An intelligent fault diagnosis system and method for laser equipment based on the Internet of Things

Abstract

The present invention discloses an intelligent fault diagnosis system and method for laser equipment based on the Internet of Things, which relates to the technical field of laser equipment fault diagnosis. The system includes: a laser cutting monitoring module, an infrared image analysis module, and a cutting path adjustment module; the laser cutting monitoring module acquires infrared image data of the laser cutting area in real time and marks the temperature distribution of each pixel point; the infrared image analysis module locates the actual laser cutting point in the infrared image through sliding window temperature integration, predicts the actual laser cutting point in the infrared image to be collected at a future moment, and is also used to analyze the temperature transfer deviation coefficient and radiation heat accumulation coefficient of the laser cutting in the infrared image; the cutting path adjustment module makes an adjustment judgment on the laser cutting path in the real-time infrared image, and corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and radiation heat accumulation coefficient of the laser cutting.

Description

Technical Field

[0001] The present invention relates to the technical field of laser equipment fault diagnosis, and specifically to an intelligent fault diagnosis system and method for laser equipment based on the Internet of Things. Background Art

[0002] With the improvement of the processing accuracy requirements for equipment workpieces in various industries, high-precision laser cutting has become a relatively mainstream technical means in the field of workpiece processing. And the diagnosis and correction of laser cutting path deviation are particularly important in high-precision processing operations.

[0003] During actual laser cutting processing, due to the local thermal conductivity differences of inhomogeneous materials, the heat melting efficiency on both sides of the laser radiation point is inconsistent, resulting in the deviation of the preset laser radiation point from the actual laser cutting seam center. In traditional laser path deviation diagnosis methods, the local thermal conductivity of inhomogeneous materials lacks real-time tracking analysis and real-time correction of the laser cutting path, resulting in redundant loss of processing materials and reduction of processing efficiency during laser cutting.

[0004] Therefore, an intelligent fault diagnosis system and method for laser equipment based on the Internet of Things are needed to solve the above technical defects. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent fault diagnosis system and method for laser equipment based on the Internet of Things to solve the problems raised in the prior art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] An intelligent fault diagnosis method for laser equipment based on the Internet of Things, comprising the following steps:

[0008] Step S100: Real-time monitoring is carried out during the laser cutting task of the laser equipment. Infrared image data of the laser cutting area is obtained in real time through an infrared thermal imaging device, and temperature distribution annotation is performed on each pixel point in the infrared image data.

[0009] Step S200: Slide window integration is performed along the direction parallel to the preset cutting path in the infrared image, the temperature integral of each pixel point in the infrared image is calculated, and then the actual laser cutting point in the infrared image is screened and located. The temperature transfer deviation coefficient of laser cutting in the infrared image is calculated according to the width of the sliding window.

[0010] Step S300: Set the laser cutting temperature calibration value, measure the maximum distance between pixel points with temperature values greater than or equal to the laser cutting temperature calibration value in the direction perpendicular to the preset laser cutting path at the actual laser cutting point in the infrared image, and calculate the radiation heat accumulation coefficient of laser cutting.

[0011] Step S400: Perform attribute annotation on the actual laser cutting points in the historical multi-frame infrared image data, construct a laser cutting data set, and train a laser cutting prediction neural network model to predict the orientation of the actual laser cutting points in the future-frame infrared image data;

[0012] Step S500: Monitor the actual laser cutting points during laser cutting in real time, and control and adjust the laser cutting path according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting in the infrared image.

[0013] In the above technical solution, the following contents are included in the step S100:

[0014] Perform real-time monitoring on the laser cutting material through an infrared thermal imaging device, obtain the infrared image data of the laser cutting area, perform noise suppression and outlier removal on the real-time infrared image data, and perform temperature distribution annotation;

[0015] For any pixel point P, the temperature distribution annotation is: P[x_P, y_P, T_P]; where, x_P is the horizontal arrangement number of the pixel point P in the infrared image data, y_P is the vertical arrangement number of the pixel point P in the infrared image data, and T_P is the temperature displayed by the pixel point P in the infrared image data.

[0016] In the above technical solution, the following analysis steps are included in the step S200:

[0017] Step S201: Set the sliding window width w and length l, and perform sliding window integration along the direction parallel to the preset cutting path in the infrared image with the pixel pitch as the sliding step;

[0018] Step S202: Calculate the sliding window integration results corresponding to each pixel point in the infrared image, and screen the pixel point corresponding to the maximum calculation result as the actual laser cutting point of the current infrared image;

[0019] Among them, if the sliding window integration calculation results of multiple pixel points are equal, the two-dimensional space clustering center of all pixel points with equal calculation results is used as the actual laser cutting point of the current infrared image data;

[0020] Step S203: Obtain the temperature data of the pixel points at an interval of l on both sides of the actual laser cutting point perpendicular to the preset cutting path direction, calculate the temperature transfer deviation coefficient k_grad of the laser cutting in the infrared image, and calculate according to the formula:

[0021] k_grad = (|△T_left| - |△T_right|) / (|△T_left| + |△T_right|);

[0022] Among them, △T_left is the temperature difference between the pixel point at a distance of l to the left of the actual laser cutting point in the direction perpendicular to the preset cutting path and the actual laser cutting point, and △T_right is the temperature difference between the pixel point at a distance of l to the right of the actual laser cutting point in the direction perpendicular to the preset cutting path and the actual laser cutting point;

[0023] By calculating the difference ratio of the temperature distribution of the processed material during laser cutting, analyzing the offset between the actual cutting seam center and the cutting point caused by the local thermal conductivity difference of the material at the laser cutting point, the rationality and scientificity of the laser cutting path adjustment are improved.

[0024] In the above technical solution, the step S300 includes the following analysis steps:

[0025] Step S301: Set the laser cutting temperature calibration value T_cal, and traverse the pixel points along the preset laser cutting path direction at the actual laser cutting point;

[0026] Step S302: Screen all pixel points with temperature values greater than or equal to T_cal, and measure the pixel spacing l_spac between the pixel points farthest apart;

[0027] Step S303: According to the measurement result of l_spac, calculate the radiation heat accumulation coefficient k_heat of laser cutting according to the formula k_heat = k_ad×(l_spac - l_slit) / l_slit;

[0028] Among them, k_ad is the heat accumulation adjustment coefficient, and its value range is (0, 1), and l_slit is the reference cutting seam width of laser cutting;

[0029] By limiting the laser cutting temperature calibration value, calculating the actual transmission range of the thermal radiation effect during laser cutting, and then analyzing the laser energy consumption usage efficiency and material processing effect during the laser cutting process, it is ensured that while meeting the processing effect, the laser extra radiation energy loss is minimized to the greatest extent.

[0030] In the above technical solution, the step S400 includes the following analysis steps:

[0031] Step S401: Obtain the historical multi-frame infrared image data of the laser cutting area, and perform attribute annotation on the screened and located actual laser cutting points;

[0032] For any frame of infrared image data i, the attribute annotation of the actual laser cutting point P_i in the image is: P_i[loc_(i - 1), loc_i, v_P, F_laser];

[0033] In the actual laser cutting point attribute annotation result, loc_(i - 1) is the orientation of the actual laser cutting point P_(i - 1) in the infrared image data of the previous frame before the infrared image data i, loc_i is the orientation of the actual laser cutting point P_i in the infrared image data i, v_P is the travel vector from the actual laser cutting point P_(i - 1) to the actual laser cutting point P_i, and F_laser is the set of laser radiation parameters; among them, the actual laser cutting point orientation includes the horizontal arrangement number and the vertical arrangement number of the actual laser cutting point in the infrared image data;

[0034] Among them, the distance between the actual laser cutting point in the infrared image and the preset cutting path is represented as a signed number; when the value is negative, it means that the actual laser cutting point in the infrared image is on the left side of the preset cutting path, and when the value is positive, it means that the actual laser cutting point in the infrared image is on the right side of the preset cutting path;

[0035] Step S402: Construct a laser cutting data set and train a laser cutting prediction neural network model; the input of the laser cutting prediction neural network model is the actual laser cutting point attribute annotation result in the current frame infrared image, and the output is the predicted orientation of the actual laser cutting point in the next frame infrared image.

[0036] In the above technical solution, the step S500 includes the following contents:

[0037] Monitor the actual laser cutting point during laser cutting in real time, use the laser cutting prediction neural network model to predict the orientation of the actual laser cutting point in the future infrared image frame, and perform laser cutting path deviation fault diagnosis;

[0038] Set a laser path deviation threshold and a radiation heat accumulation coefficient threshold, and construct a path deviation fault diagnosis paradigm as follows: (1) The distance between the predicted orientation of the actual laser cutting point and the preset laser cutting path is less than the laser path deviation threshold; (2) The radiation heat accumulation coefficient of laser cutting at any time point is less than the radiation heat accumulation coefficient threshold;

[0039] When any one of the above paradigms is not satisfied, it is determined that the current laser cutting path needs to be adjusted. For the laser cutting path at any moment, the adjustment strategy is as follows:

[0040] v_l_fix=(1 + k_heat)×v_l_ini;

[0041] v_t_fix=k_grad×dev_P / T0;

[0042] Among them, \(v_{l\_fix}\) is the correction value of the laser head moving speed in the direction parallel to the preset laser cutting path, \(k_{heat}\) is the radiation heat accumulation coefficient of laser cutting in the infrared image, \(v_{l\_ini}\) is the initial value of the laser head moving speed in the direction parallel to the preset laser cutting path, \(v_{t\_fix}\) is the correction value of the laser head moving speed in the direction perpendicular to the preset laser cutting path, \(k_{grad}\) is the temperature transfer deviation coefficient of laser cutting in the infrared image, \(dev_P\) is the distance between the actual laser cutting point and the preset cutting path in the infrared image, and \(T_0\) is the infrared image acquisition period;

[0043] Taking the preset laser cutting path direction as the reference direction, the correction values of the laser head moving speeds parallel and perpendicular to the reference direction are vectorially added as the laser head moving speed vector correction result at the current moment;

[0044] The path deviation is decomposed into speed correction in the parallel direction and fine adjustment in the perpendicular direction to avoid oscillations caused by single parameter adjustment. At the same time, according to the temperature transfer deviation coefficient and radiation heat accumulation coefficient of laser cutting, the laser cutting path is adaptively adjusted, further improving the energy utilization efficiency of the processing process and the workpiece processing effect.

[0045] For the correction value of the laser head moving speed in the direction perpendicular to the reference direction, if the value is positive, the laser head moving speed direction is to the right perpendicular to the reference direction; if the value is negative, the laser head moving speed direction is to the left perpendicular to the reference direction.

[0046] For the laser head moving speed vector correction result, if the vector modulus value is less than or equal to the maximum moving speed of the laser head, the laser head moving speed vector correction result is used as the laser head moving speed vector at the current moment, and the laser cutting path adjustment strategy decision is re-made at the next infrared image acquisition time point; if the vector modulus value is greater than the maximum moving speed of the laser head, the laser head moving speed vector correction result is proportionally reduced until the vector modulus value is equal to the maximum moving speed of the laser head, and the laser cutting path adjustment strategy decision is re-made at the next infrared image acquisition time point.

[0047] An intelligent diagnosis system for laser equipment faults based on the Internet of Things in the above technical solution, the system includes: a laser cutting monitoring module, an infrared image analysis module, and a cutting path adjustment module;

[0048] The laser cutting monitoring module obtains the infrared image data of the laser cutting area in real time and marks the temperature distribution of each pixel point; the infrared image analysis module locates the actual laser cutting point in the infrared image through sliding window temperature integration, predicts the actual laser cutting point in the infrared image to be collected at a future moment, and is also used to analyze the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting in the infrared image; the cutting path adjustment module makes an adjustment judgment on the laser cutting path in the real-time infrared image, and corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting.

[0049] In the above technical solution, the laser cutting monitoring module includes: an infrared data monitoring unit and a temperature distribution marking unit;

[0050] The infrared data monitoring unit obtains the infrared image data of the laser cutting area in real time through an infrared thermal imaging device; the temperature distribution marking unit marks the temperature distribution of each pixel point in the infrared image.

[0051] In the above technical solution, the infrared image analysis module includes: an actual laser cutting point positioning unit, a temperature transfer deviation analysis unit, a radiation heat accumulation analysis unit, and an actual laser cutting point prediction unit;

[0052] The actual laser cutting point positioning unit locates the actual laser cutting point in the infrared image through sliding window temperature integration; the temperature transfer deviation analysis unit is used to analyze the temperature transfer deviation coefficient of the laser cutting in the infrared image; the radiation heat accumulation analysis unit is used to analyze the radiation heat accumulation coefficient of the laser cutting in the infrared image; the actual laser cutting point prediction unit predicts the orientation of the actual laser cutting point in the infrared image to be collected at a future moment according to the actual laser cutting point data in the historical infrared data.

[0053] In the above technical solution, the cutting path adjustment module includes: a cutting path adjustment judgment unit and a laser head movement speed correction unit;

[0054] The cutting path adjustment judgment unit makes an adjustment judgment on the laser cutting path in the real-time infrared image; the laser head movement speed correction unit corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting.

[0055] Compared with the prior art, the beneficial effects of the present invention are:

[0056] In the present invention, the method of real-time infrared image monitoring is adopted to obtain the material temperature data in the laser cutting process, which can effectively avoid the increase of the image feature misjudgment rate caused by smoke and dust interference and material reflection in the vision scheme, and can also diagnose and correct the deviation of the laser cutting path more real-time.

[0057] In the present invention, infrared abnormal isolated points caused by material residues are removed through image outlier processing and noise suppression, and the center point of the laser cutting seam is located. The deviation between the laser radiation point and the center of the molten pool caused by the difference in local thermal conductivity of heterogeneous materials is fully considered. Compared with the acoustic vibration scheme, the accuracy and intuitiveness of identifying and locating the actual laser cutting point are significantly improved.

[0058] In the present invention, by analyzing the temperature transfer deviation coefficient and the cumulative coefficient of radiation heat in the infrared image data during laser cutting in real time, compared with the pure data-driven black box model, the interpretability of diagnosing and correcting the deviation of the laser cutting path is ensured, and at the same time, the adaptability of the system to complex working conditions is improved, and the utilization rate of the processing material and laser energy consumption during laser cutting is significantly increased. Description of the Drawings

[0059] Figure 1 It is a flowchart of an intelligent fault diagnosis method for a laser device based on the Internet of Things according to the present invention;

[0060] Figure 2 It is an organizational structure diagram of an intelligent fault diagnosis system for a laser device based on the Internet of Things according to the present invention. Detailed Embodiments

[0061] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] Embodiment: Please refer to Figure 1 - Figure 2 The present invention provides the following technical solutions:

[0063] As Figure 1 shown, the present invention provides an intelligent fault diagnosis method for a laser device based on the Internet of Things, including the following steps:

[0064] Step S100: Real-time monitoring is carried out during the laser cutting task of the laser device. Infrared image data of the laser cutting area is obtained in real time through an infrared thermal imaging device, and temperature distribution annotation is performed on each pixel point in the infrared image data;

[0065] Step S200: Perform sliding window integration along the direction parallel to the preset cutting path in the infrared image, calculate the temperature integral of each pixel point in the infrared image, and then screen and locate the actual laser cutting point in the infrared image. Calculate the temperature transfer deviation coefficient of the laser cutting in the infrared image according to the width of the sliding window;

[0066] Step S300: Set the calibration value of the laser cutting temperature, measure the maximum distance between pixel points with temperature values greater than or equal to the laser cutting temperature calibration value in the direction perpendicular to the preset laser cutting path at the actual laser cutting point in the infrared image, and calculate the radiation heat accumulation coefficient of the laser cutting;

[0067] Step S400: Perform attribute annotation on the actual laser cutting points in the historical multi-frame infrared image data, construct a laser cutting data set and train a laser cutting prediction neural network model to predict the orientation of the actual laser cutting points in the future frame infrared image data;

[0068] Step S500: Monitor the actual laser cutting point during laser cutting in real time, and control and adjust the laser cutting path according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting in the infrared image.

[0069] The following contents are included in the step S100:

[0070] Perform real-time monitoring on the laser cutting material through an infrared thermal imaging device, obtain the infrared image data of the laser cutting area, suppress noise and remove outliers from the real-time infrared image data, and perform temperature distribution annotation;

[0071] For any pixel point P, the temperature distribution annotation is: P[x_P, y_P, T_P]; where, x_P is the horizontal arrangement number of the pixel point P in the infrared image data, y_P is the vertical arrangement number of the pixel point P in the infrared image data, and T_P is the temperature displayed by the pixel point P in the infrared image data;

[0072] In specific implementation, a high-dynamic high-frame-rate infrared thermal imager is selected for real-time monitoring above the side of the laser cutting head. At the same time, a filter is used to suppress the strong light interference directly irradiated by the plasma strong light. After installing the filter, since the filter calibrates the image of the infrared thermal imager again, a pixel temperature-real temperature mapping table is constructed. Further, image local repair is used to repair the overexposure of the image in the plasma flash area, and temperature distribution annotation is performed on the repaired image according to the pixel temperature-real temperature mapping table.

[0073] The following analysis steps are included in the step S200:

[0074] Step S201: Set the sliding window width w and length l, and perform sliding window integration along the direction parallel to the preset cutting path in the infrared image with the pixel pitch as the sliding step;

[0075] Step S202: Calculate the sliding window integration results corresponding to each pixel point in the infrared image, and select the pixel point corresponding to the maximum calculation result as the actual laser cutting point of the current infrared image;

[0076] Among them, if the sliding window integration calculation results of multiple pixel points are equal, the two-dimensional spatial clustering centers of all pixel points with equal calculation results are used as the actual laser cutting points of the current infrared image data;

[0077] Step S203: Obtain the temperature data of pixel points at an interval of l on both sides of the actual laser cutting point perpendicular to the preset cutting path direction, and calculate the temperature transfer deviation coefficient k_grad of the laser cutting in the infrared image, and calculate according to the formula:

[0078] k_grad = (|△T_left| - |△T_right|) / (|△T_left| + |△T_right|);

[0079] Among them, △T_left is the temperature difference between the pixel point at an interval of l on the left side of the actual laser cutting point perpendicular to the preset cutting path direction and the actual laser cutting point, and △T_right is the temperature difference between the pixel point at an interval of l on the right side of the actual laser cutting point perpendicular to the preset cutting path direction and the actual laser cutting point;

[0080] In specific implementation, due to the possible local uniformity differences in the internal structure of the material to be processed, during laser cutting, the heat transfer efficiencies in different regions near different laser radiation points are also inconsistent. Inevitably, there is a phenomenon that some regions in the preset laser cutting path have melted and fallen off, while some regions have not melted, resulting in a deviation between the final cut seam and the preset cutting path;

[0081] Furthermore, to refine the evaluation of the heat transfer characteristics of each region during laser cutting, the infrared image method is used to compare and analyze the heat radiation efficiencies on both sides of the laser radiation point of the material to be processed, and then determine the azimuth of the center of the final cut seam, and use this to evaluate whether the laser cutting path has a lateral deviation;

[0082] Furthermore, set the sliding window l = w = 3px, corresponding to an actual distance of 0.3mm, and the sliding step is 1px, corresponding to an actual distance of 0.1mm; when performing sliding window integration, to simplify the operation, a region with a width of 3cm on each side of the preset cutting path is used as the sliding region;

[0083] Assume that the sliding window of pixel point p1 at an interval of 3px on the left side of the located actual laser cutting point contains pixel point temperature values (unit: degree Celsius) of [230, 250, 220; 1400, 1550, 1450; 1350, 1480, 1520], and the sliding window of pixel point p2 at an interval of 3px on the right side contains pixel point temperature values of [1480, 1520, 160; 1550, 1450, 170; 250, 220, 175]. Calculate that the temperature integral value of pixel point p1 is 10400, and the temperature integral value of pixel point p2 is 7065;

[0084] Assume that the actual temperature value at the laser cutting point is 1600. At this time, it can be calculated that the temperature transfer deviation coefficient k_grad of laser cutting is k_grad = (|1600 - 1550| - |1600 - 1450|) / (|1600 - 1550| + |1600 - 1450|) = -0.5.

[0085] The step S300 includes the following analysis steps:

[0086] Step S301: Set the laser cutting temperature calibration value T_cal. At the actual laser cutting point, traverse the pixel points along the preset laser cutting path direction;

[0087] Step S302: Screen all pixel points whose temperature values are greater than or equal to T_cal, and measure the pixel spacing l_spac between the pixel points farthest apart;

[0088] Step S303: According to the measurement result of l_spac, calculate the radiation heat accumulation coefficient k_heat of laser cutting according to the formula k_heat = k_ad×(l_spac - l_slit) / l_slit;

[0089] Wherein, k_ad is the heat accumulation adjustment coefficient, and its value range is (0, 1), and l_slit is the reference slit width of laser cutting;

[0090] In specific implementation, the internal structure and physical and chemical properties of the material to be processed in the area near the laser radiation point may change with the laser heat radiation. Due to the inconsistency of the local heat transfer efficiency of the material, the material ranges and orientations affected by the same heat radiation intensity in different areas are also different, resulting in uneven material ranges affected near the laser slit finally, which affects the application of the final processed product;

[0091] Furthermore, to measure the interference of laser cutting heat radiation on the material, the area with higher heat radiation near the actual laser slit is divided by an infrared image to intuitively reflect the area greatly affected by heat radiation during the laser cutting process. Then, according to the area range division results in the infrared images at different time points, the actual influence range of the laser radiation point in different areas of the material to be processed is quantified;

[0092] Further, assume that the temperature values (unit: degree Celsius) of a total of 9 pixel points on both the left and right sides of the current actual laser cutting point are, from left to right in sequence: 1100, 1250, 1400, 1550, 1600, 1600, 1500, 1400, 1300; set the laser cutting temperature calibration value to 1300, measure and obtain l_spac = 6px (0.6mm), assume the laser cutting reference slit width l_slit = 0.4mm, and the heat accumulation adjustment coefficient k_ad = 0.5, calculate the radiation heat accumulation coefficient k_heat of laser cutting = 0.25.

[0093] Step S400 includes the following analysis steps:

[0094] Step S401: Obtain historical multi-frame infrared image data of the laser cutting area, and perform attribute annotation on the actually located laser cutting point.

[0095] For any frame of infrared image data i, the attribute annotation of the actual laser cutting point P_i in the image is: P_i[loc_(i - 1), loc_i, v_P, F_laser];

[0096] In the actual laser cutting point attribute annotation result, loc_(i - 1) is the orientation of the actual laser cutting point P_(i - 1) in the infrared image data of the previous frame of infrared image data i, loc_i is the orientation of the actual laser cutting point P_i in the infrared image data i, v_P is the travel vector from the actual laser cutting point P_(i - 1) to the actual laser cutting point P_i, and F_laser is the set of laser radiation parameters; among them, the actual laser cutting point orientation includes the horizontal arrangement number and the vertical arrangement number of the actual laser cutting point in the infrared image data.

[0097] Among them, the distance between the actual laser cutting point in the infrared image and the preset cutting path is expressed as a signed number; when the value is negative, it means that the actual laser cutting point in the infrared image is on the left side of the preset cutting path, and when the value is positive, it means that the actual laser cutting point in the infrared image is on the right side of the preset cutting path.

[0098] Step S402: Construct a laser cutting data set and train a laser cutting prediction neural network model; the input of the laser cutting prediction neural network model is the actual laser cutting point attribute annotation result in the current frame of infrared image, and the output is the predicted orientation of the actual laser cutting point in the next frame of infrared image.

[0099] Step S500 includes the following content:

[0100] Monitor the actual laser cutting point during laser cutting in real time, use the laser cutting prediction neural network model to predict the orientation of the actual laser cutting point in the future infrared image frame, and perform laser cutting path deviation fault diagnosis.

[0101] Set a laser path deviation threshold and a radiation heat accumulation coefficient threshold, and construct a path deviation fault diagnosis paradigm as follows: (1) The distance between the predicted azimuth of the actual laser cutting point and the preset laser cutting path is less than the laser path deviation threshold; (2) The radiation heat accumulation coefficient of laser cutting at any time point is less than the radiation heat accumulation coefficient threshold;

[0102] In specific implementation, when the cutting line deviates greatly from the preset cutting path, secondary processing of the cutting edge line is required due to precision requirements during subsequent processing, so secondary processing losses will occur. At the same time, if the thermal radiation generated by the laser cutting on the material around the cutting line is strong, it will cause damage to the internal structure of the material and excessive use of laser energy consumption. Therefore, it is necessary to judge the path deviation and thermal radiation accumulation during laser cutting processing;

[0103] When any one of the above paradigms is not satisfied, it is determined that the current laser cutting path needs to be adjusted. For the laser cutting path at any moment, the adjustment strategy is as follows:

[0104] v_l_fix=(1 + k_heat)×v_l_ini;

[0105] v_t_fix=k_grad×dev_P / T0;

[0106] Where, v_l_fix is the correction value of the moving speed of the laser head in the direction parallel to the preset laser cutting path, k_heat is the radiation heat accumulation coefficient of laser cutting in the infrared image, v_l_ini is the initial value of the moving speed of the laser head in the direction parallel to the preset laser cutting path, v_t_fix is the correction value of the moving speed of the laser head in the direction perpendicular to the preset laser cutting path, k_grad is the temperature transfer deviation coefficient of laser cutting in the infrared image, dev_P is the distance between the actual laser cutting point and the preset cutting path in the infrared image, and T0 is the infrared image acquisition period;

[0107] In specific implementation, when the laser cut seam deviates greatly from the preset laser cutting path, negative feedback adjustment of the lateral movement of the laser head is required to ensure that the laser cut seam fits the preset laser cutting path as much as possible. At the same time, when the thermal radiation generated by the laser radiation point in the local area of the material to be processed has a strong influence, it is necessary to appropriately increase the longitudinal movement speed of the laser head to suppress the further influence of the laser cutting thermal radiation on the material to be processed near the laser cutting point and ensure the local quality consistency of the final processed product.

[0108] Further, assume that the initial moving speed of the laser head parallel to the preset laser cutting path direction is \(v_{l\_ini} = 5\mathrm{mm / s}\), the temperature transfer deviation coefficient: \(k_{grad}=-0.5\), the radiation heat accumulation coefficient: \(k_{heat}=0.25\), the path deviation distance: \(dev_P = +0.2\mathrm{mm}\), the infrared image acquisition period: \(T_0 = 0.01\mathrm{s}\), and the maximum moving speed of the laser head: \(v_{max}=20\mathrm{mm / s}\);

[0109] Calculate the correction value of the moving speed of the laser head parallel to the preset laser cutting path direction \(v_{l\_fix}=6.25\mathrm{mm / s}\), and the correction value of the moving speed of the laser head perpendicular to the preset laser cutting path direction \(v_{t\_fix}=-10\mathrm{mm / s}\), with the direction to the left. Further, vectorially superimpose the two-direction speed correction values with orthogonal vectors to synthesize the speed vector correction value of the laser head movement, and the modulus is approximately \(11.79\mathrm{mm / s}\). After comparison, no scale scaling is required.

[0110] Taking the preset laser cutting path direction as the reference direction, use vector addition of the correction values of the moving speed of the laser head parallel and perpendicular to the reference direction as the result of the speed vector correction of the laser head movement at the current moment;

[0111] For the correction value of the moving speed of the laser head perpendicular to the reference direction, if the value is positive, the moving speed direction of the laser head is perpendicular to the reference direction and to the right; if the value is negative, the moving speed direction of the laser head is perpendicular to the reference direction and to the left.

[0112] For the result of the speed vector correction of the laser head movement, if the vector modulus value is less than or equal to the maximum moving speed of the laser head, use the result of the speed vector correction of the laser head movement as the speed vector of the laser head at the current moment, and re-make the decision on the laser cutting path adjustment strategy at the next infrared image acquisition time point; if the vector modulus value is greater than the maximum moving speed of the laser head, scale down the result of the speed vector correction of the laser head movement until the vector modulus value is equal to the maximum moving speed of the laser head, and re-make the decision on the laser cutting path adjustment strategy at the next infrared image acquisition time point.

[0113] As Figure 2 shown, the present invention also provides an intelligent fault diagnosis system for a laser device based on the Internet of Things. The system includes: a laser cutting monitoring module, an infrared image analysis module, and a cutting path adjustment module;

[0114] The laser cutting monitoring module obtains the infrared image data of the laser cutting area in real time and marks the temperature distribution of each pixel point; the infrared image analysis module locates the actual laser cutting point in the infrared image through sliding window temperature integration, predicts the actual laser cutting point in the infrared image collected at a future moment, and is also used to analyze the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting in the infrared image; the cutting path adjustment module makes an adjustment judgment on the laser cutting path in the real-time infrared image, and corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting.

[0115] The laser cutting monitoring module includes: an infrared data monitoring unit and a temperature distribution marking unit;

[0116] The infrared data monitoring unit obtains the infrared image data of the laser cutting area in real time through an infrared thermal imaging device; the temperature distribution marking unit marks the temperature distribution of each pixel point in the infrared image.

[0117] The infrared image analysis module includes: an actual laser cutting point positioning unit, a temperature transfer deviation analysis unit, a radiation heat accumulation analysis unit, and an actual laser cutting point prediction unit;

[0118] The actual laser cutting point positioning unit locates the actual laser cutting point in the infrared image through sliding window temperature integration; the temperature transfer deviation analysis unit is used to analyze the temperature transfer deviation coefficient of the laser cutting in the infrared image; the radiation heat accumulation analysis unit is used to analyze the radiation heat accumulation coefficient of the laser cutting in the infrared image; the actual laser cutting point prediction unit predicts the orientation of the actual laser cutting point in the infrared image collected at a future moment according to the actual laser cutting point data in the historical infrared data.

[0119] The cutting path adjustment module includes: a cutting path adjustment judgment unit and a laser head movement speed correction unit;

[0120] The cutting path adjustment judgment unit makes an adjustment judgment on the laser cutting path in the real-time infrared image; the laser head movement speed correction unit corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting.

[0121] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. An intelligent fault diagnosis method for laser devices based on the Internet of Things, characterized in that , The method includes the following steps: Step S100: During the laser cutting task of the laser device, perform real-time monitoring, obtain the infrared image data of the laser cutting area in real time through the infrared thermal imaging device, and label the temperature distribution of each pixel point in the infrared image data; Step S200: Perform sliding window integration along the direction parallel to the preset cutting path in the infrared image, calculate the temperature integral of each pixel point in the infrared image, and then screen and locate the actual laser cutting point in the infrared image. Calculate the temperature transfer deviation coefficient of the laser cutting in the infrared image according to the sliding window width; Step S300: Set the laser cutting temperature calibration value, measure the maximum distance between pixel points with temperature values greater than or equal to the laser cutting temperature calibration value in the direction perpendicular to the preset laser cutting path at the actual laser cutting point in the infrared image, and calculate the radiation heat accumulation coefficient of the laser cutting; Step S400: Label the attributes of the actual laser cutting points in the historical multi-frame infrared image data, construct a laser cutting data set and train a laser cutting prediction neural network model to predict the orientation of the actual laser cutting points in the future frame infrared image data; Step S500: Real-time monitor the actual laser cutting point during laser cutting, and control and adjust the laser cutting path according to the temperature transfer deviation coefficient and radiation heat accumulation coefficient of the laser cutting in the infrared image.

2. The intelligent fault diagnosis method for a laser device based on the Internet of Things according to claim 1, wherein, The following content is included in the step S100: Perform real-time monitoring of the laser cutting material through the infrared thermal imaging device, obtain the infrared image data of the laser cutting area, suppress noise and remove outliers from the real-time infrared image data, and perform temperature distribution labeling; For any pixel point P, the temperature distribution labeling is: P[x_P, y_P, T_P]; where, x_P is the horizontal arrangement number of the pixel point P in the infrared image data, y_P is the vertical arrangement number of the pixel point P in the infrared image data, and T_P is the temperature displayed by the pixel point P in the infrared image data.

3. The intelligent fault diagnosis method for a laser device based on the Internet of Things according to claim 1, characterized in that, The following analysis steps are included in the step S200: Step S201: Set the sliding window width w and length l, and perform sliding window integration along the direction parallel to the preset cutting path in the infrared image with the pixel spacing as the sliding step; Step S202: Calculate the sliding window integration result corresponding to each pixel point in the infrared image, and screen the pixel point corresponding to the maximum calculation result as the actual laser cutting point of the current infrared image; Among them, if there are multiple pixel points with equal sliding window integration calculation results, the two-dimensional space clustering center of all pixel points with equal calculation results is used as the actual laser cutting point of the current infrared image data; Step S203: Obtain the temperature data of pixel points at intervals of l on both sides of the actual laser cutting point in the direction perpendicular to the preset cutting path, calculate the temperature transfer deviation coefficient k_grad of the laser cutting in the infrared image, and calculate according to the formula: k_grad = (|△T_left| - |△T_right|) / (|△T_left| + |△T_right|); Among them, △T_left is the temperature difference between the pixel point at a distance of l to the left of the actual laser cutting point in the direction perpendicular to the preset cutting path and the actual laser cutting point, and △T_right is the temperature difference between the pixel point at a distance of l to the right of the actual laser cutting point in the direction perpendicular to the preset cutting path and the actual laser cutting point.

4. The intelligent fault diagnosis method of a laser device based on the Internet of Things according to claim 1, wherein, The following analysis steps are included in the step S300: Step S301: Set the laser cutting temperature calibration value T_cal, and traverse the pixel points along the preset laser cutting path direction at the actual laser cutting point. Step S302: Screen all pixel points with temperature values greater than or equal to T_cal, and measure the pixel spacing l_spac between the pixel points with the farthest distance. Step S303: According to the measurement result of l_spac, calculate the radiation heat accumulation coefficient k_heat of laser cutting according to the formula k_heat = k_ad×(l_spac - l_slit) / l_slit. Among them, k_ad is the heat accumulation adjustment coefficient, and its value range is (0,1), and l_slit is the reference slit width of laser cutting.

5. The intelligent fault diagnosis method for a laser device based on the Internet of Things according to claim 1, characterized in that, The following analysis steps are included in the step S400: Step S401: Obtain the historical multi-frame infrared image data of the laser cutting area, and perform attribute annotation on the screened and located actual laser cutting points. For any frame of infrared image data i, the attribute annotation of the actual laser cutting point P_i in the image is: P_i[loc_(i - 1), loc_i, v_P, F_laser]; In the actual laser cutting point attribute annotation result, loc_(i - 1) is the azimuth of the actual laser cutting point P_(i - 1) in the previous frame of infrared image data of the infrared image data i, loc_i is the azimuth of the actual laser cutting point P_i in the infrared image data i, v_P is the travel vector from the actual laser cutting point P_(i - 1) to the actual laser cutting point P_i, and F_laser is a set of laser radiation parameters; among them, the actual laser cutting point azimuth includes the horizontal arrangement number and the vertical arrangement number of the actual laser cutting point in the infrared image data. Among them, the distance between the actual laser cutting point in the infrared image and the preset cutting path is expressed as a signed number; when the value is negative, it means that the actual laser cutting point in the infrared image is on the left side of the preset cutting path, and when the value is positive, it means that the actual laser cutting point in the infrared image is on the right side of the preset cutting path. Step S402: Construct a laser cutting data set and train a laser cutting prediction neural network model; the input of the laser cutting prediction neural network model is the actual laser cutting point attribute annotation result in the current frame of infrared image, and the output is the predicted azimuth of the actual laser cutting point in the next frame of infrared image.

6. The intelligent fault diagnosis method for a laser device based on the Internet of Things according to claim 1, characterized in that, The following content is included in the step S500: Monitor the actual laser cutting point during laser cutting in real time, use the laser cutting prediction neural network model to predict the azimuth of the actual laser cutting point in the future infrared image frame, and perform laser cutting path deviation fault diagnosis. Set the laser path deviation threshold and the radiation heat accumulation coefficient threshold, and construct the path deviation fault diagnosis paradigm as follows: (1) The distance between the predicted azimuth of the actual laser cutting point and the preset laser cutting path is less than the laser path deviation threshold; (2) The radiation heat accumulation coefficient of laser cutting at any time point is less than the radiation heat accumulation coefficient threshold; When any one of the above paradigms is not satisfied, it is determined that the current laser cutting path needs to be adjusted. For the laser cutting path at any moment, the adjustment strategy is as follows: v_l_fix=(1+k_heat)×v_l_ini; v_t_fix=k_grad×dev_P / T0; Where, v_l_fix is the correction value of the moving speed of the laser head in the direction parallel to the preset laser cutting path, k_heat is the radiation heat accumulation coefficient of laser cutting in the infrared image, v_l_ini is the initial value of the moving speed of the laser head in the direction parallel to the preset laser cutting path, v_t_fix is the correction value of the moving speed of the laser head in the direction perpendicular to the preset laser cutting path, k_grad is the temperature transfer deviation coefficient of laser cutting in the infrared image, dev_P is the distance between the actual laser cutting point and the preset cutting path in the infrared image, and T0 is the infrared image acquisition period; Taking the preset laser cutting path direction as the reference direction, the correction values of the moving speeds of the laser head parallel and perpendicular to the reference direction are added vectorially as the correction result of the moving speed vector of the laser head at the current moment; For the correction value of the moving speed of the laser head perpendicular to the reference direction, if the value is positive, the moving speed direction of the laser head is to the right perpendicular to the reference direction, and if the value is negative, the moving speed direction of the laser head is to the left perpendicular to the reference direction; For the correction result of the moving speed vector of the laser head, if the vector modulus value is less than or equal to the maximum moving speed of the laser head, the correction result of the moving speed vector of the laser head is used as the moving speed vector of the laser head at the current moment, and the laser cutting path adjustment strategy decision is re-made at the next infrared image acquisition time point; if the vector modulus value is greater than the maximum moving speed of the laser head, the correction result of the moving speed vector of the laser head is proportionally reduced until the vector modulus value is equal to the maximum moving speed of the laser head, and the laser cutting path adjustment strategy decision is re-made at the next infrared image acquisition time point.

7. An Internet of Things-based intelligent laser device fault diagnosis system applying the Internet of Things-based intelligent laser device fault diagnosis method according to any one of claims 1-6, characterized in that, The system includes: a laser cutting monitoring module, an infrared image analysis module, and a cutting path adjustment module; The laser cutting monitoring module obtains the infrared image data of the laser cutting area in real time and marks the temperature distribution of each pixel point; the infrared image analysis module locates the actual laser cutting point in the infrared image through sliding window temperature integration and predicts the actual laser cutting point in the infrared image to be collected at a future moment, and is also used to analyze the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of laser cutting in the infrared image; the cutting path adjustment module judges and adjusts the laser cutting path in the real-time infrared image, and corrects the moving speed vector of the laser head according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of laser cutting.

8. An intelligent fault diagnosis system for a laser device based on the Internet of Things according to claim 7, characterized in that, The laser cutting monitoring module includes: an infrared data monitoring unit and a temperature distribution marking unit; The infrared data monitoring unit obtains the infrared image data of the laser cutting area in real time through an infrared thermal imaging device; the temperature distribution annotation unit annotates the temperature distribution of each pixel point in the infrared image.

9. An intelligent fault diagnosis system for a laser device based on the Internet of Things according to claim 7, characterized in that, The infrared image analysis module includes: an actual laser cutting point positioning unit, a temperature transfer deviation analysis unit, a radiation heat accumulation analysis unit, and an actual laser cutting point prediction unit; The actual laser cutting point positioning unit locates the actual laser cutting point in the infrared image through sliding window temperature integration; the temperature transfer deviation analysis unit is used to analyze the temperature transfer deviation coefficient of the laser cutting in the infrared image; the radiation heat accumulation analysis unit is used to analyze the radiation heat accumulation coefficient of the laser cutting in the infrared image; the actual laser cutting point prediction unit predicts the orientation of the actual laser cutting point in the infrared image collected at a future moment according to the actual laser cutting point data in the historical infrared data.

10. An intelligent fault diagnosis system for a laser device based on the Internet of Things according to claim 7, characterized in that, The cutting path adjustment module includes: a cutting path adjustment judgment unit and a laser head movement speed correction unit; The cutting path adjustment judgment unit makes an adjustment judgment on the laser cutting path in the real-time infrared image; the laser head movement speed correction unit corrects the laser head movement speed vector according to the temperature transfer deviation coefficient and the radiation heat accumulation coefficient of the laser cutting.

Images