Slab defect monitoring system and method based on Internet of Things

Abstract

The invention discloses a slab defect monitoring system and method based on the Internet of Things, relates to the technical field of slab defect monitoring, and solves the technical problem that slab defect monitoring is not comprehensive or accurate enough in the prior art. The method comprises the following steps: acquiring production data of a slab and environmental data of a slab production workshop; calculating a first index coefficient based on the first index data; calculating a second index coefficient based on the second index data; calculating an environment index coefficient based on the environment data; based on the first index coefficient, the second index coefficient and the environment index coefficient, a slab production quality evaluation coefficient is obtained through calculation; training an artificial intelligence model based on the historical image data to obtain an image defect recognition model, and recognizing the image data based on the image defect recognition model to obtain a recognition result; whether early warning information is sent or not is judged based on the slab production quality evaluation coefficient; and judging whether to send early warning information based on the identification result. The technical problem is solved.

Description

Technical Field

[0001] The present invention belongs to the field of slab defect monitoring, and specifically relates to a slab defect monitoring system and method based on the Internet of Things. Background Art

[0002] In the steel production process, slabs are important semi-finished products formed after continuous casting. The quality of slabs is directly related to the performance and cost of the final product. Traditionally, defect detection of slabs relies on manual visual inspection or offline sampling, which is not only inefficient and labor-intensive, but also difficult to ensure the qualified rate of products.

[0003] The prior art (invention application with application number 2023117752349) discloses a slab production process monitoring system based on digital Li Sheng; the system includes a data acquisition module, a defect recognition module and a production management module; the data acquisition module is used to obtain the slab image of the slab to be detected during the production process and the slab production environment parameters, and the slab production environment parameters include temperature, humidity and light; the defect recognition module has a pre-trained defect recognition neural network built in, which is used to use the slab image obtained by the data acquisition module as input to obtain the defect detection result on the slab to be detected; the defect detection result includes the defect category and the defect location; the production management module A built-in defect probability prediction model; the defect probability prediction model is used to use the slab production environment parameters obtained by the data acquisition module as input to predict the probability of defects in the slab in the current production environment; this invention application can realize the functions of monitoring the slab production process, defect identification, production management, defect prediction, etc.; but the existing technology mainly relies on limited environmental parameters (such as temperature, humidity and light) and image data for defect detection and probability prediction, and fails to cover the key process parameters that affect the quality of the slab (such as casting temperature, cooling rate, pulling speed, etc.), so that the defect monitoring of the slab is not comprehensive enough, resulting in the problem that the monitoring results of slab defects are not accurate enough.

[0004] Therefore, the present invention proposes a slab defect monitoring system and method based on the Internet of Things to solve the above problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a slab defect monitoring system and method based on the Internet of Things, which is used to solve the technical problem that the prior art mainly relies on limited environmental parameters and image data for defect detection and probability prediction, fails to cover the key process parameters affecting the quality of the slab, and thus the defect monitoring of the slab is not comprehensive enough, resulting in inaccurate monitoring results of the slab defects.

[0006] To achieve the above-mentioned object, the first aspect of the present invention provides a slab defect monitoring system based on the Internet of Things, comprising: a data acquisition module, a data analysis module, an image analysis module and a monitoring and early warning module;

[0007] Data acquisition module: used for real-time acquisition of slab production data and environmental data of slab production workshop; wherein the production data includes: first indicator data, second indicator data and image data;

[0008] Data analysis module: calculates a first indicator coefficient based on the first indicator data; calculates a second indicator coefficient based on the second indicator data; calculates an environmental indicator coefficient based on the environmental data; and,

[0009] The slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient;

[0010] Image analysis module: trains an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizes image data based on the image defect recognition model to obtain recognition results;

[0011] Monitoring and early warning module: determines whether to send early warning information based on the slab production quality assessment coefficient; determines whether to send early warning information based on the identification result.

[0012] Preferably, the collecting of the production data of the slab and the environmental data of the slab production workshop includes:

[0013] The first index data, the second index data and the image data of the slab are collected in real time by the data collection device; wherein the first index data includes: the pouring temperature and the cooling rate; the second index data includes: the pulling speed and the vibration frequency; the image data refers to the image of the slab surface;

[0014] The environmental data of the slab production workshop is collected in real time through data sensors; the environmental data includes: ambient temperature, ambient humidity and dust concentration.

[0015] It should be noted that the cooling rate is calculated by collecting the temperature when the liquid steel begins to enter the crystallizer, the temperature when it is cooled to form an initial solid shell, and the time required for the liquid steel to enter the crystallizer and cool to form an initial solid shell;

[0016] That is, cooling rate = (temperature of liquid steel when it starts to enter the crystallizer - temperature when it cools to form the initial solid shell) / time required for the liquid steel to enter the crystallizer and cool to form the initial solid shell;

[0017] Pouring temperature refers to the temperature of molten steel when it enters the crystallizer, which directly affects the fluidity, solidification behavior and microstructure of the final product; pulling speed refers to the speed at which the slab passes through the crystallizer, which determines the residence time and solidification process of the molten steel in the crystallizer; cooling rate refers to the speed at which the molten steel drops from high temperature to low temperature, which directly affects the solidification behavior, microstructure and final performance of the slab; the vibration frequency of the crystallizer is used to prevent adhesion between the billet and the crystallizer and promote uniform distribution of the lubricant.

[0018] Preferably, the first indicator coefficient is calculated based on the first indicator data, including:

[0019] Label the pouring temperature as JW and the cooling rate as LV;

[0020] The first indicator coefficient is calculated by the first indicator calculation formula D1 = α1×e^[((JW-ZJW)^2) / ZJW^2]+α2×e^[((LV-ZLV)^2) / ZLV^2]; wherein D1 is the first indicator coefficient, α1 and α2 are proportional adjustment coefficients, and 0<α1<1, 0<α2<1, ZJW is the optimal pouring temperature, and ZJW is the optimal cooling rate.

[0021] It should be noted that the proportional adjustment coefficient is set by experts in this field based on experience according to the production conditions of specific slabs;

[0022] The first indicator data is directly related to the solidification behavior, microstructure and final performance of the slab, and has a significant impact on the internal quality and mechanical properties of the slab. The larger the first indicator coefficient, the greater the defect risk of the slab.

[0023] The pouring temperature determines the fluidity of the molten steel, the solidification rate and the microstructure of the final product;

[0024] The cooling rate directly affects the solidification process, thermal stress distribution and microstructure of the slab;

[0025] Too high a pouring temperature will slow down the cooling rate, increase thermal stress, easily cause cracks and other heat treatment defects, and intensify the oxidation reaction when the molten steel contacts the air, forming more oxide scale and affecting the surface quality; too low a pouring temperature will lead to uneven pouring and even pouring interruption, affecting the geometric shape and dimensional accuracy of the slab;

[0026] Too fast a cooling rate will cause greater thermal stress, especially in thick slabs, which can easily cause cracks, deformation and other problems; rapid cooling may lead to grain refinement, which can sometimes improve strength, but may also reduce the toughness of the material and increase the risk of brittle fracture; too slow a cooling rate will prolong the solidification process, increase thermal stress accumulation, and easily cause cracks or other heat treatment defects. At the same time, slow cooling may cause the surface oxide scale to thicken, affecting the surface finish.

[0027] Preferably, the calculating the second indicator coefficient based on the second indicator data includes:

[0028] Mark the pulling speed as S and the vibration frequency as F;

[0029] The second indicator coefficient is calculated by the second indicator calculation formula D2=ln[1+(S-ZS)^2]^β1+ln[1+(F-ZF)^2]^β2; wherein D2 is the second indicator coefficient, β1 and β2 are proportional adjustment coefficients, and 0<β1<1, 0<β2<1, ZS is the optimal pulling speed, ZF is the optimal vibration frequency, and ln(*) is a logarithmic function with the natural number e as the base.

[0030] It should be noted that the second indicator data mainly affects the operating conditions, equipment performance and surface quality in the production process. Although it indirectly affects the quality of the slab, it is crucial in ensuring smooth production and surface quality. The larger the second indicator coefficient, the greater the defect risk of the slab.

[0031] The pulling speed determines the residence time of the molten steel in the crystallizer and the speed of the solidification process.

[0032] The vibration frequency is mainly used to prevent adhesion between the billet and the crystallizer and to promote uniform distribution of the lubricant;

[0033] Too fast a pulling speed will shorten the residence time of the molten steel in the crystallizer, resulting in insufficient thickness of the solidified shell and increasing the risk of perforation or fracture. Too slow a pulling speed will cause the molten steel to stay in the crystallizer for too long, resulting in local overcooling, increased thermal stress concentration, and easy cracking.

[0034] Too high a vibration frequency will cause defects such as ripples and scratches on the surface of the slab. At the same time, strong vibration will increase mechanical stress and cause microcracks on the surface and inside of the slab. Insufficient vibration will cause adhesion between the billet and the crystallizer, increase problems such as sand sticking and pits. At the same time, low amplitude may not be able to effectively distribute the lubricant, resulting in poor lubrication effect, further aggravating the adhesion problem.

[0035] Preferably, the environmental index coefficient is calculated based on the environmental data, including:

[0036] The ambient temperature is marked as W, the ambient humidity is marked as H, and the dust concentration is marked as C;

[0037] The environmental index coefficient is calculated by the environmental index calculation formula D3 = θ1×e^[((W-ZW)^2) / ZW^2]+ln[((H+1)^θ2)×((C+1)^θ3)]; wherein D3 is the environmental index coefficient, θ1, θ2, θ3 are proportional adjustment coefficients, and 0<θ1<1, 0<θ2<1, 0<θ3<1, and ZW is the optimal ambient temperature of the workshop during the slab production process.

[0038] It should be noted that the ambient temperature of the production workshop will affect the cooling rate and solidification behavior of the molten steel. If the workshop temperature is too high, the slab surface will cool slowly, increasing the risk of thermal stress, which will cause cracks or deformation. On the contrary, too low an ambient temperature may accelerate cooling and cause internal defects such as shrinkage and porosity.

[0039] In a high humidity environment, moisture in the air easily reacts with the surface of the slab, accelerating the formation of oxide scale, which not only affects the surface quality of the slab, but may also fall off during subsequent processing, causing pollution or damage to other equipment;

[0040] Dust concentration will cause dust particles to adhere to the surface of the slab, forming pollutants. These particles not only affect the appearance quality of the slab, but may also fall off during subsequent processing, causing secondary pollution or damage to other equipment.

[0041] Preferably, the slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient, including:

[0042] Extracting a first index coefficient D1, a second index coefficient D2 and an environmental index coefficient D3;

[0043] The slab production quality assessment coefficient is calculated by the formula P=A1×D1+A2×D2+A3×D3; wherein P is the slab production quality assessment coefficient, A1, A2, A3 are weight coefficients, and A1+A2+A3=1.

[0044] It should be noted that the weight coefficient is set by experts in this field based on experience.

[0045] Preferably, the training of an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizing image data based on the image defect recognition model, includes:

[0046] Extracting images of slab surfaces from historical image data;

[0047] Based on the image of the slab surface in the historical image data, the corresponding slab surface defect type is matched from the database; wherein the slab surface defect types include: normal, surface cracks, scratches, pits, bubbles, etc.; the database is an existing data set storing various slab surface defect types;

[0048] Integrate the images of the slab surface in the historical image data as standard input data, and integrate the types of slab surface defects corresponding to the images of the slab surface in the historical image data as standard output data;

[0049] The artificial intelligence model is trained based on standard input data and standard output data to obtain an image defect recognition model;

[0050] The real-time collected image data is input into the image defect recognition model to obtain the type of slab surface defect corresponding to the image data and mark it as the recognition result.

[0051] Preferably, the determining whether to send warning information based on the slab production quality assessment coefficient includes:

[0052] Determine whether the slab production quality assessment coefficient is greater than the preset quality threshold; if yes, generate an early warning message and send it to the client; if not, continue to monitor and judge.

[0053] It should be noted that the preset quality threshold is set by experts in this field based on experience.

[0054] Preferably, judging whether to send warning information based on the recognition result includes:

[0055] Extract the types of slab surface defects from the recognition results;

[0056] Determine whether the type of slab surface defect is normal; if yes, continue monitoring and judging; if not, generate warning information and send the specific type of slab surface defect to the client.

[0057] A second aspect of the present invention provides a slab defect monitoring method based on the Internet of Things, comprising:

[0058] Step 1: Collect the production data of slabs and the environmental data of slab production workshops;

[0059] Step 2: Calculate the first indicator coefficient based on the first indicator data; calculate the second indicator coefficient based on the second indicator data; calculate the environmental indicator coefficient based on the environmental data;

[0060] Step 3: Calculate the slab production quality assessment coefficient based on the first index coefficient, the second index coefficient and the environmental index coefficient;

[0061] Step 4: Train an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognize the image data based on the image defect recognition model to obtain a recognition result;

[0062] Step 5: Determine whether to send a warning message based on the slab production quality assessment coefficient; determine whether to send a warning message based on the recognition result.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] 1. The present invention collects slab production data and environmental data of a slab production workshop; calculates a first indicator coefficient based on the first indicator data; calculates a second indicator coefficient based on the second indicator data; calculates an environmental indicator coefficient based on the environmental data; calculates a slab production quality assessment coefficient based on the first indicator coefficient, the second indicator coefficient and the environmental indicator coefficient; trains an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizes image data based on the image defect recognition model to obtain a recognition result; determines whether to send an early warning message based on the slab production quality assessment coefficient; determines whether to send an early warning message based on the recognition result, thereby solving the technical problem that the prior art mainly relies on limited environmental parameters and image data for defect detection and probability prediction, fails to cover the key process parameters that affect the quality of the slab, and thus the defect monitoring of the slab is not comprehensive enough, resulting in inaccurate monitoring results of the slab defects.

[0065] 2. The present invention comprehensively evaluates the quality risks in the slab production process by comprehensively calculating the first indicator coefficient, the second indicator coefficient and the environmental indicator coefficient, and generates a comprehensive slab production quality assessment coefficient; this multi-dimensional data analysis method not only takes into account the key process parameters that directly affect the internal quality and mechanical properties of the slab, but also covers the influence of production environment factors on the surface quality of the slab; specifically, the first indicator coefficient reflects the influence of casting temperature and cooling rate on the solidification behavior and microstructure of the slab; the second indicator coefficient reveals the effect of pulling speed and vibration frequency on production operating conditions and equipment performance; the environmental indicator coefficient quantifies the potential threat of workshop environment to the surface quality of the slab; by weighted summing up these coefficients to obtain a comprehensive assessment coefficient, the quality status of the slab under the current production conditions can be accurately judged, and early warning information can be issued in time, thereby realizing real-time monitoring and precise management of slab production; this method significantly improves the controllability of the production process and the consistency of product quality, effectively reduces the risk of cost increase due to quality problems, and improves overall production efficiency and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0067] Figure 1 A schematic diagram of a system module according to an embodiment of the present invention;

[0068] Figure 2 Schematic diagram of the method steps of an embodiment of the present invention. DETAILED DESCRIPTION

[0069] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] See also Figure 1 , the first aspect of the present invention provides a slab defect monitoring system based on the Internet of Things, including: a data acquisition module, a data analysis module, an image analysis module and a monitoring and early warning module;

[0071] Data acquisition module: used for real-time acquisition of slab production data and environmental data of slab production workshop; wherein the production data includes: first indicator data, second indicator data and image data;

[0072] Data analysis module: calculates a first indicator coefficient based on the first indicator data; calculates a second indicator coefficient based on the second indicator data; calculates an environmental indicator coefficient based on the environmental data; and,

[0073] The slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient;

[0074] Image analysis module: trains an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizes image data based on the image defect recognition model to obtain recognition results;

[0075] Monitoring and early warning module: determines whether to send early warning information based on the slab production quality assessment coefficient; determines whether to send early warning information based on the identification result.

[0076] Collect slab production data and slab production workshop environmental data, including:

[0077] The first index data, the second index data and the image data of the slab are collected in real time by the data collection device; wherein the first index data includes: the pouring temperature and the cooling rate; the second index data includes: the pulling speed and the vibration frequency; the image data refers to the image of the slab surface;

[0078] The environmental data of the slab production workshop is collected in real time through data sensors; the environmental data includes: ambient temperature, ambient humidity and dust concentration.

[0079] The first indicator coefficient is calculated based on the first indicator data, including:

[0080] Label the pouring temperature as JW and the cooling rate as LV;

[0081] The first indicator coefficient is calculated by the first indicator calculation formula D1 = α1×e^[((JW-ZJW)^2) / ZJW^2]+α2×e^[((LV-ZLV)^2) / ZLV^2]; wherein D1 is the first indicator coefficient, α1 and α2 are proportional adjustment coefficients, and 0<α1<1, 0<α2<1, ZJW is the optimal pouring temperature, and ZJW is the optimal cooling rate.

[0082] The second indicator coefficient is calculated based on the second indicator data, including:

[0083] Mark the pulling speed as S and the vibration frequency as F;

[0084] The second indicator coefficient is calculated by the second indicator calculation formula D2=ln[1+(S-ZS)^2]^β1+ln[1+(F-ZF)^2]^β2; wherein D2 is the second indicator coefficient, β1 and β2 are proportional adjustment coefficients, and 0<β1<1, 0<β2<1, ZS is the optimal pulling speed, ZF is the optimal vibration frequency, and ln(*) is a logarithmic function with the natural number e as the base.

[0085] Environmental index coefficients are calculated based on environmental data, including:

[0086] The ambient temperature is marked as W, the ambient humidity is marked as H, and the dust concentration is marked as C;

[0087] The environmental index coefficient is calculated by the environmental index calculation formula D3 = θ1×e^[((W-ZW)^2) / ZW^2]+ln[((H+1)^θ2)×((C+1)^θ3)]; wherein D3 is the environmental index coefficient, θ1, θ2, θ3 are proportional adjustment coefficients, and 0<θ1<1, 0<θ2<1, 0<θ3<1, and ZW is the optimal ambient temperature of the workshop during the slab production process.

[0088] The slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient, including:

[0089] Extracting a first index coefficient D1, a second index coefficient D2 and an environmental index coefficient D3;

[0090] The slab production quality assessment coefficient is calculated by the formula P=A1×D1+A2×D2+A3×D3; wherein P is the slab production quality assessment coefficient, A1, A2, A3 are weight coefficients, and A1+A2+A3=1.

[0091] An artificial intelligence model is trained based on historical image data to obtain an image defect recognition model, and image data is recognized based on the image defect recognition model, including:

[0092] Extracting images of slab surfaces from historical image data;

[0093] Based on the image of the slab surface in the historical image data, the corresponding slab surface defect type is matched from the database; wherein the slab surface defect types include: normal, surface cracks, scratches, pits, bubbles, etc.; the database is an existing data set storing various slab surface defect types;

[0094] Integrate the images of the slab surface in the historical image data as standard input data, and integrate the types of slab surface defects corresponding to the images of the slab surface in the historical image data as standard output data;

[0095] The artificial intelligence model is trained based on standard input data and standard output data to obtain an image defect recognition model;

[0096] The real-time collected image data is input into the image defect recognition model to obtain the type of slab surface defect corresponding to the image data and mark it as the recognition result.

[0097] Determine whether to send warning information based on the slab production quality assessment coefficient, including:

[0098] Determine whether the slab production quality assessment coefficient is greater than the preset quality threshold; if yes, generate an early warning message and send it to the client; if not, continue to monitor and judge.

[0099] Determine whether to send warning information based on the recognition results, including:

[0100] Extract the types of slab surface defects from the recognition results;

[0101] Determine whether the type of slab surface defect is normal; if yes, continue monitoring and judging; if not, generate warning information and send the specific type of slab surface defect to the client.

[0102] For example, suppose you are monitoring the slab production process on a steel production line and judging whether there is a risk of defects based on the data. The specific steps are as follows:

[0103] Step 1: Collect the production data of slabs and the environmental data of slab production workshops;

[0104] Assume that the following data is obtained through sensors and cameras:

[0105] The first indicator data: the pouring temperature is 1550℃ and the cooling rate is 50℃ / min.

[0106] The second indicator data: pulling speed is 2.5m / min, vibration frequency is 120Hz.

[0107] Image data: High-definition images of the slab surface are acquired through a high-resolution camera.

[0108] Environmental data: ambient temperature is 30°C, ambient humidity is 60%, and dust concentration is 0.1mg / m 3 .

[0109] Step 2: Calculate the first indicator coefficient based on the first indicator data;

[0110] The first indicator coefficient is calculated by the first indicator calculation formula D1 = α1×e^[((JW-ZJW)^2) / ZJW^2]+α2×e^[((LV-ZLV)^2) / ZLV^2];

[0111] Assume that the proportional adjustment coefficients are: α1 = 0.6, α2 = 0.4;

[0112] Assume that the optimum pouring temperature (ZJW) is 1500°C and the optimum cooling rate (ZLV) is 40°C / min;

[0113] D1=0.6×e^[((1550-1500)^2) / 1500^2]+0.4×e^[((50-40)^2) / 40^2]≈1.03;

[0114] Step 3: Calculate the second indicator coefficient based on the second indicator data;

[0115] The second indicator coefficient is calculated by the second indicator calculation formula D2=ln[1+(S-ZS)^2]^β1+ln[1+(F-ZF)^2]^β2;

[0116] Assume that the proportional adjustment coefficients β1 = 0.7, β2 = 0.3;

[0117] Assume that the optimum pulling speed (ZS) is 2.0 m / min and the optimum vibration frequency (ZF) is 100 Hz;

[0118] D2=ln[1+(2.5-2.0)^2]^0.7+ln[1+(120-100)^2]^0.3≈2.32;

[0119] Step 4: Calculate the environmental index coefficient based on the environmental data;

[0120] The environmental index coefficient is calculated by the environmental index calculation formula D3 = θ1 × e^[((W-ZW)^2) / ZW^2]+ln[((H+1)^θ2)×((C+1)^θ3)];

[0121] Assume that the proportional adjustment coefficients are: θ1 = 0.5, θ2 = 0.3, θ3 = 0.2;

[0122] Assume that the optimum ambient temperature (ZW) is 25°C;

[0123] D3=0.5×e^[((30-25)^2) / 25^2]+ln[((60+1)^0.3)×((0.1+1)^0.2)]≈1.91;

[0124] Step 5: Calculate the slab production quality assessment coefficient based on the first index coefficient, the second index coefficient and the environmental index coefficient;

[0125] The slab production quality assessment coefficient is calculated by the formula P = A1 × D1 + A2 × D2 + A3 × D3;

[0126] Assume weight coefficients: A1 = 0.4, A2 = 0.3, A3 = 0.3;

[0127] P=0.4×1.03+0.3×2.32+0.3×1.91=1.681;

[0128] Therefore, the slab production quality assessment coefficient is 1.681.

[0129] Step 6: Train an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognize the image data based on the image defect recognition model to obtain a recognition result;

[0130] Assume that the AI ​​model trained on historical image data has been completed and the following results are detected in the real-time acquired image data:

[0131] Identification result: The type of slab surface defect is scratches.

[0132] Step 7: Determine whether to send a warning message based on the slab production quality assessment coefficient; determine whether to send a warning message based on the recognition result

[0133] Assume that the preset quality threshold is 1.5;

[0134] The slab production quality assessment coefficient (P=1.681) is greater than the preset quality threshold of 1.5, and an early warning message is generated and sent to the client.

[0135] Identification result: The surface defect type of the slab is scratches, which is not a normal situation. An early warning message is generated and sent to the client.

[0136] See also Figure 2 The second aspect of the present invention provides a slab defect monitoring method based on the Internet of Things, comprising:

[0137] Step 1: Collect the production data of slabs and the environmental data of slab production workshops;

[0138] Step 2: Calculate the first indicator coefficient based on the first indicator data; calculate the second indicator coefficient based on the second indicator data; calculate the environmental indicator coefficient based on the environmental data;

[0139] Step 3: Calculate the slab production quality assessment coefficient based on the first index coefficient, the second index coefficient and the environmental index coefficient;

[0140] Step 4: Train an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognize the image data based on the image defect recognition model to obtain a recognition result;

[0141] Step 5: Determine whether to send a warning message based on the slab production quality assessment coefficient; determine whether to send a warning message based on the recognition result.

[0142] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0143] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The slab defect monitoring system based on the Internet of Things is characterized by: include: Data acquisition module, data analysis module, image analysis module and monitoring and early warning module; Data acquisition module: used for real-time acquisition of slab production data and environmental data of slab production workshop; wherein the production data includes: first indicator data, second indicator data and image data; Data analysis module: calculates a first indicator coefficient based on the first indicator data; calculates a second indicator coefficient based on the second indicator data; calculates an environmental indicator coefficient based on the environmental data; and, The slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient; Image analysis module: trains an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizes image data based on the image defect recognition model to obtain recognition results; Monitoring and early warning module: determines whether to send early warning information based on the slab production quality assessment coefficient; determines whether to send early warning information based on the identification result.

2. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The collecting of slab production data and slab production workshop environment data includes: The first index data, the second index data and the image data of the slab are collected in real time by the data collection device; wherein the first index data includes: the pouring temperature and the cooling rate; the second index data includes: the pulling speed and the vibration frequency; the image data is the image of the slab surface; The environmental data of the slab production workshop is collected in real time through data sensors; the environmental data includes: ambient temperature, ambient humidity and dust concentration.

3. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The step of calculating the first indicator coefficient based on the first indicator data includes: Label the pouring temperature as JW and the cooling rate as LV; The first indicator coefficient is calculated by the first indicator calculation formula D1 = α1×e^[((JW-ZJW)^2) / ZJW^2]+α2×e^[((LV-ZLV)^2) / ZLV^2]; wherein D1 is the first indicator coefficient, α1 and α2 are proportional adjustment coefficients, and 0<α1<1, 0<α2<1, ZJW is the optimal pouring temperature, and ZJW is the optimal cooling rate.

4. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The step of calculating the second indicator coefficient based on the second indicator data includes: Mark the pulling speed as S and the vibration frequency as F; The second indicator coefficient is calculated by the second indicator calculation formula D2=ln[1+(S-ZS)^2]^β1+ln[1+(F-ZF)^2]^β2; wherein D2 is the second indicator coefficient, β1 and β2 are proportional adjustment coefficients, and 0<β1<1, 0<β2<1, ZS is the optimal pulling speed, ZF is the optimal vibration frequency, and ln(*) is a logarithmic function with the natural number e as the base.

5. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The environmental index coefficient is calculated based on the environmental data, including: The ambient temperature is marked as W, the ambient humidity is marked as H, and the dust concentration is marked as C; The environmental index coefficient is calculated by the environmental index calculation formula D3 = θ1×e^[((W-ZW)^2) / ZW^2]+ln[((H+1)^θ2)×((C+1)^θ3)]; wherein D3 is the environmental index coefficient, θ1, θ2, θ3 are proportional adjustment coefficients, and 0<θ1<1, 0<θ2<1, 0<θ3<1, and ZW is the optimal ambient temperature of the workshop during the slab production process.

6. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The slab production quality assessment coefficient is calculated based on the first index coefficient, the second index coefficient and the environmental index coefficient, including: Extracting a first index coefficient D1, a second index coefficient D2 and an environmental index coefficient D3; The slab production quality assessment coefficient is calculated by the formula P=A1×D1+A2×D2+A3×D3; wherein P is the slab production quality assessment coefficient, A1, A2, A3 are weight coefficients, and A1+A2+A3=1.

7. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The method of training an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognizing image data based on the image defect recognition model, includes: Extracting images of slab surfaces from historical image data; Based on the image of the slab surface in the historical image data, the corresponding slab surface defect type is matched from the database; the database is an existing data set storing various slab surface defect types; Integrate the images of the slab surface in the historical image data as standard input data, and integrate the types of slab surface defects corresponding to the images of the slab surface in the historical image data as standard output data; The artificial intelligence model is trained based on standard input data and standard output data to obtain an image defect recognition model; The real-time collected image data is input into the image defect recognition model to obtain the type of slab surface defect corresponding to the image data and mark it as the recognition result.

8. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The determining whether to send warning information based on the slab production quality assessment coefficient includes: Determine whether the slab production quality assessment coefficient is greater than the preset quality threshold; if yes, generate an early warning message and send it to the client; if not, continue to monitor and judge.

9. The slab defect monitoring system based on the Internet of Things according to claim 1 is characterized in that: The determining whether to send warning information based on the recognition result includes: Extract the types of slab surface defects from the recognition results; Determine whether the type of slab surface defect is normal; if yes, continue to monitor and judge; if not, generate early warning information and send the specific type of slab surface defect to the client.

10. A slab defect monitoring method based on the Internet of Things, applied to a slab defect monitoring system based on the Internet of Things according to any one of claims 1 to 9, characterized in that: include: Step 1: Collect the production data of slabs and the environmental data of slab production workshops; Step 2: Calculate the first indicator coefficient based on the first indicator data; calculate the second indicator coefficient based on the second indicator data; calculate the environmental indicator coefficient based on the environmental data; Step 3: Calculate the slab production quality assessment coefficient based on the first index coefficient, the second index coefficient and the environmental index coefficient; Step 4: Train an artificial intelligence model based on historical image data to obtain an image defect recognition model, and recognize the image data based on the image defect recognition model to obtain a recognition result; Step 5: Determine whether to send an early warning message based on the slab production quality assessment coefficient; Determine whether to send an early warning message based on the recognition result.

Images