A data analysis method for analyzing anomalies in tire manufacturing process parameters

Abstract

This application discloses a data analysis method for anomaly analysis of tire manufacturing process parameters, belonging to the field of tire manufacturing technology. This data analysis method is implemented based on a variable analysis module and a variable anomaly period location analysis module. The variable analysis module is used to perform feature measurement and judgment analysis on abnormal signals in the tire manufacturing process; the variable anomaly period location analysis module is used to locate the occurrence period of abnormal process parameters in the tire manufacturing process under abnormal conditions. The variable analysis module includes an anomaly feature measurement module and an anomaly judgment analysis module. The anomaly feature measurement module obtains quantitative features by indexing the anomaly, and the anomaly judgment analysis module determines the threshold of the anomaly index based on the quantitative features and judges whether the variable is abnormal based on the threshold.

Description

Technical Field

[0001] This application relates to the field of tire manufacturing technology, and more specifically, to a data analysis method for analyzing anomalies in tire manufacturing process parameters. Background Technology

[0002] Tire manufacturing parameter analysis and optimization technology is one of the key technologies to ensure a smooth tire manufacturing process and qualified quality. The current development status and existing problems in this field are as follows:

[0003] Current development status:

[0004] 1. Standardization of process flow: The tire manufacturing process has formed a standardized process, including processes such as mixing, rubber component preparation, tire molding, vulcanization, final inspection and tire testing;

[0005] 2. Quantitative analysis of process parameters: Through simulation analysis, the influence of process parameters on tire performance (such as dynamic balance) can be quantified, which helps to optimize process parameters and improve tire quality;

[0006] 3. Intelligent manufacturing: The tire manufacturing industry is introducing intelligent manufacturing technologies, utilizing automated production equipment, industrial internet, modern sensing technologies, and big data analysis technologies to achieve automation, informatization, and intelligentization of the production process;

[0007] 4. Application of new materials and technologies: In the tire manufacturing process, the application of new materials and technologies continuously drives the optimization of process parameters, such as the use of high-performance rubber and steel wire;

[0008] Problems exist:

[0009] 1. Incomplete technology system: The R&D system of domestic tire companies is relatively incomplete, lacking a systematic technology system and technical roadmap, resulting in weak theoretical foundation and insufficient innovation capability;

[0010] 2. Brand value and awareness: The value and awareness of domestic tire brands are relatively low, resulting in weak pricing power and affecting corporate profits and reinvestment capabilities;

[0011] 3. The complexity of process parameter anomaly analysis: Tire manufacturing involves numerous process parameters, and there are complex interrelationships among these parameters. Optimizing these parameters requires in-depth analysis and research.

[0012] In conclusion, although some progress has been made in tire manufacturing parameter analysis and optimization, it still faces many challenges and requires further development. Summary of the Invention

[0013] To overcome a series of shortcomings in the existing technology, the purpose of this application is to provide a data analysis method for analyzing anomalies in tire manufacturing process parameters, comprising the following steps:

[0014] S1. Data collected from the tire manufacturing process is stored in the corresponding variable x. i Below, the number of variables is M, i = 1, 2, ..., M, which constitutes the variable pool X = {x i |i = 1, 2, ..., M, where M is a positive integer};

[0015] S2, Read the target variable x i In [t0,t n Construct a set of all data within a given time period:

[0016] Where t0≤t j ≤t n And t j In [t0,t n The intervals within the range are uniform, 0 ≤ j ≤ n, T 0,n For all t j The time set constituted is defined as n≥1000, where j is a natural number;

[0017] S3. Constructing a two-dimensional point set in, yes The coordinates of the variable, t j yes Time coordinates;

[0018] S4, based on C i,0,n The set sequence is obtained stepwise by using a maximum approximation method according to the number of iterations m.

[0019]

[0020] Where m is the number of layers, and m = 1, 2, ..., M;

[0021]

[0022] S5. F is established for any 1≤m≤M. i,0,n,m (t), calculate its maximum point set according to the variable coordinates. make therefore Not empty, "∶=" means "equals", where The time coordinates of the elements are as follows Represents a set The number of elements;

[0023] S6 Time coordinates of any element In T 0,n The closest time in the inner distance is t j_h According to this correspondence method, The time coordinates t of all elements j_h Converging into a collection

[0024] S7, use Adjacent time periods in T 0，n G on i，0，n Perform segmentation, that is, take any interval [t] j_h , t j_(h+1) ], according to the time sequence of the data from set X i，0，n Read x from i In [t] j_h , t j_(h+1) Construct a set containing all data in ∩T:

[0025]

[0026] S8, the point set G in the plane coordinate system i，j_h，j_(h+1)，m ,according to The set of changing peaks is pe_X i，j_h，j_(h+1)，m The set of trough points is va_X i，j_h，j_(h+1)，m ;

[0027] S9. Calculate pe_X i，j_h，j_(h+1)，m The mean is denoted as aver(pe_X). i,j_h，j_(h+1)，m ), calculate va_X i,j_h，j_(h+1)，m The mean is denoted as aver(va_X). i，j_h，j_(h+1)，m ), calculate X i,j_h，j_(h+1)，m The mean is denoted as aver(X). i，j_h，j_(h+1)，m );

[0028] S10, Calculate pe_X i,j_h，j_(h+1)，m Standard deviation, denoted as st(pe_X) i,j_h，j_(h+1)，m ), calculate va_X i,j_h，j_(h+1)，m The standard deviation is denoted as st(va_X). i,j_h，j_(h+1)，m ), calculate X i,j_h，j_(h+1)，m The standard deviation is denoted as st(X). i,j_h，j_(h+1)，m );

[0029] S11, Constructing Vectors:

[0030] pe_v i，j_h，j_(h+1)，m =(aver(pe_X) i，j_h，j_(h+1)，m ), st(pe_X i,j_h，j_(h+1)，m )), va_v i，j_h，j_(h+1)，m=(aver(ba_X) i，j_h，j_(h+1)，m ), st(va_X i，j_h，j_(h+1)，m )),

[0031] v i，j_h，j_(h+1)，m =(aver(X) i，j_h，j_(h+1)，m ), st(X i，j_h，j_(h+1)，m ));

[0032] S12, Note pe_v i，j_h，j_(h+1)，m va_v i，j_h，j_(h+1)，m v i，j_h，j_(h+1)，m The enclosed state triangle is called Q. i，j_h，j_(h+1)，m Calculate its area and denot it as s. i，j_h，j_(h+1)，m ;

[0033] S13. For each m, there is a set

[0034]

[0035] Then construct the following two sets:

[0036] MAX_S i，0,n,m =

[0037] {(max_s i，j_h，j_(h+1)，m ,m)|max_s i，j_h，j_(h+1)，m =maxS i，j_h，j_(h+1)，m},

[0038] MIN_S i，0，n,m =

[0039] {(min_s i，j_h，j_(h+1)，m ,m)|min_s i,j_h，j_(h+1)，m =minS i，j_h，j_(h+1)，m};

[0040] S14, Calculate MAX_S i，0,n，m The slope of the regression line, denoted as k. 1，i,0,n，m ; Calculate MIN_S i，0,n，m The slope of the regression line, denoted as k. 2，i,0,n,m Let k i，0,n,m =|k 2，i0，n，m -k 1，i,0,n，m |;

[0041] S15. Fix m and n, and let the natural number r ≥ 0. Then, in time T r，n+r Within, the following calculations are performed sequentially according to the aforementioned process:

[0042] ① Set X i，r,n+r ={x i，j |x i，j It is x i At time t j The value of tj ∈T r，n+r};

[0043] ② Set G i，r,n+r ={(x i,j , t j )|x i，j ∈X i，r,n+r , t j ∈T r，n+r};

[0044] ③ Set

[0045] ④ Set X i，j_h，j_(h+1)，m ={x i，p |x i，p ∈X i，r，n+r , t p ∈[t j_h , t j_(h+1) ]∩T r，n+r};

[0046] ⑤ Set G i，j_h，j_(h+1)，m ={(x i，p , t p )|x i，p ∈X i，j_h，j_(h+1) , t p ∈[t j_h , t j_(h+1) ]∩T r，n+r};

[0047] ⑥ Set MAX_S i，r，n+r，m MIN_S i，r,n+r，m ;

[0048] ⑦ Slope k 1，i，r，n+r，m k 2，i，r，n+r，m and the absolute value of the slope difference k i，r，n+r，m ;

[0049] ⑧This leads to the set

[0050] S16. The first batch of data was obtained through test production on the newly commissioned tire manufacturing line. Calculations were then performed using this first batch of data. And search results The first coordinate k of all elements in i，r，n+r，m The maximum value of max_k i，R,n，m and minimum value min_k i，R,n,m ;

[0051] S17. The tire manufacturing production line has officially started operation.

[0052] If detected, k i，r，n+r，m ∈[min_k i,R,n,mmax_k i，R,n，m The production line system was determined to be normal.

[0053] If detected, Determine if the production line system is malfunctioning;

[0054] S18, if The following values ​​are determined sequentially in reverse order:

[0055] ①From k i，r,n+r，m Read the value of r from the subscript;

[0056] ②Based on the calculation record of process S14, read k 1，i，r，n+r，m and k 2，i，r，n+r，m And compare the larger and smaller of the two;

[0057] ③If k 1，i，r，n+r，m ≥k 2，i，r，n+r，m Then in MAX_S i，r,n+r，m Find the maximum element (max_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the first time interval for investigating anomalies; simultaneously, in MIN_S i，r，n+r，m Find the minimum element (min_s) i，j_h，j_(h+1)，m S corresponding to m) i,j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the second time interval for investigating anomalies;

[0058] ④ If k 1，i，r，n+r，m <k 2，i，r，n+r，m Then in MIN_S i，r，n+r，m Find the maximum element (min_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the first time interval for anomaly detection; simultaneously, in MAX_S i，r，n+r，m Find the minimum element (max_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h，j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩Tr，n+r This is the second time interval for investigating anomalies.

[0059] Optionally, in step S4, the set sequence F is calculated. i,0,n，m The steps are as follows:

[0060] S41, G obtained based on S3 i，0,n ,make

[0061]

[0062] S42. Let m = m + 1, update m, and define a set G to iterate through it. i，0,n =G i，0,n -F i，0,n，m Update G i，0,n From X i,0，n Extract those that meet the conditions Build collection

[0063]

[0064] Among them, the extracted No replacement, update m operation until X i,0,n None of the remaining elements satisfy the condition. The maximum value of m is denoted as M.

[0065] Optionally, in step S5, its set of maximum points is calculated. The steps are as follows:

[0066] S51, based on F i,0,n,m ,make

[0067]

[0068] S52, in set F i,0,n,m In, construct a set

[0069]

[0070] Optionally, in step S8, the peak point set pe_X is calculated. i,j_h,j_(h+1),m And the valley point set va_X i,j_h,j_(h+1),m The steps are as follows:

[0071] S81, obtained from S7

[0072]

[0073] S82, Order

[0074]

[0075] S83. Therefore, we have

[0076]

[0077] S84, Order

[0078]

[0079] S85. Therefore, we have

[0080]

[0081] Optionally, calculate the mean of the peak point set aver(pe_X). i,j_h，j_(h+1)，m ), mean of the trough set aver(va_X) i,j_h，j_(h+1)，m ) and the mean of the entire sequence aver(X) i,j_h，j_(h+1)，m The steps are as follows:

[0082] S91, for all (x) i，u , t u )∈pe_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||pe_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0083]

[0084] S92, for all (x) i，u , t u )∈va_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||va_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0085]

[0086] S93, for all (x) i，u , t u )∈X i，j_h，j_(h+1)，m u = 1, 2, ..., ||X i,j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0087]

[0088] Optionally, in step S10, the standard deviation st(pe_X) of the peak point set is calculated. i,j_h，j_(h+1)，m ), Standard deviation of trough point set st(va_X) i,j_h，j_(h+1)，m ) and the standard deviation of the whole sequence st(X) i，j_h，j_(h+1)，m The steps are as follows:

[0089] S10-1, For all (x)i，u , t u )∈pe_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||pe_X i,j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0090]

[0091] S10-2, For all (x) i，u , t u )∈va_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||va_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0092]

[0093] S10-3, For all (x i，u , t u )∈X i，j_h，j_(h+1)，m u = 1, 2, ..., ||X i,j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0094]

[0095] Optionally, in step S12, the state triangle Q is calculated. i，j_h，j_(h+1)，m area s i，j_h，j_(h+1)，m The steps are as follows:

[0096] S12-1. According to Heron's formula, let the state triangle Q... i，j_h，j_(h+1)，m The lengths of the three sides are:

[0097] a = ||pe_v i，j_h，j_(h+1)，m -va_v i，j_h，j_(h+1)，m ||;

[0098] b = ||pe_v i，j_h，j_(h+1)，m -v i，j_h，j_(h+1)，m ||;

[0099] c = ||va_v i，j_h，j_(h+1)，m -v i，j_h，j_(h+1)，m ||;

[0100] S12-2, Order

[0101]

[0102] S12-3. Substituting p, a, b, and c into the following formula from the previous steps, we have:

[0103]

[0104] Optionally, in step S14, k is calculated. 1，i，0，n，m k 2，i，0，n，m k j，0，n，m The steps are as follows:

[0105] S14-1, derived from S13,

[0106] MAX_S i,0,n,m =

[0107] {(max_s i,j_h,j_(h+1),m ,m)|max_s i,j_h,j_(h+1),m =maxS i,j_h,j_(h+1),m},

[0108] Let the mean

[0109]

[0110] S14-2, From S12-1, using the least squares method, we can obtain the calculation formula.

[0111]

[0112] S14-3, derived from S13,

[0113] MIN_S i,0,n,m =

[0114] {(min_s i,j_h,j_(h+1),m ,m)|min_s i,j_h,j_(h+1),m =minS i,j_h,j_(h+1),m},

[0115] Let the mean

[0116]

[0117] S14-4, from S12-3, using the least squares method, we can obtain the calculation formula.

[0118]

[0119] S14-5. From the aforementioned steps, we obtain

[0120]

[0121] The data analysis method described above for analyzing anomalies in tire manufacturing process parameters is implemented based on a variable analysis module and a variable anomaly period location analysis module, wherein:

[0122] The variable analysis module is used to perform feature measurement and judgment analysis on abnormal signals in the tire manufacturing process;

[0123] The variable anomaly period location analysis module is used to locate the period during which abnormal process parameters occur in the tire manufacturing process under abnormal conditions.

[0124] Optionally, the variable analysis module includes an anomaly feature measurement module and an anomaly judgment analysis module. The anomaly feature measurement module obtains quantitative features by indexing anomalies, and the anomaly judgment analysis module determines the threshold of anomaly indicators based on the quantitative features and judges whether the variable is abnormal based on the threshold.

[0125] Compared with the prior art, the beneficial effects of this application are as follows:

[0126] 1) This application analyzes the variable data corresponding to tire manufacturing process parameters in segments based on the extreme time of the hierarchical extreme value sequence, and provides a calculation method: First, based on the extreme value hierarchical and time segmentation, the mean and standard deviation of the maximum value sequence, minimum value sequence and the full data sequence of the data within the segment are calculated respectively, and three state vectors are constructed based on this. Based on the state vectors, the stability of the process parameters within the segment is measured by the area of ​​the state triangle. The absolute value of the difference between the slopes of the maximum and minimum value sequences of the entire segment is used to measure the abnormal characteristics of the variables corresponding to the tire manufacturing process parameters in a specific time distribution. Then, the variables corresponding to the process parameters are allowed to slide on the time axis to obtain several abnormal characteristic indicators. Anomalies are identified from the changing trends of the abnormal characteristic indicators on the time axis. When an anomaly is determined to exist, the position of the anomaly occurrence time in the time segment is deduced to achieve anomaly time localization.

[0127] 2) The key features of this application are: the effective use of extreme value information, mean, and standard deviation to design novel analytical tools such as state vectors and state triangles, which support its effectiveness, reliability, versatility, and simplicity. Compared with existing methods, this application has advantages such as strong versatility and adaptability, good anomaly identification capability, simple calculation, and flexible deployment. In particular, it can accurately locate the time period of anomaly occurrence, promoting the analysis and optimization of process parameter anomalies. Attached Figure Description

[0128] Figure 1 This is a flowchart of a data analysis method for analyzing abnormal tire manufacturing process parameters, as disclosed in an embodiment of this application. Detailed Implementation

[0129] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0130] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0131] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0132] like Figure 1 As shown, a data analysis method for analyzing anomalies in tire manufacturing process parameters includes the following steps:

[0133] S1. Data collected from the tire manufacturing process is stored in the corresponding variable x. i Below, the number of variables is M, i = 1, 2, ..., M, which constitutes the variable pool X = {x i |i = 1, 2, ..., M, where M is a positive integer};

[0134] S2, Read the target variable x i In [t0,t n Construct a set of all data within a given time period:

[0135] Where t0≤t j ≤t n And t j In [t0,t n The intervals within the range are uniform, 0 ≤ j ≤ n, T 0,n For all t j The time set constituted is defined as n≥1000, where j is a natural number;

[0136] S3. Constructing a two-dimensional point set in, yes The coordinates of the variable, t j yes Time coordinates;

[0137] S4, based on G i,0,n The set sequence is obtained stepwise by using a maximum approximation method according to the number of iterations m.

[0138]

[0139] Where m is the number of layers, and m = 1, 2, ..., M;

[0140]

[0141] S5. F is established for any 1≤m≤M. i,0,n,m (t), calculate its maximum point set according to the variable coordinates. make therefore Not empty, "∶=" means "equals", where The time coordinates of the elements are as follows Represents a set The number of elements;

[0142] S6 Time coordinates of any element In T 0，n The closest time in the inner distance is t j_h According to this correspondence method, The time coordinates t of all elements j_h Converging into a collection

[0143] S7, use Adjacent time periods in T 0，n G on i，0，n Perform segmentation, that is, take any interval [t] j_h , t j_(h+1) ], according to the time sequence of the data from set X i，0，n Read x from i In [t] j_h , t j_(h+1) Construct a set containing all data in ∩T:

[0144]

[0145] S8, the point set G in the plane coordinate system i，j_h，j_(h+1)，m ,according to The set of changing peaks is pe_X i，j_h，j_(h+1)，m The set of trough points is va_X i，j_h，j_(h+1)，m ;

[0146] S9. Calculate pe_X i，j_h，j_(h+1)，m The mean is denoted as aver(pe_X). i,j_h，j_(h+1)，m ), calculate va_X i,j_h，j_(h+1)，m The mean is denoted as aver(va_X). i，j_h，j_(h+1)，m ), calculate X i,j_h，j_(h+1)，m The mean is denoted as aver(X). i，j_h，j_(h+1)，m );

[0147] S10, Calculate pe_X i,j_h，j_(h+1)，m Standard deviation, denoted as st(pe_X) i,j_h，j_(h+1)，m ), calculate va_X i,j_h，j_(h+1)，m The standard deviation is denoted as st(va_X). i,j_h，j_(h+1)，m ), calculate X i,j_h，j_(h+1)，m The standard deviation is denoted as st(X).i,j_h，j_(h+1)，m );

[0148] S11, Constructing Vectors:

[0149] pe_v i，j_h，j_(h+1)，m =(aver(pe_X) i，j_h，j_(h+1)，m ), st(pe_X i,j_h，j_(h+1)，m )),

[0150] va_v i，j_h，j_(h+1)，m =(aver(ba_X) i，j_h，j_(h+1)，m ), st(va_X i，j_h，j_(h+1)，m )),

[0151] v i，j_h，j_(h+1)，m =(aver(X) i，j_h，j_(h+1)，m ), st(X i，j_h，j_(h+1)，m ));

[0152] S12, Note pe_v i，j_h，j_(h+1)，m va_v i，j_h，j_(h+1)，m v i，j_h，j_(h+1)，m The enclosed state triangle is called Q. i，j_h，j_(h+1)，m Calculate its area and denot it as s. i，j_h，j_(h+1)，m ;

[0153] S13. For each m, there is a set

[0154]

[0155] Then construct the following two sets:

[0156] MAX_S i，0,n,m =

[0157] {(max_s i，j_h，j_(h+1)，m ,m)|max_s i,j_h，j_(h+1)，m =maxS i，j_h,j_(h+1)，m},

[0158] MIN_S i，0,n,m =

[0159] {(min_s i，j_h，j_(h+1)，m ,m)|min_s i,j_h，j_(h+1)，m =minS i,j_h，j_(h+1)，m};

[0160] S14, Calculate MAX_S i，0,n，m The slope of the regression line, denoted as k. 1，i,0,n，m ; Calculate MIN_S i，0,n，m The slope of the regression line, denoted as k. 2，i,0,n,m Let k i,0,n,m =|k 2，i，0，n，m -k 1，i,0,n，m |

[0161] S15. Fix m and n, and let the natural number r ≥ 0. Then, in time T r，n+r Within, the following calculations are performed sequentially according to the aforementioned process:

[0162] ① Set X i，r，n+r ={x i，j |x i，j It is x i At time t j The value of t j ∈T r，n+r};

[0163] ② Set G i，r，n+r ={(x i,j , t j )|x i，j ∈X i，r，n+r , t j ∈T r，n+r};

[0164] ③ Set

[0165] ④ Set X i，j_h，j_(h+1)，m ={x i，p |x i，p ∈X i，r,n+r , t p ∈[t j_h , t j_(h+1) ]∩T r，n+r};

[0166] ⑤ Set G i，j_h，j_(h+1)，m ={(x i,p , t p )|x i，p ∈X i，j_h，j_(h+1) , t p ∈[t j_h , t j_(h+1) ]∩T r，n+r};

[0167] ⑥ Set MAX_S i，r，n+r，m MIN_S i，r,n+r，m ;

[0168] ⑦ Slope k 1，i，r，n+r，m k 2，i，r，n+r，m and the absolute value of the slope difference k i，r，n+r，m ;

[0169] ⑧This leads to the set

[0170] S16. The first batch of data was obtained through test production on the newly commissioned tire manufacturing line. Calculations were then performed using this first batch of data. And search results The first coordinate k of all elements in i，r，n+r，m The maximum value of max_k i，R,n，m and minimum value min_k i，R,n,m .

[0171] S17. The tire manufacturing production line has officially started operation.

[0172] If detected, k i，r，n+r，m ∈[min_k i，R,n,m max_k i，R,n，m The production line system was determined to be normal.

[0173] If detected, The production line system was determined to be malfunctioning.

[0174] S18, if The following values ​​are determined sequentially in reverse order:

[0175] ①From k i，r,n+r，m Read the value of r from the subscript;

[0176] ②Based on the calculation record of process S14, read k 1，i，r，n+r，m and k 2，i，r，n+r，m And compare the larger and smaller of the two;

[0177] ③If k 1，i，r，n+r，m ≥k 2，i，r，n+r，m Then in MAX_S i，r,n+r，m Find the maximum element (max_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the first time interval for investigating anomalies; simultaneously, in MIN_S i，r，n+r，m Find the minimum element (min_s) i，j_h，j_(h+1)，m S corresponding to m) i,j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the second time interval for investigating anomalies;

[0178] ④ If k 1，i，r，n+r，m <k 2，i，r，n+r，m Then in MIN_S i，r，n+r，m Find the maximum element (min_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h,j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined.j_h , t j_(h+1) ]∩T r，n+r This is the first time interval for anomaly detection; simultaneously, in MAX_S i，r，n+r，m Find the minimum element (max_s) i，j_h，j_(h+1)，m S corresponding to m) i，j_h，j_(h+1)，m And the corresponding s i，j_h，j_(h+1)，m Thus, h is determined, and the interval [t] is determined. j_h , t j_(h+1) ]∩T r，n+r This is the second time interval for investigating anomalies.

[0179] Through the above technical solution, using the maximum approximation method, the variable x corresponding to the tire manufacturing process parameters is obtained. i In the time distribution set T 0,n The original data sequence X i,0,n Stratified extraction of maximum value sequence F i,0,n，m Where m is the number of layers, and then F i,0,n，m The time corresponding to the maximum value is at T 0,n The most recent moment on X i,0,n The elements in the data are grouped according to time, and in each time interval [t] j_h , t j_(h+1) Within ∩T, by calculating the maximum, minimum, and all values ​​respectively, the average value and standard deviation of the tire were realized. The average value and standard deviation were then considered as two dimensions of the data within the measurement group, thus constructing the state vector pe_v of the tire manufacturing process parameters for that period. i，j_h，j_(h+1)，m va_v i，j_h，j_(h+1)，m and v i，j_h，j_(h+1)，m Furthermore, based on the state triangle Q formed by the state vectors of these three process parameters... i，j_h，j_(h+1)，m area s i，j_h，j_(h+1)，m The magnitude of F is used to measure the stability of the changes in the set of process parameters. In particular, the area of ​​the state triangle is 0 when the three state vectors are collinear. Because F i,0,n，m It is extracted layer by layer according to the number of layers m, so from X i,0,n At least 1 layer of F can be obtained. i,0,n，m Typically, it is a multi-layer F i,0,n，m Therefore, in each time interval [t] j_h , t j-(h+1) Within ∩T, multiple layers of F i,0,n，m Get multiple s i，j_h，j_(h+1)，m We get from all [t] j_h , t j_(h+1) Take the maximum value s from the middle of ∩T i，j_h，j_(h+1)，m and minimum s i，j_h，j_(h+1)，m Constitute the set MAX_S i，0,n，m and MIN_S i，0,n,mCalculate the slope k of the regression line for each of the two point sets. 1，i,0,n，m and k 2，i,0,n，m And the absolute value k of the difference between these two slopes. i,0,n,m .k i,0,n，m It is the variable x that measures the tire manufacturing process parameters. i In the time distribution set T 0，n Indicators of abnormal characteristics on. Then, we let x i On the timeline, T 0,n By sliding the initial segment, several equal-length time intervals T are obtained. r，n+r and in T r，n+r The corresponding k is obtained from the above. i，r,n+r，m Specifically, we measure its k before the new production line is officially put into use, or during the trial period of the new production line. i，r，n+r，m Sequence, and obtain its initial k. i，r，n+r，m The maximum value of the sequence max_k i，R,n，m and the minimum value min_k i，R,n,m To construct k i，r,n+r，m The initial value range [min_k] i，R,n,m max_k i，R,n，m ]. Once the production line is officially put into use, if k i，r，n+r，m If the changes exceed this range, it means that the production line's state has shifted beyond its previous normal state, indicating an anomaly. Anomaly analysis and location should be performed. This completes the entire method's operation and allows it to function effectively. This methodological process preprocesses, identifies, and locates time-period anomalies in tire manufacturing process parameters. Its advantages include good versatility, adaptability, and anomaly identification capabilities; it can detect the time periods when anomaly signals occur; and it is computationally simple and flexibly deployable.

[0180] Optionally, in step S4, the set sequence F is calculated. i，0,n，m The steps are as follows:

[0181] S41, G obtained based on S3 i，0,n ,make

[0182]

[0183] S42. Let m = m + 1, update m, and define a set G to iterate through it. i，0,n =G i，0,n -F i,0,n，m Update G i，0,n From X i，0，n Extract those that meet the conditions Build collection

[0184]

[0185] Among them, the extracted No replacement, update m operation until X i,0,n None of the remaining elements satisfy the condition. The maximum value of m is denoted as M;

[0186] Through the above technical solution, using the maximum approximation method, the variable x corresponding to the tire manufacturing process parameters is obtained. i In the time distribution set T 0,n The original data sequence X i,0,n Stratified extraction of maximum value sequence F i,0,n,m Where m is the number of layers. Extreme values ​​are an important aspect of data fluctuation characteristics. Layered extreme value extraction of data can effectively characterize the changes in corresponding variables of tire manufacturing process parameters, providing good preprocessing results for subsequent analysis. This technical solution has the advantages of good versatility and applicability, and simple calculation.

[0187] Optionally, in step S5, its set of maximum points is calculated. The steps are as follows:

[0188] S51, based on F i,0,n,m ,make

[0189]

[0190] S52, in set F i,0,n,m In, construct a set

[0191]

[0192] Optionally, in step S8, the peak point set pe_X is calculated. i,j _ h,j_(h+1),m And the valley point set va_X i,j_h,j_(h+1),m The steps are as follows:

[0193] S81, obtained from S7

[0194]

[0195] S82, Order

[0196]

[0197] S83. Therefore, we have

[0198]

[0199] S84, Order

[0200]

[0201] S85. Therefore, we have

[0202]

[0203] Through the above technical solution, the peak point set pe_X is made possible. i，j_h，j_(h+1)，m And the valley point set va_X i，j_h，j_(h+1)，m All points are non-empty sets, avoiding errors in subsequent calculations. This technical solution has the advantage of computational simplicity.

[0204] Optionally, in step S9, the mean value of the peak point set aver(pe_X) is calculated. i，j_h，j_(h+1)，m ), mean of the trough set aver(va_X) i，j_h，j_(h+1)，m ) and the mean of the entire sequence aver(X) i，j_h，j_(h+1)，m The steps are as follows:

[0205] S91, for all (x) i，u , t u )∈pe_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||pe_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0206]

[0207] S92, for all (x) i，u , t u )∈va_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||va_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0208]

[0209] S93, for all (x) i，u , t u )∈X i，j_h，j_(h+1)，m u = 1, 2, ..., ||X i,j_h，j_(h+1)，m ||, variable coordinates x i，u The numerical values ​​are summed and averaged, that is:

[0210]

[0211] Optionally, in step S10, the standard deviation st(pe_X) of the peak point set is calculated. i,j_h，j_(h+1)，m ), Standard deviation of trough point set st(va_X) i,j_h，j_(h+1)，m ) and the standard deviation of the whole sequence st(X) i，j_h，j_(h+1)，m The steps are as follows:

[0212] S10-1, For all (x) i，u , t u)∈pe_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||pe_X i,j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0213]

[0214] S10-2, For all (x) i，u , t u )∈va_X i，j_h，j_(h+1)，m u = 1, 2, ..., ||va_X i，j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0215]

[0216] S10-3, For all (x i，u , t u )∈X i，j_h，j_(h+1)，m u = 1, 2, ..., ||X i,j_h，j_(h+1)，m ||, variable coordinates x i，u Standard deviation:

[0217]

[0218] Optionally, in step S12, the state triangle Q is calculated. i，j_h，j_(h+1)，m area s i，j_h，j_(h+1)，m The steps are as follows:

[0219] S12-1. According to Heron's formula, let the state triangle Q... i，j_h，j_(h+1)，m The lengths of the three sides are:

[0220] a = ||pe_v i，j_h，j_(h+1)，m -va_v i，j_h，j_(h+1)，m ||;

[0221] b = ||pe_v i，j-h，j_(h+1)，m -v i，j-h，j_(h+1)，m ||;

[0222] c = ||va_v i，j_h，j_(h+1)，m -v i，j-h，j_(h+1)，m ||;

[0223] S12-2, Order

[0224]

[0225] S12-3. Substituting p, a, b, and c into the following formula from the previous steps, we have:

[0226]

[0227] Using the above technical solution, guided by Heron's formula, and through the state triangle Q... i,j_h,j_(h+1),m The area is calculated by the difference of vertex vectors, avoiding complex calculations such as trigonometric functions. This technical solution has the advantage of simple calculation.

[0228] Optionally, in step S14, k is calculated. 1,i,0,n,m k 2,i,0,n,m k i,0,n,m The steps are as follows:

[0229] S14-1, derived from S13,

[0230] MAX_S i,0,n,m =

[0231] {(max_s i,j_h,j_(h+1),m ,m)|max_s i,j_h,j_(h+1),m =maxS i,j_h,j_(h+1),m},

[0232] Let the mean

[0233]

[0234] S14-2, From S12-1, using the least squares method, we can obtain the calculation formula.

[0235]

[0236] S14-3, derived from S13,

[0237] MIN_S i,0,n,m =

[0238] {(min_s i,j_h,j_(h+1),m ,m)|min_s i,j_h,j_(h+1),m =minS i,j_h,j_(h+1),m},

[0239] Let the mean

[0240]

[0241] S14-4, from S12-3, using the least squares method, we can obtain the calculation formula.

[0242]

[0243] S14-5. From the aforementioned steps, we obtain

[0244]

[0245] Through the above technical solution, the variable x corresponding to the tire manufacturing process parameters is... i In the time distribution set T 0,n Data sequence Xi,0,n k i,0,n,m It is an important indicator of abnormal characteristics. This technical solution has the advantages of good versatility and adaptability, and simple calculation.

[0246] The data analysis method described above for analyzing anomalies in tire manufacturing process parameters is implemented based on a variable analysis module and a variable anomaly period location analysis module, wherein:

[0247] The variable analysis module is used to perform feature measurement and judgment analysis on abnormal signals in the tire manufacturing process;

[0248] The variable anomaly period location analysis module is used to locate the period during which abnormal process parameters occur in the tire manufacturing process under abnormal conditions.

[0249] Through the collaborative work of the variable analysis module and the variable anomaly time period location analysis module, the functions of preprocessing, identifying and locating the time period of abnormal process parameters in the tire manufacturing process are realized. It not only has the advantages of versatility, adaptability and good anomaly identification ability, but also can accurately locate the time period of abnormal signal occurrence.

[0250] Optionally, the variable analysis module includes an anomaly feature measurement module and an anomaly judgment analysis module. The anomaly feature measurement module obtains quantitative features by indexing anomalies, and the anomaly judgment analysis module determines the threshold of anomaly indicators based on the quantitative features and judges whether the variable is abnormal based on the threshold.

[0251] The above technical solution enables the calculation and identification of abnormal process parameters in tire manufacturing, and has the advantages of versatility, adaptability and good abnormal identification ability.

[0252] This application addresses the anomaly analysis of tire manufacturing process parameters, focusing on resolving the shortcomings of existing calculation methods in the field, such as insufficient versatility, inability to automatically adapt to parameter feature migration as production lines age, and the inability to identify anomalous signals without providing time-segmented localization, thus hindering the analysis of process parameter anomalies and restricting continuous improvement in process parameter optimization. It creatively proposes segmenting the variable data corresponding to tire manufacturing process parameters using the extreme value time of a hierarchical extreme value sequence, and provides a calculation method. Furthermore, based on extreme value hierarchies and time segmentation, it proposes calculating the mean and standard deviation of the maximum, minimum, and full data sequences within each segment, and constructing three state vectors accordingly. Based on these state vectors, the stability of process parameters within each segment is measured using the area of ​​a state triangle. The absolute value of the difference between the slopes of the maximum and minimum area sequences of all segments is used as an indicator to measure the anomalous characteristics of the variables corresponding to tire manufacturing process parameters in a specific time distribution. Then, the variables corresponding to the process parameters are allowed to slide along a time axis to obtain several indicators of anomalous characteristics, and anomalies are identified from the changing trends of these indicators on the time axis. When an anomaly is detected, the location of the anomaly's occurrence within a time segment is determined, thus achieving anomaly time localization. The application of extreme values, the meaning extraction of mean and standard deviation, and the design of state vectors and state triangles are prominent features of this invention, playing a crucial role in supporting its effectiveness, reliability, versatility, and simplicity. The method proposed in this invention exhibits good versatility, adaptability, and anomaly identification capabilities, possessing a significant ability to detect the timing of anomaly signal occurrences, and boasts advantages such as simple computation and flexible deployment.

Claims

1. A data analysis method for analyzing anomalies in tire manufacturing process parameters, characterized in that, The data analysis method is based on a variable analysis module and a variable anomaly time period location analysis module. The variable analysis module is used to perform feature measurement and judgment analysis on abnormal signals of tire manufacturing process. The variable anomaly period location analysis module is used to locate the period of occurrence of abnormal process parameters in the tire manufacturing process under abnormal conditions. The variable analysis module includes an anomaly feature measurement module and an anomaly judgment analysis module. The anomaly feature measurement module obtains quantitative features by indexing the anomaly. The anomaly judgment analysis module determines the threshold of the anomaly index by the quantitative features and judges whether the variable is abnormal based on the threshold. The data analysis method includes the following steps: S1. Data collected from the tire manufacturing process is stored in the corresponding variable x. i Below, the number of variables is M, i = 1, 2, ..., M, which constitutes the variable pool X = {x i |i = 1, 2, ..., M, where M is a positive integer}; S2, Read the target variable x i In [t0,t n Construct a set of all data within a given time period: Where t0≤t j ≤t n And t j In [t0,t n The intervals within the range are uniform, 0 ≤ j ≤ n, T 0,n For all t j The time set constituted is defined as n≥1000, where j is a natural number; S3. Constructing a two-dimensional point set in, yes The coordinates of the variable, t j yes Time coordinates; S4, based on G i,0,n The set sequence is obtained stepwise by using a maximum approximation method according to the number of iterations m. Where m is the number of layers, and m = 1, 2, ..., M; S5. F is established for any 1≤m≤M. i,0,n,m (t), calculate its maximum point set according to the variable coordinates. make therefore Not empty, "∶=" means "equals", where The time coordinates of the elements are as follows Represents a set The number of elements; S6 Time coordinates of any element In T 0,n The closest time in the inner distance is t j_h According to this correspondence method, The time coordinates t of all elements j_h Converging into a collection S7, use Adjacent time periods in T 0,n G on i,0,n Perform segmentation, that is, take any interval [t] j_h ,t j_(h+1) ], according to the time sequence of the data from set X i,0,n Read x from i In [t] j_h ,t j_(h+1) Construct a set containing all data in ∩T: S8, the point set G in the plane coordinate system i,j_h,j_(h+1),m According to x i,tp The set of changing peaks is pe_X i,j_h,j_(h+1),m The set of trough points is va_X i,j_h,j_(h+1 ) ,m ; S9. Calculate pe_X i,j_h,j_(h+1),m The mean is denoted as aver(pe_X). i,j_h,j_(h+1),m) Calculate va_X i,j_h,j_(h+1),m The mean is denoted as aver(va_X). i,j_h,j_(h+1),m) Calculate X i,j_h,j_(h+1),m The mean is denoted as aver(X). i,j_h,j_(h+1),m) ; S10, Calculate pe_X i,j _ h,j_(h+1),m Standard deviation, denoted as st(pe_X) i,j_h,j_(h+1),m) Calculate va_X i,j_h,j_(h+1),m The standard deviation is denoted as st(va_X). i,j_h,j_(h+1),m) Calculate X i,j_h,j_(h+1),m The standard deviation is denoted as st(X). i,j_h,j_(h+1) , m) ; S11, Constructing Vectors: on_v i,j_h,j_(h+1),m =(have(on_X) i,j_h,j_(h+1),m) ,st(on_X i,j_h,j_(h+1),m )), va_v i,j_h,j_(h+1),m =(ver(ba_X i,j_h,j_(h+1),m ),st(va_X i,j_h,j_(h+1),m )), v i,j_h,j_(h+1),m =(aver(X i,j_h,j_(h+1),m ),st(X i,j_h,j_(h+1),m )); S12, Note pe_v i,j_h,j_(h+1),m va_v i,j_h,j_(h+1),m v i,j_h,j_(h+1),m The enclosed state triangle is called Q. i,j_h,j_(h+1),m Calculate its area and denot it as s. i,j_h,j_(h+1),m ; S13. For each m, there is a set Then construct the following two sets: MAX_S i,0,n,m = {(max_s i,j_h,j_(h+1),m ,m)|max_s i,j_h,j_(h+1),m =maxS i,j_h,j_(h+1),m }, MIN_S i,0,n,m = {(min_s i,j_h,j_(h+1),m ,m)|min_s i,j_h,j_(h+1),m =minS i,j_h,j_( h +1),m }; S14, Calculate MAX_S i,0,n,m The slope of the regression line, denoted as k. 1,i,0,n,m ; Calculate MIN_S i,0,n,m The slope of the regression line, denoted as k. 2,i,0,n,m Let k i,0,n,m =|k 2,i,0,n,m -k 1,i,0,n,m |; S15. Fix m and n, and let the natural number r ≥ 0. Then, in time T r,n+r Within, the following calculations are performed sequentially according to the aforementioned process: ① Set X i,r,n+r ={x i,j |x i,j It is x i At time t j The value of t j ∈T r,n+r }; ② Set G i,r,n+r ={(x i,j ,t j )|x i,j ∈X i,r,n+r ,t j ∈T r,n+r }; ③ Set ④ Set Xx ,j_h,j_(h+1),m ={x i,p |x i,p ∈X i,r,n+r ,t p ∈[t j_H ,t j_(h+1) ]∩T r,n+r }; ⑤ Set G i,j_h,j_(h+1 ) ,m ={(x i,p ,t p )|x i,p ∈X i,j_h,j_(h+1) ,t p ∈[t j_h ,t j_(h+1) ]∩T r,n+r }; ⑥ Set MAX_S i,r,n+r,m MIN_S i,r,n+r,m ; ⑦ Slope k 1,i,r,n+r,m k 2,i,r,n+r,m and the absolute value of the slope difference k i,r,n+r,m ; ⑧This leads to the set S16. The first batch of data was obtained through test production on the newly commissioned tire manufacturing line. Calculations were then performed using this first batch of data. And search results The first coordinate k of all elements in i,r,n+r,m The maximum value of max_k i,R,n,m and minimum value min_k i,R,n,m ; S17. The tire manufacturing production line has officially started operation. If detected, k i,r,n+r,m ∈[min_k i,R,n,m ,max_k i,R,n,m The production line system was determined to be normal. If detected, Determine if the production line system is malfunctioning; S18, if The following values ​​are determined sequentially in reverse order: ①From k i,r,n+r,m Read the value of r from the subscript; ②Based on the calculation record of process S14, read k 1,i,r,n+r,m and k 2,i,r,n+r,m And compare the larger and smaller of the two; ③If k 1,i,r,n+r,m ≥k 2,i,r,n+r,m Then in MAX_S i,r,n+r,m Find the maximum element (max_s) i,j_h,j_(h+1),m S corresponding to m) i,j_h,j_(h+1),m And the corresponding s i,j_h,j_(h+1),m Thus, h is determined, and the interval [t] is determined. j_h ,t j_(h+1) ]∩T r,n+r This is the first time interval for investigating anomalies; simultaneously, in MIN_S i,r,n+r,m Find the minimum element (min_s) i,j_h,j_(h+1),m S corresponding to m) i,j_h,j_(h+1),m And the corresponding s i,j_h,j_(h+1),m Thus, h is determined, and the interval [t] is determined. j_h ,t j_(h+1) ]∩T r,n+r This is the second time interval for investigating anomalies; ④ If k 1,i,r,n+r,m <k 2,i,r,n+r,m Then in MIN_S i,r,n+r,m Find the maximum element (min_s) i,j_h,j_(h+1),m S corresponding to m) i,j_h,j_(h+1),m And the corresponding s i,j_h,j_(h+1),m Thus, h is determined, and the interval [t] is determined. j_h ,t j_(h+1) ]∩T r,n+r This is the first time interval for anomaly detection; simultaneously, in MAX_S i,r,n+r,m Find the minimum element (max_s) i,j_h,j_(h+1),m S corresponding to m) i,j_h,j_(h+1),m And the corresponding s i,j_h,j_(h+1),m Thus, h is determined, and the interval [t] is determined. j_h ,t j_(h+1) ]∩T r,n+r This is the second time interval for investigating anomalies.

2. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, In step S4, the set sequence F is calculated. i,0,n,m The steps are as follows: S41, G obtained based on S3 i,0,n ,make S42. Let m = m + 1, update m, and define a set G to iterate through it. i,0,n =G i,0,n -F i,0,n,m Update G i,0,n From X i,0,n Extract those that meet the conditions Build collection Among them, the extracted No replacement, update m operation until X i,0,n None of the remaining elements satisfy the condition. The maximum value of m is denoted as M.

3. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, In step S5, the set of its maximum points is calculated. The steps are as follows: S51, based on F i,0,n,m ,make S52, in set F i,0,n,m In, construct a set 4. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, In step S8, the peak point set pe_X is calculated. i,j_h,j_(h+1),m And the valley point set va_X i,j_h,j_(h+1),m The steps are as follows: S81, obtained from S7 S82, Order S83. Therefore, we have S84, Order S85. Therefore, we have 5. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, Calculate the mean of the peak point set aver(pe_X) i,j_h,j_(h+1),m ), mean of the trough set aver(va_X) i,j_h,j_(h+1),m ) and the mean of the entire sequence aver(X) i,j_h,j_(h+1),m The steps are as follows: S91, for all (x) i,u ,t u )∈pe_X i,j_h,j_(h+1),m u = 1, 2, ..., ||pe_X i,j_h,j_(h+1),m ||, variable coordinates x i,u The numerical values ​​are summed and averaged, that is: S92, for all (x) i,u ,t u )∈va_X i,j_h,j_(h+1),m u = 1, 2, ..., ||va_X i,j_h,j_(h+1),m ||, variable coordinates x i,u The numerical values ​​are summed and averaged, that is: S93, for all (x) i,u ,t u )∈X i,j_h,j_(h+1),m u = 1, 2, ..., ||X i,j_h,j_(h+1),m ||, variable coordinates x i,u The numerical values ​​are summed and averaged, that is:

6. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, In step S10, the standard deviation of the peak point set st(pe_X) is calculated. i,j_h,j_(h+1),m ), Standard deviation of trough point set st(va_X) i,j_h,j_(h+1),m ) and the standard deviation of the whole sequence st(C) i,j_h,j_(h+1),m The steps are as follows: S10-1, For all (x) i,u ,t u )∈pe_X i,j_h,j_(h+1),m u = 1, 2, ..., ||pe_X i,j_h,j_(h+1),m ||, variable coordinates x i,u Standard deviation: S10-2, For all (x) i,u ,t u )∈va_X i,j_h,j_(h+1),m u = 1, 2, ..., ||va_X i,j_h,j_(h+1),m ||, variable coordinates x i,u Standard deviation: S10-3, For all (x i,u ,t u )∈X i,j_h,j_(h+1),m u = 1, 2, ..., ||X i,j_h,j_(h+1),m ||, variable coordinates x i,u Standard deviation:

7. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 1, characterized in that, In step S12, the state triangle Q is calculated. i,j_h,j_(h+1),m area s i,j_h,j_(h+1),m The steps are as follows: S12-1. According to Heron's formula, let the state triangle Q... i,j_h,j_(h+1),m The lengths of the three sides are: a=||on_v j,j_h,j_(h+1),m -va_v i,j_h,j_(h+1 ) ,m ||; b=||on_v i,j_h,j_(h+1 ) ,m V i,j_h,j_(h+1 ) ,m ||; c=||va_v i,j_h,j_(h+1 ) ,m -v i,j_h,j_(h+1 ) ,m ||; S12-2, Order S12-3. Substituting p, a, b, and c into the following formula from the previous steps, we have:

8. The data analysis method for analyzing anomalies in tire manufacturing process parameters according to claim 7, characterized in that, In step S14, k is calculated. 1,i,0,n,m k 2,i,0,n,m k i,0,n,m The steps are as follows: S14-1, derived from S13, MAX_S i,0,n,m = {(max_s i,j_h,j_(h+1),m ,m)|max_s i,j_h,j_(h+1),m =maxS i,j_h,j_(h+1),m }, Let the mean S14-2, From S12-1, using the least squares method, we can obtain the calculation formula. S14-3, derived from S13, MIN_S i,0,n,m = {(min_s i,j_h,j_(h+1),m ,m)|min_s i,j_h,j_(h+1),m =minS i,j_h,j_(h+1),m Let the mean S14-4, from S12-3, using the least squares method, we can obtain the calculation formula. S14-5. From the aforementioned steps, we obtain

Images