Printing quality evaluation method based on rotary screen digital printing machine
Patent Information
- Application Number
- CN202610768413.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-30
- Publication Date
- 2026-10-16
AI Technical Summary
[0003]印花质量评价在现有生产中多依赖布面检验记录、人工复核结果或单一设备参数的越限判断,评价重点通常停留在成品缺陷识别层面,难以分析缺陷在圆网版位、生产时段和传感器运行链路中的形成过程
(1)该方法通过读取待评价印花布的布面检验记录,依据成品质量标记的起始米位和成品质量标记的结束米位划分待评价布面区域,并进一步形成评价段、评价段编号、生产时段、圆网版位区间和成品质量标记,使印花质量评价具有明确的布面米位基础、生产时间基础和圆网版位基础。随后依据布速传感器、张力传感器、网筒编码器、刮刀压力传感器和烘箱温度传感器调取当前生产时段内的传感器运行数据,并对布速序列、张力序列、网筒相位序列、刮刀压力序列和烘箱温度序列进行版位截取和组合,获得评价段运行片段。相较于仅依据布面检验记录或人工经验进行质量判断的方式,本方法能够将合格标记、色差标记、重影标记、渗化标记和断纹标记与圆网仿数码印花机的实际运行状态结合分析,使成品质量现象具备生产过程追溯依据。
Smart Images

Figure CN122914653A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of textile printing inspection technology, specifically a method for evaluating printing quality based on a rotary screen digital printing machine. Background Technology
[0002] In the field of textile printing inspection technology, the quality of printed fabric is typically judged based on factors such as color consistency, pattern registration, ghosting, bleeding, and broken lines. As printing equipment has evolved from conventional rotary screen printing to machines that combine continuous rotary screen printing with digital pattern control, rotary screen simulant digital printing machines are being used in multi-pattern, multi-color, and continuous production scenarios. Because these machines involve multiple operating factors such as fabric speed, tension, screen cylinder phase, squeegee pressure, and drying temperature, the finished fabric condition is often related to the plate position, sensor timing, and process transfer during production. Therefore, establishing a production-process-oriented printing quality evaluation method for rotary screen simulant digital printing machines has become an important research direction in printing quality inspection.
[0003] In current production, printing quality evaluation largely relies on fabric inspection records, manual verification results, or judgments based on exceeding limits of single equipment parameters. The evaluation focus is typically limited to identifying finished product defects, making it difficult to analyze the formation process of defects within the rotary screen printing plate, production time, and sensor operation chain. For rotary screen digital printing machines, the sampling rhythm, plate position benchmark, and process lag state of different sensors are inconsistent. Judging solely based on the time axis or single-point parameters makes it difficult to distinguish the sources of influence of speed fluctuations, tension changes, screen cylinder phase shifts, abnormal squeegee pressure, and oven temperature fluctuations in defect formation, and also makes it difficult to determine in which type of sensor chain the anomaly first appears. Therefore, existing printing quality evaluation methods suffer from insufficient process traceability, inadequate plate position analysis, difficulty in comparing the transmission states between sensors, and insufficient basis for classifying defect causes. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a printing quality evaluation method based on a rotary screen digital printing machine, thus solving the problems mentioned in the background technology.
[0005] To achieve the above objectives, the present invention provides the following technical solution: a printing quality evaluation method based on a rotary screen digital printing machine, comprising: S1. Read the fabric inspection record, divide the evaluation segment according to the fabric inspection record and mark the production period and the circular mesh position interval, retrieve the sensor operation data of the current production period, extract and combine the data according to the circular mesh position interval to obtain the evaluation segment operation segment. S2. Construct the circular grid position coordinates and convert the sensor operation data to the circular grid position coordinates. After data processing, obtain the standard position sequence group. Select any two standard position sequences to construct sensor sequence pairs and perform dynamic time warping to obtain path codes. S3. Perform position displacement on the sensor sequence pair and calculate the cross-covariance value within the overlapping position range. Determine the position hysteresis and position hysteresis direction based on the candidate position displacement with the largest absolute value of cross-covariance, and generate the sensor transmission edge by combining the path coding. S4. Construct indirect and direct transmission paths based on the direction and connection sequence of the sensor transmission edges, and obtain closed residual terms after consistency comparison. S5. Replace the candidate transfer edge and compare the closed residual terms before and after the replacement to obtain the residual starting edge and residual migration sequence. Then, combine the finished product quality mark in the fabric inspection record to generate the residual cause type and the verification object.
[0006] Preferably, S1 includes S11; S11. Read the fabric inspection record of the printed fabric to be evaluated. The fabric inspection record includes the fabric roll number, fabric meter position, finished product quality mark, starting meter position of the finished product quality mark, and ending meter position of the finished product quality mark. The fabric area to be evaluated is determined based on the starting and ending meter positions. The fabric area to be evaluated is divided into several evaluation segments based on the length of the fabric area to be evaluated. Each evaluation segment is marked with evaluation segment information, which includes the evaluation segment number, production period, cylinder screen plate position interval, and finished product quality mark. Finished product quality markings include qualified markings, color difference markings, ghosting markings, diffusion markings, and crack markings.
[0007] Preferably, S1 further includes S12; S12. Based on the sensor group installed on the rotary digital printing machine, retrieve the sensor operation data during the current production period from the equipment control system of the rotary digital printing machine, and perform plate position interception on the sensor operation data according to the rotary screen plate position interval to obtain the evaluation segment operation segment. The sensor group includes a fabric speed sensor, a tension sensor, a mesh cylinder encoder, a scraper pressure sensor, and an oven temperature sensor. The sensor operating data includes fabric speed sequence, tension sequence, mesh cylinder phase sequence, scraper pressure sequence, and oven temperature sequence. The plate position extraction process is as follows: When the rotary screen digital printing machine is being mounted or calibrated, the reference position of the rotary screen reference point in the cylinder encoder is obtained, and the circumference of the rotary screen of the rotary screen digital printing machine is divided into several rotary screen positions. During the production period when the evaluation segment passes through the printing station, the operating data of the cylinder encoder is read. The position of the cylinder encoder at the beginning sampling time of the current production period is assigned to the starting rotary screen position, and the position of the cylinder encoder at the end sampling time of the current production period is assigned to the ending rotary screen position to obtain the rotary screen position interval. Then, the fabric speed sequence, tension sequence, cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence located between the starting and ending rotary screen positions are extracted and combined to form the evaluation segment operation segment.
[0008] Preferably, S2 includes S21 and S22; S21. Based on the sampling order of the phase values of the mesh encoder, assign a circular mesh circumferential position identifier to each sampling point; when the phase value of the mesh encoder returns from the maximum phase to the initial phase, it is recorded as the completion of a circular mesh circumferential cycle, and a new cycle number is generated. Combine the cycle number with the circular mesh circumferential position identifier in the current cycle to form the circular mesh plate position coordinates. Then, using the mesh phase sequence in the evaluation segment as the plate position reference, convert the fabric speed sequence, tension sequence, scraper pressure sequence, and oven temperature sequence to the circular mesh plate position coordinates to obtain the fabric speed plate position sequence, tension plate position sequence, scraper pressure plate position sequence, and oven temperature plate position sequence, and perform data processing to obtain a standard plate position sequence group. The data processing includes median centering and median absolute deviation scaling. The standard printing plate sequence group includes the standard printing plate sequence for fabric speed, tension, scraper pressure, and oven temperature; S22. Select any two standard plate position sequences from the standard plate position sequence group to form a sensor sequence pair. Perform dynamic time warping on each sensor sequence pair to obtain a warped path. Encode each path step in the warped path to form a path coding item. The path coding item includes a synchronization code, a first sequence retention code, and a second sequence retention code. All path coding items for the same sensor sequence pair are arranged in a regular path order to form a path code.
[0009] Preferably, S3 includes S31 and S32; S31. Based on the front and rear coordinates of the circular grid plate position coordinates, set the standard plate position sequence in the sensor sequence pair as the first standard plate position sequence and the second standard plate position sequence respectively, and perform plate position displacement. Wherein, the plate position displacement is the number of sampling points that the second standard plate position sequence moves relative to the first standard plate position sequence along the circular grid plate position coordinates; The portion of the first standard plate position sequence and the second standard plate position sequence that simultaneously have sampled values in the circular grid plate position coordinates is taken as the overlapping plate position range. Within the overlapping plate position range, the average sampled values of the first standard plate position sequence and the average sampled values of the second standard plate position sequence after displacement are calculated respectively. The deviation of the two standard plate position sequences at the same plate position sampling point is calculated based on their respective average sampled values. The two deviations at the same plate position sampling point are multiplied together, and the average of all products within the overlapping plate position range is calculated to obtain the cross-covariance value. S32. Select the candidate plate position displacement with the largest absolute value of cross-covariance as the plate position hysteresis of the current sensor sequence pair, and generate the plate position hysteresis direction based on the plate position hysteresis. When the page position lag is positive, the first standard page position sequence is the leading sequence and the second standard page position sequence is the passive sequence. At this time, the page position lag direction is marked as the first leading direction. When the page position lag is negative, the second standard page position sequence is the leading sequence and the first standard page position sequence is the passive sequence. At this time, the page position lag direction is the second leading direction. When the page position lag is zero, the first standard page position sequence and the second standard page position sequence are in the same position. At this time, the page position lag direction is marked as the same position direction.
[0010] Preferably, S3 also includes S33; S33. Combine the path encoding and plate lag direction of the same sensor sequence pair; When the lagging direction of the plate position is the first leading direction, the sensor type of the first standard plate position sequence is used as the starting sensor type, and the sensor type of the second standard plate position sequence is used as the ending sensor type to form a sensor transfer edge. When the lagging direction of the plate position is the second leading direction, the sensor type to which the second standard plate position sequence belongs is used as the starting sensor type, and the sensor type to which the first standard plate position sequence belongs is used as the ending sensor type to form a sensor transfer edge. When the lag direction of the plate position is the same as the lag direction, no sensor transmission edge is formed.
[0011] Preferably, S4 includes S41; S41. Select three types of sensors located in the same circular grid plate position interval according to the direction and connection sequence of the sensor transmission edge, and set them as the first sensor type, the second sensor type and the third sensor type respectively; When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the second sensor type, the current sensor transfer edge is set as the first transfer edge. If there is a sensor transfer edge in the sensor transfer edge group that points from the second sensor type to the third sensor type, set the current sensor transfer edge as the second transfer edge. When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the third sensor type, set the current sensor transfer edge as the third transfer edge. Among them, the first and second passing edges constitute an indirect passing path, and the third passing edge constitutes a direct passing path.
[0012] Preferably, S4 also includes S42; S42. Extract the path codes of the first and second transmission edges and combine them according to the transmission direction to form an indirect path code. Extract the path code of the third transmission edge and record it as the direct path code. Then compare the indirect path code with the direct path code. When the indirect path code and the direct path code are consistent, the transmission closed unit is recorded as a closed consistent unit. Otherwise, it is recorded as a closed residual unit. The position of the circular grid plate where the path codes of the indirect transmission path and the direct transmission path first appear different in the closed residual unit is recorded as the residual plate position. The evaluation segment number, the circular grid plate position interval, the first sensor type, the second sensor type, the third sensor type, the indirect path code, the direct path code, and the residual plate position are combined to form a closed residual term. All closed residual terms are counted to obtain a closed residual set.
[0013] Preferably, S5 includes S51; S51. Extract the sensor transfer edges that participate in the closed residual term in the closed residual set, form a candidate transfer edge set, and perform an edge replacement process on each candidate transfer edge in the candidate transfer edge set. Edge replacement processing: Replace the current candidate passing edge of the segment to be evaluated with the same passing edge of the evaluation segment with the qualified mark in the same production batch to form a replacement spectrum. Iterate through S4 according to the replacement spectrum to obtain replacement residual terms. Statistically count all replacement residual terms to obtain the replacement residual set. Match the closed residual set with the replacement residual set item by item: the closed residual items that exist in the original closed residual set but no longer exist in the replacement residual set after the candidate passing edge is replaced are recorded as the eliminated residual items; After all candidate passing edges have been replaced, the candidate passing edge with the largest number of eliminated residual terms is selected as the residual starting edge. When there are two or more candidate passing edges with the same number of eliminated residual terms, the candidate passing edge with the first position in the circle screen is selected as the residual starting edge. Starting from the residual starting edge, search for closed residual items in the closed residual set that share the same endpoint sensor type or starting sensor type as the residual starting edge, and arrange them according to the residual plate position to form a residual migration sequence.
[0014] Preferably, S5 also includes S52; S52. Generate residual cause types based on finished product quality marks, residual starting edges, and residual migration sequences; The residual starting edge includes a starting position sequence type and an ending position sequence type; the residual migration sequence includes at least one residual propagation edge, and each residual propagation edge includes a starting position sequence type and an ending position sequence type. When the finished product quality is marked as a ghost mark, and the starting point plate position sequence type of the residual starting edge is a mesh cylinder phase sequence, the residual cause type is recorded as phase closure residual type. At this time, the verification objects are determined as the circular mesh synchronous control record and the mesh cylinder encoder record. When the finished product quality is marked as a broken mark, and the starting point plate position sequence type or the ending point plate position sequence type of the residual starting edge is a scraper pressure plate position sequence, the residual cause type is recorded as imprint closure residual type, and the object to be checked is determined as scraper pressure control record. When the finished product quality is marked as a diffusion mark, and there is a first residual transfer edge and a second residual transfer edge in the residual migration sequence, wherein the starting plate position sequence type or the ending plate position sequence type of the first residual transfer edge is a doctor blade pressure plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the second residual transfer edge is an oven temperature plate position sequence, the residual cause type is recorded as the printing heat treatment residual type. At this time, the verification object is determined to be the doctor blade pressure control record and the oven temperature control record. When the finished product quality is marked as color difference, and there are a third residual transfer edge and a fourth residual transfer edge in the residual migration sequence, where the starting plate position sequence type or the ending plate position sequence type of the third residual transfer edge is the cloth speed plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the fourth residual transfer edge is the squeegee pressure plate position sequence, the residual cause type is recorded as cloth printing residual type. At this time, the verification objects are determined as cloth speed control records and squeegee pressure control records. When the residual migration sequence involves more than three types of plate position sequence, the residual cause type is recorded as multi-source closed residual type. At this time, the verification object is determined to be the control records of all sensor types in the residual migration sequence.
[0015] This invention provides a method for evaluating printing quality based on a rotary screen digital printing machine. It has the following beneficial effects: (1) This method reads the fabric inspection record of the printed fabric to be evaluated, divides the fabric area to be evaluated according to the starting meter position and the ending meter position of the finished product quality mark, and further forms the evaluation segment, evaluation segment number, production time period, rotary screen plate position interval and finished product quality mark, so that the printing quality evaluation has a clear basis of fabric meter position, production time and rotary screen plate position. Then, the sensor operation data of the current production time period is retrieved according to the fabric speed sensor, tension sensor, screen cylinder encoder, squeegee pressure sensor and oven temperature sensor, and the plate position is extracted and combined for the fabric speed sequence, tension sequence, screen cylinder phase sequence, squeegee pressure sequence and oven temperature sequence to obtain the evaluation segment operation segment. Compared with the method of judging quality based solely on fabric inspection record or manual experience, this method can combine the qualified mark, color difference mark, ghost mark, bleeding mark and broken line mark with the actual operating status of the rotary screen digital printing machine for analysis, so that the finished product quality phenomenon has a production process traceability basis.
[0016] (2) This method uses the phase sequence of the mesh cylinder in the evaluation segment as the plate position reference to generate the circular mesh plate position coordinates. The cloth speed sequence, tension sequence, scraper pressure sequence, and oven temperature sequence are converted into cloth speed plate position sequence, tension plate position sequence, scraper pressure plate position sequence, and oven temperature plate position sequence. After median centering and median absolute deviation scaling, the standard plate position sequences of cloth speed, tension, scraper pressure, and oven temperature are obtained. By selecting sensor sequence pairs from the standard plate position sequence group and performing dynamic time warping on the sensor sequence pairs, the warped path and path code can be obtained. Then, the sensor transmission edge is formed by plate position displacement, cross-covariance value within the overlapping plate position range, plate position hysteresis, and plate position hysteresis direction. Compared to current techniques that only observe fluctuations in individual equipment parameters along a time axis, this method can analyze the leading, driven, and co-located states between the standard printing speed sequence, tension sequence, squeegee pressure sequence, and oven temperature sequence in the rotary screen printing coordinate system. This transforms the evaluation of printing quality from judging single-point parameters to judging the sensor operation chain.
[0017] (3) This method constructs indirect and direct transmission paths based on the connection order of sensor transmission edges within the same circular mesh printing plate interval. It obtains closed residual terms by comparing the consistency of indirect path codes and direct path codes, and then extracts the sensor transmission edges participating in the closed residual terms to form a candidate transmission edge set. Through edge replacement processing, replacement spectrum, replacement residual terms, and replacement residual set of candidate transmission edges, the residual starting edge and residual migration sequence can be determined. Combined with the finished product quality marks in the fabric inspection record, phase closed residual type, embossing closed residual type, embossing heat treatment residual type, fabric walking embossing residual type, or multi-source closed residual type are generated. At the same time, the verification objects such as the circular mesh synchronous control record, the mesh cylinder encoder record, the doctor blade pressure control record, the oven temperature control record, and the fabric speed control record are determined. Compared to the current method of discovering color difference, ghosting, bleeding, or broken lines and then relying on experience to deduce the source of the fault, this method can incorporate fabric abnormalities, rotary screen printing plate position range, path coding, closed residual terms, residual starting edge, residual migration sequence, and verification objects into the same evaluation process. This makes the scope of equipment troubleshooting clearer, the basis for process adjustment more specific, and the printing quality evaluation results of the rotary screen imitation digital printing machine in continuous production more convenient for subsequent production management. Attached Figure Description
[0018] Figure 1 This is a schematic diagram of the printing quality evaluation method based on a rotary screen digital printing machine according to the present invention. Figure 2 This is a logic block diagram for evaluating the printing quality of the present invention based on a rotary screen digital printing machine. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. 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.
[0020] Example 1 Please see Figure 1 This invention provides a printing quality evaluation method based on a rotary screen digital printing machine. To achieve the above objectives, this invention is implemented through the following technical solution: including: S1. Read the fabric inspection record, divide the evaluation segment according to the fabric inspection record and mark the production period and the circular mesh position interval, retrieve the sensor operation data of the current production period, extract and combine the data according to the circular mesh position interval to obtain the evaluation segment operation segment. S2. Construct the circular grid position coordinates and convert the sensor operation data to the circular grid position coordinates. After data processing, obtain the standard position sequence group. Select any two standard position sequences to construct sensor sequence pairs and perform dynamic time warping to obtain path codes. S3. Perform position displacement on the sensor sequence pair and calculate the cross-covariance value within the overlapping position range. Determine the position hysteresis and position hysteresis direction based on the candidate position displacement with the largest absolute value of cross-covariance, and generate the sensor transmission edge by combining the path coding. S4. Construct indirect and direct transmission paths based on the direction and connection sequence of the sensor transmission edges, and obtain closed residual terms after consistency comparison. S5. Replace the candidate transfer edge and compare the closed residual terms before and after the replacement to obtain the residual starting edge and residual migration sequence. Then, combine the finished product quality mark in the fabric inspection record to generate the residual cause type and the verification object.
[0021] In this embodiment, the analysis first focuses on the fabric inspection records. Evaluation segments are divided based on the fabric's meter position, finished product quality mark, the starting meter position of the finished product quality mark, and the ending meter position of the finished product quality mark. Each evaluation segment is then labeled with a production period and a circular screen printing interval. Subsequently, sensor operation data for the current production period is retrieved, and the data is extracted and combined according to the circular screen printing interval to obtain the evaluation segment's operation fragment. Therefore, the quality phenomena appearing on the fabric are no longer just isolated inspection conclusions, but can be analyzed within the specific production period and specific circular screen printing interval. For example, if a fabric roll is marked with a ghosting mark between 120 meters and 125 meters, this method can first divide the 120-meter to 125-meter section into one or more evaluation segments. Then, it determines the production period of the current evaluation segment passing through the printing station of the rotary screen digital printing machine, and identifies the rotary screen printing plate interval within the current production period. Next, it retrieves the sensor operating data for the current production period and extracts the fabric speed sequence, tension sequence, screen cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence within the current rotary screen printing plate interval, combining them to form the evaluation segment's operating fragment. Through this process, finished product quality markings, evaluation segment numbers, production periods, rotary screen printing plate intervals, and evaluation segment operating fragments are incorporated into the same evaluation process, solving the problem in existing technologies where it is difficult to clearly identify the time and range of quality anomalies when judging quality anomalies solely based on fabric inspection records.
[0022] After obtaining the evaluation segment's operational data, this method uses the mesh cylinder phase sequence within the segment as the positioning reference to generate circular mesh positioning coordinates. The sensor operational data is then converted to these circular mesh positioning coordinates, and further data processing is performed to obtain a set of standard positioning sequences. Next, any two standard positioning sequences are selected from this set to construct sensor sequence pairs. Dynamic time warping is then applied to these sensor sequence pairs to obtain path codes. Simultaneously, positioning displacement is performed on the sensor sequence pairs, and the cross-covariance value is calculated within the overlapping positioning range. Based on the candidate positioning displacement with the largest absolute value of the cross-covariance, the positioning hysteresis and its direction are determined. This is then combined with the path codes to generate the sensor transmission edge. For example, within the same evaluation segment, the fabric speed standard positioning sequence and the tension standard positioning sequence can be constructed as one sensor sequence pair, and the tension standard positioning sequence and the scraper pressure standard positioning sequence can be constructed as another sensor sequence pair. If the standard printing speed sequence first fluctuates on several rotary screen printing positions, and then the standard tension sequence fluctuates similarly on subsequent rotary screen printing positions, the printing position lag and direction relative to the standard tension sequence can be determined by the printing position displacement and cross-covariance value, forming a sensor transfer edge from the printing speed sensor type to the tension sensor type. If the standard tension sequence and the standard squeegee pressure sequence also exhibit a similar transfer state, a sensor transfer edge from the tension sensor type to the squeegee pressure sensor type can be further formed. Compared to existing technologies that simply observe whether the operating data of a single sensor on the time axis fluctuates, this method can analyze the sequential state, printing position differences, and transfer direction of different sensor operating data under the rotary screen printing position coordinates. This allows printing quality evaluation to move beyond judging a single sensor operating data point and instead incorporate a process chain analysis involving sensor sequence pairs, path encoding, printing position lag, printing position lag direction, and sensor transfer edges.
[0023] After forming the sensor transfer edges, this method continues to construct indirect and direct transfer paths based on the connection order of the sensor transfer edges within the same circular grid pattern interval. The indirect and direct transfer paths are then compared for consistency to obtain closed residual terms. Subsequently, the sensor transfer edges participating in the closed residual terms are extracted to form a candidate transfer edge set. These candidate transfer edges are replaced, and the residual terms before and after replacement are compared to obtain the residual starting edge and residual migration sequence. Finally, the residual cause type and verification object are generated by combining the finished product quality mark in the fabric inspection record. For example, within the same circular grid pattern interval, if there are sensor transfer edges pointing from the fabric speed sensor type to the tension sensor type, from the tension sensor type to the scraper pressure sensor type, and from the fabric speed sensor type to the scraper pressure sensor type, the first two sensor transfer edges can constitute an indirect transfer path, and the last sensor transfer edge can constitute a direct transfer path. If the indirect path code and the direct path code become inconsistent at a certain circular grid pattern position, the current inconsistent position is recorded as the residual pattern position, and a closed residual term is formed. Next, the current candidate transfer edge is replaced with the same-named transfer edge of the evaluation segment with a qualified mark in the same production batch. If the original closed residual term no longer exists after the replacement, it indicates that the replaced candidate transfer edge is more likely to be the residual starting edge. This is then combined with the finished product quality mark for judgment. For example, if the finished product quality mark is a crack mark and the residual starting edge involves the doctor blade pressure plate position sequence, an imprint closed residual type can be generated, and the verification object is determined to be the doctor blade pressure control record. Similarly, if the finished product quality mark is a diffusion mark and the residual migration sequence involves both the doctor blade pressure plate position sequence and the oven temperature plate position sequence, an imprint thermal residual type can be generated, and the verification object is determined to be the doctor blade pressure control record and the oven temperature control record. The current process continuously utilizes closed residual terms, candidate transfer edge sets, edge replacement results, residual starting edges, residual migration sequences, residual cause types, and verification objects. This improves upon the existing technology's reliance on manual experience to troubleshoot equipment issues after discovering color difference marks, ghost marks, bleed marks, or broken lines, as well as unclear verification sequences and insufficient process analysis basis. This makes the printing quality evaluation of rotary screen digital printing machines more suitable for equipment verification and process management in continuous production scenarios.
[0024] Example 2 Please refer to Figure 2 Specifically: S1 includes S11; S11. Read the fabric inspection record of the printed fabric to be evaluated. The fabric inspection record includes the fabric roll number, fabric meter position, finished product quality mark, starting meter position of the finished product quality mark, and ending meter position of the finished product quality mark. The fabric area to be evaluated is determined based on the starting and ending meter positions. The fabric area to be evaluated is divided into several evaluation segments based on the length of the fabric area to be evaluated. Each evaluation segment is marked with evaluation segment information, which includes the evaluation segment number, production period, cylinder screen plate position interval, and finished product quality mark. The production period is the time interval during which the current evaluation segment passes through the printing station of the rotary digital printing machine. The rotary screen plate position interval is the corresponding encoder angle interval of the current evaluation segment within the production period. The finished product quality markings include qualified markings, color difference markings, ghosting markings, bleed markings, and broken lines markings.
[0025] S1 also includes S12; S12. Based on the sensor group installed on the rotary digital printing machine, retrieve the sensor operation data during the current production period from the equipment control system of the rotary digital printing machine, and perform plate position interception on the sensor operation data according to the rotary screen plate position interval to obtain the evaluation segment operation segment. The sensor group includes a fabric speed sensor, a tension sensor, a mesh cylinder encoder, a scraper pressure sensor, and an oven temperature sensor. The sensor operating data includes fabric speed sequence, tension sequence, mesh cylinder phase sequence, scraper pressure sequence, and oven temperature sequence. The plate position extraction process is as follows: When the rotary screen digital printing machine is being mounted or calibrated, the reference position of the rotary screen reference point in the cylinder encoder is obtained, and the circumference of the rotary screen of the rotary screen digital printing machine is divided into several rotary screen positions. During the production period when the evaluation segment passes through the printing station, the operating data of the cylinder encoder is read. The position of the cylinder encoder at the beginning sampling time of the current production period is assigned to the starting rotary screen position, and the position of the cylinder encoder at the end sampling time of the current production period is assigned to the ending rotary screen position to obtain the rotary screen position interval. Then, the fabric speed sequence, tension sequence, cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence located between the starting and ending rotary screen positions are extracted and combined to form the evaluation segment operation segment.
[0026] In this embodiment, the fabric inspection record of the printed fabric to be evaluated is first read, and the fabric roll number, fabric meter position, finished product quality mark, starting meter position of the finished product quality mark, and ending meter position of the finished product quality mark are extracted from the fabric inspection record. Then, the fabric area to be evaluated is determined based on the starting meter position and ending meter position of the finished product quality mark. Then, the fabric area to be evaluated is divided into several evaluation segments according to the length of the fabric area to be evaluated, and each evaluation segment is marked with an evaluation segment number, production time period, rotary screen plate position interval, and finished product quality mark. Among them, the production time period is the time period during which the current evaluation segment passes through the printing station of the rotary screen digital printing machine, the rotary screen plate position interval is the angle interval of the encoder tube mapped by the current evaluation segment during the production time period, and the finished product quality mark includes a qualified mark, a color difference mark, a ghost mark, a bleeding mark, and a broken line mark. After obtaining the evaluation segment information, the fabric speed sensor, tension sensor, encoder tube, and scraper deployed on the rotary screen digital printing machine are used to determine the evaluation segment. The blade pressure sensor and oven temperature sensor retrieve sensor operation data for the current production period from the equipment control system of the rotary digital printing machine. The sensor operation data includes the fabric speed sequence, tension sequence, screen cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence. When the rotary digital printing machine is being mounted or calibrated, the reference position of the rotary screen reference point in the screen cylinder encoder is obtained, and the circumference of the rotary screen of the rotary digital printing machine is divided into several rotary screen positions. During the production period of the evaluation segment passing through the printing station, the operation data of the screen cylinder encoder is read. The position of the screen cylinder encoder at the beginning sampling time of the current production period is assigned to the starting rotary screen position, and the position of the screen cylinder encoder at the end sampling time of the current production period is assigned to the ending rotary screen position to obtain the rotary screen position interval. Then, the fabric speed sequence, tension sequence, screen cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence located between the starting and ending rotary screen positions are extracted and combined to form the operation segment of the evaluation segment. For example, if a double image mark appears on the fabric roll A01 between the 120th and 125th meters of the fabric surface, the fabric area to be evaluated is determined by the starting meter of the finished product quality mark (120th meter) and the ending meter of the finished product quality mark (125th meter). The current fabric area to be evaluated is then divided into several evaluation segments. When one of the evaluation segments passes through the printing station of the rotary screen digital printing machine between 10:15:20 and 10:15:35, 10:15:20 to 10:15:35 is determined as the production period. The starting and ending rotary screen positions are determined based on the encoder data. Then, the fabric speed sequence, tension sequence, cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence within the rotary screen position interval are extracted to form the current evaluation segment's operating segment.Through the above implementation process, the fabric roll number, fabric meter position, finished product quality mark, starting meter position of the finished product quality mark, and ending meter position of the finished product quality mark in the fabric inspection record can participate in the printing quality evaluation together with the production period, rotary screen plate position interval, encoder angle interval, and sensor operation data of the rotary screen digital printing machine. This expands the evaluation object from simple finished fabric defects to a process object that includes fabric position, production time, rotary screen plate position, and equipment operation status. Compared with the existing technology that relies solely on manual inspection records to judge color difference marks, ghost marks, bleeding marks, or broken lines, this implementation method can clearly define the production period and rotary screen plate position interval of each evaluation segment, and retain the operating status of the fabric speed sequence, tension sequence, encoder phase sequence, squeegee pressure sequence, and oven temperature sequence within the current rotary screen plate position interval.
[0027] Example 3 Please refer to Figure 2 Specifically: S2 includes S21 and S22; S21. Based on the sampling order of the phase values of the mesh encoder, assign a circular mesh circumferential position identifier to each sampling point; when the phase value of the mesh encoder returns from the maximum phase to the initial phase, it is recorded as the completion of a circular mesh circumferential cycle, and a new cycle number is generated. Combine the cycle number with the circular mesh circumferential position identifier in the current cycle to form the circular mesh plate position coordinates. Then, using the mesh phase sequence in the evaluation segment as the plate position reference, convert the fabric speed sequence, tension sequence, scraper pressure sequence, and oven temperature sequence to the circular mesh plate position coordinates to obtain the fabric speed plate position sequence, tension plate position sequence, scraper pressure plate position sequence, and oven temperature plate position sequence, and perform data processing to obtain a standard plate position sequence group. The data processing includes median centering and median absolute deviation scaling. The median centering process is as follows: the median of the same evaluation sequence is used as the center value to shift the current evaluation sequence to eliminate the difference in the baseline position of the sequence within the same evaluation segment; The median absolute deviation scaling process is as follows: the median absolute deviation of the same plate position sequence is used as the scaling factor to perform a scaling transformation on the plate position sequence after center shift, so that different sensor sequences can enter the same analysis scale. The standard printing plate sequence group includes the standard printing plate sequence for fabric speed, tension, scraper pressure, and oven temperature; S22. Select any two standard plate position sequences from the standard plate position sequence group to form a sensor sequence pair. Perform dynamic time warping on each sensor sequence pair to obtain the warped path of the current sensor sequence pair in the circular grid plate position coordinates. Encode each path step in the warped path to form a path coding item. The path coding item includes a synchronization code, a first sequence retention code, and a second sequence retention code. All path coding entries for the same sensor sequence pair are arranged in a regular path order to form the path coding for the current sensor sequence pair; Synchronization code indicates that two standard plate positions in a sensor sequence pair simultaneously enter the next circular grid plate position in the regularization path; The first sequence retention code indicates that the first standard plate position sequence maintains the original circular grid plate position in the regularized path, and the second standard plate position sequence enters the next circular grid plate position; The second sequence retention code indicates that the second standard plate position sequence maintains the original circular grid position in the regularized path, and the first standard plate position sequence enters the next circular grid position.
[0028] In this embodiment, based on the phase sequence of the mesh cylinder in the evaluation segment, each sampling point is assigned a circumferential position identifier according to the sampling order of the mesh cylinder encoder phase values. When the mesh cylinder encoder phase value returns from the maximum phase to the initial phase, it is recorded as a circumferential cycle of the circular mesh, and a new cycle number is generated. The cycle number and the circumferential position identifier of the circular mesh are combined to form the circular mesh plate position coordinates. Subsequently, using the mesh cylinder phase sequence as the plate position reference, the fabric speed sequence, tension sequence, scraper pressure sequence, and oven temperature sequence are converted into fabric speed plate position sequences, tension plate position sequences, scraper pressure plate position sequences, and oven temperature plate position sequences, respectively. Median centering and median absolute deviation scaling are then performed to obtain standard plate position sequences for fabric speed, tension, scraper pressure, and oven temperature. For example, if the numerical reference of the fabric speed sequence is too high or the numerical dimension of the oven temperature sequence is large in the same evaluation segment, the above data processing can allow the operating data of different sensors to be compared under the same circular mesh plate position coordinates and the same analysis scale. Subsequently, any two standard plate position sequences are selected from the standard plate position sequence group to form a sensor sequence pair. For example, the fabric speed standard plate position sequence and the tension standard plate position sequence are formed into a sensor sequence pair, and dynamic time warping is performed to obtain a warped path. Then, the path steps in the warped path are marked with synchronization codes, first sequence retention codes, and second sequence retention codes to form path codes. Through the current implementation method, the fabric speed sequence, tension sequence, squeegee pressure sequence, and oven temperature sequence, which were originally scattered on the time axis, can be transformed into a standard plate position sequence unfolding around the rotary screen plate position coordinates. This makes it easier to determine the synchronization and retention states of different sensor operating data during the plate position advancement process. Compared with directly viewing the original time series, the current method can more clearly reflect the changes between the rotary screen digital printing machine operating data and the rotary screen plate position, so that the evaluation of printing quality is no longer limited to the observation of single sensor values.
[0029] Example 4 Please refer to Figure 2 Specifically: S3 includes S31 and S32; S31. Based on the front and rear coordinates of the circular grid plate position coordinates, set the standard plate position sequence in the sensor sequence pair as the first standard plate position sequence and the second standard plate position sequence respectively, and perform plate position displacement. Wherein, the plate position displacement is the number of sampling points that the second standard plate position sequence moves relative to the first standard plate position sequence along the circular grid plate position coordinates; The portion of the first standard plate position sequence and the second standard plate position sequence that simultaneously have sampled values in the circular grid plate position coordinates is taken as the overlapping plate position range. Within the overlapping plate position range, the average sampled values of the first standard plate position sequence and the average sampled values of the second standard plate position sequence after displacement are calculated respectively. The deviation of the two standard plate position sequences at the same plate position sampling point is calculated based on their respective average sampled values. The two deviations at the same plate position sampling point are multiplied together, and the average of all products within the overlapping plate position range is calculated to obtain the cross-covariance value. S32. Select the candidate plate position displacement with the largest absolute value of cross-covariance as the plate position hysteresis of the current sensor sequence pair, and generate the plate position hysteresis direction based on the plate position hysteresis. When the page position lag is positive, the first standard page position sequence is the leading sequence and the second standard page position sequence is the passive sequence. At this time, the page position lag direction is marked as the first leading direction. When the page position lag is negative, the second standard page position sequence is the leading sequence and the first standard page position sequence is the passive sequence. At this time, the page position lag direction is the second leading direction. When the page position lag is zero, the first standard page position sequence and the second standard page position sequence are in the same position. At this time, the page position lag direction is marked as the same position direction.
[0030] S3 also includes S33; S33. Combine the path encoding and plate lag direction of the same sensor sequence pair; When the lagging direction of the plate position is the first leading direction, the sensor type of the first standard plate position sequence is used as the starting sensor type, and the sensor type of the second standard plate position sequence is used as the ending sensor type to form a sensor transfer edge. When the lagging direction of the plate position is the second leading direction, the sensor type to which the second standard plate position sequence belongs is used as the starting sensor type, and the sensor type to which the first standard plate position sequence belongs is used as the ending sensor type to form a sensor transfer edge. When the lag direction of the plate position is the same as the lag direction, no sensor transmission edge is formed; The sensor transmission edge includes the starting sensor type, the ending sensor type, the circular grid plate position range, and the path code.
[0031] In this embodiment, based on the front and rear coordinates of the circular grid plate position coordinates, the two standard plate position sequences in the sensor sequence pair are first designated as the first standard plate position sequence and the second standard plate position sequence, respectively. The second standard plate position sequence is then displaced relative to the first standard plate position sequence along the circular grid plate position coordinates. After each plate position displacement, the portion of the first standard plate position sequence and the displaced second standard plate position sequence where sampled values simultaneously exist in the circular grid plate position coordinates is taken as the overlapping plate position range. The average sampled value of the first standard plate position sequence and the average sampled value of the displaced second standard plate position sequence are calculated respectively. Finally, the average product of the deviations of the two standard plate position sequences at the same plate position sampling point is calculated to obtain the cross-covariance value. Subsequently, the candidate plate position displacement with the largest absolute value of cross-covariance is taken as the plate position lag, and the plate position lag direction is determined based on the plate position lag. When the plate position lag is positive, the first standard plate position sequence is the leading sequence, and the second standard plate position sequence is the driven sequence; when the plate position lag is negative, the second standard plate position sequence is the leading sequence, and the first standard plate position sequence is the driven sequence; when the plate position lag is zero, the first standard plate position sequence and the second standard plate position sequence are co-position sequences. Then, the path encoding and plate position lag direction are combined to form a sensor transfer edge including the starting sensor type, the ending sensor type, the circular grid plate position interval, and the path encoding; when the plate position lag direction is a co-position direction, no sensor transfer edge is formed. For example, when the fabric speed standard plate position sequence is used as the first standard plate position sequence and the tension standard plate position sequence is used as the second standard plate position sequence, if the absolute value of the cross-covariance between the tension standard plate position sequence and the fabric speed standard plate position sequence is the largest after moving three sampling points along the rotary screen plate position coordinates, and the plate position hysteresis is positive, then a sensor transmission edge is formed from the fabric speed sensor type to the tension sensor type. Through the current implementation method, it is possible to determine the order of changes and plate position differences between different standard plate position sequences, so that the evaluation of printing quality is no longer limited to observing fluctuations in the operating data of a single sensor, but can identify the transmission direction between the fabric speed standard plate position sequence, the tension standard plate position sequence, the squeegee pressure standard plate position sequence, and the oven temperature standard plate position sequence.
[0032] Example 5 Please refer to Figure 2 Specifically: S4 includes S41; S41. Select three types of sensors located in the same circular grid plate position interval according to the direction and connection sequence of the sensor transmission edge, and set them as the first sensor type, the second sensor type and the third sensor type respectively; When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the second sensor type, the current sensor transfer edge is set as the first transfer edge. If there is a sensor transfer edge in the sensor transfer edge group that points from the second sensor type to the third sensor type, set the current sensor transfer edge as the second transfer edge. When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the third sensor type, set the current sensor transfer edge as the third transfer edge. Among them, the first and second passing edges constitute an indirect passing path, and the third passing edge constitutes a direct passing path.
[0033] S4 also includes S42; S42. Extract the path codes of the first and second transmission edges and combine them according to the transmission direction to form an indirect path code. Extract the path code of the third transmission edge and record it as the direct path code. Then compare the indirect path code with the direct path code. When the indirect path code and the direct path code are consistent, the transmission closed unit is recorded as a closed consistent unit. Otherwise, it is recorded as a closed residual unit. The position of the circular grid plate where the path codes of the indirect transmission path and the direct transmission path first appear different in the closed residual unit is recorded as the residual plate position. The evaluation segment number, the circular grid plate position interval, the first sensor type, the second sensor type, the third sensor type, the indirect path code, the direct path code, and the residual plate position are combined to form a closed residual term. All closed residual terms are counted to obtain a closed residual set.
[0034] In this embodiment, based on the direction and connection order of the sensor transmission edges, three types of sensors located within the same circular grid plate interval are selected and respectively designated as the first sensor type, the second sensor type, and the third sensor type. When there is a sensor transmission edge in the sensor transmission edge group that points from the first sensor type to the second sensor type, the current sensor transmission edge is designated as the first transmission edge. When there is a sensor transmission edge in the sensor transmission edge group that points from the second sensor type to the third sensor type, the current sensor transmission edge is designated as the second transmission edge. When there is a sensor transmission edge in the sensor transmission edge group that points from the first sensor type to the third sensor type, the current sensor transmission edge is designated as the third transmission edge. The first and second transmission edges constitute an indirect transmission path, and the third transmission edge constitutes a direct transmission path. Then, the path codes of the first and second transmission edges are extracted and combined according to the transmission direction to form an indirect path code. At the same time, the path code of the third transmission edge is extracted as the direct path code. The indirect path code and the direct path code are then compared for consistency. When the indirect path code and the direct path code are consistent, the transmission closure unit is recorded as a closed consistent unit. When the indirect path code and the direct path code are inconsistent, the transmission closure unit is recorded as a closed residual unit. The position of the circular mesh plate where the path codes of the indirect transmission path and the direct transmission path first differ in the closed residual unit is recorded as the residual plate position. Finally, the evaluation segment number, the circular mesh plate position interval, the first sensor type, the second sensor type, the third sensor type, the indirect path code, the direct path code, and the residual plate position are combined to form a closed residual term. All closed residual terms are counted to obtain the closed residual set. For example, within the same rotary screen printing plate area, if there are sensor transmission edges pointing from the fabric speed sensor type to the tension sensor type, from the tension sensor type to the squeegee pressure sensor type, and from the fabric speed sensor type to the squeegee pressure sensor type, then the first two sensor transmission edges constitute an indirect transmission path, and the last sensor transmission edge constitutes a direct transmission path. If the indirect path code and the direct path code become inconsistent at a certain rotary screen printing plate position, then the current rotary screen printing plate position is recorded as the residual plate position. Through this current implementation, it is possible to determine whether the sensor operating state is consistent with the direct transmission path when propagating along the indirect transmission path. This allows printing quality evaluation to no longer simply judge whether a single sensor transmission edge is abnormal, but to identify the transmission closure differences between the fabric speed sensor type, tension sensor type, squeegee pressure sensor type, and oven temperature sensor type. Compared to relying solely on manual experience to view sensor operating data, the current method can pinpoint path coding differences to specific evaluation segment numbers, rotary screen printing plate areas, and residual plate position positions, facilitating the determination of the location of quality anomalies in the sensor transmission chain.
[0035] Example 6 Please refer to Figure 2 Specifically: S5 includes S51; S51. Extract the sensor transfer edges that participate in the closed residual term in the closed residual set, form a candidate transfer edge set, and perform an edge replacement process on each candidate transfer edge in the candidate transfer edge set. Edge replacement processing: Replace the current candidate passing edge of the segment to be evaluated with the same passing edge of the evaluation segment with the qualified mark in the same production batch to form a replacement spectrum. Iterate through S4 according to the replacement spectrum to obtain replacement residual terms. Statistically count all replacement residual terms to obtain the replacement residual set. Match the closed residual set with the replacement residual set item by item: the closed residual items that exist in the original closed residual set but no longer exist in the replacement residual set after the candidate passing edge is replaced are recorded as the eliminated residual items; After all candidate passing edges have been replaced, the candidate passing edge with the largest number of eliminated residual terms is selected as the residual starting edge. When there are two or more candidate passing edges with the same number of eliminated residual terms, the candidate passing edge with the first position in the circle screen is selected as the residual starting edge. Starting from the residual starting edge, search for closed residual items in the closed residual set that share the same endpoint sensor type or starting sensor type as the residual starting edge, and arrange them according to the residual plate position to form a residual migration sequence.
[0036] S5 also includes S52; S52. Generate residual cause types based on finished product quality marks, residual starting edges, and residual migration sequences; The residual starting edge includes a starting position sequence type and an ending position sequence type; the residual migration sequence includes at least one residual propagation edge, and each residual propagation edge includes a starting position sequence type and an ending position sequence type. When the finished product quality is marked as a ghost mark, and the starting point plate position sequence type of the residual starting edge is a mesh cylinder phase sequence, the residual cause type is recorded as phase closure residual type. At this time, the verification objects are determined as the circular mesh synchronous control record and the mesh cylinder encoder record. When the finished product quality is marked as a broken mark, and the starting point plate position sequence type or the ending point plate position sequence type of the residual starting edge is a scraper pressure plate position sequence, the residual cause type is recorded as imprint closure residual type, and the object to be checked is determined as scraper pressure control record. When the finished product quality is marked as a diffusion mark, and there is a first residual transfer edge and a second residual transfer edge in the residual migration sequence, wherein the starting plate position sequence type or the ending plate position sequence type of the first residual transfer edge is a doctor blade pressure plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the second residual transfer edge is an oven temperature plate position sequence, the residual cause type is recorded as the printing heat treatment residual type. At this time, the verification object is determined to be the doctor blade pressure control record and the oven temperature control record. When the finished product quality is marked as color difference, and there are a third residual transfer edge and a fourth residual transfer edge in the residual migration sequence, where the starting plate position sequence type or the ending plate position sequence type of the third residual transfer edge is the cloth speed plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the fourth residual transfer edge is the squeegee pressure plate position sequence, the residual cause type is recorded as cloth printing residual type. At this time, the verification objects are determined as cloth speed control records and squeegee pressure control records. When the residual migration sequence involves more than three types of plate position sequence, the residual cause type is recorded as multi-source closed residual type. At this time, the verification object is determined to be the control records of all sensor types in the residual migration sequence.
[0037] In this embodiment, sensor transfer edges participating in the closed residual terms are first extracted from the closed residual set to form a candidate transfer edge set. Each candidate transfer edge in the candidate transfer edge set is then subjected to an edge replacement process. During the edge replacement process, the current candidate transfer edge of the segment to be evaluated is replaced with a transfer edge of the same name from the evaluation segment with a qualified mark in the same production batch, forming a replacement spectrum. Then, based on the replacement spectrum, the indirect and direct transfer paths are reconstructed via S4 to obtain the replacement residual terms, and a replacement residual set is formed. Subsequently, the closed residual set and the replacement residual set are matched item by item. Closed residual terms that exist in the original closed residual set but no longer exist in the replacement residual set after the candidate transfer edge replacement are recorded as eliminated residual terms. After all candidate transfer edges have undergone edge replacement processing, the candidate transfer edge with the largest number of eliminated residual terms is selected as the residual starting edge. If two or more candidate transfer edges have the same number of eliminated residual terms, the candidate transfer edge with the earlier position in the circular mesh layout is selected as the residual starting edge. Next, starting from the residual starting edge, the closed residual items sharing the same endpoint sensor type or starting sensor type as the residual starting edge are searched in the closed residual set, and arranged according to the residual plate position to form a residual migration sequence. Finally, the residual cause type is generated based on the finished product quality mark, the residual starting edge, and the residual migration sequence, and the verification object is determined. For example, if the finished product quality mark is a crack mark, and the starting plate position sequence type or the endpoint plate position sequence type of the residual starting edge is a doctor blade pressure plate position sequence, the residual cause type is recorded as an imprint closed residual type, and the verification object is determined as the doctor blade pressure control record; if the finished product quality mark is a diffusion mark, and the residual migration sequence contains both a first residual transfer edge involving the doctor blade pressure plate position sequence and a second residual transfer edge involving the oven temperature plate position sequence, the residual cause type is recorded as an imprint thermal residual type, and the verification object is determined as the doctor blade pressure control record and the oven temperature control record. The current implementation method can trace closed residual terms back to specific candidate transfer edges and determine the type of residual cause based on finished product quality marks. This allows color difference marks, ghosting marks, diffusion marks, and broken weave marks to no longer be used only as fabric surface inspection conclusions, but also to point to control records of all sensor types in fabric speed control records, circular mesh synchronization control records, mesh cylinder encoder records, doctor blade pressure control records, oven temperature control records, or residual migration sequences. Compared to the method of checking each item by item based on manual experience, the current method can provide the residual starting edge, residual migration sequence, residual cause type, and verification object, making the equipment verification sequence and process analysis direction clear.
[0038] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.
Claims
1. A method for evaluating printing quality based on a rotary screen digital printing machine, characterized in that: include: S1. Read the fabric inspection record, divide the evaluation segment according to the fabric inspection record and mark the production period and the circular mesh position interval, retrieve the sensor operation data of the current production period, extract and combine the data according to the circular mesh position interval to obtain the evaluation segment operation segment. S2. Construct the circular grid position coordinates and convert the sensor operation data to the circular grid position coordinates. After data processing, obtain the standard position sequence group. Select any two standard position sequences to construct sensor sequence pairs and perform dynamic time warping to obtain path codes. S3. Perform position displacement on the sensor sequence pair and calculate the cross-covariance value within the overlapping position range. Determine the position hysteresis and position hysteresis direction based on the candidate position displacement with the largest absolute value of cross-covariance, and generate the sensor transmission edge by combining the path coding. S4. Construct indirect and direct transmission paths based on the direction and connection sequence of the sensor transmission edges, and obtain closed residual terms after consistency comparison. S5. Replace the candidate transfer edge and compare the closed residual terms before and after the replacement to obtain the residual starting edge and residual migration sequence. Then, combine the finished product quality mark in the fabric inspection record to generate the residual cause type and the verification object.
2. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 1, characterized in that: S1 includes S11; S11. Read the fabric inspection record of the printed fabric to be evaluated. The fabric inspection record includes the fabric roll number, fabric meter position, finished product quality mark, starting meter position of the finished product quality mark, and ending meter position of the finished product quality mark. The fabric area to be evaluated is determined based on the starting and ending meter positions. The fabric area to be evaluated is divided into several evaluation segments based on the length of the fabric area to be evaluated. Each evaluation segment is marked with evaluation segment information, which includes the evaluation segment number, production period, cylinder screen plate position interval, and finished product quality mark. Finished product quality markings include qualified markings, color difference markings, ghosting markings, diffusion markings, and crack markings.
3. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 2, characterized in that: S1 also includes S12; S12. Based on the sensor group installed on the rotary digital printing machine, retrieve the sensor operation data during the current production period from the equipment control system of the rotary digital printing machine, and perform plate position interception on the sensor operation data according to the rotary screen plate position interval to obtain the evaluation segment operation segment. The sensor group includes a fabric speed sensor, a tension sensor, a mesh cylinder encoder, a scraper pressure sensor, and an oven temperature sensor. The sensor operating data includes fabric speed sequence, tension sequence, mesh cylinder phase sequence, scraper pressure sequence, and oven temperature sequence. The plate position extraction process is as follows: When the rotary screen digital printing machine is being mounted or calibrated, the reference position of the rotary screen reference point in the cylinder encoder is obtained, and the circumference of the rotary screen of the rotary screen digital printing machine is divided into several rotary screen positions. During the production period when the evaluation segment passes through the printing station, the operating data of the cylinder encoder is read. The position of the cylinder encoder at the beginning sampling time of the current production period is assigned to the starting rotary screen position, and the position of the cylinder encoder at the end sampling time of the current production period is assigned to the ending rotary screen position to obtain the rotary screen position interval. Then, the fabric speed sequence, tension sequence, cylinder phase sequence, squeegee pressure sequence, and oven temperature sequence located between the starting and ending rotary screen positions are extracted and combined to form the evaluation segment operation segment.
4. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 3, characterized in that: S2 includes S21 and S22; S21. Based on the sampling order of the phase values of the mesh encoder, assign a circular mesh circumferential position identifier to each sampling point; when the phase value of the mesh encoder returns from the maximum phase to the initial phase, it is recorded as the completion of a circular mesh circumferential cycle, and a new cycle number is generated. Combine the cycle number with the circular mesh circumferential position identifier in the current cycle to form the circular mesh plate position coordinates. Then, using the mesh phase sequence in the evaluation segment as the plate position reference, convert the fabric speed sequence, tension sequence, scraper pressure sequence, and oven temperature sequence to the circular mesh plate position coordinates to obtain the fabric speed plate position sequence, tension plate position sequence, scraper pressure plate position sequence, and oven temperature plate position sequence, and perform data processing to obtain a standard plate position sequence group. The data processing includes median centering and median absolute deviation scaling. The standard printing plate sequence group includes the standard printing plate sequence for fabric speed, tension, scraper pressure, and oven temperature; S22. Select any two standard plate position sequences from the standard plate position sequence group to form a sensor sequence pair. Perform dynamic time warping on each sensor sequence pair to obtain a warped path. Encode each path step in the warped path to form a path coding item. The path coding item includes a synchronization code, a first sequence retention code, and a second sequence retention code. All path coding items for the same sensor sequence pair are arranged in a regular path order to form a path code.
5. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 4, characterized in that: S3 includes S31 and S32; S31. Based on the front and rear coordinates of the circular grid plate position coordinates, set the standard plate position sequence in the sensor sequence pair as the first standard plate position sequence and the second standard plate position sequence respectively, and perform plate position displacement. Wherein, the plate position displacement is the number of sampling points that the second standard plate position sequence moves relative to the first standard plate position sequence along the circular grid plate position coordinates; The portion of the first standard plate position sequence and the second standard plate position sequence that simultaneously have sampled values in the circular grid plate position coordinates is taken as the overlapping plate position range. Within the overlapping plate position range, the average sampled values of the first standard plate position sequence and the average sampled values of the second standard plate position sequence after displacement are calculated respectively. The deviation of the two standard plate position sequences at the same plate position sampling point is calculated based on their respective average sampled values. The two deviations at the same plate position sampling point are multiplied together, and the average of all products within the overlapping plate position range is calculated to obtain the cross-covariance value. S32. Select the candidate plate position displacement with the largest absolute value of cross-covariance as the plate position hysteresis of the current sensor sequence pair, and generate the plate position hysteresis direction based on the plate position hysteresis. When the page position lag is positive, the first standard page position sequence is the leading sequence and the second standard page position sequence is the passive sequence. At this time, the page position lag direction is marked as the first leading direction. When the page position lag is negative, the second standard page position sequence is the leading sequence and the first standard page position sequence is the passive sequence. At this time, the page position lag direction is the second leading direction. When the page position lag is zero, the first standard page position sequence and the second standard page position sequence are in the same position. At this time, the page position lag direction is marked as the same position direction.
6. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 5, characterized in that: S3 also includes S33; S33. Combine the path encoding and plate lag direction of the same sensor sequence pair; When the lagging direction of the plate position is the first leading direction, the sensor type of the first standard plate position sequence is used as the starting sensor type, and the sensor type of the second standard plate position sequence is used as the ending sensor type to form a sensor transfer edge. When the lagging direction of the plate position is the second leading direction, the sensor type to which the second standard plate position sequence belongs is used as the starting sensor type, and the sensor type to which the first standard plate position sequence belongs is used as the ending sensor type to form a sensor transfer edge. When the lag direction of the plate position is the same as the lag direction, no sensor transmission edge is formed.
7. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 6, characterized in that: S4 includes S41; S41. Select three types of sensors located in the same circular grid plate position interval according to the direction and connection sequence of the sensor transmission edge, and set them as the first sensor type, the second sensor type and the third sensor type respectively; When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the second sensor type, the current sensor transfer edge is set as the first transfer edge. If there is a sensor transfer edge in the sensor transfer edge group that points from the second sensor type to the third sensor type, set the current sensor transfer edge as the second transfer edge. When there is a sensor transfer edge in the sensor transfer edge group that points from the first sensor type to the third sensor type, set the current sensor transfer edge as the third transfer edge. Among them, the first and second passing edges constitute an indirect passing path, and the third passing edge constitutes a direct passing path.
8. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 7, characterized in that: S4 also includes S42; S42. Extract the path codes of the first and second transmission edges and combine them according to the transmission direction to form an indirect path code. Extract the path code of the third transmission edge and record it as the direct path code. Then compare the indirect path code with the direct path code. When the indirect path code and the direct path code are consistent, the transmission closed unit is recorded as a closed consistent unit. Otherwise, it is recorded as a closed residual unit. The position of the circular grid plate where the path codes of the indirect transmission path and the direct transmission path first appear different in the closed residual unit is recorded as the residual plate position. The evaluation segment number, the circular grid plate position interval, the first sensor type, the second sensor type, the third sensor type, the indirect path code, the direct path code, and the residual plate position are combined to form a closed residual term. All closed residual terms are counted to obtain a closed residual set.
9. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 8, characterized in that: S5 includes S51; S51. Extract the sensor transfer edges that participate in the closed residual term in the closed residual set, form a candidate transfer edge set, and perform an edge replacement process on each candidate transfer edge in the candidate transfer edge set. Edge replacement processing: Replace the current candidate passing edge of the segment to be evaluated with the same passing edge of the evaluation segment with the qualified mark in the same production batch to form a replacement spectrum. Iterate through S4 according to the replacement spectrum to obtain replacement residual terms. Statistically count all replacement residual terms to obtain the replacement residual set. Match the closed residual set with the replacement residual set item by item: the closed residual items that exist in the original closed residual set but no longer exist in the replacement residual set after the candidate passing edge is replaced are recorded as the eliminated residual items; After all candidate passing edges have been replaced, the candidate passing edge with the largest number of eliminated residual terms is selected as the residual starting edge. When there are two or more candidate passing edges with the same number of eliminated residual terms, the candidate passing edge with the first position in the circle screen is selected as the residual starting edge. Starting from the residual starting edge, search for closed residual items in the closed residual set that share the same endpoint sensor type or starting sensor type as the residual starting edge, and arrange them according to the residual plate position to form a residual migration sequence.
10. The printing quality evaluation method based on a rotary screen digital printing machine according to claim 9, characterized in that: S5 also includes S52; S52. Generate residual cause types based on finished product quality marks, residual starting edges, and residual migration sequences; The residual starting edge includes a starting position sequence type and an ending position sequence type; the residual migration sequence includes at least one residual propagation edge, and each residual propagation edge includes a starting position sequence type and an ending position sequence type. When the finished product quality is marked as a ghost mark, and the starting point plate position sequence type of the residual starting edge is a mesh cylinder phase sequence, the residual cause type is recorded as phase closure residual type. At this time, the verification objects are determined as the circular mesh synchronous control record and the mesh cylinder encoder record. When the finished product quality is marked as a broken mark, and the starting point plate position sequence type or the ending point plate position sequence type of the residual starting edge is a scraper pressure plate position sequence, the residual cause type is recorded as imprint closure residual type, and the object to be checked is determined as scraper pressure control record. When the finished product quality is marked as a diffusion mark, and there is a first residual transfer edge and a second residual transfer edge in the residual migration sequence, wherein the starting plate position sequence type or the ending plate position sequence type of the first residual transfer edge is a doctor blade pressure plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the second residual transfer edge is an oven temperature plate position sequence, the residual cause type is recorded as the printing heat treatment residual type. At this time, the verification object is determined to be the doctor blade pressure control record and the oven temperature control record. When the finished product quality is marked as color difference, and there are a third residual transfer edge and a fourth residual transfer edge in the residual migration sequence, where the starting plate position sequence type or the ending plate position sequence type of the third residual transfer edge is the cloth speed plate position sequence, and the starting plate position sequence type or the ending plate position sequence type of the fourth residual transfer edge is the squeegee pressure plate position sequence, the residual cause type is recorded as cloth printing residual type. At this time, the verification objects are determined as cloth speed control records and squeegee pressure control records. When the residual migration sequence involves more than three types of plate position sequence, the residual cause type is recorded as multi-source closed residual type. At this time, the verification object is determined to be the control records of all sensor types in the residual migration sequence.