Method for filtering pipe assembly process data

Abstract

This invention discloses a data filtering and processing method for pipeline assembly processes, comprising the following steps: acquiring the original data of the pipeline assembly process; using the original angle data as the abscissa, performing fidelity preservation and correction processing on the original angle data, correcting the abscissa of the backtracking; setting iteration parameters, including half-range change rate, amplification factor, and number of iterations; filtering the assembly process data; obtaining the Frechet distance between the filtered data and the original data; using the set iteration parameters, iteratively calculating the change rate; obtaining the optimal change rate; smoothing the assembly process data; and cleaning duplicate data. By accurately filtering the extra data generated during the multi-step tightening process due to factors such as tool withdrawal and overcoming friction, the key information of the original data is completely preserved. This not only solves the data filtering problem in the prior art but also significantly improves the accuracy, precision, and reliability of assembly quality analysis.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, specifically relating to a data filtering and processing method for pipeline assembly process. Background Technology

[0002] The quality of pipeline assembly directly affects the safety and performance of various systems, and data, as the key to identifying assembly quality problems, is of great significance for improving assembly efficiency and accuracy. To obtain data during the conduit assembly process, existing tightening tools can typically record assembly process data in real time, including the rotation angle of the outer nut and the torque applied by the operator throughout the operation. By analyzing and interpreting this assembly process data, quality problems can be identified early, helping to improve the quality of conduit assembly.

[0003] When a conduit is tightened in one go, the assembly process data obtained is continuous and smooth, which can be directly used for assembly quality analysis. However, during the manual assembly of hydraulic conduits, due to factors such as limited assembly space, operators often cannot complete the tightening in one go and need to perform multiple tightening steps. In this case, the tightening tool will record irrelevant data generated by tool withdrawal, friction overcoming, and other operations, which cannot be directly used for assembly quality analysis.

[0004] Existing data filtering methods mainly include various filters such as high-pass filters, low-pass filters, and band-pass filters. They filter based on the frequency characteristics of the signal and are a global filtering method. They cannot filter irrelevant data at specific points, which leads to the loss of key information and affects the accurate assessment of assembly quality, making it difficult to meet actual usage requirements. Summary of the Invention

[0005] The purpose of this invention is to provide a data filtering and processing method for pipeline assembly process, so as to solve the problem that existing data filtering methods are prone to losing key information when processing pipeline assembly data, thus affecting the accurate assessment of pipeline assembly quality.

[0006] This invention is achieved through the following technical solution:

[0007] The data filtering and processing method for pipeline assembly includes the following steps:

[0008] S1. Obtain the raw data of the pipeline assembly process and form the raw angle data sequence and raw torque data sequence;

[0009] S2. Using the original angle data as the x-axis, perform fidelity and correction processing on the original angle data, and correct the x-axis of the backtracking.

[0010] S3. Set the iteration parameters, including half-range change rate, amplification factor, and number of iterations;

[0011] S4. Filter assembly process data, including formulating filtering criteria based on the critical position change rate, and filtering the assembly process data according to the filtering criteria to obtain filtered data.

[0012] S5. Obtain the Frechet distance between the filtered data and the original data;

[0013] S6. Using the set iteration parameters, iteratively calculate the rate of change;

[0014] S7. Obtain the optimal rate of change;

[0015] S8. Smooth the assembly process data;

[0016] S9. Clean up duplicate data.

[0017] In some embodiments, in step S1, the original angle data sequence is defined as PA, and the original torque data sequence is defined as PT; expressed as:

[0018] PA = (pa1, pa2, ..., pa) m );

[0019] PT = (pt1, pt2, ..., pt) m );

[0020] Where m represents the number of data points.

[0021] In some embodiments, in step S2, the correction angle data sequence is defined as CA, and the correction torque data sequence is defined as CT; wherein, CA = (ca1, ca2, ..., ca m ), CT = (ct1, ct2, ..., ct m ), where m represents the number of data points;

[0022] Define ca1 = pa1 and ct1 = pt1. Starting from the second data point i = 2 and ending at i = m, iterate through the original angle data sequence PA.

[0023] If pa i ≥pa i-1 The process data is preserved with high fidelity using formula 1).

[0024]

[0025] If pa i <pa i-1 The rollback data is corrected using formula 2).

[0026]

[0027] In some embodiments, in step S3, the half-range data point positions are defined, and the correction angle data sequence CA and the torque data sequence are calculated as the half-range change rate K of the CT. H ; indicates as:

[0028] ct h ca h These are the data corresponding to the half-range data points in CA and CT, respectively.

[0029] Define the amplification factor as U, and the rate of change of the upper limit of the loop as N. C =UK H Let T be the number of iterations.

[0030] In some embodiments, in step S4, starting from the second data point i=2 and ending at i=m, the rate of change K at the current position is calculated. P ; indicates as:

[0031]

[0032] Define the filter angle data sequence as SA, the filter torque data sequence as ST, the critical position change rate as s, and the initial value of s as 0.

[0033] In some embodiments, in step S5, the original data curve is defined as PC=(PA,PT)', which is represented as:

[0034] Define the filtered data curve as SC=(SA,ST)', which is expressed as:

[0035] Define PC and SC as two curves located on the metric space (S, d), where d is the distance function between two points in X, the starting and ending points of PC are in the interval [a, b], and the starting and ending points of SC are in the interval [c, d].

[0036] Define the Frechet distance between PC and SC as FD, and calculate the value of FD according to formula 5):

[0037] FD = inf α,β max t∈[0,1] {d(PC(α(t)),SC(β(t)))}……5);

[0038] α is a continuous non-decreasing function mapping [0,1] to [a,b], and β is a continuous non-decreasing function mapping [0,1] to [c,d], representing the "walking" speed along curves PC and SC, respectively; α(t) and β(t) represent the positions along curves PC and SC at time t, respectively; Inf represents the infimum, which is the minimum value of the largest d(PC(α(t)), SC(β(t))) among all possible pairs of α and β;

[0039] For any point pc on curve PC i and any point sc on curve SC j The Euclidean distance d(pc) between two points i ,sc j The calculation formula for ) is as follows:

[0040]

[0041] Where, x A and y A They represent points pc respectively i x and y coordinates B and y B They represent points sc respectively j The x and y coordinates are given by i = 1, 2, ..., m and j = 1, 2, ..., n, where i and j represent the positions of the data points and m and n represent the number of data points.

[0042] In some embodiments, in step S6, from s = N C / T begins, followed by N C / T is the increment, until s = N C End the iteration and repeat steps S4 and S5;

[0043] Define the rate of change matrix as S, the Frechet distance matrix as FD, and the iteration number as count, with an initial value of 1.

[0044] In each iteration step, the Frechet distance fd corresponding to different s values ​​is recorded, and the count value is incremented by 1; this is represented as:

[0045]

[0046] In some embodiments, in step S7, the iteration starts from count = 1 and continues until count = T, with an increment of 1.

[0047] The iteration stops when the following condition is met, and the count value at this time is recorded as b; denoted as:

[0048] abs(fd count -fd C) ≤ 0.5;

[0049] fd C is the last value of the Frechet distance matrix FD;

[0050] Define the optimal change rate as K Best , K Best = S b .

[0051] In some embodiments, in step S8, initialize s = K Best Execute step S4; after execution, obtain the smoothed assembly process data SA and ST at the optimal change rate.

[0052] In some embodiments, in step S9, define the cleaning data curve as WC = (WA, WT)', where WA is the sequence of cleaned angle data and WT is the sequence of cleaned torque data; expressed as:

[0053]

[0054] where j represents the data position and n represents the number of data points;

[0055] Starting from i = 1 until i = m, iterate with an increment of 1 to clean the duplicate data, including:

[0056] Initialize j = 0; if sa i - sa i-1 = 0, then skip this loop;

[0057] If i < m and sa i+1 - sa i ≠ 0, then set wa j = sa i , wt j = st i , and the value of j automatically increases by 1;

[0058] If i = m and satisfies sa i - sa i-1 ≠ 0, then set wa j+1 = sa i , wt j+1 = st i ;

[0059] After the loop execution is completed, obtain the cleaned data.

[0060] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0061] This invention precisely filters out extra data generated during multi-step tightening processes due to tool withdrawal, friction overcoming, and other factors, while fully preserving the key information of the original data. It not only solves the data filtering problem in existing technologies but also significantly improves the precision, accuracy, and reliability of assembly quality analysis. This is of great significance for improving the efficiency and safety of pipeline assembly and provides a foundation for data-driven assembly quality control in the aerospace manufacturing industry.

[0062] Compared to traditional global filtering methods, the method of this invention can accurately retain original data directly related to assembly quality, avoiding the loss of critical information and ensuring data integrity and authenticity. This is of great value in discovering potential assembly quality problems. Through precise data analysis, operators can quickly identify anomalies in the assembly process, adjust assembly strategies in a timely manner, effectively prevent assembly errors, and significantly improve assembly efficiency and safety, thus having a direct positive impact on flight safety and performance.

[0063] The method of this invention provides a targeted solution to the special needs of the pipeline assembly field, especially the specific problems arising in multi-step tightening operations. It is more in line with the actual needs of the aerospace manufacturing industry and has higher applicability and practicality. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart of the data filtering and processing method for pipeline assembly in an embodiment of the present invention.

[0066] Figure 2 This is a graph of the original data obtained in an embodiment of the present invention.

[0067] Figure 3 This is a data curve obtained by correcting the horizontal axis in an embodiment of the present invention.

[0068] Figure 4 This is a schematic diagram illustrating the relationship for determining the optimal rate of change in an embodiment of the present invention.

[0069] Figure 5 This is a data curve obtained from the smooth assembly process data in an embodiment of the present invention.

[0070] Figure 6 This is a data curve obtained by cleaning up duplicate data in an embodiment of the present invention. Detailed Implementation

[0071] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0072] The pipeline assembly process data filtering method of the present invention is used to filter out extra irrelevant data generated during multi-step tightening, so as to solve the problem of irrelevant data affecting assembly quality analysis, improve the accuracy and reliability of assembly quality analysis, and meet the actual use needs in the field of pipeline assembly.

[0073] In some embodiments of the present invention, the data filtering method for the pipeline assembly process is described with reference to... Figure 1 This includes the following steps:

[0074] Step S1: Obtain the original assembly process data

[0075] Define the original angle data sequence as PA and the original torque data sequence as PT; expressed as:

[0076] PA = (pa1, pa2, ..., pa) m );

[0077] PT = (pt1, pt2, ..., pt) m );

[0078] Wherein, the i-th data point of sequence PA is denoted as pa i The i-th data point of sequence PT is denoted as pt. i Here, i represents the position of the data in the sequence; m represents the number of data points. The meanings of i and m remain the same thereafter.

[0079] Step S2: Correct the horizontal coordinate of the rollback

[0080] Define the correction angle data sequence as CA and the correction torque data sequence as CT.

[0081] CA = (ca1, ca2, ..., cam );

[0082] CT = (ct1, ct2, ..., ct m );

[0083] Define ca1 = pa1 and ct1 = pt1. Starting from the second data point i = 2 and ending at i = m, iterate through the original angle data sequence PA.

[0084] If pa i ≥pa i-1 If the sequence value is in a non-decreasing state, the process data is preserved with fidelity using formula 1).

[0085]

[0086] If pa i <pa i-1 If the sequence is in a backward state, the backward data is corrected using formula 2) to make it a non-decreasing state.

[0087]

[0088] After calibration, all have ca i ≥ca i-1 The corrected sequence CA is in a non-decreasing state.

[0089] Step S3: Determine the iteration parameters

[0090] Define the position of the half-range data point as h. If m is odd, then h = (m-1) / 2 + 1; if m is even, then h = m / 2.

[0091] Calculate the half-range change rate K of the correction angle data sequence CA and the torque data sequence for CT. H ; indicates as:

[0092] ct h ca h These are the data corresponding to the half-range data points in CA and CT, respectively.

[0093] Define the amplification factor as U, then the rate of change of the upper limit of the loop is N. C =UK H .

[0094] Define the number of iterations as T, which is a positive integer.

[0095] Step S4: Filter assembly process data

[0096] Starting from the second data point i=2 and ending at i=m, calculate the rate of change K at the current position. P ; indicates as:

[0097]

[0098] Define the filter angle data sequence as SA, the filter torque data sequence as ST, the critical position change rate as s, and the initial value of s as 0.

[0099] Filter assembly process data according to the following criteria:

[0100] If 0≤K P ≤s, process data is preserved with fidelity using formula 3);

[0101]

[0102] If K P <0 or K P >s, filter the process data using formula 4);

[0103]

[0104] Step S5: Calculate the Frechet distance between the filtered data and the original data.

[0105] Define the original data curve as PC=(PA,PT)', which is represented as:

[0106]

[0107] Define the filtered data curve as SC=(SA,ST)', which is expressed as:

[0108]

[0109] Define PC and SC as two curves located on the metric space (S, d), where d is the distance function between two points in X. The starting and ending points of PC are in the interval [a, b], and the starting and ending points of SC are in the interval [c, d].

[0110] Define the Frechet distance between PC and SC as FD, and calculate the value of FD according to formula 5):

[0111] FD = inf α,β max t∈[0,1] {d(PC(α(t)),SC(β(t)))}……5);

[0112] α is a continuous non-decreasing function mapping [0,1] to [a,b], and β is a continuous non-decreasing function mapping [0,1] to [c,d], representing the "walking" speed along curves PC and SC, respectively. α(t) and β(t) represent the positions along curves PC and SC at time t, respectively. Inf denotes the infimum, which is the minimum value of the largest d(PC(α(t)), SC(β(t))) among all possible pairs of α and β.

[0113] For any point pc on curve PC i and any point sc on curve SC j The Euclidean distance d(pc) between two points i ,sc j The calculation formula for ) is as follows:

[0114]

[0115] Where, x A and y A They represent points pc respectively i x and y coordinates B and y B They represent points sc respectively j The x and y coordinates are given by i = 1, 2, ..., m and j = 1, 2, ..., n. i and j represent the positions of the data points, and m and n represent the number of data points.

[0116] Step S6: Iteratively calculate the rate of change

[0117] From s=N C / T begins, followed by N C / T is the increment, until s = N C End the iteration and repeat steps S4 and S5.

[0118] Define the rate of change matrix as S, the Frechet distance matrix as FD, and the iteration number as count, with an initial value of 1.

[0119] After the iteration is complete, there are a total of T elements in S and FD.

[0120] In each iteration step, record the Frechet distance fd corresponding to different s values. count At the same time, the value of count is incremented by 1; this is represented as:

[0121]

[0122] Step S7: Determine the optimal rate of change

[0123] The iteration starts from count = 1 and continues until count = A, with an increment of 1.

[0124] The iteration stops when the following condition is met, and the count value at this time is recorded as b; denoted as:

[0125] abs(fd count -fd C )≤0.5.

[0126] fd C This is the last value of the Frechet distance matrix FD.

[0127] Reference Figure 4 The sequence FD will gradually converge to a straight line. Therefore, when the distance fd count with fd C When the distance is less than 0.5, it is considered that the convergence position has been reached, and its subscript is b.

[0128] Define the optimal rate of change as K Best Its value is K Best =S b .

[0129] There is a one-to-one correspondence between sequence S and sequence FD. The x-axis corresponding to the optimal convergence position is the x-axis of the optimal rate of change, and its value is S. b .

[0130] Define the rate of change matrix as S count Define the Frechet distance matrix as FD count That is, the rate of change is S. count When the calculated Frechet distance is FD count .

[0131] Step S8: Smooth assembly process data

[0132] Initialize s = K Best Perform step S4. The data obtained after execution are the smoothed assembly process data SA and ST under the optimal rate of change.

[0133] Step S9: Clean up duplicate data

[0134] After smoothing the assembly process data, many duplicate data points exist in SA and ST, therefore, data cleaning is required. The data cleaning curve is defined as WC = (WA, WT)'; expressed as:

[0135]

[0136] Starting from i=1 and ending at i=m, iterate in increments of 1. Perform filtering according to the following steps:

[0137] Initialize j = 0; if sa i -sai-1 If it is equal to 0, skip this loop;

[0138] If i < m, and sa i+1 - sa i ≠ 0, then let wa j = a i , wt j = st i , and the value of j is automatically incremented by 1;

[0139] If i = m, and sa i - sa i-1 ≠ 0, then let wa j+1 = sa i , wt j+1 = st i ;

[0140] After the loop execution is completed, the cleaned data can be obtained. The cleaned angular data sequence is WA, and the cleaned torque data sequence is WT, (j = 1, 2,..., n), where j represents the data position and n represents the number of data points.

[0141] The method of the present invention can accurately identify and remove irrelevant data generated during the multi-step tightening process, such as noise caused by factors such as tool withdrawal and overcoming friction, while completely retaining the key information in the original data directly related to the assembly quality, ensuring the purity and integrity of the data, and improving the accuracy of the assembly quality analysis.

[0142] Aiming at the special problems generated during the multi-step tightening operation in the pipeline assembly field, such as multiple operations caused by space limitations, the present invention adopts a precise data filtering strategy, effectively solving the limitation of the existing global filtering method that cannot filter irrelevant data at a fixed point, avoiding the loss of key information, and significantly improving the accuracy and reliability of the assembly quality analysis.

[0143] Different from the data filtering methods widely used in the Internet industry in the prior art, the method of the present invention is specifically aimed at the special needs of the pipeline assembly field, not only solving the technical problems in a specific field, but also improving the pertinence and practicality of data processing, and better meeting the needs of the aviation manufacturing industry for high-quality assembly data.

[0144] By introducing the fidelity filtering method, the processing flow of the pipeline assembly process data is optimized, the efficiency and accuracy of the assembly quality analysis are improved, providing a more efficient and accurate assembly quality control means for the aviation manufacturing industry, which has important value for improving the dynamic performance safety and performance.

[0145] The following describes in detail the method for filtering and processing the pipeline assembly process data of the present invention in combination with embodiments.

[0146] Using the assembly process data of a hydraulic conduit with an outer diameter of 16mm as input, the data filtering and processing method for the pipeline assembly process adopted in this invention is described; including the following steps:

[0147] Step S1: Read the raw data

[0148] Define the original angle data sequence as PA and the original torque data sequence as PT. A line graph is plotted with PA as the x-axis and PT as the y-axis, as shown below. Figure 2 As shown.

[0149] PA = (pa1, pa2, ..., pa) 346 = (0.1, 0.1, ..., 67.3);

[0150] PT = (pt1, pt2, ..., pt) 346 = (1.357, 1.266, ..., 39.250);

[0151] Where i represents the data location, and m = 346 represents the number of data points.

[0152] Step S2: Correct the horizontal coordinate of the rollback

[0153] Define the correction angle data sequence as CA and the correction torque data sequence as CT.

[0154] Define ca1 = pa1 and ct1 = pt1. Starting from the second data point i = 2 and ending at i = 346, iterate through the sequence PA.

[0155] If pa i ≥pa i-1 If the sequence value is in a non-decreasing state, the process data can be preserved with high fidelity using the following formula.

[0156]

[0157] If pa i <pa i-1 If the sequence is in a backward state, the backward data can be corrected using the following formula to make it a non-decreasing state.

[0158]

[0159] After calibration, all have ca i ≥ca i-1 The corrected sequence CA is in a non-decreasing state. A line graph is plotted with the corrected CA on the x-axis and the corrected CT on the y-axis as shown below. Figure 3 As shown.

[0160] Step S3: Determine the iteration parameters

[0161] Since m = 346, the position of the half-range data point is h = m / 2 = 173. Calculate the half-range change rate K of the correction angle data sequence CA and the torque data sequence for CT. H .

[0162]

[0163] If the magnification factor U = 3, then the rate of change of the upper limit of the loop is N. C =3K H =7.8705. Define the number of iterations T = 50.

[0164] Step S4: Filter assembly process data

[0165] Starting from the second data point i=2 and ending at i=346, calculate the rate of change K at the current position. P .

[0166]

[0167] Define the filter angle data sequence as SA, the filter torque data sequence as ST, and the critical position change rate as s = 0. Filter the assembly process data using the following criteria.

[0168] If 0≤K P ≤s, the process data is preserved with fidelity using the following formula.

[0169]

[0170] If K P <0 or K P If the value is greater than 0, filter the process data using the following formula.

[0171]

[0172] Step S5: Calculate the Frechet distance between the filtered data and the original data.

[0173] Define the original data curve as PC = (PA, PT)':

[0174]

[0175] Define the filtered data curve as SC=(SA,ST)':

[0176]

[0177] Define PC and SC as two curves located on the metric space (S, d), where d is the distance function between two points in X. The starting and ending points of PC are in the interval [a, b], and the starting and ending points of SC are in the interval [c, d].

[0178] Define the Frechet distance between PC and SC as FD, and calculate the value of FD according to formula 5):

[0179] FD = inf α,β max t∈[0,1] {d(PC(α(t)),SC(β(t)))}……5);

[0180] α is a continuous non-decreasing function mapping [0,1] to [a,b], and β is a continuous non-decreasing function mapping [0,1] to [c,d], representing the "walking" speed along curves PC and SC, respectively. α(t) and β(t) represent the positions along curves PC and SC at time t, respectively. Inf denotes the infimum, which is the minimum value of the largest d(PC(α(t)), SC(β(t))) among all possible pairs of α and β.

[0181] For any point pc on curve PC i and any point sc on curve SC j The Euclidean distance d(pc) between two points i ,sc j The calculation formula for ) is as follows:

[0182]

[0183] Where, x A and y A They represent points pc respectively i x and y coordinates B and y B They represent points sc respectively j The x and y coordinates are given by i = 1, 2, ..., m and j = 1, 2, ..., n. i and j represent the positions of the data points, and m and n represent the number of data points.

[0184] The calculated FD = 97.5028.

[0185] Step S6: Iteratively calculate the rate of change

[0186] From s=N C Starting with / T=0.1574, with N C The increment is / T = 0.1574, and the iteration continues until s = 7.8705. Steps S4 and S5 are repeated.

[0187] Define the rate of change matrix as S, the Frechet distance matrix as FD, and the iteration number as count, with an initial value of 1. After iteration, S and FD will contain a total of 50 elements.

[0188] In each iteration step, the Frechet distance fd corresponding to different s values ​​is recorded, and the count value is incremented by 1; this is represented as:

[0189]

[0190] Step S7: Determine the optimal rate of change

[0191] The iteration starts from count = 1 and continues until count = 50, incrementing by 1. Iteration stops when the following condition is met, and the count value at this point is recorded as 20.

[0192] abs(fd count -fd C ≤0.5;

[0193] In this embodiment, due to fd 20 -fd 50 ≤0.5, therefore the value of b is 20;

[0194] Define the optimal rate of change as K Best Its value is K Best =S 20 =3.1482. The iteration result is shown in the figure below. Figure 4 As shown, the curve converges when b = 20.

[0195] Step S8: Smooth assembly process data

[0196] Initialize s = K Best =3.1482 Execute step 4. The data after execution is the assembly process data sa after smoothing under the optimal rate of change. i and st i (i = 1, 2, ..., 346). Plot a line graph with SA as the x-axis and ST as the y-axis, as shown below. Figure 5 As shown.

[0197]

[0198] Step S9: Clean up duplicate data

[0199] After smoothing the assembly process data, many duplicate data points exist in SA and ST, so it is necessary to clean up the duplicate data. The cleaned data curve is defined as WC=(WA,WT)'.

[0200] Starting from i=1 and ending at i=346, iterate in increments of 1. Filter according to the following calculation method.

[0201] Initialize j = 0. If sa i -sa i-1 If the value is 0, skip this iteration.

[0202] If i<346, and sa i+1 -sa i ≠0, then let wa j =sa i wt j =st i Meanwhile, the value of j increments by 1.

[0203] In particular, if i = 346, and satisfies sa i -sa i-1 ≠0, then let wa j+1 =sa i wt j+1 =st i .

[0204] After the loop completes, the cleaned data is obtained. The cleaned angle data sequence is WA, and the torque data sequence is WT (j = 1, 2, ..., 103), where j represents the data position, and the number of data points is n = 103. A line graph is plotted with WA as the x-axis and WT as the y-axis, as shown below. Figure 6 As shown, its shape is completely consistent with the smoothed assembly process data, only the data length is different.

[0205]

[0206] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for filtering and processing data during pipeline assembly, characterized in that, Includes the following steps: S1. Obtain the raw data of the pipeline assembly process and form the raw angle data sequence and raw torque data sequence; S2. Using the original angle data as the x-axis, perform fidelity and correction processing on the original angle data, and correct the x-axis of the backtracking. S3. Set the iteration parameters, including half-range change rate, amplification factor, and number of iterations; S4. Filter assembly process data, including formulating filtering criteria based on the critical position change rate, and filtering the assembly process data according to the filtering criteria to obtain filtered data. S5. Obtain the Frechet distance between the filtered data and the original data; S6. Using the set iteration parameters, iteratively calculate the rate of change; S7. Obtain the optimal rate of change; S8. Smooth the assembly process data; S9. Clean up duplicate data; In step S1, the original angle data sequence is defined as PA, and the original torque data sequence is defined as PT; expressed as: ； ； Where m represents the number of data points; In step S2, the correction angle data sequence is defined as CA, and the correction torque data sequence is defined as CT; where CA = (ca1, ca2, ..., ca m ), CT=(ct1,ct2,…,ct m ), where m represents the number of data points; Define ca1 = pa1 and ct1 = pt1. Starting from the second data point i = 2 and ending at i = m, iterate through the original angle data sequence PA. If pa i ≥pa i-1 The process data is preserved with high fidelity using formula 1). ……1）； If pa i <pa i-1 The rollback data is corrected using formula 2). ……2）； In step S4, starting from the second data point i = 2 and ending at i = m, calculate the rate of change K at the current position. P ; indicates as: ； Define the filter angle data sequence as SA, the filter torque data sequence as ST, the critical position change rate as s, and the initial value of s as 0; If 0≤K P ≤s, process data is preserved with fidelity using formula 3); ……3）； If K P <0 or K P >s, filter the process data using formula 4); ……4）。 2. The data filtering and processing method for pipeline assembly process according to claim 1, characterized in that, In step S3, the half-range data point positions are defined, and the correction angle data sequence CA and torque data sequence are calculated as the half-range change rate K of the CT. H ; indicates as: ;ct h ca h These are the data corresponding to the half-range data points in CA and CT, respectively. Define the amplification factor as U, and the rate of change of the upper limit of the loop as N. C = UK H Let T be the number of iterations.

3. The data filtering and processing method for pipeline assembly process according to claim 1, characterized in that, In step S5, the original data curve is defined as... , is represented as: ; Define the filtered data curve as , is represented as: ; Both PC and SC are defined as residing in the metric space. S , d Two curves on X, where d is the distance function between two points in X, the starting and ending intervals of PC are [a,b], and the starting and ending intervals of SC are [c,d]. Define the Frechet distance between PC and SC as FD, and calculate the value of FD according to formula 5): ……5）； α is a continuous non-decreasing function mapping [0,1] to [a,b], and β is a continuous non-decreasing function mapping [0,1] to [c,d], representing the "walking" speed along curves PC and SC, respectively; α(t) and β(t) represent the positions along curves PC and SC at time t, respectively; Inf represents the infimum, which is the minimum value of the largest d(PC(α(t)), SC(β(t))) among all possible pairs of α and β; For any point pc on curve PC i and any point sc on curve SC j Euclidean distance between two points d (pc i ,sc j The calculation formula for ) is as follows: ……6）； in, x A and y A They represent points pc respectively i x and y coordinates x B and y B They represent points sc respectively j The x and y coordinates are given by i = 1, 2, ..., m and j = 1, 2, ..., n, where i and j represent the positions of the data points and m and n represent the number of data points.

4. The data filtering and processing method for pipeline assembly process according to claim 3, characterized in that, In step S6, from s=N C / T begins, followed by N C / T is the increment, until s = N C End the iteration and repeat steps S4 and S5; Define the rate of change matrix as S, the Frechet distance matrix as FD, and the iteration number as count, with an initial value of 1. In each iteration step, the Frechet distance fd corresponding to different s values ​​is recorded, and the count value is incremented by 1; this is represented as: 。 5. The data filtering and processing method for pipeline assembly process according to claim 4, characterized in that, In step S7, the iteration starts from count = 1 and continues until count = T, with an increment of 1. The iteration stops when the following condition is met, and the count value at this time is recorded as b; denoted as: ； fd C This is the last value of the Frechet distance matrix FD; Define the optimal rate of change as K Best K Best =S b .

6. The data filtering and processing method for pipeline assembly process according to claim 5, characterized in that, In step S8, initialize s = K Best Execute step S4; after execution, obtain the smoothed assembly process data SA and ST under the optimal rate of change.

7. The data filtering and processing method for pipeline assembly process according to claim 6, characterized in that, In step S9, the cleaning data curve is defined as follows: WA is the cleaned angle data sequence, and WT is the cleaned torque data sequence; represented as: ； Where j represents the data location and n represents the number of data points; Starting from i = 1 and ending at i = m, the iteration proceeds in increments of 1, cleaning up duplicate data, including: Initialize j = 0; if sa i -sa i-1 If the value is 0, then skip this iteration. If i < m and sa i+1 -sa i ≠ 0, then let wa j = sa i , wt j = st i , and at the same time, the value of j is automatically incremented by 1; If i = m, and satisfies sa i -sa i-1 ≠ 0, then let wa j+1 = sa i wt j+1 = st i ; After the loop completes, the cleaned data is obtained.

Images