , pressure value P i Substitute: D0 i =β1×T i +β2×P i +β3, calculate the theoretical coating thickness D0 i ; Where, T i 、P i is the normalized value; Then the theoretical coating thickness D0 i Subtract the coating thickness D during the coating phase from the operating data i , get the coating thickness deviation value DC i and recorded it as an indirect characteristic of the relationship between coating thickness and temperature and pressure.

4. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 3 is characterized in that: Extract indirect features based on the relationship between cooling rate and coating quality as follows: Extract historical operation data, which also covers the cooling speed V of different batches of products during the cooling stage of the production process j ; Then, a number of fireproof aluminum-based composite panels produced at different cooling rates were extracted as samples, and multiple coating quality tests were conducted. Based on the test results, the cooling rate with the best coating quality was comprehensively evaluated and determined as the optimal cooling rate V0; where V j is the normalized value; Then extract the cooling rate V in the cooling stage from the operating data j : And through: Calculate the cooling rate deviation L i , and record it as an indirect characteristic of the relationship between cooling rate and coating quality; where V i is the normalized value.

5. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 4, characterized in that: The method for judging data anomalies is as follows: pass: Calculate the characteristic comprehensive index G i ; Where γ1, γ2, γ3, γ4, and γ5 are weight coefficients pre-set according to the degree of influence of each feature on process abnormality; Then the characteristic comprehensive index G i Compare with the pre-set normal range of characteristic comprehensive indicators: When the characteristic comprehensive index G i If the characteristic comprehensive index is not within the normal range, it is judged that the current process is abnormal.

6. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 5, characterized in that: The root cause of abnormalities is traced by comparing the characteristic data of each process link with historical normal data. The specific method is as follows: Step T1, hot pressing stage root cause determination: When an abnormality occurs in the current process, the temperature characteristic sequence T corresponding to the hot pressing stage will be i and pressure characteristic sequence P i Temperature characteristic sequence T corresponding to historical normal production j and pressure characteristic sequence P j For comparison: Then through: Calculate the temperature difference DT and pressure difference DP; Where r = 1, 2, ..., s, s is the number of comparisons between the characteristic data and the historical normal data; Then DT and DP are compared with the preset temperature difference threshold DTy and pressure difference threshold DPy respectively: If DT>DTy or DP>DPy, it is judged that there is a problem in the hot pressing stage; Step T2, Root cause determination during cooling phase: When an abnormality occurs in the current process, the cooling rate characteristic sequence V corresponding to the cooling stage will be i The cooling rate characteristic sequence V corresponding to historical normal production j For comparison: Then through: Calculate the cooling rate difference DV; Then DV is compared with the preset cooling rate difference threshold DVy: If DV>DVy, it is determined that there is a problem in the cooling stage; Step T3, Root cause determination during coating stage: When an abnormality occurs in the current process, the corresponding coating thickness deviation value DC is extracted i and cooling rate deviation L i At the same time, obtain the coating thickness deviation value DC corresponding to the historical normal data j and cooling rate deviation L j , then DC i and L i Respectively with DC j and L j For comparison: Then through: Calculate the coating thickness deviation difference DDC and cooling rate deviation difference DL; Then DDC and DL are compared with the coating thickness deviation value difference threshold DDCy and the cooling rate deviation difference threshold DLy respectively: If DDC>DDCy or DDC>DDCy, it is judged that there is a problem in the coating stage.

7. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 4, characterized in that: The method for judging data anomalies is as follows: For direct features: Extracting a normal range normalization coefficient interval pre-set according to the process standards corresponding to the temperature value, pressure value, cooling rate, and coating thickness; wherein the normal range normalization coefficient interval refers to the normal range corresponding to the temperature value, pressure value, cooling rate, and coating thickness, represented by a value in [0, 1]; When the temperature T i , pressure value P i , cooling rate V i Or coating thickness D i If it is not within the corresponding normal range normalization coefficient interval, then the corresponding direct feature is judged to be abnormal; For indirect features: In the historical operation data, the coating thickness deviation value and cooling rate deviation of the defect-free fireproof aluminum-based composite panel are statistically analyzed, and their distribution range is obtained. Then, the coating thickness deviation threshold and cooling rate deviation threshold are set accordingly according to their distribution range; When the coating thickness deviation DC i When the coating thickness deviation threshold is exceeded, or the cooling speed deviation L i When the cooling rate exceeds the deviation threshold, it is judged that the corresponding indirect feature is abnormal.

8. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 1, characterized in that: The data cleaning method is as follows: Select the operating data sequence corresponding to the temperature value; Obtaining a temperature fluctuation threshold TB preset according to the fireproof aluminum-based composite panel during the hot pressing stage; Extract the temperature values T of n-1 adjacent time points i and T i-1 , and then compare the absolute value of the difference with the temperature fluctuation threshold: When |T i -T i-1 When |>TB, the temperature value T is determined i is the noise data, and then respectively through T i The average temperature value at two adjacent time points is corrected, that is, by: The corrected T i ` to replace the temperature value T i ; According to the data cleaning method of the temperature data series, data cleaning is performed on other operating data series.

9. The method for analyzing the processing technology of fireproof aluminum-based composite panels based on big data according to claim 1, characterized in that: The normalization process is as follows: Select the temperature value; In the operating data sequence corresponding to the temperature value, extract the maximum and minimum temperature values T max and T min ; pass: Calculate the normalized temperature value T1 i .

10. A big data-based fireproof aluminum-based composite panel processing technology analysis system, which is implemented by the big data-based fireproof aluminum-based composite panel processing technology analysis method according to any one of claims 1 to 9, characterized in that: The system includes: Data acquisition module: composed of various sensors, responsible for collecting operating data corresponding to temperature, pressure, cooling rate, and coating thickness during the processing of fire-resistant aluminum-based composite panels, and transmitting the operating data to the data preprocessing module; Data preprocessing module: used to clean and normalize the collected operation data, remove noise and unify the data format, and then transmit the processed operation data to the feature extraction module; Feature extraction module: used to extract direct features and indirect features from the processed operating data, and transmit the extracted feature data to the abnormality judgment module; Abnormal judgment module: used to judge the abnormality of feature data. If abnormality is found, the abnormality information is transmitted to the root cause tracing module. Root cause tracing module: After receiving abnormal information, the characteristic data of each current process link is compared and analyzed with historical normal data to locate the specific process link where the problem occurs and output the root cause analysis results.

Images