Analysis and early warning method for strip breakage of pickling rolling unit

Abstract

The invention relates to the technical field of steel rolling, in particular to an acid rolling mill unit broken belt analysis and early warning method, a knowledge base is formed by combining a broken belt mechanism with big data classification statistics, and post analysis and pre-warning are performed on broken belts based on the knowledge base. The knowledge base is updated along with continuous accumulation of the broken belt cases; secondly, when belt breakage occurs, automatically integrating and processing relevant data of the belt breakage by the system, performing all-directional belt breakage analysis based on the knowledge base, and giving a belt breakage analysis conclusion and a parameter optimization suggestion; and thirdly, the core function of the system is that before strip steel production, strip breakage risk prediction is carried out on the steel coil in a production plan based on the knowledge base, corresponding optimization parameters and production scheduling suggestions are matched, and the strip breakage probability is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of steel rolling, and in particular to a belt break analysis and early warning method for an acid rolling mill. Background Art

[0002] Cold-rolled strip steel is made of hot-rolled steel coils as raw materials. After pickling to remove the oxide scale, cold rolling is carried out. The finished product is hard-rolled coils. The cold work hardening caused by continuous cold deformation increases the strength and hardness of the hard-rolled coils, but the toughness decreases. In cold rolling production, strip breakage is a common accident. Strip breakage refers to the strip being broken in the rolling mill frame. A mild strip breakage accident will cause the unit to shut down and affect production efficiency. A serious strip breakage accident requires a lot of time to handle the equipment, and often requires roller replacement, which seriously affects the operating rate and requires cutting, resulting in a decrease in product yield and degradation of product quality in the pickling area. When the strip breakage failure is serious, it will also damage the working rolls, intermediate rolls, and even support rolls, increasing roll consumption. The strip breakage failure in the cold rolling process will cause great harm to production safety, equipment, roll system and quality, and cause significant economic losses. Therefore, it is very important to study the causes of strip breakage in the cold rolling process and give early warnings, whether to improve the yield rate of thin plate production, reduce product costs, or extend the service life of the rolling mill.

[0003] In the cold rolling production line, strip breakage on the rolling mill is a relatively serious accident. The accident handling time is long, and it may damage the rollers, produce degraded products, and reduce the yield rate. Therefore, reducing the occurrence of strip breakage accidents and quickly diagnosing them are particularly important in cold rolling production. Summary of the invention

[0004] The present invention provides a method for analyzing and warning of belt breakage in a pickling mill. The method systematizes the expert diagnosis method and combines it with big data classification statistics to form a knowledge base. Based on the knowledge base, post-analysis and pre-warning of belt breakage are performed, and warning and optimization parameters are provided for subsequent production. This is crucial to reducing belt breakage accidents and can significantly reduce the number of belt breakages.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for analyzing and warning of belt breakage in a pickling mill comprises the following contents:

[0007] 1) Based on the mechanism of band breaking, the band breaking is classified into weld band breaking and non-weld band breaking. Each category is further divided into different steel grades. The band breaking characteristic values ​​are counted according to different types, and a band breaking knowledge base is constructed, which is updated in real time.

[0008] 2) When a belt break occurs, the system automatically collects and processes the relevant data of the belt break, performs a comprehensive belt break analysis based on the knowledge base, and gives the belt break analysis conclusion and parameter optimization suggestions;

[0009] 3) Before strip production, the risk of strip breakage in the production plan is predicted based on the knowledge base, and the corresponding optimization parameters and production scheduling suggestions are matched.

[0010] Furthermore, the specific process of building the broken belt knowledge base includes the following steps:

[0011] (1) In the initial stage of system construction, the system is configured according to the system's own knowledge base, according to the warning thresholds of hot-rolled raw material related parameters, production plan transition warning thresholds, weld quality hit rate thresholds, surface defect warning thresholds, and rolling parameter thresholds;

[0012] (2) When the weld is broken, if the steel grades of the previous and next steel coils are different, historical data are collected to obtain the minimum thickness difference and temperature difference between the head of the previous coil and the tail of the next coil corresponding to the upstream hot-rolled raw material under the same steel grade transition condition, which is used as the early warning threshold for the hot-rolled raw material breakage;

[0013] (3) When the weld is broken, if the steel grades of the front and rear coils are different, historical data are collected to obtain the same steel grade transition condition. The front and rear steel grades are combined to record the corresponding minimum transition rolling force, transition tension, and transition roll gap as the break warning threshold. At the same time, the historical data of the relevant steel grade transition without breaking is obtained, and the corresponding unit tension and the reduction rate of each stand are recorded as optimization parameter recommendations;

[0014] (4) When the weld is broken, the difference between the weld quality parameters at the time of breakage and the weld quality parameters without breakage is obtained through historical data statistics, including weld penetration rate, weld height difference, weld reinforcement height, and weld incomplete filling. The minimum hit rate and the minimum standard deviation of each parameter are used as the breakage warning threshold;

[0015] (5) When the non-weld seam breaks, the acid rolling surface inspection data is obtained. At the same time, based on historical big data statistics, the minimum defect area caused by the same type of defect is used as the surface defect warning threshold of the broken strip.

[0016] Furthermore, the broken belt analysis includes the following steps:

[0017] (1) Real-time monitoring of the L1 strip-breaking signal of the basic automation level of pickling mill. Based on the rack number where the strip-breaking signal occurs and the strip position tracking, it is determined whether the steel coil numbers before and after this rack are the same. If they are the same, it is a non-weld strip break. If they are different, it is identified as a weld strip break.

[0018] (2) Obtain the upstream hot-rolled raw material size and temperature strip length sequence data corresponding to the coil, with the data accuracy required to reach 0.5 to 1 meter per point, and analyze whether the hot-rolled raw material position data corresponding to the broken strip position is abnormal based on the threshold value of the corresponding parameter in the knowledge base;

[0019] (3) If the weld is broken, the weld quality detection data of the roll is obtained by communicating with the weld monitoring system of the welding machine, and the weld quality is analyzed;

[0020] (4) If the belt is broken at a non-weld seam, obtain the acid rolling surface inspection data and determine whether there is a serious defect at the corresponding belt break position according to the threshold of the corresponding parameter in the knowledge base;

[0021] (5) Obtain other L1-level rolling parameters related to strip breakage, including rolling force, tension, roll gap, speed setting value and actual value of the front and rear coils. The data accuracy is required to reach a frequency of 10 to 50 ms. Analyze whether there is an abnormality in the rolling parameters based on the threshold value of the corresponding parameters in the knowledge base;

[0022] (6) Based on the above analysis, the conclusion of the broken belt analysis is given.

[0023] Furthermore, the belt break warning is performed after the production plan is issued before production, and specifically includes the following steps:

[0024] (1) Receive production plan data issued by the manufacturing execution system L3;

[0025] (2) based on the broken strip knowledge base, abnormal identification is performed on the hot-rolled raw material size and temperature data corresponding to the coil;

[0026] (3) identifying abnormal defects in the hot rolling table inspection data corresponding to the coil based on the broken strip knowledge base;

[0027] (4) Conduct transition review between each roll and the next roll in the production plan, use the rolling force model to calculate the rolling force difference, roll gap difference, tension difference, and speed difference between the two rolls, and issue an early warning of the risk of strip breakage based on the strip breakage knowledge base;

[0028] (5) Based on the above analysis, the final belt break risk level is given. The risk level is represented by color, with green indicating no risk, yellow indicating a risk of less than 50%, and red indicating a risk of more than 50%;

[0029] (6) For production sequences with the risk of belt breakage, matching is performed according to the knowledge base, and optimal parameters or production scheduling suggestions are given to avoid belt breakage;

[0030] (7) When the defect location reaches the entrance uncoiler and 50 meters in front of the rolling mill, an early warning is issued again.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] Integrate all belt break cases, build a belt break knowledge base, and automatically analyze the causes of belt breakage to improve the efficiency of technicians in analyzing and solving belt breakage problems; based on the big data knowledge base, summarize excellent samples and warning thresholds for belt breakage, provide warnings and optimize parameters for subsequent production, which is crucial to reducing belt breakage accidents and can significantly reduce the number of belt breaks. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of counting the number of belt breaks according to an embodiment of the present invention.

[0034] Figure 2 It is a schematic diagram of the number of belt breaks classified according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings:

[0036] The present invention provides a method for analyzing and warning of belt breakage in a pickling mill, comprising the following contents:

[0037] 1) Based on the mechanism of band breaking, the band breaking is classified into weld band breaking and non-weld band breaking. Each category is further divided into different steel grades. The band breaking characteristic values ​​are counted according to different types, and a band breaking knowledge base is constructed, which is updated in real time.

[0038] First, the breaking of strips needs to take rolling parameters as the primary analysis point, confirm the abnormal rolling parameter items, and then further analyze the real reasons for the abnormal rolling parameters. When different specifications of strips break, the degree of abnormal rolling parameters is different, and it is necessary to use big data for statistics, including the parameter mean, limit value, range, and variance under normal production conditions, as shown in Table 1;

[0039] Table 1 Knowledge base - rolling parameter threshold construction project

[0040]

[0041] According to the strip steel grading rules, the hot-rolled raw steel coils are graded, and the strip steel grading rules are determined using a three-dimensional rule code. The three dimensions of strip steel grading are: set thickness, width, and steel type; among them, according to actual needs, the thickness is divided into 9 grades: the width is divided into 6 grades, and the steel type is divided into 10 grades. Different grades represent different intervals. Each strip steel corresponds to a unique three-dimensional grading interval, and each strip steel roll corresponds to a unique three-dimensional rule code. According to the above three-dimensional rule code, the initial threshold is given according to the system's own knowledge base, see Table 2;

[0042] Table 2 Knowledge Base - Hot Rolled Raw Material Threshold Construction Project

[0043]

[0044] The hardest and softest steel grades allowed to be connected for different steel grades (divided into 8 grades according to the yield strength from small to large), the upper limit of the transition of the thickness of the front and rear rolls, the upper limit of the transition of the width of the front and rear rolls, the upper limit of the transition of the roller gap of the front and rear rolls, the upper limit of the transition of the rolling force of the front and rear rolls, and the upper limit of the transition of the tension of the front and rear rolls are shown in Table 3;

[0045] Table 3 Knowledge Base - Planned Transition Threshold Construction Project

[0046]

[0047]

[0048] When different specifications of strip steel are connected, the upper and lower limits of weld quality parameters and the hit rate are often based on manual experience, and no effective parameter specification standards have been formed. Therefore, the weld quality statistics of different specifications are conducted based on big data statistics. According to a large amount of actual weld quality data, the upper and lower limits of weld quality parameters and the hit rate under different specifications transition conditions are statistically analyzed, as shown in Table 4.

[0049] Table 4 Knowledge base - weld quality threshold construction items

[0050]

[0051] Different steel grades have different tolerances for different types of surface defects. It is necessary to collect statistics on the lower limits of the areas of typical defects (such as edge cracks and scars) corresponding to different steel grades, as shown in Table 5.

[0052] Table 5 Knowledge base - surface defect threshold construction project

[0053]

[0054] When different steel grades are connected, belt breakage often occurs accidentally. The reason is that different operators intervene in the rolling parameters when passing the weld. The system collects the rolling parameters of the cases without belt breakage as excellent parameters, see Table 6, to guide operators to intervene in advance and reduce the probability of belt breakage.

[0055] Table 6 Knowledge Base - Excellent Rolling Parameters Project

[0056]

[0057] 2) When a belt break occurs, the system automatically collects and processes the relevant data of the belt break, performs a comprehensive belt break analysis based on the knowledge base, and gives the belt break analysis conclusion and parameter optimization suggestions;

[0058] When a belt break occurs, the system automatically integrates and processes relevant data, and analyzes the cause of the break according to the following steps based on the threshold determined by the knowledge base:

[0059] (1) First, it is necessary to monitor the L1 strip-breaking signal of the basic automation level of pickling mill in real time. The rack number where the strip-breaking signal occurs is used to determine whether the steel coil numbers before and after the rack are the same based on the strip position tracking. If they are the same, it is a non-weld strip-breaking. If they are different, it is identified as a weld strip-breaking. Based on the weld tracking signal, it is determined whether the strip-breaking position is before or after the weld and the distance from the weld.

[0060] (2) Obtain L1-level rolling parameters related to strip breakage, including actual values ​​of rolling force, tension, roll gap, and speed within 10 seconds before and after the strip breakage, with a data frequency of 50 ms, and analyze whether there are abnormalities in the rolling parameters. Analyze whether the rolling parameters 5 to 10 seconds before the strip breakage are abnormal. The judgment basis is the threshold value stored in the knowledge base.

[0061] Obtain the upstream hot-rolled raw material size and temperature strip length sequence data corresponding to the coil, requiring the data accuracy to reach 1 point per meter, and analyze whether the hot-rolled raw material position data corresponding to the broken strip position is abnormal. The analysis is based on the raw material threshold stored in the knowledge base;

[0062] (3) If the weld is broken, analyze whether the steel grade transition, thickness transition (including nominal thickness and actual thickness at the head and tail), and width transition (including nominal width and actual width at the head and tail) of the front and rear rolls are abnormal. The analysis is based on the transition parameter thresholds corresponding to different steel grades stored in the knowledge base;

[0063] (4) If the weld is broken, it is necessary to perform threshold determination on the transition of the rolling force setting value, the tension setting value, the roll gap setting value, and the speed setting value of the front and rear coils, also based on the thresholds stored in the knowledge base;

[0064] If the weld is broken, the weld quality detection data of the roll is obtained by communicating with the weld monitoring system of the welding machine, and the weld quality is analyzed based on the corresponding weld quality parameter threshold in the knowledge base;

[0065] (5) If the strip is not broken in the weld, obtain the hot rolling surface inspection data to determine whether there is a serious defect at the corresponding broken strip position, and perform abnormal diagnosis based on the defect area threshold stored in the knowledge base;

[0066] (6) Through a comprehensive analysis of the above rolling parameters and raw material quality, planned transition, weld quality, and surface defects, a conclusion on the broken strip is given, including the cause of the broken strip and a comparison with the case without broken strip.

[0067] 3) Before strip production, the risk of strip breakage in the production plan is predicted based on the knowledge base, and the corresponding optimization parameters and production scheduling suggestions are matched;

[0068] The belt break warning is to warn whether there is a belt break risk based on the knowledge base after the production plan is issued before production, so as to reduce the probability of belt break, and specifically includes the following steps:

[0069] (1) Receive production plan data issued by the manufacturing execution system L3;

[0070] (2) based on the broken strip knowledge base, abnormal identification is performed on the hot-rolled raw material size and temperature data corresponding to the coil;

[0071] (3) identifying abnormal defects in the hot rolling table inspection data corresponding to the coil based on the broken strip knowledge base;

[0072] Hot-rolled raw material audit: First, determine whether the head and tail need to be swapped, and determine whether the hot-rolled coil has been repaired. If it has not been repaired, the head and tail of the hot-rolled production should be replaced with the tail and head of the cold-rolled production. If the number of repairs is an even number, the head and tail of the data also need to be swapped;

[0073] Calculate the hit rate of strip quality data: use the actual value of the strip and the upper and lower thresholds saved in the knowledge base to calculate, then count the number of qualified and unqualified meters, and the qualified meters / total length is the hit rate;

[0074] Hot rolling surface defect audit:

[0075] Make abnormal judgment on hot rolling surface quality data based on the knowledge base;

[0076] (4) Conduct transition review between each roll and the next roll in the production plan, use the rolling force model to calculate the rolling force difference, roll gap difference, tension difference, and speed difference between the two rolls, and issue an early warning of the risk of strip breakage based on the strip breakage knowledge base;

[0077] Production plan review:

[0078] Periodically query the production plan number, serial number, cold coil number, and hot coil number of the three-level production plan table, create a new cache table, and compare the above field data with the cache table. Use a joint query method to determine whether there are any changes. Changes include adding, deleting, changing the plan number, and changing the sequence number. When changes are found, obtain the specifications of the front and rear coils of the changed steel, read the thickness, width, steel type information of the front and rear coils, and the corresponding raw material curve data and rolling model setting values. Compare the transition rules in the knowledge base and conduct a transition review of the entire production plan according to the sequence of the production plan numbers;

[0079] (5) Based on the above analysis, the final belt break risk level is given. If there are unqualified raw materials and planned transitions, a mark prompt is given in the production plan table. The risk level is represented by color, green means no risk, yellow means the risk is less than 50%, and red means the risk is greater than 50%;

[0080] (6) For production sequences with the risk of belt breakage, matching is performed according to the knowledge base, and optimal parameters or production scheduling suggestions are given to avoid belt breakage;

[0081] (7) When the defect position reaches the entrance uncoiler and 50 meters in front of the rolling mill (50 meters or within 50 meters), an early warning is issued again.

[0082] The main purpose of strip break warning is to judge whether the strip quality meets the standard, whether there are major surface defects and whether the raw material transition is qualified before the strip is produced. The risk of strip breakage for each coil is predicted through data such as the head and tail tolerance of raw material data, raw material plate shape, transition rules, and meter inspection instrument data. At the same time, it analyzes whether the transition of the secondary set value is abnormal and whether there is a risk of excessive transition of the set value.

[0083] The following examples are implemented on the premise of the technical solution of the present invention, and provide detailed implementation methods and specific operation processes, but the protection scope of the present invention is not limited to the following examples. The methods used in the following examples are conventional methods unless otherwise specified.

[0084] [Example 1]

[0085] An embodiment of the present invention provides a knowledge base-based pickling mill belt break analysis and early warning method.

[0086] In a steel plant's 2130 cold rolling production line, a weld fracture occurred when producing the front steel H0A27053000000 and the rear steel H0A29004000000, and the fracture location was 3-4 rack;

[0087] Belt break warning:

[0088] When the plan is issued, the model conducts quality review and transition review on the raw materials in the production plan, and prompts that the steel grade and export thickness transition are not suitable, as shown in Table 7:

[0089] Table 7 Basic information of steel before and after the belt break

[0090]

[0091] At 18:44:49 on October 30, the uncoiler on the hot coil gave an abnormal alarm for steel grade transition: "Planned steel grade jump is not allowed, the previous steel coil number = 'H0A27053000000'".

[0092] At 18:52:44 on October 30, a voice warning was issued that the actual steel grade transition was unqualified: "The actual steel grade transition was unqualified, the entry volume number is H0A27053000000".

[0093] Belt break analysis:

[0094] After tracking the weld into frame 3, the tension between frames 3-4 dropped from 38 tons to 0.43 tons within 2 seconds, and then triggered a belt break signal. After the belt break occurred, the model automatically performed a belt break analysis.

[0095] Within 1 second before the strip broke, the actual value of the rolling force changed from 1,368 tons to 679 tons, a change of 689 tons, which exceeded the threshold given in the knowledge base.

[0096] According to the above data analysis, the fundamental reason for the strip breakage was the abnormal transition of steel grades, which resulted in a large transition change in the actual rolling force of the 3rd stand during the weld stage, causing the strip to break near the weld.

[0097] [Example 2]

[0098] When producing the front steel H0816106000000 and the rear steel H0810078000000, the weld was broken at the 4-5 frame.

[0099] Belt break warning:

[0100] When the plan is issued, the model conducts quality review and transition review on the raw materials in the production plan, and prompts that the steel grade and export thickness transition are not suitable, as shown in Table 8:

[0101] Table 8 Basic information of steel running before and after the belt break

[0102]

[0103] Belt break analysis:

[0104] After the weld entered the S4 frame, the tension dropped from 45 tons to 0.8 tons within 2 seconds, triggering the strip break signal. Before the strip broke, the rolling force fluctuated greatly, about 200 tons, but this large fluctuation was due to the rolling of high-strength steel, not the cause of the strip break.

[0105] The weld quality was unqualified: the hit rate of the front and rear height difference index was 85.56%, which was not up to standard and was the main reason for the breakage.

[0106] See Figure 1-2, by counting the relevant belt breakage data since November 2020, the overall number of belt breaks has shown a downward trend, and the number of belt breaks has decreased by more than 50% compared with before. The reason is that the system predicts the risk level of belt breaks through production plan review and gives voice alarm prompts. From November 2020 to March 2021, a total of 144 belt breaks were recorded, of which 35 were caused by procedure transition, accounting for 24.3% of the total belt breaks. Among the 35 cases of procedure transition belt breaks, 33 voice warnings were given to slow down, which fully demonstrated that the model has achieved certain results. With the gradual operation of the model, the voice alarm has played a certain role after April, and the procedure transition belt breaks have decreased month by month. At the same time, the number of high-strength steel belt breaks has also been greatly reduced. First, through the production plan review, the unqualified materials were reworked. Secondly, the optimal unit tension and the reduction rate of each frame were given through the knowledge base to solve the problems of high-strength steel punching and high-strength steel weld belt breaks.

Claims

1. A method for analyzing and warning of belt breakage in a pickling mill, characterized in that: It includes the following: 1) Based on the mechanism of band breaking, the band breaking is classified into weld band breaking and non-weld band breaking. Each category is further divided into different steel grades. The band breaking characteristic values ​​are counted according to different types, and a band breaking knowledge base is constructed, which is updated in real time. 2) When a belt break occurs, the system automatically collects and processes the relevant data of the belt break, performs a comprehensive belt break analysis based on the knowledge base, and gives the belt break analysis conclusion and parameter optimization suggestions; 3) Before strip production, the risk of strip breakage in the production plan is predicted based on the knowledge base, and the corresponding optimization parameters and production scheduling suggestions are matched.

2. The method for analyzing and warning of belt breakage in a pickling mill according to claim 1, characterized in that: The specific construction process of the broken belt knowledge base includes the following steps: (1) In the initial stage of system construction, the system is configured according to the system's own knowledge base, according to the warning thresholds of hot-rolled raw material related parameters, production plan transition warning thresholds, weld quality hit rate thresholds, surface defect warning thresholds, and rolling parameter thresholds; (2) When the weld is broken, if the steel grades of the previous and next steel coils are different, historical data are collected to obtain the minimum thickness difference and temperature difference between the head of the previous coil and the tail of the next coil corresponding to the upstream hot-rolled raw material under the same steel grade transition condition, which is used as the early warning threshold for the hot-rolled raw material breakage; (3) When the weld is broken, if the steel grades of the front and rear coils are different, historical data are collected to obtain the same steel grade transition condition. The front and rear steel grades are combined to record the corresponding minimum transition rolling force, transition tension, and transition roll gap as the break warning threshold. At the same time, the historical data of the relevant steel grade transition without breaking is obtained, and the corresponding unit tension and the reduction rate of each stand are recorded as optimization parameter recommendations; (4) When the weld is broken, the difference between the weld quality parameters at the time of breakage and the weld quality parameters without breakage is obtained through historical data statistics, including weld penetration rate, weld height difference, weld reinforcement height, and weld incomplete filling. The minimum hit rate and the minimum standard deviation of each parameter are used as the breakage warning threshold; (5) When the non-weld seam breaks, the acid rolling surface inspection data is obtained. At the same time, based on historical big data statistics, the minimum defect area caused by the same type of defect is used as the surface defect warning threshold of the broken strip.

3. The method for analyzing and warning of belt breakage in a pickling mill according to claim 1, characterized in that: The broken belt analysis comprises the following steps: (1) Real-time monitoring of the L1 strip-breaking signal of the basic automation level of pickling mill. Based on the rack number where the strip-breaking signal occurs and the strip position tracking, it is determined whether the steel coil numbers before and after this rack are the same. If they are the same, it is a non-weld strip break. If they are different, it is identified as a weld strip break. (2) Obtain the upstream hot-rolled raw material size and temperature strip length sequence data corresponding to the coil, with the data accuracy required to reach 0.5 to 1 meter per point, and analyze whether the hot-rolled raw material position data corresponding to the broken strip position is abnormal based on the threshold value of the corresponding parameter in the knowledge base; (3) If the weld is broken, the weld quality detection data of the roll is obtained by communicating with the weld monitoring system of the welding machine, and the weld quality is analyzed; (4) If the belt is broken at a non-weld seam, obtain the acid rolling surface inspection data and determine whether there is a serious defect at the corresponding belt break position according to the threshold of the corresponding parameter in the knowledge base; (5) Obtain other L1-level rolling parameters related to strip breakage, including rolling force, tension, roll gap, speed setting value and actual value of the front and rear coils. The data accuracy is required to reach a frequency of 10 to 50 ms. Analyze whether there is an abnormality in the rolling parameters based on the threshold value of the corresponding parameters in the knowledge base; (6) Based on the above analysis, the conclusion of the broken belt analysis is given.

4. The method for analyzing and warning of belt breakage in a pickling mill according to claim 1, characterized in that: The belt break warning is carried out after the production plan is issued before production, and specifically includes the following steps: (1) Receive production plan data issued by the manufacturing execution system L3; (2) based on the broken strip knowledge base, abnormal identification is performed on the hot-rolled raw material size and temperature data corresponding to the coil; (3) identifying abnormal defects in the hot rolling table inspection data corresponding to the coil based on the broken strip knowledge base; (4) Conduct transition review between each roll and the next roll in the production plan, use the rolling force model to calculate the rolling force difference, roll gap difference, tension difference, and speed difference between the two rolls, and issue an early warning of the risk of strip breakage based on the strip breakage knowledge base; (5) Based on the above analysis, the final belt break risk level is given. The risk level is represented by color, with green indicating no risk, yellow indicating a risk of less than 50%, and red indicating a risk of more than 50%; (6) For production sequences with the risk of belt breakage, matching is performed according to the knowledge base, and optimal parameters or production scheduling suggestions are given to avoid belt breakage; (7) When the defect location reaches the entrance uncoiler and 50 meters in front of the rolling mill, an early warning is issued again.

Images